Covering America - Fighting For Our Health

Covering
America
R E A L R EM E D I ES
FOR THE UNINSURED
  
   
      
 
Jack A. Meyer
Project Director
Elliot K. Wicks
Editor and
Project Manager
Advisory Panel
David M. Carlisle
Director, California Office of Statewide Health Planning
and Development
Lynn M. Etheredge
Independent Consultant
Judith Feder
Professor and Dean of Policy Studies
Georgetown University
Christine C. Ferguson
Director, Rhode Island Department of Human Services
Jacob S. Hacker
Junior Fellow
Harvard University Society of Fellows
Robert B. Helms
Resident Scholar
American Enterprise Institute
Larry Levitt
Vice President
The Henry J. Kaiser Family Foundation
Len M. Nichols
Principal Research Associate
The Urban Institute
Mark V. Pauly
Chair, Department of Health Care Systems
The Wharton School
University of Pennsylvania
Alice M. Rivlin
Senior Fellow and Johnson Chair
The Brookings Institution
Cathy Schoen
Vice President, Research and Evaluation
The Commonwealth Fund
Gail R.Wilensky
The John M. Olin Senior Fellow
Project Hope
Chair, Medicare Payment Advisory Commission
Covering America
R E A L R E M E D I E S F O R TH E U N I N S U R E D
Jack A. Meyer
Project Director
Elliot K. Wicks
Editor and Project Manager
  
  
      
June 2001
Project Director
Economic and Social Research Institute (ESRI)
Jack A. Meyer
Editor and Project Manager, ESRI
Elliot K. Wicks
ESRI Project Staff
Sharon Silow-Carroll
Emily Waldman
Patricia Gepert-Guajardo
Stephanie Anthony
Todd Kutyla
Mark Legnini
Tarneice Hinton-Davis
Copyright  by the Economic and Social
Research Institute, Washington, D.C. All rights
reserved. No part of this publication may be used
or reproduced in any manner without permission
in writing from the Economic and Social Research
Institute, except in the case of brief quotations
embodied in news articles, critical articles, or
reviews.
Economic and Social Research Institute
 th Street, N.W., Suite 
Washington, D.C. 
() -
Website: www.esresearch.org
Copy Editor
Jean Bernard
Economic and Social Research Institute
Board of Directors
Cover and Book Design
Beth Schlenoff
Mickey Levy
Chief Economist, Bank of America Corporation
This project was made possible by a grant from
The Robert Wood Johnson Foundation, Princeton,
New Jersey.
Single volumes of this document can be ordered at
no cost by logging on to our website: www.esresearch.org. The document is also available in PDF
format from our website.
William Lilley III
President and CEO, Policy Communications Inc.
Kenneth McLennan
Retired President and Senior Consultant,
Manufacturers’ Alliance for Productivity and
Innovation
Jack A. Meyer
President, Economic and Social Research Institute
Beth Shulman
Consultant and former Vice President,
United Food and Commercial Workers Union
Sheila Zedlewski
Program Director, The Urban Institute
3
Table of Contents
4
Acknowledgements
119
Kronick and Rice Proposal: Key Elements
5
Preface
121
6
A State-Based Proposal for Achieving
Universal Coverage
Overview
13
A Comparison of Reform Plan Features
21
Butler Proposal: Key Elements
23
Reforming the Tax Treatment of
Health Care to Achieve Universal
Coverage
43
45
by Richard Kronick and Thomas Rice
135
Pauly Proposal: Key Elements
137
An Adaptive Credit Plan for Covering
the Uninsured
by Mark V. Pauly
by Stuart M. Butler
153
Singer, Garber, and Enthoven Proposal:
Key Elements
Feder, Levitt, O’Brien, and Rowland
Approach: Key Elements
155
Near-Universal Coverage Through
Health Plan Competition
   
Assessing the Combination of
Public Programs and Tax Credits
by Sara J. Singer, Alan M. Garber, and
Alain C. Enthoven
by Judith Feder, Larry Levitt, Ellen O’Brien,
and Diane Rowland
173
Weil Proposal: Key Elements
175
A Private/Public Partnership for
National Health Insurance
The Medical Security System:
A Proposal to Ensure Health Insurance
Coverage for All Americans
by Jonathan Gruber
by Alan R. Weil
57
Gruber Proposal: Key Elements
59
73
Hacker Proposal: Key Elements
193
75
Medicare Plus: Increasing Health Coverage
by Expanding Medicare
Wicks, Meyer, and Silow-Carroll Proposal:
Key Elements
195
A Plan for Achieving Universal Health
Coverage
      
 
by Jacob S. Hacker
101
103
Holahan, Nichols, and Blumberg Proposal:
Key Elements
Expanding Health Insurance Coverage
  / 
by John F. Holahan, Len M. Nichols, and
Linda J. Blumberg
by Elliot K. Wicks, Jack A. Meyer, and
Sharon Silow-Carroll
4
Acknowledgements
The Economic and Social Research Institute thanks
the following people and groups for their valuable
contributions to this project: our Advisory Panel for
guidance in the design of the project and its many
elements,as well as the critiques of the authors’ proposals; The Robert Wood Johnson Foundation, a
national philanthropy devoted exclusively to health
and health care, and especially Karen Davenport,
David Colby, and Linda Bilheimer for their oversight and valuable advice; our authors for their
thoughtful analyses and their cooperation and timeliness in supplying numerous drafts and responding
to requests for various kinds of inform a ti on ;
Burness Communications for help in publicizing
the work of the project; Jean Bernard for editing
assistance; Beth Schlenoff for cover and book design;
and ESRI staff—Sharon Silow-Carroll, Emily
Waldman, Patricia Gepert-Guajardo, Stephanie
Anthony, Todd Kutyla, Ta rn ei ce Hinton-Davis,
Laurie Rosenberg, and Mark Legnini—for reviewing drafts, offering ideas,and providing support and
encouragement.
About the Economic and
Social Research Institute
The Economic and Social Research Institute (ESRI)
is a nonprofit organization that conducts research
and policy analysis in health care and in the reform
of social services. ESRI specializes in studies aimed
at improving the way health care services are organized and delivered,making quality health care accessible and affordable, and enhancing the effectiveness
of social programs.
5
Preface
The Economic and Social Research Institute has
established a process to develop and evaluate a continuing series of major reform proposals to extend
health care coverage to most, if not all, Americans.
Supported by The Robert Wood Johnson Foundation, the project is designed to generate the kind of
serious thinking and debate that will raise the visibility of the problem of inadequate coverage and lead
to serious consideration of far-reaching soluti on s .
We hope to inform,facilitate,and help re-invigorate
debate over how best to provide health care coverage
for the uninsured in the United States.
In the past few years, bold, sweeping proposals
have been moribund; in their place have come a
stream of incremental changes and marginal thinking that leave the basic flaws of our current system
largely untouched while tinkering around the edges.
It is necessary to create some intellectual ferment
around the need for serious reforms.
While many ingredients are necessary to jumpstart the public discussion about health care reform,
including both national leadership and public support, we believe that an essential ingredient is a
forum where new and creative ideas can be developed, dissected, debated,and reconfigured. We have
put in place a structure designed to draw the best,
most knowledgeable, and most creative policy
thinkers and analysts who are concerned with fundamental reform and achieving the ultimate goal of
universal coverage. We envision this forum to be a
place where innovative thinkers can gather to subject their plans to objective critique and continuous
refinement and where policy makers, the media,
interested stakeholders, and, ultimately, the public
look for a variety of stimulating, practical proposals
for health reform.
A major activity of this effort has been the production of a series of commissioned proposals for
extending health care coverage to people who are
now uninsured. The hope is that policy makers will
be able to draw upon the proposals generated
through this process to formulate a set of reforms
that moves the country toward universal coverage.
But there is another objective: We hope that the
process of producing, assessing, and revising these
proposals will generate interest in broad-based
reform and thereby help to raise the visibility of the
issue and create awareness of the need for comprehensive reform.
This kind of systematic consideration of reform
alternatives is justified at this time for several reasons: first,the problem of the uninsured grew worse
during the s despite a strong economy; second,
the political climate seems to have warmed somewhat toward considera ti on of solutions to health
care access probl em s ;a n d ,t h i rd , we can better afford
the cost of a solution now than in recent history.
As the proposals put forth in this volume show,
the barriers to solving the problem of the uninsured
do not include a lack of viable solutions. There are a
number of reform strategies that would solve the
problem, but deciding on a particular one is difficult
because all viable solutions require making difficult
trade-offs. There are no easy solutions—no “magic
bullets”—that will solve the problem without
requiring changes that some people would prefer
not to make. However, the fact that there are no easy
or non-controversial solutions to the problem of the
uninsured should not deter us from making a commitment to find a lasting solution. We hope that the
proposals presented here will help to take us one
step closer to achieving that end.
— Jack A. Meyer, Ph.D.
President, Economic and Social Research Institute
6
Overview
by Elliot K. Wicks
The Reform Proposal Production Process
This volume is the culmination of a year-long
process designed to develop major, comprehensive
proposals to extend health coverage to most
Americans. Our objective in carrying out this project was to produce  thoughtful, contrasting
approaches to moving the country toward universal
covera ge . We wanted the papers to represent the
best, most innovative thinking on this important
policy issue.
In issuing invitations to potential authors—all
expert health analysts and researchers—we stressed
that we were interested in proposals that represent a
rethinking of our present approaches to providing
health insurance. We explained that the major question the paper should address is “How should the
nation’s health care financing and delivery system
be reformed to ensure coverage for nearly all
Americans?”
We indicated that we were looking for new, fresh
ideas to move substantially toward universal covera ge . Although political feasibility is obviously
important, we wanted authors to consider
approaches that involve fundamental reform and
perhaps even a complete overhaul of many current
structures. We did not want polemics, advoc ac y
pieces, or attacks on past or present injustices or
inefficiencies. The proposals were to be practical,
with some attention paid to addressing how we get
from where we are now to where the authors would
have us go. That is, they should acknowl ed ge the
political difficulties and consider what has to be
overcome, but they should not assume that current
views cannot be changed. Writers were told not to
assume the present political climate.
Our process was de s i gn ed to select authors to
ensure that final papers represent the spectrum of
possible approaches and philosophical perspectives.
The purpose of this project is to engender interest
and debate, not to result in advocacy for any particular strategy or approach. Thus, an essential objective was to produce proposals that draw on the
thinking of people representing the full range of
ideas.
We invited about  people to submit letters
indicating the approach they would take in tackling
this assignment. About  responded with interest
in writing a paper; from this group, we chose 
(some of whom are a team of two or more people).
We then commissioned these authors to prepare
draft papers. Once the authors had completed a
draft, they presented their reform proposals in person to a distinguished and diverse advisory panel of
health policy analysts and researchers. These people
were chosen for their broad experience, acknowledged expertise, and differing philosophical perspectives. The advisory panel provided feedback,
comments, and critiques for each paper. Their task
was to help authors clarify and strengthen their proposals; no endorsement was asked for or given, nor
should it be inferred. The authors then prepared a
new draft, taking into account the advisory panel’s
review, as well as comments from the staff of the
Economic and Social Research Institute. The results
of the process are contained in this volume.
These papers are just one step in an iterative
process. These proposals lack cost estimates. A
research firm will be commissioned to model these
proposals for the purpose of providing estimates of
their cost and impact on the uninsured. As part of
this process, authors will have the opportunity to
revise their proposals, particularly those el em en t s
that have a crucial impact on the cost. So, in one
sense, the proposals that appear here are provisional;
authors are not wedded to every particular parame-
7
ter in their proposals, and all reserve the right to
amend them if they think it is prudent to do so. The
cost estimates,along with the revised versions of the
proposals, will be disseminated at a later date.
FIGURE 1
Proportion of the Uninsured Under Age 65 in
Various Income Categories Measured as Percent
of Federal Poverty Level, 1999
The Dimensions of the Problem
Approximately  million Americans—or . percent of the non-elderly population—are without any
kind of health covera ge , either public or priva te .1
This figure is evidence of a major social problem that
needs to be resolved. It has serious negative consequences that include lack of access to needed care,
declining health, and, for some people, the assumption of crushing financial burdens. Unfortunately,
despite the recent long period of sustained prosperity and economic growth, the number of people
without health insu ra n ce protection has increased
appreciably in the last decade.
To craft sensible solutions to the problem of the
uninsured requires understanding the dimensions of
the problem. On the one hand, the  million figure
disguises the complications regarding the scope of
the problem. The number who are without health
coverage at some time during the year is higher than
 million, probably about  percent higher. On the
other hand, perhaps only half of the  million lack
coverage during the entire year. Both of these figures
reflect the fact that at any time many people are moving in and out of insurance, which clearly complicates
the task of engineering policies to cover them. Nevert h el e s s ,s ome people lack coverage for long periods,
and they are probably most vulnerable to incurring
high medical costs or going without needed care.
The characteristics of the uninsured are well
documented. They are predominantly lowerincome people, which suggests that a high proportion of the uninsured lack coverage because it is not
afford a bl e . About  percent of the non-elderly
uninsured are people with incomes that put them
below the federal poverty level, despite the availability of major public programs, primarily Medicaid
Unless otherwise noted, the data in this section are taken from
Sherry A. Glied. Challenges and Options for Increasing the Number
of Americans with Health Insurance. The Task Force on the Future of
Health Insurance, The Commonwealth Fund, January 2001.
1
Source: Glied, 2001, Table 1.
and the State Children’s Health Insurance Program
(S-CHIP), to cover low-income people. Another 
percent have incomes between the poverty level and
 percent of the poverty level. In total, people
with incomes below  percent of the poverty level
account for  percent of the uninsured (see figure
). The median family income of all uninsured families is only , (see figure ).
The uninsured are predominantly members of
families in which someone is employed. Only 
percent do not have a worker in the family. But
many of those who work ( percent) are not
offered health insurance. This leaves them with the
option of buying coverage in the individual market,
which is even more expensive than employer-sponsored coverage. Although the uninsured are disproporti on a tely employed in low-wage firms (thus,
their low average incomes), they are distributed
across firms of all sizes: in ,  percent were
employed in firms with fewer than  employees; 
percent were self-employed; but the remaining 
percent were employed in firms with  or more
employees. In fact,  percent were in firms that
employed  or more employees.
Uninsured adults come from all (non-elderly)
8
FIGURE 2
Income Characteristics of the Uninsured,
Non-Elderly Population, 1999
Family Income as
Percent of Federal
Poverty Level
Percent in Income
Group Who Are
Uninsured
Median
Family
Income
Under %
%–%
%–%
%–%
More than %
All Uninsured
%
%
%
%
%
%
 ,
 ,
,
 ,
 ,
 ,
Source: Glied, 2001.
age groups: about a quarter of all uninsured adults
are age  to ; another quarter are between  and
 years old; the remaining half are between ages 
and . Children make up about  percent of the
total uninsured.2
Many of the uninsured are eligible for some kind
of coverage. About  percent decline coverage available from their employers. Many others—about
one-quarter of uninsured adults and two-thirds of
uninsured children—are el i gi ble for public programs but fail to enroll for one reason or another.
These statistics demonstrate that making coverage
available is not enough. The uninsured have to
know about the available coverage, have to be able
to afford the portion that is not subsidized,and have
to be willing to go through the process of establishing eligibility and enrolling.
The uninsured represent a diverse, constantly
moving target. Unidimensional strategies will not
solve the problem. The proposals presented in this
volume recognize that fact.
Categories for the Papers
All the proposals that follow differ from one another in important ways. Some build on the current
Bureau of Labor Statistics. Current Population Survey. March 2000
Supplement.
2
system in one way or another, while others would
jettison major parts of it; still others fall somewhere
in between. Some would move quickly to virtually
universal covera ge ; others would progress more
gradually toward that goal.
But as different as they are, the proposals share
certain notable features. They all see affordability as
the most important barrier to broader coverage,and
they all make substantial financial subsidies available to a major portion of the population (typically
to those with income levels up to the median or up
to  percent of the federal poverty level). Except
in those instances where coverage is a matter of
“right,” so that there is no test of eligibility for subsidies, they all use income alone as the only determinant of eligibility; that is, none would confine
coverage to families, to children, to people whose
assets fall below a defined level, to those who do not
already have coverage, or to any other group. They
all seek to improve horizontal equity by making the
level of subsidies less dependent on the state where
one lives. They all look to revenues from the federal
government as the major source of financing,
though they differ with respect to how the revenue
would be collected (for example,a payroll tax versus
general revenues). On the other hand, they all give
some flexibility to states,allowing them to influence
some elements of the new system (though there is
wide difference in the level of state influence). Only
one of the proposals would eliminate employers as a
source of coverage and a means of risk pooling,
although one other might produce this result over
time. Finally, the proposed reforms (with the possible exception of one) would not dramatically
change the role of health insurers and health plans;
they would still continue to perform essentially the
functions that they do today.
Several other features are common to a significant proportion, but not all, of the proposals.Five of
the reform proposals would rely heavily on some
kind of purchasing pool or insurance exchange.
These entities would offer coverage from a variety of
insurers and health plans, with some standardization of products, and would generally be open both
to employers and individuals. Usually, individual
enrollees would be able to choose any plan partici-
9
pating in the exchange. In a number of instances,
subsidies would be available only if people bought
coverage through the exchanges. A feature common
to six of the proposals is the use of tax credits (or
vouchers) to subsidize the cost of care. Typically,
proposals that espouse this approach also propose
to alter the tax tre a tm ent of employer-paid insurance premiums: They would eliminate or reduce the
right of employees to exclude this form of compensation from their calculation of taxable income.
Although any attempt to categorize the proposals
runs the risk of trying to force a proposal into a category that does not quite fit, it is still useful to think of
them as falling into three general categories, based
on their approach to subsidizing coverage.
. The tax credit/voucher approach. Proposals that
adopt this approach provide subsidies in the form of
tax credits or publicly funded vouchers that can be
applied against the cost of buying health coverage.
Reform proposals that take this approach tend to
encourage consumers to buy coverage in the private
insurance market. That is, they may do away with
Medicaid and S-CHIP (except for continuing with
Medicaid coverage for special-needs populations
and long-term care) and, instead, have people formerly covered by these programs buy coverage in
private markets, just as the rest of the population
currently does.
. The Medicaid/S-CHIP expansion, premiumreduction approa ch . Approaches that fo ll ow this
model generally build on these existing public programs by raising the income limits to include many
more needy people, doing away with all tests of eligibility except income, and including provisions to
move toward uniform eligibility standards across
states. Another way to think of these approaches is
that they are subsidizing the purchase of coverage by
reducing the premium; for the lowest-income
group, the premium is zero, and people above this
level may pay something toward the premium but
still less than the full price. Understandably, these
approaches tend to give states substantial responsibility for designing, implementing, and operating
the subsidy programs. Although they build on the
foundation of existing public plans, these proposals
sometimes give subsidized people the option to pur-
chase a non-governmental plan.
. The social insurance/health care as a “ri gh t”
approach. The essence of this approach is to guarantee everyone—regardless of income or any other
characteristic—access to a standardized benefit
package at no direct cost to the covered person.
Health coverage is “free” in the same sense that public education is free; that is, it is publicly financed
through taxes. There are no tests of eligibility
(except, perhaps, legal residency status). This
approach can be thought of as applying the social
insurance aspects of the Medicare model to the
whole population. However, the approach is compatible with having people covered through their
employers as long as a no-cost plan is available. In
fact, several of the proposals combine this approach
with payroll-tax financing and a “play or pay”
requirement for employers.
We now turn to a brief description of each proposal, identified by author.
The Proposals in Brief
Stuart Butler. This approach provides refundable tax
credits to virtually all households. Credit amounts
would be based on need as determined by family
income and the amount of health costs (including
premiums and out-of-pocket costs). One option for
lower-income families would be a fixed credit of up
to ,. The sliding-scale credit, available to all, is
structured to limit health costs as a percent of
income: the credit covers a higher proportion of
health costs as these costs account for a higher share
of income. Availability of tax credits would enable
the gradual repeal of the current tax subsidy for
working people, so that employees could no longer
exclude from their taxable income the amount that
employers contribute on their behalf to health
insurance premiums. Workers whose employers
offer coverage would have to use their credit to get
coverage at the workplace, but other employees and
individuals could use their tax credit to buy coverage from a range of other organizations, including
other employers, churches, unions, associations
plans, and other affinity groups. To receive new federal grants to supplement the tax credits, states
10
would have to agree to take certain steps to make
coverage more affordable for low-income workers.
Mark Pauly. This tax-credit/voucher approach
emphasizes the advantage of beginning reform with
a relatively simple,financially feasible,easily modifiable intervention. The author stresses that there is
much uncertainty about the response of insurers
and consumers to major changes,so it is desirable to
adopt an approach that can be readily changed as
experi en ce shows what works and what does not.
The principal targeted group is lower-middleincome households with incomes above the poverty
line but below the median income (regardless of the
age, sex, or relationship of family members). They
would be eligible for a voucher or credit that is large
enough to cover perhaps one-half to two-thirds of
the premium for moderately comprehensive covera ge . The credits, in the form of coupons worth
, for individual coverage and , for family
coverage, could be used for any qualified insurance
that costs at least as much as the coupons. Very-lowincome households would be eligible for publicly
provided or contracted comprehensive insurance,
with no premium share required. People with
incomes above the median (with a few exceptions
for high risks and possibly those with incomes near
the threshold) will not be eligible for the new program initially, but would retain the tax exclusion for
group coverage. This approach encourages people
to acquire coverage through private markets with
the expectation that the market will adapt and
improve under the impetus of the influx of new
buyers.
Sara Singer, Alan Garber, and Alain Enthoven.
This tax-credit approach to reform is built around
purchasing pools,called insurance exchanges, which
would be established in all geographic regions and
would provide coverage on a guaranteed-issue and
community-rated basis. There are three incentives
to form exchanges: tax credit subsidies could be
used only for coverage purchased through
exchanges; exchanges and participating plans are
not su bj ect to state coverage mandates; and
exchanges and participating plans are protected
against adverse selection through premium risk
adjustment. If no exchange arises vo lu n t a ri ly, a
national exchange,similar in function to the Federal
Employees Health Benefits Program, would be
available to firms with fewer than  employees.
Medicaid and S-CHIP would still be an option for
low-income individuals, but they and families with
annual income up to , would be eligible for a
refundable tax credit equal to  percent of the
median-cost exchange plan. Low-income people
who do not use the tax credit to en ro ll in a plan
would be automatically enrolled in a default plan,
and states would be el i gi ble for federal grants to
cover part of the costs of these otherwise uninsured
people. The tax credit would continue on a phasedout basis for families with incomes up to ,;
these people would have to choose between the tax
credit and the tax exclusion for employer-paid premiums. Families with even higher incomes could
continue to enjoy the tax exclusion subsidy, but it
would be capped.
Jonathan Gruber. This proposal is a premiumreduction approach, but it is also built around voluntary purchasing pools in each state. Most if not all
insurers would be expected to participate in these
pools because subsidies are available only if coverage is purchased through the pool. For people below
 percent of the poverty level, enrollment in a
pool-based plan is automatic and free of charge. For
those between  percent and  percent of poverty, the premium would be priced to limit the cost of
coverage to  percent of income. Others receive no
subsidy. Employers have incentives to use the pool
as their coverage source, since their employees are
eligible for subsidies only if they purchase coverage
through the pool. Financing comes from general
revenues, a cap on the amount of employer-paid
premiums that can be excluded from employees’
taxable income, and the phase-out of S-CHIP and
Medicaid. Payments to health plans are paid on a
risk-adjusted basis to encourage fair competition,
discourage risk selection, and reward efficiency.
John Holahan, Len Nichols, and Linda Blumberg.
This proposal is a premium-reducti on ,s t a te - b a s ed
approach that extends S-CHIP-type coverage to a
larger portion of the population and encourages
purchasing pools by making subsidies available only
if coverage is purchased through pools. States have
11
incentives to participate because the federal share of
the per capita cost is  percent higher than Medicaid
matching rates.States also have considerable flexibility in designing and operating the subsidy program.
But the program would create an entitlement, so
enrollment could not be limited by the state, and the
federal government would specify a minimum benefits package. The standards for eligibility, based on
income alone, would be the same across all states.
People below  percent of the poverty level would
get full subsidies, while those between  percent
and  percent of poverty would get partial subsidies. The new purchasing pools would be composed
of Medicaid and S-CHIP recipients,state employees,
and anyone else who chooses to join. Anyone,
regardless of risk, could buy a standard benefits
package through the pool at a cost that would never
exceed a computed statewide community rate.
Employers may, but are not required to, buy coverage
through the pools, but they must offer employees
pool coverage as an option, because subsidies are
available only to those purchasing through the pool.
After five years,states could mandate that all individuals purchase coverage.
Judith Feder, Larry Levitt, E ll en O’ B ri en , and
Diane Rowland. This approach considers how an
extension of S-CHIP might be combined with an
employer tax credit. Given the authors’ assumption
that comprehensive reform is highly unlikely in the
foreseeable future, their emphasis is on not disrupting but expanding a system that works quite well for
low-income people, who are the main focus of the
authors’ concern. They also want to use new dollars
most efficiently, that is, to avoid having public coverage, whether credits or S-CHIP extensions, substitute for employer-based coverage. They would make
comprehensive benefits available at no cost to all
individuals with incomes below  percent of the
federal poverty level. Such coverage would also be
available to people between  percent and  percent of poverty, but they would pay some premiums
and cost sharing, up to a maximum of  percent of
income. The authors would prefer that eligibility be
an entitlement, as with Medicaid, rather than being
dependent on annual state appropriation decisions,
as with S-CHIP. The proposal looks favorably on
using tax credits to help fund coverage for those
with incomes above  percent of poverty. The
authors suggest that it is difficult to design a modest
individual tax credit that is simultaneously effective
and well-targeted, while at the same time avoiding
“crowd-out” of existing publicly sponsored and
employer-sponsored insurance. A better approach,
they indicate ,m ay be tax credits provided to employers to encourage greater offering of insurance.
E ll i ot Wicks, Jack Meyer, and Sharon SilowCarroll. This is a tax credit approach that guarantees
universal coverage by mandating that everyone buy
coverage and by having Medicare provide fallback
coverage for anyone who still remains temporarily
uninsured. The new program replaces S-CHIP and
Medicaid (except for long-term care and special
populations). Everyone is eligible for tax credits, but
the tax exclusion for employer-paid premiums is
eliminated. People at or below the poverty level get
credits adequate to purchase coverage comparable
to that provided by Medicaid. People above the
median income level get a tax credit roughly equal
to the average value of the current tax exclusion.
Those in the middle, between the poverty level and
the median income, get credits that decline in value
as income rises, starting with a credit sufficient to
buy Medicaid coverage and ending with a credit
equal to the credit that will be available to people
with incomes above the median. People who do not
have proof of private coverage automatically default
to Medicare; so all providers know that Medicare
will pay for those without other covera ge . To discourage permanent use of the Medicare default, at
tax time, people who cannot show proof of other
coverage have to pay a premium plus a penalty (to
which the tax credit can be applied) for every month
they were covered by Medicare. Administration of
all claims, coordination of benefits, etc., goes
through a centralized administrative entity to
improve efficiency and reduce administrative burdens for providers, patients, and health plans.
Jacob Hacker. This proposal is a modified “play
or pay”approach that creates incentives for individuals and employers to participate in Medicare (with
augmented benefits). Employers would be required
either to provide and pay for private coverage and
12
automatically enroll their employees in an employer-chosen plan or pay a modest payroll tax ( percent) and enroll their employees in Medicare. For
many employers, especially those with low-wage
employees or higher-risk populations, the modest
payroll contribution would be less costly and thus
more attractive than offering their own plan.
Workers would also pay a premium toward
Medicare covera ge , based on income and family
size. States would have federally financed incentives
to enroll in Medicare those people who do not
enroll through employment. S-CHIP and Medicaid
would be replaced.States would also provide “wraparound” coverage to ensure that those who would
have been in Medicaid still receive similarly comprehensive benefits. Individuals could buy into
Medicare, paying premiums based on income. The
author estimates that  percent to  percent of the
population could eventually en ro ll in Medicare
Plus, which would give the program bargaining
leverage and en su re broad pooling of risk.
Eventually, a mandate to have coverage would apply
to everyone.
Alan Weil. This proposal combines three key elements: making access to standard free health coverage a “right,” requiring employers to “play or pay,”
and allowing people not covered by employer plans
to buy coverage through large purchasing pools
called “ i n su ra n ce exchanges.” Unless an employer
offers one of the federally defined standard benefits
packages, automatically enrolls all employees, and
pays for  percent of the employer premium and 
percent of dependent covera ge , the employer and
the firm’s employees pay a payroll tax equal to
approximately what the average employer and
employee premium is today. People not covered by
employer plans get coverage through insurance
exchanges that operate in all geographic regions and
can be private or public entities licensed by the federal government. Exchanges seek bids from and
contract with a number of health plans. Individuals
can choose any health plan the exchange offers; they
have a ch oi ce of both standard and non-standard
benefit packages. All plans are offered on a commun i ty - ra ted basis. Using revenues from the payroll
tax, exchanges provide at least one insurance option
that can be obtained at no charge, but they also offer
more expensive coverage options to those willing to
pay more. One option would be a medical savings
account. S-CHIP and the portion of Medicaid that
serves low-income, non-disabled adults and children are eliminated, but states operate subsidy programs to cover the copayments, deductibles, and
uncovered services that these families face.
Richard Kronick and Tom Rice. This social insurance approach combines a number of key features:
access to health coverage as a “ri gh t ,” maximum
flexibility to states, payroll-based financing, and
elimination of employers’ roles in providing health
coverage. To be granted the federal funds that
finance most of the cost of the new system, each
state must establish a financing and delivery system
that makes access to a zero-premium, relatively
comprehensive, standard benefits plan available to
all. To get coverage,a state’s citizens need only enroll
with the state; there are no income or other eligibility tests. Individuals who do not enroll voluntarily
are assigned to a no-cost plan. More comprehensive
plans will generally also be available, but at an extra
cost.States have great latitude in designing their systems; for example, both a single-payer system or an
approach based on contracts with private health
plans would be acceptable. To get federal funding,
states would show that they will en ro ll at least 
percent of the population and meet other standards
related to maintaining quality of care and ensuring
good administra ti on . States would be expected to
contribute to the funding of the new program in an
amount equal to  percent of their present
Medicaid and S-CHIP spending (both of which
would be eliminated), but the bulk of the funds
would come from a federal payroll tax equal to 
percent of what is currently spent by employers and
employees for health coverage. Low-wage employers
and employees would pay less. The division of the
tax between employers and employees would be
based on current distribution of spending. The federal government would monitor and enforce quality
standards and provide incentives for states to
improve outcomes. n
13
A Comparison of Reform Plan Features
The following table provides a side-by-side comparison of the features of the
reform plans that are described in detail in the following chapters. The plans are
identified by the names of the authors.
A Comparison of Reform Plan Features
Butler
Feder / Levitt / O’Brien /
Rowland
Gruber
Hacker
Holahan / Nichols /
Blumberg
General
Approach
Would make refundable tax
credits available to working
households. States would
get grants to expand health
coverage to more residents
and make insurance more
affordable. Coverage obtained
at work or from a range of
other organizations such
as churches or unions.
Expand Medicaid and the
State Children’s Health
Insurance Program for lowincome people. Possible
combination with tax credit
to small, low-wage firms to
expand employer offerings.
Establishment of purchasing
pools in every state through
which households with
incomes up to 300% of the
federal poverty level would be
eligible for no-cost or
reduced-cost coverage on a
sliding-scale basis; automatic
plan enrollment for lowestincome households.
A modified “play or pay”
approach that creates incentives for workers and employers to buy into “Medicare
Plus,” a national program
based on Medicare.
Extend the type of subsidized
coverage that is currently
available under S-CHIP to all
lower-income people and
subsidize insurance for the
highest risk.
Target
Population
Working uninsured individuals and families; the plan
would achieve near-universal
coverage for all working
households of legal U.S.
residents.
People below 150% of poverty
level covered at no cost; those
between 150% and 200% of
poverty would pay some
premiums and cost sharing.
Higher-income people could
buy-in to public coverage and
pay a sliding-scale premium.
Employees of small, low-wage
firms benefit from tax credit.
Individuals and households
under 300% of the federal
poverty level would receive
subsidies. Households with
incomes below 150% of
poverty level would be eligible for no-cost coverage.
All Americans not covered
by Medicare or employersponsored insurance.
Individuals with incomes
under 250% of the federal
poverty level and those at
high health risk. Subsidies
available only to those who
enroll through the state
purchasing pool.
Form of
Public
Programs
Refundable tax credit, funded
via repeal of federal income
tax provision that makes
employer contributions to
employees’ health insurance
non-taxable income; federal
tax revenues would fund
grants to states to help lowincome families buy coverage.
S-CHIP expansion, federally
subsidized, with some state
match, for those with limited
incomes, and a federal tax
credit subsidy for small
employers to help cover
workers.
Household income determines eligibility for no-premium plans (for households
under 150% of poverty level)
or reduced-premium plans
(for households under 300%
of the federal poverty level
on a sliding-scale basis but
premium not more than 10%
of income).
Premiums for those buying
into Medicare Plus would be
scaled to income, with lowerincome citizens paying only a
small percent of income.
Employers would be eligible
for transitional subsidies and
for reductions in their contribution rate based on firm
income.
Increased federal-funding
match to participating states;
full subsidies to people below
150% of poverty; cost-sharing
up to 7% of income for people
between 150% and 200% of
poverty and to 12% for people
between 200% and 250% of
poverty. Higher-risk individuals, regardless of income, pay
no more than a statewide
community rate.
Mandates
for
Coverage
None, but to receive tax credit, individual or family would
have to buy a health plan that
included a minimum set of
benefits. High-level of voluntary compliance expected
among most workers since
employees required to tell
employers which health plan
they wished to join.
None.
None.
None initially but individual
mandate would apply eventually if a nontrivial share
of Americans remained
uninsured.
After five years, states could
mandate that everyone be
covered.
Sources of
Funding
Savings from elimination of
existing tax exclusion, and
federal general tax revenues.
Federal general revenues, with
state matching payments.
Federal general revenues,
savings from replacement of
Medicaid and S-CHIP health
programs, and limits on tax
exclusion for employer-provided insurance.
Payroll contributions and premiums, general revenues, and
other smaller sources.
Federal general revenues, and
cuts in existing programs
since the need would be
reduced as health reform is
implemented.
15
Kronick / Rice
Pauly
Singer / Garber /
Enthoven
Weil
Wicks / Meyer /
Silow-Carroll
All non-elderly legal residents
would be guaranteed comprehensive health insurance as a
“right” (at no direct cost)
through a public insurance
approach designed by each
state and monitored by the
federal government.
A refundable tax
credit/voucher system would
make some level of coverage
affordable to lower-middleincome people who currently
have no health insurance.
Very-low-income households
would initially be eligible for
publicly financed zero-premium comprehensive insurance.
Combines refundable tax
credits and insurance
exchanges to promote lowercost, higher-value health
coverage while allowing
employers and individuals
to continue current arrangements if they desire.
A new Medical Security
System would be created to
provide universal coverage,
making coverage a “right.”
Tax credits for all households,
varying by income. Universal
coverage achieved by mandating that everyone have or buy
health coverage and having
Medicare automatically
cover anyone temporarily
uninsured. Builds on present
system of private health
plans and employer-based
coverage.
All non-elderly legal residents.
Principal target group is
lower-middle income families
and individuals with incomes
above the federal poverty
line, or about half of the
uninsured. Very low-income
families covered publicly, at
least initially.
Low and moderate-income
people who are not eligible
for Medicare.
All legal U.S. residents under
age 65.
All of the uninsured.
Federal subsidies to states
to finance availability of nocost coverage to all legal
residents.
A voucher or tax credit large
enough to cover one-half to
two-thirds of the premium for
moderately comprehensive
coverage. The credits would
be in the form of coupons
worth $1,500 for individual
coverage and $3,500 for
family coverage. No-cost
publicly financed coverage for
very low income households.
Continuation of Medicaid/
S-CHIP for eligible individuals
and families who choose to
stay in these programs;
refundable tax credits equal
to 70% of median-cost health
plan; federal payments to
states equal to 50% of the tax
credit to cover the costs of
running “default plans” for
people who do not enroll.
Payroll tax, Medicaid, and
S-CHIP funds.
Refundable tax credits for all
households but varying
according to income—minimum credit approximately
$700 a year for an individual
and $1,200 a year for a family.
People below 100% of poverty
would get credit sufficient to
buy coverage comparable to
Medicaid. Those above that
level up to median income
would get gradually reduced
subsidies.
All legal residents under age
65 automatically covered by
comprehensive benefits.
Everyone would have at least
one health insurance option
that would not require payment of premiums. There
would be a mandatory payroll
tax.
None.
None.
All employers and employees
would pay a new payroll tax.
All people would have to
enroll or be enrolled by
default.
Every individual and family
would have to have health
coverage at least as comprehensive as Medicare’s, plus
prescription drugs and wellchild care. Those who fail to
show proof of purchase would
pay a premium plus a penalty
for Medicare backup coverage
for every month without other
coverage.
Primary revenue source would
be a payroll tax levied on
employers and employees,
supplemented by federal general revenues, state revenues,
and, in some states, premium
payments from individuals.
Federal budget revenues;
those who buy more expensive coverage would pay outof-pocket. Full coverage for
those with incomes below
125% of the federal poverty
level would be financed
through a combination of
state and federal revenues.
Phased-in cap on current federal tax exclusion; general
revenues; and savings over
time from changing consumer
behavior and increasing
health plan competition.
Payroll tax, premiums, and
federal subsidies.
Federal general revenues, but
partially offsetting savings
would be realized from the
elimination of Medicaid and
S-CHIP and from making
employer-paid health premiums taxable income for
employees.
Butler
Feder / Levitt / O’Brien /
Rowland
Gruber
Hacker
Holahan / Nichols /
Blumberg
Major Tax
Changes
Repeal of the federal income
tax provision that makes
employer contributions to
employees’ health insurance
a non-taxable form of income.
Explores tax credits to individuals or employers, the latter
to subsidize the offering of
coverage to uninsured workers with modest incomes.
Limits the tax exclusion for
employer-provided insurance
equal to no more than the
cost of the median-cost plan
in each purchasing pool.
Cap on tax exclusion of
employer-provided health
insurance at level of twice the
average premium of Medicare
Plus coverage.
Federal taxes would be
increased if surplus not
available.
Level of
Benefits
To qualify for the tax credit,
families would have to enroll
in a health plan that included
at least the minimum insurance package, which would
be primarily catastrophic
coverage.
Comprehensive but not
specifically delineated.
Physician services, inpatient
and outpatient hospital, prescription drugs, nominal
payments for well-child care,
prenatal care, and immunizations.
A defined benefit package
similar to Medicare plus outpatient prescription drugs,
preventive services, mental
health benefits, and maternal
and child health care.
States determine a new standard benefit package—within
federal guidelines—for everyone under 250% of poverty
and those at high health risk.
Role of
Federal
Government
Would establish a default
system of health insurance
regulation to encourage availability of affordable insurance;
would establish a benchmark
health plan with basic features and catastrophic
protection. Would monitor
state compliance and work
with states on a plan to
eliminate uninsurance.
Would make federal funds
available at enhanced
Medicaid matching rates
to states willing to cover
targeted uninsured.
Funds subsidies, sets
minimal rules, provides
oversight of purchasing
pool administration.
The Health Care Financing
Administration would have
primary responsibility for
administering Medicare Plus.
In addition to offering standard fee-for-service coverage,
Medicare Plus would also
allow beneficiaries to enroll
in private health plans that
contracted with the program.
Financial support, monitor
state compliance of minimum
rules, oversee state spending
and enforcement.
Role of
State
Government
Would develop a mechanism
to supplement federal tax
credit for eligible workers and
help cover those who did not
purchase minimum insurance.
Would have to use additional
federal funds to expand existing or develop new programs
to achieve target levels of coverage. Would work with
health insurers on insurance
reform that keeps benefits
affordable.
Would provide coverage to
low-income uninsured residents, consistent with federal
rules affecting eligibility, benefits, administration, and
other program aspects.
Not addressed, except for
continued responsibility for
remaining parts of Medicaid.
Would transform from
provider of insurance to a portal for coverage under the
new Medicare Plus system.
States would continue to
finance care for the eligible
aged, blind and disabled. In
addition, they would have to
reach out to and enroll nonworkers, provide wraparound
coverage for those who would
have been in Medicaid, and
subsidize premiums for
unemployed people.
Increases role of states significantly while granting more
flexibility.
Kronick / Rice
Pauly
Singer / Garber /
Enthoven
Weil
Wicks / Meyer /
Silow-Carroll
Payroll tax substitutes for
employer and employee
premiums, which has implications for tax exclusion
provision of employer
premium contributions.
No major tax code changes,
but tax credits in the form of
coupons would help people
purchase qualified health
insurance. The new vouchers
would be viewed and treated
as tax reductions for those
who use them.
Phased-in cap on current
federal tax exclusion for
employer-paid premiums.
New payroll tax would be
established for employers
and employees.
The tax exclusion for
employer-paid health
premiums would be
eliminated.
A federally-defined standard
benefit package. Benefits
would include prescription
drug coverage; dental and
long-term care would not be
required.
To qualify for the credit, the
plan would have to cover
effective medical and surgical
services, prescription drugs,
and medical devices based on
a standard definition. Patient
cost sharing would be permitted, as would managed care.
Generally determined by the
market, with minimum standards set by the Insurance
Exchange Commission,
including goods and services
known to be medically effective and provided at reasonable cost.
Guarantee is for basic coverage, but individual may
supplement with own funds
to buy more comprehensive.
A package of benefits comparable to Medicare’s plus a
prescription drug benefit and
well-child care coverage.
Would impose payroll taxes
on employers and employees,
calculate money needed and
provide funds to each state
health care system, monitor
state implementation of
expansions, measure quality
and health outcomes, determine and update standard
benefit package, monitor and
regulate quality of care in
states.
Would make information
about insurance purchasing
and plans available, including
price and quality and could
subsidize the production
and distribution of such information. It also would be (or
contract with) an insurer of
last resort.
Establish the Insurance
Exchange Commission to
oversee insurance exchanges,
distribute tax credits and
make default plan payments.
Establishes U.S. Insurance
Exchange as backup in
markets without private
exchanges.
Would set up and regulate
insurance exchanges, forward
tax revenues, and determine
size of payroll tax.
Would fund all tax credits.
Would establish general
guidelines for states setting
up the aggregate purchasing
arrangements (APA). Would
continue to operate Medicare,
for the elderly and as a
temporary back-up plan for
people who do not have proof
of private coverage.
States would have much flexibility in designing a system—
how to pay health care
providers (e.g., single payer
vs. competing health plans),
be responsible for raising revenue to supplement federal
financing, meet federal
requirements, and enroll residents in health plans. Would
provide information on enrollment options and procedures,
negotiate with health plans
and providers, regulate health
plans, and collect data to
evaluate the system.
Would have primary role of
selecting or managing the
public plan for poor people
not currently covered by
Medicaid. Could continue to
regulate individual insurance
and regulate risk-rating. In
addition, states could choose
to provide payments for people with high medical expenses, possibly allowing smaller
deductibles or less-constraining upper limits in low-cost
plans.
Continue to provide Medicaid
and S-CHIP; use new federal
funds to pay for care under
default plans by reimbursing
safety-net providers.
States would continue to pay
some Medicaid costs to keep
coverage at current levels;
would subsidize copayments
under basic plan for lowincome residents.
Each state would be required
to establish an aggregate purchasing arrangement through
which small employers and
individuals would purchase
coverage. In exchange for no
longer financing the acute
portion of Medicaid or S-CHIP,
states would assume greater
responsibility for long-term
care services under Medicaid.
17
Butler
Feder / Levitt / O’Brien /
Rowland
Gruber
Hacker
Holahan / Nichols /
Blumberg
Effects on
Existing
Public
Programs
Medicaid and S-CHIP would
continue as now.
Medicaid and S-CHIP would
continue and be expanded.
Gradual phase out of
Medicaid and S-CHIP (and
accompanying federal subsidies) for those families who
qualify on income alone.
Medicaid remains in place for
the elderly and disabled.
Would eventually replace
existing public programs for
the uninsured with a single
national program based
on Medicare. Medicaid and
S-CHIP would be phased out
with eligibles automatically
enrolled in the new Medicare
program or employer-sponsored plans.
Participating states would
receive enhanced federal
S-CHIP matching rate for all
current Medicaid and S-CHIP
beneficiaries under 250%
of poverty; all states must
continue smaller, residual
Medicaid program for children
and adults with special needs
as well as all long term care
services; would eliminate
federal payments to states
covering individuals with
incomes above 250% of
poverty. No change in
non-participating states.
Role of
Insurers/
Health
Plans
Would continue to be a major
source of coverage. Would
have to bring premium rates
into line with federal or state
underwriting and benefit
requirements, but would
benefit from administrative
savings associated with the
automatic enrollment system.
Would stay the same as
today, although some market
reforms might be necessary.
Could participate in stateestablished purchasing pool
or continue to operate outside of such arrangements.
Would stay the same as
today; would compete for
business from Medicare Plus
system.
Health plans participating in
the new state plan would be
required to accept all applicants, with premiums set at
a statewide community rate.
Payments to plans would be
risk adjusted. Insurers would
not be subject to any new
federal market regulations
outside the state purchasing
pool.
Role of
Employers
Similar to present but would
have to inform employees
about the tax credit program
and deliver the tax credit.
Would serve as a clearinghouse, creating automatic
enrollment mechanisms for
insurance, setting up payroll
deduction and payment
systems for employees and
providing proof of insurance
for each worker.
Similar to present. If tax credit
were pursued, small lowwage employers would be
encouraged to offer insurance
to their employees; employers would receive the
tax credit if they provided
insurance.
Would continue to offer
health coverage to workers,
but could do so within the
purchasing pool or outside
of it.
Employers would enroll workers at workplace. They could
choose to sponsor coverage
at least as generous as the
new program’s or pay a modest payroll-based contribution
to fund public coverage.
Would continue to have
choice to offer health coverage to their workers. If they
offer, they must make state
plans available, but they can
also offer plans outside the
state pool.
Risk Share/
Purchasing
Pools/
Insurance
Regulation
Insurance industry and states
would have to work together
to develop a means for
adjusting risk among plans.
Possible reforms in the individual insurance market
unless tax credits could be
applied to a publicly managed
insurance product.
Purchasing pools are foundation of proposal: subsidies
are available only for coverage purchased through the
pools.
To avoid adverse selection,
measures are imposed to
make it more difficult for
employers to shift between
public and private coverage.
50% to 70% of the population
might eventually enroll in
Medicare Plus, providing
strong bargaining leverage
and broad pooling of risk. No
new regulations are imposed
on private insurance, and
there are no insurance pools.
State-established purchasing
pools are foundation of proposal. Medicaid (except the
disabled and elderly) and
S-CHIP enrollees and state
employees would be included
in the pool. The pool would
be open to individuals and
employers, and insurers
could offer standard benefit
package at a statewide community rate, plus add-on
products priced separately.
Kronick / Rice
Pauly
Singer / Garber /
Enthoven
Weil
Wicks / Meyer /
Silow-Carroll
Would vary by state, but new
state program could replace
S-CHIP and portions of
Medicaid.
Medicaid and S-CHIP would
continue, and more lowincome people would be
subsidized to enroll in these
programs or some other
public program.
Medicare remains intact;
people enrolled in Medicaid
and S-CHIP may stay in these
programs or opt instead for
tax credits to be used in the
private market.
S-CHIP would be subsumed;
Medicaid would be mostly
subsumed.
S-CHIP and Medicaid largely
replaced, except for disabled
and elderly.
In some states, plans would
compete for business from
states and would have to
include services specified in a
federally-defined benefits
package. Some states might
choose to pay providers
directly and eliminate the role
of insurers/health plans.
Would continue to be major
source of coverage. Would be
required to guarantee
renewability in the individual
market and to set premiums
on modified community-rating
basis in the small-group market. Insurers would redeem
vouchers or certificates.
Would compete to provide
low-cost, high-quality care;
collect and report quality of
care and health outcomes
data.
Plans would contract with
health insurance exchanges
to offer range of plans, including a “no-cost” plan (that
is, no enrollee contribution);
would market plans and
monitor quality of care.
Would continue to be major
source of coverage but would
be required to offer a policy
that covers the services
comparable to Medicare plus
prescription drugs and wellchild care, to participate in
purchasing pools, and to
community rate in individual
and small-group markets.
Employers would no longer
provide or buy health
coverage for their workers.
Although employer role
would be eliminated, both
employers and employees
would have to contribute to
financing coverage.
Similar to current role.
May become their own insurance exchange; continue to
offer benefits to employees;
or purchase coverage from
exchanges.
Employers would collect
payroll tax but could opt out
by offering own generous
plans to employees.
Employers would be required
to offer (but not necessarily
pay for) coverage for employees and dependents. Benefits
must be at least comparable
to Medicare plus a prescription drug benefit and wellchild care. Employers with 10
or fewer employees would
have to offer coverage
through the purchasing pool.
Since coverage in no-cost
plan is automatic, everyone is
pooled together, though
states would have latitude to
decide specifics.
Few restrictions would be
placed on qualifying coverage. But all policies must
have a guaranteed renewability clause, and low-cost
policies must be sold under
modified community rating.
Plans with more generous
coverage could charge higher
premiums to high-risk people.
Insurers could impose modest
waiting periods for people
who did not enroll during
open season.
The Federal Insurance
Exchange Commission would
develop risk- adjustment
strategies. Payments would
be risk-adjusted both
between health plans within
an exchange and across
exchanges.
Insurers selling through
insurance exchanges would
be required to offer guaranteed-issue, community rated
standard benefit packages.
All health plans would have
to accept all individual and
small-group applicants and
provide immediate and full
coverage for all covered
benefits with no waiting
periods or exclusions for prior
conditions. Insurers selling
individual and small-group
coverage would have to price
premiums on a communityrated basis. Purchasing pools
(APAs) open to all individuals
and groups.
19
21
Butler Proposal
Key Elements
Stuart M. Butler has outlined a new proposal to achieve near universal coverage
for health insurance that is built on the following key elements:
    available to working households would replace
the current tax exclusion accorded to employees. Employees could no
longer exclude from their taxable income the amount employers contribute on their behalf to health insurance premiums.
   sponsor coverage would have to use the credit
to get coverage at work. Other employees and individuals could use their
tax credit to buy coverage from a range of additional sources. These would
include plans offered by employers, association plans, and plans offered by
affinity groups such as churches, unions, and so on.
 , regardless of whether they sponsored coverage, would have
to undertake a “clearinghouse” function, which could include adjusting
employee tax withholdings to reflect their credit, creating an automatic
enrollment mechanism for insurance, and setting up a payroll deduction
and payment system for employees to pay their chosen plan.
      to supplement the tax credits, states
would have to develop a plan acceptable to the federal government to
make coverage more affordable for low-income workers.
22
About the Author
   . ,  .., is Vice President, Domestic
and Economic Policy Studies, at The Heritage
Foundation. Born and educated in Great Britain,
where he earned a Ph.D. in American economic history from St. Andrews University in Scotland, Butler
is considered one of Washington’s most innovative
thinkers. He was named one of Washington’s “dozen
key players” on health care reform by the National
Journal, and played a key role in the intellectual and
policy revolution that produced the  welfare
reform legislation. His books include, Enterprise
Zones: Greenlining the Inner Cities; Privatizing
Federal Spending; Out of the Poverty Trap (with
Anna Kondratas); and A National Health System
for America (with Edmund Haislmaier), which in
 presented a blueprint for reform based on
consumer choice and competition.
23
Reforming the Tax Treatment of Health Care to
Achieve Universal Coverage
by Stuart M. Butler
Overview
The aim of this proposal is near-universal health
insurance for working households, including the
self-employed, using a tax-based subsidy and insurance reform to make such insu ra n ce affordable.
Virtually all legal residents would be included in the
proposed system (other than those enrolled in government programs), including those who currently
have employer-sponsored health insurance. The
place of employment would con ti nue to be the
point at which subsidies typically are delivered and
choices made for most working people, but employers would no longer have to sponsor plans for
employees to receive a tax subsidy. The proposal
envisions a range of other organizations, from
churches to unions, that would supplement traditional employer-sponsored health insurance by
sponsoring health plans in tandem with an insurer
that carries the insurance risk. Self-employed individuals or those temporarily out of the workforce,
and employees of firms that do not sponsor insurance, could use the tax benefit to en ro ll in plans
offered by such organizations or by health insurers.
In the proposed system the tax exclusion for
employer-provided health insu ra n ce and other
health care tax benefits currently available would be
replaced with a new refundable tax credit based on
income and household health costs (both insurance
premiums and out-of-pocket costs). The design of
the tax credit would provide greater assistance to
lower-income families and less assistance to higherincome families than today’s system of tax relief. It
would also give much more assistance to those
households that face unusually high medical costs,
regardless of their income.
The proposal would achieve horizontal equity:
Households with the same income and medical
expenses would receive the same tax benefit, whether
they obtained coverage through their employer or
another organization, or they purchased their own
insurance and care. Low-income households could
choose between two forms of tax benefit: a refundable sliding-scale tax credit based on total health
expenditures as a proportion of family income, or a
flat credit that,if desired, could be assigned to a chosen health plan in return for a reduced premium.
Households above a specific income threshold could
use only the sliding-scale credit. The federal credits
could be supplemented with state subsidies, and the
federal government would encourage state subsidies
through a new federal grant to states.
Money for the new system of tax credits would
come from two sources: elimination of the existing
tax exclusion and other health tax breaks and general tax revenues. Americans generally are resistant to
sweeping change, so the proposal envisions a gradual transformation of the tax treatment of health
care costs, beginning with the introduction of limited refundable tax credits for those without employer-sponsored coverage or who are unable to afford
it. Thereafter, the tax exclusion would be reduced
gradually, in line with the general availability of a
more comprehensive system of tax credits (available
for out - of - pocket medical costs, health savings
accounts, and insurance).
Employers would be required to modify their
federal tax withholding procedures to reflect each
employee’s estimate of the health credit available to
him or her. Employers would also be required to set
up an escrow account to make payments, deducted
from employee compensation, to any plan chosen
24
by the employee that met the minimum government requirement. Working families would have to
en ro ll in at least a minimum, catastrophic health
insurance plan to be el i gi ble for the tax credit for
any health costs. They could choose any approved
plan in the area, unless their employer sponsored a
plan, in which case they would have to join it to
obtain tax relief. If an employee did not choose a
plan,he or she would be enrolled in a default plan or
program determined by the state.
The federal government would establish a
“default” system of health insurance regulation to
encourage the availability of affordable insu ra n ce
that could be purchased with the credit. The default
system would include modified community rating
for all plans with federally approved minimum benefits,and a change in federal law to create new forms
of group insurance plans.States could choose either
to adopt the federal default or to agree with the federal government on a functionally similar statedesigned rate regulation system. In addition, each
state and the federal government would have to
agree on a plan to eliminate uninsurance. The proposal envisions a market in which families obtain
insurance from organizations with which they are
affiliated, such as unions, churches, and similar
groups; from their employer if that employer
decides to sponsor coverage; from large insurers or
managed care plans; or from health plans sponsored
by large employers and offered to non-employees.
Coverage and Eligibility
Today’s tax code provides a number of tax benefits
for health care intended to help working families to
1
For a discussion of the tax treatment of health care, see Grace-Marie
Arnett (ed.). Empowering Health Care Consumers through Tax Reform.
Ann Arbor, MI: University of Michigan, 1999.
John Sheils and Paul Hogan. “Cost of Tax-Exempt Health Benefits in
1998.” Health Affairs 18 (2): 178.
2
Sheils and Hogan estimate the average value of all federal health tax
benefits for 1998 at $296 for families with incomes of $15,000 to
$19,999 and $2,357 for families with incomes of $100,000 or more
(Sheils and Hogan, p. 180). Although these averages exaggerate the
differential between lower- and upper-income families with employersponsored insurance (since a higher proportion of lower-income families
receive no tax benefits), the combination of lower-cost plans and a
lower tax bracket means a large inequity for insured families. Workers
without employer-sponsored insurance typically receive no tax benefits
3
obtain health care coverage. But the current system
places severe limits on who can obtain tax help.1 By
far the largest form of assistance is the exclusion
from an employee’s income of employer-paid health
insurance benefits. The value of all health care tax
benefits (including reductions in payroll and income
taxes) has been estimated for  at . billion at
the federal level and . billion at the state level.2
The design of the exclusion has been widely criticized as a highly inefficient and inequitable method
of helping working families to afford health care.3
Since the exclusion is for insurance only, it also
leaves lower-income employees vulnerable to high
out-of-pocket expenses without any tax relief.4 And
because the exclusion is only for employer-sponsored coverage, employees of firms without a health
plan cannot claim the tax benefit if they pay for their
own plan.5
The proposal would fundamentally change this
tax treatment,using the tax code to direct assistance
in a far more equitable and progressive way. It
would create a tax subsidy system available to all
working households of legal U.S. residents (other
than Medicaid and Medicare enrollees, and those
normally enrolled in state-sponsored programs).
The tax exclusion available to employees for
employer contributions to their health care plan or
expenses would be repealed, as would other health
care tax deductions available to taxpayers, and
replaced with a fully refundable tax credit available
to all individuals for health insurance and out-of
pocket medical expenses.6 The credit could be used
to help pay the employee cost of employer-sponsored benefits. Employers would continue to deduct
health payments as a cost of labor. All individuals
unless they are self-employed.
Many employers do offer tax-free Section 125 accounts that reduce
this problem. But Section 125 requires employees to agree to a payroll
deduction to fund the accounts, and unused balances in the accounts
revert to the employer at year’s end. This discourages most employees
from using the accounts to shield from taxation all but the most predictable out-of-pocket expenses.
4
Taxpayers can claim a tax deduction for unreimbursed health expenses,
but the taxpayer must itemize deductions and can deduct only amounts
that exceed 7.5 percent of income.
5
In the proposal, fully refundable means the individual or family would
be eligible for a subsidy from the federal government if the computed
credit exceeded the family’s federal income and payroll tax liability.
6
25
currently receiving tax relief for employer-sponsored plans would be eligible for the credit,as would
those working families without access to employersponsored coverage. Thus there would be no eligibility distinction among those receiving tax relief
for health care. As noted in the discussion below
regarding administration of the program, there
would be mechanisms to deliver the credit to those
who are moving between jobs or are unemployed.
To receive the tax credit,the individual or family
would have to purchase a health insurance plan that
included at least certain minimum benefits. With
the refundable tax subsidy available to all those not
currently in a government program, along with a
grant program for states to supplement the federal
credit and a federal-state compact to make affordable group insurance more available (discussed
below), the proposal would achieve near universal
coverage.
Structure of the New Subsidy
Lower-income households would have a ch oi ce
between two forms of tax credit—a fixed credit and
a sliding-scale credit. Other households would be
able to claim a sliding-scale credit.
Fixed Credit
Low-income families could claim a fixed credit for
health care expenses, provided the family obtained
at least the minimum insurance coverage. Eligible
expenses would include insurance, direct spending
on services, the employee cost of employer-sponsored plans,and contributions to accounts intended
to cover health costs, such as medical savings
accounts and employment-based flexible spending
accounts. The credit would be , per adult and
 per child, up to a maximum of , per family. This fixed-credit option would be available to
families with inc omes of up to ,, and to singles with incomes up to ,. There would be no
phase-out for this credit.
The fixed-credit option offers simple and predictable assistance for lower-income working families,although in most cases the family would receive
more assistance by choosing the sliding-scale credit.
It would be available only for families below certain
income thresholds. The reason for this is that a fixed
credit for all households would exceed the value of
the sliding-scale credit and the value of tod ay ’s
exclusion for most upper-income families, while the
purpose of this approach is to concentrate health
tax benefits on those families facing the greatest difficulty in affording health care.
An individual or family eligible for the fixed tax
credit could choose to assign it to a chosen health
plan in return for a commensurate premium reduction. In this case the insurance company would
adjust its own tax payments to the federal government to reflect the fixed-credit amount, while for
tax purposes the enrollee would be deemed to have
received the credit. This could be arranged through
the workplace, as noted below, or directly with the
health plan for self-employed or temporarily unemployed individuals.
Of course, many families would not be certain
whether their income would be below the thresholds for the fixed credit, but this is less of a practical
concern than it might at first appear. As explained
below, the place of employment normally would be
the means through which plan choices, payments,
and tax benefits would be channeled. The employer
would inform the employee whether the basic wages
or salary of the employee would make his or her
family eligible for a fixed credit. If the employee discovered during the year that his or her income
exceeds the threshold and so becomes ineligible for
the fixed credit (because of a salary increase or overtime pay, for example), the employee would not
actually have to cancel the assigned credit with his
or her health plan, but simply factor the amount
into his or her adjusted withholdings or end-of-year
tax retu rn . In a few cases, some particular workers
with rising incomes might receive slightly too much
money under the fixed credit. But requiring individuals who just exceed the eligible threshold to calculate a sliding scale and return the difference would
not be worthwhile. It is not necessary for advance
payments made through the tax code to be reconciled perfectly at the end of the tax year, only that
they be reasonably close (as a comparison, the standard deduction is a loose reconciliation that may
26
shortchange the taxpayer or the IRS occasionally,
but has the advantage of simplicity).7
The Sliding-Scale Credit
Both high- and low-income households would have
the option to claim a fully refundable sliding-scale
credit based on health expenses as a proportion of
total income.8 Different families would qualify for a
different credit amount for a year’s health costs,
depending on their incomes, much like the child
care credit in the tax code. Expenses eligible for the
credit would be the same as in the fixed-credit
opti on . As in that opti on , health expenses would
include premium costs and out-of-pocket expenses.
It would be calculated as follows, subject to a maximum credit of , per year for families and
, for individuals: 9
Structure of the Sliding-Scale Credit Option
Health costs up to %
of adjusted gross income (AGI)
% credit
Health costs between
% and % of AGI
% credit
Health costs above %
of AGI
% credit
An income-related sliding-scale option would be
more complicated than the fixed-credit option
(although many families use such a sliding-scale
credit for the cost of child care), but it typically
would mean a larger subsidy, especially for those
whose high costs accounted for a large proportion of
their income. It can also be designed to avoid major
tax changes for middle-income Americans who
would lose their current tax exclusion. In addition,
7
For a brief discussion of the tax reconciliation issue, see Linda J.
Blumberg. “Expanding Insurance Coverage: Are Tax Credits the Right
Tack to Take?” Unpublished paper, The Urban Institute, Washington, DC,
August 12, 1999, p. 18.
The design of the proposed credit follows consideration and evaluation
of a number of tax credit proposals, including various fixed-credit and
percentage-credit as well as earlier sliding-scale credit proposals by the
author and his colleagues at The Heritage Foundation (one of which
formed the basis of legislation, S 1743, HR 3698, introduced in 1993 by
Sen. Don Nickles [R-OK] and Sen. Cliff Stearns [R-FL]). See Stuart Butler.
“A Tax Reform Strategy to Deal with the Uninsured.” Journal of the
8
the sliding-scale credit provides some age and geographic adjustment, because families in areas with
relatively high insurance and medical service costs—
or other workers facing higher costs—would be able
to claim a larger federal credit. The dollar value of
the credit also would rise over time in proportion to
medical costs. Moreover, for many families, the credit could be estimated very accurately, because heavy
medical expenses are not necessarily unpredictable.
For instance, a family with chronic medical problems
may pay the full stop-loss amount routinely each
year and be able to project its out-of-pocket costs.
Families below the income thresholds for the
fixed credit would have the option of claiming
whichever credit provided them with the most
money. Those above the thresholds could claim the
sliding-scale credit only. The thresholds are calculated to be just below the “break even” point for the
great majority of families, meaning that the value of
the fixed credit at the thresholds typically would be
only slightly less than the value of claiming the sliding-scale credit. In most instances this avoids a
sharp “cliff effect,” in which a small rise in income
means a large drop in the value of the credit.
State Subsidies
The tax credit proposal is designed to be as compatible as possible with existing state programs, such as
Medicaid, the State Children’s Health In su ra n ce
Program (S-CHIP), and high-risk pools. It is not
designed to replace them.
New Federal Grant
Under this proposal, the federal government would
provide $ bi ll i on annually to the states to assist
them in su pp l em en ting the federal tax credit for
health-related expenses. In addition, each state
American Medical Association 265 (19) (May 15, 1991); Stuart Butler
and Edmund Haislmaier. “The Consumer Choice Health Security Act.”
Issue Bulletin 186, December 1993; Stuart Butler. Expanding Health
Insurance through Tax Reform. Menlo Park, CA: The Henry J. Kaiser
Family Foundation, Kaiser Project on Incremental Health Reform,
October 1999. A combined fixed credit/sliding-scale credit similar to the
current proposal was introduced in July 2000 (HR 4925) by Rep. John
Cooksey (R-LA).
A limit is placed on the credit on the assumption that families facing
extremely high health costs are helped best by a combination of a federal credit and other means, rather than solely by a formula credit.
9
27
would receive a grant amounting to the estimated
federal taxes raised in the state from individuals
who do not enroll in a minimum health plan and
thus cannot claim the federal credit.
State supplements would be especially important
for low-income workers, such as most of those currently leaving welfare,and, when combined with the
federal credit, could enable the worker to afford a
reasonable level of health insurance. States are
required by the  welfare reform legislation to
provide Medicaid coverage to certain families even
though they do not receive cash assistance, and many
states also take advantage of the flexibility under the
legislation to continue Medicaid coverage for other
working families. But  data indicated that about
 percent of women who had left welfare since 
were uninsured.10 The federal credit, especially if
combined with state assistance, would provide significant assistance to these low-income families,
increasing their potential to afford the out-of-pocket
cost of employer-sponsored insu ra n ce ,i f offered, or
to purchase other coverage and services. With the
federal credit and grant available, states would have
greater flexibility to design the best option to deal
with various groups of working families, such as
combining the federal credit with S-CHIP or other
subsidies, and taking steps to make insurance less
expensive for workers with the tax credit.
exclusion for employer-provided insurance would
have led to an increase in federal tax revenues in
 of . billion11 and a projected . billion
in .12 Of this, approximately one-third would be
extra payroll tax revenue (. billion in Social
Security tax and . billion in Medicare Hospital
Insurance tax in ).
While the state grant in the proposal would be a
discretionary spending item, the federal credit would
be a tax entitlement and not subject to the budget
limitations that apply to discretionary programs. In
other words, Congress would not have to vote for a
specific budget amount each year for the credit. The
net cost of the tax reform and new grant is envisioned
to be between  billion and  billion annually.
    
  
Approximately one-third of the tax revenue from
the increase in taxable compensation resulting from
the reform would be in the form of Social Security
and Medicare taxes, and would not be available to
fund a new health care credit. On the other hand,
because this earmarked money would go into the
retirement trust funds, there would be a reduction
in the future amount of general revenue support
needed to fund future retirement benefits. Thus, the
equivalent amount of future general revenue could
be allocated to help fund the tax credit.
Financing
While a simu l a ti on of the proposal has not been
undertaken, analyses of other similar proposals
reveal its potential impact on government and
household finances and on coverage.
Effect on Government Finances
  
According to estimates by the Lewin Group, eliminating various health care deductions and the tax
10
Bowen Garrett and John Holahan. “Health Insurance Coverage after
Welfare.” Health Affairs 19, (January-February 2000): 177.
John Sheils, Paul Hogan, and Randall Haught. Health Insurance and
Taxes. Washington: The National Coalition on Health Care, 1999, p. 47.
This figure assumes that employer contributions to employee retirement
health plans remain tax-free.
11

State finances would be affected by the proposal in
four ways .F i rs t , those states with an income tax mirroring the federal code initially would realize a windfall increase in tax revenues, assuming they did not
change their tax rules. The proposal assumes this
m on ey is retu rn ed to state taxpayers. Second, by
reducing uninsurance, the new federal credit would
substantially reduce the burden on states of subsidizing hospitals and physicians for uncompensated
Letter from John Sheils, the Lewin Group, to Robert Moffit, The
Heritage Foundation, dated July 6, 2000, assessing the impact of legislation (HR 4925) sponsored by Rep. John Cooksey.
12
28
The proposed tax credit structure is designed,
on average, to leave most middle-income families
with little change in their tax liability.
care. Currently Medicare and Medicaid disproportionate share hospital (DSH) payments amount
to approximately  bi ll i on , so if the federal program were not reduced, the states would have a
windfall gain. Third,each state would receive a grant
based on the federal government’s estimated savings
due to individuals in the state who do not obtain the
minimum basic insurance and thus cannot claim the
federal credit. And fourth, the federal government
would provide states with a total of  billion to supplement the federal credit and reduce uninsurance.

The objective of the tax credit proposal is to make a
reasonable level of health care insurance and services affordable to all working households without
incurring strong opposition because of a large
increase in the explicit tax liability of any other
income group. Under the proposal, the net income
of a family after taxes and health expenses would be
affected by four factors. First, the family’s taxable
compensation would rise according to the size (if
any) of the employer contribution to the worker’s
health plan. Second, the family’s tax bill initially
would rise because of this increase in taxable compensation. Third, the family might increase or
decrease its health care expenses, depending on its
preferences and whether it could cash out the
employer contributi on . And fourth, the family
would qualify for one or both of the refundable tax
credit options. The bottom line for the family would
be the net effect of these four factors.
The proposed tax credit structure is designed,on
average, to leave most middle-income families with
little change in their tax liability. Upper-income
households with high medical expenses also would
not see a major change in their taxes. And many of
those households that would pay significantly higher taxes would do so because they took more of their
compensation in cash rather than health benefits—
receiving, in effect, a pay raise. Lower-income families would be as much as , better off in federal
tax benefits even if they decided to opt for the simple flat credit. But lower-income families with
above-average total medical expenses could receive
a larger amount of tax assistance through the
refundable sliding-scale credit. For example,a family of four with , in income and , in
health spending in any one year would qualify for a
federal credit of approximately ,. But this family also would qualify for supplementary assistance
financed by the federal grant to states as w ell as any
other assistance the state provided.
Administration of the Program:
A New Role for Employers
The place of employment is a particularly convenient and efficient venue through which to make
insurance payments and handle other transactions
(such as collecting federal taxes). The proposal envisions employers as the key clearinghouse for plan
choices, tax adjustments, and payments associated
with health care. But, unlike today, it would not
require employers to organize or sponsor a plan, or
make any contribution to the cost of coverage, for
the employee to obtain tax relief. As noted below, for
those heads of household who do not work for an
employer, such as the self-employed and those
between jobs, other mechanisms would apply for
receiving credits and paying premiums.
Unless the employer sponsored a plan, employees would choose their own plans,and could change
jobs without changing plans. Moreover, the tax benefits would no longer depend on insurance decisions made by the employer.
All employers would be required to undertake
two key functions—delivering the tax credit to
workers and paying premiums through payroll
deduction.
29
Delivering the Tax Credit to Workers
Employers would be required to inform the Internal
Revenue Service (IRS) and the state which health
plan (or state-sponsored program) each employee
had selected, and to adjust the employee’s tax withholding to reflect the estimated value of the credit.
This employment-based selection and financing
mechanism has been suggested by a number of
analysts.13
Withholding adjustments is the simplest way for
most workers to obtain a tax benefit prorated each
pay period, so they would not have to wait until the
end of the tax year to receive their subsidy. For
employees who elect to have their credit assigned to
a health plan in return for a reduced premium, the
tax withholding would reflect receipt of the credit.
Since assignment would necessitate informing the
IRS (via the employer) of the employee’s choice,the
agency could decide whether to audit an individual
who did not file a tax return. The great majority of
households that do not have to file a return would
fall under the maximum incomes eligible for claiming the fixed credit.
It should be noted that all employed Americans
(other than those in public programs) would be in
the tax/premium withholding system. Thus, there
would be no obvious distinction between upperincome workers receiving a small credit, and lowerpaid workers receiving more assistance through a
refundable credit. Only the withholding amount
would vary, as it does for employees today. This system would mean no separate arrangements or stigma associated with the program.
Paying Premiums through Payroll Deductions
The second legal obligation on employers would be
to institute an automatic payroll deduction system
for health insurance premium payments,structured
much like the flexible spending plans many employers now maintain vo lu n t a ri ly. Once the employee
selects a health plan and indicates the employer’s
13
For example, see Lynn Etheredge. Tax Credits for Uninsured Workers.
Paper prepared for the Health Insurance Reform Project of George
Washington University, September 1999.
14
Employers would not be required to accept amounts for out-of-pocket
premium cost, the employer would be required to
deduct a specific amount each pay period and place
that amount in an escrow account. The amount
would have to be enough to pay the premium, but
also could include additional amounts, as today, to
pay predictable out-of-pocket costs.14 Since the eligible tax credit also would be made available at each
pay period, the employee would have the necessary
subsidy available for the payroll deduction. If the
employee did not select a plan voluntarily, the
employer would assign that employee to a default
plan or government program selected by the state
and make a default payroll deduction accordingly.
Selecting a default plan or program,as noted below,
would be part of an agreement between the state
and the federal government to achieve maximum
coverage.
The employer would be responsible for providing
new employees with information from the government explaining the tax credit and payroll deduction
system. Employees would have to sign a document
stating they understood the system and indicating
the plan in which they wished to participate. If they
did not do so, the default plan assignment and payroll deduction would go into effect until the employer received information that the worker was enrolled
in a state health plan or in an insurance plan elsewhere (such as in a spouse’s plan).
There is good reason to believe that this mechanism would be efficient administratively and would
lead to a high level of employee enrollment. Lynn
Etheredge has proposed automatic workplace
enrollment for a tax credit system and estimates that
the administrative cost of insurance using such a
system could be . percent, compared with administrative costs several times that for individual and
non-employment small groups.15 In addition, evidence from savings plans suggests that an automatic
enrollment system for health insurance could have
dramatic effects on sign-up rates. Brigitte Madrian
and Dennis Shea have found that a workplace-based
costs, though employers currently offering flexible spending accounts
probably would do so.
15
Etheredge, 1999, p. 6.
30
automatic en ro ll m ent system for (k) plans—
where, to be excluded, the employee must actively
decline to be included—boosted participation rates
from  percent to  percent for such voluntary
pensions, with even sharper rises for young and
lower-paid employees (for employees with incomes
below ,, the rate increased from . percent
to . percent).16 In the health system proposed
here, of course, individuals not actively making a
ch oi ce would be assigned a plan or en ro ll ed in a
state health program.
This payment system is also very similar to the
way in which the Federal Employees Health Benefits
Program (FEHBP) enables a federal worker who
may work in a small office,such as that of a member
of Congress, to choose from dozens of plans. In the
FEHBP, the worker tells the employer which plan he
or she has chosen, but the payment details are handled by the Office of Personnel Management
(OPM), which for this purpose functions like a payroll processing firm for the individual’s immediate
employer. OPM functions as a clearinghouse,
deducting premiums from each federal employee,
pooling the money, and making payments to each
health plan based on the total number of its government enrollees.
Estimating the Cash Value of Employer-Paid
Health Benefits
Employer contributions to the employee’s health
plan would be considered taxable income in the first
instance, but also would be considered employee
expenditures on health for purposes of calculating
the credit. This raises the question of how to calculate the per capita value of a group health plan for
tax purposes. While this cannot be done perfectly—
and, of course, the current system hides large tax
benefit inequities—it can be accomplished accurately enough in one of two ways adopted in the
 Nickles-Stearns legislation. One option is for
firms to negotiate the cash value with their employees. This probably would be the preferred option in
16
Brigitte Madrian and Dennis Shea. The Power of Suggestion: Inertia in
401(k) Participation and Savings Behavior. National Bureau of Economic
Research Working Paper No. 7682, May 2000, p. 51.
unionized firms with a benefits contract or where an
employer makes a defined contribution to an
employee’s health plan. If the firm and its employees
did not choose this option, a fallback formula could
be developed by the state or federal government.
The best such formula might be a structure of relative values for various categories of household composition and risk (such as family structure, sex, age,
etc.) that matches the categories used in underwriting restrictions placed on plan premiums by the federal government (for plans covered by the Employee
Retirement Income Security Act, or ERISA) or state
governments (for plans regulated by the state). In
this way the assessed value for categories of workers
would reflect the relative premium costs of coverage
for these risk categories in the state.
This same method of calculation would be use d
when a firm and its employees chose to end a sponsored health plan in favor of turning that fringe
benefit into cash income that employees could use
to enroll in other plans. A temporary “maintenance
of effort” requirement could be applied, so that in
the first year after ending a plan the employer would
be required by law to add the plan’s value to paychecks, making the full compensation amount
explicit as the basis for future pay levels.
If an employee had chosen to be enrolled in
family coverage obtained by a spouse working for
another firm, and thus did not receive an employer
contribution and was not part of the employer’s
insurance group, there would be no taxable employer contributi on . As today, it would be up to the
employee (or his or her union) and the employer to
decide if these employees received a taxable supplement to income in lieu of the contribution. If the
firm and its employees chose to dismantle their
existing plan to permit employees to choose other
plans, the cash-out value for workers would be calculated the same way.
Employer-Sponsored Plans
This proposal envisions that those employers that
wish to sponsor insu ra n ce themselves (arranging
plans for their employees or self-insuring under
ERISA) could continue to do so. If a firm decided to
do that, its employees would have to use the plans
31
organized by the firm as their primary insurance
under the rules specified by the employer if they
wished to claim the federal tax credit. Thus, to
maintain stability in the insurance pool, employees
of such firms would not have the right to opt out of
the employer’s plan. These employers would still be
required to arrange for a payroll deduction and
adjust withholdings, as most already do.
Many smaller employers today wish to contribute to their employees’ health care, yet they face
organizational burdens and high administrative
costs in providing insurance themselves or they cannot provide affordable coverage. 17 Those that do
contribute by sponsoring plans often do so only
because of the design of the tax code. The tax features of the proposal would give many of these firms
the attractive option of making a defined contribution to an employee-chosen plan not sponsored by
the employer. To be sure, some firms now sponsoring insurance would decide—in most cases with the
support of their employees—to end plan sponsorship and switch to a defined contribution. But that
form of “crowding out” would be more efficient and
beneficial to employees, since it would entail more
choices. Less desirable instances of crowding out
that might destabilize a company’s risk pool would
be reduced in two ways. First, the proposal requires
employees to enroll in a company-sponsored plan if
it is provided to all employees. Second, employees
have the right to use their credit to offset the out-ofpocket costs of coverage for themselves or their
dependents, which would make employer-sponsored coverage more affordable for many lowerincome workers who decline it today.
A New Opportunity for Large Corporate Plans
This tax proposal would remove the current tax barrier to large corpora ti ons marketing their health
plans widely to non-employees. It is common for
17
In a recent survey, for instance, 27 percent of small employers (with
fewer than 50 employees) offering dependent coverage reported that
their employees declined it because of the cost (EBRI Issue Brief, no. 226,
October 2000).
As part of its strategy to improve and coordinate care for its own
workforce, Deere created its own HMO. To make the most efficient use
of its new health facilities, the company then contracted with other
18
large firms to take products initially developed as an
internal service to the firm and market them to
external customers, thereby deriving revenues from
what had previously been an overhead cost for the
f i rm . For example, G en eral Motors formed the
General Motors Acceptance Corporation (GMAC)
out of its huge automobile loan service developed to
help sell its cars. GMAC has now branched out into
a broad range of financial services, including home
mortgages, because the tax system does not deny the
mortgage interest deduction to someone obtaining
a mortgage from a car company. But only a few large
companies have explored marketing their health
plans to non-employees, most notably John Deere.18
The employees of firms contracting with the Deere
plan are still in an employer-sponsored plan,so they
qualify for the tax exclusion. But the tax code does
not give tax relief to individuals or non-employment groups signing up for the plan, and this has
discouraged Deere and other companies from offering such coverage. The proposed tax credit would
remove this obstacle, opening up a potentially large
new market for existing corporate plans and an
opportunity for many working families to obtain
coverage under these plans.
Incentive for Employer Contributions
Under this proposal, the employer would continue
to be the link to health coverage, and the employee
would be obligated to enroll in a plan, so the proposal would not mean a reduction in employer
involvement. Moreover, with this new tax credit for
non-employer-sponsored coverage, there would be
an incentive for many firms that do not do so today
to make a financial contribution to insurance, since
they could do so without the burden of sponsoring
insurance, while still enabling their employees to
enjoy the same tax benefits that would apply to
sponsored insu ra n ce . Thus, while the tax conse-
firms to enroll their employees in the Deere plan. The Deere plan is now
offered to federal employees in some areas under the FEHBP, as well,
and to some Medicare and Medicaid beneficiaries. In fact, out of more
than 400,000 individuals enrolled in the Deere plan, less than 20 percent
are John Deere employees. See Stuart Butler. Transcending EmploymentBased Health Insurance. Council on the Economic Impact of Health
System Change, 1999, conference paper available at http://sihp.brandeis.edu/council/pubs/Butlertx.pdf.
32
qu en ces of an individual obtaining insurance
through an employer or any other source would
cause some employers to close down their health
plan and convert the benefit to cash income, those
same consequ en ces could induce these and many
other employers to make a contribution to an “outside” health plan selected by the employee.
Self-Employment and Transitions
The credit would be adaptable for working-age individuals who are self-employed and for individuals
either tempora ri ly not working or leaving other
health programs. For example, self-employed individuals would furnish the IRS with evidence of
insurance and make appropriate adjustments to
their estimated tax payments. If a worker chose to
remain in his or her former employer’s sponsored
plan under the Consolidated Omnibus Budget
Reconciliation Act (COBRA), the credit would
apply to the cost of coverage,as it would to any normal medical cost. If such a worker were selfemployed, or worked for a firm that did not offer
insurance, he would recover the credit for COBRA
coverage through quarterly tax payments or withholding.
An unemployed person with an assigned credit
similarly would face a reduced premium. If the person did not qualify for or choose the assignment
option,he or she could obtain the value of the credit as an ad ju s tm ent to his or her unemployment
compensation. The tax credit for unemployed
workers could be paid through the unemployment
insurance system. This would require a funds transfer between the Treasury and the Department of
Labor, with the money then distributed to state
unemployment offices (similar to the supplemental
benefit programs delivered in this way since ).
The state unemployment offices also could take on
responsibility for remitting premium payments to
insurers. Early retirees would also be eligible for the
credit, while the value of their health benefits paid
by their previous employer would be taxable,unless
converted into a cash contribution and shielded
from taxation in some other way.
Administering the Program:
Working with States to Make Insurance
More Available
Reforming the federal tax subsidy system to channel
more assistance to those who need help to afford
coverage and care is only half the equ a ti on . The
other half is ensuring that attractive and affordable
health plans are available to individuals and families. In theory a credit can work in an insurance system that charges for its services according to a
market assessment of a person’s medical servi ce s
and insurance risk. But this would require huge and
carefully designed subsidies to certain individuals,
and would be impractical.
A more practical approach has two elements. The
first is to develop a subsidy system ,l i ke the one above,
designed to deal with the great majority of individuals based on their income and medical needs. The
second is for the federal government to work with
states to make sure that new and affordable plans are
available to families within a stable insurance market,
and that plans are affordable for people who remain
high-cost “outliers” despite the subsidy system.
This proposal seeks to do this through a federalstate partnership aimed at making available new
kinds of group health plans at reasonable prices. To
achieve this, the federal government would have to
do three things. First, it would amend laws to create
opportunities for new kinds of group plans. Second,
it would make available grants to states under certain conditions to enable the states to reduce uninsurance by supplementing the federal refundable
tax credit. And,third,the federal government would
negotiate with states, during which a state could
accept “default” federal insurance regulation and
federally sponsored insu ra n ce groups, or it could
implement, with federal approval, an alternative
insurance arrangement designed to achieve the federal goals of reasonably priced health plans at reasonable prices and a reduction in uninsurance.
Potential New Insurance Groups
   
One of employment-based coverage’s central claims
is that it is an effective way of forming relatively sta-
33
ble groups for insu ra n ce purposes. This is true to
some extent, especially for larger firms, in part
because the heavy tax bias today in favor of employer-sponsored insurance makes a firm’s workforce in
effect a “captive” group. But the benefits of employer-formed groups have been exaggerated ,p a rticularly for small employers, and they come at considerable cost in terms of reduced ch oi ce ,“ j ob lock,” and
other side effects. 19 Indeed, the problems of small
employer-based groups have forced many states and
insurers to create new groupings to achieve greater
insurance stability, spread risk, and provide affordable premiums regardless of health status.
This proposal thus does not envision individual
insurance as the principal alternative to tod ay ’s
employment-based coverage. For one thing, forming
people into groups for insurance purposes achieves
administrative economies and is a means of spreading risk. But, in addition, an intermediary institution
can negotiate with insurers or providers in the interests of group members. This intermediary function
is necessary to achieve efficiency and satisfaction in a
market where the consumer is not typically a sophisticated buyer. S ti ll , to achieve satisfaction and efficiency for the consumer, the goals of the intermediary must coincide with those of the consumer. For
such reasons this proposal envisions the encouragement of other kinds of groups,particularly for individuals who do not have access to employer-sponsored insurance or for whom a group centered on
the place of employment is not ideal.
Four types of groups are particularly attractive alternatives to traditional employment-based
covera ge :
Affinity Groups. Several common institutions in
American communities are well placed to serve this
function for insu ra n ce . For example, unions as
“f ri en dly societies” have had a long history of
involvement in health care in the United States and
elsewhere. Many unions are active in the organization of multi-employer health plans under the Taft-
Ha rt l ey Act. Union plans also flourish in the
FEHBP, in some cases offering associate membership to non-union members.
Many religious denominations also have a long
history of providing insurance services for their congregations.20 For lower-income African Americans
and others, the church is a far more stable institution
in the community than loc a l ,s m a ll employers, and
one that has the long-term social welfare of families
firmly in mind. In addition, the Catholic Church and
other denominations sponsor networks of hospitals.
Since churches, like unions and many other groups,
routinely communicate with their members by mail,
these intermediaries also present a lower-cost
“piggyback” means of marketing health plans and
reducing administrative costs.
This proposal does not envision these alternative
groups acting as insurers themselves, but instead as
buying agents that reach agreements with insurance
plans that actually shoulder the risk. These organizations would form the group of purchasers and
receive a fee from the insurer for performing marketing and management functions. This is the
arrangement used by such organizations operating
in the FEHBP.
Other affinity g roups, such as farm bureaus and
professional associations, exist in part to negotiate
insurance packages for their members, but such
groups face significant limitations that restrict the
role they can play. A major impediment is the tax
law, which denies tax benefits to most people who
obtain coverage through groups not closely associated with their employment. (Where that impediment does not apply, such as to union plans in the
FEHBP, alternative groups play a major role.) In
addition,state-level benefit mandates and insurance
regulations discourage many affinity groups from
offering health insurance.
Associations. Various employment-related associations have arisen to group people together to
obtain insurance without the employer directly
19
Uwe Reinhardt. “Employer-Based Health Insurance: A Balance Sheet.”
Health Affairs 18 (6): 124–32.
the 20th century, but were forced out of existence largely by tax-subsidized employer plans and, later, by Medicaid; see David Beito. From
Mutual Aid to Welfare State. Chapel Hill, NC: University of North
Carolina, 2000, chs. 9, 10.
Fraternal organizations, many of them church-affiliated, were a major
source of health insurance (and even capitated health plans) earlier in
20
34
sponsoring coverage. These include health purchasing cooperatives and coalitions (HPCs) and multiple-employer welfare arrangements (MEWAs), and
they also face restrictions at the state level that affect
their insurance arrangement and benefits.
There have been proposals in recent years at the
federal level intended to create new kinds of associations that would be free from many state restrictions, particularly state benefit mandates. The most
important of these are HealthMarts and association
health plans. While these proposals do raise concerns about their potentially disruptive impact on
existing state efforts to devise affordable insurance
groups because of possible risk selection,they could
be vehicles for lower-cost group coverage.21
Federal Em pl oyees Health Benefits Program
(FEHBP). While technically an employment-based
system, the FEHBP actually serves the equivalent of
a small country (with nearly  million covered
individuals) and offers a broad choice of plans.
While a federal worker’s immediate employer does
not sponsor plans, the place of employment is still
the “entry point” for selecting plans, much like the
process envisioned in this proposal. FEHBP plans
are regulated at the federal level, through a combination of general statutory and administrative regulation supplemented by a process of nego ti a ti on
between the Office of Personnel Management, on
behalf of the federal government,and plans wishing
to market through the FEHBP. Premiums are community rated, but costs and benefits vary widely.22
The FEHBP operates parallel to the systems for
workers outside the federal government. There have
been several proposals to open up the FEHBP to
non-federal workers under various conditions, typically using a separate insurance pool.23
Large-Employer Plans. As noted earlier, the current tax laws discriminate against large corporations
offering their health plan coverage to non-employees. Another major limitation on this opportunity is
A Federal “Default” System of Insurance Regulation
This proposal recognizes that stimulating the creation of new forms of group coverage not sponsored
by employers involves two challenges. The first is to
create an environment in which such alternatives
are fostered, principally by removing any barriers to
such plans. The second is to tackle the concern that
such groups would be disrupted by the pressures of
adverse selection.
Under the proposal the federal government
would enact legislation to help make new forms of
group insurance more widely available and affordable in each state. Each state would have the choice
of allowing this federal structure to supplement or
replace its own system of regulation, or proposing
its own package of changes to achieve the same
objectives as the federal structure, perhaps incorporating some of the federal steps. The federal government then would negotiate with the state on the
final regulatory arrangement, withholding the proposed grant and applying some or all of the federal
set of changes if there was not complete agreement.
The federal legislation would address four areas:
. Premium regulation. ERISA and the Health
Insurance Portability and Accountability Act
(HIPAA) would be amended to include limitations
on underwriting and to establish a minimum set of
benefits. Currently HIPAA establishes certain minimum protections for enrollees in federal- and stateregulated plans, including those under ERISA. It
establishes a minimum,which states can exceed,and
it allows states to propose alternatives for certain
provisions that can be accepted as sufficient by the
federal government.
21
For a discussion of the issues associated with such proposed new associations, see Elliot Wicks and Jack Meyer. Small Employer Health
Insurance Purchasing Arrangements: Can They Expand Coverage?
Washington: National Coalition on Health Care, 1999.
25–39; Stuart Butler and Robert Moffit. “The FEHBP as a Model for a
New Medicare Program.” Health Affairs 14 (4); Craig Caplan and Lisa
Foley. Structuring Health Care Benefits: A Comparison of Medicare and
the FEHBP. Washington, DC, AARP Public Policy Institute, May 2000.
For a description of the FEHBP, see Harry Cain. “Moving Medicare to
the FEHBP, or How to Make an Elephant Fly.” Health Affairs 18 (4):
23
22
that the Employee Retirement Income Security Act
(ERISA) does not protect such plans from state regulation unless they are made available to the
employees of other ERISA-protected firms, or
through federal programs such as the FEHBP.
For instance, Senator Bill Bradley made such a proposal in his presidential campaign.
35
Federal law would be amended to encourage new types of
group coverage. In particular, large affinity groups and associations
would be permitted and protected from state regulation.
The proposed federal legislation would amend
ERISA and HIPAA to require a minimum set of
benefits in all plans for which the new federal tax
credit could be used. Unless states negotiated alternative arrangements, they could not prevent plans
containing these minimum benefits from being
marketed, provided the plans met other applicable
state and federal requirements. In addition, ERISA
and HIPAA would be amended to require that all
plans, including plans in the individual market,
limit underwriting so that premiums for any particular plan option could vary at most only by age, sex,
geography, and family type, and not by such factors
as health status. As noted, this would go into effect
unless a state negotiated an alternative arrangement
with the federal government. Plans could place
more limits on premium variation either voluntarily or under state regulation. The narrower premium
variations typical in ERISA plans—chiefly by family
type—would meet that requirement.
. New types of group plans. Federal law would be
amended to encourage new types of group coverage.
In particular, large affinity groups and associations
would be permitted and protected from state regulation. Federal law governing the FEHBP would be
amended to permit a separate insurance pool for
non-federal employees. Plans currently available in
the FEHBP would be allowed to market to the new
pool, if they wished, and other plans could market
exclusively to the new pool, provided they met the
general requirements of the FEHBP. While the current FEHBP is community rated, the premiums of
plans offered to non-federal workers in particular
would have to comply with the federal underwriting
requirements specified above or a rating system
agreed to between state and federal government. In
addition, the plans offered to non-employees by
ERISA-regulated companies would be regulated
under ERISA rather than under state law.
. A new grant program for states. Congress would
enact a new grant program, discussed earlier, for
states to supplement the federal tax credit to achieve
near-universal affordable health care coverage. The
grant program would contain two sets of funds:One
would be based on the estimated value of the federal
tax benefits unclaimed by those who did not purchase minimum insurance, and the other would be a
$ bi ll i on grant each year available to states on the
basis of a federally approved state plan to reduce
uninsurance according to agreed-on goals by making choices of affordable plans available in the state.
. A federal-state compact. At least some of the
federal reforms would go into effect unless a state
proposed an agreeable alternative arrangement likely
to achieve the same ultimate goal of affordable coverage for families eligible for the federal tax credit. A
state might propose alternative premium rating
requirements for non-ERISA plans and benefit
requirements that made lower-cost plans widely
available in the state. It could create a high-risk pool,
and so reduce the need for underwriting restrictions.
Furthermore, it could propose the use of state programs, including direct services through clinics and
other facilities, to accomplish the equivalent of insurance coverage for some types of households. The
state might also propose a modification of the federal law to permit the FEHBP and association plans to
be made available in the state to integrate these new
alternatives into its current insurance system.
New group plans and premium restrictions
would raise concerns about adverse selection. For this
reason the proposal also envisions the federal government working with states to experiment with new
ways to limit this problem. This might take the form
of high-risk pools and/or a reinsurance market. In
reinsurance the health plans themselves buy insurance against ending up with an unusual portfolio of
risk. In ad d i ti on ,s t a tes and the federal government
could experiment with retrospective risk-adjustment
pools in which plans pay amounts into the pool based
36
on their enrollment and receive money from the pool
based on the actual paid claims of their enrollees.24 An
additional federally chartered national reinsurance
pool or set of regional reinsurance pools could act as
a final reinsurer between the cooperatives.
A share of the $ billion federal grant would be
available to states on reaching agreement with the
federal government. The amount would be based on
the costs of insurance and services still faced by
lower-income households after they had received
the federal credit. For these families, the grant could
be used to reduce the cost of coverage, such as by
subsidizing high-risk pools or by supplementing the
federal credit. This proposal envisions states receiving a bonus if they achieve the objectives in the
agreement, and losing some funds if they do not.
This federal-state compact approach is similar to
the approach taken in some provisions of HIPAA
and the  welfare reform legislation. In each case
states could propose alternative means of achieving
the federal objective,thereby avoiding certain provisions of the new federal law. In the welfare reform
legislation, goals were set for reductions in welfare
dependency, with bonuses available to those states
that meet or exceed the agreed-on goal.
An additional obligation of a state, as a condition
of receiving the federal grant, would be to identify a
fallback private health plan and/or government program for any employed individuals (and their
dependents) who did not select a specific plan at
their place of work. Employers would enroll automatically in the plan those employees who did not
choose an alternative plan containing the minimum
required coverage.
Are Institutions Capable of Carrying Out
their New Administrative Functions?
The federal and state governments, as well as the
private sector, would have a number of administrative obligations under the new program.
For a review of the issues involved in blended approaches to risk
adjustment, see Linda J. Blumberg and Len M. Nichols. “Health
Insurance Market Reforms: What They Can and Cannot Do.” Urban
Institute, 1998. See also Joseph Newhouse. “Risk Adjustment: Where
Are We Now?” Inquiry (Summer 1998).
24
The Federal Government
Administration of the proposed program would be
shared by federal and state governments, but with
key roles for employers and insurers. The Treasury,
through the Internal Revenue Servi ce , would
administer the federal tax credit. For the vast majority of individuals, the vehicle for distribution of the
credit would be the normal withholding process by
the employer, with the amounts reconciled in tax
returns. Some small modification of the worksheets
for withholding and tax returns would be needed.
The Treasury would need to monitor individuals
assigning their fixed credit to a health plan,especially those who did not file income tax returns, but this
would not place significant additional requirements
on the Treasury.
There would need to be coordination between
the Treasury and the Department of Labor to deliver the value of the credit to unemployed individuals
through the state unemployment insurance system.
If premium payments also were handled through
unemployment offices, the Labor Department
would have to remit credit and premium payments
to an individual’s chosen plan. This would be a new
obligation for the department, handled through
local employment offices. In addition, the Labor
Department would be the principal federal agency
monitoring employer compliance with the requirements for establishing payroll reduction plans and
remitting premiums.
The Health Care Financing Administra ti on
(HCFA) would be a logical choice to broker agreements with the states and to monitor these
agreements, and to be responsible for changes in
federal insurance rules. But it would be impossible
for the agency to undertake these responsibilities
with its current resources and bureaucratic culture—HCFA has been unable to carry out several
responsibilities assigned by Congress properly
because of a lack of resources. The experience of
implementing and enforcing HIPAA indicates
shortcomings at HCFA in data collection, oversight,
and guidance of states. HCFA also has been unable
to handle its obligations under Medicare. The ne w
requirements in this proposal represent yet another
reason to overhaul the agency. The Labor and
37
Treasury departments, responsible for some provisions of HIPAA, also have some weaknesses, though
much less serious ones.25
A possible solution to HCFA’s inability to carry
out many of the proposed new functions would be
to combine these functions with structural reforms
already proposed for Medicare. Such Medicare proposals include creation of a “Medicare Board,” separate from HCFA, to manage a system of competing
health plans. The board would be modeled on the
Office of Personnel Managem en t’s role in the
FEHBP, leaving HCFA to focus on the traditional
fee-for-service Medicare program. Establishing the
board is seen as a way of creating an agency with a
very different staff and culture from HCFA—one
that would create and manage competitive markets.
Given that the federal functions under the proposed
tax credit system would be similar to those of a
Medicare board—establishing conditions for affordable plans to exist in a competitive market—it might
be possible to include them in a widened role for
such a board.
States
New responsibilities for states would accompany the
additional funding available through the new federal grant to states. After negotiating a plan and goals
to reduce uninsurance, a state would have to develop a mechanism to supplement the federal tax credit for eligible workers and to deal with those who
did not purchase minimum insurance and had a
plan or program selected for them. States with an
income tax might choose to use the employmentbased state tax withholding system to deliver such a
supplement. Otherwise they would have to devise
an alternative automatic system, such as payments
to insurance companies, or enable households to
claim assistance directly from the state. In cooperation with the federal government, states also would
be responsible for administering the credit and premium payment functions for the unemployed.
Supplemental benefit programs and temporary programs to provide supplemental cash benefits have
Karen Pollitz, Nicole Tapay, Elizabeth Hadley, and Jalena Specht. “Early
Experience with ‘New Federalism’ in Health Insurance Regulation.”
Health Affairs 19 (4) (July/August 2000).
25
been delivered in this way since the s. A similar
process could be created with the unemployment
insurance system to pay health insurance premiums
for unemployed Americans; with additional administrative funding to defray the costs,the same system
could be used to provide con ti nu ed health insurance coverage.
To meet their agreed-on goal of reducing uninsurance,states would have to use the additional federal funds to expand existing programs—or develop
new ones—to achieve targeted levels of coverage. In
addition, they would have to work with insurance
companies to devise ways of introducing modified
community rating along with a reinsurance and/or
risk-adjustment system. This proposal envisions the
states having the flexibility to accomplish these goals.
In addition,each state would have to specify a default
enrollment system for those employees of firms in
the state who did not indicate a preferred plan to
their employer (and were not covered under a
spouse’s plan or a government program).
Employers
New obligations on employers would not be onerous. Employers would be required to inform their
employees about the tax credit program and to
make available the necessary federally produced
en ro ll m ent forms. Employers already have to
arrange tax withholding, so they merely would have
to adjust that amount according to a worksheet to
include the credit in their withholding calculations.
Employers also would be required to provide the
IRS with proof of insurance for each worker, which
could easily be done by sending on an insurance
statement received from the insurer. Fu rt h er,
employers would have to set up accounts for payroll
deductions to be placed for employees and make
payments from these accounts to health plans. For
large employers this would be similar to creating
flexible accounts. For small employers it would be a
new obligation, but a small one that could be carried out routinely by a payroll firm.26
26
See Butler, October 1999.
38
Insurers
Insurers would face several new requirements that
would add to their costs, but there would be significant administrative savings associated with the
automatic enrollment system. Among the new costs
would be the requirement to bring premium rates
into line with federal or state underwriting and benefit requirements. In addition, the insurance industry and the states would have to work together to
develop a mechanism to adjust risk among plans.
That would be a major and costly undertaking, but
in many states it would be a continuation of experiments already underway.
Encouraging Households to
Obtain Coverage
Under the proposal a condition of employment for
workers would be to tell their employers which
health plan they wished to join, or to accept enrollment in the default plan or program chosen by the
state. There are several reasons to assume that there
would be a high level of voluntary compliance with
this con d i ti on of employment, and that the selfemployed and those not regularly employed would
choose to enroll in a minimum plan at least.
Automatic Enrollment
Indicating insurance coverage and adjusting withholdings for the credit would be a routine feature of
employment, with individuals indicating their plan
ch oi ce when they are hired or during an annual
enrollment period. As noted earlier, evidence from
pension plans suggests that even a voluntary automatic enrollment system can increase participation
s h a rp ly, especially among younger and lowerincome workers. With a law requiring employers to
enroll those employees who do not choose a specific
plan in a default plan, there would be a strong
incentive for employees to take the time to make an
active choice. If unemployment offices made it simple for en ro ll m ent and tax subsides to con ti nu e
between jobs, coverage also would likely be high
among the short-term unemployed.
Loss of Tax Benefits
Any individual successfully avoiding enrollment in a
minimum plan would lose significant tax benefits,
not just for insurance, but for any health care product or service, from prescription charges to the cost
of well-baby visits. In addition, particularly for selfemployed individuals who would not be assigned a
default plan by an employer, the IRS could deny a
personal exemption to those who do not include in
their tax returns proof of year-round coverage.
Individuals Who Could not Afford Insurance
Of course, some individuals might feel unable to
afford basic covera ge , despite the subsidy and the
requirement to choose a plan or be enrolled in a
default plan or program. For example, they might
refuse to pay any costs associated with the default
plan. Others with low incomes or who are selfemployed might fail to enroll in a plan despite the
subsidies available. In these cases the state would
have to decide how best to address the needs of
these individuals. Since the state would receive a
grant to supplement the federal tax credit, it would
have funds to help families pay for insurance and
sign up at the workplace or to deal with the family’s
needs in other ways, such as enrolling the family in a
state-based program. In addition, the state would
receive a federal grant equal to the estimated value
of federal tax credits left unclaimed by those who
did not obtain coverage.
A Basic Benefits Package
To qualify for the tax credit, families would have to
enroll in a health plan that includes at least a minimum insurance package, although the credit would
be available for insurance and out-of-pocket costs
that exceed the minimum. The federal government
would establish a benchmark plan with basic features and catastrophic protection that would apply
to federally regulated plans. As part of their compact with the federal government, states would have
to incorporate this minimum into their insurance
regulations or propose an actuarially equivalent
minimum.
This proposed federal minimum should not be
39
confused with proposals to establish a federal
requirement for comprehensive coverage and specific benefits. The basic plan is intended to provide
minimum, primarily catastrophic insurance protection,not comprehensive coverage. The vast majority
of households would choose more extensive coverage, but requiring typical coverage as the required
minimum would make insu ra n ce prohibitive to
lower-income families, as the experi en ce of state
mandates has demonstrated.27 In addition, the minimum would be in the form of broad areas of covera ge , such as hospitalization and major medical,
similar to the requirements for plans in the FEHBP
or the California Public Employees’ Retirement
System (CalPERS), rather than a precisely defined
set of specific benefits, such as Medicare fee-forservice. CalPERS operates much like the FEHBP, but
for California state employees. For a minimum plan
to meet the conditions for the credit, it would have
to match the federal base plan’s broad features, and
its benefits would have to be at least equivalent to
the federal benchmark in actuarial terms.
The Process of Reform
The New Program and American Values
The proposed new system of insurance has a radically different dynamic from today’s arrangement,
yet its link with the place of employment fits in with
the familiar aspects of today’s system that most
Americans are loathe to abandon. Since enrollment
and the financial transactions associated with insurance would continue to be at the place of work, it
would look very much like today’s system.
In several respects, however, the new system
would be much more in line with the general
American view of economic rel a ti onships and
health care than is today’s employment-based sys-
Melinda Schriver and Grace-Marie Arnett. “Uninsured Rates Rise
Dramatically in States with Strictest Health Insurance Regulations.”
Backgrounder 1211, Washington: The Heritage Foundation, August 14,
1998.
27
28
A recent poll conducted for the Democratic Leadership Council, for
instance, found 72 percent support for a proposal to give employees the
option of receiving a tax break if they chose to “purchase their health
care through an outside organization such as the AARP, union, a church,
or a community-based purchasing co-op instead of through their
tem. For one thing, most workers would have far
more choices than they have now. In addition, they
would own their own policy as individuals or members of a group. This would give families far g reater
consumer control over their insurer, and insurers
generally would have the same incentive to satisfy
enrollees (rather than their employer) as plans do in
FEHBP.
Workers who are not in employer-sponsored
plans also could obtain their insurance through
intermediaries they trust to protect their interests,
rather than having to deal directly with insurers in
the individual market. Americans have indicated,
through their support for patients’ rights legislation
and in responses to surveys, that they want the government or another intermediary to exert some
control over health insurers. They also have indicated that churches, unions, and other such organizations are often more attractive intermediaries than
employers. 28 The proposed system would remove
many of the obstacles that make it difficult for families to obtain coverage through these groups.
By delinking tax benefits and employer sponsorship of i n su ra n ce , the proposed system would
achieve the true portability Americans want.
Workers typically would be able to remain with
their chosen health plan, even when they change
jobs, merely by informing their new employer of
their existing plan when they sign on for the
required payroll deduction and tax withholding.
Getting from Here to There
Americans are nervous about radical change in their
health care, including sudden changes in the tax
treatment of their benefits, even if in all probability
the change would be to their advantage. To address
this,the proposal could be introduced in stages,over
a long period, to make the transition as gradual and
employer.” If given that option, half the respondents said they would
choose it; see Mark J. Penn. “Health Care Is Back.” The New Democrat
Blueprint (Spring 2000): 70, 71. Other surveys indicate similar support
for non-employment groups as vehicles for insurance, and a recent poll
of women conducted for the Center for Policy Alternatives indicated
that 72 percent of women would like their health insurance to be independent of their employment (Women’s Voices 2000, Center for Policy
Alternatives and Lifetime Television, Washington, DC, 2000). See also
“Focus on Women.” The New York Times (September 27, 2000).
40
Because the proposed system would operate through the
place of employment, it would have the major political advantage
of not appearing to be a major departure from the existing
familiar arrangement.
politically feasible as possible. New refundable credits for households without access to employer-sponsored coverage already have been proposed in Congress and would be the logical first step. The second
step would be to introduce the new state subsidy
program, probably as an amendment to S-CHIP, to
enable states to supplement the credits with federalstate funds. The second step also would require
employers to administer the payroll deduction system. The third step would be for the federal government to modify HIPAA and ERISA, and to work
with the state governments to introduce rating
restrictions, basic benefits, and a risk-adjustment
system. With plans available based on these regulatory changes, it would be more feasible to implement
the core tax reform, which would replace the tax
exclusion with a comprehensive tax credit system.
Political Feasibility
The purpose of this proposal is to achieve universal
coverage in a way that relies on markets and choice
rather than new government programs. To do so
would change the way most health care for working
households is subsidized through a major reform of
the tax treatment of health care. But while the vision
of change is comprehensive, one must acknowledge
that Americans are reluctant to embrace radical
change that disrupts familiar arrangements.
Therefore, the actual introduction would have to
take place in stages, each one explained clearly and
extensively to build public support.
Because the proposed system would operate
through the place of employment,however, it would
have the major political advantage of not appearing
to be a major departure from the existing familiar
arrangement. But the tax credit approach has other
political advantages. For example, the basic idea
already has broad bipartisan support. Major tax
credit bills have been introduced regularly in Con-
gress since , and several were introduced in the
th Congress or put forward as proposals by
groups of members of widely differing political persuasions. Both presidential candidates in  also
embraced tax credits as a part of the solution to
uninsurance.
While the proposal would place new requirements on states, it also would help to relieve the burden on states caused by uninsurance and uncompensated care. The federal credit itself would reduce the
incidence of uninsurance, and states could combine
it with their own subsidies to enable families to afford
coverage, thanks to the new federal grant to state s .O f
co u rs e ,m a ny states might not like federal preemption of insurance regulation and the obligation to
meet targets for reductions in uninsurance, but they
would have the flexibility to modify their own regulations to achieve the targets, and the new federal grant
would be a clear inducement. The recent welfare
reform experience indicates that the combination of
agreed-on targets,flexibility, and financial awards to
states can be a political winner, and a similar
approach is incorporated into this proposal.
There no doubt would be a range of reactions
among insurers and health care providers, but the
tax credit approach has already won strong support
among such key groups as the American Medical
Association and the National Association of Health
Underwriters. This approach also has been gaining
ground among associations representing managed
care plans and insurers.
Larger employers and unions traditionally have
opposed changes in the tax code that would finance
alternatives to the traditional employer-based system, fearing they would cause employer-sponsored
insurance to unravel. But industry support is growing, in part because recent tax credit proposals have
included protections for employer-sponsored coverage, and in part because of an increasing recogni-
41
tion of the limitations of employer-sponsored coverage, especially for employees of small firms. The
U.S. Chamber of Commerce, for instance, now supports an income-related tax credit, as well as
expanding ERISA to facilitate insurance pools based
on associations and community-based organizations. The additional paperwork requirements in
the proposed system would be only a minor change
for most mid-size and larger employers, but would
be seen as burdensome by many small firms. The
requirement to assess the cash value of employersponsored benefits would be more problematic for
employers, both because of the burden of making
the calculation and because of concern about
employee reaction to varying total compensation
for employees with the same total income.
So far, the major unions have continued to
oppose tax-based initiatives, even though many of
them actually operate plans that could prosper
under a tax credit system. But emphasizing the protections for large employment-based plans negotiated by unions, and the opportunities for new unionsponsored plans in the service and small-business
sector, could reduce this opposition.
Conditioning tax relief on obtaining a minimum plan would make no practical difference to
most families, since they either have or want good
coverage. What would cause more consternation is
taxing employer-paid benefits, even with a new tax
credit. The idea of restricting or eliminating the tax
exclusion to raise government revenues has been
proposed several times and has encountered strong
opposition. But this proposal contains clear advantages to employees contemplating new taxes on a
previously tax-free fringe benefit. They would qualify for a credit that in most cases was approximately
the same as their current exclusion or larger, thanks
to the sliding-scale nature of the credit and the additional tax expenditure. Moreover, the tax change
would allow families new tax relief for out-of-pocket costs, so there would be a visible benefit every
time a person visited the doctor or paid for a prescription. And even though upper-income workers
generally would face a net tax increase, in many
cases this would be because of a conscious decision
to take more of their compensation in cash income
rather than fringe benefits. Thus the tax change
would be an attractive swap for most workers.
Other Issues
Cost and Efficiency
This proposal likely would increase pressures to use
medical services efficiently and to control costs. For
one thing, eliminating the exclusion and making
employer subsidies explicit would spur employees
to consider value for money much more carefully
and to exchange unnecessary benefits for cash. The
exclusion has long been recognized as encouraging
wasteful overinsurance by middle- and upperincome families. For another, giving workers the
opportunity to join large groups outside the place of
work would sharpen plan competi ti on and allow
workers in small firms, usually with costly and inefficient covera ge , to join more efficient plans.
Making it more possible for families to join plans
that reflect their preferences also would improve
economic efficiency.
Efficiency also would be improved by reducing
the administrative costs currently associated with
plans covering individuals or group plans serving
the employees of small firms (including the high
costs borne by small employers in arranging insurance). The proposal envisions many workers now in
very small employer groups switching to much larger non-employment groups. Small employers currently providing insu ra n ce also would face lower
costs in their new role. The potential reduction in
administrative costs is difficult to estimate; it would
depend on the types of plan that emerge. In estimating versions of the Heritage Foundation proposal,
the Lewin Group assumed administrative costs of 
percent of benefit costs, based on plans covering
large numbers of people in the individual market.29
The FEHBP market suggests that costs could be
much lower. Others have estimated that, with automatic workplace enrollment and payments, the
administrative costs of a tax credit system could be a
low as . percent.30
29
Sheils et al., 1999, p. 45.
30
Etheredge, 1999, p. 6.
42
In any event, if large groups formed in the market, with enrollees coming from today’s individual
and small-group market, the average administrative
costs of insurance would tend to fall.
Subsidies to low-income families for affordable
insurance would encourage more Americans to
leave welfare and unemployment permanently by
removing a major obstacle to long-term private
employment in the small-business sector. The system would have some marginal effects on work
effort because of the credit design. But,although the
sliding-scale credit does include a traditional phaseout mechanism, the cost/income ratio determining
the marginal credit percentage leads to only a gradual reduction in the value of the credit for a given
amount of health spending as income rises. The
equivalent effective marginal tax rate for the credit
phase-out drops at particular incomes, or bend
points,as income rises. The rate also declines for any
income level as expenses rise. For a family of four
with , worth of expenses, for instance, the rate
is . percent for incomes up to ,, declining to
. percent for incomes of ,.
Achieving Greater Equity
Tax-based approaches of this kind sometimes are
seen as less efficient because part of the tax expenditure goes to households that are insured already.
Currently some three million workers and dependents purchase some level of health insurance, even
though they are unable to claim tax relief. Moreover,
if tax credits were available to offset out-of-pocket
costs associated with employer-provided plans,
many other insured families would be able to claim
additional tax relief. Hence the cost per newly
insured individual typically would be higher than
approaches that restrict services or program eligibility only to those who are uninsured.
A key objective of tax-based approaches to uninsurance, however, is to reduce the large inequities in
the current tax treatment of health care costs. If a
policy aim is to provide similar levels of assistance
to families with the same income and health needs,
similar subsidies must be provided to families currently with or without insurance. To do otherwise
would discriminate against families that had taken
steps to protect not only themselves but also their
community by reducing the likelihood that they
would incur medical expenses they could not pay.
Moreover, the argument that a tax credit approach
would mean subsidizing people who are already
insured is not unique to this approach. Any program designed to assist individuals without insurance is bound to crowd out some existing insurance
if the program is more generous. That is why
Medicaid expansions and other efforts by states to
cover uninsured families lead to at least some erosion of employer-sponsored insurance. If the equitable allocation of subsidies is an objective, this
effect cannot be avoided.
Quality of Care
While the quality of care available to families under
the proposal would depend on many factors, it is
possible to make some observations about probable
effects. One is that the wider coverage resulting
from the reform would mean that many more
Americans would receive their care through plans
that monitor their health, rather than through
emergency rooms and occasional visits to the doctor. Second,most families would have greater choice
of plans than they do now, and thus could select
plans that better meet their medical needs. Third, a
more consumer-driven market would lead to more
usable information that enables families to pick
more appropriate plans, because plans would be
under greater pressure to satisfy enrollees (rather
than their employers). Increased choice and competi ti on in a market makes it more economical for
third parties, such as consumer organizations, to
obtain and distribute information to prospective
enrollees—a phenomenon seen in the FEHBP. n
43
Feder, Levitt, O’Brien, and Rowland Approach
Key Elements
Judith Feder, Larry Levitt, Ellen O’Brien, and Diane Rowland outline an
expansion of Medicaid and the State Children’s Health Insurance Program
(S-CHIP) and explore its interaction with tax credits for individuals or
employers. Specifically, they conclude that:
      -     is best achieved by extending eligibility
for public programs without cost sharing or premiums to all individuals
with incomes below  percent of the federal poverty level, and, as in
S-CHIP, extending eligibility for public programs with modest premiums
and cost sharing (up to a maximum of  percent of income) to people
with incomes between  percent and  percent of poverty. People with
incomes above  percent of poverty could also be allowed to “buy in” to
public coverage by paying a sliding-scale premium based on income.
        to small, low-wage employers for providing
coverage to their employees is a more effective and less disruptive complement to a public program than a tax credit directed at individuals to
purchase non-group coverage.
44
About the Authors
 , .., is Professor and Dean of Policy
Studies at Georgetown University. Over three
decades, Feder has published extensively on financing for both health and long-term care.She has also
served in policy leadership positions—specifically, as
staff director of the Congressional Pepper Commission (chaired by Senator John D. Rockefeller IV)
and, during the first Clinton Administration,as
Principal Deputy Assistant Secretary for Planning
and Evaluation at the Department of Health and
Human Services. Feder is currently Vice Chair of the
Academy of Health Services Research and Health
Policy; senior advisor to the Kaiser Commission
on Medicaid and the Uninsured;and member of
editorial boards for several health policy journals.
A political scientist (B. A.from Brandeis,Ph.D.
from Harvard),she is a member of the National
Academy of Public Administration and of the
National Academy of Social Insurance. As a senior
scholar at Georgetown’s Institute for Health Care
Research and Policy, Feder’s current research projects include co-directorship (with Sheila Burke) of
a Kaiser Family Foundation project on incremental
health reform and of a Robert Wood Johnson
Foundation project on long-term care financing.
     , ...,is Vice President and Director of
the Changing Health Care Marketplace Project and
California Health Policy for the Henry J. Kaiser
Family Foundation. Before joining the Foundation,
Mr. Levitt was a senior manager with The Lewin
Group, where he advised public and private sector
clients on health policy and financing issues. He
previously served as a Senior Health Policy Advisor
to the White House and Department of Health and
Human Services, working on the development of
President Clinton’s Health Security Act and other
health policy initiatives. Prior to that,he served as
the Special Assistant for Health Policy with California Insurance Commissioner John Garamendi—
where he co-authored Commissioner Garamendi’s
“California Health Care in the st Century” proposal—and as a medical economist with Kaiser Permanente. He also served in Massachusetts state government. He holds a bachelors degree in economics
from the University of California at Berkeley, and a
masters degree in public policy from Harvard University’s Kennedy School of Government.
 ’, .., is Assistant Research Professor at
Georgetown University’s Institute for Health Care
Research and Policy, where her current research
focuses on trends in health insurance coverage and
options for coverage expansion. In addition, she is
examining issues related to employer decision-making about health benefits and the role of health
insurance in labor markets. Dr. O’Brien is also a
member of the teaching faculty in the Georgetown
Public Policy Institute. Before coming to
Georgetown in , she was a researcher at the
Health Care Financing Administration where her
work focused on long-term care in Medicare and
Medicaid. Dr. O’Brien has an M.A. in Economics
from the University of Iowa and a Ph.D in
Economics from the University of Notre Dame.
  , .., is Executive Vice President of
the Henry J. Kaiser Family Foundation and the
Executive Director of the Kaiser Commission on
Medicaid and the Uninsured. She is an adjunct
Associate Professor in the Department of Health
Policy and Management at the School of Hygiene
and Public Health of the Johns Hopkins University.
Dr. Rowland serves on the Board of the Academy for
Health Services Research and Health Policy, the
Secretary’s Task Force on Infant Mortality, and the
Commonwealth Fund Task Force on Health
Insurance for Working Americans. She is a founding
member of the National Academy for Social
Insurance, Past President of the Association for
Health Services Research, and a Brookdale Senior
Fellow. Dr. Rowland specializes in issues related to
health insurance coverage, access to care, and health
care financing for low-income, elderly, and disabled
populations. She has published widely on these subjects and is a noted authority on health policy,
Medicare and Medicaid.
45
Assessing the Combination of
Public Programs and Tax Credits
by Judith Feder, Larry Levitt, Ellen O’Brien, and Diane Rowland
Overview
Intrinsic to any proposal to expand health insurance
coverage is taking a position regarding the coverage
that already exists.One set of proposals aims directly at replacing current coverage, creating new mechanisms that would apply to the already insured and
to people who have no insurance. These proposals
often place as much emphasis on equity—for example, ensuring that all low-income people receive
equivalent government subsidies, regardless of how
they are insured—as on decreasing the number of
people who are uninsured. Another set of proposals
aims more narrowly at the uninsured, but affects the
already insured indirectly (and sometimes unintentionally) with the incentives and mechanisms newly
put in place.
Our approach resembles the second set of proposals in its focus on the uninsured. But, unlike
either of the above sets, our objective is the retention, not disruption, of existing coverage—specifically, publicly provided and employer-sponsored
health insurance. We adopt this strategy not because
we think that these mechanisms have no flaws; we
recognize that flaws exist. However, we are concerned that there is more to lose than to gain from
disrupting them—particularly for low-income people whose coverage is our primary concern. Simply
stated, our goal is to expand coverage for those without it and to “do no harm” to coverage mechanisms
now in place.
In the absence of comprehensive health reform
aimed at universal coverage, we suggest that the following principles should guide the design of incremental efforts to decrease the number of uninsured
Americans:
• New dollars spent on health insurance should
be targeted to significantly expand coverage.
• Coverage should be expanded to the uninsured
without disrupting coverage already available in the
public and private sectors.
• Expansion should begin with,and place priority on, coverage for those uninsured who are least
able to pay.
To satisfy these principles, we argue that, for the
low-income uninsured, the most effective approach
to expanding coverage is to extend the Medicaid and
State Children’s Health Insurance Program (S-CHIP)
eligibility now available to children and some parents
to all low-income individuals. Because the lowestincome population is least able to purchase health
insurance on its own, this public program should
have the highest priority as a claim on federal dollars.
Fo ll owing the Medicaid/S-CHIP approach
would mean extending eligibility for comprehensive
benefits at no cost (as in Medicaid) to all individuals
with incomes below  percent of the federal
poverty level, and extending benefits with some premiums and cost sharing (as in S-CHIP) to individuals with incomes between  and  percent of the
federal poverty level. People with incomes above
 percent of poverty could be allowed to “buy in”
to public coverage by paying a sliding-scale premium based on income.
Although public programs are the most appropriate way to extend coverage to the low-income
population,they could be combined with tax credits
to reach the uninsured who have modest incomes.
In such a combination, careful attention is needed
to en su re that any tax credits complement rather
than su b s ti tute for existing public and private
sources of coverage.A number of policy makers and
46
analysts have proposed tax credits that could be
used by individuals to buy non-group insurance.
Making such a tax credit non-refundable—thereby
targeting it pri m a ri ly to those with incomes in
excess of  percent of the poverty level—would
mitigate conflicts with public coverage. Even so, it is
difficult to design a modest, non-refundable individual tax credit that is effective and well targeted,
while at the same time avoiding disruption of existing employer-sponsored insu ra n ce . For example,
allowing application of an individual tax credit to
employer coverage would significantly reduce the
risk of displacing that coverage, but it would also
likely substantially increase the cost of the credits
with payments to the already insured.
Because it could be better targeted, a more effective approach to combining a public program
expansion and a tax credit might be a health insurance tax credit provided to employers, rather than to
individuals. It could be targeted to those small ,l ow wage businesses least likely to offer insurance today,
maximizing the focus of public dollars on improving
access to employer-sponsored covera ge . Although
employer tax credits have the disadvantage of leaving
people with modest incomes dependent on their
employers’ willingness to expand covera ge , this
downside may be more than balanced by the upsides
of better targeting and less disruption.
The following discussion begins by explaining
the risks posed by disruption of either employersponsored or publicly sponsored covera ge , then
examines the reasons,in the current policy environment, for reliance on more than one policy instrument to expand coverage. We then make the case for
expanding public coverage and explore the issues
raised by pairing that expansion with a tax-based
approach. We conclude with a discussion of specific
design issues raised by the public and private components of a combined strategy.
Why Avoid Disruption?
It is hard to disagree with a critique of the nation’s
current mix of insurance mechanisms as insufficient,
inefficient, and inequitable. However, to advocate
replacing these mechanisms with something else pre-
sumes the political wherewithal to achieve comprehensive reform and the political and administrative
wherewithal to devise an improved system. Experience warrants skepticism on both counts. Over the
last century, periodic efforts to achieve comprehensive reform have encountered significant political
obstacles. The most recent effort, in –, was
obviously no exception. There are various obstacles,
but high on the list is the concern of those who
already have insurance coverage that they will be
worse, not better, off under reform. Proponents of
the Clinton administration’s Health Security Act
argued that the plan would secure health insurance
for all Americans. But critics successfully countered
that the plan would dramatically alter, indeed undermine, coverage of the already insured. The plan that
claimed to benefit everyone came to be seen as likely
to benefit the uninsured minority, while making the
already-insured majority worse off.
It is not clear that an alternative policy and political strategy—one that claims from the outset to
benefit only the uninsured—will be more successful. After all, it will require explicit recognition of
the need to redistribute resources from those who
have insurance to those who do not. But given the
political problems generated by the fear of disrupti on , it seems worth trying a more targeted
approach. Moreover, when the aim of policy is
incremental change—which is the most likely scenario in the near term—the need to minimize disruption becomes even more important, because the
gains are not large enough to justify the risk of losses for those who are already insured.
Minimizing displacement requires attention to
the policy and political advantages of both existing
employer-sponsored and publicly sponsored covera ge . Not only do most Americans gain coverage
through employment, but polls also indicate that
they value that approach, despite changes in the
structure of employment and the dissatisfaction
that has accompanied changes in employer-sponsored insurance over the last decade.1 From a policy
1
Kaiser Family Foundation and Harvard School of Public Health. PostElection Survey: The Public and the Health Care Agenda for the Next
Administration and Congress. Menlo Park, CA: Kaiser Family
Foundation, 2001.
47
perspective, the strongest advantages of employersponsored insurance are its administrative efficiencies (including practically automatic enrollment of
employees) and its significant,if not total,spreading
of risks across people of different incomes and
health status. Employer coverage is particularly
valuable to low- and modest-wage workers, because
it provides for easily accessible enrollment and compensates for cash flow problems that would arise if
workers had to shop on their own.
Although the tax preference for employersponsored insurance is frequently criticized as
“inequitable”—and its direct monetary benefits are,
in fact, skewed to those who are better off— it also
serves as an appropriate incentive to achieve efficient risk pooling. Despite rhetoric to the contrary,
there is little or no evidence that newly created
administrative structures can replicate the effectiveness of employers in effectively pooling risks.2 To
disrupt employer coverage without confidence in a
reasonable alternative ultimately puts the scope and
adequacy of coverage in jeopardy.
There is also a significant risk to disruption of
public coverage—that is, coverage provided by the
Medicaid and S-CHIP programs. Medicaid, in particular, has been criticized as being more discouraging than inviting of participation, whether by
beneficiaries or providers. But barriers to participation likely have more to do with Medicaid’s meanstested eligibility and the implementation of that
means test than with something peculiar to the
Medicaid program. States’ dramatic expansion of
Medicaid coverage at the end of the s, and their
more recent implementation of S-CHIP, indicates
that a public program’s attractiveness, or the ease
or difficulty of participating in it, reflects policy
ch oi ces that are an essential part of any program,
new or old.3
The fact is that low-income people will always
need more public support than the rest of the pop-
ulation if they are to have affordable access to
coverage and services. Medicaid’s -year history
of providing health insurance to segments of the
low-income population has established both
administrative and legal structures that protect
beneficiaries’ rights to benefits and care. Proposals
to replace Medicaid may offer far less support than
Medicaid currently provides—whether in benefits,
administrative arrangements, or legal foundations
(including enforcement of federal entitlements).
And creation of new federally financed subsidy
mechanisms for the uninsured—even if they ostensibly leave Medicaid untouched—may encourage
political pressure to weaken existing protections.
(As discussed below, enactment and implementation of S-CHIP raise precisely that possibility.) Just
as it is appropriate to question whether new adm i n i s tra tive stru ctu res can ef fectively rep l ace
employer-sponsored insurance, it is also appropria te to qu e s ti on wh et h er new ad m i n i s tra tive
and subsidy structures can effectively replace
Medicaid.
Jack Meyer et al. Tax Reform to Expand Health Coverage:
Administrative Issues and Challenges. Menlo Park, CA: Kaiser Family
Foundation, 2000.
The following discussion draws on: Judith Feder et al. “Covering the
Low-Income Uninsured: The Case for Expanding Public Programs.”
Health Affairs (January/February 2001); Diane Rowland, Rachel Garfield,
Christina Chang, and Barbara Lyons. Building on Medicaid to Cover the
Low-Income Uninsured. Washington: Kaiser Commission on Medicaid
and the Uninsured, forthcoming.
2
Donna Cohen Ross and Laura Cox. Making It Simple: Medicaid for
Children and CHIP Income Eligibility Guidelines and Enrollment
Procedures. Kaiser Commission on Medicaid and the Uninsured, 2000.
3
Why Rely on More than One Policy
Instrument? 4
The population without insurance is not a homogenous group. Differences in peoples’ circumstances
or characteristics do not necessarily require the use
of different policy strategies to reach them. But different policy strategies will be more or less effective
in reaching different segments of the uninsured
population (for example, those with lower versus
higher incomes).
Currently, tax policy has gained some political
popularity as a strategy to expand coverage—
targeted specifically to low- and modest-income
people. President Bush campaigned in favor of such
a policy, and various proposals for targeted credits
have been put forward by both Democrats and
4
48
Republicans in Congress.5 The appeal of tax preferences to provide subsidies appears to be the
potential to operate with minimal government
involvement. In theory, people could apply by filing
tax returns, rather than applying to a government
office, and they could choose a plan on their own,
rather than relying on plan options selected by a
government agency. However, three factors make
this strategy problematic for low-income people.
The first factor is the tax system’s limitations in
reaching the low-income uninsured. About half of
people without health insurance do not file an
income tax return or owe any income taxes.6 To
reach them at all, an income tax credit would have
to be refundable—that is, available without regard
to tax liability. The Earned Income Tax Credit
(EITC) is a refundable tax credit that has been enormously successful in enhancing income for the
working poor. However, it is harder to support the
purchase of health insurance than to b oost income.
Tax credits, including the EITC, are typically
refunds—money the taxpayer gets back at the end
of the year. To buy health insurance, people with
limited incomes need the cash in advance. Further,
they need to know they can keep the money, even if
their income changes. Advance payment and nonreconciliation of income and subsidies at year’s end
would require significant departures from current
tax practices—practices seen as ensuring the accuracy and efficiency of the tax system.
The second factor is problems with the market
or insurance products that such a credit could buy.
About  percent of the uninsured lack access to
employer-sponsored insu ra n ce , whether through
their own jobs or the jobs of family members. Credit
recipients without access to employer coverage
would be dependent on access to the non-group
insurance market to obtain coverage. But that market is riddled with problems. To avoid adverse selection,insurers use practices to avoid enrolling people
likely to use services. Except in a few states with
comprehensive regulation, insurers can deny people
access; exclude coverage for services, conditions,
body parts, or body systems; and charge whatever
premiums they deem appropriate. As a result, people pay more when they get sick and can lose access
to coverage. Overall, benefits in the non-group market are quite limited (often excluding maternity
benefits, prescription drugs, and mental health, and
typically using significant deductibles or benefit
caps). The fact that people insured in the non-group
market are no less healthy than people with employer coverage demonstrates the effectiveness of insurer practices in controlling access to coverage by
people in relatively poor health.7
The third factor is the questionable adequacy of
the tax credit. The most prominent proposals
involve tax credits that fall far short of the cost of
health insurance (for example, a , credit for a
family, when the cost of a typical family insurance
policy typically exceeds ,). Clearly, the lower a
person’s income, the less able that individual is to
make up any difference between the credit and the
cost of an insurance policy.
For some or all of these reasons, even some proponents of tax credits recognize that a public program is better than using the tax system to reach the
low-income uninsured. Building on existing public
programs has two fundamental advantages. First
and foremost is the extension of an adequate subsidy for an adequate product—that is, a subsidy for
the full cost of comprehensive insurance to people
with limited incomes. Second is the existence of an
administrative apparatus in every state to determine
eligibility for subsidies in advance and to facilitate
enrollment in health insurance plans. Medicaid and
S-CHIP programs—which now serve about  million people—have contracts in place with providers
and managed care plans (indeed, they are public
managers of private markets) and have established
mechanisms for collecting and matching funds
from the federal government. Although recent
Randall Weiss and Mark Garay. Recent Tax Proposals to Increase Health
Insurance Coverage. Menlo Park, CA: The Kaiser Family Foundation,
2000.
7
John Holahan, Unpublished analysis of Medical Expenditure Panel
Survey, 2000.
5
6
Jonathan Gruber and Larry Levitt. “Tax Subsidies for Health Insurance:
Costs and Benefits.” Health Affairs (January/Feburary 2000).
49
The conclusion that a public program is the appropriate mechanism
for reaching the low-income population means that if a tax strategy
is to be pursued for those with higher incomes, it should complement,
not substitute for, a public program.
attention has focused on barriers to participation in
public programs,a decade ago attention centered on
the speed of Medicaid enrollment expansions in
response to changes in federal law—from . million people in  to . million in . And,
although a year or two ago enrollment in S-CHIP
seemed to be expanding slowly, all states have dramatically expanded income eligibility standards for
children—above  percent of poverty in 
states—in recent years.8
The conclusion that a public program is the
appropriate mechanism for reaching the lowincome population means that if a tax strategy is to
be pursued for those with higher incomes, it should
complement, not substitute for, a public program.
Issues at the Intersection of Public
Programs and Tax Policies
The likelihood and form of a combination of a tax
credit and a public program will depend on the
political process. But the effectiveness of such a combination—in expanding coverage with minimal disrupti on — wi ll rest on answers to some strategic
design questions.
Should the Two Instruments Be Parallel or Layered?
Establishing a tax credit alongside existing or new
public coverage might seem, on the surface, to offer
individuals an attractive ch oi ce of how or where
they wish to obtain covera ge . However, given the
complexity of the health insurance market, and the
difficulty of obtaining information for meaningful
comparison-shopping within it, it is reasonable to
doubt whether competing mechanisms constitute
meaningful consumer choice.
8
Ross and Cox, 2000.
This doubt is reinforced by the incentives states
would face if tax and public programs existed side
by side. Medicaid expenditures are always high on
the list of state fiscal concerns, and Medicaid costs
are once again rising faster than state revenues.9 A
tax credit could enable states to justify a contraction
of Medicaid and S-CHIP coverage on the grounds
that alternative subsidies were available for use in
the private market. And, from the state perspective,
these subsidies would have the advantage of being
financed at federal, not state, expense. Successful
substitution of federal credits for state/federal public programs would mean both a shift from state to
federal expenditures and a decline in the benefits
and stability of health insurance provided to lowincome people.
Given the incentives, establishing tax credits
alongside public programs can be seen as establishing a choice for states,more than it does for individuals. The result may be to undermine rather than
enhance protection for low-income people, especially if (as is likely) the tax credit option is significantly less comprehensive than public coverage. To
secure and extend health coverage for low-income
people, layering a tax credit on top of a public program—that is, targeting each policy instrument to a
different income group within the uninsured population—is a more effective approach.
How High up the Income Scale Should Eligibility
for Public Coverage Extend?
Public programs for the low-income population
offer comprehensive benefits at little or no cost to
beneficiaries, reflecting an emphasis on en su ri n g
affordability of coverage and services for people
with limited ability to pay. With the cost of private
9
National Association of State Budget Officers and National Governors
Association. The Fiscal Survey of States: December 2000.
50
insurance at about , per adult and , per
family, on average, the income level at which full or
nearly full subsidies are arguably necessary to ensure
affordability is relatively high. Expenses at this level
are clearly beyond the means of people with
incomes at the federal poverty level (, for an
individual, , for a family of four). Indeed, even
at incomes of t wi ce that level, a premium would
absorb more than  percent of family income.
Hence, an argument can be made for extending
public coverage to incomes up to and even above
double the poverty level.
However, as income rises, so does the proportion of people with employer-sponsored insurance.
At incomes below  percent of the federal poverty
level, only  percent of the population has employer-sponsored insu ra n ce . By contrast, at incomes
between  percent and  percent of poverty, 
percent of the population has employer-sponsored
coverage, and at incomes between  percent and
 percent of poverty,  percent has it.10 In a sense,
these modest-income people who have coverage are
actually paying for it, whether by forgoing income
they would otherwise receive in wages or by paying
actual out-of-pocket premiums. Whether people
between one and two times poverty are perceived as
able or unable to “afford” premiums and cost sharing, therefore, depends on whether the focus is on
the  percent who actually have coverage, or on the
roughly  percent not offered coverage (who
would be expected to pay for coverage explicitly
out-of-pocket instead).11
The scope of employer-sponsored insurance
among people with incomes above the federal
poverty level raises the additional question of how
to balance affordability of coverage for low- and
modest-income individuals with displacement of
private coverage as public coverage is expanded.
While the majority of employers offer health insurance to their workers, many also complain about its
costs and administrative burdens. Some have talked
a bo ut providing cash payments in the form of a
10
Paul Fronstin. Sources of Health Insurance and Characteristics of the
Uninsured: Analysis of the March 1999 Current Population Survey.
Washington: Employee Benefit Research Institute, 2000.
“defined contribution” rather than sponsoring
health insurance coverage. Despite a likely preference among employees for employer-sponsored
over publicly sponsored coverage, availability of a
public program at higher-income levels would create incentives for employees to choose free or nearly
free public coverage over employer coverage that
might require a substantial premium contribution.
It also might create the opportunity for employers—particularly employers whose employees earn
relatively low wages—to drop coverage entirely.
Indeed, concern about crowd-out, as it is popularly
described, led Congress to limit el i gi bi l i ty for SCHIP to children in families with incomes of up to
 percent of poverty who lacked employer coverage. In establishing eligibility levels for a program
expansion, experience and analysis indicate the
importance of careful attention to the potential disruption of employer coverage.
How Big Should a Tax Credit Be, and to What Kind
of Coverage Should It Apply?
The concern about employers dropping coverage is
clearly not limited to public program expansions.
Departures from the provisions of current tax policy that favor the purchase of health insurance
through the workplace instead of coverage purchased individually in the non-group market would
perhaps be even more likely to induce employer
dropping and employee switching than would public expansions. The likelihood that dropping would
occur depends on the scope of a new tax policy.
Making premium payments for non-group coverage
tax deductible—widely advocated on equity
grounds—would partially neutralize the current tax
preference for employer coverage, though the benefits of pooling and lower administrative costs would
remain. Extension of a tax credit for non-group
insurance—more generous in many cases than a
deduction—could actually create advantages to
purchasing outside the workplace, especially for
those employees who are young and healthy and,
11
Estimates from the 1996 Medical Expenditures Panel Survey, provided
by Mark Merlis, Insitute for Health Policy Solutions.
51
thus,able to get favorable premiums. The more generous the tax credit, the more willing employees
would likely be to seek coverage outside the workplace, and the more likely employers would be to
drop sponsorship of health insurance.
The potential for employers to drop coverage
can be mitigated if new tax credits are applicable,
not just to the purchase of non-group coverage, but
also to worker payments toward employer-sponsored covera ge . The application of credits toward
employer-sponsored coverage is also advocated on
grounds of equity—treating individuals with similar incomes similarly, regardless of how they obtain
their insurance coverage. However, allowing credits
to be applied toward employer coverage will significantly increase the costs of an intervention, because
in firms that already offer coverage,credits will go to
the bulk ( percent) of workers who accept coverage,12 along with the minority who do not. Indeed,
such an approach should not be seen as preventing
substitution of public for private dollars; rather, it
constitutes an explicit substitution of public for private dollars to achieve equity and to secure existing
employer-sponsored coverage. A policy ch oi ce on
this issue will clearly depend on the total dollars
available and the willingness to spend on the already
insured, as well as the newly insured.
Can Subsidies for Public and Private Coverage Be
Integrated Smoothly?
A policy that layers a tax credit on top of a public
program must pay particular attention to administrative and equity issues that arise at the intersection
of the two policy instruments. For example,a policy
that abru pt ly terminates eligibility for relatively
comprehensive public coverage that is available at
little or no cost at a specific income level creates a
cliff: people with incomes below the specified level
get a lot, while people with incomes just ab ove that
level get nothing. That is, in fact, the way eligibility
for both Medicaid and S-CHIP currently works.
Clearly, extension of a tax credit mitigates this
12
Estimates from the 1996 Medical Expenditures Panel Survey, provided
by Mark Merlis, Insitute for Health Policy Solutions.
cliff, because it creates benefits above the eligibility
level for public coverage. The more generous the tax
credit, the less steep the cliff becomes. One way to
think about establishing the size of the credit, then,
is to set it so that the amount of out-of-pocket
spending it requires recipients who are just above
the limits of eligibility for public coverage to pay for
private insurance is similar to the out-of-pocket
spending toward public insurance expected of people whose incomes are just below the eligibility
limit. The value of the credit, relative to the cost of
premiums, could then decrease as income rises. The
desire to smooth out cliffs in subsidies, however,
must be balanced against the desire to avoid providing a tax credit—particularly one for non-group
insurance only—that risks disrupting existing
employer coverage.
Smooth integra ti on also requires attention to
the availability and characteristics of i n su ra n ce
products. If no changes are made in the private
insurance marketplace, some people who are eligible for a credit may be unable to find or afford coverage—given insurance practices that limit access to
or set prices for insurance based on people’s age,
health status, or other factors.13 One way to address
this problem would be to regulate the insurance
market by establishing rules affecting both access
and price. Another would be to establish a new publicly managed market in which insurance products
are made available to all potential purchasers (for
example, a purchasing cooperative). A third would
be to allow people above the eligibility level for public coverage to “buy in” to the public program—that
is, pay a premium from their own resources to
obtain publicly sponsored coverage (not really so
different from a purchasing cooperative, because
most Medicaid and S-CHIP programs now provide
coverage to families though private health plans).
Making the tax credit applicable toward—and,
indeed, equal to—the premium for publicly sponsored coverage would further smooth any transition. Of course, when establishing a buy-in to a
Deborah J. Chollet and Adele M. Kirk. Understanding Individual Health
Insurance Markets: Structure, Practices, and Products in Ten States.
Menlo Park, CA: Kaiser Family Foundation, 1998.
13
52
public program—or a purchasing cooperative, for
that matter—one has to pay careful atten ti on to
how it relates to the private insurance market to
ensure that it does not turn into a dumping ground
for high-cost individuals.
The design choices actually made in each of
these areas clearly have enormous implications for
the cost, effectiveness, and administrative operations of any initiative that combines a public program and tax credits. And, as is often the case in
policy making, design ch oi ces will require tradeoffs. Choices made to smooth integration, for example, may run counter to choices made to minimize
substitution or disruption. Following is an array of
possible choices.
Establishing a Public Program
The most effective way to reach the low-income people who are now uninsured would be to extend protections that are now available to some of them to all
low-income people. Currently, Medicaid concentrates primarily, and S-CHIP almost solely, on lowincome children. Although Medicaid covers women
while they are pregnant, and states have the option
to include parents, in  states uninsured working
parents are ineligible for Medicaid if they work fulltime at the minimum wage.14 Further, low-income,
childless adults, no matter how poor, are ineligible
for coverage under federal law unless they qualify as
disabled. To reach the entire low-income uninsured
population,an initiative would make income, rather
than family status,the sole criterion for eligibility.
Such a public program extension must address a
number of other policy issues, as outlined below.
Eligibility
The extension of eligibility for comprehensive benefits at virtually no cost to all individuals with
incomes below  percent of the federal poverty
level would en su re affordable coverage with little
threat to current employer coverage. Employers
Jocelyn Guyer and Cindy Mann. Employed But Not Insured: A State-byState Analysis of the Number of Low-Income Working Parents Who Lack
Health Insurance. Washington: Center on Budget and Policy Priorities,
1999.
14
now cover only about  percent of the population
with incomes below the federal poverty level and
only  percent of the population with incomes
between  percent and  percent of the federal
poverty level. Although the proportion with
employer coverage rises at higher income levels
(about half for people between  and  percent
of poverty), to truly ensure access to affordable coverage, eligibility would have to go beyond this very
poor group. One approach would be to build on
public policy decisions that have already extended
coverage to children in families with incomes up to
 percent of the federal poverty level and apply a
similar policy to their parents and other adults. As
in S-CHIP, it might be appropriate to apply some
premiums and cost sharing in the income range
between  and  percent of poverty (up to a
maximum of  percent of income). And (if
resources allow) it would be desirable—on equity
grounds—to avoid current S-CHIP rules in many
states that deny coverage for a period of time to
those who have had employer-sponsored coverage
(consistent with current Medicaid policy). Finally,
to ensure a smooth transition for people with higher incomes,it may be appropriate to allow individuals with incomes above  percent of poverty to
buy into the public program by paying a slidingscale premium based on income.
Federal/State Roles
Extension of public coverage requires consideration
of the way federal and state governments share
financing and authority. Medicaid (and S-CHIP) is
a federal/state matching program, under which the
federal government offers to match state expenditures to entice states to provide more coverage than
they would on their own. The matching formula
provides more federal money (raises the matching
rate) for states with poorer populations. States
accepting federal funds are required to abide by federal rules for eligibility, benefits, administration,
and other aspects of program operations.
Over the years, provision of federal matching
funds has helped to expand coverage. But matching
funds (without minimum federal eligibility standards) have not achieved uniform coverage across
53
Given the priority we place on covering the low-income uninsured, the most
effective approach to expanding coverage is to extend the Medicaid and S-CHIP
eligibility now available to children and some parents to all low-income individuals.
states, nor have they miti ga ted states’ discomfort
with applying rules that reflect federal priorities. In
, Congress decided to achieve greater uniformity of coverage by phasing in a floor on eligibility levels for children—requiring states to cover children
in families with incomes below  percent of
poverty as a condition for receiving any Medicaid
funds. Above the floor, eligibility levels continue to
vary. In ,  states had extended eligibility
under Medicaid or S-CHIP to children with
incomes above  percent of the federal poverty
level, while in eight states, eligibility standards were
below  percent of poverty. Variation is wider for
parents, where no federal floor exists. Ei gh teen
states extend eligibility for parents to incomes above
 percent of the poverty level, but  states limit
eligibility to parents with incomes below  percent
of the federal poverty level.
Variation in eligibility across states reflects not
only the reluctance of some states to spend, but also
their reluctance to extend programs that are expected to comply with federal rules—as a condition for
receipt of federal money. Rules affect benefits,
provider payment,a host of administrative arrangements, and—as discussed below—beneficiaries’
“entitlement” to benefits. In recent years, states have
successfully sought waivers from and elimination of
rules that limit their ability to manage the federal
dollars they receive according to state, rather than
federal, priorities. To overcome state reluctance, SCHIP legislation explicitly increased federal matching rates and expanded state flexibility (on benefits
and, as discussed below, establishment of an individual entitlement).
Creation of a new public program aimed at covering adults would confront similar issues of limits
on certain states’ willingness to spend and to operate under federal requirements. Simply making federal funds available at Medicaid matching rates to
states willing to cover childless adults—who are
ineligible under current federal law, regardless of
income—might lead to coverage expansions by
some states. But other states—particularly those
that do not even take advantage of the existing
option to cover parents of Medicaid children—are
likely to respond only to an increase in federal
matching rates or, perhaps, full federal funding. And
a uniform response undoubtedly would require
establishment of a federal floor (as former President
Clinton proposed to apply to coverage for parents,if
states failed to act). Without such action, any new
coverage initiative would likely produce considerable variation in coverage across states.
Entitlement vs. Block Grant
Perhaps the most fundamental conflict over rules
attached to federal matching funds has been
whether the new coverage constitutes an individual
entitlement (as with Medicaid) or a benefit provided at the discretion of the state (as with S-CHIP).
Medicaid funds are available only as a federal entitlement—that is, everyone who satisfies eligibility
requirements is guaranteed covera ge . Under
Medicaid, federal financing follows the individual.
Although states can establish eligibility levels, determine how easy or difficult it is for people to participate, and affect how generous or restricted benefits
and access to care are, they cannot deny coverage to
an eligible individual. By contrast,S-CHIP is a block
grant that provides capped federal funds to states
and allows them to choose whether to create an
individual entitlement. States can choose to use the
new federal funds to expand Medicaid, thereby creating Medicaid-like obligations to individuals (and
assuring access to federal funds at the regular
Medicaid matching rates if the cap is exceeded). But,
if they prefer—as many have—states can create sepa ra te programs in which they can cap enrollment
and receive a capped federal allotment to help pay
for services.
54
This aspect of S-CHIP’s design was a critical element of the political compromise believed to be necessary both to enact the S-CHIP legislation and to
ensure state participation. However, deterioration of
economic circumstances, rising health care costs,
and strained state budgets could lead states to limit
en ro ll m ent by establishing waiting lists. To states
concerned about existing, let alone new, coverage
commitments, the option under S-CHIP that permits states to receive federal funds without an openended coverage guarantee is far more attractive than
Medicaid’s open-ended entitlement. The existence of
S-CHIP makes it likely that states would seek a similar option under any extension of public programs.
It is ironic that the absence of a federal entitlement in a public program expansion would contrast
sharply with the creation of a federal entitlement to
any new tax credit. Under federal law, anyone who
qualifies for a tax credit is entitled to receive it; obligations cannot be capped. A tax credit, like
Medicaid, is a federal entitlement (albeit to a dollar
amount, rather than to a defined set of benefits). If
Congress is willing to establish tax credits as entitlements—as is the case for all other tax subsidies—
consistency would suggest a similar approach to the
expansion of public coverage.
Establishing a Tax Credit along with a
Public Program Expansion
An individual tax credit aimed at people with
incomes too high to qualify for public coverage—
above, say,  percent of poverty—could reach its
target population without requiring any significant
va ri a ti ons from standard tax practices. At this
income level, tax liabilities are generally high
enough to make refundability unnecessary, and
individuals could simply apply for the credit retrospectively when they file their taxes. Cash flow problems are less severe than for lower-income
populations, or they can be mitigated easily by making funds available through standard tax withholding mechanisms (which people in this income range
are accustomed to using, for example, to account for
mortgage interest).
If the goal were to assure individuals with in-
comes above  percent of the poverty level that
they would have to pay no more than  percent of
their income for coverage—the maximum level
under S-CHIP—tax credits in the range of ,
for an individual and , for a family (phased
out gradually as income rises) would be required.
However, credits of this amount could prove disruptive to employer covera ge , and smaller amounts
would provide a smoother transition from public
coverage than exists today.
Policy issues posed by establishing an individual
tax credit include the following:
Ensuring Access to a Market or Product
Given that the bulk of uninsured individuals with
modest as well as low incomes lack access to
employer coverage,most of the beneficiaries of a tax
credit will be dependent on the non-group or individual insurance market. Expanding that market
may mitigate, but will not eliminate, the risk selection and instabilities it creates. These probl em s
could be addressed by regulating access to and premiums in the non-group market. However, efforts
to enact such regulations have run into enormous
barriers—both political and technical—at the state
and federal levels.A frequently proposed alternative
to regulation is to make tax credits applicable to premiums paid to a publicly managed insurance market, in which access, benefits, and premiums are
regulated. Medicaid could legitimately be considered such a market, given many state programs’
reliance on private insurance plans. Allowing credits
to be used in Medicaid would assure individuals eligible for a credit that a product was indeed available
to buy. However, allowing a buy-in to Medicaid in
the absence of broader regulation—or creating a
new insurance arrangement to accomplish the same
thing—would likely increase the costs of a public
program. The program is most likely to attract higher-risk and more costly individuals who are likely to
find public protection a better buy, given the underwriting practices of insurers. An extra subsidy
toward the premium for public coverage would be
necessary to ensure affordability for this population;
indeed, it could be thought of as a mechanism for
spreading risk with a broad source of financing
55
Balancing Access and Disruption
Deciding who gets how much of a tax credit would
determine the degree to which a credit would disrupt existing coverage. With respect to public covera ge , a decision to make the tax credit non-refundable—in other words, to layer a tax credit on top of a
public program by design—avoids the problem of
encouraging states to substitute tax credits for public
programs. Hence, disruption of public coverage
would not be a major probl em . With respect to
employer covera ge , the qu e s ti on is whether a tax
credit at the levels described is high enough to promote participation, but not so high that it promotes
substantial employer dropping and worker switching. The answer is uncertain.
One way to prevent the credit from leading
employers to drop or workers to switch coverage is
to allow it to be applicable to a worker’s share of
employer-sponsored covera ge , as well as to nongroup coverage. As noted above, such an approach
would actually substitute public for private dollars
as a means to secure employer coverage but also to
promote equity. Security and equity, however, come
at considerable public cost. Although a new credit
would induce some employees to take up coverage
for the first time,the bulk of credit recipients under
these rules would likely be individuals who already
have coverage, rather than the uninsured.
Providing the credit to employers rather than to
individuals might offer a more targeted means to
prevent the credit from inducing employers to drop
coverage—especially if eligibility for credits can be
limited to a subset of employers (like small, lowwage employers) that are currently unlikely to provide covera ge . A refundable tax credit could be
provided to employers in the subset that do provide
covera ge , of fs et ting corporate income taxes.15 For
example, each eligible employer could receive a flat
dollar amount—possibly varying for single or family coverage—for each eligible employee who is covered by health insurance. Previous efforts to induce
employer offering of coverage through subsidies at
the state level or through local pilot projects have
not been successful, but this may be because these
subsidies were either too small or perceived as temporary by employers.16 In fact, recent economic
analysis indicates that small employers are at least as
responsive as individuals to changes in the price of
insurance.17
Based on that analysis, subsidy levels and coverage expectations might be similar under the two
approaches. To induce offerings, an employer credit
would have to be at least as generous in relation to
premium costs as an individual subsidy, and it
would likely need to be refundable. Though an
employer credit would not help individuals without
a connection to an employer or uninsured workers
whose employers do not offer coverage, many of
these uninsured would be eligible for or could buy
into public covera ge . Furthermore, the employer
credit could also be extended to those self-employed
who may be uninsured despite higher incomes.
Though the primary goal of an employer tax
credit would be to encourage more employers to
offer health coverage, equity and ease of administration would require that it be made available to eligible employers who already provide insurance. These
employers would likely use at least some of the proceeds of the credit to lower employee premium contributions and, therefore, increase take-up among
currently uninsured workers. But because the vast
majority of workers who have access to employer
coverage already take it up, these resources would
likely go primarily to those already insured (for
example, in the form of higher wages).
Targeting an employer tax credit to those
15
Jack A. Meyer and Elliot K. Wicks. A Federal Tax Credit to Encourage
Employers to Offer Health Coverage. New York: The Commonwealth
Fund, 2000.
Subsidizing Employment-Based Health Insurance: Results From a Pilot
Study.” Journal of the American Medical Association 267 (7) (1992):
945-48.
Sharon Silow-Carroll. “Employer Tax Credits to Expand Health
Coverage: Lessons Learned.” Unpublished paper, The Commonwealth
Fund, 2000; K. E. Thorpe et al. “Reducing the Number of Uninsured by
17
(that is, general tax revenues). Unless resources are
available to support that subsidy, however, adverse
selection is likely to make the buy-in unworkable.
16
Jonathan Gruber and Michael K. Lettau. “How Elastic is the Firm’s
Demand for Health Insurance?” NBER Working Paper W8021.
Cambridge, MA: National Bureau of Economic Research, 2000.
56
employers least likely to offer insurance today would
maximize the focus of public dollars on improving
access to employer-sponsored coverage. The credit
could be limited, for example, to small, low-wage
employers,the majority of which do not offer coverage. Among firms with between three and  workers that have  percent or more of their workers
earning less than , per year, just  percent
offer health insurance (compared to  percent of all
firms with fewer than  employees).18 Smaller,
low-wage firms are even less likely to offer coverage.
Unfortunately, limiting subsidies to firms in this
category is not a perfect solution to the problem of
substituti on . Large firms can spin off low-wage
workers to create new small, low-wage “firms,”
thereby qualifying for subsidies they would not get
otherwise. In addition, a subsidy targeted at small,
low-wage firms may be perceived as inequitable,
since it is not available to larger low-wage employers. Nevertheless, directing subsidies to small, lowwage employers offers a reasonable approach for
targeting a tax credit to the uninsured, rather than
the already insured.
Because most economic evidence suggests that
individuals rather than employers actually bear the
cost of insurance (even if the employer ostensibly
pays), a subsidy provided to employers may be virtually identical in its effect to a subsidy provided to
individuals for the purchase of employer coverage.
But, in practice, focusing on employers may facilitate the targeting of credits based on employer characteristics (like small size and low wages) that are
associated with an absence of coverage offerings. As
a result, the primary beneficiaries would be uninsured workers, rather than workers who already
have insurance. Focusing on the employer would
also ease the cash flow problems posed by individual
subsidies, because by its nature, employer insurance
would guarantee coverage prospectively.
A tax credit for employer coverage rather than
for individual insurance has clear advantages in
terms of targeting and avoiding disruption of
18
Kaiser Family Foundation and Health Research and Educational Trust.
Employer Health Benefits 2000. Menlo Park, CA: The Henry J. Kaiser
Family Foundation, 2000.
employer coverage. But these advantages can be perceived as disadvantages by advocates of individual
choice, because the availability of subsidies to individuals would depend on the action of their employers. Undoubtedly, many individuals would not
become beneficiaries of the new policy. Only if there
is a buy-in to public coverage—as discussed above—
would these individuals have access to guaranteed
support in a combined public program/tax credit
initiative of this kind.
Summary: A Viable Merger?
Given the priority we place on covering the lowincome uninsured, the most effective approach to
expanding coverage is to extend the Medicaid and SCHIP eligibility now available to children and some
parents to all low-income individuals. Although
health insurance tax credits of various kinds could
be combined with this public program expansion,
careful attention should be paid to ensure that any
tax credits complement rather than su b s ti tute for
existing public and private sources of coverage. It is
difficult to design a modest individual tax credit
that is simultaneously effective and well targeted,
while at the same time avoiding disruption of existing publicly sponsored and employer-sponsored
insurance.A better approach may be tax credits provided to employers, rather than to individuals, to
encourage greater offering of insurance.
What is probably most important to the current
policy process is attention to the questions we have
raised about the way various policy instruments
and their application affect who will benefit from
and who will be hurt by adoption of any new policy
initiative. Given how difficult it has been to obtain
the public resources that are essential to expand
coverage, it is crucial that any resources that do
become available to expand coverage be used to
achieve that goal, especially for those least able to
protect themselves.
Acknowledgements
This paper was prepared in con ju n ction with the
Kaiser Family Foundation’s Project on Incremental
Health Reform. n
57
Gruber Proposal
Key Elements
Jonathan Gruber has outlined a proposal to substantially reduce the nation’s
uninsured rate and allow nearly all households to obtain affordable health
coverage under a voluntary initiative relying heavily on the private insurance
market. The program includes the following elements:
       -      that would offer a menu of
health plan choices to all individuals and employers.
      on a sliding-scale basis to
individuals with incomes up to  percent of the federal poverty level
buying insurance in the pool; most families with higher incomes could
obtain coverage for  percent or less of their income.
         by limiting the tax exclusion for
employer-provided health insurance to the cost of a median-cost health
plan and by phasing out the Medicaid program—and accompanying
federal subsidies—for those families that qualify on the basis of income
alone (while the program remains in place for the elderly and disabled).
          on both a prospective and
retrospective, risk-adjusted basis that would spread health risk across
entire purchasing pools so that higher-risk individuals could obtain
affordable coverage.
58
About the Author
   , .., is Professor of Economics at
the Massachusetts Institute of Technology and the
Director of the Program on Children at the National
Bureau of Economic Research, where he is a
Research Associate. He is a co-editor of the Journal
of Health Economics, and an Associate Editor of the
Journal of Public Economics. Dr. Gruber received his
B.S.in Economics from MIT, and his Ph.D. in Economics from Harvard. He has received an Alfred P.
Sloan Foundation Research Fellowship, a FIRST
award from the National Institute on Aging, and the
Kenneth Arrow Award for the Best Paper in Health
Economics in . He was one of  scientists
nationwide to receive the Presidential Faculty Fellow
Award from the National Science Foundation in
. During the – academic year, Dr. Gruber was on leave as Deputy Assistant Secretary for
Economic Policy at the Treasury Department.Dr.
Gruber’s recent areas of particular research interest
include the economics of employer provided health
insurance,the efficiency of the delivery of health
care to the indigent,the effect of the Social Security
program on retirement behavior, and the economics
of smoking.
59
A Private / Public Partnership for
National Health Insurance
by Jonathan Gruber
Overview
The private/public partnership approach to health
reform proposed in this paper builds on the popularity of a voluntary, private insurance system, while
rationalizing public “wrap-around” support for this
s ys tem. The plan’s central features are discussed
below.
Purchasing Pools. The central element of the proposal is a set of  voluntary purchasing pools established throughout the United States, one in each state
and the District of Columbia. The federal government establishes each pool and documents its catchment area. Then it establishes a set of ground rules for
any health insurance plan that wishes to be part of the
local pool. Any insurance plan that meets those
ground rules is eligible to be included in the pool,and
plans in the pool offer insurance to potential enrollees
at a community-rated price (by family type).
Individuals and Employers. Individuals or
employers are eligible to purchase insurance from
any plan in their local pool. This purchase is subsidized for lower-income families. All persons in families with incomes below  percent of the federal
poverty line are enrolled automatically and free of
charge in a plan near the pool’s median-cost plan.
All persons in families with incomes between 
percent and  percent of the poverty line (roughly
, to , for a family of four) receive a
subsidy to help pay to purchase insurance from this
pool. The subsidy caps the proportion of income
that must be spent to purchase insurance from the
median-cost plan in the pool;this cap rises from  at
 percent of poverty-level income to  percent of
income at  percent of poverty. These subsidies
can apply to direct individual or employer purchase
of insurance from the pool. Persons above  percent of federal poverty-level income receive no subsidy, but their insurance may cost them less because
they can purchase it through this pool.
Employers are allowed to purchase insurance
from the pool; they receive no direct subsidy if
insurance is purchased through the pool, but lowincome families can use their subsidies to help pay
for employer-provided insurance from the pool.
Employers and their employees are eligible for these
subsidies only if they restrict the employees’ insurance ch oi ces to plans offered through the pool.
Continuation of coverage mandates (through the
Consolidated Omnibus Budget Reconciliation Act
[COBRA] and state regulation) also are removed for
employers buying all their insurance through the
pool, providing even greater incentive for employers
to join the pool.
Plan Reimbursement. A key considera ti on with
any pooling approach is adverse selecti on . Plans
compete to attract the lowest-risk enrollees, which
can raise prices significantly at the more generous
plans that some truly sick enrollees may demand.
Adverse selection in these pools is minimized
through risk-adjusted redistribution across the plans
in the pool. This risk adjustment is a mix of prospective (for example,based on demographic characteristics and long-term comorbidities) and retrospective (for example, actual cost outliers) factors.
Financing. This approach involves a significant
federal expenditure, primarily through subsidies for
the low-income insured. These costs are partly
financed from two sources. The first is a cap on the
exclusion from taxation of employer-provided
health insurance premiums. Any currently tax-preferred spending by employers and employees on
60
insurance above the cost of the median plan in that
state becomes taxable income to the employee.
After a transition period, the second financing
source is the phase-out of current public programs
that provide insurance to low-income families solely on the basis of income. Medicaid ultimately
becomes a program only for the elderly and disabled, and the remaining Medicaid and State
Children’s Health In su ra n ce Program (S-CHIP)
populations move into this new subsidy program.
The federal government then saves its share of
spending on these programs and recaptures from
states their spending on these programs.
Implications. This private/public partnership
results in a very different situation for individuals,
insurers, and employers. In su ra n ce coverage rises
significantly, as affordability, information, and stigma barriers to insurance for the lowest-income families are removed. Coverage is not universal, but
almost all families in the United States should be
able to buy insurance coverage for  percent of
their income or less. Insurers offer their products in
a competitive environment that provides strong
incentives for cost control. Individuals pay the full
costs for choosing more expensive insurance products, because the tax subsidy to purchase insurance
is capped. This leads consumers to choose costeffective plans, allowing for increased medical-sector cost control without public spending caps or
other awkward interventions.
Background
The proposal developed in this report is designed to
meet two key political constraints and to address the
two key failings of our current system. The political
constraints are that the U.S. Congress is unwilling to
expand public insurance programs massively, or to
legislate widespread new individual or employer
mandates. The first failing is that employer-provided
insurance results in incomplete access to pooling
mechanisms for such groups as the unemployed,selfemployed, and those in small businesses. The second
failing is that public safety net programs cannot provide health care to everyone who is uninsured. These
failings are discussed in more detail below.
Private Health Insurance
Several features of the current insurance environment call for the private/public partnership
approach. The first is that the primary source of
health insurance is employers. More than  percent
of the privately insured, representing  percent of
the total non-elderly population, are covered by
employer-provided insurance. Employer provision
has much to recommend it. Workplaces of sufficient
size represent pooling mechanisms that are largely
independent of underlying health status, providing
the kind of predictable distribution of costs potential insurers want. They also provide a means of
spreading the fixed costs of an insurance plan across
a number of insured persons. In addition, human
resource departments provide professionals generally dedicated to effective provision of benefits,
leading to both high satisfaction and innovation in
health insurance options.
On the other hand, employer-provided insurance as the primary mechanism of i n su ra n ce has
some failings.First, small employers provide neither
the economies of scale nor the needed predictable
distribution of expenditures that make them attractive sources of insurance. As a result, while insurance
offering is nearly universal among medium-size and
large employers, it is much less common among the
smallest employers; even in our booming economy,
only two-thirds of employers with fewer than 
employees offer health insurance.
Second,the fact that insurance is linked to work,
and is not available at all jobs,can lead to insuranceinduced immobility across jobs, or “job lock.”
Workers who value insurance coverage may not
leave their current positions for poten ti a lly more
productive ones, for fear of losing their insurance
coverage.Estimates suggest that job lock may reduce
mobility by as much as  percent among those with
employer-provided insurance. Job lock is mitigated
by the availability of continuation coverage under
state and federal mandates that allow individuals to
continue to purchase insurance from their employer
after leaving their jobs,at  percent of the employer’s full cost of insurance. While research has shown
these continuation mandates to be an effective
means of increasing insurance coverage among job
61
changers, they are unpopular with employers
because of their administrative costs and the (below
average) health of employees who choose to continue their coverage (so that employers lose money,
even though workers pay average insurance costs).
Finally, the residual nature of the non-group
market has made it an inhospitable environment for
those who leave the employer pool. The non-group
market features high prices, typically at least  percent higher than group insurance costs for healthy
employees, and much more for older and less
healthy persons. Those with expensive medical conditions may be unable to obtain any coverage, and
the coverage available in the non-group market generally is much worse than group coverage, with high
deductibles and limited benefits.
A final issue related to employer-provided insurance is that its popularity can be traced, at least in
part, to exclusion of employer-provided insurance
payments (and roughly half of employee payments
for such insurance, as well) from taxation. This tax
subsidy has been criticized as regressive: Because it
is equivalent to a tax deduction, those who pay the
highest income taxes benefit the most. It is also cited
as a source of medical cost inflation, because it subsidizes the price of health insurance and can lead
individuals to purchase excessively generous insurance plans. Finally, this is a major expenditure for
the government, more than  billion per year.
Public Health Insurance
In principle, public health insurance in the United
States is designed to insure those unable to get coverage from the private insurance market. For the elderly, who are covered by Medicare, this is largely true
(with some notable exceptions, such as the lack of
coverage for pre s c ri ption drugs). But some of the
n on - el derly still have difficulty getting covera ge ,
which results in our high and rising level of uninsurance. In , . percent of non-elderly Americans
had no health insurance. Over the next decade, the
non-elderly population without insurance coverage
grew by nearly a quarter, to . percent, so that in
  million Americans were uninsured. This
number declined to . million in . Particularly
troubling is the significant increase in the number of
children in the United States who are uninsured;
despite dramatic expansion of public health insurance since the mid-s,the share of children without health insurance has grown by more than  percent since .
Medicaid is the primary source of public insurance. Most Medicaid spending is for insurance for the
elderly and disabled, but most of the individuals covered are women and children. Traditionally, only
those on cash welfare were eligible for public insurance, but this coverage has now been extended dramatically for two groups: pregnant women (for pregnancy-related expenses only) and children.
Currently, pregnant women are automatically covered up to  percent of poverty-level income by federal mandate, and most states have extended this coverage to  percent of the poverty level or above.
Children under age  are covered up to  percent of
poverty level, as well, and most states cover all children to the poverty level or higher under their Medicaid programs. Moreover, the  S-CHIP extended
child coverage further by providing block grants to
states, which many states have used to extend child
coverage to  percent of the poverty level or even
higher.
Despite these recent expansions, however, enormous holes in the public safety net remain. First,
and most obvious, there is no source of public
insurance (other than selected small state programs) for adults, aside from pregnant wom en .
Second, even among eligible populations, the number of people taking advantage of this public insurance entitlement is low. Recent estimates suggest
that as many as  million uninsured children may be
eligible for Medicaid or S-CHIP, but are not taking
advantage of this eligibility. This is likely due in part
to the fact that entire families are not eligible, limiting the incentives for participation, and among
middle-class families there is some stigma attached
to using public programs. Despite the fact that
Medicaid remains an entitlement for low-income
families leaving welfare, recent declines in public
coverage resulting from welfare reform highlight the
difficulties facing the safety net.
62
Details of Implementation
An enormous number of details must be addressed
when implementing a plan such as a private/public
partnership. Important questions about such a plan
and at least partial answers to them follow.
Eligibility and Subsidy Structure
Based on the income reporting described below,
individuals are eligible f or one of two kinds of subsidies. If income is below  percent of the federal
poverty line for that person’s family size (“poor”
families), then the individual and his or her family
members are automatically enrolled free of charge
in an insurance plan. The plan is selected randomly
from among the plans near the median-priced plan
in the pool. The use of a default plan is critical to the
success of this approach, because it will increase
take-up of insurance by this low-income group significantly. But placing the whole group in one plan
that happens to be at the median is potentially
inequitable and problematic if the particular plan
cannot handle this many enrollees. Therefore, this
low-income group is assigned randomly to a small
number of plans near the median;the exact number
of plans depends on the size of the pool and the
range of prices around that median.Of course,these
individuals are free to choose a different plan from
among those close to the median, and to switch
among these plans once they are assigned. The pool
administrator notifies plans of any enrollees in this
group, and the plan bills the government directly,
rather than the individuals, for the premiums.
A second subsidy for individuals in this income
range is a cap on the copayments and deductibles
for which they are responsible. Individuals enrolled
in these plans are subject to maximum copayments
of  for any visit or drug purchase. In addition,
these copayments, and any deductible,are capped at
 percent of income. Providers notify the plan,
wh i ch , in tu rn , notifies the pool administrator,
whenever a copayment is charged to someone in the
poor group (who is identified by having a separate
insurance card). When persons reach the  percent
of income limit, they are sent a new card indicating
that they are no longer to be charged copayments.
Individuals in this income range are also free to
choose plans that cost more or less than the median.
If they choose plans at above-median cost, they are
bi ll ed by the plan for the difference between that
plan’s premiums and the premiums of the median
plan. If they choose lower-cost plans, they do not
receive the difference. If these cost savings were
available,there would be significant opportunity for
fraud; low-cost plans might be set up that do not
actually provide insurance, but that just allow lowincome individuals to turn their subsidies into cash.
Because copayments and deductibles are capped for
the low-income population, establishing such “cash
cows” would be very easy.
The second subsidy group, those families
between  percent and  percent of the poverty
line (the “near poor”), receives a subsidy that is
structured so that, if they sign up for the mediancost plan, they will never pay more than  percent of
their family income on insurance premiums. The
subsidy is phased in, so that there are no large reductions in subsidy as income grows (that is, to avoid
large “implicit taxes” on income generation for families in this income range). In particular, at  percent of the poverty line, the cap is zero, with full subsidies; at  percent of the poverty line, the cap is
. percent of income; at  percent of poverty,
the cap is . percent of income; and by  percent
of poverty, the cap is a full  percent of income.
For example, suppose that a family has four
m em bers and an income of , ( ro u gh ly 
percent of the poverty line), and that family coverage in the median-cost plan costs ,. That family receives a subsidy for ,, the difference
between the cost of the median plan and . percent
of the family’s income. Individuals are then free to
enroll in higher- or lower-cost plans as they wish,
but the subsidy amount remains at ,, regardless
of the plan chosen.
For this near-poor group, if individuals do enroll
in a plan that costs less than the subsidy amount,the
government will pay them  percent of the difference between the subsidy level (which is tied to the
median plan) and the premiums in the plan they
choose. This provides some incentive to choose lowcost plans, while potentially offsetting some costs of
63
Those families between 150 percent and 300 percent of the
poverty line (the “near poor”) receive a subsidy that is structured
so that, if they sign up for the median-cost plan, they will never pay
more than 10 percent of their family income on insurance premiums.
this subsidy program to the government.
For this near-poor population (and for any poor
individuals who choose above-median-cost plans),
the pool administrator would notify the plan of the
enrollees’ information and how much of a subsidy
they are entitled to receive. The plan is then responsible for collecting the difference from the individual, and the subsidy is paid directly by the
government to the plan. So, using the example
above, the insurance plan bills the individual ,
per year, and the government, ,. If the individual chooses a cheaper plan—with a premium of
, per year, for example—then the plan bills the
individual only  per year and continues to bill
the government ,. If the individual enrolls in a
plan with premiums of , per year, then the
plan bills the government ,, and the government pays the individual a subsidy of ..
Income for these purposes is a modified version
of adjusted gross income (AGI) that includes all
income elements, but does not exclude from income
deductions from AGI that are included in the current tax code (for example, the ability to deduct
contributions to retirement savings accounts). So
this corresponds to a gross income concept.
-  
A technical, but absolutely critical, issue of subsidy
design is how eligibility is determined. There are
two models to choose from. The first is a refundable
tax credit/voucher system, with reconciliation.
Under this system, individuals apply for subsidies
before the plan year, using either their previous
year’s income or a projection of their income for the
coming year. They then receive those subsidies for
that year. The following spring, there is a reconciliation process between the income they actually
received during that year and the income they anticipated  months earlier for subsidy determination.
As a refundable tax credit, this approach faces the
additional problem of advancing money to individuals  months earlier so they can purchase the
insurance. As a voucher scheme, however, the payments are advanceable by definition.
The second approach is more like welfare.
Individuals apply for subsidies more frequently, and
report their income when they apply. If they qualify,
they become eligible for that period. There is no reconciliation, although some mechanism must be in
place to catch significant dishonesty in reporting
income.
The fundamental difference between these
approaches is the reconciliation process. Reconciliation does provide a more natural means of correcting over- or underpayments than a backstop fraud
mechanism does, but it may significantly deter participation by potential enrollees. For example, taxpayers can take advance payment of their Earned
Income Tax Credit amounts, but only about  percent of potential recipients do so, partly out of fear
that they will underestimate income and owe taxes
the next April . As a result, a refundable credit/
voucher approach could deter take-up significantly,
for fear of reconciliation costs down the road.
For this reason, the private/public partnership
approach adopts a more welfare-like approach.
Every six months, at open enrollment in November
and May, individuals are asked to verify their
income. Supporting documents are required, such as
pay stubs or W forms. If individuals qualify for one
of the two subsidy programs, they are guaranteed
those subsidies, subject to penalties for fraud, but no
reconciliation. A significant enforcement program
will be in place to ensure that individuals do not
abuse this presumptive eligibility by systematically
understating their incomes. But prospective
en ro ll ees can rest assured that honest mistakes
and/or changes in income will not result in penalties.
64
Subsidies are also available on a shorter- ru n
basis for those experiencing income fluctuations
within a six-month period. If individuals can offer
proof of income loss (such as unemployment),they
can enter the subsidy pool at any point, and from
then on, are on the regular six-month schedule.
the additional costs (of course, those costs may be
passed on to employees, either directly, through premium sharing, or indirectly). For the near-poor (
percent to  percent of poverty), the employer is
billed for the cost (minus the subsidy) of the plan
chosen by the employee.
   
All employers are also allowed, but not required, to
purchase their insurance through the local pool.
Employers may provide their employees a menu of
insurance options that includes some plans from the
local pool, and some from outside the pool. There
are two significant incentives for employers to limit
their employees’ choices to plans in the pool, however. First, only those employers that restrict
employees’ ch oi ces to plans available through the
pool can access subsidies for their low-income
employees. Second, employers that purchase all of
their insurance through the pool are no longer
required to provide continuation coverage, because
individuals can now purchase insurance through
the pool on their own. These incentives for employers to en ro ll all their employees in the pool are
designed to minimize adverse selection into the
pools (discussed further below).
Employers purchasing insurance through the
pool effectively act as intermediaries for individual
purchase. That is, employees enroll through their
employers, perhaps using additional materials provided by employers to help them choose.Employers
then withhold the premiums for the chosen plan
from the worker’s paycheck and remit that amount
directly to the pool, which, in turn, reimburses the
plans. Using the pool as a middleman between
insurers and employers makes it easier for employers to take advantage of the pool, providing another
incentive for employers to use it as their source of
insurance.
Low-income employees who obtain their insurance through the workplace apply for subsidies
through their employer. For the poor (incomes
below  percent of the poverty line),the employer
pays no premiums if the employee enrolls in the
median-cost plan. If the employee en ro lls in an
above-median-cost plan, the employer has to pay
Financing: The Employer Tax Subsidy and the
Role of Public Insurance
This plan is financed from three sources. The first is
general revenue financing, one hopes,from the projected federal budget su rp lu s . But this program’s
cost to the general budget ultimately will be offset
by the following two sources of savings.
    
 
Part of the financing for this plan will come from
the limitation of the tax exclusion for employerprovided health insurance payments. This limitation will take two forms.First, for those low-income
employees receiving subsidies, net employer payments for health insurance will be lower. Both economic theory and evidence suggest that the lower
net employer payments will be passed on to workers
in the form of higher wages, and, therefore, higher
taxes. Thus, in essence, we will end the tax subsidy
for employer-provided insurance payments to poor
employees, and limit this subsidy for the near-poor.
Second, the government will limit the tax exclusion for employer-provided health insurance payments explicitly to the cost of the median-cost plan
in the local pool. All employers will track the total
payments they make and any pre-tax payments
made by their employees for health insurance. The
government tells each employer before the beginning of each year the cost of the local pool’s mediancost plan. The employer is then responsible for
reporting as part of an employee’s wages and
salaries the difference between the total pre-tax
employer/employee expenditure on health insurance and the premiums for the median-cost plan.
This difference becomes taxable income for the purposes of both the income tax and payroll tax system.
This is true regardless of whether the employer purchases insurance through the pool.
65
   
 - 
This program obviates the need for much of the
existing Medicaid program by providing insurance
for low-income populations, so the program can be
reorganized. The portion of Medicaid that provides
health insurance coverage for non-disabled and
non-elderly families—acute-care Medicaid—can be
abolished, because those families can be automatically enrolled in their local pool, or offered a significant subsidy toward enrollment. The remainder of
Medicaid that provides health insurance coverage
for the disabled and the elderly (the bulk of
Medicaid costs) remains unchanged. The S-CHIP
program also can end under this proposal, because
insurance for low-income families is available
through the subsidy mechanism described above.
This reorganization ultimately will result in considerable cost savings that can help to finance the
new program. In addition to the federal government’s savings from reducing its Medicaid and SCHIP obligation, states can save considerable
m on ey by no longer paying the Medicaid and SCHIP costs for their acute-care population, while
the insurance coverage for this low-income population actually increases. To compensate for the loss of
these payments, the federal government receives a
transfer from each state equal to the amount the
state was paying for the Medicaid and S-CHIP programs for the acute-care population in the year
before the program was enacted. This payment rises
over time with the cost of the median plan in each
state. Once again, the majority of the program that
applies to the disabled or elderly does not change,
and no recapture applies to those funds.
There are two important concerns with removing the public insurance entitlement,however. First,
the net insurance entitlement of some families
could fall, since children now receive free coverage
under Medicaid up to or above  percent of
poverty in many states. But this should be more
than of fs et by the subsidized or free coverage of
adults in most families.
Second, and more important, there could be
major disruption of insurance during the transition
from the current public system to this new private
s ys tem. Low-income individuals who are taking
advantage of the public system may be confused or
otherwise unable or unwilling to use this new private approach. An unfortunate byproduct of this
attempt to increase insurance coverage might lead
to displacement of the neediest who are currently
publicly insured. Therefore, a transition period will
be necessary during which both the existing
Medicaid/S-CHIP programs and the subsidies for
the public/private partnership are available. The
length of this transition period can be based on evidence of understanding and willingness among lowest-income families to move to the new system. The
phase-out of public programs will occur in a “topdown” fashion, with eligibility for the highestincome groups currently in the program phased out
first (starting with S-CHIP).
A concerted outreach and public relations effort
to establish an effective understanding of the new
subsidy system will help this transition. This effort
includes working through existing Medicaid offices,
schools, day care centers, and other access points to
reach low-income populations.
Administration and Regulation—Benefits and
Risk Adjustment
Of key importance is how these pools that form the
core of the private/public participation approach
are established, administered, and regulated. In
addition, a central feature of this proposal is risk
adjustment among plans in the pool. This section
addresses implementation issues in pool administration and regulation.
    
The federal government notifies all potential insurers of the option to offer insurance through the pool
one year before the local pool is open for enrollment. The government bears the full cost of this initial solicitation, screening potential insurers in line
with the conditions outlined above, and initially
assigning applicants to plans within the pool.
Continuing administrative costs of the pool are
financed by the small fee assessed on premiums
earned by plans in the pool.
The government administers the program out of
66
There is a key trade-off in setting minimum standards for plans participating in
local pools. On the one hand, the federal government is obligated to make sure
that these are real insurance products. On the other, the government should
encourage individuals to choose efficient and low-cost health insurance plans.
a new agency, the Private/Public Partnership Health
Insurance Agency (PPPHIA), that is responsible for
establishing pools and overseeing existing pool
administration, coordinating subsidy payments,
and coordinating income reconciliation.
   
There is a key trade-off in setting minimum standards for plans participating in local pools. On the
one hand, the federal government is obligated to
make sure that these are real insurance products. On
the other, the government should encourage individuals to choose, or at least should offer the option to
choose, efficient and low-cost health insurance plans.
These joint imperatives dictate a fairly minimal
set of regulations that should guarantee that all the
insurance products offered are real insurance, but
then allows free ch oi ce . In particular, regulations
should require only that each plan feature:
• guaranteed issue and guaranteed renewability;
• coverage of physician services, inpatient and
outpatient hospital services (including emergency
rooms), and prescription drugs; and
• no or nominal copayments for one well child
visit per year, prenatal care, and immunizations.
These minimum standards should generate
rough comparability across the benefit packages
offered to plan enrollees. But variation in benefits,
or in the value of insurance plans, will remain along
four dimensions. The first is copayments and
deductibles;there is no minimum standard for these
patient charges, although, as noted above, they are
subsidized for the poor. The second is variation in
benefits around these minimum standards. For
example, plans may or may not use a formulary to
dispense pre s c ri ption dru gs ; may vary and limit
their outpatient and inpatient mental health coverage; may or may not cover home health care services, etc. These are not trivial differences, but, once
again, it is critical to reflect diversity of consumer
preferences across plans. The third is va ri a ti on in
provider networks offered among managed care
plans, and the fourth is va ri a ti on in the degree of
management of managed care plans,in terms of utilization review and physician financial incentives,
among others. As noted below, information about
all of these variations is readily available to families
during open enrollment periods.
In addition, the government must develop
financial soundness criteria to ensure that the plans
can provide their promised services. Subject to these
c ri teria, any plan that wishes to offer its servi ce s
through the pool may do so.
Each plan charges community-rated premiums
for each of four distinct populations: single; single
with children; married without children; and married with children. The use of com mu n i ty - ra ted
premiums immediately raises concerns about
adverse selection, which are addressed below.
 
To miti ga te adverse selecti on , individuals are
allowed to choose a plan at only one time during the
year. Open enrollment takes place during November
for the next year, allowing pool administrators the
month of December to process enrollment applications and assign enrollees to plans. Every family in
each plan’s local catchment area receives a mailing
on November  detailing that family’s insurance
ch oi ces for the coming year. The mailing has two
components. The first documents each plan’s costs,
reported after subsidy by income level, copayments
and deductibles, and services covered. The second
component provides more detail on the plans themselves. It includes information on provision of preventive care services and consumer satisfaction. In
addition, it gives some details on provider financial
incentives. There are also links to a web site where
67
individuals can learn more about each insuring
entity and the plans themselves.
Assessing what inf ormation should be disclosed
in the mailing, on the web site, or not at all is the
su bj ect of much debate, particularly surrounding
financial incentives to providers. While detailed disclosure of provider incentives can improve the
information available to very educated consumers
significantly, these details may be more than most
consumers need or want to know. Moreover, there
are competitive concerns in mandating too much
detail on such provider compen s a ti on arrangements, because part of how plans compete is over
their provider incentive structures. As part of establishing these local pools, therefore, a commission of
experts should meet to decide on the appropriate
amount of disclosure.
Plan switching is allowed during the year.
Individuals may switch during open enrollment
peri od , but if they do not return their forms
expressing their desire to change,they are automatically en ro ll ed in the same plan for the fo ll owing
year. Individuals who join the pool during the year
can choose their plan at that point. But to minimize
churning, any individual who leaves the pool at any
point during the year cannot reenter the pool until
the next open enrollment period.
   
The pool administrator co ll ects the open enrollment forms and informs each plan of the pool of
enrollees for the coming year. Plans are responsible
for billing enrollees and are entitled to terminate
coverage of any en ro ll ee who does not pay after
three months (for example,if a bill is sent at the end
of January, and is not paid by the end of April, the
family can be disenrolled). Disenrolled families are
barred from reentering the pool for three months.
Plans also are entitled to charge interest on all premium payments not remitted within one month, at
a rate set by the government to represent the borrowing costs of insurers.
After forms are co ll ected, the administrator is
responsible for notifying the government of all subsidy payments. The administrator also maintains a
database of information on all enrollees,plan choic-
es, and reported incomes, which is also shared with
the government.
-  
As noted earlier, risk adjustment is redistributed
across the plans in the pool to minimize adverse
selection. This risk adjustment represents a mix of
prospective (for example, based on demographic
characteristics and long-term comorbidities) and retrospective (for example, actual cost outliers) factors.
More specifically, when individuals enroll each
November, they provide information about their
a ge , sex, and incidence of a set of chronic or past
major illnesses (for example,diabetes, hypertension,
heart disease or stroke, etc.). This information is not
given to plans, but is maintained by the government. Even if individuals enroll through their
employer, they send this information directly to the
government to maintain confidentiality. In addition, at the end of each year, each plan reports the
costs for each enrollee to the government.
Based on these two sets of data, the government
applies a formula to determine a set of cross-subsidy
payments that flows across plans. This formula uses
demographic and comorbidity information to form
a predicted average health expenditure. It then takes
a weighted average of that predicted health expenditure and the actual health expenditure per capita
(the “cost index”). For each plan, the government
tabulates its cost index and redistributes funds from
the low-cost- to high-cost-index plans. Plans that
leave the pool are still eligible to receive payments
and are responsible for making payments for services incurred the previous year. The magnitude of the
redistribution is determined by technical government analysis.
This technical analysis trades off two considerations: more redistribution means less adverse selection, but also lower incentives for cost control. The
optimal redistribution scheme does not compensate
plans fully for differences in expenditure patterns,
but does compensate them enough to limit incentives for adverse selection. For example, the optimal
plan could state that any expenditures that are more
than one standard deviation from the mean for an
age/gender category will be reimbursed through this
68
redistribution system. This sort of approach still
offers incentives to keep costs down when close to
the mean, while “insuring” firms that en ro ll cases
with very high costs.
This risk adjustment does not address a different
type of adverse selection risk: adverse selection into
the pools themselves (“inside/outside adverse selection”). Since insurance is community rated, and
plans that are less comprehensive are taxed by the
risk-adjustment mechanism (higher premiums),
there is a strong incentive for healthy individuals to
remain outside this pool and in groups of healthy
persons with very low insurance premiums. By the
same logic, there is strong incentive for the sickest
individuals to get into the pool, where insurance is
subsidized for them more than it would be in the
experience-rated and non-risk-adjusted private
market. Enough adverse selection of this type could
destroy this pooling mechanism; if only the sickest
persons in society end up in the pools,the insurance
in these pools will be so expensive that they will be
unattractive to all but the most highly subsidized
poor.
This approach is designed to minimize this type
of inside/outside selection, however, because the
nature of subsidies (and discontinuation of the
coverage mandate) provides a strong incentive to
be in the pool. In the long run, all of the poor and
most of the near-poor should be in the pool.
Moreover, any employer that has a sizable share of
its workforce in the income range to which subsidies apply (which should be most employers) will
forgo a significant financial subsidy to its employees by not joining the pool. Remember, employers
have to be entirely in the pool for their workers to
receive subsidies, so they cannot “dump” their sickest workers into the pool while keeping healthy
workers outside it. Low-income employees ultimately should choose to leave employers that are
unwilling to join the pool, thereby putting pressure
on firms to enroll. And the removal of the deeply
unpopular COBRA mandate provides an additional incentive to pool.
It is difficult to assess whether this impetus is
powerful enough to get a critical mass of healthy
persons into these pools. But it seems quite likely
that it is, given the sheer size of the population to
which subsidies apply. In short, the goal here is to
use these subsidies to boost pool size to the minimum level necessary to lower costs and miti ga te
inside/outside selection significantly.

The transition to the private/public partnership is
fairly straightforward, since the pools are voluntary.
For example, suppose implementing legislation is
passed in October  that establishes the PPHIA,
which immediately begins drawing up local boundaries for the pools and gathering data from plans
that want to participate for calendar year .
Then, in November  the first enrollment period
can open.
Beginning January , , the government
begins to phase out Medicaid coverage for the nondisabled and non-elderly. Also on that date, employers begin to include the “excess” (above-areamedian) costs of their spending on health insurance
as part of taxable wages.
Advertising to individuals about this new insurance system begins immediately with passage of the
legislation in late . Medicaid administrators are
responsible for ensuring that all non-elderly/nondisabled enrollees are aware that their Medicaid
entitlement is terminating, and for introducing this
new alternative.
Implications of the Private/Public
Partnership Approach
This approach represents a fairly radical departure
from the current private and public systems of providing health insu ra n ce . While the impact of this
new system on the scope and shape of the health
care delivery system is difficult to predict, this section discusses some likely implications of this type
of reform.
Politics
The primary implication is the political dynamic
surrounding the type of major reform envisioned by
this approach. Obviously, any intervention of this
magnitude faces a daunting legislative process.
69
There are likely to be concerns from at least five
stakeholders about this approach. First, firms and
unions will be upset about the reduction in the tax
subsidy to employer-provided health insurance.
There will be a “camel’s nose under the tent” concern about capping this deduction, even if the cap is
relatively modest at first. Second, fiscal conservatives will object to the net price tag of this interventi on , particularly given the tenuous nature of
current surplus projections. Conservatives may also
be upset about the attendant increase in government bureaucracy. Third, insurers that focus on the
non-group market, but do not feel that they can
compete effectively in this new pooled group market, will protest their loss of market share. Fourth,
advocates of the traditional Medicaid program,
rather than private market solutions to the uninsured, will raise concerns about the loss of Medicaid
entitlement. Finally, tax administrators may oppose
the expanded use of the tax system under this plan
and, in particular, the introduction of area-specific
adjustments to the tax subsidy to health insurance
(which is capped at the median-cost plan in the
area).
But important stakeholders will support this
plan. Foremost will be reformers who see this as a
means of reducing the number of uninsured significantly. In addition, this approach will be validated
by market-minded advocates of competition as the
best source of health care cost control. And advocates for the poor will recognize the important
income redistribution of this approach. While upset
about the limitation of the tax subsidy as a means of
financing, employers (particularly small employers)
ultimately may approve of this approach, because it
allows them to shed their insurance provision obligations or b uy into a more effective mechanism for
purchasing insurance,and to discontinue an obligation they consider a significant burden. Group
insurers should also approve of a system that
expands the reach of their products, both locally
and nationwide.
More relevant is that this approach has more
po l i ti c a lly attractive fe a tu res than many other
alternatives. The significant expansion in public
spending programs necessary to cover an enormous
share of the existing uninsured is not feasible in
today’s pro-private solution climate. And alternative private-sector solutions, such as expanded tax
credits, face very high costs because of the limitations of the non-group market in which they would
be spent.
Implications for Health Care Costs
The lull in health care cost inflation in the United
States over the past few years has dictated a focus on
uninsurance and a backlash against the stringency
of managed care. But significant increases in health
care costs over the past year once again have raised
concerns about cost containment. At the same time,
there is little taste among the public or policy makers for a public cost-containment strategy for the
privately insured.
A key advantage of the public/private partnership is that it uses a competitive mechanism to assist
in cost containment, while balancing the selection
incentives inherent in competition through a mixed
prospective/retrospective risk-selection adjustment.
Unlike today’s insurance marketplace, individuals
will face the full marginal cost of moving from less
to more generous insurance plans. This makes these
individuals more cost-conscious shoppers, which,in
turn,puts pressure on insurance plans to lower their
costs to attract new enrollees.
Concern about competition as a source of cost
control in the health care sector is twofold. First,
there is some fear that competition will lead to inappropriate reductions in the quality of care. The best
safeguard against this is providing complete information on plan characteristics, financial incentives,
and con su m er satisfacti on , which will be done
through the open enrollment mailing. Second,plans
may compete,not to provide the most efficient care,
but to select the best risks. This concern will be miti ga ted through the risk-adjustment mechanism
described above.
The net impact of improved competi ti on on
health care cost growth is unknown. But competition, with full information, risk adjustment, and a
level playing field across plans and consumers,
remains the best politically feasible option for controlling costs.
70
Equity
This approach has significant implications for both
“vertical” equity (redistribution) and “horizontal”
equity (fairness). In terms of vertical equity, as
described above, this program involves significant
redistribution from average-income taxpayers to
those below median income (and primarily below
 percent of the poverty line). In terms of horizontal equity, this program removes many of the
inequities that now haunt our private insurance system.Large employers,small employers,and individuals will be able to purchase insurance on the same
basis, removing the existing enormous differences
across these groups because of administrative loads
and adverse selection premiums. Likewise, healthy
and sick individuals have equal opportunities to
purchase insurance from the same pool of
providers. Sicker individuals may end up buying the
more expensive plans in the pool, but this is by their
own choice, not by the active selection efforts of
insurers.
Implications for Income Generation
A concern with any program that includes incomerelated subsidies is the distortions toward income
generation. The proposed system includes two such
distortions. First, there is an additional implicit tax
on income earned above the poverty line as the full
subsidy available to those below poverty is phased
out. The magnitude of this implicit tax is modest,
however, amounting to only a . percent additional tax on income generation. For example,
consider a family of four with an income at  percent of the poverty line (roughly ,) that faces
group premiums of ,. That family’s subsidy is
the full group premium, or ,. If the family’s
income rose by  percent of the poverty line, to
,, their subsidy would fall by ., or .
percent of the income rise. This is a very modest
additional tax and is likely to cause little distortion
to income generation.
The second distortion is where the subsidy ends,
at  percent of the poverty line. If, at this point,
any families have premiums that exceed  percent
of their income, then raising incomes above 
percent of the poverty line could lower net resources
significantly. If, for example, group premiums for
the family of four were ,, then at  percent
of poverty (,), they would pay only , for
their coverage, and receive a subsidy of . But,
when the family earns the next dollar of income,
that en ti re  subsidy disappears, which could
present a significant disincentive to moving out of
the subsidized range.
This concern raises a t rade-off between limiting
the subsidies and extending them further up the
income range, which would mitigate this distortion
but also would raise costs. I propose erring on the
side of saving costs and limiting the subsidies for
two reasons.First,at higher income levels,income is
more dynamic (and less easily tracked, since higher
shares are from non-wage sources),so that administering these subsidies is more difficult. Second, the
magnitude of this distortion is likely to be relatively
small. But the implementing legislation for this policy should include a mandate to study this issue and,
in particular, keep track of the rise of health care
costs relative to incomes and how this affects the
magnitude of the “notch” at  percent of poverty.
Regional Variation
Another concern about income subsidies is that the
federal poverty line is not tied to regional variations
in cost of living. It seems highly unlikely that there
would ever be regionally rated subsidies, given the
enormous difficulties of assessing the correct
regional adjuster. But the structure of the subsidies
in this proposal provides implicit ad ju s tm ent for
regional cost variation, because the amount of the
subsidy is tied to the cost of the median-cost plan,
which reflects regional variations in the cost of living. That is, a much larger share of the population
will be subsidized in Mississippi than in
Massachusetts, but the subsidies for which they are
eligible will be much smaller.
Conclusions
Radical change of the U.S. health care system
involves a set of difficult trade-offs from both an
economic and political perspective. The private/
public partnership approach laid out in this propos-
71
al is designed to respect the demand for voluntary,
private solutions to the problem of the uninsured in
the United States in a way that could provide insurance to the vast majority of needy uninsured in this
country.
But it is important to recognize that this proposal will not lead to universal health insurance
coverage in the United States. Even with significant
subsidies to the poor and near-poor, some still will
choose to remain uninsured. In addition, a not
inconsequential number of well-off people who can
afford health insurance will continue to choose not
to buy it. This raises a critical question of how far
we are willing to go with health policy in the
United States. This proposal guarantees that coverage is universally affordable, but not universally
adopted. Is the role of the government to go beyond
this to ensure universal coverage, or is universal
access sufficient?
The reform has a number of additional virtues.
It involves significant income redistribution and a
leveling of the playing field on which individuals
purchase insurance. And it uses the powers of competition effectively to address rising health care premiums. There are additional complications, as well,
and these need to be addressed if the program is
implemented. But the key political economy advantages of a voluntary system suggest that this type of
approach has promise as a means of addressing the
failings of the current health care system.
Acknowledgements
I am grateful to The Robert Wood Johnson Foundation for support, to Barrett Kirwan for research
assistance, and to Larry Levitt, Jack Meyer, E ll i o t
Wicks, seminar participants at the National Bureau
of Economic Research, and members of the advisory panel for their helpful comments. n
73
Hacker Proposal
Key Elements
Jacob S. Hacker has proposed a plan to achieve universal coverage by building
on the Medicare program. The proposal includes the following elements:
       either to offer and automatically enroll employees in a plan at least as generous as that available under an enhanced
Medicare benefits package or to pay a modest payroll-based contribution
to help fund enrollment of their employees in Medicare Plus.
    the contribution instead of providing
their own plan would be enrolled automatically in Medicare Plus at their
workplace, although they could use their employers’ contributions (minus
a penalty) to purchase other coverage that met the same standards as the
workplace plans.
       to use outreach
efforts to enroll non-workers in Medicare Plus, which would effectively
replace Medicaid and the State Children’s Health Insurance Program
(S-CHIP).
                who are not enrolled by their
state would have an individual buy-in option available, with the premium
based on income.
74
About the Author
 . , .., is a Junior Fellow of the
Harvard University Society of Fellows. Before
coming to Harvard, he was a Fellow at the New
America Foundation and, prior to that, a Guest
Scholar and Research Fellow at the Brookings
Institution. A political scientist who studies social
policy, he is the author of The Road to Nowhere:
The Genesis of President Clinton’s Plan for Health
Security (Princeton University Press), which
co-won the  Louis Brownlow Book Award of
the National Academy of Public Administration.
His articles and opinion pieces have appeared in
the British Journal of Political Science; Studies in
American Political Development; the Journal of
Health Politics, Policy and Law; the New Republic;
the Nation; the Los Angeles Times; the Boston Globe;
and the Washington Post. His second book, The
Divided Welfare State: Government, Business, and
Social Policy in the United States, will be published
in .
75
Medicare Plus: Increasing Health Coverage
by Expanding Medicare
by Jacob S. Hacker
Overview
Motive and Rationale
Universal health insurance has been the great unfulfilled hope of American health care reformers, a
siren call luring countless victims to the shoals of
political defeat. One reason for this has been the
inherent difficulty of outlining a compelling series
of self-reinforcing policy changes that would simultaneously move the nation toward universal coverage and build political support and administrative
capacity for further steps in that direction. Instead,
the political constraints that American reformers
inevitably confront have repeatedly pushed advocates to embrace highly categorical and complex
programs that have proved du ra bly resistant to
expansion beyond their target populations.
To suggest a way out of this persistent trap, this
proposal outlines a sequential approach to universal
coverage—or near-universal coverage, depending on
how many of the steps are taken. It takes an established program, Medicare, and shows how it might
be expanded through a series of measures designed
to minimize short-term disruptions to existing coverage while creating strong incentives for the formation of an inclusive social insurance program.
Although constructed on familiar foundations,this
proposal combines el em ents usually viewed as distinct: a “single-payer” plan that pools risks broadly,
an “individual mandate” on Americans to obtain
coverage,and a modified “play-or-pay” requirement
waiving a modest levy on employers if they provide
coverage. Together, these elements would encourage,
rather than compel, working Americans to obtain
their insurance through a common framework—a
strategy that might foster a gradual movement away
from employment-based insurance.
Asked to design an ideal, or even broadly acceptable,
health financing structu re , few would pick the
patchwork of private coverage and public residual
programs that exists in the United States today.
Unlike citizens of other affluent democracies, most
Americans rely for their health security on voluntary employer plans, with public programs only partially filling the gaps left behind. Although the tax
code encourages firms to offer insurance,a substantial share of smaller and lower-wage firms do not.
Even when employers sponsor health benefits, an
increasing number of workers decline coverage
because the expense is too great. Equally important,
economists generally agree that Americans who are
covered pay for workplace insurance in the form of
lower wages, and this forgone income represents a
growing hardship. Over the past two decades, real
premiums have nearly tripled, even as most workers’
real wages have risen only modestly.
These voluntary arrangements leave more than
 million Americans without insurance, almost 
percent of them in families headed by workers.
These arrangements leave  to  million more
Americans insufficiently protected against medical
costs and a third of Americans without insurance at
some point during a two-year peri od .1 These
arrangements saddle Americans with persistent
uncertainty about how they will obtain protection if
they change jobs or if their employer reduces cover-
Stuart H. Altman, Uwe E. Reinhardt, and Alexandra E. Shields. The
Future U.S. Healthcare System: Who Will Care for the Poor and
Uninsured? Chicago: Health Administration Press, 1998; The Kaiser
Commission on Medicaid and the Uninsured, Uninsured in America:
A Chart Book, Menlo Park: Henry J. Kaiser Family Foundation, 2000.
1
76
FIGURE 1
Share of Employees* with Health Insurance from Their Own Employers, by Wage Quintile, 1979-1998
l
l
6
6
F
F
s
l
l
6
l
F
6
s
s
l
6
F
F
n
n
F
Fourth quintile
Third quintile
s
Second quintile
s
s
n
Highest quintile
6
n
Lowest quintile
n
n
Source: James L. Medoff, Michael Calabrese, and
Howard Shapiro. “Impact of Labor Market Trends on
Health Care Coverage and Inequality.” Washington:
Center for National Policy, October 4, 2000, table
H23.
*Private, nonagricultural wage and salary workers age 21– 64.
a ge . These arrangements channel the largest tax
breaks to affluent workers with generous insurance
while placing the greatest burdens on lower-income
workers and people in poor health. And during a
period of notable prosperity, these arrangements are
eroding. Ten million more Americans were without
insurance in  than in , with the most precipitous decline occurring among low-wage workers
(see figure ). Over the past two decades, Americans
have been required to pay more for their coverage
and to join plans that restrict their ch oi ce of
provider or limit their benefits. The insurance market continues to fragment while the financial “slack”
that underwrites America’s tattered medical safety
net grows tighter. If no action is taken, the situation
will only grow worse.
America’s public insurance programs, valuable
as they are for many, do not represent an effective
response to the weaknesses of workplace coverage.
Although Medicare reaches virtually all elderly and
disabled Americans through an inclusive national
program, public programs for people under the age
of  are scattered,incomplete,and often stigmatizing. Aimed at the indigent and the especially vulnerable, these state-based programs offer meager
assistance to working-poor parents and childless
adults, and virtually none to workers who have coverage but risk losing it or who can barely afford the
cost. Eligibility requirements and the availability of
care differ dramatically across states,and millions of
uninsured Americans who are el i gi ble for public
help do not obtain it, including hundreds of thousands of lower-income families who have lost insurance in the transition from welfare to work. Despite
the push to enroll children in Medicaid and programs set up under the State Children’s Health
Insurance Program (S-CHIP), public coverage of
children has actually dropped. Indeed, as late as
September , almost half of the federal money
allotted for S-CHIP had gone unspent.2
For all the shortcomings of the American
approach to the uninsured, however, the political
barriers to reform are a direct outgrowth of the
incomplete financing system that has arisen in the
United States. As the Clinton administration learned
to its misfortu n e ,a ny plan that can be portrayed as a
direct threat to the private protections that many
Americans enjoy faces a steep uphill journey. At the
same time, the existing melange of public programs
2
Robert Pear. “40 States Forfeit Health Care Funds for Poor Children.”
New York Times (September 24, 2000), A1.
77
also vastly complicates the challenge of reform.
Because programs for the non-elderly are meant to
fill gaps in private coverage, there is constant reluctance to streamline enrollment or extend public
insurance up the income ladder, lest public programs
displace private insurance. Moreover, because these
programs often provide generous coverage to recipients (at least on paper), reformers legitimately worry
that replacing them with a more extensive and uniform plan will deprive some current enrollees of the
benefits they now receive. Little wonder then that the
road to reform has proved so rocky and the proposals put forward so maddeningly complex.
And yet most Americans believe that the current
financing system fails to live up to the nation’s
ideals.3 Most also deplore the significant financial
burden and insecurity that this system imposes on
lower-income families, the ill, and those without
insurance. The proposal outlined in the following
pages presents a long-term policy approach to these
fundamental problems based on three principles:
. Affordable, guaranteed coverage offering a
defined package of benefits should be available to all
Americans, regardless of wh et h er their employers
sponsor it.
. Such coverage should be designed to uphold
social insurance precepts, spreading risks broadly
through an inclusive plan that is available to all and
within the financial reach of the less wealthy and the
less healthy.
. Basic insurance coverage should be expected
of all Americans as long as all have access to an
affordable plan.
These principles are goals, not methods—
guideposts, not pre s c ri ptions. They must be tempered by judgments about the political and administrative constraints that a successful proposal must
overcome. Five such judgments guide this proposal:
. Universal health insurance will most likely be
achieved through a series of large-scale but nonetheless partial steps that will need to be calibrated over
time to the responses of other actors and institutions
with influence over the breadth and depth of health
coverage.
. Any proposal that is perceived as taking away
or significantly raising the cost of existing private
coverage, imposing huge new costs on employers or
individuals, or significantly reducing the benefits of
public programs, is unlikely to be enacted.
. A successful plan must be seen as beneficial
and potentially available to all Americans, not as a
form of organized charity through which the many
aid the few.
. Targeting coverage narrowly on the uninsured
will likely be self-defeating, reinforcing the current
confused welter of programs and leaving unprotected many Americans who are uninsured, insufficiently insured, at risk of becoming uninsured, or
under serious financial pressure because of the cost
of coverage.
. Any plan should be simple to understand and
to enroll in, based on popularly understood and
time-tested institutions, relatively stra i gh tforward
to finance and administer, and subject to democratic control.
These principles and judgments underlie the
proposal that fo ll ows. It envisions the sequential
replacement of most state-federal public insurance
programs with a nationwide program modeled after
Medicare that all Americans without private coverage could enter by paying an income-related premium. The first part of the exposition lays out the
fundamental features of the proposal as they would
operate once fully implemented. The second part
usefully complicates this neat picture by considering
how the proposal might be phased in over time, and
how it might attract the necessary political support
to become robust legislation.
To be sure, Americans’ views are complex and multi-faceted.
Nonetheless, a wealth of opinion research indicates that while quite satisfied with the quality (if not the cost) of personal health care, the public
is quite dissatisfied with the overall structure of American health financing. See, in particular, Rosita M. Thomas. Health Care in America: An
Analysis of Public Opinion. CRS Report for Congress 92-769 GOV,
Washington: Congressional Research Service, 1992; Lawrence R. Jacobs
and Robert Y. Shapiro. Politicians Don’t Pander: Political Manipulation
and the Loss of Democratic Responsiveness. Chicago: University of
Chicago Press, 2000, 232–60; “A Survey of American Attitudes on
Health Care Reform,” conducted by the Program on Public Opinion and
Health Care and Marttila & Kiley, Inc., for The Robert Wood Johnson
Foundation; Altman, Reinhardt, and Shields, 1998, 27–28; Robert J.
Blendon et al. “Who Has the Best Health Care System? A Second Look.”
Health Affairs 14 (4): 220–30; Robert J. Blendon et al. “Satisfaction with
Health Systems in Ten Nations.” Health Affairs 9 (2): 185–92; and Karen
R. Donelan et al. “All Payer, Single Payer, Managed Care, No Payer:
Patients’ Perspectives in Three Nations” Health Affairs 15 (2): 254–65.
3
78
The Proposal in Brief
In broadest outline, the proposal has three central
components:
. All Americans not covered by Medicare or
employer-sponsored insu ra n ce would buy into an
expanded Medicare program, called “Medicare
Plus,” by paying an income-related premium. Most
would be en ro ll ed automatically, either at their
place of work or by state outreach efforts. Existing
public insurance programs for the non-elderly poor
and near-poor would be phased out.
. Employers could choose to sponsor coverage
at least as generous as that available under Medicare
Plus or pay a modest payroll-based contribution to
help fund public coverage. Workers whose employers paid the contribution would be enrolled in
Medicare Plus automatically, although they could
use their employers’ contributions (minus a penalty) to purchase private coverage that met the same
standards as workplace plans. With the exception of
extremely lavish plans, employer-sponsored insurance would retain its favored tax status.
. All Americans would eventually be asked to
show proof of coverage. This could be largely done
at existing decision points rather than through a
s ep a ra te process—when employers demonstrated
that they covered their workers, when citizens were
automatically enrolled in Medicare Plus, or when
those automatically enrolled in Medicare Plus asked
to opt out.
Before reviewing these three core el em ents in
more depth, it is worth noting some features that
the plan does not contain. The proposal does not
mandate that employers provide comprehensive
coverage. It imposes no massive payroll tax on
employers that do not sponsor insu ra n ce . It does
not require any employer to enroll its workers in a
public plan or compel any individual to take public
coverage. It does not place vast new requirements on
employer-sponsored plans and,indeed, would mean
higher costs for few employers that currently offer
covera ge . It does not eliminate the favorable tax
treatment of health benefits. It does not impose
major new regulations on private insurance. It has
no new insurance pools, no complicated new tax
subsidies, no complex new system of contracting or
risk adjustment. It does not restrict the growth of
private health spending. Indeed, it forswears essentially all new direct interven ti ons in the private
insurance sector.
This is not to deny that this proposal would
bring a sea change in U.S.health financing—and for
the better. All Americans currently eligible for public coverage would join an inclusive insurance pool
that would allow them to obtain services from nearly all providers in their region (or to enroll in a qualified private health plan that contracted with
Medicare Plus, much as private plans contract with
Medicare today). Most smaller and lower-wage
employers would likely decide that they would
rather pay a modest payroll assessment than sponsor coverage,allowing their workers to receive subsidized coverage through Medicare Plus. Employers
that still provided coverage would have to meet
minimum standards, and any American without
workplace coverage, even those turned down by private insurers, could buy into Medicare Plus. The
plan’s simple structure would make it easy to understand, efficient to administer, and a visible target for
en ro ll m ent efforts. As a very large payer using
Medicare’s basic instruments, Medicare Plus would
also have the capacity to ensure that expenditures
were contro ll ed and that enrollees always had the
option of maintaining free choice of provider. In
short, the nation’s fragmented strategy for plugging
holes in coverage would be replaced with a simple,
inclusive,and familiar public plan through which all
Americans without workplace insurance could buy
coverage at affordable rates.
Of course, this brief sketch leaves out many
thorny details that must be tackled by any reform
plan. Accordingly, the following sections take up a
series of critical topics: benefits, coverage, contributions and premiums, employer and state duties,
administration, financing, and horizontal equity.
After examining these crucial features of the proposal, I then suggest how they might be put into
place through a process of “large-scale incrementalism” that reinforces, rather than retards, political
support for further steps toward full implementation. Finally, I consider the reasons why—and the
79
conditions under which—this long-awaited journey
to universal insurance coverage could become a
possibility.
Benefits
The Medicare Plus benefits package would be simi lar to the current Medicare package. It would, however, be expected to include several benefits now
fully or partially excluded from Medicare coverage — most critically, outpatient prescription drugs,
preventive services, mental health services, and
maternal and child health. In addition, Medicare
Plus would not emulate Medicare in two respects. It
would su b s ti tute a single deductible and coinsurance rate for Medicare’s extremely high cost-sharing
requirements for inpatient hospital care, and it
would include a maximum cap on out - of - pocket
spending. Finally, wraparound programs (either
state or federal) would continue coverage for additional services now provided by state Medicaid programs to low-income families and children and to
the working disabled. The proposal would not disturb current Medicaid arrangements for the nonworking disabled and the elderly.
Taken as a whole,therefore, Medicare Plus’s benefits would be relatively modest compared with large
employers’ health plans, mainly because of h i gh er
cost sharing. Yet, unlike most private plans today,
Medicare Plus would provide free choice of provider
and place no limit on maximum benefits. And it
would still be considerably more expansive than the
current Medicare program. Although it would be
desirable if these “extra” benefits were also gradually
incorporated into Medicare, this proposal does not
call for any specific changes in the Medicare benefit
package. It seems likely, however, that the most
important of these extra benefits—prescription
drug coverage—will be incorporated into Medicare
in some form within the next decade. Over time,
moreover, political pressure from Medicare beneficiaries would likely push toward convergence of benefit levels across Medicare and Medicare Plus,creating
the opportunity for an eventual merger. The proposal would also create a new Medicare Benefits Advisory Committee (MedBAC), an expert advisory body
that would assist in designing the initial Medicare
Plus benefit package, and would review the Medicare
Plus and Medicare packages annually thereafter.
Harmonizing the two benefit packages would be one
of MedBAC’s goals.
Whatever the exact benefit package chosen, it
must balance competing objectives. Because
Medicare Plus would be expected to enroll a large—
and, most likely, growing—share of the population,
the benefits that it provides need to be generous
enough that enrolled Americans would enjoy adequate coverage for the services they need. If the benefits package were too minimal, moreover, all but
the poorest and sickest Americans would be reluctant to en ro ll in the program, turning it into a
poten ti a lly stigmatizing insurer of last resort.
Weighing in the other directi on , however, is the
need to minimize both the new revenues necessary
to establish the plan and the new burdens on
employers that sponsor insurance. Since firms that
wished to sponsor tax-favored insu ra n ce would
need to provide benefits at least as good as Medicare
Plus’s, the benefits package should not be so generous as to require a major upgrade of most employers’ health plans.
Coverage
All legal U.S. residents without qualified private
coverage would be automatically entitled to
Medicare Plus.4 Enrollment would occur through
three principal channels:
. Workplace enrollment. The bulk of Medicare
Plus beneficiaries would be enrolled automatically
at their place of work when employers elected to
make the payroll-based contribution instead of
sponsoring tax-favored covera ge . Similarly, selfemployed workers would be required to show proof
4
Arguably, undocumented workers should be treated in the same fashion as other workers, with their employers either providing coverage
directly or contributing to Medicare Plus. Yet, it is difficult to conceive of
a way in which to levy payroll contributions without relying on Social
Security numbers, and this would seem to preclude allowing undocumented workers into Medicare Plus. Nonetheless, undocumented residents would continue to be able to use emergency and other services
provided for under current law, and a portion of disproportionate share
hospital payments would be preserved to compensate medical institutions that serve large numbers of undocumented patients.
80
All workers and their families would be enrolled automatically
in either Medicare Plus or employer-sponsored plans, so the
individual mandate would have true significance only for those
without ties to the workforce.
of coverage or make the payroll-based contribution
themselves.
. State enrollment. State outreach and enrollment efforts to sign up non-workers (particularly
those el i gi ble for Medicaid and other state programs) would be the second major way in which
Americans joined Medicare Plus.
. Individual buy-in. For those Americans outside the workforce who were not signed up by the
states, an individual buy-in option would be available, with premiums scaled to income (but not to
health risk).
Because all Americans would be el i gi ble for
Medicare Plus, the availability (in contrast to the
cost) of coverage would not hinge on income or
assets. Premiums would be based on income reported for tax purposes and con du cted by Medicare
Plus’s enrollment division in coordination with the
Internal Revenue Service (IRS). Thus, states would
no longer have a reason to impose cumbersome and
stigmatizing means tests on potential enrollees.
Coverage of an individual under Medicare Plus
would be con ti nuous unless one of two events
occurred: the individual was hired by a firm that
sponsored qualified covera ge , or the individual
chose to opt out of Medicare Plus to purchase private coverage individually. In the latter case, the
individual would be asked to present an annual
insurance contract that met the same minimum
standards as those required of workplace coverage.
Opting out would be allowed only at the time of initial enrollment in Medicare Plus and during an
annual enrollment period, and exemptions would
have to be renewed. If the individual worked, the
mandatory employer contributions made on his or
her behalf, minus  percent to compensate for the
cost of healthy people disproportionately opting
o ut , would be forwarded directly to the qualified
plan. The individual would be expected to pay the
difference with after-tax do ll a rs . In the unlikely
event that the employer payments exceeded the private premium, private plans would be allowed to
rebate the difference.
All Americans would eventually be required to
show proof of public or private covera ge , which
could be done by attaching to a federal tax return a
standard form supplied by Medicare Plus and qualified private plans. All workers and their families
would be enrolled automatically in either Medicare
Plus or employer-sponsored plans, so the individual
mandate would have true significance only for those
without ties to the workforce. To reach the many in
this population who do not file tax returns, states
would be given powerful incentives to enroll nonworkers in Medicare Plus. States would also be
encouraged to subsidize Medicare Plus coverage for
the tempora ri ly unemployed, and to establish
mechanisms for en ro lling the uninsured in
Medicare Plus when they sought care.
Contributions and Premiums
Most enrollees in Medicare Plus would be workers
whose employers elected to make the payroll-based
contribution rather than provide qualified coverage.
This contribution would equal a percentage of
wages, tips, and salaries up to the Social Sec u ri ty
wage base (roughly , in ). The level of the
contribution would be dictated by three considerations.First, the payroll-based contribution should be
low enough that it does not impose an undue burden on low-wage firms, which are least likely to
sponsor coverage. Second, the level of the contribution should ensure that Medicare Plus has substantial enrollment, with the majority of en ro ll ees not
previously enrolled in public programs. Medicare
Plus would not be a public assistance program, but
rather the primary source of coverage for working
81
Americans who now struggle to obtain or afford
insurance. Finally, the employer contribution rate
should be not be so high that it would impose a large
new cost (in the form of lower cash wages) on poorer citizens who previously enjoyed public protections or on wealthier citizens who previously
enjoyed heavily tax-subsidized coverage. These considera ti ons all point to a contribution rate that is
substantially below the average amount that employers now spend for health benefits.
The payroll-based contribution suggested here —
for esthetic as well policy reasons—is  percent. This
would be the maximum share of taxable payroll that
any firm would be required to pay for health insurance. As described more fully in the section on
employers, firms insuring their workers for the first
time,and with very low average wages, would be eligible for steep reductions in their initial contribution rate.
For those who wish to recall the bitter debates of
the early s,  percent will seem an extremely
modest levy, well below the  percent to  percent
that was common in proposals then. It should be
noted that these earlier rates were usually divided
between employer and worker, meaning that the
portion paid directly by employers was lower.
Moreover, the share of wages and salaries that
employers spend on health benefits has actually fallen by about a percentage point since the early s.
Still, a  percent contribution rate is significantly
lower than the average share of payroll that employers now pay for health benefits.5 This is intentional:
A play-or-pay requirement with a modest contribution rate is an en ti rely different policy approach
from the same requirement with a higher rate.A low
contribution rate would protect against many of the
risks and problems correctly identified with the
play-or-pay design (though, of course, it does raise
other concerns), and it would create powerful selfreinforcing effects that would serve to bolster the
5
According to the Bureau of Labor Statistics, the average private
employer’s health spending represents between 7 percent and 8
percent of wages and salaries. Among firms with more than 500
employees, virtually all of which sponsor coverage, the cost is closer
to 9 percent of payroll. U.S. Census Bureau. Statistical Abstract of the
United States. Washington: U.S. GPO, 1999, 331.
position of Medicare Plus. Nonetheless, the contribution rate should be thought of as a variable rather
than an exact value. What is crucial is that the rate
chosen is consistent with the goals of the proposal
and with the considerations just discussed.
In addition to the  percent employer contribution,many workers enrolled in Medicare Plus would
also be assessed a premium that would vary with
income and family size. With regard to all but the
poorest Americans,the level of this premium would
be set so as to keep the nominal division of employer and worker responsibilities relatively similar to
what it is today within firms that sponsor coverage.
Economists are surely correct that, in general and
over the long term, workers will end up paying
much of the “employer share” through lower cash
wages. Yet this impeccable economic logic runs up
against the reality that most workers today do not
recognize the extent to which they pay for health
insurance through forgone wages, and most also
seem to treat the distinction between employer and
employee contributions as meaningful. If preserving the largely fictitious notion of employer contributions is the price that must be paid for
significantly expanded covera ge , it seems a small
price indeed.
Medicare Plus would offer four types of coverage:
single (individual), couple (individual and spouse),
single-parent family (individual and children), and
family (individual, spouse, and children). The
amount of the premium for each type of coverage
would vary with income and mechanism of enrollment and would be deducted directly from a worker’s
paycheck. Although there are several possible premium structu re s ,t a ble  describes a simple framework
that meshes closely with the proposal’s goals.
These premiums are higher than the average
monthly amount that insured workers pay today for
single coverage (, down from  in ). But
they are close to the average employee payment for
family coverage and to the . monthly premium
that Medicare enrollees pay for Part B coverage. And
many, if not most, workers enrolled in Medicare
Plus would be eligible for subsidies and thus not pay
the full premium.
For individuals en ro ll ed in Medicare Plus
82
TABLE 1
A Possible Framework for Monthly Premiums, by Income and Coverage Type*
Income Relative
to FPL**
Single
Couple
Single-Parent
Family
Family
Less than %
No premium; cost sharing limited
Less than %
No premium; cost-sharing subsidies gradually phased out
Less than %
No premium
Less than %
–
–
–
–
% or more




*As with employee payments to private health plans, Medicare Plus premiums would generally not receive favorable tax treatment.The main
exceptions would be cases in which premiums plus other health expenses exceeded 7.5 percent of adjusted gross income (under the incometax deduction for extraordinary medical expenses) or in which premiums were paid with funds from a qualified "flexible spending account"
set up by an employer.
** In 2000, the federal poverty level (FPL) was roughly $8,500 for an individual and $17,000 for a family of four.
through the workplace, determining premium levels
would be relatively simple, because payroll-based
contributions would be based on wages and thus
could be used as a proxy for income. This would
also be true, of course, of single-parent families
en ro ll ed through the workplace. The situation
would be somewhat more complicated for cou pl e
and family coverage. If both members of the couple
were enrolled in Medicare Plus through the workplace, determining eligibility would be straightforward. A married individual en ro ll ed in Medicare
Plus would simply indicate that his or her spouse
was also covered, and the program would bill them
as a unit for the couple or family premium, basing
subsidies on combined income.6 If, by contrast, one
member of the couple worked for an employer with
private insurance while the other was en ro ll ed in
Medicare Plus, estimates of joint income would be
based on a combination of the employer’s wage and
tax filings and on self-reported income, with a reconciliation process at the time of annual tax filing.
Families in this situation could elect to receive fami-
6
It should be noted that the net cost would be the same if, instead of
obtaining couple coverage, the couple paid for two single policies. And
it would also be the same if, instead of purchasing family coverage as a
unit, one member of the couple signed up for single-parent family coverage, and the other paid for single coverage.
ly coverage through the private employer, in which
case the payroll-based contributions made would be
rebated directly to the employer. Or they could elect
to buy into Medicare Plus, in which case the firm
with private coverage would contribute  percent of
payroll to defray the cost of coverage under
Medicare Plus.
In all cases,there would be an annual reconciliation process based on tax information to ensure that
Medicare Plus en ro ll ees had paid the appropriate
premium in the previous year. Workers who overpaid would be refunded the difference, while workers who had underpaid would be charged the additional amount, which could be taken directly out of
a tax refund, if applicable. In the case of major
underpayments, a penalty would apply. Premiums
would be based solely on income, so there would be
no requirement to demonstrate limited assets to justify subsidized premiums or cost sharing.
For those without ties to the workforce, both the
method for determining income and the level of
premiums would necessarily differ. All non-workers, including the unemployed and recipients of
public assistance, would receive subsidized coverage
if their income fell below  percent of the federal
poverty level (FPL). As with working Americans,
limits on out-of-pocket spending would apply to
83
TABLE 2
Employer Contributions, Monthly Premiums, and Rebates in Four Examples
Enrollee(s)
Income
(relative
to FPL)
Payroll-Based
Contribution*
Monthly
Premium
Rebate for
Private
Coverage
Cost
Sharing
Limited?
Single worker
, (%)
,
 (/yr)

No
Family of four
, (%)
,


Yes
Family of four
, (%)
,
 (,)
,
No
Non-worker
, (%)

** (,)

No
* If the employing firm paid the maximum 5 percent contribution.
non-workers with incomes up to  percent of the
FPL, phasing out between  percent and  percent. For non-workers with incomes above  percent of the FPL, premiums would rise on a sliding
scale from zero for non-workers at  percent of
the FPL to the average actuarial cost of coverage at
 percent. As described later, states would be given
incentives to establish assistance programs for the
tempora ri ly unemployed. Non-workers could
amend their self-reported income at any point during the year if their circumstances changed, and the
same end-of-year reconciliation process would
apply to non-workers as to workers.
The four simple examples in table  clarify these
guidelines.
The Role of Employers
Although employers would be asked to take on new
responsibilities, their obligations under this proposal would be inherently limited.7 Indeed,this proposal should actually reduce costs for many large and
high-wage employers by obligating all firms to pay a
share of the expense of covering working spouses
and by reducing cost shifting from uninsured to
insured patients. As for smaller and lower-wage
I assume that public employers would be treated in the same way as
private ones, but they could simply be required to sponsor coverage
meeting minimum standards, since essentially all do now.
7
** Assuming the average actuarial cost of single coverage was $2,200.
employers, Medicare Plus would offer an inexpensive and simple option for insuring their workers,
guaranteeing that no firm would have to pay more
than  percent of covered payroll for insurance.
Many firms,however, would pay much less. This
is because lower-wage firms and firms that had not
previously offered insurance would be eligible for
significant reductions in their contribution rate. For
lower-wage firms, the payroll-based contribution
would be just . percent if average annual wages
were below ,,  percent if average wages were
between , and ,, . percent if average
wages were between , and ,, and the
full  percent if average wages were , or more.
Firms that had not offered insurance before would
also be eligible for additional rate reductions during
the transition period. These transitional reductions
would equal . percentage points and would be on
top of the discounts for low-wage firms. Thus a lowwage firm that had not offered insurance in the past
could pay as little as  percent of payroll for
Medicare Plus coverage. To be eligible for this transitional reduction, firms would have to be in existence at the time the legislation was passed and not
have sponsored insurance in any of the prior five
years. This reduction would phase out over  years,
falling from . percent in year one to . percent in
year two and so on.
Besides the level of the payroll-based contributions, the potential burdens on business depend
84
upon three key aspects of the proposal: () the
nature of the minimum requirements on coverage;
() the complexity of administration and compliance; and () the rules for coverage of p a rt - ti m e
workers, employees’ dependents, and departing
workers. In each area, this proposal offers the maximum possible flexibility consistent with the goal of
en su ring that all workers and their dependents
receive health insurance.
Minimum Requirements on Coverage
The relatively modest Medicare Plus benefits package would become the benchmark against which
employer-provided health plans were judged. Firms
that did not offer coverage that was at least equivalent to Medicare Plus’s protections would be
required to make payroll-based contributions to
Medicare Plus. Such coverage would not receive special tax treatment.
“At least equivalent” does not mean “identical.”
Indeed, most employer-sponsored health plans are
already more generous than Medicare Plus would be
and are likely to remain so. Rather, employer-sponsored plans would have to meet criteria similar to
those that currently apply to private health plans
contracting with Medicare: Private plans would be
required to include all Medicare Plus covered services, but could include additional benefits. Compliance would be assessed by the Department of Labor
and the IRS, using congressionally approved guidelines. These enforcement procedures would build to
a certain extent on current law. Federal regulations
already require that plans furnish a description of
benefits to the Secretary of Labor, and that plans
with more than  participants file a detailed annual return with the IRS (Form ).
Administrative and Compliance Requirements
To the fullest extent possible, this proposal would rely
on existing administrative and enforcement mechanisms rather than new ones. In particular, the choice
of payroll-based contributions as a major financing
mechanism builds on the well-developed system for
collecting Federal Insurance Contributions Act
(FICA) taxes for old-age and survivors’ insurance,
disability insurance, and Medicare. On top of the .
percent FICA tax that employers already pay, firms
that did not sponsor private coverage would contribute an additional  percent to  percent of covered wages. Somewhat more complicated, they
would also have to deduct the Medicare Plus premium from paychecks and report these withheld contributions on tax statements. Because premiums
would be based on income, formulas for estimating
premiums based on current wages and self-reported
income would be straightforward to develop. For
instance, a new line could be added to worksheets for
calculating workers’ automatic tax withholding (W-
forms) to determine if workers had additional
income that might change their premium.
Existing mechanisms would also be expanded to
assess whether employers were exempt from the payroll-based contributions. The IRS is already charged
with determining whether employer-provided
health benefits qualify for favorable tax treatment.
Under this proposal, the IRS’s role would expand in
three main directions. First, the IRS would assess
whether employers actually sponsored insurance for
their workers and dependents. Second, the IRS (in
concert with the Department of Labor) would judge
whether this coverage met the minimum standards
necessary to allow the employer to be exempt from
payroll-based contributions for the fo ll owing tax
year. Third, the IRS would become the con du i t
between enrolled workers and Medicare Plus, collecting payroll-based contributions and premiums
while recording plan exit and entry. Accordingly,
firms would be required to notify the IRS when
workers entered or left their employ. All these duties
would be compatible with existing IRS procedures
and could be carried out using the normal tax calendar and reporting mechanisms.
The favorable tax treatment of health insurance
would be largely preserved under this proposal, with
two important exceptions.First, supplemental plans
provided by employers whose workers enrolled in
Medicare Plus would not be tax-exempt. This not
only would protect against large revenue losses, but
also would discourage employers from economizing
by dropping insurance and providing wraparound
benefits to bring coverage back up to previous levels. Second, the tax exclusion for health benefits
85
would be capped at a level twice that of the imputed
Medicare Plus premium. (A more complicated
alternative, which would account for regional premium variations, would be to cap the level at twice
the average amount paid by Medicare Plus to contracting private plans in a given geographic area.
Given how high the cap is, this seems unnecessarily
complex.) All other benefits currently receiving
favorable tax treatment—or slated to receive it, as
with the deductibility of premiums for the selfemployed—would continue to do so.
Coverage Rules
Employers that did not make payroll-based contributions to Medicare Plus would be required to offer
comparable insurance coverage to all of their
employees as well as their employees’ spouses and
non-working children under the age of . They
would also be required to automatically enroll all of
their workers in one or more qualified health plans,
a ll owing individuals to opt out only if they had
alternative coverage. In practice, this would mean
that employers could only exclude from coverage a
worker who had a family policy through Medicare
Plus (in which case the firm would contribute  percent of the worker’s payroll to Medicare Plus) or
through another employer (in which case the worker would show proof of private coverage under
another employer’s plan, and the two firms would
be free to arrange transfer payments, if they
desired). Again, if an employer covered a spouse
working for a firm that did not sponsor private
insurance,the non-sponsoring firm’s  percent contribution would be transferred to the other employer. Similar rules would apply to workers with more
8
These rules are less complex than they may appear at first. Recall that
a worker whose employer did not offer coverage would be automatically enrolled in Medicare Plus. If that worker had coverage from
another employer, he or she would simply notify Medicare Plus and
show proof of coverage, and Medicare Plus would transfer the payroll
contribution made by the first employer to the second. The same is true
in the case of spousal coverage. A worker enrolled in Medicare Plus
who had alternative coverage through a spouse would ask to decline
Medicare Plus coverage and show proof of private coverage, and
Medicare Plus would transfer the worker ’s contribution to the sponsor
of the spouse’s plan. To be sure, this approach carries the risk that some
workers will not report alternative coverage and simply remain in
Medicare Plus (or take coverage from both sources). At least among
those who are paying a premium to Medicare Plus, however, this would
than one job. For example, if one employer of a
worker with two jobs provided private coverage and
the other did not, the latter employer would contribute  percent of covered payroll to pay for the
private policy.8
The minimum share of private premiums that
employers would be required to contribute would
vary with the type of coverage and the number of
hours an employee worked.9 For employees who
worked more than half-time, the minimum share
would be three-quarters of the premium for single
coverage and two-thirds for family coverage.10 (If an
employer offered multiple qualified plans, the contribution floor would apply to the lowest-cost plan.)
For employees who worked between one and 
hours a week, the minimum share would be lower,
though exactly how much lower is hard to say,
because the requirement would need to be sensitive
to estimated effects on employer and worker spending. One possible approach would be to have the
share fall by a percentage point for each hour fewer
than  worked per week. An employee who worked
 hours a week, for example, could be asked to pay
as much as  percent of the premium for single coverage and  percent for family coverage. This would
lower the expense of part-time workers without creating incentives for firms to increase reliance on such
workers to evade coverage requirements. A weakness
of many past plans that relied on an employer mandate or play-or-pay requirement is that they would
have drastically reduced the cost of insuring parttime workers, en co u ra ging firms to hire part-time
workers or limit existing workers’ hours.
This proposal would maintain the Consolidated
Omnibus Budget Reconciliation Act (COBRA) con-
seem a relatively unattractive option.
Independent contractors would be treated as self-employed workers,
while temporary and contract workers would have to be insured by the
firm judged to be their primary employer—either the firm to which they
provided services or the firm that arranged their employment.
9
10
Across all employers sponsoring plans in 2000, the average share of
the premium paid was 86 percent for single coverage and 73 percent for
family coverage, although the average shares have been as low as 79
percent and 68 percent, respectively, over the past decade. Jon Gabel et
al. “Job-Based Health Insurance in 2000: Premiums Rise Sharply While
Coverage Grows.” Health Affairs 19 (5): 147–48. Under the Federal
Employees Health Benefits Program, government premium contributions
average just over 70 percent.
86
TABLE 3
Costs and Benefits of Employers’ New Role
New Costs
Firms that Do Not Sponsor Insurance
Firms that Sponsor Insurance
At least  percent of covered payroll for
health benefits
New requirements governing level and
breadth of coverage
Reporting and compliance expenses
Cap on tax exclusion
Reporting and compliance expenses
New Benefits
Access to low-cost coverage for workers
through Medicare Plus
New payments to cover working
spouses
Ten-year transitional reductions in
Medicare Plus contribution rate
Opportunity to limit cost of health
benefits by paying into Medicare Plus
Reduced cost of COBRA requirement
makes coverage more affordable
Reduced cost of COBRA requirement
makes coverage more affordable
Reduction in unpaid medical bills
makes coverage more affordable
Reduction in unpaid medical bills
makes coverage more affordable
ti nu a ti on requirement, but would slash its costs.
Although displaced workers could obtain COBRA
covera ge , most undoubtedly would choose subsidized insurance under Medicare Plus, especially if
state unemployment insurance programs helped
with the premiums. For this reason, employers
would have markedly greater reason to inform
departing workers of their statutory rights than they
do now, because many employees who might have
taken up COBRA coverage in the past would choose
Medicare Plus instead. As under COBRA, employers
would be required to provide standard information
describing departing workers’ options.
The new costs and benefits for employers are
summed up in table . On the cost side, employers
that do not sponsor health insurance now will have
to pay at least  percent of covered payroll for health
benefits. Employers that do sponsor health insurance will have to bear new costs if their coverage
fails to meet the minimum standards and they
choose to upgrade it, or in the rarer event that they
provide health coverage so expensive that it runs
afoul of the new cap on the tax exclusion. All firms
will have to comply with the proposal’s modest new
administrative and compliance requirements. On
the benefit side, firms that do not now provide coverage will be able to purchase low-cost coverage for
their workers through Medicare Plus and will be eligible for transitional reductions in the Medicare
Plus payroll contribution. Many firms that provide
coverage for working dependents of their employees
will receive a new transfer payment to of fs et the
cost. Some firms that now provide coverage may
also benefit from the option of enrolling their workers in Medicare Plus, which would effectively cap
their direct obligations. The cost of COBRA continuation coverage will also be dramatically reduced,
and all firms will benefit from the reduction in
unpaid medical bills incurred by the uninsured.
The Role of the States
This proposal would end Medicaid and S-CHIP as
we know them. Those now eligible for Medicaid and
S-CHIP with ties to the workforce would be covered
by Medicare Plus or employer-sponsored plans. For
87
eligible non-workers, the role of the states would
largely be transformed from a provider of insurance
into a portal for coverage under Medicare Plus.
Nonetheless, states would still retain a substantial
role. This role would, in important respects, be
financial.States would be required to make “maintenance-of-effort” payments to Medicare Plus equal
to their existing and projected spending on
Medicaid and S-CHIP benefits for children, n on elderly adults, and the working disabled. Yet states
would also serve vital administrative and enrollment functions. Not only would they be expected to
continue to finance long-term care for eligible populations, they also would be responsible for ()
enrollment of non-workers, () provision of wraparound coverage, and () subsidization of Medicare
Plus premiums for the unemployed.
Enrollment of Non-Workers
Although many Americans eligible for Medicaid and
S-CHIP would be automatically en ro ll ed in
Medicare Plus through their place of work, states
would be given strong incentives to enroll the
remaining uninsured in the program. These incentives would come in the form of reductions in the
state’s maintenance-of-effort requirement. For each
uninsured person previously eligible for Medicaid or
S-CHIP that a state enrolled in Medicare Plu s ,t h e
state’s maintenance-of-effort payments would be
reduced by an amount proportional to previous
state per capita spending. This credit, in effect,
would give states that had expanded coverage in earlier years comparatively greater scope to reduce their
financial responsibilities. In addition, of course,
states would directly reduce their spending on Medicaid or S-CHIP by shifting beneficiaries into
Medicare Plus. For non-workers not previously eligible for Medicaid or S-CHIP, a smaller credit would
be awarded. A portion of prior Medicaid spending
would also be earmarked for ongoing outreach
efforts, with costs shared between the federal government and the states,as under current law.
Provision of Wraparound Coverage
An important principle of this proposal is that cur rent recipients of public coverage should not receive
less generous benefits than they do now—although,
of course, some would be paying a portion of the
cost of coverage that they currently receive free or
virtually free. It is important, therefore, that wraparound coverage be available to provide those eligible for Medicaid with benefits that are not included
in Medicare Plus or are not commonly covered by
employer-sponsored health plans. This wraparound
coverage would be provided to former recipients
who continued to meet their state’s income eligibility criteria. It would also be provided to all those
automatically eligible for Medicaid coverage under
federal law.11 Employers of workers likely to fall into
either category would be informed that no-cost supplemental policies were available.
In providing wraparound coverage, the states
would have two options. They could con ti nue to
operate state programs to provide these benefits,
with the state share of spending credited against
maintenance-of-effort payments. Alternatively, they
could agree to have these populations enrolled in a
set of standardized federal supplements to Medicare
Plus. States would be encouraged to choose the second option by generous terms: For each en ro ll ee ,
the state’s contribution would be equal to the
(regionally adjusted) average cost of the supplemental package multiplied by  percent of the state’s
previous matching share of Medicaid spending.
Thus,if a state previously paid  percent of expenditures, it would pay  percent of the cost of the
supplemental package.12 Both the state’s contribution and the  percent savings would be credited
against maintenance-of-effort payments.
Subsidization of Coverage for the Unemployed
The tempora ri ly unemployed represent a distinct
population that states could do much to help.
Although workers between jobs may be eligible for
COBRA coverage, they must finance the entire cost
at a time when they are often under financial strain.
Moreover, such workers may end up having non11
Over the long term, it would be desirable if the eligibility requirements
for wraparound coverage were standardized and federalized to ensure
equal treatment of citizens across states.
This formula is similar to the enhanced federal matching rate under SCHIP.
12
88
trivial incomes for the year and thus be reluctant to
participate in Medicare Plus out of fear that they
will be asked to repay subsidies or incur a penalty.
(This concern would be largely unfounded because
workers could amend their expected income during
the year.) Those who are between jobs and uninsured are, in short, precisely the sort of group for
which a social insurance system requiring small
sums to be put aside for future contingencies would
be well suited. Currently, however, no such system
exists.
As a response to this problem, states would be
encouraged to develop a framework of subsidies for
the tempora ri ly unemployed. This system could
operate in conjunction with unemployment insurance and be financed by a small surcharge on state
unemployment taxes. Alternatively, states could
operate federally backed low-interest loan programs
paying the Medicare Plus premiums of the unemployed, with full or partial forgiveness offered to
workers whose income after regaining employment
fell below a specified level. At a minimum, states
would be required to provide enrollment information to all recipients of jobless benefits. States that
operated such programs, which would be largely
self-financing, would receive a reduction in their
contribution requirement proportional to the number of unemployed residents assisted.
The replacement of state insurance programs
with alternative arrangements poses unavoidable
challenges. This is as true of many tax credit proposals as it is of the Medicare expansion outlined here.
Replacing Medicaid and S-CHIP with a refundable
tax credit, for example, could easily cause some former beneficiaries to lose benefits, face higher costs,
or fail to obtain coverage. By comparison, the transition process envisioned under this proposal contains only minor risks. All current Medicaid and
S-CHIP beneficiaries with ties to the workforce
would be automatically guaranteed coverage, and
most would be el i gi ble for generous subsidies.
Virtually all of the remaining beneficiaries would be
members of populations that states are required by
federal law to cover—most notably, individuals who
meet eligibility criteria for the former Aid to
Families with Dependent Children program, preg-
nant wom en , and very poor children. At the same
time, states would have overwhelming financial
incentives to move these recipients into Medicare
Plus. Finally, all former Medicaid beneficiaries
would be guaranteed supplemental benefits through
state or federal wraparound programs.
Administration
A crucial virtue of this proposal—or crucial vice, for
those who harbor deep-seated animus toward the
Health Care Financing Administration (HCFA)—
is that it builds on well-established and time-tested
institutions of administra ti on and finance. This
choice is not incidental: Medicare is a familiar and
overwhelmingly popular program, Americans’
famed distrust of government notwithstanding.13
Judged across many dimensions, moreover, it succeeds admirably in covering a high-needs population that private insurers shunned almost entirely
before its implementation. It has a well-developed,
if not well-funded, administrative infrastructure; an
established system of private administrative intermediaries; a sophisticated hospital and physician
payment schedule;an improving procedure for contracting with private health plans; and,above all, an
accepted and longstanding place in American medical finance. Any health policy analyst could no
doubt dream up health financing systems that work
far better than Medicare does. But existing and necessarily imperfect institutions — wh et h er public
programs or private benefits—should first be
judged,not against rosy visions of ideal reforms, but
against concrete alternatives. Expanding Medicare is
the best route to inclusive and nationally comparable coverage that reaches all Americans not insured
through employment.
That said, the new responsibilities inherent in
this proposal will require improvements in MediNinety-five percent of Americans believe it is “important” or “very
important” that Medicare be preserved, putting it alongside Social
Security as the cherished core of American social insurance. And among
Americans within 15 years of entering the program, Medicare is more
trusted as a provider of high-quality and easily accessible care than
either employer-sponsored insurance or privately purchased individual
coverage. Cathy Schoen et al. Counting on Medicare: Perspectives and
Concerns of Americans Age 50 to 70. New York: The Commonwealth
Fund, 1999.
13
89
Expanding Medicare is the best route to inclusive and
nationally comparable coverage that reaches all Americans
not insured through employment.
care’s occasionally creaky administrative machinery,
if only because this proposal envisions a major
expansion of the program. Because Medicare’s
administrative costs are already quite low (less than 
percent of total program expenditures), administrative spending could be boosted without bri n gi n g
expenditures anywhere close to the administrative
costs of private health plans.
Under this proposal,HCFA would have primary
responsibility for Medicare Plus, as it does now for
Medicare. The Social Security Administration and
Treasury and Labor departments would see the
expansion of their historical roles as managers of
the FICA tax system and regulators of the operation
and tax treatment of private health plans. As already
discussed, states would also have important administrative duties related to the en ro ll m ent of n on workers in Medicare Plus and, if they so chose, the
provision of wraparound coverage. With few exceptions, then, Medicare Plus would expand existing
administrative institutions rather than create new
ones.
Like Medicare, Medicare Plus would consist of a
default fee-for-service program coupled with a system of contracts with private health plans. Medicare
Plus would use existing Medicare rates, ad ju s ted
when necessary for the lower costs of the non-elderly. In service areas where Medicare coverage does
not now extend, new rate schedules would be developed based on existing methodologies. Medicare
Plus would contract with private health plans using
procedures similar to those now used or under
development by HCFA. Because enrollees in
Medicare Plus are more likely to be familiar with
health maintenance organizations and other private
plans, and less likely to have chronic conditions,
they will probably be more interested in private
plans than are elderly Medicare beneficiaries. On
the other hand, because Medicare Plus will cover a
more comprehensive range of benefits than
Medicare, Medicare Plus en ro ll ees may have less
incentive than Medicare beneficiaries to join private
plans offering benefits not covered under the feefor-service program.
Financing
Neither the payroll-based contributions nor the
premiums are designed to cover the full cost of providing Medicare Plus benefits. The chief reasons for
this are fivefold:
• The opposition of businesses that do not provide insurance would likely be overwhelming if the
payroll-based contributions were raised to a level
closer to full funding. Under this proposal, all
employers would make some contribution to the
cost of i n su ra n ce , but none is forced to assume a
large new burden.
• A higher tax, even when proportional to
income, also places a significant burden on lowwage workers, who are not only least capable of paying any new levy, but also most likely to have
earnings at or near the minimum wage. For these
workers, new labor costs cannot be fully of fs et by
wage cuts and may increase unemployment.
• The premiums and payroll-based contributions are set so that virtually all working and nonworking participants in Medicare Plus would
purchase coverage at highly subsidized rates, with
lower-income participants receiving a large subsidy.
This cannot be done without tapping into other
funding sources.
• Relying on multiple sources of financing guards
against the disruptions that might occur if any one
source failed to produce the expected revenue.
• Most important from a political standpoint,
multiple-source financing is essential if legislators
are to minimize the visible new costs that Americans
90
will face. Current financing arrangements fundamentally obscure the real costs to Americans of
health insurance. Any transition away from this system will require careful policy choices designed to
avoid imposing large, visible, and immediate losses,
and this entails sacrificing some degree of clarity in
the interest of political feasibility.
The magnitude of the funding shortfall depends
on three principal factors: () the generosity of
Medicare Plus coverage, () the number of workers
whose employers choose to make payroll-based
contributions rather than to sponsor qualified coverage, and () the income and health characteristics
of those who en ro ll in the program, which would
influence both revenues and costs. Since the benefit
package has already been discussed, the brief analysis that fo ll ows examines the second and third of
these three crucial factors.
How Many Americans Will Be Enrolled in
Medicare Plus?
This proposal requires firms to provide basic health
coverage, but allows them to limit their financial
obligations by enrolling workers in Medicare Plus. It
thus incorporates the core element of the (now
much-maligned) play-or-pay plans of the early
s, which were subject to a number of separate
estimations.14 Although this proposal differs markedly from these earlier plans, the various analyses conducted nonetheless provide some guidance as to its
likely effects.
The key va ri a ble in estimating the number of
working Americans who will be covered under a
play-or-pay proposal is the “pay” requirement—
that is, the share of payroll that employers are
required to contribute to the public plan if they
choose not to insure their workers. A lower contribution requirement will result in more workers
For example, The Pepper Commission. A Call for Action. Washington:
U.S. GPO, 1990; Jack A. Meyer and Sharon Silow-Carroll (eds.,). Building
Blocks for Change: How Health Care Reform Affects Our Future. Reston,
VA: ESRI, 1993; John Holahan, Marilyn Moon, W. Pete Welch, and
Stephen Zuckerman. Balancing Access, Costs, and Politics: The American
Context for Health System Reform. Washington: Urban Institute, 1991;
Sheila R. Zedlewski, Gregory P. Acs, and Colin W. Winterbottom. “PlayOr-Pay Employer Mandates: Potential Effects.” Health Affairs 11
(1)(2000): 61–83.
14
en ro lling in the public plan.15 Typically, analysts
have assumed that employers (and, by implication,
workers) do not care whether coverage is public or
private, and will choose the option that minimizes
costs. This assumption seems persuasive with regard
to firms that do not sponsor insurance but somewhat more questionable with regard to firms that
do. After all, these latter firms sponsor insurance in
the absence of any coverage requirement, presumably because of labor competition and tax policy. If
workers in this firm’s labor pool continued to desire
private coverage, and if tax policy remained essentially the same, these employers might continue to
sponsor private insurance despite being able to
obtain public protections for a lower cost. Because
of the significant uncertainty surrounding employer
responses, forecasts based on the assumption that
employers minimize costs should be considered
upper-bound estimates of public plan enrollment.
Previous estimates suggest that at a  percent to
 percent contribution rate, between  percent and
 percent of the non-elderly population would be
covered by public insurance, compared with
approximately  percent today. It is important to
recognize, however, that most of this increase is due
to the movement of Americans who are uninsured
or have non-group insurance into public coverage,
not the transfer of workers out of employmentbased health plans. Indeed,earlier estimates suggest
that even at contribution rates as low as  percent,
the share of Americans en ro ll ed in employmentbased plans would most likely increase as the uninsured and non-group insured moved into both
public coverage and workplace plans.
This would certainly not be the case if the payroll
contribution rate were  percent or lower. Still, the
shift from private group coverage to Medicare Plus
might not be as large as earlier estimates suggest. In
15
Many of these earlier estimates further assume that employers minimize their direct contributions under each option (rather than total premiums, including the share paid by workers). Yet it is not clear why the
respective shares of the premium paid by employers and workers should
make a difference for firm decisions (as long, of course, as both receive
favorable tax treatment). Rather, employers would be expected to compare the total cost required to provide coverage via private insurance to
the total cost of providing insurance under public auspices.
91
the first place ,s i n ce the early s, employers have
kept health costs in check during a period of impressive wage growth. As already noted, employer health
spending is lower as a share of wages now than it was
a decade ago. Moreover, the recent upswing in the
incidence of employment-based coverage masks a
long-term decline in rates of plan sponsorship and
coverage among low-wage employers .G iven that the
employers most likely to take advantage of the “pay”
option increasingly do not provide health benefits,
the expected shift from public to private coverage
may well be less significant today than it was a
decade ago. Finally,  percent represents only the
mandatory employer contribution, not the total cost
of Medicare Plus covera ge . Workers with incomes
above  percent of the federal poverty line would
have to pay additional premiums, so the effective
contribution could exceed  percent for workers
obtaining single coverage and  percent for workers
obtaining family coverage.16
Although reliable forecasts will require
microsimulation modeling, a very rough high-end
estimate based on earlier studies and current data is
that approximately  percent to  percent of the
n on - el derly population would be enrolled in
Medicare Plus when the program was fully implemented.17 Put more simply, the plan would be very
large—certainly larger than was contemplated (at
least openly) by any of the sponsors of play-or-pay
proposals in the past, when critics loudly charged
that a public plan with a third of the non-elderly
population was an abandonment of the American
way.18 These critics will resurface whatever the size
of the public plan. But this is an area where an intuitive and widely held notion—that displacement of
employment-based coverage should be avoided at
all costs—is fundamentally at odds with good public policy. A large public plan should be embraced,
not avoided. It is,in fact, key to fulfilling the goals of
this proposal.
The virtues of a large public plan are multiple
and compelling. First,a large plan enrolling a signif icant share of the population ensures that a diverse
cross-section of Americans will be within a common insurance pool, which is essential both for the
political strength of the program and to guard
against the prospect that the public plan will be saddled with the highest-risk groups. Second, a large
public plan facilitates cost control by simultaneously increasing the bargaining power of the public
plan and the share of health costs paid by it. Third,
and perhaps most often overlooked, a large public
plan ensures that subsidies for coverage are available, not just to the very poor and the previously
uninsured, but to near-poor and lower-middle-class
workers, who are burdened the most by high premiums and at the greatest risk of losing covera ge .
Because the premiums of employment-based plans
essentially constitute a regressive head tax, a low
contribution rate and large public plan inevitably
produce a more progressive distribution of the costs
of health coverage than a higher contribution rate
and smaller public plan.
Although anathema to the philosophy of some
on the left, the notion that the public sector should
cover lower- and middle-income Americans while
leaving the more affluent in private arrangements is
not really so exotic or threatening. Social Security’s
benefit formula, after all, provides income replacement rates that are far lower for high-income workers than for low-income workers, and few are surprised or dismayed that, as a result, most upper-
16
The highest effective rate would be paid by a worker at exactly 300
percent of the federal poverty level—roughly, $25,500 for an individual
and $51,000 for a family of four. With the full Medicare Plus premium
set at $600 per year for an individual and $1,680 for a family, the effective rates would go up by 2.3 percentage points (600 divided by 25,500)
and 3.3 percentage points (1,680 divided by 51,000), respectively.
who are currently covered by employment-based coverage (22 percent
to 39 percent of the non-elderly population). These expected enrollment
rates use the high-end estimates of Zedlewski, Acs, and Winterbottom,
2000, as the low end of the range of possible effects. Data on the current distribution of coverage among the non-elderly come from the
Current Population Survey tabulations of John Holahan and Johnny Kim.
“Why Does the Number of Uninsured Americans Continue to Grow?”
Health Affairs 19 (4): 189.
17
This would include all or virtually all former recipients of Medicaid and
other public insurance programs (approximately 9 percent to 10 percent
of the non-elderly population); nearly all of the previously uninsured (15
percent to 18 percent of the non-elderly population); 80 percent to 100
percent of Americans with non-group insurance (4 percent to 5 percent
of the non-elderly population); and a third to three-fifths of Americans
18
Confusingly, a chorus of criticism also raised the concern that the public plan would be too small—an unpopular, stigmatizing, residual plan
enrolling only those on the periphery of the economy or dependent
upon public assistance.
92
income workers obtain private supplements. In the
Netherlands, about a third of citizens are allowed to
remain outside the statutory health program, with
no dire effects. Similar rules apply in Germany. In
contrast, the traditional American approach has been
to target new coverage at the bottom third of the
income scale while trying not to disturb coverage
among the top two-thirds. But the better approach,
on both political and distributive grounds, is quite
the opposite: Provide good, affordable coverage to
the bottom two-thirds of Americans, and let the top
one-third essentially do what they please.
Of course, critics will decry this approach as an
unstoppable expansion of government’s purview.
But they will also argue for a minimal tax on struggling small and low-wage firms,and a lower tax will
mean a larger public program. Moreover, the initial
enrollment in the public plan will not be etched in
stone. If the private sector keeps insurance costs
down,then an increasing number of workers will be
shifted back into employment-based plans. Nothing
in this plan prevents employers from providing
qualified coverage on their own. In contending that
the scope of the public plan will inexorably grow,
opponents must essentially concede that employers
cannot be counted on to provide insurance or
restrain the growth of private premiums.
What Will Be the Characteristics of Medicare Plus
Enrollees?
In addition to the virtues just enumerated, a large
public plan also addresses the second major criticism of past play-or-pay proposals: that the public
plan will be su bj ect to a devastating influx of the
unhealthy. Such adverse selection is always a potential problem, but much less so when half or more of
Americans under age  are enrolled in the public
plan.
To begin with, those most likely to be moved into
public coverage are probably not much more costly
than average.19 Although some of the uninsured are
in poor health (in part because they lack insurance),
many are young and inexpensive to insure. The same
is true of lower-wage workers. Past estimates suggest
that the overall costs of uninsured Americans should
be about equal to the rest of the population once
they are covered. If Medicare Plus simply enrolled all
citizens up to a given wage level — say, three times
the poverty level—there would be little reason to
expect significant adverse selection.
But, of course, that is not what Medicare Plus
would do. Employers that can obtain lower rates in
the private market are free to opt out of the program,
and individuals and families can elect to get back 
percent of employer payroll contributions to purchase
coverage on their own. As a result, employers with
high-risk workforces would have a stronger financial
incentive to pay the fixed payroll contribution than
similarly situated employers with low-risk workforces, and high-risk individuals and families would
be less likely to opt out of Medicare Plus than lowrisk individuals and families. These responses would
raise the average risk of enrollees in Medicare Plus.
For several reasons, the probable effect appears
manageable. With regard to employers’ behavior,
the overriding determinant of the ch oi ce between
public and private coverage would be average wages
rather than the health characteristics of workers,
because the cost of coverage is so heavily subsidized
at lower wage levels. In addition, nearly all smaller
firms would probably enroll their workers in
Medicare Plus to avoid the large administrative and
loading costs of private insu ra n ce in the smallgroup market. Thus, most firms that opted out
would be larger groups, which would be expected to
have more heterogeneous risk profiles. Moreover, to
the extent that there was moderate adverse selection
against Medicare Plus because high-risk groups
opted in,this could be defended as the equivalent of
a federal high-risk pool, protecting the private
group market from the destabilizing effects of an
extremely skewed distribution of risks.
19
The only possible exception is the 12 million Americans with nongroup insurance, who do appear to be above-average risks. Mark Pauly
and Bradley Herring, Pooling Health Insurance Risks, Washington: AEI,
1999, ch. 6; Alice M. Rivlin, David M. Cutler, and Len M. Nichols. “Cost
Estimates: Authors Respond.” Health Affairs Supplement (Spring 1994):
55; P. Anthony Hammond. “Actuarial Memorandum: Premiums in
Regional Health Alliances under the Clinton Administration’s Proposed
Health Security Act.” Health Insurance Market Reform, Hearing before
the Committee on Finance, United States Senate, 103rd Congress, 2nd
Session, February 1, 1994, Washington: U.S. GPO, 1994, pp. 102–4.
93
If adverse selection were more severe, several
possible remedies could be adopted. An obvious
solution would be to impose community rating and
other reforms on the private insurance sector. But
this would inspire considerable political opposition
and interfere with the practices of self-insured
health plans. Instead, an attractive halfway measure
would be to make it more difficult for employers to
shift between public and private covera ge . For
example, exemptions from the payroll-based contributions could be renewed on a five- or -year basis
rather than annually, forcing employers to stay in or
out of Medicare Plus for a long, continuous period.
Employers could also be penalized for opting into
the plan in proportion to the number of years they
had stayed outside it, discouraging opportunistic
switching as workforces age.
Similar remedies could be used to reduce
adverse selection caused by the individual opt-out
provision, but fewer problems should arise here for
at least four reasons. F i rs t , those who opt out of
Medicare Plus receive back only the amount that
their employer contributed to the program,minus a
 percent penalty designed to reduce adverse selecti on . In other words, no income-related subsidies
are available for the purchase of non-group coverage. Second, private non-group coverage is, in most
cases, not eligible for special tax treatment.20 For all
but very-high-income buyers,this means that a fairly substantial portion of private premiums will have
to be paid with after-tax dollars. Third, opting out of
Medicare Plus will require an affirmative decision
and a potentially costly search. Evidence on private
benefit plans suggests that, even with minimal
switching costs, levels of participant inertia are very
high.21 Fourth, and finally, non-group premiums are
comparatively expensive and, absent insurance
reform, undoubtedly will remain so.
For political reasons, it seems prudent to allow
20
The exceptions are when it is purchased by the self-employed, or
when premiums plus health costs exceed 7.5 percent of income, under
the deduction for extraordinary medical expenses. If the opt-out provision were destabilizing Medicare Plus, however, the extraordinary
expense deduction could be amended to exclude private non-group
insurance premiums.
Brigitte C. Madrian and Dennis F. Shea. “The Power of Suggestion:
Inertia in 401(k) Participation and Savings Behavior.” National Bureau of
21
people to leave Medicare Plus if they have qualified
private coverage. Some affluent Americans will surely seize this expensive exit option,as do affluent citizens in other social insurance systems that allow it.
But in practice very few Americans will find the
rebate attractive.
Additional Financing
Despite the advantages of a lower contribution rate,
it does have the disadvantage that it brings in less
revenue per enrollee, even as it increases the size of
the public plan. The amount of revenue forgone at
lower contribution rates, however, is not proportional to the reduction in the rate. This is because a
significant proportion of enrollees in Medicare
Plus — public assistance recipients, non-workers,
very-low-wage workers—would pay little or nothing toward the cost of coverage at any tax rate.
Furthermore, as the tax rate goes up, fewer employers contribute to the plan, and they contribute less
on avera ge . Finally, as already discussed, adverse
selection is likely to be a greater problem at higher
contribution rates than at lower ones.
The contribution rate and premiums envisioned
in this proposal clearly imply a significant net government cost.22 How significant a cost is the big
question. It bears noting that several features of the
proposal help to keep net expenditures in check.
F i rs t , all workers enrolled in Medicare Plus have a
share of their wages automatically devoted to health
insurance. Although this proportion is relatively
modest and will cover only a small fraction of the
cost of coverage at the lowest wage levels, s om e
workers who previously received public coverage for
free will now make a contribution, while some with
h i gh er earnings will pay all or most of the cost of
their coverage. This differentiates this proposal from
plans relying on refundable tax credits or expansions
of low-income programs, because these plans typi-
Economic Research Working Paper No. W7682, Cambridge: NBER, May
2000.
22
This is to be distinguished from a net cost to the economy as a
whole. In the short term, new spending on the uninsured would raise
economy-wide costs, but if Medicare Plus effectively slowed the rate of
increase in health expenditures, these costs would be outweighed by
savings quickly.
94
TABLE 4
Other Potential Sources of Revenue, with CBO Estimates
Source
Revenue, -
Rationale
Expand Medicare to include
uncovered public employees
. billion
Most excluded workers qualify for
coverage on the basis of another job
or their spouse’s employment.
One to  percent “health tax”
on private insurance payments
or provider revenues from
private sources
No CBO estimates
available
Would compensate for adverse
selection against Medicare Plus and
ensure that high-income workers
made some contribution to the program.
Would also reduce externality costs of
rising private health costs
Increase the cigarette tax
up to  cents/pack
Up to  billion
Increasing excise taxes would reduce
the health costs of smoking.
Increase the federal alcohol tax
up to  per proof gallon
Up to  billion
Increasing excise taxes would reduce
the health costs of alcohol consumption.
Three percent excise tax on
non-retirement fringe benefits
– billion
Would reduce the subsidy for benefits
without requiring the valuation of each
employee’s benefits.
Sources: Congressional Budget Office. Budget Options. Washington: Author, March 2000; Sherry Glied. Chronic Condition: Why Health Reform
Fails. Cambridge: Harvard University Press, 1997, ch. 8.; Sheldon D. Pollack. “It’s Alive.” American Prospect 11 (17) (July 31, 2000).
cally raise little or nothing directly from recipients.
Second, this proposal limits new subsidies for
coverage to those insured by Medicare Plus.
Although the rationale for this feature is discussed in
the next section, the important point to note is that
it further reduces the budgetary costs of the plan.
Third, the proposal eliminates federal spending
for S-CHIP and for some portions of the Medicaid
program, while recapturing a large share of previous
state spending for these purposes (as well as unspent
S-CHIP funds). By vastly reducing the strain on
providers that serve low-income communities, it
would also allow for a significant reduction in federal and state spending on disproportionate share
hospital payments.
Fourth, the proposal will naturally create three
positive revenue effects: a reduction in tax subsidies
for workers who shift from tax-favored private cov-
erage into Medicare Plus, a reduction in tax subsidies now realized by people who purchase very
expensive health coverage, and a long-term increase
in payroll and income tax receipts caused by the
substitution of wages for health benefits among
firms that pay less for insurance than they would
have without reform.
Nonetheless, additional financing will be needed. A significant share could c ome from the -year
budget surplus, roughly  trillion of which is due to
the recent slowdown in the growth of Medicare and
Medicaid.23 Table  lists a handful of ad d i ti on a l
sources of revenue that could potentially be tapped.
(The list is illustrative, not exhaustive.) All receipts
Karen Davis, Cathy Schoen, and Stephen C. Schoenbaum. “A 2020
Vision for American Health Care.” Annals of Internal Medicine 160 (22)
(2000).
23
95
would be transferred directly into a dedicated
Medicare Plus trust fund that would have access to
permanent budget authority, a ll owing budget
resources to be spent without new legislation.24
Fiscal Sustainability
Programs need to do more than get up and running;
they must also be fiscally sustainable over time.
Given that medical cost escalation has fairly consistently outpaced wage growth, is there any reason to
expect that a public program covering a defined set
of benefits would be able to con ti nue into the
future, absent huge tax increases? And what assurances are there that the surplus that now makes general revenues look attractive won’t evaporate in the
future, forcing a Darwinian struggle among a rash
of underfunded programs? The second question has
an easy answer: There is no assurance that any program, even one financed solely by earmarked contributions and possessing entitlement status, will
not confront future fiscal pressures. The suggestion
that general revenues should finance a portion of
Medicare Plus reflects twin policy judgments: first,
that the surplus is due in substantial part to reductions in Medicare and Medicaid spending; and, second,that the program should not be financed solely
by payroll levies, which would impose excessive
costs on lower-wage workers and the firms that
employ them. It also reflects a political ju d gm en t
that Medicare Plus will be large and popular enough
to weather potential fiscal storms.
The broader concern about fiscal sustainability
is more serious. Here too, however, important features of Medicare Plus promise to keep its costs
manageable over time. In the first place, affected
businesses will put heavy political pressure on
Congress to minimize the payroll-based contributions. Although these contributions are not the only
financing source, keeping them at modest levels will
act as an effective restraint on program growth. It
24
Some may find it strange to create a trust fund when full funding for
Medicare Plus would not come from earmarked payroll taxes. But this is
quite consistent with the financing of many existing trust funds, including Medicare’s Part B trust fund, which is financed principally by general
revenues. See Eric M. Patashnik. Putting Trust in the U.S. Budget:
Federal Trust Funds and the Politics of Commitment. New York:
Cambridge University Press, 2000.
will also ensure that Medicare Plus enrolls a significant share of the workforce, maximizing the program’s ability to use its concentrated purchasing
power to keep medical expenditures in check.
Indeed, if sustainability is a concern, it applies
equally well to private-sector spending. Medical
costs that rise consistently faster than the economy
as a whole translate into a growing share of national
income devoted to health care, whether those costs
are financed by public or private sources. There is
no compelling reason to expect that Medicare Plus
would be more profligate than the private sector. To
the contrary, if Medicare’s experience since the early
s is any guide, Medicare Plus should be at least
as capable of controlling costs as private health
plans — indeed, based on the past few years of
essentially zero growth, significantly more capable.
But if employers did hold down costs more effectively, then private insurance would become
increasingly attractive compared with Medicare
Plus. And if, by contrast, private health premiums
were not kept in line,an increasing share of employers would enroll their workers in Medicare Plus. In
short, the structure of the proposal ensures that the
sector most capable of controlling costs gains a larger share of the population. It does so, moreover, in a
context in which minimum standards for coverage
and strong protections for lower-wage workers
would ensure that cost containment did not simply
equal cost shifting.
Horizontal Equity
The financial architecture of this proposal raises the
important issue of “horizontal equity,” the notion
that similarly situated Americans should be treated
similarly. Under this proposal, Americans with
incomes below  percent of the poverty level
would receive highly subsidized coverage if their
employer elected to make payroll-based contributions to Medicare Plus, but only existing federal tax
subsidies if they did not. Thus, a low-income worker whose employer sponsors coverage would receive
significantly less in federal subsidies than a similarly
situated worker enrolled in Medicare Plus.
This problem could be rectified by providing
96
In practice, nearly all smaller and lower-wage employers
would likely choose to contribute to Medicare Plus
rather than provide coverage on their own.
additional subsidies for health coverage to lowincome workers who receive coverage from their
employer—or, if desired, who purchase coverage on
their own. But providing income-related subsidies
to all workers would be complicated (would subsidy
amounts vary by regi on , for example, or with the
health characteristics of workers?),and it could raise
costs dramatically. It would also, of course, discourage employers with low-wage workers from
enrolling in Medicare Plus by reducing the expense
of providing employment-based coverage. This, in
turn, would reduce the ability of Medicare Plus to
pool risk and provide common protection to lowwage workers. In addition, there would be no guarantee that subsidies would cover a reasonable
portion of the costs of coverage, as there would be
under Medicare Plus. Thus, if the problem were
judged sufficiently pressing, a better solution might
be to further lower the contribution rate for
employers insuring their workers for the first time
and perhaps also for employers with very small
groups or very-low-wage workforces.
In practice, nearly all smaller and lower-wage
employers would likely choose to contribute to
Medicare Plus rather than provide coverage on
their own. For firms that had low average wages
or faced very high private premiums, the potential
savings offered by Medicare Plus’s fixed contribution rate would be overwhelming. Low-wage
workers who obtained private coverage from their
employers would therefore be principally concentrated in high-wage sectors in which the payrollbased contributions were larger and employers saw
advantages in providing insurance. Available evidence suggests that low-wage workers in such
industries are already treated comparatively well;
their rates of coverage are higher and their avera ge con tri buti ons lower than other low - w a ge
employees.25
It also should be noted that those covered by pri-
vate employment-based plans would con ti nue to
receive federal tax subsidies. These subsidies are
worth little to workers who face low marginal tax
rates, but they are quite valuable for the higherincome workers who are most likely to remain in
private workplace plans under this proposal. The
break-even point at which existing federal tax subsidies would be greater than federal subsidies for
Medicare Plus coverage is probably around ,
to , in annual income for a single worker and
, to , for a family of four. (By way of
comparison, the median household income in the
United States is about ,.) In effect, then, this
proposal would replace the regressive tax subsidy in
current law with a subsidy structure whose benefits
were distributed more equally across the income
ladder, even without creating a complex new system
of tax credits.
Implementation
Because this proposal builds on existing institutions, it could be implemented swiftly. Certainly
the task would be no more daunting than the original establishment of Medicare, which, in the
pre-digital era, went from passage to the payment
of benefits in a single year. It is more realistic to
expect, however, that implementation would be
extended over several years and broken down into
a series of discrete steps. In that context, the key
issue would be whether each step increases the
chance of future progress toward full implementation or, to the contrary, blunts political interest in
further movement or even allows opponents to
scuttle progress altogether.26
Jon Gabel et al. “Class and Benefits at the Workplace.” Health Affairs
18 (3) (1999): 144–50.
25
26
Alan Weil. “Increments Toward What?” Health Affairs 20 (1) (2001):
68–82.
97
The typical American approach is to extend coverage by population group: first, the elderly; then,
poor children; then, the near-elderly; then, the parents of poor children; and so on. The advantage of
this approach is that it fixes political attention and
public sympathy on specific vulnerable groups that,
as a rule,are not expected to be able to obtain coverage on their own. But the advantages of this
approach are counterbalanced by a principal disadvantage: The divide-and-cover strategy does not
lend itself to ready movement beyond the target
population. When public assistance for a residual
group is defended as a special exception, everyone
else is presumed to be well served by the status quo.
At the same time, addressing the plight of the most
sympathetic groups further reduces interest in a
general solution. Thus, while Medicare Plus could
well begin by covering a specific population, such as
children, there would be no assurance that this initial step would be fo ll owed by others, and, in fact,
some reason to think it would not.
A more promising route would be to phase in the
components of the proposal on an extended
timetable. This could be done by, for example, softening the initial requirements for employers, subsidizing firms being asked to provide coverage for the
first time, exempting certain employers at the outset,
or gradually moving non-workers and low-wage
workers from state programs into Medicare Plus.
Two features of the proposal, in particular, could be
delayed or softened substantially without sacrificing
its fundamental goals. The first is the limit on taxfree employer-sponsored health benefits, which
would affect few existing plans but could be made
even less threatening without crippling the proposal.
The second is the individual mandate. Because all
workers would be enrolled in private plans or
Medicare Plus, while many non-workers would be
former recipients of state programs who would be
automatically enrolled as well, the remaining pool of
the uninsured would consist principally of people
who are between jobs and higher-income citizens
who have chosen to forgo insurance. Covering Americans who continue to go without insurance even
after being informed of the highly subsidized protections available under Medicare Plus would naturally
be a lower priority than guaranteeing basic health
security for low-wage workers who now lack it.
The approach I have just outlined is best termed
“large-scale incrementalism,” and it is similar to the
strategy pursued by Medicare’s original architects,
who saw public coverage for the elderly as merely the
first step toward universal health insurance.27 Without adopting the highly categorical route that
Medicare’s architects took, it is possible to outline a
comparable step-by-step agenda for full realization
of this proposal that would play out over five or  or
even  years following initial legislative enactment.
Step : Laying the Foundation. In the initial year
after enactment, employers and individuals would
be allowed to buy into Medicare Plus, with employers that enrolled their workers in the program making payroll-based contributions. Transitional rate
reductions for employers newly insuring their
workers would be made available and widely publicized. Because their value would phase out over 
years, there would be strong incentives for employers to take advantage of these reductions immediately. States would begin the process of moving
Medicaid and S-CHIP beneficiaries into Medicare
Plus.
Step : Requiring Em pl oyer Sponsorship. In the
second phase of implementation, employers that
did not enroll their workers in Medicare Plus would
be required to offer, but not contribute to the cost
of, at least one private health plan for their workers
that would have to meet very minimal standards to
receive favorable tax treatment.
Step : Requiring Employer Contributions. In the
third phase, employers that did not contribute at
least half of the cost of private coverage for full-time
workers would be required to do so. Over a period
of years, the contribution requirement would be
raised to the full cost. Simultaneously, the minimum
standards for private coverage would be upgraded
gradually to ensure that private plans were at least as
generous as Medicare Plus.
Step : Closing Gaps. If a nontrivial proportion
For a fuller explanation of the strategy behind Medicare and its only
partial success, see Theodore R. Marmor. The Politics of Medicare, 2d ed.
Hawthorne, NY: Aldine de Gruyter, 2000.
27
98
of Americans remained uninsured, the individual
mandate would go into effect. The final two pieces
of the proposal, the cap on tax-free health benefits
and the tax on supplementary coverage, would also
be put in place; yet these elements could be postponed indefinitely if circumstances required it.
This step-by-step approach is not, of course,
without risks. Perhaps the most threatening is that
opponents of the proposal, beaten in the initial legislative round, will nonetheless rise anew to scuttle
Medicare Plus later. This possibility, made vivid by
the repeal of the Medicare Catastrophic Coverage
Act, would be far less likely if initial en ro ll m ent in
Medicare Plus were substantial and broad-based.
Not only would significant initial enrollment create
a large pool of beneficiaries ready to mobilize against
backtracking; it would also create a new business
constituency in favor of continued progress. Just as
large employers have typically been more favorable
toward employer mandates, employers who enrolled
their workers in Medicare Plus would likely become
more supportive of measures requiring their competitors to assume similar burdens.
In comparison, other risks seem more tractable.
It is true, for instance, that adverse selection is likely
to be significant with a voluntary buy-in option. Yet,
if implementation proceeded as envisioned, this
would be a temporary phenomenon. Initial unfavorable selection could be treated as a transition cost
and financed through short-term transfers from
general revenues. More troublesome perhaps is the
possibility that some employers will drop coverage
and encourage their workers to purchase Medicare
Plus on their own. Some monitoring would probably be necessary to prevent flagrant abuses of this
option,and some “crowd-out” is to be expected. But
unless employers were already planning to eliminate
coverage (in which case Medicare Plus would be
cushioning an inevitable blow), it seems unlikely
that they would wish to suddenly confront their
workers with the fairly large direct costs of buying
into Medicare Plus individually, especially because
these firms would be forfeiting the goodwill that
employer-sponsored coverage buys.
Political Robustness
This proposal runs against the grain of current
political debate in a number of key respects. It presents an integrated approach to universal coverage in
an era in which incremental steps have become the
modus operandi of U.S. health policy. It shifts the
primary locus of public insurance protections from
the states to the federal government after more than
two decades in which new coverage has been
achieved under state auspices. Perhaps most politically challenging, it suggests a dramatic expansion
of an established federal program at a time when
America’s political class firmly believes that the only
route to broadened coverage is via modest expansions of state-based programs or a bevy of new tax
subsidies for private health insurance. Surely this
proposal defies all reasonable standards of political
feasibility.
In the near term, this is undoubtedly true: The
prospect that the current President and Congress
will follow the map outlined here is nil, and,in fact,
there is little prospect that either will take substantial steps toward universal coverage by any means.
Some believe that a tax credit approach will attract
widespread support, but the magnitude of credits
required to reduce significantly the number of
uninsured is unlikely to materialize anytime soon.
Far more likely is a modest new package of credits
and deductions that throws more money into the
private insurance sector but achieves limited tangible results.
In the long term, however, the tides of American
politics are more difficult to foretell. Just a decade
ago, it is worth recalling, leading corporations, trade
associations, and professional organizations were
widely convinced that major reforms would be
enacted, and many entered the debate favoring an
approach bearing some similarity to the one
de s c ri bed here. Historically, health reform has
become a major issue about every  years, after a
period of dormancy during which politicians and
private leaders celebrated private market solutions.
If the past is prelude to the future, the next big tide
of health policy ferment is set to roll into
Washington sometime near the end of this decade.
99
At the moment,this tide may seem unthinkable.
Yet the same was thought in the early s, when,as
now, tax-based reform was the rage, and employers
seemed sold on the “new” inventions of utilization
review and managed care. It is instructive to recall
what happened next. Faced with a significant recession, many large firms became alarmed about the
cost of providing health insurance, and a significant
number stridently argued that smaller firms (whose
workers were often insured by large firms through
family policies) should be required to provide coverage.State governments, too, faced hard times,feeling the pinch of an expanded Medicaid program at a
time of fiscal distress. And as employers tried to
control costs, Americans grew alarmed about the
uncertainty and increasingly visible costs of health
benefits that many had once taken for granted.28
Today’s preferred policy options—inacti on ,
symbolic gestures, or, at most, modest tax credits or
small program expansions—are a reflection not just
of the long shadow of the Clinton health plan’s failure, but also of good economic times and the historically low rate of increase in private premiums that
has come with them. If the economy continues to
weaken, if the cost of private health insurance
returns to the high growth rates of the past (as it has
over the past few years), and if the states begin to
face new fiscal pressures (as they already have started to), the interest of employers and the states in
strict cost containment may revive once again. Yet it
will do so after a decade during which both players
have used nearly all the managerial and administrative tools at their disposal, in the process prompting
a public backlash against some of the most restrictive of their practices. The alternatives left will be
much less attractive: for the states,cutting Medicaid
benefits for the very poor to retain or expand coverage for the near-poor; for employers, switching to
so-called defined contribution arrangements in
which workers are given a fixed amount for medical
costs that is pegged to inflation or company revenues rather than health costs. Judged against these
options, Medicare Plus may become not simply a
viable alternative, but an attractive solution for
many key stakeholders—not least the American
public.
If it does, this proposal will have at least three
important political advantages. First, it builds on
positive aspects of broadly popular and widely
understood institutions—namely, Medicare and
employment-based health insurance. Group health
plans work well for better-off workers in large firms.
This proposal allows them to con ti nue to do so,
while at the same time using the workplace as a conduit for public coverage for the rest of the working
population. Medicare Plus also immediately gains
the legitimacy and familiarity of a well-liked program. No great leaps of faith are required to anticipate how it will opera te . Nor are tales of sinister
bureaucracies quite so fearsome when the bureaucracy in question already takes care of grandparents,
parents, neighbors, and friends.29
Second, this proposal is designed to impose
minimal new costs on employers that do not now
provide insurance and to reduce costs for most
employers that do sponsor covera ge . At the same
time, it provides new subsidies for lower-wage
workers without imposing large, visible new costs
on other Americans.
Third, this proposal promises to retain a substantial role for private health insu ra n ce and to
spare insurers that choose not to contract with the
program from extensive regulation of their business. Implicitly, it strikes a bargain: Insurers that can
thrive in a market dominated by large employers
and a public contracting regime will be free to operate as they have before. Those whose competitive
advantage rests on cherry-picking healthy small
groups will face harder times.
To be sure,the political challenges will be formi-
Jacob S. Hacker. The Road to Nowhere: The Genesis of President
Clinton’s Plan for Health Security. Princeton: Princeton University Press,
1997, ch. 1.
demand that her elected representatives “keep government’s hands off
my Medicare.” Commentators quickly go on to conclude that Medicare
is in woeful shape: Even its beneficiaries don’t know that the federal
government runs it. Yet, if anything, this seems another strong political
count in the program’s favor—and a strong reason to base an expanded
public program on it.
28
It is often noted that some do not recognize that Medicare is a government program. A frequently repeated story, for example, describes an
elderly woman leaping up at a congressional town hall meeting to
29
100
dable. Yet this proposal holds a final advantage over
other likely contenders—what might be termed
“political robustness.” Political robustness refers to
the ability of a proposal to function successfully
despite alternative specifications and to sustain
itself politically over time. This proposal is politically robust in both senses. First, a number of its
features could be altered (or new features could be
added) without fundamentally compromising its
effectiveness. For example, this proposal would be
compatible with a system of refundable tax credits
targeted at lower-wage workers, as long as the credits could be applied to Medicare Plus as well as to
private insurance premiums, and the default
option for workers without employer-sponsored
insurance was Medicare Plus. Indeed, one possible
first step in implementing this proposal would be
to provide a refundable tax credit of a relatively
low value (say, , per person), but make it fully
convertible into coverage under Medicare Plus.
(The same could be done, of course, if a refundable
tax credit were already in place.) This would allow
Americans to apply the credit to employer-sponsored or non-group insurance if they chose, but also
would ensure that they could buy into a comprehensive program by paying a modest, income-related
premium.
By com p a ri s on , refundable tax credit schemes
are not so robust: It is difficult to guarantee, for
instance, that a tax credit will purchase adequate
benefits over time, and because take-up of private
insurance is highly sensitive to the level of the subsidy, the effects of a tax credit program on coverage
are dramatically different at alternative levels.
Second, and more important, the proposal creates several powerful self-reinforcing processes that
are likely to facilitate its expansion over time. Of
these, one that has already been mentioned is the
level of the payroll-based contribution. There will
be strong political pressure to keep the contribution
rate low. Yet, paradoxically, a low contribution rate
will bolster the size and bargaining power of
Medicare Plus. Similarly, if private health premiums
rise more quickly than the payroll-based contribution, then the share of the population enrolled in
Medicare Plus will also rise over time.
Another source of political robustness is the
structure of the premiums. As enrollment reaches up
the income ladder, Medicare Plus will naturally bring
in more revenue per enrollee. As the income of new
enrollees reaches the break-even point, the subsidies
required for new enrollees will be offset almost
entirely by the savings in federal tax subsidies for private health coverage. Moreover, controlling Medicare
Plus’s spending will also automatically reduce the tab
for existing tax subsidies by lowering the level of the
tax cap on private workplace coverage.
A final source of political robustness is the character and source of the subsidies. By targeting new
subsidies on workers enrolled in Medicare Plus, the
proposal ensures that the bulk of new spending goes
to the most vulnerable, and that these subsidies are
spent on a defined-benefit package that spreads risk
broadly across a large population. Not only does this
mean substantial savings over more scattershot subsidy arrangements; it also means clear lines of political accountability that would generally be lacking
in proposals that simply offered new tax credits for
private health plans. And because these subsidies are
financed almost en ti rely at the federal level, there
would be no concern that states would be under
competitive pressure to keep spending low, or that
responsibility for the program would be lost in the
interstices of state and federal duties.
In sum, this proposal offers medium-term
promise of political success. More important, it
holds out the hope that the next big step in
American health politics will lead not to another
political dead end but to the beginning of the difficult yet necessary journey toward universal health
insurance in the United States.
Acknowledgements
For valuable advice, I thank all the participants in
this project. Special thanks go to Judith Feder, who
offered useful suggestions at the pre-writing stage,
and to Elliot Wicks, who reviewed an early draft.
Gina Kramer provided able research assistance,
financed by the William Milton Fund of the
Harvard Medical School. These individuals and
institutions are, of course,absolved of responsibility
for the argument and analysis contained herein. n
101
Holahan, Nichols, and Blumberg Proposal
Key Elements
John F. Holahan, Len M. Nichols, and Linda J. Blumberg have outlined a new
proposal to cover the uninsured that would extend the subsidized coverage that
is available under the State Children’s Health Insurance Program (S-CHIP) to
all lower-income people. The proposal is built on the following key elements:
     financial incentives to states
to expand health coverage subsidies to all families and individuals with
incomes below  percent of the federal poverty level and those facing
higher-than-average health expenses, regardless of income. Subsidized
coverage could be purchased only through purchasing pools, which states
would have broad discretion to design.
       of a match amounting to 
percent more than the current Medicaid match, provided to states in
exchange for meeting minimum federal standards. States choosing to
participate would have to provide coverage to those meeting the eligibility
rules, but could specify a minimum benefits package consistent with federal guidelines, as in the S-CHIP program. Non-participating states would
continue their Medicaid and S-CHIP programs.
      Medicaid and S-CHIP.
People below  percent of poverty would get full subsidies, while those
between  and  percent of poverty would get partial subsidies, and the
high-risk (regardless of income) would be subsidized as well. A set of uniform federal rules would apply nationally.
 ,   with high health costs, could buy
insurance through a state-designed purchasing pool at a premium no
higher than a statewide community rate for a standard benefit package.
Employers must offer employees the state pool coverage as an option, but
they could choose whether to buy coverage exclusively through the pool.
      would combine existing Medicaid, S-CHIP,
state employees’ purchasing programs, and willing participants from the
private sector to create administrative efficiencies, pool insurance risks,
and improve bargaining clout for those within the pool.
102
About the Authors
 . , .., is Director of the Health
Policy Research Center at The Urban Institute. He
has authored numerous publications on the impact
of public policy and market forces on the health sector, with a particular focus on effects on low-income
populations. These include analyses of the recent
growth in Medicaid expenditures, variations across
states in Medicaid expenditures, and the implications of block grants and expenditure cap proposals
on states. He has also published research on the
effects of expanding Medicaid on the number of
uninsured and the cost to federal and state governments. He directs the health policy component of
the Assessing the New Federalism project which has
produced several reports on health policy for low
income populations. Other research interests
include health system reform, changes in health
insurance coverage, and managed care payment
arrangements. Dr. Holahan has a Ph.D. in economics from Georgetown University.
 . , .., is an economist and Principal
Research Associate at the Urban Institute who studies private health insurance markets and how they
work in response to decisions by employers, individuals, regulators, and public insurance programs. He
has written extensively on these and other issues
under the general rubric of health reform at both
the national and state levels. He has also been a consultant on health policy for the World Bank and the
Pan American Health Organization. Dr. Nichols is
currently a member of the Competitive Pricing
Advisory Commission (CPAC) for the Medicare
program as well the Technical Review Panel for
the Medicare Trustees Reports. He was the Senior
Advisor for Health Policy at the Office of Management and Budget (OMB) during the development
of and debate over the Clinton health reform proposals of –. Prior to OMB, Dr. Nichols was a
visiting Public Health Service Fellow at the Agency
for Health Care Policy and Research, and prior to
that he was an Associate Professor and Economics
Department Chair at Wellesley College. He received
his Ph.D. in economics from the University of
Illinois in .
    .     , .., is Senior Research
Associate at the Urban Institute. She is currently
working on a variety of projects related to private
health insurance and health care financing: estimating the coverage and risk pool impacts of tax credit
proposals, estimating employer price elasticities of
offering health insurance, the effects of insurance
market reforms on the risk pool of the privately
insured, and a series of analyses of the working
uninsured. While at the Urban Institute, she has
also done research on insurance market reforms,
the effects of the Medicaid expansions for children
on private health insurance, tax credits as tools for
expanding health insurance coverage, provision of
health insurance coverage by small employers, the
Health Insurance Portability and Accountability Act
of , the District of Columbia’s Medicaid program, reform of the Medicare AAPCC, and health
insurance purchasing cooperatives. From August
 through October  Dr. Blumberg served as
health policy advisor to the Clinton Administration
during its initial health care reform effort. She received her Ph.D. in economics from the University
of Michigan.
103
Expanding Health Insurance Coverage
A New Federal / State Approach
by John F. Holahan, Len M. Nichols, and Linda J. Blumberg
Overview
We propose a new consolidated federal-state health
insurance program based on five principles:
. Substantial new federal subsidies to finance expansion of coverage with state discretion to participate and to modify rules.
. Equity among individuals with similar incomes
and among states in a new program that would also
end the complexity in current public programs.
. Spread excess health risks broadly across the
general population.
. Organize the purchasing of subsidized health
insurance.
. Choice—all privately insured can keep their
current arrangement if they prefer.
These principles would be implemented in the
following ways.
Substantial new federal subsidies to finance expansion of coverage with state discretion. In those states
that choose to participate,full subsidies are provided
for those with incomes below  percent of the federal poverty level (FPL), and partial subsidies are
made available for those with incomes between 
percent and  percent of FPL. Within minimal federal standards—for example, minimum benefit
packages—states are given broad leeway to design
and or ga n i ze purchasing arrangements that work
best for their local conditions. Our federalism model
is the current State Children’s Health Insurance Program (S-CHIP), with high federal contributions and
considerable state flexibility. The federal share of the
total subsidy in our plan is  percent higher than
today’s Medicaid matching rates. Our model would
differ from S-CHIP in that there would be no fixed
budget allocations and states could not limit enroll-
ment arbitra ri ly. State participation is voluntary,
a n d ,a f ter five ye a rs ,s t a tes are free to impose an individual mandate if they choose. Participating states
also receive the higher match on their residual Medicaid progra m ,i n cluding long-term care benefits to
the elderly, and wraparound benefits for the nonelderly. As discussed below, participating states can
keep their disabled residents in the residual Medicaid program or bring them into the new program
(under special arrangements). Non-participating
states would continue to receive their current Medicaid and S-CHIP matching rates. The higher federal
share of expenditures in the proposed program is
designed to ensure that the net incremental cost to
states is relatively low and, thus, is intended to give
strong incentives for state participation.
Equity. All people at the same income levels in
each participating state are eligible for equal subsidies, whether they are currently en ro ll ed in
Medicaid, S-CHIP, or private plans, or they are
uninsured. This simplifies complex and conflicting
eligibility rules that discourage coverage expansion.
The primary condition for receiving subsidies is to
purchase insurance through a state-organized purchasing pool. States receive the same federal matching funds for all of their enrollees,unlike the current
system, which provides higher federal shares for
h i gh er-income S-CHIP children than for lowerincome Medicaid children. Finally, individuals with
incomes below  percent of FPL are treated equally in every participating state,and those b etween 
percent and  percent are treated similarly across
the states as well. The federal government imposes a
floor on the subsidy levels between  percent and
 percent of FPL.
Spread excess health risks broadly. No one,
104
regardless of his or her personal health risk, is
charged more for insurance than the hypothetical
(and computed) statewide community rate,and this
guarantee is financed with public (federal plus state)
do ll a rs . The statewide community rate is the rate
that would be charged by a competitive insurer for a
person of average risk for the standard benefits
package. This is a key new concept, and we devote
considerable attention to it. This explicit subsidy for
h i gh er-risk individuals is available only to those
who purchase through the new state-organized purchasing pool. At the same time, private insurers
would be free to price products outside the pool as
they see fit. The key design feature is that the financial consequences for higher-than-average risks are
shared across all citizens, because those inside and
outside the state pool pay taxes, the source of
financing for our proposed subsidies.
Organized purchasing of subsidized health insurance. The primary new institution is a purchasing
pool composed of Medicaid and S-CHIP recipients
and all others who choose to join. Federal law
requires this pool to be open to all at community
rates. Administrative efficiencies, risk pooling, bargaining power, and data collection for risk adjustment and community rate determinations are all
enhanced through this kind of purchasing en ti ty
and risk pool. States have broad discretion to design
their purchasing pools within federal guidelines,
and they can choose to en ro ll all state and local
employees in the pool as well, or give them the
ch oi ce of enrolling, as might any other employer.
Certainly the uninsured, but those with private
insurance who might prefer to purchase standardized benefits packages through this pool at a
statewide community rate as well,are free to join, or
they can maintain their existing arrangements. This
choice—to join the state pool or make their own
private arrangement—is a central el em ent in our
plan and our fifth principle, explained below. The
state-organized purchasing pool—which can be
operated by a private vendor—is required to operate (or contract for) its own managed fee-for-service (FFS) health plan and may choose to manage
competition among private plans as well. The purpose of requiring a state FFS plan is to ensure suffi-
cient enrollment capacity and to provide the state
with a basis of comparison for premium rates submitted by private insurers who wish to sell their
plans through the pool. The state assumes the insurance risk for this FFS plan.
Choice. All of the privately insured who want to
can keep their current arrangements may do so. No
new regulations are imposed on insurers outside the
state’s risk pool, and states are free to repeal insurance market regulations not required for compliance with the Health Insurance Portability and
Accountability Act (HIPAA). Employers are free to
offer coverage or not. If they do, they can offer their
coverage exclusively through the state pool, or they
can offer it both inside and outside the state pool. In
either case,the employer selects its preferred contribution level. All workers must have access to the
pool because subsidies will be available only to pool
en ro ll ee s . In addition, anti-discrimination rules
require that firms offering coverage inside and ou tside the pool must make equal contributions to
both. Any individual in the state is always free to
join the state pool during open enrollment.
Rationale
Our proposal is based on the premise that a purely
federal expansion of coverage is politically impossible. In our view, broad expansions based on a federal-state partnership are much more likely to earn
political support. The enthusiastic response to the
S-CHIP program suggests that public support for
coverage expansions should follow a joint federalstate model.
This plan also recognizes that individual and
small-group insurance markets have serious flaws
that are difficult to overcome and plague all proposals that rely on the p rivate market to expand coverage. Simply put, insurers and managed care
organizations have strong incentives to market to
the healthy and avoid the sick. State efforts to
reform small-group and individual insurance markets may have been somewhat successful at spreading risk, but they have not expanded coverage.
Guaranteed issue improves access for high risks, but
may increase premiums, thereby reducing the
105
attractiveness of coverage for low risks. Efforts to
force community rating reduce premiums for the
sick, but increase them for the healthy, leading to
possible reductions in the overall numbers covered
under a voluntary system.Our proposal spreads the
costs of h i gh - risk individuals more broadly and
uncouples that support from premiums paid by the
relatively healthy.
Our program is explicitly designed to recognize
that the system of coverage for low-income Americans has become increasingly complex and highly
inequitable. Medicaid eligibility rules are unfathomable to all but a handful of experts. Medicaid eligibility has become even more complex with the delinking of Medicaid and cash assistance fo ll owing
welfare reform.Enactment of S-CHIP with different
rules for income eligibility has added to the complexity that potential eligibles must navigate.
The current subsidy system is highly inequitable
because higher-income S-CHIP children receive
greater federal support than do lower-income children and their families who receive Medicaid. These
differentials have provided incentives for states to
favor S-CHIP enrollees or try to maximize S-CHIP
participation at the expense of Medicaid. Moreover,
within S-CHIP, states that have adopted broad
expansions (for example, Minnesota and Washington) could get a higher S-CHIP matching rate only if
they extended coverage further. In effect, there is a
financial penalty for having already enacted a broad
expansion. Our proposal eliminates this feature and
treats all states with similar income levels identically.
Finally, while no doubt providing some fiscal
benefits, current efforts in S-CHIP to prevent displacement of private coverage have led to significant
inequities. By limiting eligibility to the uninsured,SCHIP denies subsidies to families paying considerable amounts for individual coverage or for the
employee’s share of an employer’s plan,even though
their incomes are low enough by the program’s
standards to merit assistance. By providing subsidies
based on income and regardless of current coverage,
this proposal provides financial relief to those lowincome individuals who pay a lot for coverage in
addition to those we hope to encourage to purchase
health insurance for the first time. At the same time,
we preserve incentives for employers to con ti nu e
offering and helping to pay for coverage.
We have consciously decided against using tax
credits to expand coverage for four reasons.First, we
believe that subsidies for expansions of coverage
should be income-related. The administrative barriers to effectively providing income-related tax credits are tremendous when credits are provided at the
same time that required payments to insurers and
advance payments are reconciled with year-end taxable income. Credits would have to be provided at
the beginning of the year to ensure that low-income
persons had the liquidity necessary to purchase
health insurance. Because the credit amount would
be based on taxable income, advance payments
would have to be based on current or expected
income. If these estimates of actual full-year taxable
income were incorrect,the Internal Revenue Service
(IRS) would need to reconcile the amounts at the
end of the year. This “recapture” of overpayments of
credits would be costly, because low-income families tend to be inconsistent in filing income tax
returns.1 In fact, the cost of reconciling the credits is
likely to exceed the amount of m on ey actually
recouped. In addition, the risk that the credits or a
portion of them might have to be returned at yearend is a significant disincentive for low-income
individuals to participate. Second,and related to the
first issue, tax credits are particularly difficult to
administer for low-income individuals and families.
Such people are more likely to change jobs and have
gaps in employment, and may not be consistent tax
return filers. Thus, it is difficult to reach them with
advanced payments of credits made through
employers. Consequently, alternative administrative
structures would be required to serve this purpose,
but the need for such additional structures diminishes any efficiency to be gained by implementing
such a subsidy through the tax system.
Third, part of the subsidy we have designed is
Experience with the earned income tax credit shows that individuals
using the advanced payment feature of the credit have high rates of
non-filing of tax returns and not reporting receipt of the advanced payment if they do file tax returns. If this experience is consistent with what
would occur under a health insurance tax credit, the administrative costs
of correcting such errors and attempting to recover even modest underpayments from low-income persons may be significant.
1
106
explicitly related to individual health risk.
Determination of health care costs relative to the
state average is outside the scope of Treasury
Department activity and expertise. Appropriate premium contributions by individuals and appropriate
payments to insurance plans are best determined by
administrative entities at the state/local level, aided
by specific insurer, provider, and enrollee data and
information that our proposed state purchasing
authority would collect.
Fourth, most tax credit proposals allow people
to use their tax subsidies only in unorganized insurance markets. Indeed, this lack of interference in
markets is often viewed as a strength. We believe
that some structuring of the market in which subsidies are used is necessary to ensure efficiency, equity,
and access.
Coverage and Subsidies
In this program all participating states agree to provide full subsidies to all of those living below 
percent of poverty who enroll in the state purchasing pool. Further, participating states extend partial
subsidies to those between  percent and  percent of FPL. For those in this income range, states
can set the premium schedule, up to federal limits,
on overall cost-sharing burdens—that is, premiums
plus deductibles and coinsurance. These limits on
cost sharing are no more than  percent of family
income for those between  percent and  percent of poverty, and no more than  percent of
income for those between  percent and  percent of poverty. (Limits on cost sharing for children’s only coverage are limited to  percent of
income, as in S-CHIP). States can use their own
funds to subsidize individuals or families above 
percent of poverty if they choose. Participating
states also permit anyone, regardless of income, to
buy into the state pool at a premium that reflects a
statewide community rate.2 No subsidies (beyond
the current tax exemption for employer-sponsored
insurance) are extended to anyone purchasing cov2
The specific meaning of the statewide community rate is explained in
detail in the section on the state pool.
erage outside of the pool.
Participating states can choose to keep their disabled residents in the residual Medicaid program or
bring them into the new program. As participating
states, they receive the  percent higher federal
matching rate in either case. In the new program,
states are required to provide access for the disabled
to the state fee-for-service plan or to special managed care plans designed for the disabled.
The state program has to provide for guaranteed
issue—that is, anyone can sign up during open
enrollment. Anyone who does not sign up can enroll
retroactively by paying a full year’s premium plus a
 percent penalty. (Note that those below  percent of FPL face no premiums and, thus, have no
retroactive obligation, and those between  percent and  percent of FPL face reduced penalties).
The intent here is to avoid the severe adverse selection problems that result when an an individual
signs up after being diagnosed with a serious illness.
Given our income-based subsidy scheme, this
penalty is less serious than a -month pre-existing
condition exclusion and, in most cases, is less burdensome than the medically needy provisions of
Medicaid. Therefore, the medically needy path to
eligibility is eliminated for participating states.
    
 -
The crowding out of private insu ra n ce can be
thought of in a number of ways:
• employers dropping or not beginning to offer
employer-sponsored coverage given the availability
of public alternatives;
• workers dropping employer-sponsored coverage to enroll in public alternatives; or
• public spending on health care replacing current private spending.
Obviously, all of these are interrelated; however,
it is helpful to keep each one in mind, because our
program has different implications for each.
While some employers may drop coverage as a
result of our program, we do not expect that many
will. The current tax exemption for employer-sponsored insurance coverage continues to apply only to
those enrolling in insurance coverage through their
107
The administrative costs of plans offering coverage inside the pool
will be significantly below the administrative costs of existing
non-group policies, and such savings alone will likely be sufficient
to induce the vast majority of those in the non-group market to enroll.
employers, maintaining the incentive for workers to
purchase their insurance through their existing
group arrangements (if employers do drop coverage
and wages are increased as a consequence, those
increased wages become taxable income, while the
contributions to health insurance were not).
Competition for higher-wage workers who have
strong demand for employment-based tax preferences for health insurance will keep most firms
offering health insurance to be competitive in the
labor market. Although subsidies are available only
to those purchasing coverage through the state
insurance pool, employers are permitted to buy coverage for their employees in that pool. The state
pool premiums charged to employer groups are
based on a statewide community rate; the level of
the employer contribution to the premium is left up
to the employer. Firms competing for workers have
to maintain reasonably high employer shares to
attract workers who are ineligible for subsidies.
These provisions serve two purposes with regard
to the crowd-out noted above.First,they ensure that
there is no incentive for individuals to drop out of
employer-sponsored insurance arrangements.
Second,they limit the amount of private dollars displaced by public dollars, particularly in firms with
workers who earn vastly different amounts. In firms
with high-wage and low-wage workers, high-wage
workers will continue to want employer-sponsored
coverage to take advantage of the current system’s
tax subsidy, in addition to the convenience and
administrative economies of scale and risk-pooling
advantages of employment-based insurance coverage. Their interests have to be taken into account by
employers when the employers set their premium
contribution levels for the employer-sponsored
plans they offer (including plans in the state pool).
We do not purport to have developed an ironclad
approach to avoiding crowd-out; this was not our
intent. We do believe, however, that our approach
creates a reasonable balance between maintaining
much of the existing employer-based system and
generating much more equity by income class.
   —  
    
Below  percent of poverty most individuals and
families have an incentive to join the state pool.
Most of those with incomes less than  percent of
FPL who currently have employer-sponsored coverage will receive subsidies. Because of the offer of at
least partial subsidies, they will choose to obtain
coverage inside the pool. For those workers whose
firms drop coverage, presumably small firms with
low-wage workers,many will also enroll in the plans
offered by the pool.
Those currently in Medicaid and S-CHIP are
enrolled automatically. If the state chooses, state
employees can be automatically en ro ll ed , as well.
Among the low-income uninsured, incentives to
join are strong. Lack of information and indifference to health insurance are the greatest barriers;
there is plenty of evidence of non-participation by
those eligible for current public programs. With all
state residents eligible for enrollment, however, we
expect the stigma witnessed under the Medicaid
program to be reduced substantially.
Those who currently have private non-group
coverage are also likely to sign up because the pool
will offer more comprehensive coverage at lower
cost. The administrative costs of plans offering coverage inside the pool will be significantly below the
administrative costs of existing non-group policies,
and such savings alone will likely be sufficient to
induce the vast majority of those in the non-group
market to enroll. The income-related subsidies
available to this population increase the incentives
to join even more. Seasonal workers and those who
108
tend to change jobs frequently also may find the
pool attractive, because participating in it means
that changing jobs does not mean changing insurance plans. However, some non-group purchasers
who are eligible for only partial subsidies, and who
can obtain coverage at low rates due to excellent
health and/or the desire for less generous benefit
packages, may continue to purchase coverage outside the pool.
   —   
     
Those with incomes at or above  percent of FPL,
and who face high health insurance premiums either
because of above-average administrative costs (in
small groups or for individuals) or above-average
health risks due to poor health status, will find the
state plan to be attractive. This includes individuals
in firms with employer policies that have high premiums for either reason. They are able to purchase
some plans inside the state purchasing pool at a price
no higher than the statewide community rate. Those
benefiting from good experience rating or who are
willing to purchase less generous benefits packages
are less likely to enroll. Finally, while it is unclear
how many of the non-income-subsidized uninsured
will enroll, many of the uninsured are likely to find
the plans in the state pool more attractive than what
is available in the current non-group market, with its
extensive underwriting and high costs of comparing
benefits across insurers and plans
 
After a period of five years, states are permitted to
mandate that each individual obtain health insurance covera ge , either inside or outside the state
pool, individually or through an employer. This
delay is necessary to establish en ro ll m ent procedures, ensure efficient operation of the pools, refine
procedures for determining the statewide community rate, etc. The federal government will support
the mandate with the same schedule of subsidies
outlined above.
A mandate, or any serious expansion of coverage, permits the federal government to scale back its
support of acute-care activities that are now per-
formed outside the insurance-based system (for
example, disproportionate share hospital payments
[DSH]). We discuss this more fully in the section on
financing. In addition,a mandate is not likely to cost
substantially more than if a state adopted the voluntary version we have outlined. This is because, after
some years of the voluntary program’s operation,
most of those brought into coverage by the mandate
will be in households with incomes of more than
 percent of poverty, and thus would not receive
income subsidies. Those below  percent who
would come in only under a mandate are likely to be
the healthiest members of this group, so their per
capita cost should be lower than average, as well.
Federal/State Relations
States obviously have considerable responsibility
under this program in exchange for a large amount
of federal funding. States are required to meet federal standards for eligibility determination, outreach,
and enrollment.States have to incorporate Medicaid
and S-CHIP recipients into the purchasing pools
(which are de s c ri bed in detail below), along with
subsidized low-income individuals, employer
groups,and others who choose to enroll.States may
choose to incorporate state employees into these
pools, as well, or they can maintain a separate system. As with any other employer, states must offer
access to the pool and make the same contribution
to coverage in the pool. In addition, their workers
are eligible for low-income subsidies only inside the
pool. We expect that states will find it most efficient
to integrate current administrative structures for
purchasing insurance for state employees with the
new state purchasing entity. In some states,however,
this integration will take time to achieve politically.
To counter this,the federal government can provide
financial incentives for states to integrate their
employees early in the implementation of the program.
States have to establish procedures for informing
enrollees about their choices of plans and establishing standards of quality, provider payment,and risk
adjustment. States also have to develop a standard
benefits package that meets or exceeds federal bene-
109
fits package requirements. Federal standards will
include local flexibility in the spirit of the S-CHIP
program. For example, under S-CHIP, states must
establish a benefits package equal in actuarial value
to one of several benchmark plans, such as the standard Blue Cross and Blue Shield plan offered under
the Federal Employees Health Benefits Plan
(FEHBP),the health plan offered to state employees,
or the benefit plan offered by the health maintenance organization (HMO) with the largest market
share in the state. This flexible standard gives states a
great deal of leeway.
States are responsible for establishing fair and
equitable subsidy schedules so that premiums do not
exceed established maximum payments for individuals of particular income gro u p s .S t a tes are responsible for organizing purchasing pools, establishing
reporting and dissemination requirements, and
negotiating with plans over price or establishing
competitive bidding mechanisms.States are required
to operate (or contract for) a discounted fee-forservice plan to further ensure beneficiary choice and
provide an outlet for those worried about managed
care plans’ quality. States are not required to pay the
full cost of the discounted fee-for-service plan for
low-income enrollees if enough capacity and choice
is available in managed care arrangements.
Finally, states are responsible for operating a
residual Medicaid program. This would continue to
cover all groups (the elderly and,if the state chooses,
the disabled) and benefits (for example, nursing
home care) now required as part of Medicaid that
are not incorporated into the new program.
Optional groups and optional benefits can still be
provided at state discretion at the new higher
matching rate.
The federal government monitors state compliance with program rules. This includes ensuring
that states are meeting federal standards for eligibility determination, outreach and enrollment, and
some provider payment levels. The federal government monitors state procedures for calculating the
statewide community rate to ensure that it is fair to
beneficiaries—that is,not too high—and fair to the
federal government—that is, not too low (further
detail about the statewide community rate is pro-
vided in the next section). The federal government
also monitors state efforts to organize markets and
engender efficiency in their competitive bidding
processes or nego ti a ti ons with health plans. This
again is necessary to ensure that the federal government does not pay more than necessary to obtain
the coverage it seeks, and to disseminate lessons
learned by the federal government to inform other
states and improve performance nationwide.
The federal government monitors state spending, including supervision of subsidy calculations
for low-income people and for those with aboveaverage risk. The federal government also strictly
enforces provisions to avoid the financial manipulations that have occurred in Medicaid.3 For example,
the federal government might have to establish rules
on the maximum payments that can be made to
particular classes of providers. It may also be necessary to monitor payments made by health plans to
specific classes of providers. Both of these strategies
are possible mechanisms for preventing states from
encouraging providers to set their charges high,
thereby allowing the state to leverage more federal
matching funds. There is one natural limit to the
ability of states to en ga ge in these arrangements:
payments by plans to providers have to be covered
by a plan’s capitation rate. If the capitation rate is
too high,the plan has to charge premiums in excess
of the amount subsidized by the state. Given the
sensitivity of lower-middle-class individuals to premiums, plans should be reluctant to raise rates.
Why Rely on States?
There are several problems with a model that relies
so heavily on states.First,states differ widely in their
performance of current programs under current
arrangements. Among the  individual states represented in the National Survey of America’s Families
(NSAF), uninsurance rates for low-income children
vary from a low of  percent in Massachusetts to a
high of  percent in Texas, and, for low-income
These include disproportionate share hospital payments, supplemental
payments made to public hospitals and nursing homes financed with
intergovernment transfers, and other arrangements that have had
the effect of obtaining federal funds with little or no state and local
matching funds.
3
110
adults, from  percent in Massachusetts to  percent in Texas.4 According to the Current Population
Survey, Medicaid coverage varies from  percent of
low-income adults in Nevada and  percent in Idaho
to  percent in Tennessee and  percent in Vermont. In , Medicaid spending per child enrollee
va ri ed from , in New Hampshire to  in
Mississippi.5 Eligibility standards for children under
S-CHIP ranges from a maximum of  percent of
poverty in South Dakota and North Dakota to 
percent of poverty in New Jersey.6 Reliance on states
alone can result in Americans with com p a ra bl e
incomes being treated quite differently.
Second, as referenced above, a range of financial
abuses, including disproportionate share hospital
(DSH) payments and, more recently, supplemental
payment programs, have allowed states,at their discretion, to increase their effective matching rates.
These financing abuses have led to widespread skepticism about state discretion at the federal level, and
have threatened the viability of federal-state financial relations.
Third, there is extreme variation in administrative capacity at the state level. States such as New
York, Massachusetts, Minnesota, and Washington
have far more health policy expertise than do many
of the smaller states in the south and west.
Despite these problems, we believe that it is
politically unrealistic to enact a broad expansion of
coverage at the federal level at this point in time.
Right or wrong, the mom en tum in the nation is
toward greater reliance on state government. We
also believe that S-CHIP offers a fundamentally different model from Medicaid.S-CHIP has combined
higher federal matching payments with more state
flexibility. The higher federal matching rates have
made coverage expansions for children much more
4
Urban Institute tabulations of 1999 National Survey of America’s
Families data. Details of all Urban Institute calculations are available on
request.
Urban Institute calculations based on Health Care Financing
Administration (HCFA) 2082 data and the March Current Population
Survey. Form 2082 financial and enrollment data are supplied by the
state to HCFA.
5
F. Ullman, I. Hill, and R. Almeida. CHIP: A Look at Emerging State
Programs, New Federalism Issue Brief Series A, no. A-35, Washington:
Urban Institute, 1999.
6
financially attractive to states, and state governors
have been able to receive credit for reducing the
number of uninsured children. All states have
adopted S-CHIP (and the majority have extended
coverage to at least  percent of poverty),and several have expressed interest in extending coverage to
include parents. With a high level of federal matching funds it will be hard for states to walk away from
the opportunity to expand coverage to low-income
and high-health-risk individuals. For all these reasons, we believe that S-CHIP provides a good model
of federalism to follow.
One concern with state stewardship is inadequacy of funding. However, with about  percent of a
state’s population enrolled in the program, program
beneficiaries should have sufficient political power
to avoid chronic underfunding. While there is more
state flexibility under S-CHIP than under Medicaid,
a range of federal standards is essential. For example
under S-CHIP, there are rules for minimum benefits
packages. As described above, we believe the federal
government will need to set rules for benefits packages,minimum provider payment standards, operation of the pools, and to avoid the financial abuses
of Medicaid. A different concern is that states will
not be able to control the growth of costs. However,
because state expenditures will still be large compared with other state spending, states will have an
incentive to control costs.
Organization of the State Purchasing
Pools and the General Insurance Market
Development of State Health Insurance
Purchasing Pools
Under our reform proposal,each state is required to
construct a single purchasing authority and risk
pool through which insurance coverage is provided
for those who are subsidized because of low
incomes or above-average health risks and for those
with higher incomes who want to take advantage of
the choices and efficiencies inherent in large-group
purchasing mechanisms. Those subsidized because
of income include most of those currently eligible
for or enrolled in Medicaid or S-CHIP and many
who are not currently eligible for those programs.
111
To the extent that a state has existing state-only
comprehensive insurance coverage programs (for
example, Washington State’s Basic Health Plan), we
expect that these states will integrate these programs
into the state purchasing pool to secure the large
federal share of subsidy dollars for this population.
Our equity principle requires that all subsidized
enrollees with equivalent incomes be treated in the
same manner inside the pool—that is, the distinctions between Medicaid and S-CHIP and other
types of enrollees are erased. In addition, individuals exceeding the income eligibility cutoffs for subsidies and private employer groups may purchase
coverage through this pool. All purchasers/enrollees
in a given geographic coverage area have access to
the same health plans and enrollment options,
regardless of whether they are subsidized (for some
plans, additional payments will be required—see
below).
States have the option to contract with private
insurers and managed care plans to provide coverage through the state pool. However, even if private
insurers are used in this manner, each state is
required to operate its own managed FFS plan. The
state can either run this plan directly or contract
with an insurer or a third-party administrator for
this purpose. In addition to FFS being an increasingly popular option for Medicaid managed care
programs, one important benefit of this managed
plan is that it serves as a safety valve and as a check
on proprietary health plans’ bidding and contracting strategies. For example, if plans all submit relatively high bids, the state may be able to operate a
managed fee-for-service program, one that monitors use and nego ti a tes fee discounts from
providers, at a lower cost. Our vision of this new
health insurance program is sufficiently general that
one could imagine a state qualifying for the subsidy
payments with just this managed FFS plan—that is,
without trying to manage competition among proprietary health plans. An FFS plan may also be
essential in states or substate areas where sparse
populations prevent managed competi ti on from
developing. The state purchasing pools can build on
existing structures for state employees’ plans,
Medicaid, S-CHIP, or other state purchasing pro-
grams, or entirely new entities may be created at the
state’s discretion.
As noted earlier, the new program provides two
types of subsidies. The first provides premium subsidies to those below  percent of poverty, with
those in families below  percent of poverty
receiving subsidies sufficient to cover the full cost of
a comprehensive plan. The second subsidizes individuals with above-average health risks so that the
premium they face is equivalent to a community
rate calculated over the entire insured population in
the state. Both subsidies are available only to those
en ro lling in coverage inside the state purchasing
pool to minimize administrative complexity.
General Features of State Pools
Once the state determines the benefit package, subject to federal minimum requirements, private
insurers may decide to sell supplemental benefits in
the pool. But these benefits must be priced separately and treated as add-ons to, not replacements for,
the standard package. Those enrollees who are entitled to additional benefits because of Medicaid eligibility under current law (for example, children with
special needs and the disabled) continue to receive
those benefits as wraparounds through a residual
Medicaid program (which also continues to provide
long-term care as under current law). The state purchasing pools operate under guaranteed issue and
guaranteed renewal for all groups and individuals.
Each year the state purchasing pool holds an open
enrollment period at least one month long.
Enrollment will be permitted at non-open enrollment times; however, late enrollees (regardless of
the month they enroll) will be required to pay what
they would have paid in premiums for one full year
plus a  percent penalty. As de s c ri bed above, for
workers with employer offers, the income subsidy
inside the pool can be applied only to the employee
share. No additional insurance market reforms are
required outside the state purchasing pool. We
assume that HIPAA remains in place in non-participating states,and that participation in the program
satisfies HIPAA’s requirements for an individual
market mechanism.
112
Premium Payment Details
We have four goals for our premium payment
mechanics: () spread excess health risk broadly
throughout the population; () provide subsidies
for those with low incomes; () create incentives for
efficiencies in care delivery and insurance administration; and () ensure that private insurers remain
willing sellers—that is, that they co ll ect enough
money for the risks they are bearing. This section
and the more technical appendix explain the pricing
mechanics through which we achieve these goals.
One of our core principles is to set premium
payments so as to spread excess (above-average)
health risk across all taxpayers. No person enrolling
in the state purchasing pool has to pay more than he
or she would under a statewide community rate for
the state-determined standard benefits package.
Low-income persons are subsidized either completely or partially, depending on their income.
To show how this works, we present two examples. The simplest case is represented by a state using
only a managed FFS arrangement for its risk pool—
that is, there are no competing private health plans.
In this case, the state determines the premium it
charges all enrollees, based on expected use of an
average person in the entire state as well as negotiated price and utilization management incentives it
has agreed to with participating providers. In this
case, the FFS plan is the benchmark plan, by definition. Those with incomes below  percent of FPL
can enroll in this plan without cost. Those with
incomes between  percent and  percent of
FPL pay a share of the statewide community rate
according to the schedule determined by the state.
Individuals with incomes above  percent of FPL
are able to buy into this managed FFS plan by making a payment equal to the statewide community
rate, which could entail some premium variation by
age at the state’s discretion.
Federal and state governments share in the costs
of the low-income subsidy, according to the relevant
matching ra te . The governments also jointly bear
the difference between the statewide average cost
and the expected cost of providing care to those
enrolled in the state pool’s FFS plan. This includes
the higher costs due to the pool’s attracting higher-
than-average-risk enrollees.
The managed competition regime, under which
competing private health plans join the state-run
FFS plan as an option for en ro ll ee s , is somewhat
more complicated. Its details are laid out in the
technical appendix. In this case, the state must perform four key functions to determine appropriate
payment amounts.
. Elicit “bids” from plans interested in participating in the pool.
. Choose and implement a risk-adjustment
mechanism from the choices approved by the federal government.
. Set the benchmark within-pool rate.
. Determine the statewide community rate.
We describe these functions in the fo ll owing
paragraphs.
Elicit bids from plans. Insurers that want to compete for the substantial business inside the purchasing pool must submit premium bids for a standard
risk enrollee—that is, a healthy young adult male.
The healthy young adult male is chosen so that
insurers do not base their bids on different expectations about the people who are likely to enroll in the
state pool. There are two advantages to having plans
bid this way. First,it is standard procedure for insurers to determine premiums for the standard risk.
Second, it allows the state to have a uniform measure for comparing plan bids to non-pool rates,
because plans typically file rates for standard risks
with state insurance departments. While competitive bidding is one clear mechanism for eliciting
such bids, the state does not have to use this
approach. It can also set or negotiate standard rates
with the plans. In addition, insurers include an
administrative load in their standard risk bids.
Choose and implement a risk-adjustment mechanism. The state must also establish a risk-adjustment system to compensate plans appropriately if
they enroll above- or below-average-risk individuals. The federal government provides a list of
acceptable risk-adjustment methods from which
states select one. States can also request approval for
their own risk-adjustment method. The approval
process is intended to guarantee some consistency
and quality control across the nation. States use the
113
risk-adjustment mechanism to compare average
risks enrolled in each plan with average risks
throughout the state. These relative risk rankings
are necessary to determine premiums to be paid by
enrollees and the state and federal governments,and
to be received by plans.
Set the ben ch m a rk within-pool ra te . We expect
plans to submit or nego ti a te a range of bids. The
government then determines or selects a benchmark
bid. This can either be the median bid or it can be
set at a pegged level—for example,  percent of the
lowest bid. The benchmark bid must be set high
enough to engender sufficient plan capacity to
en su re that those who are fully subsidized have
some choice of plans, and low enough so that all
plans have incentives to become efficient and bid
low. The benchmark bid is used in setting the level
of the full income-related subsidy, the excess risk
subsidy, and the statewide community rate. Because
the government cannot commit financially to providing low-income persons a complete subsidy for
any plan that they might choose, the benchmark is
used to define a reasonably efficient premium level.
If any plan’s bid is above the benchmark, the individual choosing to enroll in that plan (regardless of
income) is responsible for paying the excess.
Determine the statewide community rate. The
state must compute the expected average premium
that is sufficient to cover health services plus administrative loading costs as if all those insured in the
state were enrolled in the in-pool benchmark (efficient) plan. This premium is what we refer to as the
statewide community rate. It is calculated by adjusting the benchmark bid for the standard-risk person
to the level of average risk of all those insured
throughout the state (not just in the pool). This
requires insurers to report average risk scores both
inside and outside the state pool. Because this is a
difficult calculation and necessarily an approximation, it may be necessary for the states to supplement insurer reports with other information. Two
representative data sources come readily to mind:
the Medical Expenditure Panel Survey of employers
collects state-specific premium data for most states,
and actual provider revenue data collected to support the Commerce Department’s estimates of gross
domestic product (GDP) are used in HCFA’s ongoing state and national health expenditure estimates.
Both of these sources can be used to develop alternative ways to calculate the statewide community
rate for adults and children. It may also be politically necessary to use panels of experts,including actuaries, plan representatives, beneficiary advocates,
and representatives of federal and state government,
to decide on the reasonableness of the statewide
community rate estimate.
Calculating the statewide community rate reasonably accurately is important. If it is estimated too
high, potential beneficiaries will have to pay more to
en ro ll in a state pool than we intend, and too few
will want coverage inside the pool. If it is estimated
too low, cost to the federal and state government
will be higher than necessary with more enrollees
coming into the pool for the subsidies, and the per
capita excess risk subsidy being too large.
Because calculating the statewide community
rate is so politically charged, it is probably desirable
to err on the high side in the beginning; this maximizes the chances that the system will be up and
running without implacable insurer opposition.
The extra cost this imposes will be low in the early
years because only a few states are likely to implement the program initially. Once calculation of the
statewide average cost is institutionalized, fluctuations and uncertainty in this estimate will be
reduced, and lessons learned in the early years can
be applied in the bulk of states that will implement
the program at a more measured pace, and in all
states over time.
When these calculations are completed, we have
several pieces of essential information. We have bids
from each plan, risk adjustments reflecting the
health status of enrollees of each plan in the pool, as
well as in the en ti re state, and calculations of the
statewide community rate. The state can then determine the required contributions from beneficiaries,
payments to plans, and payments by government.
What enrollees will pay. E n ro ll ees in the state
pool are responsible for paying the statewide community rate,less any applicable income-related subsidies. Employer contributions count toward an
enrollee’s obligation. In addition, if the beneficiary
114
chooses to en ro ll in a plan whose bid exceeds the
benchmark bid, he or she will also have to pay the
difference between the two bids. If the beneficiary
chooses a plan whose premium falls below the
benchmark, the plan may choose to rebate the difference to the enrollee as an incentive. The enrollee’s
payment is never affected by the mix of risks in the
plan he or she selects.
What plans are paid. Plans receive a premium
payment equivalent to their bid for the standard
risk, adjusted for the relative risk of those individuals actually enrolling in their plan. If a plan’s bid is
below the benchmark bid, the plan can offer a discount to the beneficiary. Thus, plans receive the full
costs they expect to incur and that their competitive
bids reflect. But, in general, it is important to note
that plans get part of their payment from enrollees
and part from the government in the form of
income subsidies, excess risk subsidies, and riskadjusted premium payments.
What the government pays. The government is
responsible for paying two types of subsidies. These
subsidy dollars flow directly to plans; they are not
transferred through enrollees themselves. First, the
government pays all of the applicable low-income
subsidies attributable to enrollees in each plan.
Second,the government pays the difference between
the benchmark standard-risk bid, adjusted for the
relative risk of those enrolling in a particular plan,
and the statewide community rate. This second type
of payment compensates plans for the difference in
the health risk of their en ro ll ees compared to all
insureds in the state. This payment does not subsidize a plan’s inefficiency, however, because the
health risk subsidy is based on the cost of the plans’
enrollees as if they were enrolled in a plan priced at
the benchmark. Pure inefficiency—higher bids than
the benchmark bid for the standard risk—must be
recouped from enrollees.
State options. Each participating state has a number of choices that are sometimes limited by federal
floor or minimums. Nevert h el e s s ,t h ere is considerable state discretion to affect the character of the
purchasing pool. In addition to choosing whether to
contract with private plans, some of these choices
result from asking the following questions:
• Will employers choose a plan on behalf of
employees, or will employees have full ch oi ce of
plans in the pool?
• Will state and local government employees be
incorporated fully into the pool, or will they maintain separate coverage with the option for workers
to enroll in the pool?
• Will the pool be publicly managed, or will it be
run through a public-private partnership?
• What will the premium structure be—single
vs. family; single, couple, single-parent, two-parent
(from a menu, or with federal approval)? Each state
will be required to offer a child-only premium.
• Will any age adjusters be used for setting the
unsubsidized premiums within the pool?
• What will be the pool’s approach to private
insurer premium determination within the pool?
Will the state rely on pure competitive bidding,
negotiation, or some hybrid? (All pools will have the
power to exclude plans deemed unacceptable by reason of quality and/or price.)
• What will be the exact risk-adjustment method
(from a menu or with federal approval)?
• How large a geographic area will be served by
each pool (statewide, substate, etc.)?
Financing — Who Will Pay?
This program is paid for through a combination of
federal and state funds.The proposal is expensive, to
be sure, as is any serious proposal to expand covera ge . There are currently  million non-elderly
adults and  million children living below  percent of poverty. Of these,  million adults and 
million children currently have employer-sponsored coverage. Another . million adults and .
million children have private, non-group coverage.7
Even though many of these individuals currently
have covera ge , they can receive subsidies, if they
choose to join the state plan. Another . million
adults and . million children have Medicaid;
under our proposal the federal government pays a
h i gh er share of expenditures on this population.
Urban Institute analysis based on data from the 1998 March Current
Population Survey.
7
115
We are not substantially increasing the total number of dollars
spent on health care services; rather, we are changing the
financing source of those dollars to be more heavily weighted
to federal, progressive income taxes.
Finally, . million low-income adults and . million low-income children are uninsured and can
begin receiving coverage.
The large number of low-income adults and
children who already have coverage but can
nonetheless join the state pool means that the price
of achieving equity—that is, treating individuals
with similar incomes the same—is high. It is important to remember, however, that we are proposing to
change the ways in which health insurance is paid
for with regard to low-income people and people
with above-average health risks. We are not substantially increasing the total number of dollars spent on
health care services; rather, we are changing the
financing source of those dollars to be more heavily
weighted to federal, progressive income taxes.
Our proposal involves considerable expenses
attributable to health risk subsidies for those above
 percent of poverty. Because health expenditures
are very unevenly distributed (that is, a large share
of expenditures is attributable to a small percentage
of the population), a relatively small proportion of
the population benefits greatly from the risk-related
subsidy, but this subsidy is likely to account for a
large share of premium dollars.
These costs are borne by both federal and state
governments, but primarily the former. Compared
to proposals that would be completely federal (for
example, refundable tax credits that could achieve
the same degree of coverage, if generous enough),
the federal expenditures for our proposal are lower
because of the state contribution. The federal share
is paid for through federal general revenues—and by
cuts in existing programs,the need for which would
be reduced under such a reform. The increase in federal expenditures is also limited by the strong incentives for employers to con ti nue contributions to
their employees’ coverage. These incentives include
the tax advantage to higher-wage workers, market
pressures to compete for workers who demand
health insurance at most or all jobs, and anti-discrimination rules that make it difficult to exclude
subsets of workers from health benefits that some
workers value highly enough for the firm to offer.
We fully expect that some time after full implementation of the program, the federal government
and the participating states will be able to begin to
reduce other direct payments made directly to
providers (for example, DSH payments) to fund
care for the uninsured and low-income populations.
While the need for such extra sources of funding
should be lower under the proposed program, we
do not anticipate eliminating these payments completely. Even in states that do adopt an individual
mandate, a residual number of uninsured are sure
to persist,and some support for providers in particularly high-need areas, such as inner cities, may still
be necessary. In those states not choosing to participate in the program, the costs associated with
providers in high-need areas can be expected to
continue to grow. Changes in federal programs will
have to be sensitive to the different needs in participating and non-participating states.
Under our proposal, states pay lower matching
rates than they do under Medicaid, but they will pay
on many more lives. At the same time,they get more
federal dollars for those they currently cover under
Medicaid. Moreover, they receive the higher federal
match on residual Medicaid benefits. This is a large
substitution of federal funds for existing state obligations. But federal contributions for those above
 percent of poverty under the current S-CHIP
program are eliminated; therefore, those states that
have extended S-CHIP coverage to higher-income
levels lose some federal benefits. Still, remembering
the savings states get from higher matching ra te s
across the board in Medicaid, states are free to use
some of their savings to maintain or increase subsi-
116
dies for those above  percent of FPL with stateonly money, or to supplement the federal subsidies
between  percent and  percent if they prefer.
In addition,states that have existing high-risk insurance pools can eliminate them and send enrollees
into the new state combined pool.
There are several sources of potential savings to
different groups that offset much of the necessary
tax increase. Many individuals and businesses will
pay lower premiums because of lower administrative costs that result from switching into the pool
from the current small-group and non-group markets. In addition, many of those who remain in the
private market will pay lower premiums because
many above-average risks will enter the pool
because of the community rate. To the extent that
insurers currently price small-group and individual
policies at inefficiently high levels because of fears of
adverse selection that do not actually come to
fruition,this means real system savings. The consolidated purchasing power of the new pool should
engender new efficiencies among health plans and
throughout the health care delivery system. Lower
premium payments will also lead to higher wages
and more tax revenue than would have been the case
at the same income and payroll tax rates.
Political Feasibility
This proposal has a number of political strengths,
and its share of weaknesses. From the perspective of
federal executive and congressional policy makers,
the plan offers a way to attack a major nati on a l
probl em . Political leaders will receive credit for
offering a solution to the problem without a major
expansion of the federal bureaucracy. On the other
hand,leaders at the federal level will have to bear the
political burden of using surplus revenues or raising
taxes to finance the program. Further, while federal
political leaders will provide general oversight, most
of the credit for the program’s success will go to
state political leaders.
At the state level, political leaders have the
opportunity to receive the credit for solving a serious
issue for their constituents with very little increase in
state revenues. They have the luxury of building con-
sensus and adopting a program only after that consensus exists. However, there will be some increased
financing at the state level. Moreover, states will have
a large number of new administrative challenges,
and state leaders will bear the brunt of criticism for
operational failures.
Liberals are likely to applaud the coverage
expansion that will result, but they will be dissatisfied that the proposal leaves us short of universal
coverage and grants so much state discretion. There
may be opposition to segmentation of low-income
and less-healthy Americans into state purchasing
pools, separate from plans serving higher-income
and healthier Americans. There is also likely to be
distrust and lack of confidence in states’ ability to
administer such a complex program and concerns
about chronic underfunding.
Conservatives will oppose a large government
initiative and the introduction of what is essentially
an expanded entitlement program. They will view
the large amount of spending on those who currently have coverage as inefficient. Conservatives are
likely to find favor with the large state role,and they
cannot argue that it is a “one size fits all” program.
The fact that any individual or family can continue
to choose a private plan should mollify some conservative critics.
The business community should clearly benefit.
No firm will be worse off, and many will have new,
less expensive options. The administrative burdens
on many smaller firms will be reduced,and the costs
of coverage will fall because of the movement of
above-average risks into the pool. Some of the savings, however, might be offset with wage increases
over time.
Insurers and providers are likely to find the plan
a mixed blessing. While it will greatly increase the
number of covered lives, insurers are likely to be
concerned about so many lives being within the
scope of the state-organized pool. They are likely to
fear what organized purchasing and bargaining
power can do to their traditional discretion to segment markets and earn very healthy profits. The fact
that a voluntary market outside the state pool is not
regulated beyond current law will be viewed as a
positive, however. Similarly, providers will also see
117
the benefits of more covered lives, but they will be
concerned about government bargaining power
being used to lower capitation rates, which can
affect their revenues.
Health insu ra n ce consumers would either be
better off or, at least, no worse off. They can retain
existing arrangements, which may become cheaper,
or seek a new offering in the state pool. People with
low incomes, those in poor health, and those in
small firms would clearly benefit. To the extent that
increased competition holds down the cost of premiums, those working for employers that offer coverage will see higher wages in the long run. Of
course, taxes will be higher than otherwise, offsetting these benefits to some extent.
Transitions
Our general principle is to leave the pace of the transition to the new subsidy and pooling structure to
the individual states, with financial incentives to
start early. We offer two-year planning grants to
states that are willing to start the first year. These
planning grants include enough money to fund
development of standardized reporting formats for
risk adjustment and statewide community rate cal-
culation purposes. One advantage of our system is
that it builds on existing institutions: state Medicaid
agencies, S-CHIP, and state employee health insurance programs all perform many, if not all, of the
functions we have in mind, except for calculating the
statewide community rate. Several states—for
example, Minnesota, Washington, and Massachusetts—have many of the features of the system we
are proposing already in place. We anticipate that
other states will choose to participate in the program
at varying rates. An advantage of delayed state entry
is the opportunity for those entering later to learn
from the experi en ce of the pioneers. Our recomm en ded pace is two years of planning, combining
Medicaid and S-CHIP in year two, and then adding
free choice of the pool in year three.
At least five years of state operation under the
new system are required before the state could opt
for an individual mandate. This waiting period
serves two functions: first, it ensures that the state’s
pool is operating effectively and efficiently; and,second, it permits federal and state governments to
generate more accurate estimates of the cost of the
mandate for budgeting purposes. Enforcement
mechanisms are up to the states to devise, and they
need to be approved by the federal government.
Technical Appendix
Premium Payment Details, Managed
Competition Case
This appendix explains how we determine beneficiaries’ obl i ga ti on s , government payments, and net
payments to health plans,and achieve our objectives
of equity and efficiency. Initial bids will be requested from all private plans participating in the state
pool; bids would reflect the insurer’s price for a
standard risk, including administrative load.
“Standard risk” is used to define the premium cost
for an individual of a specified age group and gender, usually a young adult male, the lowest-cost
insured adult. It does not mean the average risk for
the population insured in the pool. Expected health
costs and administrative costs are embedded in this
bid for the standard risk.
A given plan’s bid is denoted as SRi. From the
distribution of bids, the state purchasing authority
will determine a benchmark bid, SR b. As noted in
the text, states may decide to set premiums through
a pure competitive bidding process, they may use a
more interactive negotiation approach, or they may
set the “bid” unilaterally and then accept only those
plans willing to accept that rate, or some alternative
method. This standard risk bid will be converted to
a premium level by using a risk adjuster that takes
into account the difference in expected health costs
118
between actual enrollees in an insurer’s plan within
the pool and the expected cost of the standard risk.
In general,the premium for the i-th plan is Pi = SRi *
(1 + RAi), where Pi is the premium, SRi is the i-th
plan’s bid for a standard risk enrollee, and RAi is the
risk adjuster or risk score for all the enrollees in plan
i, relative to the standard risk. RAi = 0 implies that
the i-th health plan has drawn all standard risks,and
RAi = .12 means that the plan’s enrollees are expected on average to be  percent more costly than the
standard risk.
We do not have strong opinions about which
risk-adjustment method states decide to use,
though the federal government should approve the
method chosen to guarantee some consistency and
quality control across the nation. Any number of
specific techniques will produce risk scores relative
to the . standard risk that is bid,and it is these relative scores that make up RA.Each person, given his
or her characteristics and recent medical history, has
an RA score, and average RAs can be computed for
any group of persons.
Two other concepts are prerequisites to a full
explanation of our payment mechanics. The first is
the risk adjuster score that reflects the average risk
in the state as a whole, RAst , and the other, which
builds on this, is the statewide community rate, or
CRst = SRb * (1 + RAst). That is, the statewide community rate is the product of the benchmark bid
and the average risk statewide, incorporating those
both inside and outside the pool.
Now, to remain in business, each health plan
must collect its expected costs, given the actual average health risk it happens to enroll, or Pi = SRi * (1 +
RAi). Our equity principle says that no person
should pay more than the statewide community rate
CRst , except for an inefficient plan, one that bids
above SRb. This CRst is also the amount of the full
income-based subsidy, so that an en ro ll ee in a
household living at less than  percent of poverty
can choose an efficient plan and pay nothing out-ofpocket. Partial income subsidies are fractions of 
percent, s, (s £  percent), times CRst.
Thus, the community-rated premium for plan i
would be CRi = SRi * (1 + RAst). This is the standard
risk bid by plan i, adjusted for the risk of all insured
individuals in the state. CRi is the premium facing
an individual enrollee in plan i who is not eligible
for an income-related subsidy. Those who are eligible for low-income subsidies would pay:
CRi - sCRst = SRi * (1 + RAst) - s * SRb * (1 + RAst)
= (SRi - s * SRb) * (1 + RAst).
When s = 100 percent,and SRi = SRb, those with
incomes below  percent of poverty pay nothing.
If a plan bids above the benchmark, or enrollees’
income exceeds  percent of poverty so that s <
1.0, then they must pay something out-of-pocket
for that plan.
The government must pay the income subsidy
and make sure the plan is adequately compensated
for the risk profile it actually attracts. The simplest
way to accomplish both these goals is to set the government amount as a residual, the difference
between what the plan requires to stay in business
and what the enrollee pays. The government then
would pay:
SRi * (1 + RAi) - (SRi - s * SRb) * (1 + RAst),
which can be rearranged to yield:
SRi * (RAi - RAst) + s * SRb * (1 + RAst).
The first term in the government obligation is
the combination risk-adjustment and risk subsidy
payment that compensates plan i for its enrollees’
risk relative to that in the state as a whole. The second term is the income subsidy amount appropriate
for an enrollee with a particular subsidy level, s.
Obviously, aggregate government payments will be
determined by the average risk and subsidy levels of
in-pool enrollees, in addition to the efficiency of
plans that compete in the pool. Thus, both governments and beneficiaries gain if most plans are efficient and, therefore, have strong incentives to foster
a competitive climate and bidding mechanism that
encourages health plan efficiency. Health plan efficiency is key to minimizing the cost of achieving our
goals of coverage expansion and equitable access for
the low-income and the high-risk. n
119
Kronick and Rice Proposal
Key Elements
Richard Kronick and Thomas Rice propose that the United States adopt a
health care financing system that would:
   ..  a “right” to comprehensive health insurance
coverage. Health insurance would be a social insurance program, not a
means-tested program.
   for designing and administering the health
care financing system, allowing them the flexibility to create systems
that meet the needs of their residents. To receive federal funding, states
would need to assure that nearly all legal residents would be covered and
have access to at least one zero-premium plan that includes a federally
defined standardized benefits package.
            , which
largely relies on employment-based health insurance, with one relying
on a payroll tax levied on employers and employees. This tax would be
supplemented by federal general revenues, state revenues, and, possibly,
individual contributions for plans or benefits beyond those in the
standardized benefits package.
    - long-term care.
       and make a substantial investment in
measuring quality and outcomes, particularly for vulnerable groups.
120
About the Authors
 , .., is Associate Professor in the
Department of Family and Preventive Medicine at
the University of California at San Diego. His work
focuses on understanding how and whether markets
can be made to work well in health care,particularly
for vulnerable populations. He recently co-authored,
with Joy de Beyer, Medicare HMOs: Making Them
Work for the Chronically Ill (Health Administration
Press, ). He has developed and helped state Medicaid programs implement risk-adjusted payment
systems. In – he was a Senior Health Policy
Advisor in the Clinton Administration, where he
contributed to the design of the Administration’s
health care reform proposal. In the late s he coauthored, with Alain Enthoven,a proposal to achieve
universal coverage in the U.S.,and contributed to the
theory of ‘managed competition’. Prior to that he
served as the Director of Policy and Reimbursement
in the Massachusetts Medicaid program.Dr. Kronick
received his Ph.D. in Political Science from the University of Rochester.
     , .., is Professor in the Department
of Health Services at the School of Public Health,
University of California at Los Angeles. He was
Department Chair from  to . Dr. Rice is a
health economist, having received his doctorate in
the Department of Economics at the University of
California at Berkeley in . Before joining the
faculty at UCLA in , he taught at the University
of North Carolina School of Public Health. His
areas of interest include health insurance, competition, physicians’ economic behavior, and Medicare.
Dr. Rice has received the Young Investigator Award
() and the Article of the Year Award () from
the Association for Health Services Research. His
book The Economics of Health Reconsidered was
published in . He served as editor of the journal
Medical Care Research and Review from  to
.
121
A State-Based Proposal for Achieving
Universal Coverage
by Richard Kronick and Thomas Rice
Overview
We propose that the United States adopt a health
care financing system that provides comprehensive
health insurance to all legal residents. The system
will be administered by the states and overseen by
the federal government.Employers will no longer be
involved in providing health insurance covera ge ,
although both employers and employees will contribute. Other sources of financing include the federal and state governments and, in some instances,
individuals and families. The proposed system will
replace most of the major components of the U.S.
health care financing system; the two exceptions are
Medicare and Medicaid-financed long-term care.
Each state will administer its own delivery and
financing systems. The federal government will contribute to a state’s system as long as several requirements are met, including coverage by and access to
at least one zero-premium plan that includes a standard benefits package for nearly all legal residents.
Beyond that, states will have wide latitude in crafting their own particular systems. For example, a
state can choose to establish a “Canadian-style” single-payer system, in which hospitals are paid based
on a negotiated budget and physicians on a fee-forservice basis (presumably with an aggregate expenditure cap). Alternatively, and perhaps more likely, a
state can contract with health plans that compete
for enrollees through low premiums, high quality,
and/or service, with individuals who choose more
expensive plans paying for them with ad d i ti on a l
premiums.
All health insurance ch oi ces offered must
include the services specified in a federally defined
benefits package that states may choose to augment.
Services to be included are inpatient and outpatient
care, skilled nursing and home health care, mental
health care, preventive services, prescription drugs,
and durable medical equipment. To receive coverage, individuals and families will enroll with their
states. They only need to provide evidence that they
are legal residents of the state; no income verification is required. People living in states that choose to
contract with private health plans also must choose
a plan. Those failing to do so will be placed in a plan
requiring no premium payment.
The plan will be financed through a variety of
sources. The primary revenue source, a payroll tax
levied on both employers and employees, will be
supplemented by general federal revenues,state revenues, and, potentially, by individual contributions
for certain plans and/or benefits beyond those
included in the standard benefits pack a ge . Employers with predominantly low-wage workers will
pay a lower tax rate than other employers. The proposal provides a number of assurances that important con s ti tu ency groups, such as states, larger
employers, small employers, and families, will pay
no more, on avera ge , than they currently do for
their health insurance coverage.
The proposal will help en su re that the United
States meets health care access, cost, and quality
goals. All Americans under age  who live in the
country legally will be provided with health insurance that includes coverage for all major health care
services except long-term care. Those eligible will
have at least one health insurance option that does
not require them to pay any premiums. Thus, neither financial considera ti ons nor burdensome
application processes will present barriers to obtaining health insurance, thereby ensuring that the goal
122
of universal coverage is met.
Costs are likely to be contained because nearly
all participants in the health care marketplace will
have an incentive to control them.States will receive
an annual fixed-dollar contribution from the federal government; thus they are at risk for additional
spending and will have a strong incentive to spend
wisely. If they rely on competing health plans to
provide insurance packages,these plans will vie with
each other for enrollees, recognizing the strong role
that premiums play in consumer health plan choice.
The plans, in turn, will continue to pay providers in
ways designed to reduce excessive and unnecessary
use of services.
A major ch a ll en ge will be ensuring that goodquality care is provided. Markets, by themselves,
cannot ensure this, particularly in light of the difficulties consumers have in obtaining and evaluating
the necessary information. The federal government
will need to invest substantial resources in measuring health care quality and health outcomes across
states, with a sp ecial emphasis on vulnerable populations. As part of the proposal, the federal government will closely monitor the quality of the care
provided under each state system, and provide
financial rewards to states that improve quality.
A second challenge is that states will be taking
over a number of responsibilities for administering
health insurance that previously were carried out by
employers. Depending on how a state chooses to
organize, these responsibilities may include, for
example, outreach, enrollment, nego ti a ti ons with
health plans and providers, data compilation, and
quality assu ra n ce . States will need technical and
financial assistance to carry out their enhanced
roles.
Detailed Description of Proposal
Federal Requirements
To receive federal contributions, states must
demonstrate that they have met and continue to
meet specific criteria. There are two aspects to this:
having their initial program approved,and continuing to meet coverage and quality requirements. This
section addresses the initial requirements. Ongoing
requirements are addressed later. To obtain approval
for their initial plans, states must submit their proposal to a federal agency.1 The proposal will have to
demonstrate the following:2
• Nearly all (at least  percent) of the legal residential population3 in the state will be enrolled.
• All localities will offer at least one plan that
does not require any premiums, and that, like all
plans, includes the federally mandated benefits
package.
• New state residents will be covered in a timely
fashion.
• A plan is in place to monitor and en su re the
quality of services provided to all state residents.
• There is also a plan in place for collecting and
compiling the data necessary to evaluate the system;
this is likely to include information on enrollment,
utilization (perhaps through encounter data), costs,
quality, and satisfaction.
• The state has a reasonable plan for carrying out
all of the necessary activities to implement its proposed system.
States are not required to implement universal
coverage; such approval is necessary only if the state
wishes to obtain federal contributions. But given the
fact that the vast majority of revenues will come
from the federal government through the payroll
tax and general revenues,it is anticipated that states
will be anxious to obtain timely approval of t h ei r
proposals. Nevertheless, a state may choose not to
participate, in which case it forfeits the federal
money that would have been available.4
We assume that this would be an executive branch agency, but we do
not take a position on which one; it could be the Health Care Financing
Administration, another agency in the Department of Health and
Human Services, or a newly established agency.
1
2
This is not unlike the system established in Canada. Under the 1984
Canada Health Act, provinces must demonstrate that the following five
provisions are met to receive federal health contributions: public administration, comprehensiveness, universality, portability, and accessibility.
The residential population excludes prisoners, those residing in institutions, and the homeless. Although states are encouraged to enroll individuals in these populations, we recognize that this will be a difficult
task. Some safety net system of care will still be necessary to care for
some of these individuals, as well as those who do not reside legally in
the United States.
3
A state that chooses not to participate could continue to receive federal matching money for the Medicaid and State Children’s Health
Insurance Program (S-CHIP).
4
123
Eligibility
All legal residents of a state are eligible for the state’s
program.5 To sign up, they simply have to demonstrate that they live in the state and are residing legally in the United States. No intrusive inform a ti on ,
including income, will be solicited. One exception,
discussed below, is if states chose to supplement federally mandated benefits with additional ones based
on such criteria as low income.
Each state will be responsible for en ro lling its
residents in a particular health insurance plan.
Under a system such as single-payer, there probably
will be a single plan for which everyone in the state
is eligible. Services could be received from any certified provider simply by showing one’s eligibility
card. Under systems that contract with private
plans, individuals will ne ed to enroll in a particular
plan. This can be done when a person registers for
coverage or, alternatively, within a designated period
of time after registration.
Benefits will be fully portable. A person who is
visiting another state will be eligible to be reimbu rs ed for urgent and em er gency services during
that visit, and those moving to another state will also
be covered during the period in which their eligibility is being transferred from one state to another.
Benefits
There will be a federally determined standardized
benefits package to which all states, and all plans
within states, must conform. We propose that this
include such elements as medically necessary acute
inpatient care, outpatient care, acute nursing and
home health care, mental health care (in parity with
physical health care), preventive services, prescription dru gs , and du ra ble medical equipment.
Specifically excluded are dental care and long-term
5
It would be desirable to grant eligibility to all state residents, legal or
not, but we have not proposed this because of the political difficulties
inherent in granting full benefits to those living in the country illegally.
Nevertheless, there are several reasons to cover them: (a) they have
health needs like everyone else; (b) their inclusion will lead to true universality, removing the need for a safety net system that could be quite
expensive in areas with many such residents; and (c) their inclusion could
improve social cohesiveness. As an alternative to covering these individuals, we suggest that the federal government make direct payments to
states with significant concentrations of undocumented persons, with
the stipulation that these monies be used by the states to support those
providers that serve this population.
care (both of which states may choose to provide to
certain residents). In general, the benefits package
should be similar to that currently provided by most
large employers.
In those states relying on competing health
plans, determination of medical necessity will be
carried out by the plans initially. As noted below,
however, states will oversee the quality of care provided by plans.A second “check and balance” is that
the federal government will monitor and, when necessary, regulate the quality of care provided in the
state—including whether plans in the state are
defining medical necessity too narrowly.
As experimental therapies and technologies are
developed, we foresee a multi-level decision-making
process. In states that contract with competing
health plans, decisions about whether and when to
provide the new therapy will likely be made by the
health plans initially, although certainly in consultation with the state. The state will have the option of
being more prescriptive and requiring each plan to
provide covera ge . In states that make direct payments to doctors and hospitals, the state certainly
needs to be directly involved in deciding whether to
pay for new therapies. The federal government will
have the option of intervening and being more prescriptive if it does not like the choices made collectively by states and health plans.
The exact set of benefits to be covered will be
established by Congress. Of major importance will
be how these are to be updated over time. We suggest
that recommendations for updating the benefits
package be made by the Medicare Payment Advisory
Commission (or a new agency) for Congress,and by
the Health Care Financing Administration (or a new
agency) for the President.
We anticipate that states may wish to provide
more benefits than those included in the standardized package—for example, Title V services to children with special health care needs. As with the
status quo, states could establish their own eligibility criteria for such additional services, with funding
largely provided by the federal government.
One key issue concerns patient cost-sharing
requirements under the standardized benefits package. Some advocate substantial cost sharing to make
124
people think twice before using services, while others believe that this is inequitable and will reduce
the use of necessary services. Our proposal follows
the lead established by the great majority of developed countries: they have instituted low or no
copayment requirements on basic services such as
inpatient, outpatient, and preventive care. Low
copayments also follow the precedent established in
most employer-sponsored insurance.
Cost-sharing requirements should be kept low
for several reasons:
• Cost sharing represents a larger share of income for those individuals with lower incomes,thus
it is regressive.
• Similarly, low-income people are, on average,
in poorer health, so requiring them to pay a good
deal for services is doubly inequitable.
• There is evidence to indicate that the health of
low-income families is adversely affected by cost
sharing.6
• There is also evidence that cost sharing is indiscriminate in that it reduces the use of necessary and
unnecessary services.7
• Low copayments will obviate demand for supplemental insurance that covers these copayments.
Such policies may be viewed as undesirable by states
because, by increasing utilization, program costs
will rise, as well.
• Although it is clear that an individual facing
substantial copayments will use less health care than
an individual for whom care is nearly free, it is far
from clear that a health care system in which many
people face copayments (but are fully insured for
care beyond some catastrophic limit) will provide
less care than a system in which copayments are low.
In the absence of supply side constraints, if
providers are paid mo re for delivering more servic-
6
R. H. Brook et al. “Does Free Care Improve Adults’ Health?” New England Journal of Medicine 309: 1426–34; M. F. Shapiro. “Effects of Cost
Sharing on Seeking Care for Serious and Minor Symptoms.” Annals of
Internal Medicine 104: 246–51; R. O. Valdez. The Effects of Cost Sharing
on the Health of Children, Santa Monica, CA: RAND Corp., 1986.
K. N. Lohr, “Effect of Cost Sharing on Use of Medically Effective and
Less Effective Care.” Medical Care 24 (supplement): S31–S38.
7
Under the proposal, acute care services cur rently paid for by Medicaid
will be folded into the standard benefits package. Average unit payment rates to providers are likely to be higher under the standard
8
es, then they are likely to respond to lower levels of
consumer demand by adjusting standards of care.
We propose that copayments under the standard
benefits package be similar to those in policies currently offered by typical large employers. Even these
relatively low copayments will disadvantage those
who are currently covered by Medicaid and, therefore,now have no copayments.States may choose to
provide supplemental coverage for low-income persons; this supplemental coverage might cover copayments as well as other services not included in the
standard benefits package, such as dental care.
Financing
The system will be financed according to the following principles:
    
 ,  
The federal government will estimate the cost of
providing the standard benefits package to all legal
residents in each state. First, it will estimate current
health expenditures in each state using data from
the Na ti onal Health Accounts, su pp l em en ted by
survey data on employer health insurance costs and
health maintenance organization (HMO) premiums. The estimate of current expenditures will be
adjusted upward for expected utilization increases
from insuring the uninsured (and underinsured).8 It
will be adjusted downward for expected efficiencies
that will result from universal coverage.
 ’  
State governments will be expected to contribute 
percent of their current Medicaid and S-CHIP
spending on services that are included in the standard benefits package for covered populations (that
benefits package in many states than the rates currently paid by
Medicaid; however, we do not suggest adjusting current expenditure
levels upward to account for the difference. Rather, we suggest assuming that average payment rates for non-Medicaid services will decline
slightly, resulting in overall average unit payment rates (combining
Medicaid and non-Medicaid) that will be similar after universal coverage
as they are under the status quo. To the extent that there is strong evidence that utilization rates of cur rent Medicaid recipients will change
when they are covered by so-called mainstream plans, modelers might
want to assume some utilization response; however, we are doubtful
that such evidence exists.
125
All legal residents of a state are eligible for the state’s
program. To sign up, they simply have to demonstrate
that they live in the state and are residing legally in
the United States.
is, state Medicaid spending on dental services,longterm care, undocumented persons, and the elderly
would not be included in this amount).
 ’ 

The federal government will pay to state governments that operate a qualified plan an amount equal
to the estimated cost of providing the standard benefits package to all residents (estimated as discussed
above) minus the state government financial responsibility amount. The state government financial
responsibility is not a “maintenance of effort”
requirement. If states can provide the standard benefits package to all legal residents at lower-thanexpected cost, the state contribution will be smaller
than expected. Alternatively, if more money is needed, it is state governments’ responsibility to raise a
portion of the necessary revenues.
If we could be confident that the state-specific
estimates of expected expenditures after universal
coverage would closely approximate the expenditures needed to maintain the status quo in the health
care system ,t h en we would be comfortable placing
financial responsibility for the marginal health care
dollar entirely in the hands of state governments.
State governments will be making important decisions about health care financing, and they should
be accountable for the outcomes of these decisions.
However, uncertainties in the accuracy of the
state health accounts data and the ad ju s tm en t s
needed to move from the status quo to universal
coverage create uncertainty about the precision of
the estimates of state-level post-universal coverage
expenditures. As a result, it is possible that expenditures in a given state will be substantially greater
than expected expenditures, not because the state
has chosen to be generous in its payments to
providers, but rather because the expected expendi-
tures were an underestimate of the expenditures
needed to maintain the status quo in the state’s
health care system.
If a state’s expenditure for services included in
the standard benefits package differs substantially
from the expected state government financial
responsibility amount, we suggest that the federal
and state governments share in the su rp lus or
deficit.A reasonable approach would be to make the
state fully responsible for spending that is  percent
more or less than the expected state government
responsibility; have the state and federal governments share the next  percent surplus or deficit
-, and have the federal government be responsible for  percent of the su rp lus or deficit if state
spending is more than  percent of the expected
amount, or less than  percent of the expected
amount. For example, if a state is expected to spend
 billion but actually spends . billion, the federal
government would increase its contribution to the
state by  million—that is,  percent of the difference between actual state spending and  percent of expected spending.
Inevitably, there will be disagreements between
the federal and state governments about which state
expenditures should be included in this calculation.9
In the long term, a system in which the federal contribution is fixed and states are financially responsible for their decisions is preferable to a system that
encourages disputes about matching payments.
However, in the short term some sharing of financial responsibility for deviations from projections is
sensible. We suggest below that Congress appoint a
commission to make recommendations on realigning federal contributions across states; the work of
9
Witness the disagreements in the Medicaid program about disproportionate share hospital payments, and, more recently, about
payments to nursing homes and hospitals under the Upper Payment
Limit regulations.
126
this commission would facilitate the transition to
full state financial responsibility.
 
The federal government will raise most of the
money needed to finance its contribution to state
programs through a payroll tax on employers and
employees. The amount of money raised by the payroll tax will be equal to  percent of the total
amount currently spent by employers and employees for health insurance for covered benefits (that is,
it would not include amounts spent for dental care
or other services not included in the standard benefits package). This amount would include expenditures for non-group coverage and out-of-pocket
payments for health care, to the extent that payments for these services are expected to be included
in the standard benefits package. The tax rate will be
uniform throughout the country.
Following the model of the Medicare payroll tax,
we suggest that the health care payroll tax be applied
to all wages; however, if there is a desire to limit the
progressivity of the financing system, the payroll tax
could be imposed only on wages up to the Social
Security wage base.
In a full proposal,a payroll tax rate will be specified. Rather than providing our own back-of-theenvelope estimate of what this tax rate will be, we
leave this to the modelers. For purposes of discussion below, we assume that the total payroll tax will
be  percent—that is, we assume that an  percent
tax will generate an amount equal to  percent of
what is currently spent by employers and employees
(including non-group coverage and out-of-pocket
payments that would be covered).
The total payroll tax will be divided between
employers and employees based on the current distribution of spending between employers and
employees. If, for example, the modelers estimate
that  percent of current spending is done by
employers, and if the total tax is  percent, then
employers would pay a  percent tax,and employees
would pay a  percent tax. We propose that the
employee portion of the payroll tax be treated as
pre-tax income, as much of employee spending for
health insurance is now.
To increase the progressivity of the financing
system and minimize negative effects on the level of
employment, we propose a lower tax rate for
employers with predominantly low-wage workers.
For example, employers whose average employee
makes less than  per hour would pay  percent of
covered payroll rather than  percent of covered
payro ll . Self-employed persons will pay both the
employer and employee portions.
   
The difference between the amount of money the
federal government is obligated to pay to the states
and the amount raised by the payroll tax raised will
be financed by general revenues. We expect that
much of the general revenue obligation will be a
transfer from federal Medicaid funds currently used
to support Medicaid and S-CHIP. However, some of
the general revenue obligation will require new federal funds, presumably drawn from the anticipated
budget surplus.
    
The federal payment to the states will grow at a rate
to be specified annually by Congress, after receiving
a recommendation from the President and the
Medicare Payment Advisory Commission (MedPAC) or a new advisory agency to be established. In
making its recommendation, the advisory agency
will consider factors similar to those that MedPAC
considers in recommending payment updates for
Medicare. In addition, the agency must consider the
effects of any changes in the standard benefits package. To provide some protection to state governors
and health care providers, the five-year growth rate
of per capita federal payments to the states will not
be less than the five-year growth rate of per capita
Medicare expenditures, unless compelling rationales
are advanced that health care needs are increasing
much more quickly in the over- population than
in the under- population.10
10
The rationale is that if Congress has the political will to get tough
with providers in the Medicare program, then it is reasonable to ask
governors to be similarly hard-nosed; alternatively, if Congress decides
that Medicare needs additional funds to provide high-quality care to
seniors, it is not reasonable to ask governors to be significantly more
frugal for their states’ under-65 population.
127
   
  
The proposed financing system is intended to leave
the status quo largely intact for health care
providers—revenues to the health care financing
system in each state after universal coverage is
implemented should be similar to current revenues
(adjusted for the expected effects of universal coverage, as discussed above). In addition, state government payments are intended to be similar to the
current system—with a  percent savings in each
state to encourage governors and state legislatures to
support the proposal. As a result, there will be wide
disparities across states in the amount of federal
support; in states that have high per capita health
care costs, federal support will be greater than in
states with low per capita cost health care costs.
This is likely to be seen as unfair by those people
living in states that historically have spent relatively
little on health care. We suggest that Congress
appoint a commission to make recommendations
on long-term realignment of federal contribution
levels.11
    -
Federal government payments to the states will be
contingent on the states satisfying basic requirements. To the extent these requirements are not satisfied, graduated financial penalties will be applied.
For example,if a state does not provide the standard
benefits package to  percent of its legal residents,
the federal contribution is reduced—certainly by
the per capita contribution amount and, we propose, with an additional penalty. A state that fails to
meet quality standards or to provide required information will also be subject to financial sanctions.
11
A number of factors should be considered in thinking about equity.
Federal contributions might be compared to the amount of payroll tax
revenue coming from the state; one might focus only on the federal per
capita contribution from general revenue as the amount to be concerned with when considering fairness. Alternatively, if we think of current federal Medicaid spending as an entitlement to the states (or, at
least, as part of the status quo), then the federal contribution to focus
on might be simply the new general revenue payments (that is, treating
the displaced Medicaid funds, like the payroll tax, as belonging to the
states). One also might want to consider the size of the per capita state
contribution in considering equity, as well as the contribution of the
state to medical education, and, potentially, research.
   
To avoid a large windfall to shareholders of companies with retiree health obligations, a tax on corpora ti ons that show Federal Accounting Standards
Board Section  obligations on their balance
sheets will be enacted. The government would estimate the portion of a corporation’s Section  liability that would be assumed by the states under
reform. Employers would be assessed  percent of
this amount, and would pay a special “retiree health
assessment tax.” The total retiree health assessment
would be paid over  years.
Administration and Regulation
Once the new syst em has b een implemented, it will
be necessary to monitor its performance and make
necessary corrections. There is a trade-off, of course,
between accountability and state autonomy. On the
one hand, consumer protection is especially critical
in health care, where much is at stake and consumers face severe information problems. On the
other hand, for a proposal like this to be successful,
states need to tailor a system that best fits their particular circumstances.
The main concern is quality—whether the average quality is sufficiently high and whether disadvantaged populations face particular quality
barriers. Most parties will have strong incentives to
control costs, which obviously raises issues about
sacrificing quality. These concerns are discussed in
the following section.
As noted, the federal bureaucracy envisioned is
minimal. The executive branch will be advised by
the Health Care Financing Administration or a new
agency, and Congress by the Medicare Payment
Advisory Commission or a new agency. Once the
system has been implemented, some of the ongoing
duties of these two agencies will include making recommendations on:
• updating the benefits package;
• revising the formula for allotting contributions
to each of the states; and
• deciding how much federal contributions will
increase over time.
128
Quality Assurance
The major concern with the proposal is ensuring
that high-quality services are provided. There are
two somewhat independent aspects to this. First,
plans—particularly in states that emphasize price
competition—are likely to be under strong pressure
to keep costs down, potentially threatening quality.
Second, even if plans do attempt to provide goodquality care, they may lack the tools to do so.
Regulatory oversight is necessary to address the
first issue, because in markets where full consumer
information is probl em a ti c , there is no assurance
that the services provided will be of the quality
sought by consumers. Fu rt h ermore, some states
may not meet their responsibilities to ensure the
provision of good-quality care. This is particularly a
concern for services delivered to vulnerable groups
such as racial and ethnic minorities and the poor. As
a result, quality assurance is the only major area of
the proposal in which strong federal regulation is
called for.
We propose that a quasi-independent, bipartisan federal quality commission be established to
monitor the quality of health care delivered in each
of the states, and to provide financial incentives to
states to improve quality. (Quality will be defined as
encompassing access and quality.) The commission
needs to be given a sufficient budget for staffing
adequately and collecting the data necessary to fulfill its mission, which will include assessment of:
• whether nearly all individuals have health
insurance;
• whether the most vulnerable individuals face
any unduly significant barriers to accessing necessary preventive and acute care services;
• the technical quality of care provided to
patients through assessments of appropriate
processes and desirable health outcomes;
• patient satisfaction; and
• state-provided data to evaluate access, costs,
and quality. As noted earlier, these data would likely
include information on en ro ll m en t , utilization
(perhaps through encounter data), costs, quality,
and satisfaction.
In addition to measuring quality, we propose
that the quality commission be empowered to
financially reward states that perform the best in
improving quality of care. A portion of the federal
payments to states would be deducted from the
direct federal payments and provided to the quality
commission to distribute as a “quality bonus” to
states that improve their performance. The bonus
pool might begin at . percent of federal support,
and grow to  percent after  years. The proposed
growth in the size of the pool reflects anticipated
improvements in our ability to measure meaningful
aspects of quality. The movement of funds to the
quality improvement pool should lag the implementation of universal coverage—perhaps by three
to four years, to give the quality commission time to
implement measurement tools and establish baseline performance levels.
Of course, some aspects of a health system (for
example, adequately trained physicians and other
personnel, well-staffed and equipped hospitals)
are critical to the provision of quality care. These
structural aspects, however, are left to the states to
regulate.
The second issue is that many states, health
plans, providers, and consumers lack the necessary
tools to ensure adequate quality of care. The “freerider effect” will result in too small an investment in
the necessary data and research on quality. States
will want to take advantage of investments made by
other states, and health plans of investments made
by other health plans. To illustrate, suppose that one
health plan is considering investing in a system that
evaluates the quality of primary physician care,
feeding back information to providers so they can
improve. In most cases, however, providers are
enrolled in many managed care plans. If one plan
invests in improving quality, the other plans in
which the provider participates will gain without
making a concomitant investment. As a result, all
plans will underinvest in quality improvement.
The best way to deal with this problem is to treat
research on quality as a public good and have the
federal government invest in it. This is done in clinical care through the National Institutes of Health
and, to some extent, in health services research. But
the amount of investment in the latter is much too
low. In , the United States spent . trillion on
129
health care.12 In contrast, the budget of the Agency
for Healthcare Research and Quality (AHRQ), the
federal agency mainly responsible for the accumulation and dissemination of this knowledge, was only
about  million—a mere . percent of health
care dollars spent.13 We believe that this figure should
be at least  percent of total health spending,
although this enhanced funding should be phased in.
Some of the actions to be emphasized in this
research and dissemination effort should include:
• reducing medical errors;
• ensuring that providers in all medical specialties use state-of-the art technologies;
• enhanced measurement of health care processes and outcomes;
• focusing on the health status of disadvantaged
population groups; and
• creating the necessary databases to accomplish
these tasks.
Integration with the Current Health Care
Financing System

The Medicaid program for long-term care services
would remain. For aged or disabled persons who
meet Medicaid financial eligibility standards, reimbursement for institutional and home and community-based long-term care would continue, with the
current federal formula for matching payments.
Two areas deserve special consideration—a
va ri ety of services outside the standard benefits
package that are currently provided under Medicaid
to children, and non-long-term care payments currently made for those who are dually eligible for
Medicaid and Medicare (for example, prescription
d ru gs , copayments, deducti bl e s , Part B prem iu m
payments.)
For children, there are a va ri ety of services—
dental care is a prominent example—currently
provided by Medicaid that are not part of the standard benefits package. There are at least two options
here: first, maintain Medicaid financing for dental
www.cdc.gov /nchs /products /pubs /pubd/hus /00tables.htm#
National Health Expenditures
12
13
www.ahrq.gov/about /profile.htm
services under current Medicaid rules, or, second,
provide funds under a block grant to the states, with
the stipulation that the funds be used to support
health care for low-income children. While maintaining the Medicaid entitlement and open-ended
matching funds may sound attractive to advocates,
we doubt that it is a good idea. First, some services—such as dental care—are currently optional for
states under Medicaid. More important, in most
states access to dental care is extremely problematic.
Even under the current system, most states have
done a very poor job in providing dental services to
low-income children.
There are other federal programs—such as Title
V (providing funds for services for children with
special health care needs), Section  funds for
community health centers, support for community
mental health centers, and support for rural and
migrant health centers—that should continue, even
under universal covera ge . Eventually the need for
these funds might be reassessed, but this is not
included in the proposal. Similarly, although on a
much larger scale, the Department of Veterans
Affairs health care system might eventually change
in response to an environment of universal coverage
for the standard benefits package, but changes to the
VA are not part of the proposal.
Medicaid currently pays for prescription drugs,
copayments, deductibles,and Part B p remiums for a
variety of low-income Medicare recipients. Under
this proposal, the federal government will administer this Medigap-like coverage directly, rather than
relying on the states to fill gaps in a federally run
program. Ninety percent of the financial burden
lifted from the states by federal assumption of current Medicaid responsibilities would be added to
states’ financial responsibility.

The standard benefits package for persons under
age  will be richer and deeper than the benefits
currently available under Medicare. This will add to
the already existing pressure to improve the
Medicare benefits package, but changes to Medicare
are not included as part of this proposal.
130
We have chosen to rely on states to make major decisions
about health care financing and delivery because
we think they are the appropriate locus of financial
and political decision making.
  
Protecting safety net providers would be largely a
state responsibility. As discussed above, a variety of
Health Resources and Services Administra ti on
(HRSA) programs in support of providers such as
community health centers and rural and migrant
health centers would be expected to continue.
However, since the Medicaid acute care program
would be folded into the universal coverage system,
Medicaid disproporti on a te share hospital (DSH)
payments would end. Consistent with the federal
requirement to provide the standard benefits package to all state residents, state governments could
choose to con ti nue direct payments to hospitals
currently receiving DSH funds; however, these payments would compete directly with money that
might be used to pay for health insurance premiums
(in a system with competing health plans) or to support other providers (in a single-payer system).
These are problems for the states to work out,
although state plans submitted for approval to the
federal government must demonstrate that the
needs of traditional community providers have
been taken into account in designing the state
financing system.
The Central Role for States
As the title of this paper suggests, this is a “state-based
proposal.” This section discusses some of the advantages and disadvantages of a state-based proposal,
and indicates why we have chosen this direction.
We have chosen to rely on states to make major
decisions about health care financing and delivery
because we think they are the appropriate locus of
financial and political decision making. We see three
major substantive problems with a system run solely by the federal government.First,it will be difficult
for the federal government to do a good job of figuring out, in total, how much money should go into
health care. Providers will always argue for more
and complain that quality will suffer without more
resources. Some providers will go out of business.
Decision makers in Washington are so far away from
local conditions that it will be difficult for them to
determine whether the providers are correct and
how much to be concerned that some providers may
go out of business. Imagine, for example, that
Medicare was expanded to cover the entire population. Suppose now that some hospitals or physicians
in Boston or St. Louis have negative margins and
face the prospect of c ut ting back services or, perhaps, closing their doors entirely. Alternatively, these
hospitals might argue that they are unable to adopt
quality-enhancing technology because of inadequate reimbursement. The providers will certainly
appeal to the federal government for help. How are
decision makers in Washington going to determine
whether such help is needed? It is difficult to imagine that they will do a good job. We have concerns
about the ability of state governments to make good
decisions here, as well, but there is greater financial
and political accountability with decisions made
closer to the bedside (and closer to patients and
providers and taxpayers!).
Second, the federal government, under intense
scrutiny and pressure, has a very difficult time
experimenting with new forms of finance and delivery, while states are able to move somewhat more
nimbly. For example, the federal government has
tried unsuccessfully a number of times to implement a competitive bidding demonstra ti on for
Medicare. In contrast, state governments, in purchasing benefits for both state employees and
Medicaid beneficiaries, have been able to adopt
innovative purchasing practices with much less
political resistance. Federal government control is
likely to lead to a system that is less innovative and
flexible over time than a system in which state gov-
131
ernments have major responsibility for financing
and delivery decisions. Related to this point, state
diversity and experimentation will allow for the dissemination of best practices.
Third, we continue, collectively, to disagree about
whether “competitive” or “regulatory ”a pproaches to
health care financing will produce the desired combination of quality and economy. These fights are
unlikely to be resolved any time soon, and preferences are likely to differ across states. A state-based
approach allows for diversity, while a federal
approach would be likely to impose a uniform solution. (One could imagine a federal approach with
substantial waiver authority to allow state diversity,
but this is likely to be difficult to implement if state
governments do not have financial responsibility, as
well.)
We have justified our choice of state rather than
federal control of major health care financing decisions. In the concluding secti on , we present our
ra ti onale for proposing increased public-sector
involvement in health care financing and eliminating the role of employers as middlemen in our
financing system.
In this proposal, states will have substantially
expanded roles and responsibilities, including:
• choosing the financing and service delivery systems used (for example, single-payer, competing
private health plans) to provide health insurance
coverage to nearly all state residents;
• providing information on enrollment options
and procedures to all state residents;
• carrying out the enrollment process, which will
often entail having all residents choose a health
plan;
• negotiating with health plans and providers
over services, payments, and quality;
• ensuring that the necessary data are collected,
both for estimating the cost of the system and for
evaluating its performance;
• ensuring that insurers and providers do their
part to provide good-quality care; and
• sharing with the federal government in the
financial risk associated with financing health care
services for the population.
This reliance on states undoubtedly is contro-
versial; it was not chosen as the basis of our proposal lightly. There are two major concerns about placing such key responsibilities with the states. First,
until now they have not had the main responsibility
for administering health insurance for their populations. This lack of experience will be costly if they
are not able to carry out the many roles listed above.
Second, it can be argued that some states have not
been as “trustworthy” as others in implementing
their primary health insurance program, Medicaid,
and that some states have done a poor job of protecting disadvantaged groups such as racial minorities. In the case of Medicaid, some states have
extremely stringent eligibility criteria and pay
providers so poorly that few want to treat patients in
the program.
With respect to the first problem—lack of experience—there is no doubt that most states will need
both technical and financial assistance to carry our
their enhanced roles. Congress needs to allot sufficient funding for these activities. Nevertheless, one
should not be too pessimistic about states’ abilities to
carry out these roles because they are already have a
great deal of experience in regulating private insurance and HMOs; enrolling state residents in programs
such as Medicaid and S-CHIP; contracting with
health plans and third-party administrators as part of
Medicaid; and operating public health systems.
The second problem is more worrisome: some
states have done a poor job of providing health
insurance coverage in the past, and these problems
could be magnified under the proposal, which provides strong financial incentives to states for controlling costs. Aggressive cost control, in turn, could
harm quality.
Several aspects of the proposal address these
potential problems. First, all non-Medicare-eligible
state residents will be part of the system. State officials whose systems provide poor-quality or inefficient care will be under strong political pressure to
improve them. Second, states will find it difficult to
skimp on providing adequate access because this is a
federal government requirement. Those states that
do not enroll at least  percent of their population
will face severe financial penalties. Third, only those
state plans that are initially approved by the federal
132
government will become operational. Fourth, there
is a strong federal quality monitoring system built
into the proposal (with sp ecial emphasis on quality
and outcomes for vulnerable groups), coupled with
financial incentives that will reward states that offer
good-quality care and penalize those that do not.
Transition
The following steps need to be followed in implementing the proposal. We also provide a suggested
timetable for the transition from our current
hodgepodge to the reformed system.
. Congress passes a law enacting the reforms on
July , .
. The federal agency responsibility for oversight
of the new system, in consultation with governors
and other interested parties, develops guidelines for
the state plans.Guidelines are promulgated by July ,
. The federal government makes planning
grants to the states.
. States enact enabling legislation, and submit
state plans to the federal government no later than
December , . The federal government makes
funds available to the states for transitional assistance and provides technical assistance and training.
. On January , , the federal government
starts co ll ecting new payroll taxes. States with
approved state plans start receiving payments from
the federal government and providing coverage to
all under-, non-Medicare residents of the state.
Employers presumably stop providing health benefits to employees for services included in the standard benefits package.
If a state does not implement a plan, then
employers and employees will be paying a substantial payroll tax, but not receiving federally subsidized health benefits. This is likely to create
extremely strong pressure on state governments to
design and implement acceptable plans.
Impact of Proposal on Society’s Goals
Access
The proposal will help ensure that the United States
attains the three major policy goals of universal
access, cost control, and good quality of care. With
regard to access,all Americans who live in the country legally will, as a right, be provided with health
insurance that includes coverage for all major health
care services except long-term care. Everyone will
have at least one health insurance option that does
not require payment of any premiums. Thus, financial considera ti ons will no longer be a barrier to
obtaining health insurance, nor will stigma be an
issue because everyone will receive coverage
through the state. States will be given strong financial incentives from the federal government to
en su re that they meet coverage requirements.
Combined,these should ensure that the goal of universal coverage is met.
This is not to imply that access issues will be
solved. There are many barriers to access, and financial impediments are only one. In addition, to the
extent that there are some copayments under the
standardized benefits package, some individuals
currently enrolled in Medicaid may find themselves
worse off. Similarly, the benefits package excludes
various services covered by state Medicaid programs, which could also impede access. States may
wish, therefore, to consider providing additional
coverage to poor and vulnerable individuals to compensate for this shortfall.
Costs
Costs are likely to be contained because nearly all
participants in the health care marketplace will have
an incentive to control them. Because states will
receive an annual fixed-dollar contribution from the
federal government, they are at risk for additional
spending, so they have a strong incentive to spend
wisely. If they rely on competing health plans to
provide insurance packages,these plans will vie with
each other for enrollees, recognizing the strong role
that premiums play in consumer health plan choice.
The plans, in turn, will continue to pay providers in
ways designed to reduce excessive and unnecessary
use of services.
As in the case of access, cost control is hardly
guaranteed, nor should it be. In fact, as medical
therapies become more efficacious and the population ages, the United States may wish to spend more
on health care.
133
Some elements of the proposal will almost certainly increase health care costs. These include universal coverage and comprehensive benefits (the
latter will increase utilization for those who are currently underinsured).
Quality
The biggest ch a ll en ge will be en su ring that goodquality care is provided. We do not believe that markets, by themselves,can assure this,particularly with
the difficulties consumers have in obtaining and
evaluating the necessary inform a ti on . Nor can
states—some of which may be under tight fiscal
constraints—necessarily be depended on. If the federal government is going to turn over large amounts
of revenue to the states, it will need to invest substantial resources in measuring health care quality
and health outcomes across states, with special
emphasis on vulnerable populations. The types of
investments and oversight needed to accomplish
this goal were discussed above.
Political Feasibility
We are aware that our proposal represents a substantial departure from the status quo and will face
significant political opposition. Two main features
of the proposal are likely to create opposition: first,
transforming current voluntary employer and
employee payments into a mandatory payroll tax,
and, second, turning over the money and the
responsibility to state governments. We address each
of these concerns below.
There are two reasons to propose a payroll tax.
First and most important, we know of no way to get
to universal coverage, or anywhere close to it, without required contributions. The main alternatives to
the payroll tax for the required contributions are a
value-added tax or an increase in the income tax
(accompanied by an individual mandate). The payroll tax is closer to the way the United States currently finances health care and is arguably politically
more palatable than either of these alternatives.
Proposals for expanding subsidized coverage (for
example, expanding S-CHIP to parents in families
with incomes less than  percent of federal pover-
ty level) are not likely to do much to reduce the
number of uninsured. The S-CHIP-like expansion
proposals may be the best we can hope to accomplish, but we should be clear about how far they
leave us from universal coverage.
The second reason we propose public-sector
decisions about health care financing is that we
think employers have added relatively little value to
health care purchasing. Employers are in business to
make a product or provide a service,not to purchase
health care. Since most actions that employers
might take in an attempt to purchase better health
care will only be effective if they are joined by many
other employers, they face a significant co ll ective
action problem. While a few employer groups, such
as the Pacific Business Group on Health, have made
some progress in overcoming the collective action
probl em , progress has been limited, and effective
purchasing coalitions are the exception,not the rule.
Certainly the tribulations of Medicare and state
Medicaid programs give us pause, as well, but these
organizations have the potential, partially actualized, to be prudent intermediaries in health care
finance.
Some who support public financing and the
elimination of the employer role in purchasing care
will be upset that our proposal calls for state governments to set the rules of health care financing,
rather than simply expanding the federal Medicare
program to cover all Americans. Certainly Medicare
is an extremely popular program and has the advantage of an existing and concrete framework. Our
proposal envisions new state structures and suffers
from uncertainty—what exactly will health care
look like in my state after Congress passes universal
coverage legislation? The answer must await state
action. However, the difficulty of having a uniform
set of federal rules to respond flexibly to the needs of
the residents and providers in each state outweighs,
in our judgment, the potential advantages of simply
expanding Medicare.
We are aware that our approach relies on substantially more public-sector involvement in health
care financing and delivery than the country has
been comfortable with in the past. There is likely to
be concern about this proposal from a variety of
134
important constituencies—some employers with
relatively young and highly paid employees might
end up paying more in payroll tax than they now
pay for health insurance, as would employers who
currently do not offer health benefits at all. Some
providers and insurers will be nervous about concentrated purchasing power in the hands of state
government; some governors will be reluctant to
absorb substantial new responsibilities and the
political demands these responsibilities create; and
some citizens will be concerned about giving state
governments greater control over the financing and
delivery of health care. However, the proposal will
have appeal, as well: The left should find mandated
universal coverage with no required premiums
attractive, and the right should like the opportunity
for states to tailor their own health care system, and
the likelihood that many of these state systems will
embrace market-based approaches over administered pricing approaches. As we approach  million
uninsured (the  Current Population Survey
results showing a decline in the number of uninsured are most likely a short-term blip in a longerterm trend), rising costs (again) for employers and
employees, increasing dissatisfaction among physicians and hospitals, managed care backlash from
patients, and a poor outlook for profit growth for
health insurers reveal at least the potential that the
country might be ready to experiment with
approaches that are more revolutionary than evolutionary. Our proposal provides a sensible policy prescription that could be successfully implemented if
the political window opens more widely than its
customary narrow slit.
Acknowledgements
We would like to thank Harold Luft for helpful comments on a draft of this proposal. All opinions are
solely those of the authors and do not reflect those
of the reviewers, sponsor, funding agency, or our
respective universities. n
135
Pauly Proposal
Key Elements
Mark V. Pauly has outlined a new proposal to reduce the number of uninsured
that would:
                  that could
be easily modified in response to lessons learned from this new approach.
     -       -     with flexible, refundable
tax credits or “coupons,” redeemable for insurance premiums or a reduction in taxes, for any licensed medical-surgical insurance policy with a
premium at least as large as the credit.
   -     eligible for publicly provided or
contracted insurance, or for equivalent-cost private insurance, with no
premium share required.
          to retain the
tax exclusion for group coverage, until mandated to participate in the
new program at some point in the future.
       are renewable.
136
About the Author
 .  , .., is Bendheim Professor and Chair
of the Department of Health Care Systems, Professor
of Health Care Systems, Insurance and Risk Management and Public Policy and Management at the
Wharton School,and Professor of Economics in
the School of Arts and Sciences at the University of
Pennsylvania.Dr. Pauly served as commissioner on
the Physician Payment Review Commission,and he
is an active member of the Institute of Medicine.
One of the nation’s leading health economists,Dr.
Pauly has made significant contributions to the
fields of medical economics and health insurance.
His classic study on the economics of moral hazard
was the first to point out how health insurance coverage may affect patients’ use of medical services.
He has explored the impact of conventional insurance coverage on preventive care, on outpatient
care,and on prescription drug use in managed care,
as well as studying the influences that determine
whether insurance coverage is available and the
effect of insurance use on health outcomes and
cost.Dr. Pauly is co-editor-in-chief of the International Journal of Health Finance and Economics and
an associate editor of the Journal of Risk and
Uncertainty.
137
An Adaptive Credit Plan for Covering the Uninsured
by Mark V. Pauly
Introduction
To deal with the problem of large numbers of persons without health insurance, this paper outlines a
flexible and adaptive method of using refundable
tax credits, supplemented by a full publicly subsidized program for very-low-income households.
This flexibility and adaptability is necessary for two
reasons: () The availability of credits will transform
insurance markets in ways that eventually (but not
initially) will change how credits might be used.
Specifically, credits will cause some uninsured persons to seek coverage, and that influx of new customers may help to transform the markets they
enter. () There is substantial and (currently) irreducible uncertainty about two key aspects of this
system’s response to the availability of partial credits. How many will take up insurance for a given
credit policy, and how the markets in which they
will use their credits will be transformed,are subject
to enormous uncertainty that cannot be addressed
by better current data collection and/or analysis or
simulation. The reason for this uncertainty is that
there has been no experience with such a large-scale
system of credits or subsidies offered to the target
populations; we simply don’t know what will happen because we have never seen anything like it.
Attempts to extrapolate from other situations we
have observed (tax subsidies for the self-employed,
behavior of Medicaid recipients) can offer some
hints about direction and size of relative effects, but
they cannot substitute for actual experience.
Policy analysts typically deal with such behavioral uncertainty in federal programs in one of three
ways. One way is to acknowl ed ge the uncertainty
f ra n k ly, present ranges of values for possible out-
comes, with no pretense that some value in the
range is a “best” (or even “better”) guess than others, and then suggest a plan that deals with uncertainty and learns from its resoluti on . Such a plan
ordinarily will not be best for any one particular scenario for the unknown variables, but it will be good
on average for a wide range of possible scenarios. A
second approach is to design an interven ti on that
leaves it to the states to resolve some of the uncertainty, and counts on them to offer different
approaches. The third approach is to pick one possible behavioral response, declare it to be virtually
certain, and design a policy that fits it. This last
approach in policy for the uninsured leads to stalemate, because, for any policy that inspires optimism
in some, there will be others with worries and concerns who can block the proposal. Until now we
have done virtually nothing about the bulk of the
uninsured, because the outcomes are uncertain and
undesirable results cannot be ruled out.
The strategy here is to follow the first approach,
and to proceed in two stages or phases. We begin
with a relatively simple, financially feasible, easily
reversible or modifiable intervention,targeted at the
uninsured for whom it is most suitable. We embed a
s ch eme to learn from that intervention to alter
aspects of the program according to a fixed, transparent, and comfortable process. The first phase of
the plan also is intended to be easy to administer
and understand, and permissive and encouraging
rather than restrictive and intimidating in determining el i gi ble persons and qualified insurance
policies. The program initially targets one subset of
the uninsured population for credits. As a result,
this simple initial program is less than comprehensive and less than perfect. But it is a step forward,
138
and one that is better than the current situation. In
the second phase, the observed outcomes from that
intervention set the stage for generalizing the program, and inform a ti on from that interven ti on is
used to determine the best way to generalize. In this
process, change is an indication of learning, not
mistakes, and we do not have to wait until every
possible glitch is anticipated before moving ahead.
In what follows, we primarily provide details on
the Phase I proposal. The alterations to this structure
that might take place during Phase II are described
later.
Overview
In Phase I,the flexible credit plan divides the under U. S. population into three groups, each treated
differently. Lower-middle-income households,
which make up two-fifths of the uninsured, are the
primary initial target group. All families with
incomes above the poverty line but below the median (regardless of the age,sex, or relationship of family members) are made eligible for a voucher or
credit of a given amount, varying only by whether it
is used for individual or family coverage, that can be
used to purchase insurance. For budgetary and
political reasons,credits are likely to be less than the
full premium for a comprehensive insurance policy
for all eligible persons. So to maximize the use of the
subsidy, few restrictions are placed on the type of
insurance for which it can be used, or on the cost of
that insurance, but a publicly provided or contracted fallback insurance plan will offer policies
financed with the same subsidy on the same terms
as private plans. The subsidies take the form of
refundable credits or “coupons” redeemable against
either all or part of the insu ra n ce premium, or
redeemable as a tax reduction on presentation of
proof of insurance purchase. A key assumption is
that an influx of new buyers will improve the functioning of private insurance markets substantially,
especially individual markets.
In addition, in Phase I, very-low-income households will become eligible for publicly provided or
contracted comprehensive insurance, with no premium share required. People with incomes above
the median (with a few exceptions for high risks
and, possibly, those with incomes near the threshold) will not be eligible for the new program initially, but may con ti nue to use the new individual
insurance plans, and may retain the tax exclusion
for group coverage.
The most important behavior to monitor in
planning adjustments for Phase II is that of the
emerging private insurance markets, both individual
and group other than employment. If these markets
are functioning reasonably well, the contribution
toward insurance for very-low-income households
will be converted in Phase II into a voucher or credit
that these households also can use for private insurance, with coverage similar to that of the Medicaidtype plan. If many of the uninsured with incomes
above the median also use the new private market,it
should be possible to implement a mandate requiring the remaining tiny minority to buy some coverage,and to cap the value of the exclusion,thus turning the exclusion into an adjustment of the income
tax base for almost everyone. Finally, the coverage
that can be obtained with the credit or coupon alone
to buy a “low-priced” plan will be examined to determine if there are any substantial health benefits to
lower-middle-income people from adding coverage
(at added budget cost). If such benefits do exist,
stricter minimum standards and financing to pay
their additional cost would be added in Phase II.
There are two novel features in this plan. First,
using coupons as a vehicle for credits should
improve the take-up rate greatly. Coupons are
attractive, solve cash flow problems, and are easy to
administer. Second, permission to use the credit for
insurance plans with premiums no greater than the
credit ensures that virtually everyone in the target
group will end up with at least some coverage. There
will be universal coverage for everyone, though not
for all expenses; non-poor people should be allowed
to use cost sharing. Public financial constraints and
efficient cost containment mean that such coverage
will not fully cover all medical services. There will,
however, be some public subsidy for insurance for
people who currently receive no subsidy toward any
insurance.
To deal with valuation in risk across households
139
These new credit vouchers or coupons should be
thought of as tax reductions for the lower middle class
and, thus, treated as tax reductions for all who use them.
and over time, all policies will be guaranteed to be
renewable (at premiums that are no higher than
those charged to average risks). In addition, plans
requiring no additional premiums must be rated on
an adjusted community rating basis, but premiums
for coverage beyond this level may be risk rated if
insurers wish to do so. A high-risk pool is an option
in Phase II if many high-risk individuals are still
paying very high prem iu m s , but this outcome is
unlikely (for reasons discussed below). The public
alternative plan(s) will be regulated, rated, and subsidized on the same basis as private insurance.States
may continue to regulate individual insurance, and
may add additional subsidies to the credit (in various ways), but the states will be responsible if regulation results in fewer people being uninsured than
expected.
Objectives, Assumptions, and Rationale
To design a system that adapts to different outcomes, we need to set priorities among these different outcomes. The two most important of these are
fairness and efficiency. The problems associated
with these are: everyone is for fairness but defines it
differently, and there is only one definition of efficiency but not everyone is for it.
An efficient outcome is one that matches consumers with the insurance policy they prefer (given
the subsidy) and whose costs and premiums are
minimized. With respect to the effect of the subsidies, the objectives and goals of the providers of
those subsidies—taxpayers—are important. One
would usually include as social objectives improved
health status, control of infectious diseases, and
l on ger life expectancy, but goals dealing with the
dignity, privacy, convenience, and satisfaction with
care and coverage may also matter. We assume that
the key to efficiency is to offer neutral incentives to
both low- and middle-income people for choosing
among insurance plans, and to offer targeted subsidies for those poor enough to need help.
A fair outcome is one in which people who are
similar in income and health status are treated the
same, but the amount lower-income families must
spend on care and insurance should be limited.
How limited, and how low is “low-income,” are
political decisions.
We do not enshrine or condemn particular ways
of providing insurance or even particular types of
insurance (as long as they do not lead to worse
health outcomes). Specifically, there is no intrinsic
merit to employment-based group insurance or
insurance provided by the government, non-profit
firms, or for-profit firms, and no intrinsic flaw in
HMO coverage or in rationing care (which logically
must occur) by one means over another.
We want to outline a tax credit proposal for the
uninsured that uses private market arrangements as
much as possible. In other words, we give private
markets the benefit of the doubt.
The strongest benefit from private markets is
their ability to satisfy consumers with varying
desires. If everyone wants the same amount of some
product or activity, government provision can work
reasonably well; the “public good,” in which all must
consume the same qu a n ti ty, provides the classic
example. Health insurance is not a “public good”
that all must share equally. Rather, people differ in
how much insurance they want,how they want their
insurer to perform,and how much they value different aspects of plan performance. These differences
in preferences apply both to the level of financial
protection and to the degree to which physicianpatient decisions are constrained by insurers. In
addition, these preferences are not entirely dictated
by a family’s income; people at given income levels
will still choose different covera ge . One key
unknown is how much variation there is in private
demands.
140
The production efficiency benefits that competitive markets can furnish also matter. Even here we
assume that the key issue is not so much current
cost but rather the rate at which costs will grow over
time, and that government provision or production
can be as efficient as private-sector provision or
production.
For various reasons, not enough tax reduction
can or will be made available to fully subsidize
insurance that makes all care free at the point of
service for everyone. There need to be limits, either
on what fraction of the market price insurance covers or on the amount of services providers are permitted/encouraged to supply. We do not advocate
such constraints, or the partial coverage they imply,
but we do recognize reality.
Related to these constraint and diversity issues,
we posit that not all undesirable possible behaviors
should be regulated. Put slightly differently, we
assert that regulations forbidding things that almost
no one ever does are not appropriate. One reason
for not trying to prohibit everything undesirable
that happens only occasionally is the cost of monitoring and administration and the desire for administrative simplicity; the other reason is that passing
such regulations often opens new avenues for additional political influence, lobbying, and legal action.
Coverage and Eligibility
The proposed credit will be a fixed-dollar amount
for persons with incomes at certain levels. (We tentatively suggest , for individuals and , for
families.) The key design parameters that will need
to be specified in a political process are the dollar
amount of the credit, the definition of the income
levels (and possibly other characteristics) that trigger eligibility, and what kinds of insurance are eligible for the credit. Some assumptions about plausible
values for these parameters are made below, but the
choice of their levels is ultimately a political one.
One uncertainty, if we specify insurance coverage and eligibility, is how many of the eligible persons will apply for the credits, and, of those who
obtain the credit, how many were formerly uninsured. Will many of the eligible uninsured pass up
the credit? Will many of the insured claim the credit? Will there be some uninsured people who are not
eligible for the credit?
The other uncertainty is the performance of
markets in which consumers use the credit. The
individual and small-group markets will be most
affected. We know that how markets behave is influenced by the demands, information, and tastes of
buyers who use them, which implies that the influx
of a large number of insurance buyers who are different in important ways from buyers already in
these markets will virtually guarantee that the market will change. But how? Will insurers be more or
less concerned about the risk level of new buyers?
The Health Insurance Portability and Accountability
Act (HIPAA) now requires that individual and
small-group insurers treat those already insured who
become higher risk the same as all others they
insure, so will the additional premium for this “guaranteed renewability” feature rise or fall? Will the
Internet allow new produ ct s ,m ore choice, and lower
premiums, or will it just add to the confusion?
In the current private health insurance market,
only relatively high-income people receive substantial tax subsidies. In current public insurance markets,especially Medicaid, the subsidy is so large that
the net premium is zero or close to it. We have no
experience with significant tax credits for lowermiddle-income families. This means that we cannot
predict with any degree of accuracy how they will
respond, and that no additional information is
available to refine these estimates.
With regard to insurance markets, most potential tax credit recipients are of approximately average risk and would have to use the individual or
small-group markets to obtain covera ge , but we
have no experience with the consequence for such
markets of an influx of a large number of averagerisk buyers in those markets subsidized to pay moderately high premiums. In current markets potential
customers are not willing to pay high premiums or,
if they are,it is because they are unusually high-risk.
Our approach here is to begin with the subset of
the uninsured likely to be most in need of and
responsive to the new credit. (If this policy is effective in prompting the previously uninsured to seek
141
TABLE 1
Segment
Income as
% Poverty
Proportion of
Uninsured
% with Private or
Employment-Based
Insurance*
Low
Less than %
%
%
Medium
–%
%
%
High
Greater than %
%
%
*Includes private individual insurance, private group insurance, and insurance furnished by state and federal governments to employees
and dependents (including military).
Source: Current Population Survey, March Supplement, available at bls.census.gov/cps/ads/adsmain.htm, 1999.
coverage, we think that the presence of large numbers of new demanders will make it easier—and
perhaps inevitable—to change the way in which
insurance is produced and priced for everyone.)
As shown in table , we divide (for a sample
plan) the current population of uninsured
Americans into three groups: low-income households with family incomes below  percent of the
poverty line (about  percent of the uninsured);
those with family incomes above  percent of the
poverty line (about  percent),and those with family incomes between  percent and  percent of
poverty (about  percent of the uninsured). The
last group will be the primary target of a creditbased intervention.
We recommend that eligibility for fully subsidized, complete comprehensive coverage at a publicly chosen insurer (Medicaid, Children’s Health
Insurance Program [CHIP], the insurance plan for
state employees, or an insurance with the same coverage and policies as Medicare) be extended to all
low-income households, regardless of whether they
include children or able-bodied adults. All poor
people would receive free comprehensive insurance.
(These households will also be el i gi ble for the
coupons described below, but we would expect few
to use them.) Households with incomes between 
percent and  percent of poverty (roughly median
income) would receive a credit, but it would be large
enough to cover only part of the premium for complete and comprehensive coverage.
At the upper end of the distribution of el i gi bl e
incomes, the voucher could be “phased down” (over
a range from  percent of poverty to  percent of
poverty) to equal the average value of the employment-based group insurance exclusion. For instance,
for a typical family at  percent of poverty with an
employment-based policy with a premium of ,
and a  percent marginal income tax rate combined
with the payroll tax, the value of the exclusion would
be ,; if the marginal income tax rate were 
percent,it would be ,. Therefore, phasing down
the family voucher to a value of about , might
be reasonable.
The key group might be described as “lowermiddle income people” or “middle-income people
below the median.” As we have defined the groups,
this group contains the most uninsured people.
However, about  percent of people in this group
are already insured, with private or employmentbased coverage; even at the low end of the range,
most people obtain such insu ra n ce . That is, most
people in this group are able to “afford” coverage,
but some (for a variety of reasons) do not purchase
it.1 More important, for reasons we describe below,
this group’s characteristics suggest that it would be
most affected by and would most benefit from a
moderate but not complete insurance tax credit.
1
M. V. Pauly, and M. K. Bundorf. “Is Health Insurance Affordable for the
Uninsured?” Working paper, Stanford University School of Medicine,
November 2000.
142
Finally, while it is difficult (and awkward) to make
distinctions here, one might argue that, relative to
those with higher incom e s ,c redits will have a greater
impact on the health of this population.
This group is very heterogeneous with regard to
almost all characteristics. Whether we look at the
total lower-middle-income population or the subset
of that population that is uninsured, there is substantial representation of all age groups under , all
employment options, and all expense risk level s .O n e
common characteristic that will be important is that
more than  percent of this population is in families with at least one worker; lower-middle-income
families in America (whether insured or not) work
for a living. But for other characteristics, there are
variable (and offsetting) influences. For example,the
uninsured are more likely, other things being equal
(including age), to be in fair or poor health, but they
are more likely than the insured to be young. The
effect of lower average age offsets the effect of the
greater likelihood of being in poor health, so, overall,
their risk level is no higher than average.
Most of the uninsured in this subpopulation (as
in the overall set of uninsureds) have been privately
insured before and have been uninsured for less
than a year (although some are uninsured for a long
time). Thus, this subpopulation has had access to
and experi en ce with private insurance, and most
would be expected to have access again.A potentially important implication of this typical orientation
toward the private insurance market is that single
publicly run insurance programs might not be best
for (or strongly desired by) this group. If buyers will
have to pay some part of the premium for a comprehensive policy (as they must), they may not want to
be channeled into Medicaid, S-CHIP, or even a
Medicare clone. Instead, we hypothesize that, other
things being equal (including the depth of coverage
and the premium the person must pay), most people in this subpopulation would opt for private
insurance, and for a variety of plans, ranging from
reasonably costly but permissive at one extreme, to
high-deductible, low out-of-pocket limit, and
down ri ght cheap at the other. This hypothesis, in
tu rn , implies that more of these buyers would be
more likely to be willing to be insured if the plan
could be private (as well as public) than if it has to
be public only. Some people in this population
almost surely would prefer a public insurer offering
a public plan: one chosen somehow in the political
process and managed at some level by government
employees. Others will prefer a private plan that
they can stay with, even if their incomes rise or they
get decent jobs.
These considera ti ons direct us away from a
strategy of extending Medicaid/CHIP-type plans up
the income scale to the income range for this subpopulation of lower-middle-income persons, but
push us toward making sure that one of the available plans is a public one. Thus, in our ideal
arrangement a public plan of some type is an
opti on , but not an obligation, and the “terms of
trade” among all plans should be neutral and reflect
only true cost and quality differences. This public
insurer could offer a variety of plans, including the
“low-cost” plan described below.
Subsidies
Credits toward purchasing qualified health insurance should be made available to all lower-middleincome, legal residents of the United States. In the
ideal arrangement, one might prefer to vary the
value of the credit with family income in a continuous fashion. However, in the interest of administrative simplicity and making the process of using the
credit as easy as possible, we think it de s i ra ble to
start with a single value for lower-middle-income
families with a given number of dependents. As a
rough rule o f thumb, we think the credit should be
somewhere between half and two-thirds of the premium for a decent basic policy. Both the proportion
of premium covered and the definition of a decent
policy are subject to adjustment (although they are
obviously related).
Ro u gh ly speaking, one might assume that the
average premium for a decent family coverage policy would be , to , per year; therefore, a
credit averaging about , would fit the specifications; the analogous amount for a self-only policy
would be about ,. Credits would be updated in
proportion to the growth in actual premiums paid
143
in transactions. At the high end of the income range
(say, between  percent and  percent of poverty), the credit might be phased down gradually to
equal the average value of the exclusion. This phasedown could be accomplished by end-of-the-year
adjustments in income taxes.
However, the actual premiums that would be
quoted in the individual market would vary with
age. The self-only premium for people under age 
would be less than ,, for example, whereas that
for someone age  (even in reasonably good health)
would be about ,. Should these effects of age be
of fs et? There is no easy answer, nor are there hard
data, for this question. A fixed-dollar (unadjusted)
approach will pay for decent coverage for more
people, but they will be young. A proportional
approach, or some other method that increases the
credit when the premium increases because of age,
will provide better incentives for decent coverage for
older people. However, evidence suggests that older
people are willing to pay more for covera ge ; they
have a clearer idea of their need for benefits, and
they behave more responsibly. In contrast, younger
people are most likely to be uninsured. The real issue
here is precisely what is not known—how many
people of which ages would buy coverage under different plans, and how many older people with difficulty in obtaining coverage would compensate us for
covering a large number of young people?
In the first phase,in the interest of simplicity, we
propose a uniform credit, independent of age. We
can then see what pattern of insurance purchasing
emerges. If this program appears to be ineffective in
affecting the coverage choices of middle-aged people,the credit could be age-adjusted to some extent.
Doing so, however, would cause some potential conflicts with employment-based insurance, as we discuss below.
Financing
Financing for the tax reductions or credits for the
lower-middle-income population ( percent to
 percent of poverty) will be accommodated by
federal budget revenues. Those who purchase coverage more generous than the low-cost plan will use
their own resources, either through direct payment
of premiums or through indirect payment by
employees as part of their total compensation in lieu
of money wages. Financing for full coverage for
those below  percent of the poverty line will be
provided by a combination of state and federal revenues, with existing levels of state payments for
Medicaid and S-CHIP to be retained.
These new credit vouchers or coupons should be
thought of as tax reductions for the lower middle
class and, thus, treated as tax reductions for all who
use them. Since virtually all eligible persons should
be expected to claim and use their credit coupons,it
should be possible to estimate the gross value of the
tax reduction fairly precisely. However, since use of
the credit offsets the value of the tax exclusion, the
net tax reduction will be somewhat more difficult to
estimate, because it requires knowledge of the distribution of the value of the exclusion and, more
important, because persons at the upper end of the
income eligibility range who now benefit from the
exclusions may or may not prefer to use the credit
coupon rather than continuing to use the exclusion.
Insurance and Risk
The two most common ways in which private insurance is provided in the United States both have
problems. The most dominant form, employmentbased group health insurance, suffers from lack of
portability across jobs and lack of good matching
between the types of plans employers or unions
choose to offer and what each employee (or perhaps
even most employees) really want. We know that
workers are dissatisfied with employer-chosen managed care plans (and,to a much greater extent, when
there is only one managed care offering), and we
know that employers do not feel that they can take
enough out of wages to pay for unrestricted indemnity insurance plans. The result is a compromise
between offering restrictive inexpensive plans and
more costly plans with “parachutes” in the form of
point-of-service features or preferred provider
organizations (PPOs) with very large provider networks.On the other hand,the non-group insurance
market uses a much larger share of the prem iu m
144
dollar to cover administrative costs, and may make
it difficult for some “non-average” risks to obtain
coverage at premiums they find acceptable. Policy
makers, therefore, tend to dislike both currently
available vehicles for providing health insurance,
and often imagine that there could be potential
group-purchasing arrangements, rather than the
job-based setting for individuals or small firms,that
could offer a wide range of choices to voluntary participants and keep premiums low. At a minimum,
many feel that the market in which credits may be
used should be organized into some type of “quasigroup” setting and regulated to prevent risk rating.
This hope may well not be realized, however, so we
need to design a plan that does not depend on it.
There are two reasons why individual plans are
so expensive that reinforce each other. It is costly to
offer a large variety of plans to one buyer at a time;
the plan finally selected will be costly because it is
custom-designed. But when the plan is costly for the
benefits it provides, more effort has to be made to
persuade people to buy it, so commissions are generous. There is a kind of catch- in that insurance is
hard to sell because it is expensive, and then it
becomes expensive because it is hard to sell.
Currently about  million people under age 
buy coverage in this customized and costly individual market. Ei gh teen million more people would
receive credits and become potential customers in a
reformed situation. If the credit is set on the generous side and a wide variety of policies at different
costs are eligible, many of these buyers will be eager
to obtain insurance. It is possible that the flow of
large numbers of heavily subsidized demanders who
are known to be average risks (or better) could help
individual markets to function better. The subsidy
itself should simplify one of the most difficult and
costly tasks in the current individual market.
Commissions (and some part of general administrative expense related to billing) are substantial,
and they serve primarily to compensate agents and
brokers for persuading people to buy insurance.
(Since brokers typically offer plans from a variety of
companies, they put little effort into selling one
firm’s product rather than another’s). But with a significant subsidy, there should be much less need for
an expensive sales effort. The credit or coupon only
becomes valuable if it is used for insurance, so people will want to use it.
There is evidence that when significant subsidies
to purchase are offered, private administrative costs
are reduced substantially. When Blue Cross of New
Jersey was required to heavily subsidize individual
coverage, administrative expenses were only  percent of premiums. Private insurers in Chile sell to
customers who receive credit for the payroll taxes
they have already paid; even though the coverage
has had upper limits until recently, the loading is
still about  percent.2
Not only might the creation of a mass of new
demanders allow for lower administrative costs, it
also might help provide direct help to avoid the
other serious problem with individual insurance:
difficulties in dealing with risk variation. If individual insurers do not charge higher premiums to
higher-risk individuals, they will have to set (average) premiums so that insurance is too expensive
for average-risk individuals compared with what it
provides. A credit will greatly diminish this problem, for several reasons. As noted above, the lowermiddle-income uninsured are reasonably good
risks, so the proportion of high risks in the pool of
potential individual market buyers will greatly
diminish. With fewer high risks to worry abo ut ,
firms would rationally put less effort into trying to
identify high risks to charge them higher premiums.
The benefit from identifying a high risk is the avoidance of high claims that only a small minority
incurs. But if the proportion of such risks drops in
half, say, the screening cost to identify one such risk
doubles, which is bound to lead to less underwriting
at the margin. Even though insurers know that the
bulk of benefits will be paid to a few insured persons,if the proportion of such persons is few, it does
not pay to incur underwriting expenses to discover
who they are. Moreover, with a generous enough
subsidy, many lower risks will still find insurance a
good deal, and will stay in the pool rather than for2
C. Sapelli and A. Torche. “The Mandatory Health Insurance System in
Chile: Explaining the Choice between Public and Private Insurance.”
Submitted for publication, International Journal of Health Care Finance
and Economics (2000).
145
going coverage. The threshold level of new buyers
that can alter insurer underwriting practices is not
known.
This vision of what the insurance market might
look like obviously is not guaranteed to materialize.
However, if we offer generous credits with only
modest amounts of premium rating regulation, we
will soon know whether enough good-risk buyers
will enter the market to make it a reality. Of course,
if we are right, even fairly strict premium regulation
would not be constraining. If the behavior a regulation is designed to constrain is going to be rare anyway, why not play it safe and regulate it? The most
obvious answer is that writing, monitoring, and
reporting to comply with regulations have a cost of
their own.
So there are a number of regulatory options
here, none of which is fully satisfactory. In Phase I
we would propose a compromise strategy in which
any low-cost plan must be sold under modified
community rating, while plans with more generous
coverage could charge higher premiums to those
(given age, sex, and location) whom they identify as
high risks. In all cases, insurers would be permitted
to impose modest waiting periods for people who
did not enroll during an initial “open season.”
However, we think that this will be rare, because all
persons would, at a minimum, be enrolled initially
in a low-cost plan.
An important reason why few high risks should
need new coverage is that a kind of risk regulation
already exists in federal law that helps to provide
substantial protection to high risks. This is the
requirement that all non-group insurance be sold
with a guaranteed renewability provision (section
. of HIPAA),in which the insurer must promise not to raise premiums selectively for those
already insured who become high risk;in return,the
initial premium to low risks is slightly higher than it
would be otherwise. Over time, unless there is enormous turnover among insurance plans, this provision should result in almost all high risks paying
average-risk premiums.
Guaranteed renewability protects people against
increases in premiums because of the onset of highrisk conditions. It does this by offering insurance
that guarantees that a person’s premium will only
increase at the same rate as the premiums of all who
buy that insurance plan.Such a promise is financially feasible for insurers because they charge what is,
in effect, a “two part” premium—one part to pay for
current-period expenses and the other to cover any
above-average premium for those who began as
average or low risks but became high risk. Both
common sense and economic theory 3 suggest that a
risk-averse low risk should prefer to stay with such a
plan rather than switch to one in which the insurer
has a reputation for increasing the insurance premium of people who become sick. For such a guaranteed renewability arrangement to work, however,
people have to be willing to buy insurance, even
when they are not high risk. If they choose to buy
insurance only after they get sick, such behavior will
prevent any insurer, regulated or not, from being
able to cover its benefits costs with moderate premiums. For any voluntary market to work,there needs
to be a penalty on low risks who try to stay uninsured until they get sick. If a credit makes insurance
affordable for lower risks,there is no longer any justification for such irresponsible behavior, and
penalizing those who wait to buy insu ra n ce until
they become high risks before seeking insurance
makes sense.
A guaranteed renewability requirement should
require minimal enforcement and, indeed, almost
all individual policies contained this feature even
before HIPAA required it. Nothing is completely
without cost; guaranteed renewability will lock
h i gh er risks into particular insurance companies,
but that prospect should make prospective buyers
more careful in the first place. The key point is that,
with guaranteed renewability, all persons who keep
buying health insurance will be protected against
high premiums, even if they become high risks.
Finally, since almost everyone can buy at least a
low-cost policy at modified community rates, there
should be few high-risk customers who were formerly uninsured. To the extent that the low-cost
insurance already covers catastrophic expenses, the
3
M. V. Pauly, H. Kunreuther, and R. Hirth. “Guaranteed Renewability in
Insurance.” Journal of Risk and Uncertainty 10 (1995): 143–56.
146
If credits or coupons are widely offered, it is reasonable to assume that profit-seeking
insurers will try to sell their insurance to people who are receiving substantial subsidies
to pay for it. Thus, our strategy is to rely on insurers to market insurance to their customers,
rather than use government agencies to enroll clients.
impact of risk differences on the cost of incremental
coverage should be smaller. There would still be a
modest reward for buying more comprehensive
coverage with guaranteed renewability, however, in
the form of no waiting period and, possibly, no
higher premiums. Indeed, theoretical research suggests that a good way to deal with adverse selection
is to combine a subsidized community-rated base
policy with risk rated add-ons.
Administration and Regulation
The Role of Employers
The great majority of people with private insurance
in the United States obtain it in part because they
receive a portion of their compensation in the form
of tax-shielded health insurance premiums. That is,
the employment contracts for most American
workers entail a compulsory diversion of compensation to health insurance; the wage-benefits package includes a partially paid health insurance
premium available to all workers. A worker who
declines to use this employer-enforced contribution
may be able to save the other part of the premium,
labeled “employee share,” but the “employer” portion generally is not returned to individual workers
who decline coverage.
Employees and employers both prefer to arrange
compensation for employees in this fashion because
it is the only simple way to obtain a substantial
reduction in the amount of employee income subject to income and payroll taxes, and because it
tends to lower insurer administrative cost (though
limiting plan choice).
We propose that any tax credit, in contrast, not
be limited to workers who choose this form of
group insurance. In s te ad , all people at the target
income levels would be eligible for the credit if they
obtain insurance at least as costly as the credit. The
credit would go to the newly insured who were formerly uninsured, those who obtained their insurance by spending their wages on non-group
covera ge , those who divert part of their wages to
group covera ge , and those who formerly chose
employment-based coverage and now have some
other method of arranging insurance. The size of
and eligibility for the credit would depend only on
whether the person obtained insurance,not on how
he or she obtained it.
In contrast to the current situation in which
employment-based group insurance is subsidized,
but other ways of obtaining insurance are not,
incentives would be neutral. Therefore, credits
would be available to those with “employer paid”
insurance; neutrality would be achieved if credits
were set equal to the difference between the value of
the exclusion and the value of the credit (or, in the
alternative, workers using the credit could simply
report total employer premiums as taxable income).
For those firms able to offer attractive plans at lower
premiums than their employees would pay in the
non-group market, we expect that employer provision would continue.
There would be some potential shifting of workers out of the group market into the individual market, compensated by increases in money wages,
especially for very small and poorly run groups.
These workers indeed might be better off with individual insurance if it were not much more expensive
than their (small) g roup coverage and offered them
a va ri ety of plans that better matched their needs
and were portable across jobs. A defined contribution strategy is another good way to implement
such a transition. This change would make many
lower-wage workers better off than they are now,
either because they would obtain coverage for the
first time or because they are able to choose the coverage they prefer.
147
Some fear that allowing insurance to have the
same tax tre a tm ent in both group and non-group
s et ti n gs would som eh ow “break the pool” in
employment-based group insurance and harm
higher-risk working families. We think that this is
extremely unlikely. Employers would not be
required to let workers (individually or collectively)
out of the pool, so there would be no way for individual low-risk workers to impose costs on fellow
workers. Moreover, the empirical evidence strongly
suggests that the current reduction in money wages
to pay for group insurance is greater for older workers and women, other things being equal. Younger
workers who seek a relatively inexpensive individual
policy would not expect (if their company dropped
its coverage) to get back very much in money wages,
so that individual insurance would not be a very
good deal.
Finally, the proportion of people in employmentbased group insurance who are truly high risk is
quite low. The reason is simple: to get such insurance, someone in the household must be able to
work. Among workers themselves, very few people
have high-cost chronic conditions. There are highcost dependents, however, but the availability of
public programs for the disabled also draws off
most of the high risks. In effect, people in employment-based pools have already been pre-screened,
and their premiums (in virtually all regulatory
regimes) depend on the risk le vels of insured workers and their dependents, not on the average risk
level in the population or even the average risk level
among all workers and their dependents.
This pre-screening helps to keep down the possibility of adverse selection, so there is some merit to
linking insurance to the employment relationship
for at least some workers (though not necessarily
all). There are some other reasons to maintain a
connection. Employers have an interest in seeing
that workers do not miss work because of illness
and,thus,disrupt the production process; for workers who are key members of production teams, lost
output can be substantial. Much of the long-term
incidence of improved productivity through
reduced absenteeism and greater “presenteeism”
(higher productivity on the job) will be transformed
into higher worker wages. The payroll mechanism
may be a good vehicle for ensuring regular deductions from salary to pay premiums, and starting a
new job may be a good time for people to think
about adding insurance benefits. But the best way to
decide whether these advantages of fs et the disadvantages of letting the employer select and manage
the worker’s insurance is to let workers make neutral choices.
Marketing Subsidized Insurance
One of the ch a ll en ges to government-managed
insurance plans such as Medicaid or S-CHIP is to
get poor people who are eligible for subsidized coverage to enroll and accept their subsidies. Even now,
nearly a third of those eligible for Medicaid fail to
obtain it, and, after its recent “success,” S-CHIP
picks up barely a third of the remaining uninsured
children. This is unacceptable. Sometimes, one suspects, state governments with budget concerns
might not be too worried about a low “take-up rate,”
but even when efforts are made, the rate is often
(though by no means always) low. It is the low takeup rate for such free insurance that causes some
estimates of the impact of credits on coverage to be
relatively low.4
If credits or coupons are widely offered, it is reasonable to assume that profit-seeking insurers will
try to sell their insurance to people who are receiving
substantial subsidies to pay for it. Thus,our strategy
is to rely on insurers to market insurance to their
customers, rather than use government agencies to
enroll clients. The main problem that these agencies
have faced is the difficulty of limiting enrollment to
those eligible, while at the same time encouraging
people to reveal the financial information needed to
determine eligibility. To do so, they have felt it necessary to impose onerous burdens on potential eligibles, such as New York’s annual obligation to provide
face-to-face evidence of income and family composition qualifications to a government official. Our
strategy, even in this first phase, but more effectively
in the next,is to achieve a high take-up rate by sepaJ. Gruber and L. Levitt. “Tax Subsidies for Health Insurance: Costs and
Benefits.” Health Affairs 19 (1) (January-February 2000): 72–85.
4
148
rating the purchase of insurance from the process of
establishing eligibility. To do this, we propose to provide potential clients with vouchers or certificates
good for , or , off an insurance plan premium,and sending them (through the mail or some
other device) to people with incomes close to what
would make them eligible.
The certificate could be transferred to the insurer who provides coverage,along with any initial premium,and then redeemed by the insurer for its face
value after a period of con ti nuous covera ge .
However, the certificate would be coded with the
person’s Social Security or other taxpayer identification number, and anyone who used the certificate
but was ineligible to do so (because family income
was too high) would have to make up the difference
(and pay a penalty) on his or her income tax return.
This device also would provide a convenient way to
subsidize people who are eligible for part of the year,
and it would provide a way to obtain insurance that
would not require the lower-middle-income buyer
to advance the full premium before seeking reimbursement. Lower-income families (below  percent of poverty), technically eligible for free publicly
managed insurance, also would be permitted to use
the certificates if they preferred private insurance to
Medicaid or S-CHIP.
For people with no “employer-paid” coverage,
the credit coupon could simply be tu rn ed over to
the insurer, which would then redeem it with the
government. For those whose employers paid for
premiums as part of compensation, eligible employees could attach the coupon to their tax returns,and
calculate the net credit as the difference between the
value of the credit and taxes on the employer payment. While precise details depend on parameter
values, these ad ju s tm ents also would permit the
value of the credit to equal the value of the exclusion
for incomes near or slightly above the upper limit.
Initial distribution of the coupon could be done
in several ways. Coupons could simply be mailed to
those indicating a low expected wage on their tax
withholding (W-) form. These also could be available at post offices or other convenient sites.
Insurers could furnish coupons,as well. The coupon
would clearly state who is eligible to use it, and
impose end-of-year tax penalties on those who are
ineligible. If Publishers Clearinghouse can reach
nearly everyone,there must be a program that could
work for health insurance.
Finally, those taxpayers with incomes in the
range at which they are eligible for the credit, who
nevertheless fail to redeem their credit, could be
enrolled automatically in a low-cost insurance plan
with a premium no greater than the credit. Such a
process will require a monitoring mechanism.
Benefits
An important design issue concerns specifying the
policy for which the credit may be used. Defining
“minimum benefits” is always politically troublesome, because every supplier of medical services
will lobby to have generous coverage of its services
included. But requiring that everything be covered
means that the premium will be very high—higher
than any feasible credit, higher than many persons
are willing to pay, and higher than taxpayers are
willing to subsidize.
A solution to this problem would be to have a
fairly inclusive definition of covered services, but
permit policies to hold down premiums through
deductibles, coinsurance, and upper limits. That is,
at least initially, any policy that would qualify for the
credit would have to pay for all medical and surgical
services and all prescription drugs and medical
devices, based on some commonly accepted definition of what constitutes standard (non-experimental) care. However, cost sharing could be imposed,
as could managed care rules and incentives for
providers. The definition (for medical and surgical
services) used for the traditional Medicare plan
probably could be used here, but, even if some of
the services not typically covered were included in
the definition of covered services, the presence of
cost sharing would limit the extent to which they
would be used as well as their additional cost.
Citizens could be free to choose, at additional cost,
plans that cover these services with lower levels of
patient cost sharing. In effect, we tell people that
they may use their individual , credit to pay
for any policy covering a set of medical and surgical
149
goods and services that has a premium of   ,  
or more.
However, one can be certain that, unless the
credit is as large as the premium for a generous
insurance plan, some people will not be willing to
make additional payments to buy some specific generous plans, but will choose either no insurance or a
partial-coverage plan inste ad . Different analysts
(and different citizens) have different views on how
much coverage they think a family at some income
level ought to have. There is a trade-off here. We
could put stricter bounds on qualified covera ge ,
making such coverage more adequate. But the credit would not cover the full premium, and then some
people would refuse the credit, preferring to remain
uninsured rather than pay an additional premium.
In line with our primary objective of covering
the uninsured, we think that,at first,there should be
virtually no rules about cost sharing. The only rule
would be that the policy provide dollar benefits
appropriate to its premium. Rather than attempt the
surely difficult and probably impossible task of
specifying the appropriate levels of cost sharing for
every credit recipient, we propose to begin by letting
the credit recipients themselves decide what kind of
coverage they prefer. Once we know the pattern of
coverage, and can check to see what effect it has on
access to care and health levels, we can judge
whether stricter regulation of the package is needed.
We start in an unrestricted way (and try to stay
with light restrictions) because we think that the
highest-priority objective ought to be to get at least
some health insurance coverage to every American
who is not high-income, even if the level of coverage
cannot be adjusted (or financed) to be what some
regard as perfect in the initial round. We postulate
that the law of diminishing returns holds here as
everywhere else: the first infusion of coverage will
do the most good, and the benefit of latter additions, though probably positive, will be smaller. The
way to get almost everyone to buy coverage in
response to a credit is to permit people to use the
credit as they like, with as much or as little of their
own contribution as they prefer. It would be irrational for someone to refuse to use his or her credit
on a (private or public) insurance policy that costs
no more than the credit and provides at least some
protection against what otherwise would be out-ofpocket payments the person would be forced to
make and/or a policy that improves access to some
types of care.
Finally, if insurance did take the form of full coverage above a deductible, the RAND health insurance experiment reassures us that cost sharing will
not have substantial adverse effects on most measures of health status for the non-poor. In the experiment, the only adverse health outcome (beyond
some minor effects on vision correction and oral
health) was for people initially at high risk for
hypertension. For the majority at normal risk and
for other high-risk conditions, cost sharing does not
appear to harm health, as long as there is catastrophic coverage.5
Fit with the Current System
The Role of States
States would have the primary role in administering
the Medicaid-like coverage for poor adults. The
bulk of the additional cost for this coverage would
be provided by the federal government.
States would be permitted to regulate the extent
to which premiums for coverage in excess of the lowcost policy vary with risk, but would be penalized if
this regulation caused people to remain uninsured
or caused disproportionate numbers to choose minimal-coverage (low-cost) policies. Finally, states
could provide payments for people with very high
costs or chronic illness. Doing so would mean that
the low-cost coverage in such states could have a
smaller deductible or less-constraining upper limits.
What Improved Private Insurance
Markets Might Look Like
(and How We Can Get There from Here)
It seems eminently plausible that the influx of large
numbers of lower-middle-income workers and their
families into non-group insurance markets would
5
J. P. Newhouse. Free for All?: Lessons from the RAND Health Insurance
Experiment. Cambridge: Harvard University Press, 1993.
150
change those markets. We have offered some optimistic views of how they might change, but the
truth is that no one knows for sure what might happen. The usual response of policy designers or policy makers is to abhor uncertainty (as definitely
worse than a vacuum) and, therefore, to design
many rules and some incentives to configure the
outcome. For reasons stated earlier, we are skeptical
of such efforts, not only because there really are no
experts who know what outcomes will be produced
by what rules, but also more generally because we
think it important to give the widest scope to possible innova ti on s . This said, we do think that some
probable contours of a new market may be worthwhile to forecast and even to contemplate encouraging (or at least not deterring).
The most common reaction to proposals that
employees be armed with credits that they can use
for different insurance from that which their
employers provide (if the employers provide any at
all) is to begin to imagine the design of a new and
improved type of group insurance, with many choices (all good), well administered. We do not share this
vision of a single tidy market for all. We remain
skeptical that such quasi-groups generally can
achieve anything like the administrative savings now
obtained by large groups with compulsory participants. We think there ought to be a “HealthMart”
option, but that we ought to think of (and permit or
foster) other alternatives. The key to considering
alternatives is to note that there must be trade-offs: a
scheme can offer a limited set of choices priced in a
transparent fashion, but that scheme surely will have
a cost in terms of the scope of options (especially
new options) offered and, probably, in terms of
administrative cost. Another unrealistic view is to
imagine that the Internet alone can cut administrative cost and expand choice without being bedeviled
by adverse selection and mass confusion (not to
mention an inability to generate revenues).
Our proposal in this case is to impose a modest
amount of voluntary standardization on plans and
plan types, and then allow the market itself to sort
out how they should be offered. The standardization
would allow plans to be designated as meeting certain model types, to facilitate shopping and compar-
ison, but a plan could be non-standard as long as
buyers were clearly warned about its type. “Site” or
program sponsors could offer assurances or guarantees about the plans they list, but they would then be
responsible for the performance of the plans they
list. Conversely, an eBay of health insurance could
list all legitimate offers cheaply, but make no representation about their quality beyond what is
embodied in state regulations.
The public sector would have a role in making
information about insurance purchasing and plans
available to potential consumers. Information about
quality and price is in the nature of a public good,
and would not be supplied adequately or in an
unbiased fashion in a private market (although
development of helpful and accurate guides for federal employees suggests that the same thing could
happen for  million new insu ra n ce customers).
Subsidizing the production and distribution of such
information would be a good role for government;
one vehicle would be to distribute publicly financed
vouchers that could be redeemed toward purchasing
the buyers’ guide of one’s choice.
Monitoring, Adapting, and Phase II
We recognize that the Phase I plan we have suggested will not achieve universal comprehensive covera ge , nor will it necessarily perfect how insurance
markets function. That is why we envision a formal
monitoring process and a plan to bring the other
two groups in the population—people below 
percent of poverty and people above  percent—
into the system in a second phase.
What would be the best way to monitor this system’s performance? Possibilities might include the
fo ll owing: State insurance or health departments
(which often regulate health insurance anyway)
might be one expert entity. Or there might be a federal advisory commission, as there has been for
Medicare. It would be desirable to offer incentives to
any oversight body to make the process work. For
example,states might be rewarded if the number of
uninsured in their state in the target income category fell more than expected,and they might be penalized in some fashion if the number fell short, espe-
151
cially if the shortfall could be attributed to state regulation of rating or costly mandated coverage provisions in the individual or small-group markets.
States also might be given the task of monitoring
the effect of credit-subsidized insurance on health
status. In addition, they should determine whether
the partial coverage plans affordable with the credit
amount alone adversely affect health status more
than plans with more generous covera ge . State
health departments already exist, and it is likely that
effects vary by state.
The key point here is that monitoring should
focus on changes in use of services and, ideally, on
changes in health outcomes. In the state-monitored
scheme, for example, if a state were able to encour age the use of more services that are effective for
health, we would not be especially concerned about
the distribution of people who switch from being
uninsured to insured. Because the availability of
credit certificates increases the competitiveness of
firms not offering group insurance in the labor market, those firms that did offer some coverage would
be motivated to manage it more effectively to continue to be competitive, and these changes might
help health outcomes substantially. Attention also
might be given to other consumer goals, such as
financial stability and relief of anxiety.
The specific issues to be monitored most closely
are the two major policy uncertainties: how many
people of what type will take the credit,and how the
group market based on groups other than employment will be transformed. If substantial numbers of
large firms dropped group covera ge , or if small
firms with high-risk workers did so, that would
indicate that something is wrong. (Remember: the
fallback insurer is always available for a person at
any risk level, and the low-cost coverage has adjusted community-rated premiums.) Enhancement of a
high-risk pool would be the p roper response to any
evidence of increased risk segmentation. If credits
were not claimed in adequate numbers, that would
imply that the value of the credits was too small
compared with the premium for a good policy. The
supply, purchase, and form of any “zero-premium”
partial-coverage policies should be monitored, as
well.
Quality: What Would a Good (but
Imperfect) Low-Cost Plan Look Like?
One of the most controversial aspects of our plan is
its acceptance that people may use a moderate-size
credit to buy a less than comprehensive policy
whose premium is close to the credit. It is easy to
argue that this is bad idea if one assumes that people
will make poor choices in the plans they choose, or
if one believes that nothing less than full coverage
will do. (To achieve “first best” optimal covera ge ,
either a mandate or a lavish public budget would be
necessary. While we favor a mandate in a politically
unconstrained world,6 in this paper we assume that
mandates are not feasible.) At one extreme, the previously uninsured person could buy a policy with
full coverage above a substantial deductible. While,
according to insurance theory, this makes perfect
sense in an otherwise perfect world, in reality it is
likely to be unattractive to some lower-middleincome people because they would expect much of
the benefit to substitute for charity care they might
have received for free if they contracted a serious illness (and were able to obtain care without insurance coverage).
On the other hand, a first-dollar policy with a
very low upper limit would not be especially attractive either (although it would be better than nothing
and would appeal to those consumers, fervently
believed by politicians to exist, who only want
insurance if they can be assured of collecting some
money from it). The best cost-sharing pattern might
be one with a moderate but not trivial deductible
(say, ), and as high an upper limit as the credit
will buy. Our own analyses suggest that, for the average worker, the upper limit might be on the order of
, to , per year.7 Illnesses costing more
than , could be covered by public insurance
for the chronically ill, if states chose to do so. The
fallback insurer could assist people who do not
understand which low-cost coverage is best.
M. V. Pauly, P. Danzon, P. J. Feldstein, and J. Hoff. Responsible National
Health Insurance. Washington: AEI Press, 1992.
6
7
M. V. Pauly and B. Herring. “Expanding Insurance Coverage through
Tax Credits: Tradeoffs and Options.” Health Affairs 20 (1) (JanuaryFebruary 2001): 1–18.
152
The other strategy this insurance might follow is
aggressive use of closed panels and managed care,
but with the option of going “out of plan” with outof-pocket payments. Given the oversupply of physicians in the United States and the wide range of
hospital costs, an effective discount network might
be established to fit within or close to the total value
of the credit. However, we strongly suspect that,
once a person signs up for a frugal plan, the experience of high out-of-pocket payment or strict
rationing will prompt that person to put a moderate
amount of his or her own money on the line and
buy the next step up in insurance policies in the
future. There is no doubt, however, that the mar ket
for low-cost plans would be stimulated.
Phase II Administrative Changes
Now suppose that a new and improved type of nongroup health insurance market does em er ge , and
that credits substantially reduce the number of
uninsured. The next step is to allow those with
incomes below  percent of poverty to use such
private insurers. This would be relatively easy to do;
we only need to permit poor people currently using
the public fallback insurer to use the same money to
buy private insurance that non-poor people do.
That is, the comprehensive Medicaid-like coverage
already being provided to the poor would be converted into a premium support system, with expenditures converted into vouchers usable for private
alternatives.
The other change is that it should be possible to
abolish, or at least radically transform,the safety net
system. This system is intended to help people who
“fall through the cracks” through no fault of their
own. If credits for full coverage are offered to people
with low incomes,and if lower-middle-income people receive credits large enough to allow them to
afford insurance,there should be very few who qualify for the safety net system (perhaps only non-regis-
tered aliens and people who become ill in the midst
of a transition from one plan to another). One might
convert payment for the services provided by public
hospitals and clinics into a pre-paid plan, and allow
people to sign up for this type of insurance.
What of the people with incomes above  percent of poverty? If a properly functioning private
insurance system em er ge s , we think it might be a
good idea to mandate that they buy covera ge . The
simplest way of doing this would be to levy a tax surcharge equal to the premium for the fallback coverage on people with incomes above the median who
are not insured. This system could be put in place
easily, and it would affect only a tiny minority of the
population who could hardly claim to be financially
strained to pay the tax. Mandating coverage for
lower-middle-income people (as already noted) will
prove to be a more serious problem, and we do not
advocate it now. The tax exclusion could be extended at a capped level to all high-income people (that
is, to the small minority who currently buy nongroup coverage, possibly as part of a tax law change
capping the value of the exclusion for all).
Conclusion
This proposal suggests addressing the lower-middle-income uninsured first with a system of generous tax credits to purchase insurance of their choice
in a lightly regulated competitive market. If the
good outcome that is possible does emerge,then the
poor, near-poor, and the well-off could be invited to
join this system, with substantial subsidies for the
poor and substantial good wishes for the well-off.
If the plan does not work, we could hardly be
worse off for trying. We will have settled the controversy over what kind of markets private insurers can
expect to offer, and what kinds of roles private
insurers should be expected to play in helping to
deal with the uninsured. n
153
Singer, Garber, and Enthoven Proposal
Key Elements
Sara J. Singer, Alan M. Garber, and Alain C. Enthoven have designed a comprehensive, new approach for expanding access to health insurance. The proposal
is built on the following key elements:
       near-universal coverage by making private
plans more affordable and helping low- and middle-income people buy
coverage. This would be accomplished though tax credits and by creating
“insurance exchanges” that would provide health insurance choices and
promote competition among health plans.
         by public or private entities
or employers (for their own employees). Exchanges would offer individuals
a choice of at least two health plans in every geographic region at community-rated premiums. The “U.S. Insurance Exchange” would be established
to serve individuals and companies with fewer than  employees in areas
where private exchanges do not emerge. Coverage purchased through
exchanges would be exempt from state small-group reform laws and
insurance mandates.
 -  -  who purchase insurance through
an exchange would receive refundable tax credits valued at  percent of
the median-cost plan. The credits would apply only for coverage purchased
through the exchanges. Eligible low-income individuals who did not enroll
in a health plan would be automatically enrolled in a federally funded
default plan organized by the state. Other individuals would continue to
exclude from taxable income their individual or employer-paid health
insurance contributions, but a phased-in cap would limit this exclusion.
  “          ” would be created. It would be
similar to the Securities and Exchange Commission—having authority to
distribute tax credits and default payments, accredit insurance exchanges,
risk-adjust premiums across insurance exchanges, and serve as an information clearinghouse for consumers.
154
About the Authors
 . , ..., is Executive Director of
Stanford University’s Center for Health Policy and
a Senior Research Scholar at Stanford’s Institute for
International Studies. She is program director for
the Center for Demography and Economics of
Health and Aging, project director for the “Project
on Global Healthcare Productivity,” “Developing
Best Practices for Patient Safety,” and “Managed
Care and Health Care Markets,” and she was lead
investigator for “Decreasing Variation in Medical
Necessity Decision-Making.” Ms. Singer is also a
member of Stanford’s University Committee on
Faculty and Staff Benefits. In , Ms. Singer
served as staff director of the California Governor’s
Managed Health Care Improvement Task Force.
She developed legislative strategy for moderate
Democrats on the Ways and Means Committee
during the committee’s debate of the Health Security
Act of . In , Ms. Singer was a health policy
analyst in the Executive Office of the President,
Office of Management and Budget. She holds an
A.B. degree from Princeton University and
an M.B.A. degree with a Certificate in Public
Management from Stanford University.
  . , .., .., is the Henry J. Kaiser, Jr.
Professor and Professor of Medicine at Stanford
University, where he is also Professor of Economics,
Professor of Health Research and Policy, and
Professor of Economics in the Graduate School of
Business (courtesy). He is the founding director of
both the university’s Center for Health Policy and
the Center for Primary Care and Outcomes
Research at the School of Medicine. He is a Staff
Physician at the Veterans Affairs Palo Alto Health
Care System and Director, Health Care Program, of
the National Bureau of Economic Research, Inc.
(NBER). He serves as Chair of the Medical and
Surgical Procedures Panel of the Medicare Coverage
Advisory Committee and is a member of the
national Blue Cross and Blue Shield Association’s
Medical Advisory Panel, the American Society for
Clinical Investigation, and the Institute of Medicine
of the National Academy of Sciences. A summa cum
laude graduate of Harvard College, he received his
Ph.D. in economics from Harvard and an M.D. with
research honors from Stanford. His research is
directed toward methods for improving health care
delivery and financing, particularly for the elderly,
in settings of limited resources.
  . , .., is a Senior Fellow, Center
for Health Policy, Institute for International Studies,
and the Marriner S. Eccles Professor of Public and
Private Management, Emeritus, in the Graduate
School of Business at Stanford University. He holds
degrees in Economics from Stanford, Oxford and
MIT. He has been an Economist with the RAND
Corporation, Assistant Secretary of Defense, and
President of Litton Medical Products. In , he
received the President’s Award for Distinguished
Federal Civilian Service from John F. Kennedy.
He is a member of the Institute of Medicine of
the National Academy of Sciences and a fellow of
the American Academy of Arts and Sciences. He
is a consultant to Kaiser Permanente, the former
Chairman of the Health Benefits Advisory Council
for CalPERS, and former Chairman of Stanford’s
University Committee on Faculty and Staff Benefits.
He has been a director of the Jackson Hole Group
and PCS. He was the  winner of the Baxter Prize
for Health Services Research and the  Board of
Directors Award, Healthcare Financial Management
Association. His latest book, published in , is In
Pursuit of an Improving National Health Service.
155
Near-Universal Coverage Through
Health Plan Competition
An Insurance Exchange Approach
by Sara J. Singer, Alan M. Garber, and Alain C. Enthoven
Overview
We propose to expand access to private health insurance among the non-elderly population ineligible
for Medicare. Our plan will accomplish this goal by
making private plans more affordable for low- and
middle-income households and by promoting competition to increase the value of insurance offerings.
The proposed approach will promote the collection
and dissemination of information on the quality of
health plans and their providers. The plan promotes
higher-value health insurance coverage by exposing
consumers to price differences and better information about plan quality. Although the plan can
accom m od a te Medicare beneficiaries with little
structural modification, we do not propose to
replace Medicare during the plan’s initial implementation. Beneficiaries currently enrolled in
Medicaid, State Children’s Health In su ra n ce
Programs (S-CHIPs), or other government programs could remain in these programs, but could
choose instead to enroll in private plans. The plan
also provides new funding and incentives for states
to improve access to basic health services for the
uninsured,and to improve insurance coverage rates.
The plan does not impose new mandates on
employers to pay for coverage.
The proposed plan has these key elements:
• Insurance exchanges are public or private entities, including certified employers, that serve as the
vehicle through which most individuals acquire
health insurance. They offer individual choice of a
minimum of two different health plans on a guaranteed-issue and community-rated basis, with incentives to choose high-value plans. At least one
insurance exchange generally is available in each
geographic region. Exchanges help make coverage
affordable by being large enough to achieve
economies of scale in brokering plans and in providing information to enable people to make choices among plans. The principal incentives to support
establishment of insurance exchanges are () new
tax credits for low- and middle-income households
that could be used only for coverage purchased
through a qualified insurance exchange, () preemption from state insurance mandates (that is,
Employee Retirement Income Security Act of 
[ERISA] protection), and () protection from the
effects of adverse selection for exchanges and participating health plans.
• U.S. In su ra n ce Exchange (USIX), a national
program parallel to the Federal Em pl oyees Health
Benefits Program (FEHBP), will serve as an insurance exchange for individuals and firms with fewer
than  employees, in areas in which no private
health insurance exchange has emerged.
• Subsidies in the form of refundable tax credits for
health insurance will be available for low- and middle-income Americans who purchase qualifying
health insurance plans. In contrast to families in
higher tax brackets, such households today have
limited financial incentives to purchase private
health insurance plans.
• New financing for “default plans” and basic
health care services will be provided for low-income
individuals who are eligible for the refundable tax
credit, but who do not choose to enroll in a health
plan. Each state will receive new grants to provide a
default plan in each geographic area within its jurisdiction; people who do not choose their own health
156
plan will be enrolled automatically in the default
plan. Many states will provide new financing for
public hospitals, clinics, and other providers that
meet open-access standards, as part of their default
plan.Each state would receive a payment equal to 
percent of the new tax credits for individuals who
are eligible for such credits but who remain uninsured.States will receive incentive bonuses or reductions based on the extent to which they improve
performance of a set of preventive care measu re s
(for example, childhood vaccinations, firsttrimester pregnancy visits, hypertension control)
and reduce the percentage of the population that
remains uninsured. The goal is to ensure that every
eligible individual is enrolled in a health plan.
• There would be a phased-in cap on the currently
unlimited health insurance exclusion from taxable
income for health insurance benefits paid by
employers or individuals. Individuals eligible for
both the exclusion and the subsidy could choose
which of the two tax benefits to use.The dollar value
of the cap would be set high enough to represent a
substantial subsidy, yet low enough to provide substantial new financing for expanding health insurance coverage and other uses.
• A new, independent Insurance Exchange Commission (IEC) with narrow, specific powers would be
created to accredit insurance exchanges, conduct risk
adjustment across insurance exchanges, and serve as
a clearinghouse for public information on the quality of health plans. This agency would have an
appointment procedure and organizational structure similar to that of the Securities and Exchange
Commission (SEC), and would have a similar function—to encourage smooth information flow and
functioning of insurance exchange markets.
No single component of this proposal is likely to
achieve near-universal health insurance by itself.
Effective cost containment is essential to the expansion of health insurance in the long term, which otherwise would require prohibitively costly subsidies.
Moreover, tax credits will be ineffective if beneficiaries cannot use the credits to obtain coverage. Thus
they need access to competitive ,h i gh - va lue health
insurance plans with guaranteed issue (that is, the
requirement to offer health insurance, regardless of
the applicant’s medical history). For these reasons,
we propose to implement a coordinated policy of
targeted subsidies, consumer choice, and incentives
to offer and choose high-value health care coverage.
A summary o f the proposed subsidies and associated requirements follows (see next page).
Coverage/Eligibility: Establishing
Insurance Exchanges to Expand Plan
Choice
Central to the proposal is individual choice of subsidized plans through insurance exchanges. Ideally,
almost everyone would be covered through insurance exchanges large enough to achieve economies
of scale in brokering plans and capable of providing
information about plans to individuals and businesses. Like the FEHBP, the California Public
Employees Retirement System (CalPERS), and a
number of recently formed private purchasing
groups that serve employers, the insurance
exchanges would offer a ch oi ce among multiple
plans, with incentives for individuals to choose highvalue ones.
Advantages of such insurance exchanges
include:
• The exchanges would provide reasonably
priced coverage for the self-employed, n on - poor
unemployed, people between jobs, and employees
who currently lack access to affordable,high-quality
health insurance.
• Insurance exchanges would facilitate continuity of plan covera ge . Exchanges are likely to offer
many of the health plans operating in an area, so
that most people could keep their health plan membership when they change jobs.
• The exchanges would serve as an entry point
for low-income, uninsured individuals, who would
become eligible for substantial new subsidies to
purchase coverage.
• The exchanges would miti ga te many of the
market imperfections that plague the small-group
market (for example,through risk pooling, community rating, guaranteed issue, and competi ti on ) ,
making it easier for small employers to offer a
choice of plans.
157
S U M M A RY OF PROPOSED SUBSIDIES AND REQUIREMENTS
Eligibility Category
Proposed Subsidy
Requirements
Medicare beneficiaries
Medicare
No alternative proposed
Medicaid and S-CHIP eligibles
Medicaid and S-CHIP, respectively,
or full refundable tax credit
If refundable tax credit is chosen,
forgo Medicaid and S-CHIP
benefits and must purchase
through an insurance exchange
Low-income individuals up to
, and families up to
,
Full refundable tax credit equal to
 percent of median-cost plan
Must choose between the tax credit
and the capped exclusion; if tax
credit is chosen, must purchase
through an insurance exchange
Low-income individuals up
to , and families up to
, who do not enroll
 percent of the full refundable
tax credit, paid to state
Automatic enrollment in default
plan
Middle-income individuals
up to , and families
up to ,
Partial refundable tax credit,
phased out to 
Must choose between the partial
tax credit and the capped
exclusion; if tax credit is chosen,
must purchase through an
insurance exchange
Individuals not eligible for
the refundable tax credit
Capped exclusion
NA
Federal, s t a te , and private group-purchasing
arrangements operate in many markets today, with
varying success. Insurance exchanges will share
some fe a tu res with these organizations, but will
improve on existing arrangements in several important ways. The most important problems associated
with existing group-purchasing arrangements
include the inability to gain market share and
achieve administrative savings; adverse selecti on ,
either as a natural feature of plan competition or as
a result of regulatory and legal constraints; and
resistance to or opposition from health plans and
insurance brokers.1
From the individual’s perspective, the exchanges
offer far more choice than typically would be available today. In contrast to the current system, when
individuals enroll in an HMO through an exchange,
Elliot Wicks, Mark Hall, and Jack Meyer. “Barriers to Small-Group
Purchasing Cooperatives.” Washington: Economic and Social Research
Institute, March 2000.
1
it would be because they chose to enroll voluntarily,
rather than because they lacked an alternative insurance option. Furthermore, choices would be determined by value as perceived by the consumer, since
the consumer who chooses a plan with a higher premium than the low-priced plan will pay for the
added cost of the choice.Good information, such as
our proposal would generate, is critical to the operation of the exchanges; people are more likely to be
satisfied with their choices if they know what they
are getting.
A key challenge will be the formation and
growth of the exchanges. Our proposal inclu de s
substantial incentives to promote private insurance
exchanges.
• People are el i gi ble for the substantial new
refundable tax credits for low-income individuals
only if they purchase health insurance through certified insurance exchanges.
• Health plans offered through an insurance
exchange are exempt from state small-group market
158
reforms and laws mandating health plan benefits
for enrollees (that is, ERISA preemptions). These
preemptions are necessary to allow insurance
exchanges to form and operate across state lines, to
enable insurance exchanges to become a competitive option for employers who now self-insure, and
to give employees of small firms and the selfemployed access to flexible insurance plans that are
currently available to employees of many large
firms.
• Exchanges and health plans participating in
insurance exchanges are protected from adverse
selection (see “Insurance, Risk Selection, and Risk
Adjustment” below).
Insurance exchanges could be private or public,
for-profit or non-profit organizations, electronic or
traditional. Large and mid-size employers also
could be designated as insurance exchanges serving
employees of the firm. The IEC will certify that
insurance exchanges meet minimum standards for
eligibility for tax credits and ERISA preemption.
Exchanges would be certified through annual filings
with the IEC, which would be empowered to deny
or withdraw certification for exchanges that failed
to meet these requirements. Certified insurance
exchanges will be required to meet the fo ll owing
basic requirements:
• Non-employer exchanges must accept all individuals not eligible for Medicare and groups in their
service area (guaranteed issue) at a flat premium rate
(community rating), with adjustments only for covering additional people, such as a spouse or dependents. Beyond these requirements, non-employer
exchanges would have flexibility in formulating eligibility rules (that is, employer size maximum) and
underwriting policies (for example, waiting periods
and open en ro ll m ent practices for individuals).
Employers can also qualify as exchanges if they
accept all employees, except part-time workers, at a
flat premium rate.
• Exchanges must offer a “meaningful choice” of
plans, defined as the offer of a minimum of two
products from a minimum of two independent
companies,2 though considerably more ch oi ce s
would be desirable, including point-of-service
(POS) or preferred provider organization (PPO)
products as well as closed-panel health maintenance
organizations (HMOs) and newer alternatives such
as defined-contribution “care groups.” Such offers
must include at least one product that provi de s
some coverage for treatment by most providers in a
region (for example, a POS-type plan), and a lowpriced alternative (perhaps with more restrictive
choices or catastrophic coverage).
• Exchanges must require participating plans to
offer some standardized basic benefits to facilitate
plan comparison and discourage plans from segm en ting markets by health risk. However, plans
would be allowed to offer enhancements to the basic
features.
• Exchanges must perform at least minimal risk
adjustment (initial risk adjustment would be based
on age) and/or rely on other mechanisms to limit
the financial rewards to plans for engaging in practices that encourage risk selecti on , to preserve
choice among plan types and create incentives for
plans to enroll and care for high-cost patients.
Exchanges also must participate in risk adjustment
between insurance exchanges in a region or state.
• Exchanges must require participating plans
and providers to meet minimum standards for
measuring quality.
• Exchanges must make available comparative
information on plan benefits, pricing, quality measurement, quality improvement initiatives,and other
aspects of plan performance in an effort to help
members make informed, high-value choices.
Employers that would prefer not to fulfill these
minimum requirements can choose not to become
an insu ra n ce exchange. If they chose this option,
they would continue to be regulated by ERISA, and
they could not participate in the tax credit program.
States would continue to regulate the non-insurance
exchange market; most states have guaranteed issue
and some rating requirements in the small-group
market.
For purposes of defining “companies” and “products,” exchanges that
contract directly with provider groups would be counted as contracting
with multiple companies, as long as a sufficiently large number of different risk-bearing provider groups in an area could be chosen. For example, Buyers Health Care Action Group’s (BHCAG’s) “care groups” would
meet this definition.
2
159
Group-purchasing arrangements and many
employers that provide health insurance today meet
most of these requirements. Most offer a choice of
plans and products and provide guaranteed issue to
all participating employees. Many independent
insurance brokers and/or consortiums of brokers
could meet this requirement, as well. PacAdvantage,
a small group-purchasing organization in California, performs risk adjustment, using demographic
and administrative health data, to protect products
and plans that attract high-risk individuals. Benefits
Alliance, a group-purchasing organization for midsize employers in California, helps to en su re that
health risks are spread evenly among participating
plans by requiring that each plan offer both an HMO
and an open-network produ ct . Many such groups
offer comparative inform a ti on that is becoming
more sophisticated as it migrates toward electronic
forms. For example, California Consumer HealthScope, a web site by Pacific Business Group on
Health, provides consumers with comparative quality information and the ability to search for information about physicians and the health plans through
which they are accessible. HealthScope also provides
comparative information about drugs available
through health plan formularies.
Our proposal includes incentives to en co u ra ge
the growth of electronic insurance exchanges. Today,
such exchanges offer electronic procurement, enrollment, administration (for example, eligibility verification and bill payment), and inform a ti on about
health insurance options. For instance, Ehealthinsurance. com claims to offer products from about 
carriers and sells insurance to individuals and small
employers in  states.3 EbenX and Sageo, two additional electronic insurance exchange companies,
facilitate transactions for employees on behalf of
employers and health plans. According to industry
analysts, online individual and small-group sales
currently represent approximately  percent of total
policy sales.4 Additional e-commerce companies,
such as TriZetto and HealthAxis, provide traditional
insurers,insurance exchanges, brokers,and employers with electronic capabilities to purchase health
i n su ra n ce . Six of the nation’s largest insurers have
formed MedUnite to develop online enrollment,
physician selection, claims approval and processing,
and prescription services. Individual insurers are
developing similar capabilities.
Even in the absence of new subsidies to help
them increase en ro ll m ent and, thus, spread fixed
costs, electronic insurance exchanges and e-commerce companies may be able to achieve the administrative savings that many group-purchasing
arrangements have failed to achieve thus far.
Projected long-term administrative savings from
electronic insurance exchanges range from  percent to  percent of administrative costs.5 In a survey of large employers,  percent of respondents
indicated interest in using the Internet to help
employees enroll in a health plan or choose a physician group.6 According to industry research, many
consumers also strongly prefer to purchase health
insurance online.7
The advantages to enrollees of joining insurance
exchanges (that is, subsidies, choice, and lower-cost
coverage) likely will drive demand for such services;
this demand, in turn, will encourage entrepreneurs
and employers to seek certification as insurance
exchanges. Brokers, who have traditionally served
the small and mid-size market, but who have often
been denied the ability to form purchasing groups,
may find this a particularly attractive opportunity.
We would not prohibit insurers from sponsoring an
insurance exchange, nor would we prohibit
exchanges from becoming insurers themselves.
However, the IEC would monitor exchange sponsorship and report concern about abuses to the
Frank Cerne. “Reaching Out on the Web.” Insurance Networking
(April 2000).
Journal Sentinel (October 4, 2000).
3
4
Ibid.
5
Tara Ashish et al. “Opportunity for Health Care Savings through
Internet Technology.” Stanford, CA: Graduate School of Business.
Independent student research, May 24, 2000; Jason Gertzen. “Blue
Cross Steps into Internet Territory to Sell Health Insurance.” Milwaukee
“Despite Costs, Employers Stick with Health Benefits.” Reuters Health
(October 13, 2000).
6
7
Cybercitizen Health. “Internet Users Want to Manage Health Insurance
Benefits Online.” Press release, New York: CyberDialogue, www.cyberdialogue.com/resource/press/releases/1999/08-25-cch-insurance.html,
August 25, 1999.
160
The benefits of purchasing through the exchanges likely will make them
the predominant mode of health insurance purchase, especially for
employees of small and mid-size firms and for individuals.
Department of Justice and the Federal Trade
Commission, and could establish firewalls against
anti-competitive practices if necessary. Given the
high degree of flexibility in sponsorship and benefit
design, we expect that many individuals and
employees will use an insurance exchange to enroll
in a health plan.
Our proposal requires that at least one insurance
exchange serve every geographic region in the
United States. With new subsidies and flexibility in
sponsorship and benefit offerings, we expect existing private purchasing group arrangements to
expand to meet this requirement. The benefits of
purchasing through the exchanges likely will make
them the predominant mode of health insurance
purchase, especially for employees of small and
mid-size firms and for individuals. Most employees
of large employers will purchase through exchanges,
as well, because their employers are likely to seek
qualification as exchanges to make it possible for
low- and middle-income employees to obtain subsidies (in the form of refundable tax credits). If,
despite these advantages, insurance exchanges do
not materialize in parts or all of a state within several years, that state can work with the IEC to develop
alternatives. The IEC can waive the insurance
exchange requirement and authorize one or more
alternatives. For example, exchanges may develop
more slowly in rural areas. To serve residents in
these areas, the IEC could work with states to promote expansion of insurance exchanges specializing
in plans for rural areas to cover multiple states.
Alternatively, if no private insurance exchanges
are available in particular regions of a state by three
years after enactment of the proposal, a state can
request authorization from the IEC to implement a
national program parallel to the FEHBP, called the
U.S. In su ra n ce Exchange (USIX), in these areas.
Individuals and employers with up to  employees
would be eligible to participate in USIX; at the dis-
cretion of the states (and with authorization from
the IEC), the maximum size of participating
employers could be incre a s ed . Much like FEHBP,
USIX would be required to offer all plans in the
region that meet specified, reasonable standards
(negotiated with USIX), and each plan would set its
own price. Start-up administrative costs in each
state for USIX would be financed primarily by each
state, but states also could receive some limited federal funding. After a start-up period of,at most,several years, administrative costs would be
incorporated into the premiums charged by USIX
in the state.
Coverage/Eligibility: Coverage of
Unaffiliated Individuals through
Default Plans
Low-income individuals who are eligible for full
subsidies (see below), but who are ineligible to purchase through an insu ra n ce exchange, or who are
eligible but fail to purchase a health insurance plan
during a defined en ro ll m ent peri od , would be
enrolled automatically in a “default plan” developed
by the state.8 Individuals who are eligible for both
the subsidies and Medicaid or S-CHIP, and who fail
to en ro ll in any program or plan, also would be
enrolled in the default plan. States would identify
the default plan providers and distribute payments
to them. We expect that default providers in most
states will be public hospitals, community clinics,
and other “safety net” providers. Because states
would have considerable flexibility in targeting
default plan payments to providers, they might
8
B. Madrian and D. Shea. “The Power of Suggestion: Inertia in 401(k)
Participation and Savings Behavior.” NBER Working Paper 7682, May
2000. This research found that an automatic enrollment approach
raised pension participation rates from 37 percent to 86 percent. For
lower-wage ($20,000-$29,000) and younger (age 20–29) workers,
participation increased from 25 percent to 83 percent.
161
make other choices, such as low-cost private plans.
Default plans would be expected to meet certain
minimum standards in order to receive federal payments. For example, default plans would be
required to conduct outreach to default plan members, encourage and provide primary and preventive
services, and en co u ra ge eligible individuals and
families to enroll in private plans offered through
insurance exchanges or employers. Our plan provides for considerable state flexibility and financial
incentives to achieve these goals.
New federal payments to states for the default
plan would be set equal to  percent of the value of
the refundable tax credit (see below) multiplied by
the number of individuals presumed eligible for the
full tax credit who do not actively enroll in a health
insurance plan. A  percent subsidy for these individuals is generous compared with current payment
levels for the uninsured and their expected use of
services. A  percent subsidy also preserves incentives for individuals to join conven ti onal health
insurance plans.States that increase this group’s rate
of active enrollment in insurance plans would retain
a portion of the affected default plan payments (for
example,the payments would be reduced by  percent multiplied by the improvement in covera ge ,
rather than  percent).
The default payments to states also would
include incentive components related to a state’s
performance in providing clearly effective health
care to its population, particularly higher-risk population groups (for example, low-income children
and adults). States that improved performance
(after accounting for trends in income and state
economic performance) would receive ad d i ti on a l
incremental payments; states with worsening performance would face incremental payment reductions. Initially, measures could be selected from
health measures already being co ll ected to track
Healthy People  performance goals, which are
likely to be particularly sensitive to uninsurance
rates and/or the quality of public outreach programs. Such measures include rates of pre-natal care
(collected from birth records), vaccination rates
(obtained from state public health records), and
avoidable hospitalization rates (collected from hos-
pital discharge data). Additional measures might be
added from the Na ti onal Health Care Quality
Report Card, which probably will be published
annually beginning in several years. As a condition
for receiving default plan payments, states also
might be required to co ll ect some measures of the
services they are providing to uninsured patients.
Changes in Tax Incentives to Improve
Equity and Affordability in Health
Insurance Purchasing
Current federal tax law does not count employerpaid health premiums as taxable income for
employees. This unlimited tax exclusion has helped
to promote the purchase and availability of health
insurance, particularly employment-based insurance, but it also weakens incentives to control health
plan costs, because the added costs are in pre-tax
dollars. Over time, our proposal would transform
the unlimited exclusion into a capped exclusion.
Individuals could take either the capped exclusion
or, if they are eligible, the new refundable tax credit
(described below).
In year one, the exclusion would be capped at
double the price of the median-cost plan premium
in the previous year. (Depending on data availability, the median-cost plan premium of a representative sample of plans analyzed by actuaries working
with the IEC, or the median-cost plan of FEHBP
premiums, would be used for this calculation.) For
the next eight years, the exclusion would be cap ped
at whichever is lower: the level of the cap in the previous year or  percent,  percent,  percent,
etc., of the price of the median-cost plan premium
in the previous year (adjusted for any demographic
changes). In the th and subsequent years, the cap
would be equal to the median-cost plan premium in
the previous year, plus  percent.
Implementation will require employers to
impute employer premium payments (which would
be su bj ect to Internal Revenue Service [IRS] and
Department of Labor [DOL] audits) and to report
employer-paid premiums that they have excluded
from taxable business income (up to the cap) on the
employee’s W-. Employees then could exclude any
162
SCHEDULE FOR THE HEALTH INSURANCE EXCLUSION CAP PHASE-IN
Year
Percent of median plan premium in the previous year

 percent

Lesser of  percent, or level of cap in Year 

Lesser of  percent, or level of cap in Year 

Lesser of  percent, or level of cap in Year 

Lesser of  percent, or level of cap in Year 

Lesser of  percent, or level of cap in Year 

Lesser of  percent, or level of cap in Year 

Lesser of  percent, or level of cap in Year 

Lesser of  percent, or level of cap in Year 
 and following
 percent
additional premium payment, up to the cap on the
exclusion.
We would expect the IEC, with the technical
assistance and advice of accountants, to develop
standards for imputation of employer health care
expenses to employees. A question of regional cost
variations would ari s e .G en erally, the tax laws do not
provide for regional variations, but they could. As is
the case for other things (for example, sales taxes,
when they were deductible, or business meals), the
Treasury could publish tables with applicable caps by
ZIP code. There are arguments on both sides of this
issue. We would favor allowing regional variations
based on factor prices, analogous to those used in
the Medicare Prospective Payment System.
The cap on the exclusion will be adjusted geographically, using a formula determined by
Congress and administered by the IEC. A formula
based on insurance premium variations by metropolitan statistical area (MSA) and non-MSA regions
is a potential starting point. These geographic
ad ju s tm ents could be based on actual prem iu m
costs in the area (for example, demographically
adjusted FEHBP premiums) or on geographic cost
adjusters, such as those used in the Health Care
Financing Administra ti on (HCFA) Prospective
Payment System (PPS). If the ad ju s tm ent is to be
based only or primarily on geographic cost differ-
ences, and not on health plan cost differences across
areas after accounting for differences in input costs,
then a long transition period would be required.
The transformation of the unlimited employer
exclusion into a capped exclusion is attractive for
several reasons. First, the capped exclusion discourages employer contribution policies that inhibit cost
consciousness. Second, since the value of the tax
exclusion rises with the marginal tax rate, capping
the exclusion makes it less regressive. Third, the cap
provides a significant source of financing for the
proposed health insurance tax credits, which also
will contribute to greater equity in government subsidies for private health insu ra n ce purchases.
Without this provision, our proposal still could be
enacted. However, the new budget costs of the proposal would be significantly higher, and/or the new
tax credits would be significantly smaller or restricted to a lower income range.
Subsidies
Low- and Middle-Income Individuals and Families
Our proposal would create a new refundable tax
credit for low- and middle-income individuals and
families. The base credit amount would be equal to
 percent of the median-cost plan premium for single coverage (adjusted to the demographic charac-
163
teristics of the el i gi ble population) in the previous
year. Tax credits could be taken on a single, dual, or
family basis. We envision a credit equal to twice the
single credit for dual coverage,and . times the single credit for family coverage (that is, equivalent to
the relative single, dual, and family premiums of
many large employers). Subsidies would vary by
region, and they would be adjusted geographically in
the same manner as the tax exclusion cap described
above. Tax credits would not vary by age (see following discussion of risk adjustment of premiums).
Individuals eligible for Medicare would not be
eligible for the new tax credit. Individuals eligible for
Medicaid or S-CHIP could take the credit and enroll
in a certified exchange plan or an employer-covered
plan,though this would require that they relinquish
their Medicaid or S-CHIP benefits. Thus,individuals
or families currently eligible for Medicaid or S-CHIP
could continue to participate in these programs if
they chose, and they would have an incentive to continue, because the average subsidy is considerably
greater. However, financial disincentives to enroll in
a private plan provided by an exchange or employer
would be reduced substantially. If many Medicaidor S-CHIP-eligible individuals and families chose to
en ro ll in such a private plan, the state and federal
government likely would realize significant net cost
reductions. Given the difference in the level of subsidy between the average Medicaid and S-CHIP benefit and the proposed tax credit, this would be true
even if relatively healthy individuals switched out,as
is likely to be the case. States would be required to
maintain support for Medicaid and S-CHIP beneficiaries who enroll in a private plan, contributing 
percent of the median-cost plan.Eligible individuals
in employer-provided plans could choose either the
tax credit or the capped exclusion.
The tax credits would begin to be phased out at
incomes of , (single) and , (couples
and families). They would be phased out fully at
income levels of , (single) and , (couples and families). These amounts would be indexed
in the same way as tax brackets. The phase-out is
structured to begin at income levels above the
phase-out of earned income tax credits and most
other means-tested benefits, so that the implicit
marginal tax rate applied to an increased income
does not rise steeply.
Setting the subsidy equal to  percent of the
median-cost plan premium, rather than an alternative, such as the full price of the low-priced plan,
maintains an incentive to limit the prices of lowpriced plans. In many markets today, several health
insurance plans are available at  percent of the
median-cost plan. We assume that health plans will
offer products priced to meet the needs of this new,
large p opulation of potential enrollees, even if this
means adjusting benefits or changing cost-sharing
provisions to reduce premiums. Insurance exchanges and USIX could encourage insurers to offer
such plans. Because we propose only limited standardization of health plan benefits, setting the tax
credit as a percentage of the median-cost plan provides some assurance that the reference benefits
package provides a reasonable level of coverage.
Employed individuals would claim the credit
through an additional few lines of paperwork for
their current W- filing, which determines their tax
withholding. Individuals who do not receive the
credit through their employer (that is, individuals
who purchase through an insurance exchange that
they choose individually and, thus, to which they
pay premiums directly) would attach proof of coverage through a certified exchange to a W- to claim
the credit. States would be allowed to use some of
their uninsured funds to develop capacity to assist
unemployed individuals with the credit. And, along
with Consolidated Omnibus Budget Reconciliation
Act (COBRA) notification, those terminated from a
job would be notified of potential eligibility for this
assistance. All individuals also would be required to
provide similar documentation of coverage on their
tax return. In the absence of a presumption of eligibility, insufficient income tax withholding due to
improper credit claims would be subject to the usual
IRS interest and penalties.
To increase take-up rates,simple prospective criteria would be used to create a presumption of eligibility, based on wages and hours worked (for
example, full-time workers earning less than  an
hour who purchase a family policy would be presumed eligible), and based on last year’s family tax
164
returns (for example, workers eligible in the previous year who have not indicated new employment
or earnings on their current W- form — which
they would need to fill out to claim the credit in
advance — also would be presumed el i gi bl e ) . Individuals who meet the eligibility presumption, but
who turn out to be ineligible in their end-of-year tax
filings, would not be subject to penalties; they simply would be required to repay the amount credited
on a reasonable repayment schedule. Precedent for
this policy comes from the IRS, which forgives tax
penalties for insufficient withholding for those
whose tax liabilities are much higher than they were
the year before. Some state health insurance programs for lower-income families, for example,
Wisconsin, also have implemented steps similar in
spirit to this presumption.
For employed individuals, credits would be
transferred directly to employers. The credit would
be made available to employers in advance for
individuals with insufficient tax liability. Making
these payments to firms rather than individuals (as
in Earned Income Tax Credit [EITC] payments)
would improve accuracy and simplify administration for the IRS. Also, relatively few employers
would owe negative taxes as a result of any incorrect
payments.
Certified insurance exchanges also could collect
credit-based payments monthly or quarterly on
behalf of the individuals enrolled through their
exchanges via the IEC. We expect this option to be
used most often by individuals (along with their
families) who do not receive coverage through their
employers. The IEC would receive an annual appropriation from Congress for this purpose. In this
case,the certified insurance exchange would submit
proof of coverage on behalf of the eligible individual. Thus, these individuals would have to pay only
the portion of their premium not covered by the
credit. Such individuals would not be allowed to
receive the credit through their employer also. The
IEC would forward the tax identification numbers
of all individuals receiving subsidies through insurance exchanges to the IRS for audit purposes, and
would assist the IRS in detecting other types of
fraud. The IEC and the IRS would have the authori-
ty to impose sanctions on exchanges that use tax
credits fraudulently.
Although recent studies have suggested a significant fraud rate with the EITC, it seems unlikely that
fraud will occur at a similar rate here. First, to be eligi bl e , an indivi dual must doc u m ent covera ge in
a certified exchange. The number of certified
exchanges is likely to be small compared with the
number of individuals and businesses filing tax documents related to the EITC, thus creating simpler
oversight. Second, sanctions against exchanges —
including losing certification and criminal penalties—presumably would deter their participation in
fraud. Third, in contrast to dollars, individuals are
not likely to want more than one health insurance
plan.
Administra ti on of the tax credits will require
employers to collect some ad d i ti onal information
from their employees. However, the process represents a relatively minor addition to the information
on wages and other benefits employers currently
provide to the federal government, so this should
create a relatively minimal burden. No other private
entity is as well situated to provide this needed information about employees, and any additional preparation costs could be deducted as a business expense.
Payments to Default Plans
Payments to states for individuals enrolled in their
default plan will be administered by the IEC. The
default plan payment will be equal to  percent of
the value of the tax credit for eligible individuals in
the state who do not actively enroll in an insurance
plan. These payments, and the associated incentive
payments, were described earlier.
Insurance, Risk Selection, and Risk
Adjustment
Our proposal attempts to strike a balance between
protecting health plans and insurance exchanges
from adverse risk selection and stifling innovation, variation, and flexibility.
Because we have attempted to minimize complexity by maintaining tax credits that do not vary
with risk status, and by fixing premiums that indi-
165
Critical to this proposal is an ongoing quality measurement
and public reporting program, including risk-adjusted
outcome studies and comparison of actual care patterns
with recommended guidelines.
viduals pay when they purchase insurance through
an exchange, there is substantial potential for
adverse selecti on . Some aspects of our proposal,
such as the large subsidies available up to relatively
high-income levels, will mitigate adverse selection
by attracting large numbers of average and low-risk
enrollees. However, additional efforts are likely to be
necessary. Consequently, the IEC will develop minimum standards for risk adjustment of plan premiums within insurance exchanges, and the IEC will
provide risk adjustment among insurance exchanges
in a particular region or state.
Among Plans within Insurance Exchanges
A key ch a ll en ge is to ensure that plans do not face
financial penalties for attracting enrollees who are
likely to have above-average health expenditures, or,
conversely, are not rewarded for attracting low-cost
enrollees. An ideal risk-adjustment procedure would
remove the disincentives to attract high-cost
en ro ll ees without rewarding health plans whose
costs are high because they are inefficient or unable
to limit use appropriately. Obviously, such an ideal
system does not exist;thus considerable flexibility in
dealing with risk-selection problems within exchanges is desirable.
Because of variation in plan features, such as the
groups of participating providers, services reimbursed, and breadth of choice of prescription drugs,
different plans within an exchange are likely to attract
enrollees who would be expected to generate different
levels of expenditures. Insurance exchanges would be
required to meet minimum standards for risk adjustment of payments to their participating plans, based
on differences in the expected use of populations of
enrollees that they attract. The IEC would specify
minimum standards. Individual exchanges and states
would be free to use additional methods, such as partial reinsurance. Initially, the risk adjustment is likely
to be based on age alone. As risk-adjustment technology improves, and as experience with other methods
accumulates, the IEC may implement alternative
standards. Note that within-exchange risk adjustment will redistribute payments within the exchange
from lower-risk to higher-risk plans; it does not
include any cross-exchange subsidies.
Among Insurance Exchanges in a Region or State
The IEC would oversee risk adjustment across insurance exchanges in each region or state. The IEC
would develop methods for measuring selection
effect s ,b a s ed on data provided by health plans and
exchanges. If necessary, the IEC would con du ct
demographic risk ad ju s tm ent and, possibly, more
sophisticated risk adjustment, to redistribute premiums among insu ra n ce exchanges. In effect, plans
with higher-risk demographics would be subsidized
by lower-risk plans in the same region. The IEC also
would have some authority to work with states to
adapt high-risk pools and other state initiatives to
the insurance exchange program. Adverse selection
could affect exchanges, despite these measu re s ,i f , for
example,non-exchange employers encouraged sicker employees to seek coverage through exchanges as
individuals. Exchanges also could attract high-risk
en ro ll ee s , who are more likely to be sick, simply
because they are open to individuals. Consequently,
the IEC would monitor adverse selection between
insurance exchanges and the non-exchange market
a n d ,i f necessary, would recommend the inclusion of
the non-exchange market in the risk-adjustment calculations. An alternative solution could be to provide stop-loss protection for employers functioning
as exchanges that had very high-cost employees.
Note that for these purposes, all employers that
qualify as insurance exchanges would be included in
the risk-adjustment calculations and prem iu m
redistribution.
166
STRUCTURE OF THE INSURANCE EXCHANGE COMMISSION
IEC Steering
Committee
Coverage
Committee
Risk-Adjustment
Committee
Oversight of Insurance Exchanges: The
Insurance Exchange Commission
The federal Insurance Exchange Commission (IEC)
will be created to oversee the proposed new subsidies
and the insurance exchanges eligible to benefit from
them. Its broad mission is to help the market provide
access to high-quality health care; however, its powers for achieving this goal would be relatively narrow
and tailored to en su ring that insurance exchanges
and the competition they foster function effectively
and with minimal intervention. The IEC would be
an independent agency structured like the Securities
and Exchange Commission. An appointed board of
directors, whose members would be selected for
their professional qualifications, would serve for
fixed, staggered five-year terms. Board members
would be appointed by the President, with the advice
and consent of the Senate.
A Steering Committee, made up of individuals
with experience and expertise in health care financing and organization, would direct the IEC’s activities. In addition, the IEC would operate four standing committees: () Coverage, () Risk-Adjustment,
() In su ra n ce Exchange Opera ti ons and Compliance, and () Quality Measurement and Improvement. Committee membership would include indi-
Insurance Exchange
Operations and
Compliance
Committee
Quality Management
and Improvement
Committee
viduals from the payer and provider communities,
industry (including pharmaceutical and device
manufacturers), consumers,and health care experts.
Some of the IEC’s responsibilities could be contracted out to other agencies, such as the Agency for
Healthcare Research and Quality (AHRQ), or private-sector organizations, such as the National Committee for Quality Assurance (NCQA). Committee
members would meet usual conflict-of-interest standards for senior government officials.
Among the main functions of the IEC would be
distributing tax credit payments toward premiums
to insurance exchanges and distributing default
plan payments to states. By assigning these responsibilities to the IEC,our proposal minimizes new burdens for the IRS.
The four standing committees of the IEC,
shown in the chart ab ove, would have the following
responsibilities.
The Coverage Committee would issue recommendations and set minimum standards for benefits covered by health plans offered through certified
insurance exchanges. The minimum standards will
be designed to ensure that par ticipating plans cover
medical goods and services that are known to be
effective and that are provided at reasonable cost,
but they also will be sufficiently general and flexible
167
to allow plans to create a wide range of coverage
options.
The Risk-Adjustment Committee would be
responsible for developing and implementing new
approaches to risk ad ju s tm en t . They would be
expected to draw on a wide range of expertise and
consult broadly in developing and testing new methods. Although risk adjustment will be limited to simple age adjustment initially, the Risk-Adjustment
Committee will review existing and developing riskadjustment methods on an ongoing basis, test such
methods, and implement the best ones, based on
their feasibility and their ability to overcome adverse
selection within and among insurance exchanges.
The Insurance Exchange Operations and Compliance Committee would en co u ra ge development of
insurance exchanges, and would develop and
administer incentives to create and con ti nue the
exchanges. It also would establish and enforce minimum standards for the formation and operation of
insurance exchanges to en su re that they serve the
interests of m em bers. The new federal minimum
standards would replace state laws for plans offered
by insurance exchanges.
This committee would certify private insurance
exchanges as eligible to receive subsidies. Based on
proposals from affected states, it also would be
responsible for ensuring the development of at least
one insurance exchange to cover every geographic
region in the country. In addition, it would make
sure that residents of regions that failed to establish
exchanges within three years would be able to enroll
in health care plans through USIX.
The committee also would be responsible for
monitoring market concentra ti on and detecting
abuses of either monopoly or monopsony power
that an insurance exchange might develop. The committee also would have the ability to obtain price
information from the exchanges to detect evidence
of abuse of m on op s ony power, such as contract
prices with plans that fall well below the prices paid
to plans in other markets. The committee would
monitor exchanges for abuses by sponsors who also
offer insurance through their exchange. It would
provide information on such questionable competitive conditions to the Department of Justice and the
Federal Trade Commission. To date, group-purchasing arrangements and similar en ti ties have not
accounted for a large share of i n su red lives in any
geographic area, so concern about market power is
based on the potential growth of the exchanges,
rather than on current problems.
We would not prohibit insurers from sponsoring
an insurance exchange, nor would we prohibit
exchanges from becoming insurers themselves.
However, the IEC would monitor exchange sponsorship and report concern about abuses to the
Department of Justice and the Federal Trade
Commission, and could establish firewalls against
anti-competitive practices if necessary.
The Quality Measurement and Improvem en t
Committee would establish minimum quality measurement and reporting standards for health plans
participating in insu ra n ce exchanges and for those
acting as default plans. Health plans would report
quality data directly to the IEC. Insurance exchanges
also might be required to report some measures of
quality, for example, disenrollment, complaint,and
satisfaction rates. The IEC would ensure that such
data could not be used to compromise individual
patient confidentiality and would provide these data
for use by government agencies, consumer groups,
consultants, benefit managers, and others in evaluating the quality of insurance products and exchanges.
The committees’ operating budget would be
determined and appropriated by Congress. Staffing
for the IEC would include the IEC director and
seven members of the steering committee, full-time
chairs for the four standing committees, and fulltime staff supporting the steering committee and
each of the standing committees. The budget also
would include funds for operations, an annual budget for development activities, and incentive funds for
exchanges. The operating budget,as shown below, is
approximately  million annually.
Financing the Proposal
Costs of the proposal include the new tax credits for
low- and middle-income Americans, ad d i ti on a l
payments to the IRS to administer the tax aspects of
the plan (with assistance from the IEC), limited
168
ANNUAL BUDGET* OF THE INSURANCE
EXCHANGE COMMISSION
Staffing
(s)
Steering Committee
Director and  members
Salary + full benefits
Senior (professional) staff ()
,
,
Standing Committees
Average . FTE per committee member
Director and  members of each
,
committee
Senior (professional) staff
,
( for each committee)
Administrative staff ( total)
,
Office expenses
,
Travel
,
Risk-adjustment development
Quality measurement and improvement
,
,
Exchange Incentives Fund*
,

,
*Initial-year budget; will be reduced in subsequent years.
start-up costs for USIX,and an operating budget for
the IEC. These costs will be financed by:
• capping the current unlimited exclusion of
employer-paid health insurance;
• savings over time through behavioral effects
among consumers and health plans because of
increased cost-consciousness and improved valuebased competition among health plans; and
• general revenues.
Fit within Existing System and Transition
from Present to Future
This proposed plan for near-universal health insurance relies primarily on existing institutions, and
preserves the best features of the existing health
insurance system while closing gaps in covera ge .O u r
proposal retains the employer-based system and the
option for all those eligible for low-income state and
federal health care programs to continue in them if
they so desire. Our proposed default plan would
improve support for “safety net” providers and provide new incentives for preventive services that
might help lower their costs. This proposal also
builds on what we know works best today and is
most acceptable to the American public, that is,
large-scale group-purchasing arrangements such as
FEHBP and CalPERS, and independent agencies
with limited authority to help competition work well
in complex industries, such as the SEC. Competition
among private health plans and among private
insurance exchanges with minimal federal oversight
fits American values be s t .E conomic theory also suggests that this is the most effective way to expand
ch oi ce ,l ower prices, and improve quality of care and
service. Where such arrangements fail to develop,
our proposal offers USIX as a backup program that
also would create competition and choice.
Where we recommend a significant departure
from the current system, we propose a gradual transition. In particular, we recommend a -year trans i ti on period to adjust the tax exclusion of health
insurance premiums fully. We also recommend a
minimum of three years for development of private
insurance exchanges before introducing USIX in a
market.
Political Feasibility
Opposition to such a plan is likely to emerge from
these major features:
. The cost of the plan is substantial; the bulk of
its costs come from the tax credits used to subsidize
insurance coverage. Furthermore, some opponents
would claim that a tax credit mechanism like the one
proposed here is “inefficient” because it provi de s
new subsidies for many low- and middle-income
families that are currently purchasing private health
insurance. The claim of inefficiency is easily misunderstood, and applies only in the narrow sense of
government expenditures for health insurance. The
system of tax credits proposed here gives tax credits
to low- and middle-income individuals, regardless of
whether they already have insurance individually or
through their employer, as long as the employer is
certified as an insurance exchange. Even non-eligible
169
employers may encourage low- and middle-income
individuals to seek coverage outside the firm to avail
themselves of the tax credit. The purpose of giving
tax credits to all low- and middle-income individuals is to compensate those low- and middle-income
individuals and families who purchase insurance (or
obtain it through an employer), giving up either premium payments or wages to do so. Offering the subsidy only to those who lack insurance is,in our view,
short-sighted and inappropriate policy, since it
strongly en co u ra ges employers to drop coverage,
thus “crowding out” private insurance and distorting
incentives. Moreover, it is unfair to the many families
that con ti nue to struggle to make their monthly
health insu ra n ce premiums. Funds spent on lowand middle-income individuals under this plan are
transfer payments, and they provide social benefits
even if the low-income individuals would have purchased insurance in the absence of the subsidy. The
cost per additional insured individual may be higher
than under plans that crowd out private insurance,
but the added costs represent socially desirable
transfer payments (like the EITC) rather than waste.
This tax reduction also would achieve a second
goal—improving health insurance coverage while
reducing incentives to switch to even more heavily
subsidized government plans. Educating the public
and policy makers about the reasons for the costs
and the advantages of the plan will be challenging.
. The plan limits the tax exclusion for health
insurance expenditures. The current tax exclusion is
popular, and particularly benefits high-income
individuals and families who purchase high-cost
plans. Such people will not favor the cap on the
exclusion, which reduces the overall cost of the plan.
We believe that the gradual phase-in of the cap, and
pegging the cap to the cost of a reasonably representative and generous private health plan, will help to
overcome objections to it.
. The plan creates an independent federal
agency to oversee health insurance. Some critics will
object to creation of the IEC, claiming that it will be
another federal bureaucracy that imposes undue
burdens on employers,health plans, and health care
providers and provides poor service to the public.
Some also may object to what they perceive to be the
IEC’s complexity. However, the IEC’s authority is
limited to a small set of specifically designated powers, and its main functions are to assist in implementing the tax credits and developing better data
on risk adjustment and health plan quality. In setting minimum standards for use of the tax credits,
the IEC will play a relatively hands-off role, setting
standards more like those used by the FEHBP, rather
than, for example, the very detailed recommendations of President Clinton’s  health care task
force. At the same time, by extending the ERISA preemption to insurance e xchanges and plans that will
serve many small employers and individuals, we
remove an inequitable set of regulatory burdens that
currently face small employers and individuals seeking coverage.
. The plan will not provide coverage for every
single American. Although our goal is universal
health insurance coverage, this plan does not contain a mandate to cover every American. Thus,those
who seek immediate universal coverage will object
that this plan falls short. We believe that plans that
propose  percent coverage typically do not
achieve access for all individuals, and that attempts
to do so require some combination of high costs,
restrictions on choice of plans or providers, limited
coverage, and a constrained role for the private sector. We also believe that our proposal to offer new
funding to states to provide “basic” and preventive
care to individuals who do not choose to enroll in a
subsidized plan (and so may be less likely to use
health care until they really need it) will improve
their access to care, as well. We believe that the political objections to a plan that would come closer to
immediate universal coverage would be far greater
than to our plan.
. Groups that favor either a single-payer system
or another form of comprehensive government
interven ti on in health care will oppose the heavy
reliance on existing private institutions. Some of
them will argue that administrative costs will be
lower with a government-run plan, and that equity
in access to health care will be put at risk by a plan
that promotes choice among private health insurers.
We believe that many of those groups, however, will
see the appeal of our plan, which achieves much
170
broader health insurance covera ge , despite their
objections to specific features, and that this proposal will enjoy far broader political support than the
alternatives they favor.
. The plan will change the mechanism through
which safety net providers receive payment. Under
this proposal,safety net providers would receive payments for care for the uninsured by participating in
default plans. In addition, to receive payments,states
will be required to meet minimum requirements and
will have financial incentives to conduct outreach to
facilitate enrollment of unaffiliated individuals in
private plans. The payments outlined are more generous than safety net providers receive today, and
they promote high-quality care and expanded coverage. However, many safety net providers like the current system and may resist change.
Overall, the plan represents little threat to existing interests and little change in familiar institutions
and structures. Its key features—preserva ti on of
private health insurance, expansion of choices and
flexibility, use of targeted tax credits, improvement
of markets for individual insurance, and protection
of safety net providers like public hospitals—are
also features of plans that already have demonstrated bipartisan support. Like those plans, it does not
attempt to solve every problem in the current health
insurance system. Because our proposal represents a
fine balance between the achievable and the ideal, it
has the potential to attract broad political support.
Quality, Cost, and Efficiency
Critical to this proposal is an ongoing quality measurement and public reporting program, including
risk-adjusted outcome studies and comparison of
actual care patterns with recommended guidelines.
Much of this effort would be conducted or promoted by the IEC, which would set standards for information collection and dissemination. It would build
on existing best practices in quality measurement,
whether public- (for example, Medicare) or privatesector (for example, the Consumer Assessments of
Health Plans Study [CAHPS] and the Foundation
for Accountability [FACCT]), and would sponsor
development of new measures. It also would
encourage standardization of data collection across
exchanges and other public or private purchasers.
These efforts would help plans and providers to
develop improved measures of processes and outcomes of care by plans and providers. Finally, the
IEC would develop and enforce standards for data
security and confidentiality.
Also critical to this proposal is effective cost containment, requiring conditions in which all
Americans have a personal reason to care about
health care costs, support serious cost-containment
efforts, and economize. This is partly a matter of
cultural attitudes, but it is also a function of the economic structure of health plan choice. Our strategy
for containing costs and increasing value rests on
competi ti on among health care organizations to
serve price-sensitive consumers.
Recognizing wide variations in quality and economy of health care delivery systems, the proposed
economic structure of health plan ch oi ce would
encourage greater cost consciousness, and the proposed quality measurement program would increase
consumers’ confidence in their choices. We anticipate that informed and cost-conscious consumers
will migra te , gradually and voluntarily, to models
that offer the greatest value, as they have done in
exchange models such as FEHBP and CalPERS.
Equity
Use of the tax credit to subsidize low-income individuals and families promotes equity by direct
income transfer. Qualification for the tax credit is
broad,and it is available to low- and middle-income
individuals who are not enrolled in Medicare,
Medicaid, or S-CHIP. Individuals with qualifying
income who are now adequately insured,along with
those who are uninsured or underinsured, would
qualify for the tax credits. In addition, the middleincome individuals who would purchase insurance
if they could do so at costs similar to those available
to groups would now be able to obtain insurance at
more favorable rates via the insurance exchange.
Making the tax exclusion for health insurance universal will lower the after-tax cost to those lowerincome individuals who currently must pay health
171
insurance premiums with after-tax dollars. The cap
on the tax exclusion will affect primarily highincome individuals who purchase very generous
health insurance policies.
An additional protection for at-risk populations,
most of which are lower-income,is explicit designation of the default plan. We believe that most default
plans will provide access to county and other public
hospitals, which will then have a reliable source of
revenue that can be used to subsidize the costs of
uncompensated care. Such institutions routinely
provide care to individuals who are uninsured and
unable to pay for medical services,and undoubtedly
have mitigated the adverse consequences of lack of
insurance. The continued viability of such institutions is critical unless and until it is possible to
ensure that all Americans can receive care at other
institutions.
An additional source of equity concerns is the
treatment of persons with chronic illnesses and others who are expected to have disproportionately
high health expenses. Many such individuals now
face exceedingly high health insurance premiums or
may not be able to purchase insurance at all. The
proposed system emphasizes development of riskadjustment mechanisms for health insurance premiums that would make community rating feasible
and thus enable such high-risk individuals to purchase insurance at the same rates as other people.
Overall,the proposed plan provides substantial protections to those who are unable to obtain insurance
at reasonable rates, either because their incomes are
too low or their expected medical costs are too high.
Acknowledgments
This paper was prepared in consultation with Mark
B. McClellan of the Department of Economics and
the Center for Health Policy, Stanford University. n
173
Weil Proposal
Key Elements
Alan R. Weil proposes creating a Medical Security System (MSS) to provide
health insurance coverage to every legal resident under age . His plan would:
   -  access to a basic health coverage
package at no cost to them, by requiring employers either to provide coverage or pay a payroll tax to finance coverage purchased through insurance
exchanges.
    designed to organize the insurance market. The exchanges would operate in defined geographic areas
and contract with health plans that would offer all their products on a
guaranteed-issue, community-rated basis.
      through a health insurance
exchange or their employer. At their option, employers could opt out of the
MSS financing system by providing part- and full-time employees with a
health plan equal to or greater in value than one of the standard benefits
packages.
        ’  Insurance Program (S-CHIP)
and the low-income and adult components of Medicaid, folding those
beneficiaries into the MSS, while waiving copayments and providing
wraparound coverage for services not included in the basic package.
       , existing government funding sources, individual
premium payments, and additional appropriations for financing.
174
About the Author
  . , .., ..., is Director of the Assessing
the New Federalism project at the Urban Institute.
This project—the largest in the Institute’s -year
history—monitors, describes and assesses the effects
of changes in federal and state health, welfare, and
social services programs. Mr. Weil was formerly
Executive Director of the Colorado Department
of Health Care Policy and Financing. This cabinet
position is responsible for Colorado’s Medicaid
and Medically Indigent programs, health policy
development, and health data collection. Mr. Weil
was a member of President Clinton’s Advisory
Commission on Consumer Protection and Quality
in the Health Care Industry and served on Mrs.
Clinton’s health care task force. He was the lead
Democratic staff during negotiations of the 
National Governors’ Association policy on Medicaid
reform and the  NGA policy in support of universal health insurance. Mr. Weil received his bachelor’s degree in economics and political science from
the University of California at Berkeley. He holds a
master of public policy degree from the John F.
Kennedy School of Government at Harvard
University, and a J.D. from Harvard Law School.
175
The Medical Security System: A Proposal To Ensure
Health Insurance Coverage For All Americans
by Alan R. Weil
System Overview
This paper sets forth a proposed Medical Security
System (MSS) that would provide health insurance
coverage to all non-elderly Americans.
System Design
The Medical Security System combines funds from
a payroll tax, existing government funding sources,
and additional appropriations to provide a basic
insurance plan to all individuals. Private health
insurance exchanges structure the insurance market, allowing individuals to select a basic plan at no
charge or pay an additional amount to obtain
enhanced covera ge . A medical savings account
option is also available. Employers can avoid the
payroll tax (their own and their employees’) if they
provide and pay for a significant portion of a comprehensive insurance option for their employees
and their dependents. All legal residents, regardless
of employment status, who do not obtain coverage
through their employer have free access to a plan
through an exchange, and can purchase higher-cost
coverage if they desire. Waivers of cost sharing are
available to low-income families through a stateadministered system.
The Medical Security System is financed primarily through a payroll tax. Tax rates are set originally
to approximate current private insurance spending.
For illustration, tax rates of . percent for employers and . percent for employees applied to the
Social Security wage base would generate approximately the amount spent on private insurance premiums and would distribute the costs between
employers and employees in accordance with the
national average contributions for family coverage.
A portion of existing state and federal Medicaid and
State Children’s Health In su ra n ce Program (SCHIP) funds for poverty-related eligible populations, along with additional general revenues, will
also be required.
A set of nationally standardized model benefits
packages will be developed (for example, tightly
managed with limited copayments,less tightly managed with higher copayments). One benefit option
will be structured as a medical savings account
(MSA), with a high-deductible health plan combined with mandatory contributions to a savings
account that can be used only for medical costs. The
standard benefits package with the lowest actuarial
value (other than the MSA) is referred to as the
benchmark pack a ge . The value of the benchmark
package determines the revenue necessary to provide universal coverage under the MSS. This
amount, less other funding streams, will ultimately
determine the actual payroll tax rate.
The MSS operates through health insurance
exchanges that serve to organize the insurance market. Exchanges operate in defined geographic areas
designed to encompass one or more health care
markets. Exchanges are assumed to be private, but
there are no requirements as to their form of governance or ownership. Any exchange that meets basic
requirements can obtain a license to operate,meaning that multiple, competing exchanges may exist
within any given market. Exchanges have no regulatory power.
Exchanges bear no insurance risk; instead, they
contract with licensed health plans. To operate, an
exchange must offer all standard benefits plans;
however, exchanges may also offer additional plan
designs. All products must be offered to all partici-
176
pants on a community-rated basis. At least one plan
must be offered free-of-charge. Through the MSS,
exchanges receive a fixed amount per enrollee
adjusted for the age and gender composition of the
en ro ll ee s . On average, the MSS provides each
exchange with sufficient funds to offer the benchmark plan at no cost to the enrollee. As a practical
matter, the specifics of the no-cost plan in any
exchange will be determined by the bids the
exchange receives. It is important to emphasize that
the actual design of the no-cost plan may or may
not match the benchmark plan, and it may vary
across exchanges. Exchanges also co ll ect premium
payments from individuals who select enhanced
plans, and they provide information to participants
about their plan options.
The notion of a health insu ra n ce exchange is
borrowed loosely from the stock exchange. Stock
exchanges create a marketplace for highly regulated
goods (securities). To succeed,these exchanges must
attract buyers and sellers (firms and shareholders).
While exchanges impose significant contractual
requirements on buyers and sellers,they do not have
any regulatory authority themselves. This analogy
has its limits, but it does suggest the type of role
envisioned for the exchanges.
Employers may be exempt from the payroll tax
s ys tem if they provide all part-time and full-time
employees with a health insurance pa ckage equal to
or greater than one of the standard benefits packages described above. Employers must contribute a
minimum of  percent of the cost of individual
coverage and  percent of the cost of dependent
coverage, and employees are required to participate
in coverage offered by exempt employers. Employer
exemptions are given for three-year periods and
cannot be revoked during that time. The continuation coverage requirement under the Consolidated
Omnibus Budget Reconciliation Act of 
(COBRA) is eliminated.
Every legal American resident, regardless of
work status or earnings, may obtain any no-cost
insurance package through any exchange operating
in his or her area (unless the resident is covered by
an exempt employer). At the time of enrollment,the
person may select a plan that requires a premium
payment, in which case premium billing is handled
by the exchange. Non-payment results in the
exchange moving the person into a no-cost plan.
Enrollment cycles are annual, with exceptions for
people who move into or out of the region. Health
care providers may arrange with exchanges and
health plans to provide for enrollment at the point
of service for anyone who has not gone through the
enrollment process. Any person may elect the medical savings account option, but the election is irrevocable for five years.
Families with income below current mandatory
Medicaid eligibility income standards qualify for a
copayment waiver, allowing them to obtain medical
services without making copayments or deductibles
and to obtain certain otherwise uncovered services.
Waiver applications are processed by states based on
recent earnings data. Copayment waivers are valid
for one year and must be renewed annually.
The low-income child and adult components of
Medicaid are eliminated, as is the State Children’s
Health In su ra n ce Program (S-CHIP). Medicare
remains intact. The portion of Medicaid that serves
people with disabilities is reconfigured as wraparound coverage beyond the basic benefits available
through the MSS.
System Principles
The Medical Sec u ri ty Sys tem is built on three
principles.

Health insurance coverage should be universal.
Putting this principle into effect requires two
important conceptual shifts within the health care
system.First, we must acknowledge that no amount
of public covera ge , whether in the form of public
programs such as Medicaid or financial support
such as tax credits,can expand from the base of our
existing voluntary, employer-based insu ra n ce system into a universal system. Efforts to build from
the voluntary employer-sponsored insu ra n ce base
have yielded many innovative approaches, such as
sliding premium scales, waiting periods for enrollment in public programs, and tax credits to purchase coverage in the non-group market or to apply
177
toward the employee’s share of the health insurance
premium. While each incremental step is important
and can benefit many people, this approach cannot
reach universality. Substantively, the two systems
cannot mesh to provide true universal covera ge ,
and, politically, sharing financial responsibility
between the government and employers without
clearly defined roles for each creates a constant battle to shift costs to the other payer.
Second, we must separate financing from enrollment. A variety of financing sources must be called
on to support the health care system. In addition,
effective enrollment mechanisms must reach all
people. But the two systems must be separated, so
that income tests and applications with cumbersome verification requirements are no longer barriers to program enrollment. Similarly, we should
encourage the separation of enrollment and health
benefit administration from the workplace, thereby
facilitating more efficient labor markets and eliminating employers’ access to personal health information about their employees.
 
The MSS is built around the critical American value
of choice, which is available at three level s .F i rs t ,i n
contrast to the circumstances facing most Americans
today, individuals are able to choose their health plan
and delivery system. Secon d ,i n d ividuals are able to
choose their health care provider. They can choose a
plan based at least in part on whether that plan
includes the providers on whom the individual relies.
They may always obtain services outside of the MSS
if they are willing to pay for those services. Third,
individuals can determine the level of financial risk
they are willing to bear. Individuals can choose from
a range of insurance structu re s ,f rom catastrophic to
comprehensive. They pay according to the risk they
are willing to take, and they are able to join a risk
pool with others with similar tastes for risk.
  
The political process will determine the parameters
of the benchmark plan and the funds necessary to
support universal access to that plan. The definition
of the guarantee is fundamentally a social decision
that should be made in the political arena. The combination of individual choice and competi ti on
among health plans will yield efficiencies within
that core system. Spending decisions beyond the
core are made by individuals, without tax subsidies,
reflecting their desire to obtain more health insurance coverage than the MSS provides. The MSS
retains the third-party payment system. While
third-party payment is inherently inflationary, it
provides a very real value to people by reducing the
financial risks they face.
Coverage and Eligibility
The Medical Security System is universal for legal
residents of the United States under age . The system fully uncouples financing and enrollment,making a basic health insurance package available to
every American, regardless of employment status or
income. Some people will continue to receive their
insurance coverage through their employer. Others
will obtain coverage through new health insurance
exchanges that structure the market for health
insurance. Everyone will have an insurance option
available at no charge, but will also be able to use his
or her own money to purchase a higher-cost, more
comprehensive product.
Design
  
All Americans (except those covered by an exempt
employer) will have the option of obtaining health
insurance through new entities called health insurance exchanges. A health insurance exchange is a
market organizer for health insurance products. The
exchange receives funds (as described below) and
enters into contracts with multiple insurance companies. Exchanges can take any ownership or governance form (for example, private, public, corporate,
not-for-profit). Health insurance exchanges are
licensed by the federal government.
Each exchange must offer all standard plans (as
described below) and at least one health plan that
can be obtained free-of-charge by the enrollee. The
exchange may also offer additional benefit designs.
All plans, standard or otherwise, must be offered at
178
All Americans will have the option of obtaining health insurance
through new entities called health insurance exchanges. Each
exchange must offer at least one health plan that can be obtained
free-of-charge by the enrollee.
a community rate to all exchange participants with
no underwriting. Exchanges may not tie participation in the exchange to purchase of any other goods
or services,and exchanges may not sell any goods or
services other than health insurance.
The federal government will define a set of geographic regions, known as catchment areas,
designed to approximate health care markets. An
exchange that wishes to operate within a catchment
area must serve the entire area. A single exchange
can serve one or more catchment areas, or even
operate nationwide. There is no restriction on the
number of exchanges that may operate; in fact,
competing exchanges may operate in any given
catchment area. An exchange operating in a catchment area must accept enrollment from any resident
in that catchment area.
Exchanges contract with health plans and pay
them for each enrollee in the exchange who selects
that plan. The exchange bundles the funds it
receives through the Medical Security System with
the premium contributions made by individuals
and passes them along to health plans.
An exchange that performs its functions well—
contracts with a variety of plans, informs enrollees
of their choices, handles premium co ll ections—
should obtain a sufficient enrollment base to be
financially self-sustaining. However, since the MSS
cannot function without the exchanges, states will
be required to create an exchange if none exists in
the state.
 
A national board will create a series of standard
benefits plans. These plans will reflect a range of
options with respect to delivery system (for example, tightly managed,loosely managed,unmanaged)
and cost sharing. As noted above, all health
exchanges must offer all standard plans.
Based on the funds available to it through the
MSS, each exchange must also offer at least one nocost plan.A no-cost plan is one in which any person
may enroll without being required to make any premium payment from his or her own funds. A nocost plan may or may not correspond with one of
the standard plans. As a practical matter, the
exchange, knowing the funds it will receive per
member, will solicit bids from participating health
plans at a price that matches those funds (less
administrative costs absorbed by the exchange). The
benefit design of the received bids will determine
what the exchange can offer at no cost. Given this
method, it is important to note that the specifics of
the no-cost plan may vary among exchanges.
Every exchange must offer an MSA option.
Under this option,a high-deductible plan is provided, with the balance of the funds placed into a savings account. Funds in the account roll forward
indefinitely and are available only for medical costs.
The MSA option is available to all MSS participants,
but a participant who elects that option must
remain in the MSA for five years.
No rebates are permitted for below-cost plans;
that is, exchanges may not offer plans that cost less
than the funds they have available, and then refund
those excess funds to the enrollee. The only exception to this is the MSA. In the MSA option,all funds
other than those used to purchase the highdeductible insu ra n ce are deposited into the MSA,
which is available to the en ro ll ee only for healthrelated purposes.
 
Employers may con ti nue to provide health insurance to their employees. An employer that provides
all part-time and full-time employees with a health
insurance package equal to or greater in value than
one of the standard benefits packages, and con-
179
tributes a minimum of  percent of the cost of
individual coverage and  percent of the cost of
family coverage, can become exempt from the MSS.
This exemption means that the employer does not
pay the payroll tax into the MSS. Exempt employers
must make employee participation in the company’s
health plan automatic and mandatory. Employer
exemptions are granted for renewable three-year
periods, and the employer must abide by the terms
of the exemption for the entire period.
Rationale
The MSS relies heavily on health insurance
exchanges because of the demonstrated value of
pooled purchasing arrangements. These arrangements pool risk, offer a choice of plans, and organize the insurance market in a manner that increases
competition. The limited success of purchasing
pools in the current health insurance system is primarily a reflection of the environment in which
these pools must opera te . The MSS offers these
pools a large number of members and a set of market rules under which they do not face any competitive disadvantages. Under these conditions, health
insurance exchanges can provide value and choice.
In the MSS,health insurance exchanges can take
any organizational form, and any number of
exchanges can exist side-by-side with identical or
overlapping catchment areas. There is no reason to
restrict organizational form or create monopsony
purchasers as long as all exchanges are required to
fo ll ow appropriate market rules. Competi ti on
among exchanges will occur on the basis of information provided to enrollees,the range of insurance
options offered to enrollees, and the overall quality
of service provided. In the short run, exchanges will
be created and some will fail, imposing a cost on the
health care system. Over time, a limited number of
exchanges are likely to survive. The process of
exchange competition, and the possibility of new
entrants if existing exchanges do not meet the needs
of their customers, should yield good-quality service. A natural corollary of having multiple, competing exchanges is that those exchanges have no
government or regulatory power.
The series of standard benefits packages is
designed to enhance the efficiency of the market.
The goal is to create sufficient standardization that
consumers can evaluate the relative value of various
options. One alternative would have been to permit
only standard benefits plans to be sold. This option
was rejected because the process that defines these
plans will inevitably be political and potentially
slow-moving and could prevent the adoption of
innova ti ons (such as the emergence of point-ofservice plans).
To ensure the viability of the standard plans,the
MSS requires that all benefit designs offered by the
exchange, standard or not, be offered on a community-rated basis without underwriting. One alternative would have been to permit risk rating for
non-standard products. This option was rejected
because of the concern that a health plan could offer
a minor va ri a ti on on a standard plan, underwrite
the plan, and offer it at a low cost while higher-risk
populations are placed in the almost identical standard plan. This would yield the sort of risk segmentation the MSS must avoid to be successful.A more
appealing option would be to permit underwriting
for any plan with a premium of, for example, more
than  percent above the benchmark plan, as long
as the plan is guaranteed renewable. This option
would allow people who want to bear less financial
risk to enroll in a more comprehensive plan without
facing the risk that premiums for this plan will be
artificially high because of people switching into the
plan at the last minute when they anticipate needing
health care services. If stable, such a structure could
segment people with a low tolerance for financial
risk without fully segmenting health care risk. This
option was rejected for its complexity, but is worth
more explora ti on as a vehicle for offering more
insurance options. Nothing in the MSS explicitly
bars the con ti nu ed existence of the non-group
insurance market. If state regulations permit it, that
market could offer Medigap-style coverage without
conforming to the precise requirements of the MSS.
In the MSS, neither exchanges nor health plans
are permitted to refund to the individual any funds
that may be available because the enrollee selected a
low-cost plan. This provision is designed to create a
true floor for insurance covera ge . This provision
180
does limit the extent of health plan competition in a
very low-cost market. That is, since rebates are not
available, all an efficient plan can do is add benefits,
which may have an inflationary effect on overall
health care spending. Despite this risk,the provision
was adopted to en su re that competitive pressure
does not result in poorer people, who may prefer
cash in hand to better insurance covera ge , from
becoming concentrated in a very low-cost, very lowquality plan that refunds a large portion of the premium to the enrollee, but offers little in the way of
health insurance protection.
The MSA option offers a realistic insurance
option for people who are willing to take significant
personal financial risk, while limiting the likelihood
of risk segm en t a ti on . Strong opponents of MSAs
argue that people willing to accept financial risk
have disproportionately high incomes, and, because
income is correlated with health status, they are
likely to have lower-than-average costs. Removing
this population from the larger risk pool yi el d s
higher premiums for everyone else. Strong proponents of MSAs argue that they are the only realistic
mechanism for creating price sensitivity and full
choice of provider among individual users of health
care services. While MSA proponents claim that
MSAs have not shown great success in the market
because of the regulatory burdens associated with
the existing MSA demonstration, evidence from the
rest of the health insurance market suggests that the
more likely reason is that only a relatively small portion of the population is willing to bear the amount
of financial risk inherent in the MSA st ructure. The
MSS includes an MSA option because it is believed
that the cost to the rest of the population is small
compared with the gain that will accrue to MSA
participants. The five-year lock-in requirement for
MSAs is designed to reduce their risk segmentation
aspects,although the requirement will not eliminate
such aspects entirely.
Employers may continue to offer health insurance benefits to their employees even after establishment of the MSS for the same reasons that they
choose to do so today: to gain a competitive advantage when recruiting employees, or to pursue particular wellness goals among their workforce.
Employers should be encouraged to play this role if
they can offer their employees benefits that go
beyond those available from health exchanges. Some
of these benefits could include improved health
plan selecti on , oversight, and information, or
reduced employee cost sharing.
Employees working for exempt employers are
barred from the health insurance exchange system.
This provision exists to prevent employers from
en co u ra ging their highest-cost employees to shift
their costs to the publicly financed program, either
through the structure of their employee subsidies or
through more direct pressure on the employee. This
restriction on employee choice is unfortunate, but it
is necessary to protect the integrity of the system.
The existence of this restriction provides a strong
justification for enforcing the standards imposed on
employers that wish to be exempt. It also creates a
need for individual premium subsidies, discussed
below.
Financing
The Medical Security System relies on three sources
for financing. The primary source of funding is a
payroll tax. Additional government revenues supplement the payroll tax, and individual prem iu m
contributions by some participants also finance the
system.
Design
 
The payroll tax forms the core of MSS financing. It
is designed to emulate the current system by having
employers co ll ect the portion paid by employees
and combine that amount with the employer’s contribution. However, as a tax, the system operates in
an equitable manner by requiring the participation
of all firms, in contrast to the current voluntary
structure.
The precise structure of the payroll tax is not
critical to the design. To illustrate, a financing system is presented with the split between employer
and employee contributions divided  percent/
percent, approximately the same division that exists
today for family coverage at the t ypical firm. Again,
181
for illustration, the proposal applies the tax to the
current Social Security wage base. A combined
employer/employee tax rate of  percent applied to
the Social Security wage base in  would yield
approximately the  billion spent on private
health insurance premiums that year. While participation in part of the Social Security system is currently optional for state and local government
employees, that would not be the case for the MSS,
thereby providing a larger tax base. Consistent with
how Social Security taxes currently opera te , selfemployed persons would pay the combined
employer and employee rate.
 
Public financing beyond the payroll tax will be necessary to generate sufficient funds to provide coverage to all Americans. If the payroll tax genera te s
approximately the resources currently spent on private coverage, additional resources will be needed to
cover those currently without health insurance and
those covered through public programs.
One source of public financing is a portion of
current expenditures on Medicaid and the State
Children’s Health In su ra n ce Program (S-CHIP).
Since coverage for these populations comes through
the MSS, appropriations to these programs can be
redirected to the MSS. Some current public expenditures will be needed for special subsidies, discussed below. Thus, current spending on Medicaid
and S-CHIP, less the amount needed to provide
other subsidies, will be available to fund the MSS.
 
The MSS provides every individual with the opportunity to participate in any health plan offered by
any health exchange. However, only some insurance
options available through the exchange will be available at no cost. For higher-cost plans, individual
contributions will be necessary. These funds will be
paid to the exchange and passed on to the appropriate health plan.
Rationale
The MSS fully uncouples financing from enrollment
in health insurance. Thus, the MSS could function
using any funding source, ranging from a portion of
general tax revenues to a per capita assessment. The
decision to rely primarily on a payroll tax is an effort
to balance various implications of these different
funding mechanisms.
The payroll tax has the advantage of emulating
the current system of financing health insurance reasonably closely. Most non-elderly Americans receive
coverage through work, meaning that we already rely
on an employment-based financing system. Despite
the view of most economists that employees pay the
full cost of insurance provided through the workplace, as a practical matter employees perceive that
they share costs with their employer. To minimize
disruption,the MSS emulates this division.
The MSS includes a major shift by imposing a
flat tax rate on a capped portion of earnings. This
stands in contrast to how health plans charge firms,
which varies by employee age and family structure,
but not by income. It also stands in contrast to how
employees observe their payments, which are generally based only on family structure, with no variation in employee (or employer) contribution based
on the employee’s salary. How the underlying costs
of health insurance are actually borne by employees
within a firm depends at least on labor market conditions, the complex tax treatment of health benefits, and, probably, on the market for the firm’s
goods. It is impossible (and not necessarily desirable) to determine the precise incidence of health
care costs today and create a tax system that emulates it. In the MSS, the goal is to approximate the
current structure while achieving the added goals of
equity and universality.
Setting aside the problem of disrupti on , and,
therefore, of likely political opposition, some would
argue that all funding for the MSS should come
from general revenues, generated from the relatively
progressive federal income tax. This funding option
has the advantage of greater vertical equity (higherincome people pay a much larger share of total
income taxes collected than they do of total Social
Security taxes collected). It also presumably has less
of a negative effect on job creation than a significant
payroll tax. Yet, it is difficult to imagine this large a
shift in the financial burden associated with health
182
insurance. The salary base of the MSS payroll tax is
capped for the same reason: concern that a new
financing system with dramatically increased costs
for high earners is not politically viable.
All payments into the MSS are based on earnings
and decisions about the type of coverage an individual desires; none is based on the enrollee’s health
status. Some people believe this is an inefficient
design, because it reduces the financial incentive for
people to adopt healthful behaviors. The MSS
rejects incorporating health status into the financing mechanism, based on the notion that other
financial consequences associated with unhealthful
behavior remain,and that individuals should not be
penalized for incurring health care costs that are not
attributable to individual behavior.
Total funding for the MSS is determined by the
cost of providing the benchmark plan to all eligible
people. The amount of revenue the payroll tax generates will vary with the cycles of the economy. Since
general tax revenues must make up the balance, the
MSS design creates some uncertainty in projecting
future federal budget demands.Given the size of the
MSS, even a modest degree of error could have significant implications for the overall federal budget.
The MSS retains a significant role for individual
expenditures for health insurance. This design feature is important for two reasons.First, while health
insurance exchanges can exert some bargaining
pressure on health plans and providers to hold
down costs, real pressure for efficiency will come
from individuals selecting from among their plan
options to obtain their preferred combination of
price and quality. Second, to be affordable,a universal coverage system must guarantee to everyone a
level of coverage that is less generous than some
people will desire. The individual purchasing option
allows those who wish to spend more than average
on health care to do so.
Relationship between Financing
and Benefits
From among the standard benefits plans, the one
with the lowest actuarial value is termed the benchmark plan. The total resources necessary to pay pre-
miums for all Americans en ro ll ed in the MSS are
determined by multiplying the cost of the benchmark plan (including health care and administrative
costs) by the number of Americans under age ,
less those covered by exempt employers. The premium do ll a rs are allocated to each health exchange
based on the number of enrollees in that exchange,
with allocations weighted to reflect the age and gender mix of each exchange’s enrollees.
Knowing the per capita premium dollars available, each exchange contracts with participating
health plans. The exchange must offer at least one
no-cost plan, so the exchange must solicit bids that
exactly equal the per capita premium dollars available. By design, the no-cost bids on average should
precisely match the benchmark plan. However,
depending on health care costs in that market and
the efficiency of the plans that operate there,the nocost bids could be for a plan somewhat ri ch er or
somewhat leaner than the benchmark plan. Plans
will also bid to offer higher-cost benefit designs,
with the individual enrollee required to pay the cost
differential between the benchmark and the highercost plan.
It must be emphasized that the MSS opera te s
with a defined contribution design. That is, while
various benefit packages are defined,and the level of
funding is designed to be sufficient to cover the
benchmark package for everyone, the actual coverage available to any person will be whatever health
plans can offer, given the premium payments they
receive on behalf of their enrollees.
Other Subsidies
Two groups will find that a standard package of
insurance benefits does not meet their needs. One is
made up of people with very low incomes for whom
the modest copayment and deductible provisions of
a standard package create barriers to access to care.
The second group includes people with disabilities
who need services that go beyond those provided by
traditional health insurance benefits. The MSS
includes sep a ra te subsidy systems to address the
needs of these two groups. In addition, the tax system subsidizes low-income employees of exempt
183
employers whose premium burden may be more
than they can afford to bear.
Design
Since at least one plan in every exchange is offered at
no cost,there are no premium subsidies to individuals based on income. However, in the MSS, states
will administer a system that permits individuals in
low-income families to obtain a copayment waiver,
allowing them to obtain medical services without
making copayments or meeting deductibles.
Eligibility for waivers is guaranteed for everyone
with income below poverty and all other current
mandatory Medicaid eligible populations, but states
may adopt broader eligibility standards.Eligibility is
for one year and must be renewed annually.
While individuals who qualify for the waiver will
likely be concentrated in no-cost plans, some may
choose to enroll in plans that charge premiums.
These plans could have larger cost-sharing requirements than the no-cost plan. Complete forgiveness
from cost-sharing requirements could create a
windfall to the enrollee and add incentives for plan
design and selection. Therefore, the waiver will be
designed based on its application to the benchmark
plan. For qualifying individuals in other plans, the
waiver will be a credit against cost sharing designed
to have the same actuarial value as the complete
waiver for the benchmark plan. All administration
of the cost-sharing waiver will be the responsibility
of the health plans and the state, with no financial
risk to or involvement of the provider or the
exchange.
Under Medicaid, low-income children are currently el i gi ble for a broader range of health care
services than are found in a typical private insurance plan. As a part of the MSS,all children in families that receive the cost-sharing waiver will also
receive state-administered fee-for-service coverage
for that additional set of services.
People who qualify for Medicaid on the basis of
disability currently receive a set of medical and supportive services that are not found in private insurance plans. This portion of the Medicaid program
will continue to exist. However, since everyone will
be eligible to participate in the MSS, the disability
component of Medicaid will function as a wraparound to the core set of benefits.
Financing for all of these subsidy programs will
retain the existing Medicaid matching structure
between the federal government and the states.
Matching funds are available to states that extend
copayment subsidies beyond the mandatory populations. Matching funds (for premium costs only)
are also available if a state chooses to offer premium
subsidies to low-income individuals, even though
the MSS has no provision for such subsidies.
Beyond these requirements, states may use their
own funds to offer whatever subsidies they wish to
low- and moderate-income individuals or people
with special health care needs.
Low-income employees of exempt employers
are required to participate in their employer’s health
plan. While an exemption is available only if the
employer pays a large portion of the premium, the
employee may still face a significant financial burden to pay for coverage—a burden that low-income
employees whose firms are not exempt do not face.
Therefore, the MSS includes a refundable tax credit
for these employees. The credit would cover  percent of the actual cost to the employee of enrolling
in the employer’s plan for people with adjusted
gross incomes below the poverty line, with the
amount of the credit phasing out to zero when
income reaches  percent of the poverty line.
Rationale
The goal of these provisions is to ensure that lowincome populations can participate in the mainstream health care delivery system. Underlying the
MSS design is the notion that all Americans should
have access to a good health insurance plan,and that
empirical evidence shows that all Americans are sufficiently price sensitive that a relatively heterogeneous group of people will select a no-cost plan.
Cost-sharing exemptions and wraparound benefits
permit low-income populations to participate in the
same health plans as the rest of the population.
These provisions create some administrative complexity and raise concerns about coordination of
benefits and service delivery systems. However,
those risks are viewed as less substantial than the
184
risk of an inadequate or substandard health care
system for low-income people and/or for people
with disabilities if these groups are segregated from
the general system.
The decision to fold Medicaid and S-CHIP into
the MSS is sure to be controversial. There has always
been a tension within the Medicaid program based
on whether it should seek to provide mainstream
health care to its enrollees or cultivate relationships
with a limited set of providers that have the experience and cultural competence to serve a financially
disadvantaged population. A case can be made for
either approach; however, in the context of true universal coverage, the former approach is more sustainable and defensible. In a universal system, a
special set of rules for a needy group will always be
under attack. The MSS seeks to eliminate these
boundaries and keep the en ti re population in one
system. It is likely that within the MSS, groups of
what we now think of as safety net providers will
form health plans, as they have done in the
Medicaid program, to compete for the population
accustomed to obtaining services from them. This is
consistent with the benefits of keeping the en ti re
population in a single financing system.
Of particular importance is the decision to have
no direct special subsidies of premiums for lowincome individuals. Isolating such subsidies from
the rest of the health care system creates serious
risks that the subsidy will decline over time or in
times of budget pressure, leaving this group in a de
facto segregated system. It creates arbitrary distinctions among people with different incomes who
must be deemed deserving of the subsidy, often on
the basis of very limited information about their
real health care needs or the real costs they face purchasing health care. If they are large, such subsidies
can also create the risk of significant work disincentives as income increases and the subsidy is phased
out. Such subsidies also create tremendous administrative burdens.
The decision to include a refundable tax credit
for low-wage workers in exempt firms is an effort to
make the best out of a difficult challenge. For riskselection reasons, employees in exempt firms are not
permitted to participate in the health exchange.
However, they will then face premium costs that
may be prohibitive. Subsidizing employees based on
their salary is inefficient, because it adds an administrative burden for the employer, and an individual’s salary may not reflect his or her family
resources. Therefore,the tax system is used to transfer funds to truly needy families who are required to
purchase coverage through their employers.
The System as Viewed by Various Actors
Individual Enrollees
Individuals may choose to obtain their coverage
through any exchange operating in their area, and
they may choose from among any of the plans
offered by the exchange. There will always be at least
one no-cost option, but other plans will require a
premium payment on the part of the consumer.
Most people will presumably take active steps to
select an exchange and a source of insurance coverage from among the various options available. However, some people will fail to fo ll ow this process.
Health care providers will have the option (and the
strong financial incentive) to enter into an arrangement with one or more exchanges and health plans
to permit people who present for services and are
not yet enrolled in a plan to be enrolled at the time of
servi ce . Unlike existing public programs, the MSS
has no recertification or reapplication processes.
Enrollment in an exchange and a health plan are
continuous until an active step is taken to enroll elsewhere. Therefore, over time, enrollment in the system will be essentially universal.
While this paper uses the term,“individuals,” to
refer to enrollees,most people will enroll as families.
A series of issues arises when a family has one member who works for an exempt employer, and another member who either does not work or works for a
firm without an exemption. I have not attempted to
work out the details of the family choices and cash
flows in such a circumstance; however, this is an
important area for further work.
Employers
Employers must choose whether to seek an exemption. The system is designed to make exemption a
185
In a universal system, a special set of rules for a needy group
will always be under attack. The MSS seeks to eliminate
these boundaries and keep the entire population in one system.
reasonably attractive option, as large employers can
serve as innovators in the health care system.
Employers may also wish to offer a higher level of
benefits than the MSS does, and there should be no
barriers to a firm that wishes to do so.
Non-exempt employers must collect and pay the
payroll tax. This function will be similar to the
activities currently performed in the Federal Insurance Contributions Act (FICA) system, and will
add administrative and financial burdens to firms.
Exempt employers may contract with health plans
or self-insure, as they may do today.
Health Plans
Insurance plans face a very different environment
under the MSS than they do today. The MSS brings
into the insurance system the tens of millions of
Americans who are currently uninsured and whose
demand for health care services is not fully known.
Guaranteed issue and community rating for all
products is an environment unfamiliar to most
plans. Contracting with many potential health
insurance exchanges creates new responsibilities.
The possibility of large swings in plan enrollment in
the early years creates significant risks. At the same
time,plans operating in a universal-coverage system
have a large new pool of potential customers.
Disruption could be significant in the short run.
The va ri ety of new risks taken on by health plans
comes at a cost, suggesting the need for relatively
generous assumptions about program financing in
early years.
After a transition peri od , it is reasonable to
expect that the MSS actually would reduce the risk
health plans face. Churning of en ro ll ees should
decline substantially as insurance provision is
uncoupled from the workplace. Continuous insurance coverage should reduce the cost health plans
face because the build-up of demand for servi ce s
during spells without insurance should disappear,
and increased disease severity due to delayed medical treatment should decline. Movement toward
community rating should force plans that rely heavily on underwriting out of the market, leaving a
more heterogeneous risk pool for the remaining
plans.
In the early years, plans can expect their contracts and nego ti a ti ons with health insurance
exchanges to be similar to current practices among
purchasing pools. For example, payments to the
exchanges through the MSS will be adjusted for age
and gender, so plans will submit bids to health
insurance exchanges that emulate the age and gender categories used by the MSS. If plan bids are
higher than the amount available through the MSS,
the balance will be charged to the individual
enrollee as a premium.
Over time these relationships may change. For
example, one could imagine exchanges developing a
cost-neutral risk-adjustment payment system for
health plans. If effective, such systems would benefit
the plans, since premium payments received would
reflect expected costs more accurately. The value of
this benefit would be reflected in lower bids for coverage through the innovative exchange, attracting
more en ro ll ees and ultimately ben ef i ting the
consumer.
Federal and State Governments
The federal government has five responsibilities in
the MSS. First, it must collect payroll taxes. With a
structure parallel to that used in FICA,it seems natural to give this task to the Social Sec u ri ty
Administra ti on , or a partner agency operating in
conjunction with the SSA. Tax receipts would function in essentially the same manner as Social
Security receipts.
Second, the government must define distinct
geographic areas that serve as the catchment areas
for health insurance exchanges. This is an extremely
186
complex and important task. The goal is to balance
two competing interests: On the one hand, all
insurance products are community rated within
the area, and, as discussed below, payments to
exchanges may be adjusted by region. Therefore,the
areas must correspond somewhat with existing
health care market boundaries. On the other hand,
exchanges benefit from scale and a minimum number of borders with other areas. This argues for
relatively large regions.
Third, the government must license health
insurance exchanges. While the formal requirements for exchanges are limited, they can create
trem en dous problems if they fail, misappropriate
funds, or use risk-se gmenting behaviors. Therefore,
oversight of the exchanges will be required.A federal license will be necessary to receive any funds from
the MSS.
Fourth, the government must define the process
for making payments to licensed exchanges. Among
other items, the government must determine the
relative payments to be made for enrollees in each
age and gender grouping used to define payments.
Fifth, the government must grant exemptions to
qualifying firms.
State governments have roles, as described
above, in operating subsidy systems for low-income
enrollees and people with disabilities. State governments are responsible for establishing one or more
health insurance exchanges if the private market
does not do so. In addition, states retain their existing authority to license health plans that operate in
their states and enforce market con du ct , solvency,
and network adequacy rules.
Health Care Providers
As a modification of the financing system, the MSS
has no direct effect on health care providers, but
many implications flow from the system. Under the
MSS, levels of uncompensated care should fall significantly, affecting the financial strength of many
health care institutions. By uncoupling financing
from enrollment, changes in health insurance coverage should be less frequent, thereby creating more
stable medical and financial relationships between
patients and providers. The MSS raises the possibil-
ity of increased consolidation in the health plan
market with its attendant benefits of reduced complexity, but perhaps lower reimbursement ra te s
because of the reduced relative negotiating power of
providers.
System Calibration
While the paper thus far has set forth the general
structure of the Medical Sec u ri ty System, certain
details are crucial to the program’s success. In addition, in a dynamic economy and health care environment, the MSS must have built into it balancing
tensions that will ensure the system’s proper functioning. This section examines in more detail three
aspects of the MSS: the benchmark benefits package,the balance between payroll tax and other funding sources, and regional variation.
The Benchmark Benefits Package
The various standard benefits packages exist to create a series of comparable options that facilitate an
effective market for health insurance. These packages will be most helpful in this role if they represent typical plans that people obtain today through
their place of employment or in the individual market. Beyond that, the details of each package are
unimportant, because plans are free to offer any
benefit design. Failure to specify these plans well will
have limited effect on the overall MSS.
The one standard package that serves as the
benchmark plan plays a very substantial role. The
cost of this plan for an average-risk person multiplied by the number of people in the MSS determines the total funding necessary to support the
system. On avera ge , the benchmark plan is what
participants in the MSS will receive at no cost.
Proper definition of the benchmark plan is critical for political reasons. It is not credible to propose
a universal health insurance plan where the guaranteed coverage is far below what most people experience today. Even if, in the end, the payroll tax were
set lower than the current contribution toward coverage through employer-sponsored insurance, and
individuals had sufficient resources left over to buy
their way into a plan comparable to what they have
187
today, the politics of a scaled-back standard package
would be unacceptable.
An adequate benchmark plan is also critical for
substantive reasons. If the benchmark plan is too
lean, most Americans will want to purchase a richer
plan, leaving the lowest-income group that cannot
afford to pay for premiums out of their own pockets
segregated in inadequate, no-cost plans. An inadequate benchmark plan also creates concerns about
risk selection, discussed below.
Similar problems arise if the benchmark plan is
too rich. The political problem is that the tax levels
necessary to support the MSS will be high—more
for many people than they are paying for premiums
today. Substantively, an excessively rich benefits
package will stifle competi ti on among plans and
allocate society’s resources toward health care above
the value those resources provide.
Between these two bounds, the appropriate level
of benefits in the benchmark plan is a matter of
social choice: How much of the cost of health care
should be spread across all Americans, and how
much should be borne by the individual? At the outset,the benchmark plan and its actuarial value must
be determined through an administrative process.
But the political process must be relied on to maintain the appropriate balance between the extremes.
The benchmark package is very visible. Failure of
the MSS to provide most Americans with benefits at
least as good as the benchmark package will create
political pressure to expand the resources available
to purchase the package. Health plans will compete
to offer the most desirable no-cost plan, helping to
maintain the strength of the package available to
enrollees at no charge. Yet, as the package becomes
richer, there will be strong pressures to reduce the
payroll tax. The universality of the system, in terms
of benefits received and taxes paid, helps prevent
significant erosion or expansion of the benefits.
Balance Across Funding Sources
Most firms that choose to be exempt are likely to
make this decision based on negotiated benefits in a
labor contract or a desire for greater control over
benefit costs. However, two groups of employers
could select exempt status for simple financial rea-
sons: firms with a healthier-than-average workforce
and firms with higher-than-average salaries. For the
first group, insurance purchased on the open market will cost less than average. Since the payroll tax
assessment is based on broad averages,there is a cost
advantage to having an exemption. For the second
group, the payroll tax assessment will impose a
higher-than-average burden that can be avoided if
the firm purchases its own coverage.
For both groups,the problem is the gap between
the cost of participating in the payroll tax-based system and the benefits the firm’s employees gain from
the insurance provided in that system. Of course, a
firm takes on substantial administrative costs when
it administers its own health insurance benefits, and
the three-year lock-in for the exemption means the
firm must be confident that its status is fairly stable.
Still, if a significant number of firms are in either of
these positions,they could start a death spiral in the
MSS in which lower-risk and higher-revenue populations withdraw, leaving high-cost and low-revenue
people in the system.
The best mechanism to prevent this dynamic is
to set the payroll tax to collect only a portion of the
program’s costs and rely on general revenues to fill
in the balance. Under this design,a firm that chooses to be exempt avoids the cost of the payroll tax, but
gives up a benefit that is greater than the cost to the
average firm. The proper mixture of revenue
sources is difficult to determine in advance. If a large
number of employers seek to change their exemption status, that may be a sign that the system is out
of balance, and the funding mixture needs to be
adjusted.
Regional Variation
As described thus far, the funding provided to
health insurance exchanges for the average-risk
enrollee is the same around the country. While the
system could work this way, the result would be a
no-cost plan that is much ri ch er than the benchmark plan in low-cost regions, and a very lean nocost plan in high-cost areas. This outcome has
political and practical problems. One solution is to
adjust payments to exchanges by region based on
va ri a ti on in health care costs. This would lead to
188
more comparable benefits around the country. An
alternative is to adjust payment to exchanges based
on general salary levels. This provides regional equity, since each region in essence keeps the funds it
raises through the payroll tax.
While either solution assures some comparability of benefits around the country, neither addresses
a more complex issue: the possibility that people in
different regions place a different value on health
care spending. These different values could reflect
the wealth of the region—people in wealthier
regions may feel they can devote more of t h ei r
resources to health than to other priorities.
Alternatively, these values could reflect the efficiency
of the health care system, with people in some
regions feeling that an inefficient health care delivery system does not yield solid value for the money.
If there were an effective political process that could
capture these regional va ri a ti on s , it would make
sense for the payroll tax rate to vary by region to
reflect these differences. As a practical matter, this
process could be undertaken by states, but state
boundaries generally do not correspond with health
insurance market boundaries. Therefore, while
desirable in theory, it may not be possible in practice
to design a system that varies taxing and spending
levels around the country.
Thus, the MSS as proposed provides areaadjusted payments to health insurance exchanges.
This is not a perfect solution to the dilemmas discussed above, but it does assure a reasonable
amount of equity until a better design emerges.
Other Issues
Risk Selection
One purpose of a universal health care system is to
spread the risk of health costs across a broad population. If the system leads to a concentration of relatively high- or low-risk groups into one health plan,
the benefits of that risk sharing are lost.
One opportunity for risk selection is through
s el ective marketing of insurance products and/or
benefit packages designed to appeal particularly to
certain low-risk populations. Through these techniques, a health plan, or a plan in conjunction with
an exchange, could direct low-risk populations to
specific plans where they would have low premiums, forcing higher premiums on participants in
other plans.
Three features of the MSS are designed to reduce
this risk. First, all products offered by all exchanges
must be community rated and must be offered on a
guaranteed-issue basis. At the outset, this feature
prevents explicit underwriting that would guide
only certain populations into certain plans. In addition, it creates instability for any successful risk segmentation endeavors. If an effort to select favorable
risk succeeds,the benefits of that selection are available to anyone who wishes to enroll. This reduces
the incentive for plans to select, because their benefits are transitory, and it reduces the likelihood that
such efforts would succeed even if they were to be
undertaken.
Second, a plan is limited in how much financial
reward it can provide its enrollees as a result of successful risk selection. That is,no rebates are available
to program participants, so a plan with very low
costs can reward its low-risk enrollees with lower
premiums only to a certain point. But that point is
not likely to be particularly low, because the benchmark plan will represent a decent benefits package
for a mixed-risk population. Beyond that point, all
the plan has to offer its enrollees is richer benefits—
precisely the sort of behavior that runs counter to its
risk-selecting objectives. Of course, the plan can
benefit from selection, by making profits on the gap
between the premiums it receives and the health
care costs it incurs. But it is not clear why even
healthy people would choose a lean benefits package
over a richer one. Thus, even though plans have an
incentive to obtain low-risk enrollees,plans will find
it difficult to target this population.
Third, size requirements argue against extensive
risk-selection activities. It is reasonable to project
that a modest number of fairly large health plans
will dominate each health care market under the
MSS. These plans require such a large number of
en ro ll ees to succeed that it is difficult to imagine
them having significant, effective risk-selection
efforts. Smaller plans, to succeed, will need to differentiate themselves. While marketing based on net-
189
work quality or service might be successful for a
small plan,it is difficult to imagine a successful risksegmentation campaign on behalf of a small plan.
A related concern is that a health insurance
exchange could become a front for an aggressive
risk-selecting health plan. While complying with
exchange requirements by offering a full range of
plans, the exchange could market very selectively,
perform some informal underwriting activities, and
channel low-risk people to a preferred health plan.
As with the plans themselves, exchanges would find
that the potential rewards of this behavior are quite
small, yielding very little incentive to seek out lowrisk enrollees.
Aside from inten ti onal efforts by plans or
exchanges to find low-risk enrollees, selection can
occur as a natural event when enrollees are offered
different health plans. People with greater health
care needs have good reason to select plans with
more comprehensive benefits. Evidence from the
Federal Employees Health Benefits Program
(FEHBP) and other large employers confirms that
health plans offering more comprehensive benefits
experience adverse risk selection.
The best buffer against these risk-selection
problems is to develop and maintain a strong
benchmark plan. The benchmark plan must represent good enough coverage that a heterogeneous
profile of individuals will enroll in it. While it is
unavoidable that some share of higher-risk individuals will seek more comprehensive covera ge , the
incentives to do so must not be so great that they
result in total risk segmentation in the market.
Annual open enrollment creates some barrier to
people moving to richer plans when they anticipate
having higher health care costs. However, beyond
annual enrollment and a good benchmark, the MSS
has no formal mechanism to prevent risk selection
or to compensate plans or enrollees for risk selection
taking place. Despite the likelihood that this will
occur, no obvious alternative presents itself. Given
the state of the art in risk-adjustment techniques,the
administrative complexity of operating such systems
for a large population, and their susceptibility to
political manipulation,I conclude that no mandatory risk-adjustment system should be imposed as part
of the MSS. (The incentive for health insurance
exchanges to develop risk adjustment systems voluntarily is discussed above.) If the MSS succeeds in
providing universal coverage at a cost to the individual that is primarily, but not entirely, separated from
his or her personal risk of incurring health care
costs, it will represent a significant step forward that
can be improved on over time.
Transition
The MSS is a comprehensive insurance system that
differs fundamentally from the current health care
system in the United States. As such, movement to
the MSS would require a substantial transition. The
MSS was designed to illustrate a model, not to offer
a straightforward path from the current system to
one like it. There are two incremental steps that
could be taken, however, that would represent significant movement toward the MSS, thereby
smoothing the prospects for the transition.
First, incremental steps could be taken to
en co u ra ge the creation of entities like health care
exchanges. Specifically, new requirements could be
placed on employer-sponsored health insurance
payments for them to retain their tax-exempt status.
Modest steps would include requirements that
employers offer standardized benefits packages or a
ch oi ce of plans. A more substantial step would be
permitting the tax exemption only if firms purchased insurance through some sort of a pooled
purchasing arrangement. Alternatively, if a new system of individual tax credits for health insurance
were developed, those credits could be made larger
for people who purchase insurance through a pooled
arrangement.Of course, obtaining support for these
measures would be difficult, but they would begin to
restructure the market in a manner consistent with
how it would function under the MSS.
Second,incremental steps could be taken to create a more equitable financing system for employersponsored health insurance. Employers could be
rewarded for offering a larger subsidy to their lowerwage employees than they do to their higher-wage
employees. Employers could be required to pay at
least a specific portion of the cost of insurance for
any of their expenditures to retain their tax-exempt
190
status. Employers that do not provide health insurance to their employees could face other requirements, such as a higher minimum wage.
Setting aside incremental steps, a few features of
the MSS are designed to minimize the disruption
involved in the transition. Specifically, the level and
split of the employer and employee payroll taxes are
designed to emulate the dominant practice in the
marketplace. The use of age- and gender-adjusted
premium payments to exchanges, and then passing
them along to health plans,is designed to retain the
approximate structure of risk in the group insurance market. The employer exemption provisions
are designed to encourage continuity among
employers that are adding significant value to the
health care system, not only by providing their
employees with coverage, but also by promoting
quality, choice, and information.
These provisions, however, leave a substantial
transition burden. Three major (and many minor)
areas of risk arise in the transition.
First, as noted above, in the early years of the
MSS substantial risks are associated with premium
and enrollment levels for health plans. Aside from
the cost implications of these risks,there are broader system implications that create the possibility of
program failure. The system will require sizable
administrative systems to manage the many tasks
associated with the program.
Second, projections of tax revenue could be erroneous. Financing certainty is essential for effective
opera ti on of the system. Health plans must know
what resources are available when constructing their
bids. The government may need to use general
appropriations to cover the possibility of errors in
this area. However, very large potential costs and
much budgeting uncertainty are associated with the
government taking on this responsibility.
Third, the MSS shifts the cash flow of hundreds
of billions of dollars in the health care system. Rather
than making direct, monthly payments to health
plans, employers will pay taxes to the government,
which will transfer those funds to exchanges, which
will then pass the funds to health plans. These additional steps introduce time and, given the sums
involved ,s i gnificant cash flow costs.Once the system
has been operating for a time, cash flow expectations
can be adjusted. However, at the outset, the potential
delay in cash flow could create significant costs.
Cost Containment
The MSS relies on two types of forces to contain
costs. Political forces will determine the funds devoted to the guaranteed benefits package, with tax rates
and appropriations to the program set through the
political process. The program is designed to create
broad interest among the general population that will
ensure its continued political support. The expectation is that a large number of Americans will enroll in
no-cost or very-low-cost plans, thereby ensuring that
the value of the benchmark plan does not erode. In
addition, the benchmark plan will form the base for
people purchasing more expensive coverage, so all
participants will have an interest in retaining a solid
base. At the same time, aggressive competition
among plans to serve that group will ensure that the
benchmark plan is valued appropriately. With a
defined contribution design,there is no direct translation between loading up the benchmark benefits
package with additional services or providers and
having those providers receive any benefit.
Market forces will determine people’s willingness to pay for coverage beyond the basic package.
This will be a large, contested market with plans
charging premiums ranging in cost from zero to
substantial. Competition among plans to provide an
attractive no-cost plan will be intense. The limited
evidence available today suggests that employees at
all income levels are quite sensitive to price when
selecting health plans if they have to bear the full
cost of their decisions. All funding for health insurance beyond the no-cost plan will come from individuals making their own choices about the relative
value of an additional bit of health insurance compared with their other priorities.
Total health care spending will reflect the outcome of a political process that probably has somewhat expansionary tendencies, combined with a
market process that creates much greater price sensitivity among individuals than exists today. It is
impossible to know whether this total is larger or
smaller than current health care spending.
191
Political Feasibility
It is difficult to evaluate the political feasibility of
the MSS in the current environment, in which serious health reforms are not even being discussed.
Based on recent history, opposition to the plan
would be strongest from the employer community,
which would resist the payroll tax-based financing
system. In the political debates, it would be easy to
demonize the proposal by portraying it as highly
disruptive to people’s current covera ge , as representing a federal takeover of the health care system,
and as encouraging employers to drop covera ge ,
leaving people at the mercy of an untested government program.
These rhetorical devices simply show that any
serious reform of the health care system can be criticized. In fact, the MSS was designed with an eye
toward minimizing disruption, while pursuing principles, such as equity and ch oi ce , that Americans
value. Whether it has achieved these goals is a reasonable question.
One area of political feasibility requires a bit
more attention: the creation of horizontal equity. In
the current voluntary, employer-based insurance system, employers can offer their employees anything
from no coverage to very comprehensive coverage. A
payroll tax shifts all employers to an equal financial
burden (except those that choose to be exempt, but
they retain a significant financial burden, as well).
Proponents of market-based efficiency should
applaud this move toward equity, because it will prevent one firm from gaining a competitive advantage
in its product market at the expense of imposing a
social burden by leaving its employees without
health insurance. However, despite the logical argument behind horizontal equity, we can expect significant opposition to this change in practice.
One reason for the opposition is the fear that
firms have of moving from a system where they can
control their health expenditures, scaling them up
in good economic times and when the labor market
is tight,and scaling them down when the opposite is
the case. Unfortunately, data on the range of costs
employers incur for health care are limited. Based
on the available data, we can say that the average
insurance cost per employee across various meas-
ures, such as firm size, industry, and average wage,
rarely varies from the national mean by more than
 percent. These data do not tell us how much variance there is around the mean within any category.
However, they do offer some evidence that, despite
the control employers have over their costs, employers that offer coverage tend to spend (combining
employer and employee contributions) within a
reasonably narrow range. This suggests that
enforced horizontal equity, while having a very real
effect on those firms that do not offer coverage at
all, will have a modest effect on firms that already
provide coverage.
Equity, Efficiency, and Choice
A fair system has horizontal equity—people in like
circumstances are treated the same—and vertical
equity—people with more ability to bear a burden
take on a larger share. Equitable systems, however,
often come at the cost of efficient use of resources.
And when discussing health insurance where certain actors can benefit financially from risk segmentation, it is tempting to achieve equity at the cost of
reducing or eliminating choice.
No single plan can achieve perfect equity, efficiency, and choice, but the MSS is designed to maximize all three va lu e s . The system has horizontal
equity by requiring financial participation from all
actors. It has vertical equity through the payroll tax
financing system. Efficiency emerges from the economic pressures inherent in health plan and health
insurance exchange competition for enrollees, and
individual price sensitivity for benefits beyond the
benchmark level. Ch oi ce is constrained in some
regards, but most people will observe much greater
health plan and provider ch oi ce than they have
today.
Quality and Access
The MSS is structured around health plans. I consider this a positive step for pursuit of health care quality. Despite heated debates about HMOs and the
quality of care they provide, it is difficult to dispute
the notion that organized systems of care are essential to creating data and systems that have the potential to improve health care quality. These systems are
192
a necessary, although not a sufficient, condition for
improving quality. Because the MSS encourages
development of those systems, it encourages the conditions that support quality improvement efforts.
By structuring market competi ti on around
health plans,the MSS does more to promote quality
than do reforms that rely on market forces at the
point of service. With extremely poor data currently
available on individual providers, reforms that
expect individuals to make value-sensitive choices
every time they select a provider are certain to
encourage competition almost exclusively on price.
While we are far from the potential in this area, it is
possible that competitive health plan selection could
occur on the basis of quality as well as price.
A uniform payroll tax applied around the
country with proceeds directed locally will tend to
push health care spending to a uniform proportion
of the economy. Other funding sources, such as
general federal appropriations and federal matching funds to assist people with low incomes and
with disabilities, may have a more skewed distribution. Still, the payroll tax base could be a significant
force for generating more equitable distribution of
health care resources.
In the end, the MSS is designed as a health care
financing system, not as a mechanism for directly
addressing the distribution of health care resources.
Eliminating the problem of uninsurance, and dramatically reducing the phenomenon of underinsurance, should yield substantial improvements in
access and quality. However, these areas require
additional atten ti on beyond reform of the health
care financing system.
Conclusion
The Medical Security System is offered as a proposal
for achieving true universal health insurance coverage in the United States. It is an imperfect proposal,
with many complex areas still to be defined. The
trem en dous importance people place on health
security makes any transition from the current system difficult to achieve. Yet, a vision of a fair and efficient system can help us think about the direction
incremental steps should take today, and help prepare us for a time when universal coverage returns to
the center of the American political agenda. n
193
Wicks, Meyer, and Silow-Carroll Proposal
Key Elements
Elliot K. Wicks, Jack A. Meyer, and Sharon Silow-Carroll have outlined a proposal to achieve universal health coverage while maintaining a market-based
system and simplifying administration. Key elements of the proposal include
the following:
       , payable in advance, for all households, with the
credit varying by income, sufficient for those below the federal poverty
level to cover the full cost of coverage comparable to Medicaid and gradually reduced for higher-income people.
           at least as comprehensive
as Medicare plus drugs and well-child care. Those not meeting the requirement would be automatically covered by Medicare as a backup but would
have to pay a premium plus a penalty (at tax time) for every month without private coverage.
      (but not necessarily pay for)
a minimum benefits plan no less comprehensive than Medicare.
           that permits employees to
exclude from their taxable income the amount that their employer pays
for health coverage.
      , or aggregate purchasing arrangements, to serve as a source of health coverage for individuals and small
employers. Insurers are required to participate and offer a standard
benefits plan comparable to Medicare.
      , coordination of benefits,
etc., to reduce administrative duplication and inefficiency.
194
About the Authors
     . , .., is Senior Fellow at the
Economic and Social Research Institute and Senior
Consultant with Health Management Associates.
Dr. Wicks specializes in analysis of policy reforms
to help bring affordable health coverage to more
Americans,especially workers in small firms. He
has extensive knowledge of arrangements of pooled
purchasing of health coverage and recently directed
a project to investigate the barriers to the success
of health purchasing cooperatives. Other recent
research includes a study of small-group market
reform and employers’ and consumers’ use of
health plan report cards. Dr. Wicks is the author of
numerous articles and monographs on these and
other subjects related to health care financing and
delivery. Dr. Wicks has worked for health care
consulting firms, policy research institutes, a trade
association, and state government, and he was
formerly a faculty member at Michigan State University. He has a Ph.D. in Economics and Social
Policy from Syracuse University and an M.A.in
Economics from Northwestern University.
 . , .., is the founder and President of
the Economic and Social Research Institute. Dr.
Meyer has conducted policy analysis and directed
research on health care issues for several major
foundations as well as federal and state government.
He has led projects developing policy options for
reforming the overall health care system and directed research on community-wide reforms covering
all regions of the U.S. Many of these projects have
highlighted new strategies for overcoming barriers
to health care access and innovative designs for
extending health insurance coverage to the unin-
sured. Dr. Meyer is the author of numerous books,
monographs, and articles on topics including health
care, welfare reform, and policies to reduce poverty.
He has also directed recent studies on the viability
of safety net providers, Medicaid managed care for
people with disabilities, the conversion of public
hospitals to private status, and assessments of
reform proposals to extend health coverage to
workers in small firms. Dr. Meyer holds a Ph.D.
in Economics from Ohio State University.
     -    , ..., ..., is Senior
Research Manager at the Economic and Social
Research Institute. She has conducted health policy
analysis for more than ten years, specializing in
assessing health care reform strategies that expand
coverage to vulnerable populations. Recent projects
include examining employers’ attitudes about their
current and future involvement in providing health
coverage, reviewing community-based health plans
for uninsured individuals, and profiling state and
local initiatives to expand employer-based health
coverage among the working uninsured. Ms. SilowCarroll is the author of In Sickness and In Health?
The Marriage Between Employers and Health Care,
which analyzes the corporate/employer role in providing health care coverage from economic, social
and cultural perspectives. She has written numerous
reports and articles reviewing public and private
sector programs aimed at enhancing access, containing costs, and improving quality of care. Ms.
Silow-Caroll received an M.B.A. in Health Care
Management from the Wharton School, and an
M.S.W. from the University of Pennsylvania School
of Social Work.
195
A Plan for Achieving Universal Health Coverage
Combining the New with the Best of the Past
by Elliot K. Wicks, Jack A. Meyer, and Sharon Silow-Carroll
No Americans should be denied access to needed
medical care because they lack health insurance coverage, and no health care providers should go unpaid
because they treat people who lack the means to pay
for care. This proposition is the guiding principle
underlying the proposal for universal health coverage
that we develop in this paper. We have designed a system that achieves universal coverage by (a) providing
generous subsidies in the form of tax credits for
those with limited ability to pay, (b) mandating that
everyone buy coverage from one source or another,
(c) establishing Medicare as a temporary backup
payer for those who fail to purchase coverage, and
(d) establishing aggregate purchasing arrangements.
The system is built on the foundation of current private health plans and employment-based coverage.
Our plan addresses only the non-elderly population;
the Medicare program for the elderly would remain
as a separate program.
Objectives
Our approach is based on a vision of the way health
care financing should look. It seeks to achieve the
following objectives.
Universal Coverage
Large numbers of Americans are without health
coverage—about  million by the latest count—
and this number has been rising, even in the face
of the longest period of sustained prosperity in
U.S. history. Many other citizens have inadequate
coverage that does not protect them from incurring
unaffordable medical bills in the case of a serious
illness or injury or does not encourage use of costeffective preventive and primary care services.
Many people lose coverage when they lose or
change jobs. Although there are programs to cover
poor children and families, many low-income individuals and working families are not eligible for
public subsidy programs and cannot afford to buy
coverage privately.
We are proposing a plan to correct these problems. We believe that any plan to achieve universal
coverage must include two features: a federal mandate that everyone have covera ge , and substantial
subsidies to make coverage affordable for everyone.
Those features are already embodied in Medicare,
the plan that covers essentially everyone over the age
of . In enacting Medicare, we decided as a society
that ensuring access to needed medical care for the
elderly population was so important that we were
willing to impose a degree of compulsion to achieve
universal covera ge . Using the same rationale, we
believe that a mandate for coverage can be justified
as a way to ensure access for people of all ages and to
achieve fairness—by eliminating “free-riders,” those
who do not buy coverage but use the medical system’s resources in em er gencies. Our plan would
en su re that no one is ever “between” coverage or
otherwise falls through the cracks. This objective is
achieved by establishing an individual mandate and
by making Medicare the default payer for those who
nevertheless fail to get coverage (with disincentives
for individuals to rely permanently on Medicare
coverage).
Reduced Fragmentation, Duplication, and Inequities
of Public Subsidy Programs
The current public financing system is highly fragmented and unduly complicated. There are many
different types of subsidy programs, each with dif-
196
ferent el i gi bi l i ty criteria and benefit structures.
As people’s circumstances change, they become
ineligible for one program but may not be eligible
for others, or they may be eligible but not know
they are. Some people are reluctant to enroll
because of the stigma associated with public programs. As a consequence, people fall through the
cracks. The subsidy system is far from seamless. Few
people can keep track of all the system’s features,
especially because of constant policy changes,and it
inevitably is bureaucratic, duplicative, and expensive to administer.
Current subsidies are also unfair. Because the
major public financing progra m s ,e s pecially Medicaid, permit substantial local discretion in setting eligibility standard s ,i n equities abound. People in equal
circumstances are not treated equally. Needy people
in some parts of the country have no coverage, while
similarly situated people in other parts of the country have comprehensive coverage. We have a multitiered system of care based in part on income and in
part on where people live and their state’s eligibility
and benefits standards. If the objective is to provide
access to appropriate care to all needy people, it is
hard to defend the current system.
In addition, the subsidy system for the nonpoor—the income tax provisions that allow employer-paid health premiums to be excluded from
employees’ taxable income—favors higher-income
people over lower-income people. Higher-income
people often work for employers who pay more
toward coverage, so more income is tax-exem pt ;a n d
because they have higher marginal tax rates, the tax
exclusion is worth more to them. Government’s taxexpenditure cost is very high—. billion in federal money and . billion in state tax losses in .1
The approach we propose eliminates the patchwork of subsidy approaches and multiple public
programs, the burdensome and expensive administrative procedures for determining eligibility, and
the myriad complicated and constantly changing
regulations. This objective is achieved by replacing
1
John Sheils, Paul Hogan, and Randall Haught. Health Insurance and
Taxes: The Impact of Proposed Changes in Current Federal Policy. Report
of The National Coalition on Health Care, October 18, 1999.
most subsidy programs with a tax credit, basing eligibility for the credit on income alone.
Simplified Administration
The current administrative system for private
health insurance is unnecessarily inefficient, wasteful, and burdensome for patients and providers.
Time and resources are wasted because there is
no centralized system for determining eligibility,
identifying benefit limits, submitting claims for
payment, and coordinating benefits. Many of the
costs are borne by patients as they try to wend their
way through the maze of claims submission and
administration, and surely many costs that are legal
obligations of insurers are actually paid out of
pocket by patients because claims are never submitted or are not settled accurately. The system we
propose would substantially reduce the burdens
and costs borne by providers and patients related
to determining the limits on benefits, filing claims,
and coordinating benefits. This objective is
ach i eved by establishing a nati onal cen tra l i zed
electronic mechanism for paying claims and coordinating benefits.
Maintaining the Role of Private Health Plans and
Insurers as Sellers of Health Insurance Coverage
The current system of private health plans and
insurers works well for most Americans. Competition among health plans and insurers for business helps to promote efficiency and better service.
In addition, recent changes in state and federal laws
governing the sale of insurance to small employers
have improved some aspects of performance.
Though further changes may be needed, there
seems to be little reason or political desire to abandon the basic structure of a private insurance system, and there is no obvious alternative on the
horizon. The case for retaining the basics of the current system is strong.
Continued Reliance on Employer-Sponsored
Coverage and the Role of Employers as Poolers of Risk
Employers play a major role in pooling risk; that is,
they bring together people with different levels of
risk who all pay a similar premium. Avoiding seg-
197
mentation of risk is a major challenge for any system, so it seems wise not to abandon a system that
meets the need for risk pooling for a large portion of
the population. In addition, much of the pressure
and many of the ideas for cost control and quality
improvement have origins in the employer community. The case is strong for con ti nuing to have
employers be advocates for employees in purchasing
coverage that offers good value.
Features of the System
Subsidies
 
Every (non-elderly) American would be eligible for
a health coverage subsidy in the form of a “refundable” tax credit that could be used to offset costs of
covera ge , whether the premium is paid by the
insured person (including those buying individual
coverage) or by an employer on behalf of an
employee. That is, a family’s income tax liability
would be reduced by the amount of the credit, as
long as the total of the employer and employee premium was equal to or greater than the credit.
However, any premiums paid by the employer
would be considered taxable income to the employee. Although the size of the tax credit would be larger for lower-income people, everyone would be
eligible for the minimum credit. That minimum
credit would be equal to the average value of the
current federal income tax exclusion to those who
have employer-sponsored coverage, or approximately  per year for an individual or , per
year for a family.2 Larger subsidies would be available to people below the median family income
(currently about , per year). People with
incomes at or below the federal poverty level would
get a total tax credit sufficient to pay the full premium for coverage comparable to the costs of effi-
In 2000 the average employer premium contribution was about $4,600
a year for family coverage and about $2,100 for single coverage.
(Computed from Jon Gabel et al. “Job-Based Health Insurance in 2000:
Premiums Rise Sharply While Coverage Grows.” Health Affairs 19 [5]
[September/October 2000]: 147). For someone in the 32 percent marginal tax bracket, the tax savings are about $1,500 for family coverage
and $700 for single coverage.
2
ciently provided Medicaid benefits.3 For those
between the poverty level and the median income,
subsidies would be reduced gradually as income
rises so that those at the median income would
receive the minimum subsidies indicated above.
The tax credit subsidy needs to be “refundable”
and payable in advance. That is, people whose
income tax liability is less than their credit would
receive the difference in the form of a “refund.” And
because premiums have to be paid monthly beginning more than a year before tax time,a mechanism
is needed to make coverage affordable during the
year as premiums come due. We propose that the
amount of the subsidy be based on the previous
year’s reportable income. People whose reportable
income is low enough to qualify them for advance
payments would receive a federal voucher every
month that could be applied to the cost of coverage.
Vouchers could be transferred to and redeemed by
either employers or insurers, depending on whether
the person has coverage through an employer or in
the individual market. (This voucher process and
redemption generally would be handled electronically through the centralized administration system
described later.) If a family that is not receiving
vouchers experiences a decline in income that
would make it el i gi ble for advanced payments, it
could apply for eligibility at that time. Any government overpayments would be reconciled at the next
tax filing, with minor amounts (for example, under
 per year) being forgiven.
An important issue is how to adjust the subsidies
over time as the cost of medical services increases.
For people below the poverty level, the subsidy
should be adjusted so that it is always adequate to
purchase coverage equivalent to current Medicaid
benefits. We propose that the Office of the Actuary
in the Health Care Financing Administra ti on be
assigned the task of developing an appropriate
3
Because Medicaid-covered services vary from state to state, the federal
enabling legislation would define a uniform benefits package. But the
intent is that the covered services would be equivalent to what is now
typically available to Medicaid enrollees. Any changes in benefits over
time would be defined by changes in federal law or regulations.
198
The most compelling reason for mandating
that everyone be covered is that this requirement
is necessary to ensure universal coverage.
index to increase the subsidy over time. For people
earning above the median income, we would propose that the subsidy not be indexed to increase
automatically but could be increased at the discretion of Congress. For those between the poverty
level and median income, the subsidy would
increase automatically when it was increased for
those below the poverty level.
 
The current tax policy that excludes employer premium contributions from employees’ taxable
income is widely acknowledged to be an inequitable
and inefficient way to subsidize the purchase of
health insurance. It is inequitable because many
low-income people do not have employer-sponsored covera ge , so they get no benefit; h i gh erincome people tend to have more comprehensive
coverage and thus more excludable income; and
higher-income people have higher marginal tax
rates, so they benefit more from every dollar that is
excluded from tax. The subsidy is inefficient—that
is, its cost is high relative to the objective—because
much of the forgone tax revenue pays for subsidies
to people who could afford to pay for coverage out
of pocket. It is also inefficient because it encourages
people to consume more health care services relative
to other goods and services than they would if
employer-paid premiums were not “tax sheltered.”
Substituting a tax credit for the tax exclusion
substantially reduces the inequity, especially
because the credit we propose is larger for lowerincome people but fixed for those with incomes
above the median. Though a case could be made for
entirely phasing out the credit for higher-income
people on equity and efficiency grounds, this would
produce a large tax increase for politically influential middle- and high-income people. Retaining a
substantial subsidy for these income groups should
lessen political opposition to the change, although
high-income people whose employers contribute
generously to comprehensive coverage would still
have a higher tax liability with the tax credit than
with the tax exclusion.
The current Medicaid program provides comprehensive medical coverage to families that qualify.
(It also includes coverage of some non-medical
services, about which more is said later.) We believe
that such comprehensive medical coverage should
be subsidized fully for all families below the poverty
level,and that substantial though gradually decreasing subsidies should be available to families up to
the median income level. Without such subsidies,
many will find the cost of paying for the nowmandatory coverage to be burdensome if not
impossible. The graduated phasing-out of lowincome subsidies is justified on the grounds of ability to pay and to ensure that the system does not
include strong work disincentives. If the subsidy
were reduced too quickly, some people would be
reluctant to take higher-paying jobs that would
make them ineligible for the low-income subsidy
because they actually might have lower net incomes
after paying more for health insurance.
The feature of the proposal that may need more
explanation is the provision that allows the credit to
be applied not only to premiums paid by employees
or individuals but also to those paid by employers.
This provision would appear to make the budgetary
cost of the subsidy higher, but that is not likely to be
the case in the long run. Since employer-paid premiums would be taxable income to employees,there is
no reason for employees to prefer being compensated in the form of employer-paid premiums rather
than money wages. If the tax credit could not be
applied to employer-paid prem iu m s , employees
would urge employers to stop paying the premium
and give them the equivalent in money wages.
Employees then could pay for coverage themselves
and use the tax credit to offset the cost,leaving them
199
with more net income. Over time employers likely
would stop paying anything for premiums because
doing so would benefit their employees without
adding to the employers’ costs. (Economists argue
that employers are largely indifferent to the form of
compensation; what counts is the cost of the total
compensation package,including money wages plus
employee benefit costs.) At the point where employees pay all of the prem iu m , the credit could be
applied to the entire premium;so the budgetary cost
is the same as if the credit could be applied toward
both the employer and employee portions of the
premium. But if employers stopped paying for premiums, they might be tempted simply to abandon
en ti rely their role as purchasing agents acting on
behalf of their employees. For reasons noted earlier,
we think this would be a bad result. But if the credit
applies equally to employer-paid or employee-paid
premiums, employers may continue to pay a portion of the premium and to pursue good value in
purchasing health coverage for their employees.
Regarding changes in the subsidy over time, the
rationale for increasing the subsidy for impoverished
people as medical costs rise is straightforward: We
want to ensure that they can afford the coverage they
are required to buy. For the group between the
poverty level and median income, the same rationale
justifies increasing their subsidy, which would happen more or less automatically, since it is tied to the
subsidy for those below the poverty level, with a
phase-out as income rises. We do not propose, however, to automatically increase the tax credit subsidy
for people above median income. The main reason
for creating the credit for them in the first place was
not because the subsidy was needed to make coverage affordable, but to avoid the political objections
that would occur if this group had to give up its current tax exclusion subsidy without having anything
else in its place. On equity grounds, a case could be
made for having the real (after-inflation) value of
the tax credit subsidy decline over time. If the credit
is increased for this higher-income group, we favor
having Congress explicitly decide to raise it rather
than having it increase automatically as medical
costs rise.
Individual Mandate
 
Every individual and family would be required to
have health coverage—that is, at a minimum, as
comprehensive as Medicare benefits (Parts A and B)
with the addition of a drug benefit and well-child
care.4 To en su re that such plans are available, all
insurers offering health coverage would be required
to offer a policy that includes the services covered by
Medicare plus prescription drugs and well-child
care and to price the policy on an actuarially defensible basis.
A mandate without effective enforcement would
not achieve the desired result. We propose that
everyone be required to show proof of purchase of
coverage as part of his or her annual filing of federal
income tax forms. In the case of families, proof of
coverage would be required for the person filing the
return, his or her spouse, and all dependents listed
on the tax return. Insurers, health plans, and selfinsured employers would be required to issue a
standard form to all policy holders that serves as the
proof of purchase and is attached to income tax
returns (comparable to W- forms now issued by
employers to show earnings). The forms would
indicate the months during which each person is
insured. Individuals who have no taxable income
and, therefore, do not now file a tax retu rn , s ti ll
would be required to send in proof of coverage
when federal tax returns are due.
Those who fail to show proof of coverage incur
the fo ll owing penalty: for every month they are
without coverage, they would be required to pay a
fee, to be included with their tax retu rn , that is
equivalent to the monthly cost of coverage for
Medicare benefits plus a  percent surcharge. The
fee would be levied whether or not the individuals
use any medical services during the time they are
uninsured. The fee would be based on the actuarial
cost of providing the augmented Medicare benefits
package to a non-Medicare population under age
4
In referring to Medicare benefits, we are referring only to the services
covered, not to the kind of system that delivers these services. An individual could meet the mandate requirement by choosing an indemnity
plan, a preferred provider organization (PPO), a health maintenance
organization (HMO), etc., as long as the benefits were as extensive as
those covered under Medicare.
200
. The premium assessment would be adjusted to
reflect family size and composition and regional differences in medical costs, but not age of the adults
(explained under “Insurance Regulation” below).
The tax credit applicable to the family or individual
(discussed in the previous section) can be applied
against this liability.
 
To achieve the objective of universal coverage
requires that one of two conditions be met: either
everyone must be required to purchase coverage, or
a mechanism must be in place that automatically
covers everyone (as is the case with social insurance
systems). In a sense, we have chosen to build in both
conditions. We would mandate that everyone buy
private coverage, but, in addition, we propose that
anyone who is not privately covered, for whatever
reason, would default into Medicare coverage and
pay a premium for the time he or she is covered
under that system. In a sense, this approach can be
thought of as an individual (as contrasted with an
employer) “play or pay”mandate: one either “plays”
by purchasing private coverage or “pays” by being
assessed for Medicare coverage. (The mechanism by
which this is accomplished is discussed in the next
section.)
The most compelling reason for mandating that
everyone be covered is that this requirement is necessary to en su re universal covera ge . But there are
other reasons,as well. Even if they have the means to
buy coverage, some people will choose not to do so
if there is no mandate;yet when they need expensive
care for life-threatening or emergent conditions,
society is not willing to deny them access to essential
services. They then become “free-riders” who do not
bear their fair share of the costs.Our approach prevents this.
In ad d i ti on , no one has been able to devise a
practical mechanism for making the individual
insurance market work well without mandating that
everyone have coverage. The individual market falters without a mandate because individuals can predict when they will need certain expensive kinds of
medical care—for example, elective surgery or
maternity benefits. Some people will choose to buy
coverage only when they expect to need care, which
creates severe problems of adverse selection. The
ability to buy coverage only when the insured person is likely to incur expenses negates the insurance
principle, which involves pooling of risk among
individuals who cannot predict when they will need
expensive services. Moreover, it is not fair to require
people who want to stay insured permanently to pay
for the costs of care provided to individuals who
become part of the insurance pool only when they
know they will be incurring major medical expenses. Individuals who go in and out of coverage on
that basis are not paying their fair share.
Since our approach includes the “mandate” that
everyone who does not buy private coverage is automatically covered by Medicare and must pay a premium for that coverage, why, then, do we also
propose to mandate the purchase of private coverage? Without such a mandate,the number of people
who default into Medicare coverage would almost
surely be greater (even though they must pay a premium plus a penalty for the time they are covered
by Medicare). Because we seek to encourage coverage acquired in the private market rather than having large numbers of people defaulting into
Medicare, we propose a mandate to buy private coverage to promote that objective. On the other hand,
our proposal would still achieve universal coverage,
even without this requirement. The mandate to purchase private coverage is desirable but not necessary.
We propose to use the federal income tax filing
mechanism to enforce the mandate because it is a
relatively simple approach and would be an add-on
to a process that most people complete routinely
each year. Of course, some people who do not file
returns now would have to do so, but they also
would be required to do so to verify the amount of
tax credit subsidy for which they are eligible
(explained above). Because virtually all of these
people would have very little income, the form
could be very short and easy to complete. Some people would fail to comply, and they would be subject
to the same penalties as people who do not file
returns: if they owed an obligation—in this case, if
they failed to buy coverage, and the cost of coverage
exceeds their tax credit—they would be subject to
201
interest penalties identical to those for unpaid or
overdue taxes. Some people—for example, the
h om el e ss—might fail to file, but because of their
very low income, they normally would have no premium obligation anyway.
The  percent penalty above and beyond the
actuarial cost of coverage for those who fail to get
coverage on their own is imposed to create incentives to buy private coverage. Because we want to
maintain the privately based insurance system, we
want to make the default position of having
Medicare pay the bills more expensive than getting
coverage through the private system. The base premium amount is based on a community-rated premium, with ad ju s tm ents for family size and
differences in regional costs. The ra ti onale is to
make the rate comparable (not counting the surcharge) to what could be purchased privately in the
region. (As we explain later, we propose that community rating be required in the private insurance
market for individuals and small groups.)
A requirement that individuals be insured must
be coupled with a definition of a minimum benefit
package that fulfills the requirement. We have chosen Medicare coverage (Parts A and B) plus a drug
benefit and well-child care. We add drug benefits
because we believe that prescription drugs should
be covered, and because there seems to be strong
support for adding such a benefit to Medicare. We
add well-child care because it is cost-effective coverage that is not applicable,and therefore not covered,
under Medicare. We favor this definition of minimum benefits for several reasons.First,this is a benefit package that already applies to the elderly
population. If it is good enough for the elderly, it
should be good enough for everyone else. It would
be inappropriate to have a benefits package minimum for the non - el derly that was more compreh en s ive than that available under Medicare for
seniors. Second, the Medicare benefits package
reflects a political decision about what constitutes
an appropriate level of services. Relying on that
decision avoids going through the very controversial
and politically charged process of defining a new
benefits package, with the inevitable intense lobbying from disease-specific advocates and provider
groups. Third, Medicare benefits are not so comprehensive as to make the cost of the minimum package very high, nor is the coverage so generous as to
encourage “excess” consumption of medical services. Fourth,having the minimum benefits package be
the same as Medicare benefits simplifies administration of the “fallback” coverage system for people
who fail to get private coverage (explained later).
Procedures related to claims review, reimbursement
of providers, cost-control measures, etc., would be
identical for the fallback system as for the existing
Medicare system. No new mechanisms or bureaucracies would be required.
Even though we favor using Medicare benefits as
the minimum standard for everyone, our approach
would work well even if some other benefit standard
were adopted. The merits of our approach do not
depend on using Medicare coverage to define minimum benefits for the individual mandate. If it is
decided that the peculiarities of the Medicare benefits pack a ge , even with the additions we suggest,
make it unsuitable for use as the standard,a number
of alternatives could be substituted. For example,
the minimum coverage package could be based on
the services covered under the Federal Employees
Health Benefits Plan (FEHBP). Or Congress could
devise an en ti rely new set of benefits just for this
program. If it is simply the association with
Medicare that carries negative political connotations, the program could be identified with an
entirely different name.
Medicare as a Source of Backup Coverage
 
Anyone who lacks coverage at any point for whatever reason would be automatically covered by
Medicare. The benefits package available would be
identical to the minimum benefits required under
the individual mandate, which in this proposal are
the equivalent of Medicare coverage plus drug coverage and well-child care. People who default into
this arrangement do not become Medicare
enrollees, but Medicare is responsible for any (covered) medical expenses they incur during the time
they lack other coverage. Medicare pays providers
on the same basis and through the same mecha-
202
nisms that are used currently, but the money does
not come out of the present Medicare trust fund.
Instead, a new fund would be established for this
purpose. This fund would be financed primarily by
the assessments imposed on people for the months
they lack coverage and that are paid as part of the
federal income tax filing process, as explained earlier. Even though the assessment is equal to an actuarially determined premium plus a  percent penalty,
some funding shortfall is likely because some people
covered through this system will not actually pay
any assessment. Their income will be so low that the
tax credit for which they are eligible will fully offset
the full cost of their assessment. Funds to cover this
shortfall would have to come from other sources,
essentially general tax revenues.
 
Even when everyone is required to have coverage,
some people inevitably will not be en ro ll ed in a
health plan for some period of time. They will be
between jobs and fail to get individual covera ge ,
they will fail to pay a premium and subsequently
will be disenrolled, or they simply will fail to sign
up, even though they are required to by law. These
people still need to have a source of coverage if they
are to get the care they need, and if providers are to
be paid for the services they provide to them.
Having Medicare provide backup coverage solves
this problem. We would not make these people
Medicare enrollees because we want to encourage
them to get private coverage. That is also the reason
for imposing the  percent penalty above and
beyond the actuarial cost of coverage. Some people
probably will default into Medicare coverage for
long periods—the homeless, for example—and
they are likely to be people of above-average risk.
But this is probably an appropriate way to spread
risk.
We propose Medicare benefits as the coverage
package available to those who default to the fallback system, but there are obviously other alternatives. The coverage could be more or less generous.
For example, the FEHBP benefits package could be
used as a model. But whatever the benefits, they
should be identical to the minimum benefits pack-
age required for the individual mandate. The coverage should be no less comprehensive because the
notion of universal coverage implicitly requires that
everyone have access to a societally determined
minimum set of benefits. On the other hand, if the
fallback benefits were more generous, people would
have incentives to default to the fallback system
rather than buying coverage on their own.
Even under our preferred option of using the
Medicare benefits package, it would be possible to
choose an administrator for the default system
other than Medicare—for example, FEHBP—if
that is thought to be politically desirable. The
advantage of having Medicare as the administrator
is that the program already has contractual agreements with nearly all providers and has mechanisms
in place for administering claims,setting reimbursement amounts, and making payments. Thus, no
new bureaucracy and few administrative changes
would be necessary to administer the system.
We would not propose to cover undocumented
immigrants through this system. The incentives for
people to enter the country illegally to get access to
treatment would be too strong. We understand that
some undocumented peop l e ,e s pecially children ,a re
covered by Medicaid or the State Children’s Health
Insurance Program (S-CHIP), even though they are
not officially eligible for these programs. Some of
them would lose coverage if this proposal were
adopted. We acknowledge that meeting the health
needs of undocumented people, especially for primary and preventive care, deserves attention and that
better ways need to be found to cope with the problem. But finding a solution is beyond the scope of this
proposal, because the problem has many dimensions
other than those related to health coverage.
Employer Mandate to Offer (But Not Pay for)
Coverage
 
We would require employers to offer coverage to
their employees and employees’ dependents, but
they would not be required to pay anything toward
the premium, though many would choose to do so
just as they do now. Employers would be free to
design any benefits package they thought appropri-
203
ate, as long as the coverage was at least as comprehensive as Medicare coverage plus a drug benefit
and well-child care. The requirement to offer coverage could be met by offering it through an aggregate
purchasing arrangement (APA) that would be available in each state (as explained below) and would
offer benefits packages equivalent to Medicare and
Medicaid, among others.
Employers that offer their own plan rather than
purchasing through the APA would be required to
a ll ow employees who are eligible for the lowerincome tax credit (those with incomes below the
family median) to purchase coverage equivalent to
Medicaid coverage through the APA. That is,if eligible lower-income employees request it, employers
would be required to withhold premiums from paychecks and send the withheld amount to the APA,
along with the same dollar contribution that the
employer makes for the firm’s standard plan.
 
Inclusion of this mandate to offer coverage ensures
that every employee would have the option of
being covered by an employer-sponsored plan, with
the risk-pooling advantages and administrative
economies of scale that group purchasing achieves.
Employers would have a reason to seek a good value
for their employees’ benefit, even if the employer
pays nothing. And having the employer withhold
premiums from paychecks and pay the insurer or
health plan is more efficient and less expensive for
health plans and relieves employees of the burden of
doing this themselves. Having a plan available
through an employer makes it easy for most people
to meet the requirement that they buy covera ge .
They avoid the need to incur burdensome transaction costs—finding a plan for themselves or finding
an agent who will help them do so, making difficult
ju d gm ents about the value of various plans, and
then paying individual premiums to the plan they
choose.
We see no substantial advantage to requiring
employers to pay for coverage.Economists generally
believe that employers of fs et the cost of paying
insurance premiums by paying lower money wages
than they otherwise would. If that is so, there is no
advantage to requiring employers to nominally pay
the cost of the premium (now that employer-paid
premiums would not be tax-excludable income).
Moreover, a requirement to pay for coverage likely
would cause some employers who pay only the minimum wage to lay off some workers because the
workers’ contribution to productivity would not be
great enough to justify paying the now-higher total
compensation.
We allow employers to choose the aggregate purchasing arrangement (described next) as the vehicle
for offering employees coverage. This makes it easy
for employers that do not offer coverage now to
meet their obligation, but still ensures that employees are not forced into the less efficient and potentially confusing individual market. We require
employers to allow lower-income employees to purchase coverage equivalent to Medicaid coverage
through the APA because the employer’s own benefit package may not be suitable for lower-income
employees. For example, the deductibles and copayments may be unaffordable, or primary care
services may not be covered adequately.
Aggregate Purchasing Arrangements
 
Each state would be required to establish an aggregate purchasing arrangement that would serve small
employers and individuals (though larger employers could opt to purchase coverage through this
mechanism). The federal government would establish general guidelines for these organizations, but
states would be given wide flexibility in deciding
what kind of arrangement to establish. For example,
they could establish a traditional health purchasing
cooperative,a HealthMart, or a similar organization
of their own design. They could establish just one or
as many as they thought feasible. They could be private, quasi-public, or public entities, at the state’s
opti on . As an alternative, the state could allow
employers and individuals to buy into the state’s
existing employee plan or a different state plan
designed for this purpose.
Each aggregate purchasing arrangement would
be required to offer, at a minimum,a benefits package equal to Medicare benefits plus drug coverage
204
and well-child care and another equal to Medicaid
coverage. Each participating health plan would be
required to offer at least these two benefits packages.
No insurer or health plan offering insured plans in
the state could refuse to offer coverage through the
APA, but the APA (under state-determined regulations) would establish its own criteria for deciding
which plans to include.
People getting coverage through the APA,
whether individuals buying for themselves or
employees in a group, would individually be able to
choose any health plan that offers coverage through
the APA. Each insurer selling through the APA
would price coverage on a community-rated basis
(with the minimal adjustments explained below),
and that coverage would be available at that price to
all individuals and all groups with  or fewer
employees. The rates health plans charge inside the
APA could be no higher than the rates they offer for
comparable coverage to groups outside the APA.
But the APA could negotiate with insurers for a
lower price than their outside pri ce , reflecting
administrative savings, volume discounts, or other
efficiencies that insurers and health plans realize by
selling coverage through the APA. Insurers contracting with the APA would be subject to all the rules
that apply in the small-group and individual markets, as explained below.
Employers with  or fewer employees would be
required to offer coverage exclusively through the
APA. Groups of any size could participate, but the
state would have the option of choosing to establish
a mechanism for the APA that would allow insurers
and health plans to charge groups with more than
 employees a premium that reflected the specific
group’s risk and administrative costs. Individuals
could also buy coverage through the APA. No group
or individual seeking coverage through the APA
could be excluded; the APA and all participating
insurers would have to accept all applicants.
 
When individuals and small employers want to purchase health coverage on their own,they are at a disadvantage. They lack the specialized knowledge and
resources that large employers can allocate to this
task,so they are not in a good position to determine
whether they are receiving good value and buying a
plan that best fits their needs. In addition, because
of the diseconomies that health plans face in serving
small groups and individuals, most notably high
marketing and administrative costs, these buyers
pay more for coverage than large employers.
Aggregate purchasing arrangements have the potential for giving individuals and small employers some
of the advantages that large employers enjoy—not
only lower administrative costs, but also the power
to nego ti a te with health plans to ensure that purchasers are buying high-value coverage. Fu rt h er,
APAs could provide the kinds of cost and quality
comparisons that individuals and employees need
to choose wisely among health plans.
But previous experience with purchasing cooperatives and similar organizations has been discouraging. For the most part, they have not realized the
expected economies, and they have had trouble
attracting sufficient numbers of employers and
maintaining health plan participation. Most
observers agree that the problems of aggregate purchasing arrangements would be largely solved if they
could become big enough, and if they could be
assured of health plan participation. We attempt to
solve the first problem by requiring all very small
employers to participate. These are the employers
least capable of buying cost-effective health coverage
on their own, but there are enough of them to create
a large pool of business when all of them participate
and all their employees get coverage, as required. The
size of this market might be sufficient by itself to
attract many health plans, but to ensure continued
participation, we would require that all plans participate if asked to do so by the APA. States might
decide to establish criteria for limiting the number of
plans that participate in the APA, because,if all plans
were to participate, administrative costs might be
excessive and the range of choices might be overwhelming. Further, APAs might want to limit participation as a nego ti a ting ploy: they could bargain
with plans to give all of their business to the few
health plans that offer the best deal.
We propose to allow every individual or employee buying coverage through the APA to choose any
205
The problems of aggregate purchasing arrangements
would be largely solved if they could become big enough,
and if they could be assured of health plan participation.
plan that participates in the APA. Especially in the
era of managed care, we think people should be able
to select the plan that best matches their needs, and
they should be able to switch plans periodically (for
example, once a year) if they become dissatisfied.
This individual-choice provision also puts competitive pressure on plans to perform well. Moreover,
when employees can choose their own plan, they
will often be able to stay in the same plan when they
change employers.
We allow groups of any size to buy coverage
through the APA if they wish to do so. Having larger
groups be part of the APA could help it to achieve
greater economies of scale and give it more negotiating clout when dealing with health plans. The APA
would also serve as a convenient vehicle for providing coverage for those employers that wish to adopt
a defined-contribution approach to paying for
health coverage. But if large groups are community
rated with the smaller groups in the plan, there is a
danger that the APA will be adversely selected
against: higher- ri s k , larger groups would have an
incentive to join the APA to get a lower premium
rate. We therefore propose that these larger groups
be separately rated if the state chooses to take this
approach.
Insurance Regulation
 
Federal law would require all health plans to accept
all individual and small-group applicants (a guaranteed-issue requirement) and to provide immediate
and full coverage for all covered benefits. In other
words,there could be no waiting periods, exclusions
for prior conditions, or other limits on coverage that
would be applied differently for new enrollees.
In su rers and health plans selling coverage to
individuals and groups with  or fewer enrollees
would be required to price premiums on a commun i ty - ra ted basis. Adjustments would be permitted
only for family size and composition and for regional differences in medical expenses. In other words,
each insurer would put all individuals and groups of
 or fewer in a single pool for a defined geographic area, and the insurer’s premium would be based
on the medical claims experience of all of the people
in that pool.
 
The guaranteed-issue requirements are consistent
with current federal policy in the small-group market, but,more important,are absolutely necessary to
en su re universal covera ge . Justification for any of
the current limits on coverage having to do with
prior conditions, waiting periods, coverage portability, and so forth is negated by the individual mandate requirement. Insurers traditionally have
included these provisions to protect themselves
against people who would wait to buy coverage until
they knew or suspected they would be incurring
major medical expenses. This problem has been
acute in the individual market, but insurers believe
it is also a problem in the market composed of very
small groups (those with no more than four or five
employees). Under our proposal, however, everyone
will have coverage because of the individual mandate, so this justification for limiting coverage for
new enrollees disappears. Insurers, however, could
be permitted to establish reasonable waiting periods
for conditions not covered under the Medicarebased minimum benefits package but covered by an
optional, more comprehensive benefits package.
Since coverage for these additional benefits would
be voluntary, adverse selection problems could arise
if insurers were not allowed to impose any restrictions on access to coverage. In the interest of avoiding churning and associated administrative costs, it
also would make sense to limit plan switching to an
open enrollment period or at the time of s om e
change in job status or family condition, such as
206
marriage or the birth of a child.
The decision to require community rating is
based in part on a value judgment and in part on
efficiency grounds. The value judgment is that people should not be rewarded or penalized (in the
form of premium differences) for risk characteristics over which they have little or no control. People
cannot change or influ en ce their age, gender, or
genetic predisposition, all of which make them
more or less vulnerable to illness and injury. They
cannot affect their past medical experience and
associated medical expenses. They can influence
current behavior, which can affect future medical
expenses and,therefore, risk, but health insurers seldom consider personal behavior, aside from smoking, in assessing risk. And if they were to consider
personal behavior and life ch oi ce s , the practice
would raise difficult ethical issues about what is
“good” behavior and what is “bad.” (For example,
should overweight people and sky divers be charged
higher rates than thin people and long-distance
runners or skiers, and would that be fair?)
On the whole,it seems more fair to use community rating than risk rating if there are no practical
reasons not to do so. The essence of the insurance
principle is to share risk—those who do not need
expensive medical care during a period subsidize
those who do. If the risk is not shared “at the front
end” through community rating, then it has to be
shared “at the back end”through some other mechanism, such as a more direct subsidy to those who
have incurred very high medical expenses. But the
cost of financing that subsidy has to come from the
people who do not incur high medical expenses in
either case. Community rating is simpler and fairer
than the alternative ways of sharing risk.
Objections to community rating that have
merit in today’s insurance market are largely
negated by the provisions of our approach. People
argue that community rating raises rates for lowrisk populations (as it undoubtedly does to some
degree), and that, as a result, some of these people
find that the higher price exceeds the value they
attach to having coverage. So they drop coverage,
which increases the number of uninsured. But our
proposal would require that everyone buy coverage
and would help them to do so with subsidies based
on financial need.
People also argue that the current age-rated system achieves a kind of rough equity: low-risk people
are more likely to be young, and younger people
have lower incomes, on average; so it is fair that they
should pay less, because they have less ability to pay.
But our approach addresses this concern by linking
subsidies to income so that everyone has the ability
to pay.
The efficiency argument for community rating
and against risk rating is that risk rating is a wasteful
process. It requires expenditure of resources to segregate people into risk categories, a process that
does nothing to enhance the welfare of insured people. It does not expand the amount of medical care,
improve quality, or enhance efficiency.
If risk rating were allowed, some higher- ri s k
groups and individuals would face rates that could
make coverage unaffordable unless the subsidies
were va ri ed according to risk. Linking the size of
subsidies to the level of risk would pose immense
administrative complications and would be an
expensive undertaking.
On the other hand, a practical argument for risk
rating is that it allows individual insurers to offset at
least partially the effects of drawing a population
whose risk is not representative. Insurers that happen
to attract higher-risk people can afford to cover the
higher medical costs they incur by charging enrollees
above-average premiums. But this is not a practical
long-run solution because, as they raise their premiums, insurers will lose the lower-risk people they
cover because those people will switch to less expensive health plans. This practical argument needs to be
addressed through a risk-adjustment process that
somehow compensates insurers that cover a disproportionate number of higher-risk people.
We would allow premiums to be adjusted for
family size and composition, a provision that is virtually a universal practice now and is presumably
not controversial. The other rating factor we would
a ll ow is for differences in regional medical costs.
This also fo ll ows current practice and is en ti rely
consistent with the community rating principle. The
very term implies that uniform rates should apply to
207
a limited geographic area. Most purchasers tie the
rates they pay HMOs to local medical costs. Since
medical costs vary significantly from region to
region,and insurers’ business is concentrated in different regi on s , insurers need to be able to protect
themselves by charging rates that reflect those
regional differences. Further, it could be argued that
people who live in higher-cost areas should have to
pay more so that they have incentives to seek ways to
hold costs down. Their costs should not be subsidized by people who live in lower-cost areas.
Although the more common practice is to define
small groups as firms with  or fewer employees,
we choose to include groups with  or fewer
employees within the community rating pool. The
inclusion of larger employers broadens the risk pool
substantially, which means that the costs of covering
higher-risk individuals and small groups is spread
over a larger portion of the pop u l a ti on . It is also
questionable whether groups of between  and 
employees are large enough to be risk rated separately or to self-insure and be at risk for the costs of
their employees’ medical expenses.
Elimination of Medicaid, S-CHIP, and Other
Public Programs as a Source of Coverage
 
Medicaid, S-CHIP, and similar programs to fund
coverage for low-income people would be eliminated
and would be replaced with subsidies in the form of
refundable tax credits (described earlier) that allow
people who otherwise would be eligible for these
programs to purchase coverage in private markets
just as everybody else does. (The important exception would be that Medicaid would continue to fund
and administer the long-term care portion of the
program. States would have increased responsibility
for financing long-term care, as described later.)
We recognize that not all low-income people
will be well equipped to deal with the private market, especially because some will be unemployed
and thus will not have coverage through their jobs.
For these people, the APAs that each state must
establish can serve as an appropriate source of coverage. In fact, the state may decide to make an APA
responsible for negotiating with health plans, per-
haps through a competitive bidding process, to offer
managed care coverage that is specially tailored to
the needs of this population, but open to anyone
who chooses it.
States would be required to con ti nue to have
mechanisms to integrate services, provide case management, and otherwise meet the unique needs of
many of the people with disabilities receiving
Supplemental Security Income (SSI) cash assistance
and Medicaid. They also would need effective outreach programs to identify these special populations — which might be somewhat more difficult,
because they could no longer be identified as they
en ro ll in Medicaid. In most instances, existing
Medicaid program structures would con ti nue to
serve these functions, but Medicaid would no longer
be the source of funding. Some supplemental funding might be necessary from the states to cover services that normally are not considered to be medical
in nature but that these people need if they are to
improve their health. Examples include speech therapy and transportation services.
 
Because our proposed system provides low-income
people with tax credits adequate to purchase coverage equivalent to the benefits provided by Medicaid,
there is no ongoing need for Medicaid, S-CHIP, and
other similar programs that subsidize care for the
poor. An argument could be made for phasing out
these programs in stages, especially Medicaid, to
ensure a smooth transition for vulnerable populations that may not be well equipped to find private
coverage in a private system that may not be fully
prepared to meet their needs. On the other hand,
phasing out one system while phasing in another
adds a layer of complexity—for example, the need
to establish a mechanism to let people choose
between getting tax credits or temporarily retaining
coverage under Medicaid or S-CHIP.
Programs providing highly specialized services
that aid very specific populations, such as Maternal
and Child Health Block Grants and the Ryan White
program, could be retained.
The tax credit subsidy makes private marketbased coverage affordable for low-income people.
208
Having them purchase coverage just as higherincome people do eliminates the stigma often associated with medical “welfare,” and it gives these people
access to the same providers available to the rest of
the population. The multi-tiered arrangement that
characterizes our current system is eliminated. Lowincome people get “mainstream” care in the same
way that elderly people do. They will not face the discrimination from providers and lack of access that is
now often the lot of Medicaid recipients, because,as
far as providers are concerned, they will be indistinguishable from the population that receives smaller
subsidies (the tax credit available to everybody).
Elimination of these low-income programs also
eliminates the expensive and burdensome process of
determining and redetermining eligibility and moving people from one insurance system to another as
their eligibility status changes. Everyone is eligible
for tax credits,and the amount of the credit is determined through the income tax reporting system.
Low-income people are like everybody else except
that they receive a larger tax credit.
Substituting private insurance for Medicaid
coverage will increase the budget cost, at least in the
short run, because providers will be reimbursed at
market rates rather than lower government-constrained rates—though presumably health plans
and employers will negotiate vigorously to keep
provider rates as low as possible throughout the
system, just as they do now. But in terms of equity
and ensuring access, there is no defensible justification for paying providers less for serving the
low-income population than for serving any other
population.
The extra cost associated with higher provider
reimbursement rates may be at least partially offset,
for two reasons. First, providers currently try to
recoup some of the fee discounts they have to accept
from Medicaid by charging other payers more. Such
attempts at cost shifting will no longer be necessary,
which should benefit current non-Medicaid payers.
Second, people who are now uninsured will get better care, which should have a favorable impact on
costs. They will have access to primary care and preventive services, which not only will improve their
health status, but also reduce costly use of hospital
em er gency rooms and prevent simple probl em s
from developing into acute problems that are
expensive to treat.
Centralized Electronic Administration
 
Determination of eligibility, claims submission,
coordination of benefits, and similar administrative
processes would be channeled through a centralized
electronic clearinghouse that would serve all insurers and health plans. The role of this entity would be
analogous to the role the Federal Reserve System
fills for the nation’s banking system plus the role
served by centralized administration of credit card
transactions. All health plans and insurers would be
required to participate, to accept the common data
format and procedures of the system, and to share
in the costs of the clearinghouse.
All transactions would be handled electronically.
Every person en ro ll ed in a health plan would be
issued a card (comparable to an automatic teller
machine [ATM] card or “smart card”) with electronically embedded information sufficient to serve
the functions of the system. Every health care
provider would have a card reader connected to the
centralized system (just as most retailers are connected to a centralized administra ti on for credit
card transactions), and every patient encounter
would begin with reading the pati en t’s card. That
reading would provide all necessary inform a ti on
about eligibility, covered benefits, amount of copayment, whether the deductible has been met, etc. Any
payments made by the patient at the time of the visit
would be entered. Although it might be desirable to
include information of a medical nature as part of
this smart-card mechanism, this raises important
privacy and confidentiality issues that are beyond
the scope of this analysis.
Patients who mispla ce, forget, or lack an identifying card would be entered into the system by
name, Social Security number, and mother’s maiden
name. Services could not be denied to patients
because they lack a card, and Medicare would guarantee that providers are paid for covered services if
no other insurer can be identified for the patient.
Federal legislation would be required to put this
209
system in place to ensure that the infrastructure is
created and that health plans and providers comply
with the requirements for standardization. We have
no strong preference about whether the organization that carries out these activities is government
or private. We noted that somewhat analogous institutions are the Federal Reserve check-clearing system, which is public, and the credit card clearing
system, which is private. Both seem to work well. In
either case, the ongoing costs of operating the system should be recovered through fees levied on
health plans, insurers, and other risk-bearing entities, including self-insured employers. The federal
government probably would need to appropriate
funds to cover some of the initial costs of establishing the system.
 
The current system for administering submission of
claims, determining el i gi bi l i ty, calculating copayment obligations, and coordinating benefits is woefully deficient, duplicative, and inefficient. Patients
often are responsible for keeping track of and submitting claims from many different providers for
each serious episode of care. They get bi ll ed for
services that are obligations of insurers, and they
neglect to submit many claims that insurers should
be paying. Providers waste huge amounts of time
and money submitting and resubmitting claims and
billing patients and insurers, processes that are especially inefficient when the patient is covered by multiple insurers. Insurers have to send multiple reports
to patients and providers indicating what has been
paid and what is the patient’s obligation.
Using a centralized clearinghouse along with
electronic submission of information would greatly
reduce these administrative burdens and costs borne
by patients, providers, and insurers. Administration
still would be complicated and expensive: insurers
still would have to approve treatment plans,authorize services, etc. But much of the inefficiency and
cost and many of the hassles of the current system
could be eliminated. In particular, patients would be
relieved of the administrative burdens that now rest
on them.
Financing
 
Health care coverage would continue to be financed
by employers ,i n d ividuals and families, and government, but the ways in which the federal and state governments finance subsidies would change radically.
Programs of medical coverage for specific population
groups now subsidized by federal and state governments—most notably, Medicaid and S-CHIP—
would be eliminated, with the important exceptions
of Medicare for the elderly and Medicaid for longterm care. Medicaid and S-CHIP would be replaced
with tax credit subsidies, as explained earlier.
The revenues to finance tax subsidies (which are
available to everyone but at different levels, depending on income) would come from two federal
sources—general tax revenues and the special tax
assessments on individuals defaulting to Medicare
coverage. The general revenue money would not be
pri m a ri ly new net spending, however. The federal
government would experience large revenue increases by eliminating the tax exclusion of income that
employees receive in the form of employer-paid
health insurance premiums. Both income tax and
payroll tax revenues would rise when the exclusion
provision is removed. In addition, the federal government would no longer finance Medicaid or SCHIP. These changes would result in more money in
the general fund, which would go far toward financing the new income tax credits. Some additional new
public resources probably would be required, however—a political challenge that is easier to manage in
a period of large budget surpluses.
Eliminating Medicaid and S-CHIP would relieve
states of substantial funding burdens. We would
propose that, in exchange, states be required to
assume greater responsibility for current long-term
care services provided under Medicaid. States initially would be required to finance the costs of this
program fully, with the benefit limits as currently
defined, if they can do so without exceeding their
present Medicaid and S-CHIP obligation. They also
would be required to finance certain non-medical
but medically related services now covered by
Medicaid for special needs populations and the
near-poor elderly. Initially states would be required
210
to maintain their current level of Medicaid effort,
and, if the cost exceeded that level, the federal government would make up the difference. Over time,
states’ obligation would be phased gradually to a
system in which their share is based on some measure of ability to pay, such as state per capita income,
rather than on previous levels of funding. The effect
would be to cap their obligation, with the federal
government assuming costs beyond the states’ levels
of obligation.
 
The changes in funding proposed are based on two
judgments: () that income redistribution activities
(which any subsidy program is) should be the
responsibility of primarily the federal government,
and () that the levels of subsidies should be based
on need, not on the recipient’s place of residence.
We take the position that, for equity reasons, the
amount an individual contributes in the form of
taxes to fund the subsidy program should be based
on ability to pay (that is, income), not on the state
where the individual lives. Likewise, the amount of
subsidy provided to an individual should be based
on that person’s ability to pay for coverage, not on
where he or she lives or on a state’s ability and willingness to provide such subsidies. The capacity of
states to fund subsidies varies widely, and their ability to pay is likely to fluctuate widely with changes in
state or regional economic conditions. This proposition leads to the conclusion that establishing subsidy standards and funding the cost is primarily a
federal responsibility.
It is not unreasonable,however, to require states
to maintain some level of effort rather than experiencing a large windfall gain, with all of the subsidy
cost falling on the federal government. Thus we propose increased state responsibility for long-term care,
but, consistent with our earlier line of reasoning, we
would propose a gradual transition from requiring
states to maintain current levels of effort to contributing according to their citizens’ ability to pay,
which we measure by state per capita income.States
may be wary of accepting the responsibility for funding long-term care, because, given the aging of the
population and other trends,long-term care costs are
likely to rise more rapidly than most other parts of
the health care system. A federal cap on states’ longterm obligation, with the federal government picking
up the excess, may be a reasonable trade for states’
continued acceptance of meeting federal standards
for quality and other aspects of long-term care.
Risk Adjustment
Because our proposal requires insurers to use community rating for the individual and small-group
markets,they cannot use rate adjustments to protect
themselves against getting a disproportionate number of h i gh - risk en ro ll ee s . Under these circumstances, a strong case can be made for developing a
mechanism to compensate insurers operating in
these markets for differences in the risk profiles of
the people they insure. The public interest requires
that insurers be rewarded for being efficient in
administering and providing high-quality medical
services, not for being skillful in selecting and
attracting low-risk populations and avoiding highrisk populations. Without some method for compensating insurers for differences in the risk of the
populations they cover, insurers have strong incentives to risk-select; and experience suggests that it is
very difficult to prevent risk selection through legislative prohibitions. But even if insurers did not
inten ti on a lly seek to attract low-risk people and
avoid high-risk people, some would get more than
their fair share of high-risk people,partly because of
random factors and partly because certain kinds of
plans appeal to people with certain kinds of risk
profiles. Health plans that gain a rep ut a ti on for
being particularly skilled at treating people with certain kinds of severe medical conditions, for example, could be especially vulnerable to adverse
selection. In su rers drawing higher-risk people
would be at a competitive disadvantage unless a
mechanism were in place to compensate them,
essentially through some kind of money transfer
from insurers with a relatively high proportion of
low-risk enrollees to insurers with a disproportionate share of high-risk enrollees.
The problem is that the state of the art in risk
adjustment is still in the developmental stage. Much
conceptual work is being done, and a number of
211
experiments are underway, but there is still some
question whether these techniques are sufficient for
the job. There is no way to predict risk completely so
that insurers can be compensated fully and accurately before the fact, and techniques that compensate on the basis of after-the-fact incurred expenses
decrease incentives for health plans to contain costs.
We are not prepared to endorse any particular riskadjustment mechanism or approach; we leave that
task to people with expertise in this area. But we do
think that efforts to develop workable approaches
should be continued and accelerated.
It is uncertain how critical a risk-adjustment
process is to the success of our approach (or, for that
matter, to most other approaches that depend on
private insurers). Most of the market operates now
without risk adjustment, with at least some success.
The fact that everyone would be required to have
coverage under our plan helps in some respects,
because it en su res that everyone, both high- and
low-risk people, will be in some pool. The problem
of having people buy insu ra n ce only when they
anticipate needing expensive care also is eliminated.
Transition
The reform proposed here is not an incremental
change. It does not build on existing programs of
public coverage. It requires some major restructuring, particularly at the government level, and
imposes new obligations on individuals, employers,
and health plans. Designing a gradual transition
from the current system to the new system is not an
easy task because many of the changes must become
operational at the same time.
Some steps can be taken before the program is
implemented fully. The APAs can be established in
each state, and, once they are ready to begin operations,all small employers with fewer than  employees could be required to use them as the vehicle for
offering coverage. Those that provide coverage
already would switch to the APA when their existing
health plan comes up for renewal. At the same time,
the requirement that all employers offer (but not
necessarily pay for) coverage for their employees
could be implemented .E m p l oyers with  or fewer
employees would be required to use the APA, and
many others probably would choose to do so.
The requirement that all health plans price premiums on a community-rated basis for employers
with  or fewer employees can and should be
implemented gradually. States vary considerably in
the extent to which they limit health plans’ ability to
vary premium rates for small employers, and it
would be too disruptive to require them all to move
from their current position to full community rating over a short period of time. In fact,implementation of full community rating should not begin
until the individual mandate is in place. Otherwise,
premium rates for low-risk groups may rise to such
an extent that significant numbers of them would
drop coverage. Thus the requirement for community rating should be phased in over a period of several years for the small-group market. Any movement
toward community rating and guaranteed-issue
requirements for the individual market, however,
probably will have to wait until the individual mandate and tax credits are in place. Otherwise, health
plans likely would suffer adverse selection because
sicker people would take advantage of the community rates to buy covera ge , while healthier people
would wait until the individual mandate required
them to do so.
A number of the most important features of the
reform must become operational at the same time
to avoid creating severe problems. They include the
individual mandate, full community rating, tax
credit subsidies, elimination of the tax exclusion of
employer-paid premiums,health plan premium risk
adjustment,and requirements that individuals show
proof of coverage as part of the tax-filing process. It
would be highly desirable to implement the electronic system for paying and reconciling medical
claims at the same time these other features of the
s ys tem are put in place, but it probably would be
possible to begin the rest of the new program without having the electronic system fully operational.
Cost Containment
Because this proposal would extend coverage to
everyone currently without insurance, it likely
212
would increase the demand for medical services and
raise the total level of health expenditures. But apart
from that, nothing in the proposal should be a
strong force to increase costs. Nevertheless, private
funders of insurance coverage obviously would continue to be concerned about costs, and cost escalation would have special implications for the federal
government,since rising costs would create pressure
to increase the size of the tax credits and, thus, the
revenue loss associated with this form of subsidy.
Without tax credit increases, coverage would
become unaffordable for many people if health
costs rose appreciably. Clearly, continued attention
will need to be directed to efforts to contain medical
cost escalation. The proposal does not include any
new forms of cost control, but it does incorporate
features that should strengthen existing competitive
market forces and create stronger incentives for
consumers to be cost-conscious.
One necessary condition for market forces to
work to contain costs is the presence of incentives
that encourage cost-conscious behavior. This proposal helps to create appropriate incentives by eliminating the tax exclusion for employer-paid health
premiums, which en co u ra ges people to buy more
comprehensive health insurance than they would
otherwise. The current tax exclusion subsidizes any
level of health coverage paid for by the employer
and, thus, encourages employees to prefer extensive
coverage, making it almost costless for them to consume any well-insured health care services. The proposed reform provides subsidies only for
purchasing the standard coverage package; that is,
individuals who choose to buy more comprehensive
coverage would pay all of the cost difference
between that benefit package and the standard coverage pack a ge . As a consequence, they would be
more likely to carefully weigh the benefits against
the costs,and fewer people would buy very compreh en s ive covera ge . Because they would be paying
more out of pocket than they do now for at least
some services, people are likely to reduce their rates
of utilization, especially for services that are only
marginally beneficial.
For the same reason that people would tend to
choose less comprehensive plans, they also would
have strong incentives to choose plans that are efficient and offer high value. The new tax credit subsidy would be a fixed amount unaffected by the cost
of the health plan chosen. Thus people who chose a
plan that costs more than their subsidy (which
would be most people) would have to pay the full
extra cost out of pocket. That is a strong incentive
not to choose an inefficient, costly plan.
Of course, incentives to choose a higher-value
plan have little effect when people have few plans
from which to choose, as is often true today. The
proposed approach would give many people more
plan options: everyone acquiring coverage through
the APA could choose from a number of plans. To a
greater extent than currently, health plans would be
in head-to-head competition for consumers’ business, so they would have stronger incentives to offer
plans that provide high value. Moreover, health
plans would not have the option available to them
now of being able to compete on the basis of risk
selection, because for employers with fewer than 
employees, premiums would be determined on a
guaranteed-issue, community-rated basis; and the
risk-adjustment process, to the extent that it is accurate and effective, would greatly reduce the rewards
associated with being skillful at risk segmentation.
Fewer resources would be devoted to finding ways
to avoid high-risk enrollees, an effort that produces
no real social benefit. The only remaining basis for
competi ti on would be to offer good-quality care
and high levels of service at reasonable prices.
Apart from the fact that people previously uninsured would use more medical services than before,
there is nothing obvious in this proposal that is likely to cause cost escalation. It is true that a significant
portion of medical payments now subject to government price administration—namely, for services
covered through Medicaid—would be reimbursed
at market rates, which are likely to be higher than
administered-price rates. A one-time price increase
is a likely result, but over the longer term, reimbursement rates are not likely to rise at a greater rate
than they would under an administered-price
arrangement, because even administered pri ce s
have to rise at roughly the same rate as market prices
to induce providers to offer services.
213
Whether market forces will be adequate to contain costs remains to be seen, but nothing in this proposal is likely to reduce the prospects for success—
quite the contrary. But if additional cost containment
strategies have to be pursued, it is likely that they
would have been necessary even in the absence of the
implementation of this proposed reform.
Quality of Care
Implementation of this proposal is likely to have a
positive effect on quality of care in two ways. First
and most obvious, by ensuring that everyone is covered, the reform eliminates financial barriers to
access. People who now defer or deny themselves
care because they lack coverage will no longer have
any reason to do so, so problems can be detected
sooner, when treatment is more effective and less
expensive. People now enrolled in Medicaid and
similar public programs often have difficulty finding providers willing to accept them, either because
the providers do not participate or because they
already have a full roster of public-program
patients. This constraint would be greatly reduced.
Now covered by private insurance, these people
would have access to the wider range of providers
serving their area.
Second,the increased head-to-head competition
among health plans just described will force them to
prove that they provide good-quality care as part o f
their efforts to convince potential enrollees that the
plans offer a high-value product.
Implementation of this reform is likely to leave
largely unchanged the other forces that influ en ce
quality. Employers still will have reason to be concerned about quality, since they will be the source of
coverage for much of the population. Medicaid no
longer will fund care for low-income people (apart
from long-term care), so government’s responsibility for quality will need to be extended beyond concerns about low-income populations to the entire
population. Activities of government organizations
such as the Agency for Healthcare Research and
Quality will be even more vital than they are now.
Special attention should be directed to the effects of
the reform on low-income populations formerly
served by Medicaid and S-CHIP to ensure that the
new system meets their needs,as well as those of the
general population.
Political Feasibility
Like any proposal that represents a major departure
from the status quo, this approach to reform would
not be free of opposition. Nevertheless, the proposal does offer a number of significant political
advantages.
Advantages
The reform model presented here should appeal to
traditional conservatives for several reasons. First, it
puts everyone into the mainstream medical system
rather than into a government-run “bureaucratic”
system. Second, it eliminates existing governmentrun programs like Medicaid and S-CHIP, which,
because they are subsidy programs directed essentially at the poor, are often perceived negatively as
“welfare” programs. In place of these programs, the
proposal substitutes tax credits, which conservatives
tend to favor over direct government-financed programs and which can legitimately be sold as a form
of tax reduction. Third,the proposal places few constraints on employers. Apart from being required to
allow their employees to opt into the aggregate purchasing arrangement and to offer but not pay for
health coverage, employers’ role in providing coverage for their employees remains largely unchanged.
Fourth, the proposal relies on competitive market
forces to contain health care costs, rather than introducing new forms of government-administered
price regulation or cost controls. Moreover, by eliminating the tax exclusion for employer-financed
health premiums, the proposal would require people to bear more responsibility for their health care
costs.
Major stakeholders who are often threatened by
proposals that would extend coverage broadly may
find that this proposal is more attractive than other
reform alternatives. Insurers and health plans would
continue to play the role they do now, and, in fact,
they would have as new customers large numbers of
people who currently are covered by government
214
programs. Providers, too, should find many aspects
of this proposal appealing. Most notably, providers
that now serve large numbers of Medicaid patients
would be paid at market rates rather than at the normally lower government-determined rei m bu rs ement rates. In addition, there would no longer be
any uncompensated care, because everyone would
be required to have insurance coverage, and even
those individuals who failed to meet the requirement would have their medical bills paid by
Medicare. Providers who now serve recipients of
public programs also would be freed from the
administrative burdens of dealing with the government bureaucracies that administer programs for
low-income patients. And all providers would be
freed of many administrative burdens when the integrated electronic claims handling system is in place.
Advocates for low-income and disadvantaged
populations obviously would find the universal coverage feature of this program to be highly desirable.
They also would likely look favorably on the feature
that provides tax credit subsidies to everyone, so that
there is no income test for eligibility and, therefore,
no stigma for the poor in accepting such subsidies.
The multi-tiered system of care that characteri ze s
our current system would be eliminated. Everyone
would be served by “mainstream” providers, and
providers would be unable to distinguish among
patients according to the subsidy they receive. Access
for low-income people also would be improved.
Those whose access is now limited to providers who
accept patients from public programs would benefit
by being able to choose from the same range of
providers as the rest of the population. Extension of
choice is obviously desirable for its own sake, but the
greater choice of providers also should enhance the
quality of care available to people who are now
served by public programs.
Disadvantages
Although the proposal is likely to appeal to many
groups, it also faces formidable political obstacles.
First, the program involves a large tax expenditure.
Even though the tax credits that subsidize purchases
of health coverage for the entire population can be
characterized as a form of tax reducti on , it is still
true that the subsidies under this program would
cause large losses in tax revenue. While it is probably
easier for politicians to defend off-budget tax
expenditures than on-budget appropriations, people who look with disfavor on policies that expand
the role of government are not likely to be mollified.
And, of course, to the extent that this proposal
reduces tax revenues, it reduces the revenue available to spend on other public programs. Offsetting
the negative effects of increased tax expenditures for
tax credit subsidies is the fact that other changes,
particularly taxation of employer-paid health insurance premiums and elimination of Medicaid and SCHIP, will have a large positive effect on the budget
position of the federal and state governments (and,
incidentally, alleviate some of the revenue problems
of Medicare and Social Security, since payroll tax
revenues will also increase).
Some people will view this program as more
“big government.” Even though it requires no significant buildup of new government bureaucracy or
personnel, relies on private-sector forces, and actually increases the number of people who will be
buying coverage in the private sector, it will be characterized by some as a major expansion of governm en t’s role and influ en ce . Some will criticize the
idea of using Medicare as a fallback for those who
have no coverage, even though, in most cases, this
would be a temporary arrangement. Vexed by the
current rules and regulations of the Health Care
Financing Administration, they would not eagerly
greet the idea of having the agency administer coverage for even more people. Critics with this perspective also may fear that if cost increases begin to
appear in the health care sector, government cost
controls will not be far behind.
This approach also involves a degree of compulsion that some will find objectionable. The requirement that everyone purchase health care coverage is
a restriction on individual liberty, and while it is certainly possible to marshal persuasive arguments to
justify this level of compulsion, the arguments will
not convince everyone. The proposal also requires
employers to take certain actions. Specifically, they
are required to offer (but not pay for) coverage to all
employees and to allow individual employees to opt
215
into the aggregate purchasing arrangement.Employers that do not offer coverage now, or those who
have well-established insurance programs of their
own, will have to make changes that may appear
burdensome to them.
Like any major reform,this proposal creates various kinds of financial redistribution that are likely
to be opposed by people who view themselves as
being worse off as a result of the change.
Eliminating the tax exclusion for employer-paid
premiums is likely to cost higher-income people
with comprehensive employer-paid coverage more
than they gain in the way of a tax credit. Thus, they
may oppose this change in the approach to financing subsidies for health covera ge . Opposition to
eliminating the tax exclusion is likely to be particularly strong from labor unions that have bargained
successfully for generous health coverage programs.
The requirement that all firms with fewer than 
employees be part of a single risk pool will raise the
cost of health coverage for lower-risk employers and
their employees: Because all employers in the pool
will be charged the same community rate, low-risk
employers and their employees will no longer be
able to realize their current risk advantage.
Although insurers and health plans would play a
larger role in the revised system, some will still
object to specific provisions. In general, insurers are
not enthusiastic about community rating because
they believe that constraints on their ability to
adjust rates in accordance with the risk of the populations they insure can jeopardize their ability to
remain profitable. Even if an appropriate riskadjustment process is in place, as proposed, it may
be hard to convince insurers that their worries are
groundless. Moreover, some insurers have carved
out a profitable niche for themselves that is based on
their effectiveness in selecting low-risk populations.
Since their competitive advantage up to now has
been attributable to their being more skillful in
selecting risk, rather than in being cost-effective in
managing care, these insurers and health plans may
oppose the changes envisioned in this proposal.
Advocates for low-income populations may be
wary of the proposal because it does away with
much of Medicaid. They have worked long and hard
to en su re that Medicaid incorporates fe a tu res to
protect lower-income populations and to be attentive and responsive to their special needs. Even
though we propose retaining certain elements of
Medicaid to meet these special needs, advocates may
still worry that placing these vulnerable populations
into mainstream care may leave them without all
the protections and special services they need. n