Reforming Old Age Security: Effects and Alternatives

Transcription

canadian tax journal / revue fiscale canadienne (2015) 63:2, 357 - 73
Reforming Old Age Security:
Effects and Alternatives
Nicholas-James Clavet, Jean-Yves Duclos, Bernard Fortin,
and Steeve Marchand*
Précis
Dans le budget de 2012, le gouvernement fédéral a annoncé son intention de hausser
l’âge d’admissibilité à la Sécurité de la vieillesse et au supplément de revenu garanti de
65 à 67 ans. La hausse sera instaurée graduellement à compter de 2023. Lorsque la
politique sera entièrement mise en œuvre en 2030, elle accroîtra les revenus nets du
gouvernement fédéral de 7,1 milliards $ par année, mais réduira les revenus nets
provinciaux de 638 millions $ (en dollars constants de 2014). En supposant qu’il n’y a
aucun changement dans les comportements liés à la main-d’œuvre et à l’épargne, le fait
de reporter la date d’admissibilité fera passer de 6 à 17 pour cent le nombre de particuliers
âgés de 65 et 66 ans faisant partie du groupe à faibles revenus (pour 100 000 aînés à
faibles revenus de plus dans cette catégorie d’âge), et sera le plus néfaste pour les aînés
à faibles revenus et les femmes. Des réformes différentes à la Sécurité de la vieillesse
pourraient permettre de réaliser des gains nets pour les finances publiques sans avoir de
telles incidences négatives sur le taux de faibles revenus parmi les aînés.
Abstract
The federal government announced in its 2012 budget its intention to increase the age of
eligibility for old age security and the guaranteed income supplement from 65 to 67 years.
The increase will be introduced gradually, beginning in 2023. When the policy is fully
implemented in 2030, it will increase the net revenues of the federal government by
$7.1 billion per year, but reduce net provincial revenues by $638 million (in constant 2014
dollars). Assuming no change in labour and savings behaviour, delaying the date of
eligibility will also increase the percentage of individuals aged 65 and 66 years who are
in the low-income group, from 6 percent to 17 percent (for an additional 100,000 lowincome seniors in this age group), and will be most harmful to low-income seniors and to
women. Alternative reforms to old age security could make it possible to achieve net
gains for public finances without having such large negative impacts on the low-income
rate among seniors.
Keywords: Protection n reforms n Old Age Security n poverty n public finance n Canada
* Of Université Laval, Quebec (e-mail: [email protected]; jean-yves.duclos@
ecn.ulaval.ca; [email protected]; [email protected]).
 357
358  n  canadian tax journal / revue fiscale canadienne
(2015) 63:2
Contents
Introduction
Effects of the Announced Reform
Effects on Public Finances
Effects on Low Income
Alternative Scenarios
Scenario 1: A Decrease in the OAS Clawback Threshold
Scenario 2: A Uniform Decrease in OAS Benefits
Scenario 3: A Progressive Increase in Benefits
Can Individuals Affected by the Reform Avoid Falling Below the Low-Income
Threshold?
Conclusion
Appendix
Gross Old Age Security (OAS) Payments
Guaranteed Income Supplement (GIS) Payments
OAS Clawback
Federal Income Tax
Provincial Income Tax
Social Assistance
358
360
361
361
365
365
366
368
368
371
372
372
372
372
373
373
373
Introduc tion
The government of Canada announced a major reform to old age security (oas) and
the guaranteed income supplement (gis) in its 2012 budget.1 The oas regime (as of
2014) makes annual payments to nearly every Canadian aged 65 and over, up to a
maximum of $6,7652 per year, subject to a clawback at the rate of 15 percent beginning once net income exceeds $71,592. Seniors residing in Canada for less than
10 years do not receive any oas, and those who have lived in the country for 10 to 40
years receive partial payments. gis payments are made to Canadians who qualify for
oas and whose other sources of income are limited. The maximum amount of the gis
is $9,173 for single people and $12,165 for couples, including a new increase introduced in 2011. The amount of the benefit declines with family income, reaching zero
when total pre-oas family income is $17,064 for singles and $22,512 for couples.
In the 2012 budget, the government announced that, starting in 2023, the age
to qualify for oas and gis payments would be increased by four months annually to
reach 67 years of age by 2029. This article analyzes the effects that this policy is likely
to have on federal and provincial public finances as well as on poverty among seniors.
The effects of the reform on federal public finances have been partially studied by
the Office of the Chief Actuary (oca).3 The oca uses demographic forecasts and
1See Canada, Department of Finance, 2012 Budget, Budget Speech, March 29, 2012, at 6.
2All amounts in this paragraph are annual payments based on October 2014 effective amounts.
3 Office of the Superintendent of Financial Institutions, Office of the Chief Actuary, Actuarial
Report (11th) Supplementing the Actuarial Report on the Old Age Security Program as at
31 December 2009 (Ottawa: OSFI, 2012).
reforming old age security: effects and alternatives  n  359
extrapolates from past tendencies to predict the resulting decreases in oas and gis
payments in upcoming decades. Our goal in this article is to supplement the oca’s
analysis using a microeconomic approach that can estimate the direct and indirect
effects of the reform on public finances as well as on household incomes. We also
consider the effects that the announced reform would have on provincial and federal
income taxes, oas clawback payments, provincial social assistance, and low-income
rates. Moreover, we estimate the impact of three alternative reform scenarios that
produce similar impacts on public finances; these scenarios are compared with the
reform scenario proposed in the 2012 budget in terms of both their effects on public finances and their impact on poverty among seniors.
The analysis of these changes to oas and gis is of considerable importance.
These two programs represent 17.3 percent of federal government program
expenses and 2.4 percent of gross domestic product.
n Changes in oas and gis payments are likely to generate substantial impacts
on tax revenues for both the federal and provincial governments.
n The design of oas and gis clawbacks is of primary tax-planning importance
for middle-income Canadians.
n Finally, from a social-planning perspective, the increase in life expectancy has
led to a generational increase in oas and gis claims, and this trend is expected
to continue with the continued aging of the population.4
n
Our calculations are made using SIMUL, a dynamic microsimulation model developed by members of the Industrial Alliance Research Chair on the Economics
of Demographic Change.5 Dynamic microsimulation models are used in many
countries to predict long-run distributions of incomes as a function of individual
characteristics 6 or to predict economic effects of different redistribution policies.7
They can also be used to assess how different retirement schemes may affect households in the long run.8
4 Frank T. Denton and Byron Grant Spencer, “Age of Pension Eligibility, Gains in Life
Expectancy, and Social Policy” (2011) 37:2 Canadian Public Policy 183-99.
5See www.cedia.ca.
6See, for instance, Marike Knoef, Rob J.M. Alessie, and Adriaan Kalwij, “Changes in the
Income Distribution of the Dutch Elderly Between 1989 and 2020: Dynamic Microsimulation”
(2013) 59:3 Review of Income and Wealth 460-85.
7See Ramses H. Abul Naga, Christophe Kolodziejczyk, and Tobias Müller, “The Redistributive
Impact of Alternative Income Maintenance Schemes: A Microsimulation Study Using Swiss
Household Data” (2008) 54:2 Review of Income and Wealth 193-219.
8See Geof Rowe and Michael Wolfson, “Public Pensions—Canadian Analysis Based on the
Lifepaths Generational Accounting Framework,” paper presented at the 6th Nordic Seminar
on Microsimulation Models, Copenhagen, June 8-9, 2000; and Jad-Maarten Van Sonsbeek,
“Micro Simulations on the Effects of Ageing-Related Policy Measures” (2010) 27:5 Economic
Modelling 968-79.
(2015) 63:2
Our model simulates demographic, economic, and fiscal evolution with regard to
the population of Quebec up to 2030. Details of the model can be found in a paper
we published in June 2012.9 The model can also predict the long-term effects of fiscal
reforms, accounting for anticipated demographic changes and the evolution of income distribution, personal income taxes, and social transfers. Because the model’s
calculations are made at the micro level, all of the relevant individual characteristics
that are needed to calculate oas and gis are taken into account. As mentioned
above, these characteristics include age, individual income, total family income, and
years of residence in Canada. A number of income sources are modelled for each
member of the family (labour income, pension income, investment income, etc.).
The estimates of the model, generated using Quebec data, are generalized for Canada
assuming constant demographic, economic, and fiscal ratios between Quebec and
Canada as a whole (see the appendix to this article for details). While imperfect, this
extrapolation seems reasonable, given the presence of other sources of error inherent in long-term forecasting.
The next section details the estimated effects of the announced changes to oas
and gis on federal and provincial finances, as well as the impact on poverty in Canada. We then examine three alternative reform scenarios and compare their effects
with those of the announced reform. We show that while these scenarios would have
a comparable effect on public finances, the impact on poverty would be reduced.
Effec ts of the Announced Reform
Increasing the age of eligibility for oas will obviously decrease direct program costs
for the federal government. This gain in terms of public finances will nevertheless
be partially counterbalanced by two factors. First, individuals aged 65 and 66 years
who will no longer receive oas and gis payments will (if their income is too low) be
eligible for provincial social assistance, and this will increase the costs of these benefits
to the provinces. Second, since oas benefits enter into calculations of taxable income,
federal and provincial taxes paid by individuals aged 65 and 66 will decrease.
These factors have impacts on both public finances and personal incomes. As
indicated above, the size of these impacts is estimated using a microsimulation model
that specifies demographic and economic characteristics of Canadians over the next
20 years and accounts for a major share of the elements of the fiscal system and
transfers to individuals. Note that the model assumes that individuals do not change
their work or savings behaviour as a result of the reform. The likelihood and consequences of the effects of such behavioural changes are discussed in a later section
of the article (see “Can Individuals Affected by the Reform Avoid Falling Below the
Low-Income Threshold?”).
9 Nicholas-James Clavet, Jean-Yves Duclos, Bernard Fortin, and Steeve Marchand, “Le Québec,
2004-2030: Une Analyse De Micro-Simulation,” rapport de project, Centre interuniversitaire
de recherche en analyse des organisations (CIRANO) ( Juin 2012).
Effects on Public Finances
Table 1 presents the estimated effects of the announced reform in 2030, the first year
in which the reform will be fully implemented. The gross effect (without accounting
for impacts on taxes and social assistance) of increasing the age of eligibility for oas
benefits would be to reduce federal spending by about $6.4 billion (in constant 2014
dollars). The corresponding amount for gis would be about $2.1 billion.10 As a result, overall, increasing the age of eligibility for oas and gis would reduce federal
expenditures by about $8.5 billion per year.
These positive effects on federal finances would be partially countered by a decrease in federal income tax of $950 million, and a $388 million decline in oas
clawbacks from those with high income. The net result of increasing the age of eligibility would be to reduce federal expenditures by $7.1 billion in 2030. Public finances
of the provinces would be negatively affected by the reform as a result of the two
effects mentioned above. Since some individuals aged 65 and 66 would have to turn
to social assistance rather than oas and gis benefits, provincial spending on social
assistance would be about $169 million higher in 2030. The provinces would also
experience a $469 million decline in income tax levied on oas payments. The total
cost to the provinces would thus be in the range of $638 million annually. The
overall effect on federal and provincial public finances is therefore estimated to be
about $6.5 billion annually.
Effects on Low Income
The announced reform would have a greater impact on certain categories of individuals. To start with, annual income for the 65-66 age group is less than the population
average. Within that age group, the individuals who would be most affected are
those who would lose their eligibility for oas and gis at the same time. The loss of
gis payments would be only partially counterbalanced by greater reliance on social
assistance payments, because these are less generous than the gis.
Figure 1 presents the concentration curves of the effect of the reform in 2030 by
percentile of disposable income. The concentration curves show the cumulative
proportion of the total decline in revenues affecting a certain lower-income percentile of individuals. For example, about 60 percent of the personal income losses will
be felt by the poorest 50 percent of individuals among the population aged 65 and 66.
Thus, the negative impact of the reform is to a large extent targeted toward the
10 Our OAS predictions compare generally well with those of the OCA. The OCA predicts a fall
of $9.2 billion in 2030 nominal dollars, which, according to its inflation assumptions, corresponds
to a $6.4 billion decrease in 2014 dollars—a figure that is very close to our own estimate. Our
predicted decrease in GIS payments is, however, $890 million higher than the OCA’s 2030
forecast in 2014 dollars. The source of the difference is that GIS predictions depend significantly
on trends in long-run income growth, and our trends (unlike those of the OCA, which rest on
aggregate extrapolations) are based on microeconomic changes in population composition,
human capital, and labour force behaviour.
(2015) 63:2
Table 1 Estimated Effects of the Announced Reform on Public Finances in 2030
millions of constant 2014 dollars
Federal finances
Gross OAS payments . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
OAS clawback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
GIS payments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Federal income tax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6,354
(388)
2,109
(950)
Net federal revenues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7,123
Provincial finances
Social assistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (169)
Provincial income tax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (469)
Net provincial revenues . . . . . . . . . . . . . . . . . . . . . . . . . . . Net federal and provincial revenues . . . . . . . . . . . . . . . . . . (638)
6,485
GIS = guaranteed income supplement.
OAS = old age security.
Source: Authors’ calculations based on the SIMUL model; Office of the Superintendent of
Financial Institutions, Office of the Chief Actuary, Actuarial Report (11th) Supplementing the
Actuarial Report on the Old Age Security Program as at 31 December 2009 (Ottawa: OSFI, 2012); and
Statistics Canada’s social policy simulation database (SPSD), version 21.
Concentration of disposable income loss
FIGURE 1 Concentration Curve of the Effect of the Announced Reform
on Disposable Income in 2030 (Individuals Aged 16 Years and Over
and Those Aged 65-66 Years)
1.0
0.8
0.6
0.4
0.2
0.0
0
0.2
0.4
0.6
0.8
Percentile of disposable income before the reform
16 years and older
65-66 years
Source: Authors’ calculations based on the SIMUL model.
45 degree line
1
lower-income population among those aged 65 and 66. It is also useful to look at
how the reform affects different percentiles of a broader population. If we consider
those 16 years of age and over who have completed their studies, we observe a higher
concentration of the reform. This is so because 65- and 66-year-olds tend to belong
to the lowest percentiles of the income distribution in that broader population. We
see in this case that 70 percent of the income loss would be felt by the poorest
50 percent of individuals.
We also find that the poorest individuals (the poorest 5 percent) are relatively
unaffected by the reform because these poor do not receive much oas, but instead
receive more social assistance. The concentration of the effect increases strongly in
the next percentiles, however. We see that the 20 percent poorest of those aged 16
and over face nearly 40 percent of the total cost of the reform, and that the share of
the decline in income to be faced by the richest 10 percent is just over 5 percent.
Figure 2 presents the predictions for the low-income rate of individuals from
2012 to 2030 among those aged 65 and 66 years with and without the announced
reform. The low-income rate is defined as the proportion of individuals whose family income, adjusted for family size, is below the market basket measure (mbm). This
measure defines low income in relation to the cost of a pre-established set of goods
and services.11
With the announced reform, the age of eligibility for oas pensions and the gis
will increase gradually in each year starting in 2023; this explains the growing difference between these curves starting in that year. We observe that, in the absence
of reform, the low-income rate among those aged 65 and 66 years predicted by the
model goes from 10 percent in 2012 to about 6 percent in 2030. These results flow
in particular both from the expected effect of technological progress and from
growth in the level of education of those aged 65 and 66 years between 2014 and
2030. Since more highly educated individuals tend to receive larger private pensions,
the model predicts an increase in the pension income of the 65-66 age group.12
With the announced reform, individuals 65 and 66 years of age will gradually become ineligible for the oas and gis starting in 2023, and their low-income rate will
increase rapidly to stabilize at about 17 percent in 2030, when the reform will be
fully implemented in its final form. The announced reform will thus have the effect
of increasing the low-income rate among those aged 65 and 66 years from 6 percent
to 17 percent. The reform will thus lead to 100,000 or so additional individuals falling into a low-income situation.
11Since the data used to construct our model do not distinguish between regions within the same
province, we use the Montreal MBM across the board. This measure is adjusted for family size
in order to account for the differences in family needs related to family size. Note that the
Montreal MBM is one of the lowest in Canada and that poverty measures would have been
higher if we had used other regions as a reference.
12It is important to note that our model is better suited to predicting long-run effects of
demographic changes than short-term fluctuations in economic variables. For instance, we
believe we should not put too much faith in figure 2’s shorter-term variations.
FIGURE 2
(2015) 63:2
Low-Income Rate in 2030 Among Individuals Aged 65-66
Years, With and Without the Announced Reform
0.20
0.18
Low-income rate
0.16
0.14
0.12
0.10
0.08
0.06
0.04
0.02
Year
Without reform
With reform
FIGURE 3 Effects of the Reform on Disposable Income of Individuals
Aged 65-66 Years in 2030, by Income Quintile and Gender
0
−5
Loss (%)
−10
−15
−20
−25
−30
−35
−40
1
2
3
Quintile
Men
Women
4
30
29
20
28
20
27
20
26
20
25
20
24
20
23
20
22
20
21
20
20
20
19
20
18
20
17
20
16
20
15
20
14
20
13
20
20
20
12
0.00
5
Figure 3 shows the effects of the reform on the income of individuals aged 65 and
66 years in 2030, by quintile of income (the first quintile groups together the poorest
20 percent of individuals) and by gender. The lower income quintiles lose more than
the higher quintiles. For example, men in the poorest 20 percent lose 35 percent of
their income, while men in the highest quintile lose less than 5 percent of their income as a result of the reform. The effect is broadly similar among women; however,
women systematically lose more than men in every quintile. The difference is considerable in some cases: men in the middle quintile lose about 11 percent of their
income, while women in the same quintile lose 32 percent of their income.
Table 2 summarizes the effects of the reform on disposable income and the lowincome rate, and now differentiates by gender. Women aged 65 and 66 years have
lower disposable income than men of the same age group before the reform (a total
of $12.2 billion versus $13.6 billion for men) but experience greater losses from the
reform. Similarly, the proportion of women who live in a low-income household is
higher than the proportion of men (7.0 percent for women versus 6.0 percent for
men), but this rate increases by more among women than among men (to 19.3 percent for women versus 16.7 percent for men).
A lt e r n at i v e S c e n a r i o s
Considering that the announced reform risks having significant effects on the lowincome rate, is it possible to consider alternative reform scenarios that would have
similar impacts on public finances but less negative impacts on low-income individuals? The effects on public finances and the low-income rate are discussed below for
three alternative scenarios.
Scenario 1: A Decrease in the OAS Clawback Threshold
According to the fiscal parameters used, only taxpayers with net income before
adjustments greater than $71,592 face the clawback of at least a portion of their oas
payments. The clawback rate is 15 percent, implying that the entire payment is reimbursed when the taxpayer’s net income before adjustments reaches $114,815.
These thresholds may seem high, so it seems natural to analyze the effects of extending the clawback over a larger share of the population without affecting low-income
individuals.
A $36,000 decrease in the clawback threshold for all oas recipients aged 65 and
over (not only those aged 65 and 66) would be necessary to generate approximately
the same overall gains in public finances (federal and provincial) in 2030 as the announced reform. As a result of this change, the threshold at which additional income
leads to oas clawbacks would be reduced to $35,592, and the threshold at which
oas payments are fully reimbursed would be lowered to $80,194. The effects of this
scenario are summarized in column 2 of table 3, and can be compared with the effects of the announced reform in column 1.
It may seem surprising that the decline in the threshold required to have the
same impact on finances is so large. Table 3 shows that the amount that the federal
(2015) 63:2
Table 2 Estimated Effects of Reform on Disposable Income
and the Low-Income Rate, 2030
2030, no reform
2030, with reform
Change
Disposable income (millions of constant 2014 dollars)
Men . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13,967
11,040
Women . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12,545 8,987
(2,927)
(3,558)
Proportion living in a low-income household (%)
Men . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.0
Women . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.0
16.7
19.3
10.7
12.1
Source: Authors’ calculations based on the SIMUL model; Office of the Superintendent of
Financial Institutions, Office of the Chief Actuary, Actuarial Report (11th) Supplementing the
Actuarial Report on the Old Age Security Program as at 31 December 2009 (Ottawa: OSFI, 2012); and
Statistics Canada’s social policy simulation database (SPSD), version 21.
government can recover by this approach would be substantial (around $9.5 billion). However, the amounts recovered would be deducted from taxable income,
and since the marginal tax rates for individuals affected by this alternative reform
would be much higher than the rates for those affected by the announced reform,
taxes on income would fall significantly. Federal taxes would decline by $1.95 billion
and provincial taxes by $934 million. We thus find a total impact on public finances
comparable to that expected by the announced reform. The difference is that the
low-income rate would remain unchanged, whereas under the announced reform it
would increase by 11.42 percentage points.
Scenario 2: A Uniform Decrease in OAS Benefits
An alternative scenario would be to implement a uniform decrease in oas payments
for all eligible individuals. According to our model, we would need to reduce benefits by $926 annually to achieve total gains in public finances comparable to those
under the announced reform. This amount may seem rather high; but, as with the
preceding reform, the direct gain would be countered by a major decrease in taxes
on income. In effect, the taxable revenues that would be affected by this reform
would be subject to marginal tax rates that would on average be higher than those
of the announced reform, given that they would apply to individuals in higher income segments.
Thus, the reform in scenario 2 would have the direct effect of decreasing total
oas payments by $9.6 billion in 2030. The effect on clawbacks for high income and
on social assistance payments would be comparatively small ($254 million and
$7 million respectively), while income taxes would decline substantially ($1.76 billion for federal and $920 million for provincial). While the total gains would be
comparable to those under the announced reform, the increase in the low-income
rate would be low (0.65 percentage points for those aged 65 and 66 years and 0.39
percentage points for those aged 67 to 69).
11.42
0
Effect on low-income rate, age 65-66 years (%)a . . . . . . . . . Effect on low-income rate, age 67-69 years (%)a . . . . . . . . . (934)
0
0
6,623
0.65
0.39
6,644
(926)
(7)
(920)
7,570
9,584
(254)
0
(1,759)
9.51
3.99
6,659
(632)
(69)
(563)
7,291
6,764
(361)
2,014
(1,127)
Scenario 3:
progressive increase
in amount of benefit
GIS = guaranteed income supplement.
OAS = old age security.
Note: Columns may not add because of rounding.
a Increase in percentage points with respect to the no-reform scenario.
Source: Authors’ calculations based on the SIMUL model; Office of the Superintendent of Financial Institutions, Office of the Chief Actuary, Actuarial
Report (11th) Supplementing the Actuarial Report on the Old Age Security Program as at 31 December 2009 (Ottawa: OSFI, 2012); and Statistics Canada’s social
policy simulation database (SPSD), version 21.
(638)
6,485
Net federal and provincial revenues . . . . . . . . . . . . . . . . . 0
(934)
(169)
(469)
Net provincial revenues . . . . . . . . . . . . . . . . . . . . . . . . . . 7,556
0
9,507
0
(1,951)
7,123
6,354
(388)
2,109
(950)
Scenario 2:
uniform $926
decrease in OAS
millions of constant 2014 dollars
Scenario 1:
$36,000 decrease in
clawback threshold
Net federal revenues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Provincial finances
Social assistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Provincial income tax . . . . . . . . . . . . . . . . . . . . . . . . . . . . Federal finances
Gross OAS payments . . . . . . . . . . . . . . . . . . . . . . . . . . . . OAS clawback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . GIS payments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Federal income tax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Announced
reform
TABLE 3 Comparison of Announced Reform and Alternative Reform Scenarios—Effects on Public Finances
and on the Low-Income Rate in 2030
(2015) 63:2
Scenario 3: A Progressive Increase in Benefits
Scenarios 1 and 2 have the advantage of not significantly increasing the low-income
rate among seniors while having a comparable fiscal impact. One possible drawback
is that, as opposed to the announced reform, those scenarios would affect some individuals who have already retired and hence may be unable to adjust their work
behaviour.13 This concern would be avoided under the third alternative scenario,
which consists of a gradual increase in oas and gis benefits each year, starting at the
age of 65. Our calculations suggest that the effect on public finances would be comparable to those for the announced reform when the following formula is used for
the sum of oas and gis payments:
 R ( Agei − 64 ) if 65 ≥ Agei ≤ 70,
Ri* =  i
6

if Agei > 70,
Ri
where ri is the sum of oas and gis payments that individual i would receive without
the reform and Agei is the age of the individual. The individual would thus receive the
amount ri* , which corresponds to one-sixth of the normal payment at age 65. In
each of the following years, the individual’s payment would increase by one-sixth of
the normal payment, to reach the full payment amount by the age of 70 years.
The last column of table 3 presents the effects of this reform. Since the impact
of the reform would be distributed over a longer period of time, the effect on social
assistance would be lower than for the announced reform. However, the low-income
rate would increase by 9.51 percentage points for those aged 65 and 66 years and by
3.99 percentage points for those aged 67 to 69.
Ca n I n d i v i d u a l s A f f e c t e d b y t h e R e f o r m
Av o i d Fa l l i n g B e l o w t h e L o w - I n co m e
T h r e s h o l d?
The results presented above assume that economic behaviour is not affected by the
reform scenarios. It is not difficult to imagine that individuals would react to the announced modifications to the oas and gis payment regime: participation in the
labour market could be extended, and savings could become higher as a result of
increasing the eligibility age. These adjustments in behaviour are at least one of the
major motivations for the announced reform put forward by the Canadian government, in particular in a context of an aging population and growing pressures on
public finances.
13It should also be noted that from 1952 to 1971 OAS was funded by three taxes: a tax on
personal income, a sales tax, and a corporate income tax. Because a high proportion of current
seniors paid a nominally dedicated tax for OAS, there is some degree of benefit-tax linkage that
could play a role in how people would view any move to limit OAS benefits for older retirees.
We thank the editors of this journal for this useful remark.
The behavioural effects of different scenarios are difficult to predict quantitatively. However, economic theory can be used to support some conjectures regarding
how the reform may alter behaviour qualitatively. Let us first focus on the impact of
the reform on individual savings. One important role of savings is to smooth the
consumption profile over an individual’s life cycle, given the variability of income
patterns and particularly the reduction in labour earnings after retirement.14 Given
that the announced reform may reduce expected income at 65 and 66 years of age
for individuals eligible for oas and gis payments, those individuals may be led to
increase their savings before and after this age interval in order to prevent too large
a negative shock in their consumption. Milligan15 shows that even if poverty measured by income is spiking during years prior 65, the pattern for poverty measured
by consumption is different. Also, income effects may encourage an increase in the
labour supply (in terms of both participation in the labour market and hours of
work) as long as leisure is a normal good. In particular, the reform may stimulate
some oas and gis recipients who would otherwise have planned to retire at the age
of 65 or 66 to delay their retirement.
On the other hand, as regards substitution effects, the reform could have the
perverse effect of reducing labour supply and savings of those who become recipients of social assistance, because the implicit tax rate of social assistance payments
is much higher than that of the gis. This last effect is nevertheless ambiguous, because the implicit tax rate of social assistance payments influences individuals less
than the rate faced by those who receive the gis, which applies to a larger range of
income. All in all, as long as these substitution effects are not too high, one can expect the reform to increase overall labour participation rate and savings. Gustman
and Steinmeier16 and Klaauw and Wolpin17 estimate that increasing the social security
retirement age in the United States would lead some people to delay their retirement and to increase hours of work before becoming eligible for social security.
Compared to the announced reform, scenario 1 (decreasing the clawback threshold by $36,000) could, via substitution effects, discourage labour supply and savings
among individuals earning between $35,592 and $80,194; it would, however, have
the opposite effect on those earning between $80,194 and $114,815. Owing to the
greater redistributive effect of this scenario, its income effect of increasing labour
and savings would be weaker among low-income individuals and stronger among
high-income individuals.
14See Albert Ando and Franco Modigliani, “The ‘Life Cycle’ Hypothesis of Saving: Aggregate
Implications and Tests” (1963) 53:1, part 1 American Economic Review 55-84.
15 Kevin Milligan, “The Evolution of Elderly Poverty in Canada” (2008) 34:4 Canadian Public
Policy 79-94.
16Alan Gustman and Thomas Steinmeier, “The Social Security Early Entitlement Age in a
Structural Model of Retirement and Wealth” (2005) 89:2-3 Journal of Public Economics 441-63.
17 Wilbert van der Klaauw and Kenneth Wolpin, “Social Security and the Retirement and Savings
Behavior of Low-Income Households” (2008) 145:1-2 Journal of Econometrics 21-42.
(2015) 63:2
Again, as compared with the announced reform, scenario 2 (a uniform reduction
in oas payments) would have positive income effects on labour and savings, but
those effects would be weaker among low-income individuals and stronger among
high-income individuals. As in the case of the reduced clawback threshold, a uniform reduction of $926 in payments would displace toward the lower end of the
income distribution the disincentive effects on labour supply and savings associated
with the 15 percent oas clawback rate.
As for scenario 3 (a gradual oas payment increase), this alternative would have
similar effects to those of the announced reform, but they would be distributed
gradually between the ages of 65 and 70. Given its more redistributive character, the
income effects of this scenario would nevertheless be weaker among low-income
individuals and stronger among higher-income individuals.
These behavioural effects would obviously affect not only the low-income rate,
but also all of the public finance results presented in tables 1 and 3. Higher incentives
to work would increase the reform’s and the scenarios’ effects on public finances
through a smaller decrease in income taxes and a smaller increase in social assistance.
This discussion suggests that individuals could in principle adjust their behaviour
following the announced reform, and thereby possibly avoid falling into a significantly lower standard of living. It is not certain, however, that this adjustment would
be sufficiently strong to prevent shifting so many people below the low-income
threshold. Certain individuals have difficulty maintaining a rigorous savings discipline in order to counter anticipated future losses in living standards. Research in
behavioural economics suggests, for example, that the most elderly individuals make
inconsistent decisions more often18 and that the poorest individuals are more negligent in their financial management than those in higher income groups.19
A simple analysis of the current distribution of income also shows that many individuals are unable to maintain sufficient living standards before the age of 65.
Figure 4 shows the low-income rates of individuals between 50 and 70 years of age
in 2012. We observe that the rate increases considerably with age until it stabilizes
around 26 percent just before the age of 65. With full annual payments of oas and
gis being applicable as of the age of 66, the low-income rate is low after this age.
(For the year in which an individual turns 65, the annual payment may be smaller,
depending on the month in which that individual’s birthday falls.) The years leading
up to the age of 65 thus seem to be the most difficult in terms of low income.
The foregoing analysis suggests that, in many cases, individuals do not manage to
work enough before the current age of eligibility or to save enough in anticipation
of future income loss. The currently elevated low-income rate among those aged 64
18See Sumit Agarwal, John C. Driscoll, Xavier Gabaix, and David Laibson, “The Age of Reason:
Financial Decisions over the Life-Cycle with Implications for Regulation” [2009] Brookings
Papers on Economic Activity 51-117.
19See Barry Scholnick, Nadia Massoud, and Antony Saunders, “The Impact of Wealth on
Financial Mistakes: Evidence from Credit Card Non-Payment” (2013) 9:1 Journal of Financial
Stability 26-37.
FIGURE 4 Low-Income Rate by Age in 2012
0.20
Low-income rate
0.15
0.10
0.05
0
50
55
60
65
70
Age
Source: Authors’ calculations using Statistics Canada’s social policy simulation database
(SPSD), version 21, and the 2009 Montreal market consumption basket; see
www.statcan.gc.ca/pub/75f0002m/2012002/tbl/tbl04-eng.htm.
years suggests that individuals at the age of 65, deprived of their oas and gis, will
also have an elevated low-income rate. The announced reform is thus likely to extend the low-income rates currently observed prior to the age of 65.
Co n c l u s i o n
The dynamic microsimulation model of the Industrial Alliance Research Chair on
the Economics of Demographic Change allows us to predict the effects of the Canadian government’s announced reform of the oas and gis regimes on federal and
provincial public finances. The model accounts simultaneously for predicted demographic change, income distribution, taxes on individual incomes, and social
transfers. Thus, it allows us to estimate not only the direct effects of the reform
(decreases in oas and gis payments), but also the effects on personal incomes and
taxes as well as on provincial and federal social transfers.
Considering all of these effects, the announced reform would decrease net federal outlays in 2030 by $7.1 billion but would generate fiscal losses of $638 million
for the provinces (in constant 2014 dollars). The overall gain (federal and provincial)
would thus be about $6.5 billion. This reform would have the greatest impact on the
least wealthy Canadians. About 60 percent of the losses in living standards would be
experienced by the poorest 50 percent of the population aged 65 and 66, or by the
poorest 40 percent of the population aged 16 and over who have completed their
(2015) 63:2
studies. The low-income rate of individuals aged 65 and 66 years in 2030 would rise
from 6 percent to 17 percent in 2030 as a result of the reform.
We have also studied the effects of three alternative reform scenarios. These
scenarios have comparable effects on overall public finances (federal and provincial)
but different impacts on low-income individuals.
Scenario 1 proposes the reduction of the oas clawback threshold by $36,000.
This would bring the threshold at which oas begins to be reimbursed to $35,592
and would reduce the income at which the payment would be completely clawed
back to $80,194. This scenario would not result in an increase in the low-income
rate. Scenario 2 proposes a uniform $926 decrease in the annual payment for all
eligible individuals; it would have a negligible impact on the low-income rate. Scenario 3 would gradually increase payments from the age of 65 to the age of 70. This
scenario leads to a higher low-income rate, by 9.51 percentage points for those aged
65 and 66 and by 3.99 percentage points for those aged 67 to 69. Thus, relative to
the announced reform, scenarios 1 and 2 would have the advantage of not significantly increasing the low-income rate among seniors while having a comparable
fiscal impact.
A pp e n d i x
The estimates of the model are generated using Quebec data. They are then generalized for Canada as a whole, using the rules set out below.
Gross Old Age Security (OAS) Payments
The model predicts that the government’s announced reform will diminish total
gross oas payments by 10.9 percent in Quebec in 2030. We apply this proportion
to the gross oas payments of $58.29 billion (in constant 2014 dollars) predicted by
the Office of the Chief Actuary (oca) for Canada in 2030. This results in a decrease
of $6.354 billion. For each of the alternative scenarios, a proportion is calculated by
the same method and is used to extrapolate the scenarios’ effects on oas to Canada.
Guaranteed Income Supplement (GIS) Payments
Our model predicts that the reform will diminish total gis by 13.23 percent in
Quebec in 2030. We apply this proportion to the oca’s prediction of total gis payments of $15.94 billion (in constant 2014 dollars) in 2030. This results in a decrease
of $2.109 billion. For each of the alternative scenarios, a proportion is calculated by
the same method and is used to extrapolate the scenarios’ effects on gis to Canada.
OAS Clawback
We use Statistics Canada’s social policy simulation database (spsd) to estimate the
share of gross oas payments that is paid back through clawbacks in Quebec and in
Canada. The proportion is 2.20 percent in Quebec and 3.42 percent in Canada, so
the ratio is 1.556 times higher in Canada than in Quebec. We then apply this factor
to our predicted Quebec ratio of oas clawback on gross oas.
Federal Income Tax
We use the spsd to estimate the share of total federal income tax in Canada that
comes from Quebec. That share is 16.58 percent. We then use this percentage to
adjust our predictions of federal income tax.
Provincial Income Tax
We use the spsd to estimate the share of total provincial income tax in Canada that
is paid in Quebec. That share is 32.22 percent. We then adjust our predictions of
provincial income tax by dividing our provincial income tax predictions for Quebec
by 0.322.
Social Assistance
We use the spsd to estimate the share of total social assistance in Canada that is
paid in Quebec. That share is 26.1 percent. We then adjust our predictions of social
assistance by dividing our predictions for Quebec by 0.261.
Is There a Sixth Comparability Factor in
Canadian Transfer Pricing?
Robert Robillard*
Précis
Cet article examine les décisions marquantes récentes sur les prix de transfert au
Canada. Il suggère que les tribunaux canadiens aient pu donner naissance à un sixième
facteur de comparabilité pour l’application du principe de pleine concurrence qui
contrevient à l’intention de l’article 247 de la Loi de l’impôt sur le revenu. Ce nouveau
facteur de comparabilité est désigné comme la pertinence des facteurs de dépendance
entourant la relation entre les parties liées.
D’abord, l’article récapitule brièvement le processus connu sous le nom d’analyse de
comparabilité, tel qu’énoncé dans les lignes directrices publiées par l’Organisation de
coopération et de développement économiques et dans l’article 247 de la Loi. Puis, il
examine la signification donnée par les tribunaux canadiens à l’analyse de comparabilité.
L’auteur souligne comment les tribunaux, par leurs décisions, ont créé un sixième facteur
de comparabilité aux fins des prix de transfert. L’article se termine par une brève analyse
des conséquences inattendues que ce nouveau facteur de comparabilité pourrait avoir
relativement à ce qui constitue des « efforts raisonnables » en vue d’utiliser des
attributions ou des prix de pleine concurrence, et de l’application possible de pénalités
relatives aux prix de transfert.
Abstract
This article reviews the recent landmark transfer-pricing case law in Canada. It suggests
that the Canadian courts may have given birth to a sixth comparability factor for the
application of the arm’s-length principle that contravenes the intent of section 247 of the
Income Tax Act. This new comparability factor is referred to as the relevance of non-arm’slength factors surrounding the relationship between related parties.
First, the article briefly summarizes the process known as the comparability analysis,
as set out in the guidelines issued by the Organisation for Economic Co-operation and
Development and in section 247 of the Act. Second, the meaning given by the Canadian
courts to the comparability analysis is examined. The author highlights how the courts,
through their decisions, have created a sixth comparability factor for transfer-pricing
purposes. The article concludes with a brief discussion of the unintended consequences
* Of DRTP Consulting/Université du Québec à Montréal (e-mail: [email protected]).
The opinions expressed in this document are my own.
 375
(2015) 63:2
that this new comparability factor may have for what constitutes “reasonable efforts” to
use arm’s-length allocations or prices, and thus for the possible application of transferpricing penalties.
Keywords: Transfer pricing n comparability n comparable uncontrolled price n
transactional net margin n tax policy n fiscal policy
Contents
Introduction
The OECD Comparability Analysis
Comparability Analysis and the Canadian Courts
What the Tax Court of Canada First Said on the Relevance of Contractual Terms
Conceptual Tensions Arise on the Relevance of Non-Arm’s-Length Factors
The Relevance of Non-Arm’s-Length Factors Is Put to the Test
The Supreme Court Officially Gives Birth to a Sixth Comparability Factor
A Sixth Comparability Factor: Non-Arm’s-Length Factors Surrounding the
Relationship
Conclusion
376
380
382
382
385
387
389
392
394
Introduc tion
In Canada, section 247 of the Income Tax Act1 pertains to transfer pricing. It has
replaced former subsection 69(2). In 1998, the government of Canada modified its
transfer-pricing legislation in order “to harmonize the standard contained in section
69 of the Act with the arm’s length principle as defined in the revised [1995] oecd
guidelines.”2 At the time, the amendments to the Act implemented a “transfer pricing regime based explicitly on the arm’s length principle.”3 For greater certainty,
subsection 247(1) included definitions of “arm’s length allocation” (of profit or loss)
and “arm’s length transfer price” for the purposes of the transfer-pricing rules. In
each case, the allocation or price arranged between the participants in a transaction
must reflect what would have occurred “if the participants had been dealing at arm’s
length with each other.”4
Accordingly, under Canadian tax legislation, the determination of a transfer price
for a transaction between a Canadian taxpayer and a non-arm’s-length non-resident
1 RSC 1985, c. 1 (5th Supp.), as amended (herein referred to as “the Act”). Unless otherwise
stated, statutory references in this article are to the Act.
2 Canada, Department of Finance, 1997 Budget, Budget Plan, February 18, 1997, at 203. See
also Organisation for Economic Co-operation and Development, Transfer Pricing Guidelines for
Multinational Enterprises and Tax Administrations (Paris: OECD, 1995) (herein referred to as
“the 1995 OECD guidelines”).
3 See the summary of the Income Tax Amendments Act, 1997 in (1998) 21:2 Canada Gazette
Part III, at 2a.
4 Canada, Department of Finance, Explanatory Notes Relating to Income Tax (Ottawa: Department
of Finance, December 1997), at 505.
is there a sixth comparability factor in canadian transfer pricing?  n  377
must usually abide by the arm’s-length principle as defined in the oecd Transfer Pricing
Guidelines for Multinational Enterprises and Tax Administrations.5 Subsection 247(2)
prescribes a transfer-pricing adjustment when “the terms or conditions made or
imposed, in respect of the transaction or series, between any of the participants in
the transaction or series differ from those that would have been made between
persons dealing at arm’s length.”6
At the onset of a transfer-pricing compliance audit, subsection 247(4) is invoked by
the Canada Revenue Agency (cra) in any requests for contemporaneous documentation.7 Specifically, reference to the information requirements listed in subparagraphs
247(4)(a)(i) through (vi) must now be included in the contemporaneous documentation letter “at the stage of initial contact with the taxpayer in all audits” that involve
controlled transactions.8 Subparagraphs 247(4)(a)(i) through (vi) are extremely important to Canadian taxpayers. They play a crucial role in the eventual application
of a transfer-pricing penalty under subsection 247(3) where reasonable efforts were
not made to determine an arm’s-length price.
Subparagraphs 247(4)(a)(i) through (vi) bear a resemblance to the comparability
factors listed in the oecd guidelines. Part d.1 of chapter i of the oecd guidelines
provides taxpayers and tax administrations alike with five categories of comparability
factors to ensure the proper application of the arm’s-length principle,9 in a procedure
that is widely known as “the oecd comparability analysis.” The five categories are
1. the characteristics of property or services,
2. the functional analysis,
3. the contractual terms,
4. the economic circumstances, and
5. the business strategies.
5 Organisation for Economic Co-operation and Development, OECD Transfer Pricing Guidelines
for Multinational Enterprises and Tax Administrations (Paris: OECD, 2010) (herein referred to as
“the OECD guidelines”). This is an updated version of the 1995 guidelines cited in note 2,
supra. See also Information Circular 87-2R, “International Transfer Pricing,” September 27,
1999, at part 3.
6 The Act also includes provisions such as subsections 17(8) (dealing with non-interest-bearing
loans to controlled foreign affiliates) and 18(4) (containing thin capitalization rules) governing
specific transactions between related parties.
7See Transfer Pricing Memorandum TPM-05R, “Requests for Contemporaneous Documentation,”
March 28, 2014, appendix A, “Sample Contemporaneous Documentation Letter.”
8 Ibid., at paragraph 9.
9 It should be noted that the OECD has released a public discussion draft that proposes
modifications to part D of chapter I of the guidelines: See Organisation for Economic
Co-operation and Development, BEPS Actions 8, 9 and 10: Discussion Draft on Revisions to
Chapter I of the Transfer Pricing Guidelines (Including Risk, Recharacterisation, and Special Measures)
(Paris: OECD, December 2014). The proposed changes would have no impact on the analysis
provided in this article.
(2015) 63:2
Paragraph 1.15 of the 1995 oecd guidelines (paragraph 1.33 of the 2010 version) suggests that the first step in the comparability analysis is to undertake a
“comparison of the conditions in a controlled transaction with the conditions in
transactions between independent enterprises.”10 This requires the specific comparison of every economically relevant characteristic in the controlled transaction
and in an arm’s-length transaction.11 But in order for that comparison to make any
conceptual sense with respect to the determination of an arm’s-length transfer price,
it is solely the arm’s-length factors that need to be taken into account. Article 9 (associated enterprises) of the oecd model tax convention provides for adjustments
when “conditions . . . made or imposed between the two enterprises in their commercial or financial relations . . . differ from those which would be made between
independent enterprises.”12 That is, any such adjustment is meant to correspond to
the conditions that would have existed between parties dealing at arm’s length.
Paragraph 1.38 of the oecd guidelines states that
the examination of the five comparability factors is by nature two-fold, i.e. it includes an
examination of the factors affecting the taxpayer’s controlled transactions and an examination of the factors affecting uncontrolled transactions.13
Such an approach enables the evaluation, as also pointed out in paragraph 1.38, of
the “relative importance of any missing piece of information on possible comparables, which can vary on a case-by-case basis.”14
That methodological precept is at the heart of the arm’s-length principle, which
treats the members of a multinational enterprise (mne) group as “operating as separate entities rather than as inseparable parts of a single unified business.”15 In reality,
non-arm’s-length factors would arise only if the parties were not operating as separate entities. In other words, non-arm’s-length factors are rendered non-existent
for the purpose of the application of the arm’s-length principle—that is, for the
determination of an arm’s-length transfer price. However, non-arm’s-length factors
are not simply ignored or overlooked; they are eliminated by virtue of the comparability analysis carried out with respect to the relevant arm’s-length factors. This is
in fact the basis of the arm’s-length principle. More precisely, paragraph 1.6 of the
oecd guidelines states:
10 Supra note 2 (1995 guidelines) and note 5 (2010 guidelines).
11 See also the OECD discussion draft, supra note 9, at paragraph 1.
12 Organisation for Economic Co-operation and Development, Model Tax Convention on Income
and on Capital: Condensed Version 2014 (Paris: OECD, July 2014) (herein referred to as “the
OECD model tax convention”).
13 Supra note 5 (2010 guidelines).
14 Ibid.
15 Ibid., at paragraph 1.6.
Because the separate entity approach treats the members of an mne group as if they
were independent entities, attention is focused on the nature of the transactions between those members and on whether the conditions thereof differ from the conditions
that would be obtained in comparable uncontrolled transactions.16
Such a methodological focus on the nature of the transactions should clearly prevent the inclusion of any non-arm’s-length factors in the comparability analysis for
transfer-pricing purposes.17 However, as discussed in detail below, the Canadian
courts have taken a different approach to the determination of an arm’s-length transfer price for the purpose of former subsection 69(2), and subsequently section 247.18
On the one hand, the Canadian courts have recognized the relevance of the
oecd guidelines in carrying out the comparability analysis; indeed, those guidelines
have become part of the Canadian transfer-pricing landscape.19 On the other hand,
as we shall see, the Canadian courts have given birth to a sixth comparability factor,
namely, the relevance of non-arm’s-length factors surrounding the relationship
between related parties.
I argue in this article that in no way were non-arm’s-length factors contemplated
as either being part of the oecd comparability analysis or being relevant to the application of subsection 69(2) or section 247.20 This methodological quarrel is not
simply a theoretical one. It could give rise to serious tax litigation in the future as
the gap between the Canadian courts’ interpretation of the arm’s-length principle
and the Canadian legislation on transfer pricing slowly but surely widens.
16 Ibid.
17 This conclusion is also supported by paragraph 1.11 of the OECD guidelines, supra note 5,
which addresses the issues arising from any transaction that “may not be found between
independent parties.” As paragraph 1.11 explains, such a circumstance would not in “itself
mean that it [the transaction] is not arm’s length.” However, the arm’s-length principle would
be more difficult to apply “because there is little or no direct evidence of what conditions
would have been established by independent enterprises.” (Ibid.) In other words, despite these
difficulties, the emphasis must remain solely on the arm’s-length factors for the purposes of the
comparability analysis, according to the OECD.
18 In Canada v. General Electric Capital Canada Inc., 2010 FCA 344, at paragraph 12, Noël JA,
writing for a unanimous panel, indicated that “there is no meaningful difference” between
former subsection 69(2) and section 247. That interpretation permeates this article, although
nuances may indeed be possible. Accordingly, in the discussion that follows, the two provisions
are analyzed as a single provision—that is, as the Canadian representation of the application of
the arm’s-length principle in the transfer-pricing context.
19 In the case law, the pertinence of the OECD guidelines for Canadian transfer pricing was first
stated in Smithkline Beecham Animal Health Inc. v. Canada, 2002 FCA 229, at paragraphs 6-9.
20 IC 87-2R, supra note 5, also does not consider the inclusion of non-arm’s-length factors in the
determination of an arm’s-length price.
(2015) 63:2
T h e OECD Co m pa r a b i l i t y A n a ly s i s
As discussed above, the oecd comparability analysis comprises five categories of
factors:
1. the characteristics of property or services,
2. the functional analysis,
3. the contractual terms,
4. the economic circumstances, and
5. the business strategies.
As indicated in paragraph 1.33 of the oecd guidelines, these factors permit “comparison of the conditions in a controlled transaction with the conditions in transactions
between independent enterprises.”21 Each of these categories is central to the application of the arm’s-length principle. They ultimately lead to the selection and
application of a transfer-pricing method, which in turn enables the determination
of an arm’s-length price for a specific transaction between parties not dealing at
arm’s length.22 As clarified in Information Circular 87-2r, “selecting the most appropriate pricing method depends largely on the assessment of the comparability of
transactions.”23
At the abstract level, comparison implies not only the evaluation of at least two
objects considered as wholes but also the examination of the major components of
those objects. This, in a nutshell, is what the oecd comparability analysis is all about;
it involves an examination of any relevant arm’s-length comparability factor.24 In an
analysis based on the comparison of every relevant arm’s-length factor in the controlled transaction and an arm’s-length transaction, any non-arm’s-length factors
that are part of the controlled transaction are rendered ineffective.
In other words, the precise purpose of the arm’s-length principle is ultimately to
adjust the terms and conditions of any controlled transaction to the terms and conditions of an arm’s-length transaction (obviously, not the converse). Canada’s
transfer-pricing legislation—in particular, paragraphs 247(2)(a) and (c)—is explicit
to that effect.25 Therefore, non-arm’s-length factors are clearly irrelevant to the
comparability analysis in transfer pricing—that is, for the application of the arm’slength principle.
21 Supra note 5 (OECD guidelines).
22 Paragraph 2.2 of the OECD guidelines, ibid., explains the relationship between the
comparability analysis and the selection and application of a transfer-pricing method.
23 IC 87-2R, supra note 5, at paragraph 33.
24 Paragraphs 1.33 to 1.35 of the OECD guidelines, supra note 5, are explicit on that matter.
25 The analysis in this article does not include paragraphs 247(2)(b) and (d), which pertain to the
recharacterization of any transaction “entered into primarily for bona fide purposes other than
to obtain a tax benefit.”
If any non-arm’s-length factor is present in the controlled transaction, it must be
discarded, since it would logically render an alleged arm’s-length transaction fundamentally non-arm’s-length. In truth, it is difficult to see how a factor that existed
solely in a non-arm’s-length environment could be included in a comparison with
an arm’s-length arrangement, since that factor could not exist in the arm’s-length
environment. In such a case, comparison could never be claimed, even remotely, for
the application of the arm’s-length principle in determining an appropriate transfer
price.
Contractual terms are among the five categories of factors included in the oecd
comparability analysis. In the commercial world, the contractual terms of a transaction are of paramount importance, in particular in ensuring that the interests of
every party involved in the transaction are honoured in an orderly fashion. For
transfer-pricing purposes, contractual terms as they are defined enable taxpayers to
determine the allocation of various “risks” (which is one of the components of the
functional analysis). These risks may include market risk; risk associated with property, plant, and equipment; commercial investment risk; financial risk; and credit
risk, among others.26
For tax administrations, the examination of the contractual terms in a given
controlled transaction or relationship helps in ascertaining whether the “purported
allocation of risk is consistent with the economic substance of the transaction.”27
Contractual terms are examined by tax administrations because in any arm’s-length
transaction they “generally define explicitly or implicitly how the responsibilities,
risks and benefits are to be divided between the parties.”28 In the context of Canada’s
transfer-pricing rules, ic 87-2r recognizes the relevance of contractual terms as
they “may influence the degree of comparability of transactions.”29
In Canadian case law, the courts have postulated that both arm’s-length and nonarm’s-length economically relevant factors in a specific controlled transaction must
be taken into account for the purpose of the comparability analysis—that is, for the
application of the arm’s-length principle under section 247 (and former subsection
69(2)). In the following section, I will show how the Canadian courts ended up
drawing this remarkable conclusion, which infers language that is in fact absent
from the oecd guidelines and departs from the intent of the Act.
26 OECD guidelines, supra note 5, at paragraph 1.46. The OECD discussion draft, supra note 9,
deepens the understanding of risks in part D.2.
27 OECD guidelines, supra note 5, at paragraph 1.48.
28 Ibid., at paragraph 1.52.
29 IC 87-2R, supra note 5, at paragraph 32.
(2015) 63:2
Co m pa r a b i l i t y A n a ly s i s a n d
t h e Ca n a d i a n Co u r t s
What the Tax Court of Canada First Said on
the Relevance of Contractual Terms
The relevance of the contractual terms for the determination of an arm’s-length
transfer price was first analyzed by the Tax Court of Canada in GlaxoSmithKline Inc.
v. The Queen30 and General Electric Capital Canada Inc. v. The Queen.31
In GlaxoSmithKline, the Canadian taxpayer appealed income tax assessments by
the cra in a dispute over the price that the taxpayer had paid to its related suppliers
for the purchase of ranitidine (an active pharmaceutical ingredient). The taxpayer
contended that its dealings with its suppliers were at arm’s length on the basis of
both the facts and the circumstances surrounding these purchases, and that the
amounts paid were “reasonable in the circumstances.”32 Applying the comparable
uncontrolled price (cup) method to the transactions at issue, the court essentially
dismissed the taxpayer’s argument and concurred with the position of the cra.
The cra’s argument, as it pertains to the comparability analysis, was that the
supply agreement and the licence agreement concluded by the Canadian taxpayer
with its related suppliers were to be examined separately for the determination of
the arm’s-length price for the purchases of ranitidine. According to the cra, this
was the correct method for determining an appropriate price since it was consistent
with the transaction-by-transaction approach suggested by the oecd guidelines.33
The cra also relied on Singleton v. Canada34 to support its position. In Singleton, the
Supreme Court of Canada explained that unless a provision of the Act is ambiguous,
a court must apply the statute as it is written “rather than search for the economic
realities of the transaction.”35 Accordingly, the cra argued on the basis of Singleton
that
one must look at the transaction in issue and not the surrounding circumstances, other
transactions or other realities because in order to give effect to the legal relations, one
has to view the agreements independently.36
30 2008 TCC 324.
31 2009 TCC 563.
32 GlaxoSmithKline, supra note 30, at paragraph 8.
33 See paragraphs 1.42-1.44 of the 1995 OECD guidelines, supra note 2 (paragraphs 3.9-3.12 of
the 2010 version, supra note 5). However, it must be pointed out that in the 2010 guidelines,
paragraph 3.9 also highlights cases where “separate transactions are so closely linked or
continuous that they cannot be evaluated adequately on a separate basis.” Such transactions
include the “licensing of manufacturing know-how and the supply of vital components,” a
situation highly similar to the purchases of ranitidine in the GlaxoSmithKline case.
34 2001 SCC 61.
The cra argued that the licence agreement was irrelevant since in other instances
the Canadian taxpayer had concluded specific arm’s-length supply agreements for the
purchase of other pharmaceutical ingredients. Those agreements were independent
from any licence agreement. The cra consequently argued that the specific licence
agreement concluded by the taxpayer was irrelevant for the determination of the arm’slength price of the purchases of ranitidine. Only the supply agreement mattered.
The Canadian taxpayer contended that the two commercial agreements were in
fact specific “circumstances” that surrounded the controlled transaction and as such
both had to be considered in accordance with the oecd guidelines. The supply
agreement defined the terms and conditions of the purchases of ranitidine and also
the payment of royalties (through the licence agreement), among other things.
From that point of view, both agreements had to be included in the comparability
analysis, and not just the supply agreement.
Ultimately, the Tax Court agreed with the cra’s position. Although the licence
agreement concluded by the taxpayer pertained indirectly to the purchases of ranitidine, it had to be ignored in determining an appropriate arm’s-length transfer price.
The agreement did not constitute a relevant “circumstance” in that determination.
Rip acj concurred with the cra that “the two agreements covered distinct subject
matters.”37 Moreover, taking into account the distinct tax treatment of the purchases of ranitidine and the payments of royalties, the court could not reach any
other conclusion.38 In short, an arm’s-length factor (the licence agreement) was
deemed irrelevant to the comparability analysis both on the contractual terms and
for the determination of an arm’s-length price. The court’s direct application of the
cup method was sufficient to quash the taxpayer’s position.39
In General Electric Capital Canada, the Canadian taxpayer appealed the cra’s income tax assessments on the basis that the guarantee fee paid by the taxpayer to its
us parent company in respect of debts owing to third-party creditors represented
the actual arm’s-length price of that service.40 The appeals were allowed and the
cra assessments ultimately vacated.
With regard to the comparability analysis, the issue pertained to the relevance of
a non-arm’s-length factor for the determination of an arm’s-length price. At the Tax
Court of Canada, the cra suggested that the taxpayer had “implicit support” from
its parent company, an obvious condition arising from the non-arm’s-length nature
of the relationship. According to the cra, this in itself should have precluded the
37 Ibid.
38 Ibid., at paragraphs 77-78.
39 See ibid., at paragraphs 119-161, in particular paragraphs 141 and 161. I will return to
GlaxoSmithKline below, since the decision was appealed first to the Federal Court of Appeal
and then to the Supreme Court of Canada, before a confidential settlement was announced in
early January 2015 pending the return of the case to the Tax Court of Canada.
40 General Electric Capital Canada, supra note 31, at paragraph 1.
(2015) 63:2
relevance of any payment of fees by the taxpayer to its parent company to guarantee
its debts owed to third-party creditors.41
Of note, the cra relied on paragraph 7.13 of the oecd guidelines, which discusses
the “incidental benefits attributable solely to being part of a larger concern.”42 The
cra contended that the taxpayer would not have entered into the guarantee contract
if the parties had been dealing at arm’s length. The implicit support was a specific
circumstance (that is, a non-arm’s-length factor) that surrounded the controlled
transaction, according to the cra. However, it should be pointed out that the oecd
guidelines state that in such cases no transaction has occurred; in short, no service
has been received.
The taxpayer disagreed with the cra’s position, arguing that the proper application of the arm’s-length principle required that “all distortions that arise from the
parties’ relationship . . . be eliminated to arrive at an arm’s length result.”43 The
comparability analysis can then be carried out to determine an arm’s-length price,
which may in fact be nil if paragraph 7.13 were indeed to apply to the transaction.
After a careful examination of the meaning of “arm’s length,”44 Hogan j concluded that the concept of implicit support could not be ignored in determining an
arm’s-length price. In his view, “[t]hat concept has nothing to do with the exercise
of de facto or de jure control which defines a non-arm’s length relationship.”45 In
other words, the implicit support provided by the parent company to its subsidiary,
a recognizable non-arm’s-length factor, was considered relevant by the court for the
determination of the arm’s-length price in a commercial transaction between parties
not dealing at arm’s length.
On the one hand, in GlaxoSmithKline the Tax Court rejected an explicit commercial agreement between the related parties because it did not constitute a relevant
circumstance in determining an arm’s-length price. On the other hand, in General
Electric Capital Canada the court included in the relevant circumstances a factor that
clearly arose from the non-arm’s-length relationship between the taxpayer and its
parent company. Both decisions were appealed, as will be discussed below.
These two decisions of the Tax Court of Canada set the path for a deeper examination of what may constitute “relevant circumstances” in order to carry out the
analysis of the contractual terms between parties not dealing at arm’s length and
the comparability analysis applied to transfer pricing.
42 OECD guidelines, supra note 5. An OECD public discussion draft issued in 2014 does not
modify this position: see Organisation for Economic Co-operation and Development, BEPS
Action 10: Proposed Modifications to Chapter VII of the Transfer Pricing Guidelines Relating to Low
Value-Adding Intra-Group Services (Paris: OECD, November 2014), at paragraph 7.14.
43 General Electric Capital Canada, supra note 31, at paragraph 173.
Conceptual Tensions Arise on the Relevance
of Non-Arm’s-Length Factors
In GlaxoSmithKline, the Federal Court of Appeal46 was asked if the two agreements
(the licence agreement and the supply agreement) were indeed relevant circumstances
to be considered in the comparability analysis and, subsequently, in determining an
arm’s-length price for the purchases of ranitidine. Nadon ja, writing for a unanimous panel, enumerated five “circumstances” that showed that the licence agreement
was clearly useful and central to the determination of the arm’s-length price of the
ranitidine.47
The Federal Court of Appeal rejected the relevance of the Singleton case for an
arm’s-length transfer-pricing determination. The court basically indicated that
there was no similarity between subparagraph 20(1)(c)(i) and subsection 69(2) (the
predecessor of section 247). Accordingly, the decision of the Supreme Court in
Singleton not to recharacterize a transaction in light of its economic realities was
irrelevant for transfer-pricing purposes.48
Instead, the court relied upon Gabco Limited v. mnr.49 The court noted that the
“reasonable business person” standard was relevant for the determination of an arm’slength transfer price.50 In other words, an examination of the business circumstances
was required in order to determine an arm’s-length price. As the court stated,
the test set out in Gabco, supra, requires an inquiry into those circumstances which an
arm’s length purchaser, standing in the shoes of the appellant, would consider relevant
in deciding whether it should pay the price paid by the appellant to Adechsa [the supplier] for its ranitidine.51
The two agreements were therefore part of the comparability analysis. They
“[did] not arise from the non-arm’s length relationship between the appellant and
Adechsa or between the appellant and Glaxo Group.”52 By extension, the supply
agreement was also considered pertinent since it put forward the terms and conditions of the ranitidine purchases, including the benefits and rights obtained by the
taxpayer through the licence agreement. In other words, every arm’s-length circumstance considered to be relevant from the taxpayer’s perspective was truly appropriate
for the determination of the arm’s-length price.
46 GlaxoSmithKline Inc. v. Canada, 2010 FCA 201.
49 68 DTC 5210 (Ex. Ct.).
50 GlaxoSmithKline, supra note 46, at paragraphs 69-72.
(2015) 63:2
In General Electric Capital Canada, the Federal Court of Appeal53 reanalyzed the
relevance of the inclusion of any non-arm’s-length factor for the determination of
an arm’s-length price. The issue of the “implicit support” received by the taxpayer
from its parent company was back on stage. It was at the heart of the argument
submitted by the cra in its appeal. Writing for a unanimous panel, Noël ja explained that the inclusion of the implicit support as an economically relevant factor
in determining an arm’s-length price “gives rise to a pure question of statutory
construction which must be assessed on a standard of correctness.”54 After examining paragraphs 247(2)(a) and (c), the court stated:
The concept underlying subsection 69(2) and paragraphs 247(2)(a) and (c) is simple.
The task in any given case is to ascertain the price that would have been paid in the
same circumstances if the parties had been dealing at arm’s length. This involves taking
into account all the circumstances which bear on the price whether they arise from the
relationship or otherwise.
This interpretation flows from the normal use of the words as well as the statutory
objective which is to prevent the avoidance of tax resulting from price distortions which
can arise in the context of non arm’s length relationships by reason of the community
of interest shared by related parties. The elimination of these distortions by reference
to objective benchmarks is all that is required to achieve the statutory objective.
Otherwise all the factors which an arm’s length person in the same circumstances as
the respondent would consider relevant should be taken into account.55
Simply put, the Federal Court of Appeal in General Electric Capital Canada stated
that “any” factor deemed relevant, whether by the taxpayer or by the tax administration, may be included in the comparability analysis for the determination of an
arm’s-length price. Remarkably, the court pointed out that this comparability analysis could include factors that existed solely because the parties were not dealing at
arm’s length if those factors were considered pertinent to the determination of an
arm’s-length price.56 The court did not consider that the sole purpose of the comparability analysis was in fact to render ineffective the non-arm’s-length factors in
the determination of an arm’s-length price.
The court indicated that its conclusion on this point was based on the “test” set
out by the Federal Court of Appeal in GlaxoSmithKline.57 According to the court, a
person deemed to deal at arm’s length—or the court, for that matter—should have
to consider any relevant non-arm’s-length factor in addition to relevant arm’slength factors in order to properly apply the arm’s-length principle as prescribed by
53 General Electric Capital Canada, supra note 18.
57 Ibid., at paragraphs 58-59, in reference to GlaxoSmithKline, supra note 46.
subsection 247(2). In General Electric Capital Canada, the implicit support provided
by the parent company to its subsidiary hence became a relevant factor in the characterization of the contractual terms and, for that reason, in the determination of an
arm’s-length price.
However, it should be noted that in GlaxoSmithKline, the Federal Court of Appeal
never suggested that non-arm’s-length factors were relevant for the determination
of an arm’s-length price, but quite the opposite. The court clearly indicated that the
two agreements were relevant circumstances. They unequivocally “[did] not arise
from the non-arm’s length relationship between the appellant and Adechsa or between the appellant and Glaxo Group.”58 The court also added:
I again wish to emphasize that the above circumstances were circumstances that would
have been present even if the appellant had been dealing at arm’s length with Adechsa
and Glaxo Group. Consequently, an arm’s length appellant would necessarily have had
to consider those circumstances in deciding whether it was willing to pay the price
asked for by Adechsa for the sale of the Zantac ranitidine.59
In short, obvious conceptual tensions were persisting in light of these two decisions
of the Federal Court of Appeal with respect to what were “relevant circumstances”
in carrying out the analysis of the contractual terms between parties not dealing at
arm’s length, in performing the comparability analysis, and ultimately in determining an arm’s-length price.
The Relevance of Non-Arm’s-Length Factors Is Put to the Test
In Alberta Printed Circuits Ltd. v. The Queen,60 the main issue before the Tax Court
pertained to arm’s-length fees paid in relation to prototype circuit boards. The
taxpayer was disbursing a setup fee to its related counterparty located in Barbados
for the “functions of validation, formatting, and panelizing” provided through specialized software.61 Although the taxpayer’s appeals were allowed, approximately
25 percent of the transfer-pricing adjustment made by the cra was sustained on the
basis that the taxpayer had “not met the onus of establishing that it did not overpay”
its related counterparty for the services provided.62
With respect to the comparability analysis, the court examined every function
performed in Barbados in return for the setup fees.63 Regarding the contractual
terms, Pizzitelli j remarked that the annual contracts between the taxpayer and its
60 2011 TCC 232.
62 Ibid., at paragraphs 234-245; the quoted text is taken from paragraph 243.
(2015) 63:2
counterparty in Barbados “did not contain many material terms, such as terms relating to liability or warranty with respect to poor work, terms dealing with dispute or
jurisdiction issues, or other usual contractual terms one would expect in an international agreement.”64 In short, the contractual terms demonstrated that the parties
were clearly not dealing at arm’s length.
It followed that the Tax Court again had to identify the comparability factors
that were relevant to the determination of an actual arm’s-length price. That determination was also required because the taxpayer and the cra had made different
assumptions prior to the selection and application of a transfer-pricing method.65
For the determination of the relevant comparability factors, Pizzitelli j relied on
the “guidelines” put forward in GlaxoSmithKline by the Federal Court of Appeal.66
According to the court, it followed that the contractual terms should reflect the
“real business world.”67 Therefore, as reiterated by the court, “the test requires a
consideration of all relevant factors that a reasonable business person in the Appellant’s shoes would consider.”68
After listing the five oecd comparability factors, Pizzitelli j stated that the list was
not intended to be exhaustive, citing the decisions in GlaxoSmithKline and General
Electric Capital Canada.69 Interestingly, the court emphasized that this meant that the
reasonable business person may have to consider relevant “factors or circumstances
that exist solely because of the non-arm’s length relationship of the parties.”70 On that
point, Pizzitelli j deferred unequivocally to the Federal Court of Appeal in General
Electric Capital Canada, which had stated that there was no error in law in taking into
consideration the implicit relationship between the non-arm’s-length parties.71
Pizzitelli j reiterated that “in short, all circumstances means ‘all ’ the circumstances an appellant finds himself in before a reasonable businessman steps into his
shoes” when the time comes to determine an arm’s-length price.72 As previously
discussed, this interpretation contradicts what was established by the Federal Court
of Appeal in GlaxoSmithKline.
70 Ibid., at paragraph 160. The court added that this had been demonstrated in GlaxoSmithKline
(FCA), supra note 46.
The Supreme Court Officially Gives Birth
to a Sixth Comparability Factor
Following the Federal Court of Appeal’s decision in GlaxoSmithKline, the cra appealed to the Supreme Court of Canada73 to inquire about the correct application
of subsection 69(2): “in particular, what circumstances are to be taken into account
in determining the reasonable arm’s length price against which to compare the nonarm’s length transfer price.”74 The taxpayer also appealed to the court, to contest
the decision to remit the transfer-pricing matter to the Tax Court of Canada, on the
basis that the taxpayer had “successfully demolished the Minister’s assumptions,
thus fully discharging the taxpayer’s burden in appealing the reassessment.”75 Both
appeals were dismissed.76
Although this is not explicitly mentioned in the decision, the Supreme Court was
likely aware of the conceptual tensions between the decisions rendered by the Federal Court of Appeal in General Electric Capital Canada and GlaxoSmithKline, with
respect to the exact nature of “economically relevant characteristics” to be included
in the comparability analysis and ultimately in the determination of an arm’s-length
price. The court aptly remarked that
section 69(2) does not, itself, offer guidance as to how to determine the “reasonable
amount” that would have been payable had the parties been dealing at arm’s length.
However, the [oecd] Guidelines suggest a number of methods for determining whether
transfer prices are consistent with prices determined between parties dealing at arm’s
length.77
Starting from that premise, the court took upon itself the task of identifying the meaning of “comparability” with the respect to the application of the oecd guidelines.78
Essentially, the Supreme Court confirmed that other transactions may be useful
to the determination of an arm’s-length price. In other words, other transactions
may be relevant to “determine whether the actual price was or was not greater than
the amount that would have been reasonable had the parties been dealing at arm’s
length.”79 In GlaxoSmithKline, the relevance of the two agreements concluded by
73 Canada v. GlaxoSmithKline Inc., 2012 SCC 52.
75 Ibid.
76 The case was returned to the Tax Court of Canada (see file 98-712(IT)G) but was settled prior
to trial, although (according to the Financial Post) the details remain confidential: see Julius
Melnitzer, “GlaxoSmithKline Transfer Pricing Case Settled,” Financial Post, January 12, 2015
(http://business.financialpost.com/legal-post/glaxosmithkline-transfer-pricing-case-settled).
(2015) 63:2
the taxpayer (the licence agreement and the supply agreement) for the purpose of
the comparability analysis was clear.80
The Supreme Court indicated that other transactions could be used to carry out
the comparability analysis if they were “economically relevant circumstances” in the
determination of an arm’s-length price—that is, if they were part of the contractual
relationships between the parties not dealing at arm’s length, as stated in the oecd
guidelines.81
The court concurred with the reasoning of the Federal Court of Appeal and
rejected the applicability of Singleton.82 The court based its conclusion solely on
paragraph 1.15 of the 1995 oecd guidelines, which it quoted as follows:
Application of the arm’s length principle is generally based on a comparison of the
conditions in a controlled transaction with the conditions in transactions between independent enterprises. In order for such comparisons to be useful, the economically relevant
characteristics of the situations being compared must be sufficiently comparable. To be comparable means that none of the differences (if any) between the situations being compared
could materially affect the condition being examined in the methodology (e.g. price or
margin), or that reasonably accurate adjustments can be made to eliminate the effect
of any such differences.83
This led the court to find that the licence agreement was relevant.84 But in a somewhat
surprising turn of events, the court went further and also specified that “according
to the 1995 Guidelines, a proper application of the arm’s length principle requires
that regard be had for the ‘economically relevant characteristics’ of the arm’s length
and non-arm’s length circumstances to ensure they are ‘sufficiently comparable.’ ”85
In essence, the Supreme Court was following the somewhat flawed reasoning
put forward by the Federal Court of Appeal in General Electric Capital Canada. It is
worth emphasizing that there is no mention whatsoever of the relevance of nonarm’s-length circumstances or factors in either paragraph 1.15 or anywhere else in
the oecd guidelines (both the 1995 and the 2010 versions), for the purposes of the
comparability analysis or for the selection and application of a transfer-pricing
method. Paragraph 1.15 of the 1995 oecd guidelines (paragraph 1.33 of the 2010
version) does suggest that the comparability analysis starts with the “comparison of
80 Although not mentioned by the court, this conclusion also loosely ties in with paragraph 3.9 of
the OECD guidelines, supra note 5, which highlights cases where “separate transactions are so
closely linked or continuous that they cannot be evaluated adequately on a separate basis.”
81 See ibid., at paragraph 1.52.
82 Shell Canada Ltd. v. Canada, [1999] 3 SCR, was also rejected on the same grounds—namely,
because there was no similarity between subparagraph 20(1)(c)(i) and subsection 69(2).
83 GlaxoSmithKline, supra note 73, at paragraph 41 (emphasis added). See the 1995 OECD
guidelines, supra note 2, and paragraph 1.33 of the 2010 version, supra note 5.
84 GlaxoSmithKline, supra note 73, at paragraphs 43-65, in particular paragraphs 55-60.
the conditions in a controlled transaction with the conditions in transactions between
independent enterprises.” But as I pointed out in the introduction to this article, in
order for that comparison to make any conceptual sense, it is solely the arm’s-length
factors that should be considered with respect to the determination of an arm’slength transfer price.
I also noted in the introduction that this methodological approach pertaining to
the comparability analysis is aligned with article 9 of the oecd model tax convention, which provides for transfer-pricing adjustments when “conditions . . . made or
imposed between the two [associated] enterprises in their commercial or financial
relations . . . differ from those which would be made between independent enterprises.”86 That is, that any such adjustment is meant to correspond to the conditions
that would have existed between parties dealing at arm’s length.
Non-arm’s-length conditions, circumstances, and factors must be excluded for
the purpose of the application of the arm’s-length principle and the determination
of an arm’s-length transfer price. Non-arm’s-length factors are eliminated by virtue
of the comparability analysis carried out, an analysis that directly pertains to the
relevant arm’s-length factors in the transaction itself, the parties involved in the
transaction, and the conditions and circumstances surrounding the transaction. In
this respect, it is worth repeating the following statement in paragraph 1.6 of the
oecd guidelines:
Because the separate entity approach treats the members of an mne group as if they
were independent entities, attention is focused on the nature of the transactions between those members and on whether the conditions thereof differ from the conditions
that would be obtained in comparable uncontrolled transactions.87
As discussed above, a methodological focus on the nature of the transaction
clearly should prevent the inclusion of any non-arm’s-length factor in the comparability analysis for transfer-pricing purposes. After all, in the end, the objective of
the comparability analysis is to properly apply the arm’s-length principle—that is, to
identify a transfer price that reflects the terms and conditions that would be found
between unrelated parties. It is therefore those terms and conditions found between
unrelated parties that must be the definitive focus of the comparability analysis for the
determination of any transfer price in accordance with the arm’s-length principle.
This specific point was introduced by the Federal Court of Appeal in GlaxoSmith
Kline, where the court stated unequivocally that the two agreements between the
parties not dealing at arm’s length had to be considered in determining an arm’slength price because they were circumstances that did “not arise from the non-arm’s
length relationship between the appellant and Adechsa or between the appellant
86 Supra note 12.
87 OECD guidelines, supra note 5.
(2015) 63:2
and Glaxo Group.”88 This was, and still is, the proper application of the “economically relevant characteristics” test suggested by the Supreme Court of Canada in
GlaxoSmithKline, in carrying out the comparability analysis applied in transfer pricing. That test reflects the intent and spirit of the oecd guidelines with respect to
the comparability analysis and the intent of section 247.89
As discussed above, the non-arm’s-length factors are rendered ineffective in the
arm’s-length price determination. It is worth repeating that the purpose of the arm’slength principle in international taxation is to adjust the terms and conditions of any
controlled transaction to the terms and conditions of an arm’s-length transaction.
Accordingly, non-arm’s-length factors included in the controlled transaction must be
discarded. A factor that exists solely in a non-arm’s-length environment cannot be compared with an arm’s-length arrangement since that specific factor does not exist in the
arm’s-length environment.
All of this to say that in GlaxoSmithKline the Supreme Court went too far. After
inferring language that is absent from the oecd guidelines and departs from the
intent of the Act, the court indicated that the application of the arm’s-length principle through the lens of subsection 69(2) (current section 247) required that both
the arm’s-length and the non-arm’s-length economically relevant factors be taken
into account for the purpose of the comparability analysis.
Like the Federal Court of Appeal in General Electric Capital Canada, the Supreme
Court did not appreciate that the sole purpose of the comparability analysis for
transfer-pricing purposes is to render ineffective the non-arm’s-length factors in the
determination of an arm’s-length transfer price.
A Sixth Comparability Factor: Non-Arm’s-Length
Factors Surrounding the Relationship
While the Supreme Court’s conclusions may be unfortunate, they cannot be ignored
by lower courts. The quarrel over the correct transfer-pricing methodology is therefore not just theoretical in nature. In McKesson Canada Corporation v. The Queen,90
89 The Federal Court of Appeal in GlaxoSmithKline was basically confirming the government’s
intent in modifying its transfer-pricing legislation in 1998 in order “to harmonize the standard
contained in section 69 of the Act with the arm’s length principle as defined in the revised
OECD guidelines” (1997 Budget Plan, supra note 2, at 203). As discussed in the introduction
to this article, the amendments implemented a transfer-pricing regime based explicitly on the
arm’s-length principle. New subsection 247(1) included definitions of “arm’s length allocation”
(of profit or loss) and “arm’s length transfer price,” to ensure that transactions between Canadian
taxpayers and non-arm’s-length non-residents would reflect what would have occurred if those
parties had been dealing at arm’s length. The inclusion of any non-arm’s-length factor in the
determination of an arm’s-length price is not considered in these definitions for the application
of section 247.
90 2013 TCC 404 (appeal filed at the Federal Court of Appeal on 10 January 2014: files A-48-14
and A-49-14).
the Tax Court had to determine whether the value of pools of receivables sold at a
discount by the Canadian taxpayer to its Luxembourg parent company met the
arm’s-length standard in section 247. The transactions were regulated by a receivable sales agreement and a servicing agreement—that is, by alleged arm’s-length
contractual terms.91
Deferring to the Supreme Court’s decision in GlaxoSmithKline, Boyle j stated that
[a] judge is to take into account all transactions, characteristics and circumstances that
are relevant (including economically relevant) in determining whether the terms and
conditions of the transactions or series in question differ from the terms and conditions to which arm’s length parties would have agreed.92
Boyle j also noted that the oecd guidelines were not written by legislators. Accordingly, the legal provisions of the Act were predominant in the “fact finding and
evaluation mission by the Court.”93 Furthermore, the determination of an arm’slength price required an examination of the contractual terms from the perspectives
of both the Canadian taxpayer and its related counterparty.94
Going one step further, in McKesson Canada Corporation, the court took it upon
itself to address the issue of “factors that exist only because of the non-arm’s length
relationship.”95 Following in the footsteps of Pizzitelli j in Alberta Printed Circuits,
Boyle j pointed out that the question whether non-arm’s-length factors are relevant
in an arm’s-length analysis may not arise in the context of a single transaction.
However, “[t]he question does appear significant in the context of a long-term commitment to do things over a period of time.”96
This assertion crystallized what appears to be characterized as a sixth distinctive
comparability factor in Canadian transfer pricing—the relevance of non-arm’slength factors surrounding the relationship between related parties. Boyle j justified
his conclusions on the scope of the comparability analysis by inferring that without
a thorough examination of these relevant non-arm’s-length factors,
companies within wholly controlled corporate groups could enter into skeletal agreements conferring few rights and obligations to the non-resident participant . . . all with
the view to obtaining a more favourable transfer price to reduce Canadian taxes.”97
93 Ibid.
94 Ibid.
(2015) 63:2
However, in offering this rationale, Boyle j seemed to disregard his own analysis of
paragraphs 247(2)(b) and (d),98 which were included in the legislation specifically
for the purpose of addressing non-arm’s-length arrangements entered into primarily to obtain a tax benefit.99
In McKesson Canada Corporation, the court was seemingly drawing transfer-pricing
policy from the conclusions of the Supreme Court in GlaxoSmithKline.
Co n c l u s i o n
The sixth distinctive comparability factor in Canadian transfer pricing created by
the courts does not reflect the actual intent of the Act. Although the government
modified its transfer-pricing legislation in 1998 to implement changes aligned with
the 1995 oecd guidelines, the arm’s-length principle as a transfer-pricing policy
principle in Canada has never been challenged since its original inception in section 23b of the Income War Tax Act in 1939.100
The 1998 amendments reaffirmed a “transfer pricing regime based explicitly on
the arm’s length principle.”101 In no way are non-arm’s-length factors part of the
arm’s-length principle; in fact, quite the opposite. From a Canadian transfer-pricing
perspective, it follows that non-arm’s-length factors should not be included in the
comparability analysis or in the examination of the “reasonable efforts” made by
the taxpayer to use arm’s-length allocations or prices.
The latter point may lead to some controversy in future Canadian transfer-pricing
cases, since “reasonable efforts” are of the utmost importance to the taxpayer wary
of potential transfer-pricing penalties under subsection 247(3). Transfer Pricing Memorandum tpm-09 clearly states:
A reasonable effort means the degree of effort that an independent and competent
person engaged in the same line of business or endeavour would exercise under similar
circumstances. What is reasonable is based on what a reasonable business person in
the taxpayer’s circumstances would do, having regard to the complexity and importance of the transfer pricing issues that arise in the taxpayer’s case.102
As discussed above, for Canadian transfer-pricing purposes, the Canadian courts
have adopted an interpretation of “reasonable business person” that now includes
the consideration of non-arm’s-length factors in the determination of an arm’slength price. But that interpretation substantially differs from the views of the
oecd and the cra on the matter.
99 See Explanatory Notes, supra note 4, at 508-15.
100 Income War Tax Act, 1939, SC 1939, c. 46, section 13; repealed.
101 See the summary of the Income Tax Amendments Act, 1997, supra note 3.
102 Transfer Pricing Memorandum TPM-09, “Reasonable Efforts Under Section 247 of the Income
Tax Act,” September 18, 2006, at 4.
By way of its decision in GlaxoSmithKline, the Supreme Court of Canada simply
made new legislation. This singular interpretation of the oecd guidelines by the
Supreme Court introduced a sixth comparability factor in Canadian transfer pricing—
the relevance of non-arm’s-length factors surrounding the relationship between
related parties.
In short, what seems to have been a methodological oversight by the Supreme
Court of Canada in regard to the proper application of the arm’s-length principle
through the comparability analysis is more than just unfortunate. It may have repercussions on what constitutes “reasonable efforts” and the ensuing application of
transfer-pricing penalties. It is to be hoped that future transfer-pricing cases will
help to correct, and not intensify, this misunderstanding of the comparability analysis in applying the arm’s-length standard.
Risk-Based Overrides of Share Ownership
as Specific Anti-Avoidance Rules
Tim Edgar*
Précis
Le présent article permet de conceptualiser les dispositions législatives récemment
adoptées relativement aux « contrats dérivés à terme » (cdt), aux « arrangements de
disposition factice » (adf) et aux dispositions proposées ultérieurement quant aux
« arrangements de capitaux propres synthétiques » (acps), en tant qu’exemples de
règles anti-évitement qui constituent une dérogation fondée sur le risque au concept
de l’actionnariat aux fins d’une loi d’intérêt privé, au moment de conférer des
attributs fiscaux. Même si ces règles anti-évitement visent des opérations distinctes
sans lien entre elles, elles partagent certaines grandes caractéristiques. Avant toute
chose, elles prennent la forme de règles détaillées, plutôt que de normes exprimées
de façon générale, et elles comportent une exposition au risque, de manière explicite
ou implicite, en tant que procuration pour le contribuable, dans le cadre d’un éventail
d’opérations ciblées. L’auteur soutient que cet ensemble de dérogations à la
propriété fondées sur le risque, dans la Loi de l’impôt sur le revenu, pourrait être
refondu en approchant de manière uniforme le ciblage de chacune des opérations
pertinentes. Cette approche exige que l’exposition au risque soit spécifiée
explicitement en tant que pourcentage, ce qui rend possible un ciblage plus précis
(particulièrement si cette spécification diffère en fonction des dérogations) qu’à
l’heure actuelle, alors que sont utilisées des formules verbales imprécises ou des
procurations en matière d’exposition au risque. Mais des spécifications précises
nécessitent une évaluation précise de l’exposition au risque, ce qui pose certains
défis pratiques et laisse aussi entendre certaines limites pratiques. De plus, les
limites établies dans la Loi de l’impôt sur le revenu qui sont la source de la substitution
fondée sur des considérations fiscales posent particulièrement problème lorsqu’il
* Of Osgoode Hall Law School, York University, Toronto (e-mail: [email protected]).
Earlier versions of this article were presented at the Tax Law and Policy Workshop, Faculty
of Law, University of British Columbia and the Tax Policy and Research Symposium—
Perspectives from Law and Accounting, held at the Waterloo Centre for Taxation in a
Global Economy, University of Waterloo and sponsored by Deloitte LLP. The author
would like to thank workshop and symposium participants for comments and suggestions.
The article benefited in particular from the input of Neil Brisley, Hugh Chasmar, Wei Cui,
Ranjini Jha, Alan Macnaughton, Gordon Mackenzie, and Lindsay Tedds as well as two
anonymous reviewers.
 397
(2015) 63:2
s’agit de faire une distinction nette entre les positions fondées sur des considérations
fiscales et celles qui ne le sont pas, concernant une propriété caractéristique d’une
spécification particulière et qui devrait, idéalement, être appuyée par une norme
visant une certaine fin, plutôt que de reposer sur la règle générale anti-évitement de
l’article 245 de la Loi. L’auteur laisse entendre qu’en dépit du scepticisme présent
dans la documentation, le concept financier de delta fournit un outil de mesure
utilisable avec les positions des actions négociées. Ce potentiel n’est pas négligeable,
étant donné que les actions négociées seraient le point de mire des opérations
d’évitement pertinentes visées par les dérogations fondées sur le risque. En ce qui
concerne les autres types de propriété, y compris les actions non négociables, une
détermination imprécise de l’exposition au risque faisant appel à des formules
verbales et à une norme fondée sur la fin peut être utilisée de préférence à l’ensemble
actuel d’approches éclectiques du ciblage des dérogations fondées sur le risque en
tant que règles anti-évitement particulières.
Abstract
This article conceptualizes recently enacted legislation addressing “derivative forward
agreements” (dfas) and “synthetic disposition arrangements” (sdas), as well as the
subsequently proposed provisions addressing “synthetic equity arrangements” (seas),
as examples of specific anti-avoidance rules that are risk-based overrides of the concept
of share ownership for private-law purposes in the assignment of income tax attributes.
Although these specific anti-avoidance rules address discrete and unrelated
transactions, they share some broad design features. Most importantly, they take the
form of detailed rules rather than generally expressed standards, and they incorporate
risk exposure, either explicitly or implicitly, as a proxy for taxpayer purpose for a range of
targeted transactions. The author argues that this set of risk-based overrides of
ownership in the Income Tax Act could be recast using a uniform approach to the
targeting of each of the relevant transactions. This approach involves the specification
of a level of risk exposure explicitly as a percentage amount, which holds the possibility
of more accurate targeting—particularly if the specification is tailored differently for
different overrides—than is currently the case using either imprecise verbal formulas or
proxies for risk exposure. Precise specification requires, however, precise measurement
of risk exposure, which presents certain practical challenges and suggests some
practical limitations. Moreover, boundaries in the tax law that are the source of taxdriven substitution are especially problematic with respect to a bright-line distinction
between tax-driven and non-tax-driven positions in property that is characteristic of
precise specification and should ideally be backstopped with a purpose-based standard
rather than reliance on the general anti-avoidance rule in section 245 of the Act. The
author suggests that, despite skepticism in the literature, the financial concept of delta
provides a measurement tool that is feasible with positions in traded shares. This
potential is not inconsiderable, since traded shares tend to be a particular focus of the
relevant set of avoidance transactions addressed by risk-based overrides. For other
types of property, including non-traded shares, an imprecise specification of risk
exposure using verbal formulas and a purpose-based standard may be used in
preference to the existing set of eclectic approaches to the targeting of risk-based
overrides as specific anti-avoidance rules.
Keywords: Risk n tax avoidance n hedging n derivatives n shares n ownership
risk-based overrides of share ownership as specific anti-avoidance rules  n  399
Contents
Introduction
Legislative Pattern Assigning Tax Attributes to Positions in Shares
Assignment of Tax Attributes Based on Share Ownership and the Policy
Significance of Risk
Form of Legislative Overrides: Rules Versus Standards
Specification of Risk Exposure
Line Drawing and the Legislative Expression of a Level of Risk Exposure
Specifying Risk Exposure Differently for Different Overrides
Synthetic Dispositions
Dividend Stop-Loss Rules and Foreign Tax Credit Trading
Dividend Rental Arrangements and Synthetic Equity Arrangements
Character Conversion Transactions—Derivative Forward Agreements
Wash Sales
Suppressing Discontinuities Using a Purpose-Based Standard
Discontinuities as a Focus of Tax-Driven Substitution
A Purpose-Based Standard as the Preferable Policy Instrument To Suppress
Discontinuities
Application of a Purpose-Based Standard for Different Specifications of
Risk Exposure
Measurement of Risk Exposure
Methodological Indeterminacy as an Approach to Risk Measurement
Use of a Delta-Based Measure as an Exception to Methodological
Indeterminacy
Conclusion
399
405
406
412
417
418
422
423
425
428
430
435
437
437
441
449
450
451
455
465
Introduc tion
The March 21, 2013 federal budget included proposals for specific anti-avoidance
rules intended to address what the budget documents refer to as “character conversion transactions” and “synthetic dispositions.”1 Draft legislative rules were released
on September 13, 2013 and subsequently enacted on December 12, 2013.2 The
April 21, 2015 federal budget followed these two legislative initiatives with a similar
1 Canada, Department of Finance, 2013 Budget, Budget Plan, March 21, 2013, at 341-43, and
resolution 19 of the accompanying Notice of Ways and Means Motion To Amend the Income
Tax Act and Other Tax Legislation.
2 Canada, Department of Finance, “Government of Canada Moving Forward with Its Plan for
Jobs, Growth and Long-Term Prosperity,” News Release 2013-117, September 13, 2013, and
the accompanying Legislative Proposals Relating to the Income Tax Act, the Excise Tax Act, and the
Income Tax Regulations (Ottawa: Department of Finance, September 2013), at clauses 2, 9(3),
24, and 50, proposing the following amendments to the Income Tax Act (RSC 1985, c. 1
(5th Supp.), as amended; herein referred to as “the Act”): paragraphs 12(1)(z.7) and 20(1)(xx);
section 80.6; and subsection 248(1) definitions of “derivative forward agreement,” “synthetic
disposition,” and “synthetic disposition period”; enacted by SC 2013, c. 40; royal assent
December 12, 2013. (Unless otherwise stated, statutory references in this article are to the Act.)
(2015) 63:2
proposal for a specific anti-avoidance rule intended to address “dividend rental arrangements” (dras).3 The definition of a “derivative forward agreement” (dfa) targets
character conversion transactions.4 The definition of a “synthetic disposition arrangement” (sda) targets synthetic dispositions.5 The proposed definition of a “synthetic
equity arrangement” (sea) targets dras.6
Broadly speaking, dfas, sdas, and seas are examples of synthetic financial instruments, consisting of the combination of long and short positions in legally distinct
instruments that provide offsetting cash flows replicating the cash flows associated
with a single instrument or transaction the income tax treatment of which is often
inconsistent with the sum of the treatments applied to the components of the synthetic.7 dfas purport to convert what would otherwise be a non-capital amount associated with a single instrument to a capital amount associated with the synthetic.
sdas purport to lock in unrealized appreciation on an asset without attracting a
disposition for income tax purposes. seas purport to provide the benefit of dividend
tax relief to taxpayers for whom it would otherwise be unavailable. The apparent
motivation for the dfa legislation was some well-known structured transactions
involving either forward sales or forward purchases of publicly traded shares by
mutual funds, in which the price was determined by reference to a portfolio of
fixed-income securities.8 The immediate transactional motivation for the sda legislation is unclear, given that a variety of transactions purporting to lock in gain
without a disposition have been used since at least the mid- to late-1990s.9 The
3 Canada, Department of Finance, 2015 Budget, Budget Plan, April 21, 2015, at 461-63, and
resolutions 31 and 32 of the accompanying Notice of Ways and Means Motion To Amend the
Income Tax Act and Other Tax Legislation.
4 Subsection 248(1), definition of “derivative forward agreement.”
5 Subsection 248(1), definition of “synthetic disposition arrangement.”
6 See the notice of ways and means motion, supra note 3, at resolution 32(2).
7 Tim Edgar, The Income Tax Treatment of Financial Instruments: Theory and Practice, Canadian Tax
Paper no. 105 (Toronto: Canadian Tax Foundation, 2000), at 313-14.
8 See, for example, Jerald M. Wortsman, Christopher S. Parks, and Leonard Nesbitt, “Complex
Financial Instruments: An Industry and Practitioner Perspective on Equity Derivatives,” in
Report of Proceedings of the Sixty-First Tax Conference, 2009 Conference Report (Toronto:
Canadian Tax Foundation, 2010), 8:1-41, at 8:21-24 (reviewing reasons for the use of “share
basket forward transactions” and describing a generic transactional example). See also Edward
Miller and Matias Milet, “Derivative Forward Agreements and Synthetic Disposition
Arrangements,” in Report of Proceedings of the Sixty-Fifth Tax Conference, 2013 Conference
Report (Toronto: Canadian Tax Foundation, 2014), 10:1-50, at 10:2-3 (describing generic
forward sale and forward purchase agreements).
9 See Wortsman et al., supra note 8, at 8:3 (noting that equity monetization transactions have
been used since the 1980s, with a surge in volume associated with rising equity values in the
mid- to late-1990s that subsided after the market crash but increased again with a rising equity
market just before the financial crisis of 2007-2009). See also Paul D. Hayward, “Monetization,
Realization, and Statutory Interpretation” (2003) 51:5 Canadian Tax Journal 1761-1824, at
1769-72 (providing a brief history of the use of monetization transactions in the United States);
apparent motivation for the sea proposals is certain dividend equivalent payment
transactions undertaken by some Canadian banks with the purported result of
avoiding the existing purpose-based legislation addressing dras.10 The Department
of Finance suggests in the budget documents that all such transactions could be
challenged under existing provisions of the Act, presumably including the general
anti-avoidance rule (gaar) in section 245, but that any challenge would be timeconsuming and costly, and specific anti-avoidance rules are preferable “to ensure
that the appropriate tax consequences apply.”11
Both the dfa and the sda legislation, as well as the sea proposals, incorporate
features of existing legislative templates such as the conversion transaction legislation12 and constructive sale legislation13 in the United States, and draft anti-synthetic
rules that were proposed, but not enacted, in Australia as part of a larger project on
the taxation of financial arrangements.14 This kind of anti-synthetic legislation ignores the legal form of separate financial instruments as tax-driven structuring in
which offsetting cash flows substantially replicate the risk exposure associated with
a particular position.15 Offsetting positions are defined largely through a description
of a set of paradigm transactions providing cash flows that are perfect or near-perfect
substitutes for the cash flows associated with fixed-payment debt, a disposition of
property with accrued gain, or share ownership. The paradigm transactions are defined in terms of a level of risk exposure that is considered indicative of the tax-driven
use of a synthetic instrument or transaction. A character conversion rule typically
and Miller and Milet, supra note 8, at 10:20 (citing anecdotal evidence that Canadian financial
institutions have promoted equity monetizations to high net worth individuals at least since the
beginning of the past decade).
10 2015 Budget Plan, supra note 3, at 461. See Tim Kiladze, “Why Canada’s Banks Care About
‘Synthetic’ Trade Crackdown,” Globe and Mail, April 22, 2015; and Barbara Shecter, “Some of
Ottawa’s Budget Windfall Could Come from Canada’s Big Banks,” Financial Post, April 22, 2015.
11 2013 Budget Plan, supra note 1, at 341 and 343; and 2015 Budget Plan, supra note 3, at 461.
The same motivation is cited for a series of unrelated specific anti-avoidance rules proposed in
the March 2013 budget as well as amendments to section 55 in the 2015 budget to address
certain corporate capital gains stripping transactions. For detailed analyses of the possible
application of GAAR to DFAs and SDAs, see Wortsman et al., supra note 8, at 8:18-21 and
8:32-37. See also Hayward, supra note 9, at 1802-12 (discussing the possible application of
GAAR to equity monetization transactions).
12 Internal Revenue Code of 1986, as amended (herein referred to as “IRC”), section 1258.
13 IRC section 1259.
14 Australia, Department of the Treasury, “Exposure Draft: Income Tax Assessment Act 1997,”
subdivisions 230-J (deemed disposal), 230-JA (deemed non-disposal), and 230-JB (deemed gain
from financial arrangement). See, in this respect, Australia, Department of the Treasury,
“Taxation of Financial Arrangements—Synthetic and Complex Arrangements,” Media Release
no. 022, March 26, 2009 (acknowledging that, after consultations on synthetic arrangements,
draft provisions released in 2007 for comment were not included in implementing legislation,
but stating that the need for special integrity measures to address synthetic arrangements
would be monitored against the background of the Australian GAAR).
15 Edgar, supra note 7, at 350-53.
(2015) 63:2
recharacterizes realized capital gain as a non-capital amount. A constructive sale rule
typically deems the taxpayer to have disposed of the relevant long position in the appreciated asset for proceeds equal to fair market value, and to have reacquired the
asset at a cost equal to those proceeds. Dividend tax relief is typically denied for dras
and similar transactions that attempt to transfer such relief between taxpayers.
In a recent conference paper, Miller and Milet16 provide a thorough technical
analysis of the dfa and sda legislation. This article focuses on an issue that they
canvass17 and that is the principal targeting feature of the legislation and the sea
proposals: that is, the specification and measurement of a level of risk exposure.
This issue is examined here from a wider policy perspective focused on the specification of risk exposure for the purpose of what may be seen as an entire family of
specific anti-avoidance rules in the Act that override share ownership as the basis for
what Thompson and Weisbach label “the assignment of tax attributes.”18 In addition to the dfa and sda legislation and the sea proposals, risk-based rules in the Act
that override the concept of share ownership include the anti-foreign-tax-credittrading rule applicable to short-term securities acquisitions,19 the dividend stop-loss
rules,20 the dra rules,21 the wash sale rules,22 and the rules governing securities
lending arrangements (slas).23 With the exception of the sla legislation, these rules
use risk exposure, either explicitly or implicitly, primarily as a proxy for taxpayer
purpose in identifying positions in shares that are considered to be tax-avoidance
transactions intended to separate private-law ownership from economic exposure,
but otherwise leave the assignment of tax attributes following ownership as a matter
of private law.
This article is motivated by Thompson and Weisbach’s critique of some recent
us case law that departs from a formalistic concept of securities ownership as the
basis for the assignment of tax attributes. They prefer that such departures be im-
16 Miller and Milet, supra note 8.
17 Ibid., at 10:22-33.
18 Reid Thompson and David Weisbach, “Attributes of Ownership” (2014) 67:2 Tax Law Review
249-96.
19 Subsection 126(4.2).
20 Subsections 112(3) through (7).
21 Subsections 82(1) and 112(2.3), and subsection 248(1), definition of “dividend rental
arrangement.”
22 The wash sale rules also extend to property other than shares. See subsection 13(21.2) (terminal
loss denied on a disposition of depreciable property by a person or partnership); subsections
14(12) and (13) (loss denied on a disposition of eligible capital property by a corporation,
partnership, or trust); subsections 18(13) through (16) (loss on a disposition of inventory
denied); subsections 40(3.3) through (3.5) (capital loss denied on a disposition by a corporation,
partnership, or trust); and subparagraph 40(2)(g)(i) and section 54, definition of “superficial
loss” (capital loss denied on a disposition by an individual other than a trust).
23 Section 260.
plemented through targeted legislative overrides, but they do not explore the design
of those overrides. In the context of hedged and derivative positions in shares, this
article considers how the category of such legislative overrides that are risk-based
can be specified in an admittedly tentative effort to improve target-effectiveness
with a tolerable level of administrative and compliance intensity. In this respect, the
article explores three claims.
The first is that a single approach could be developed for the articulation of the
set of risk-based anti-avoidance rules in the Act, in an attempt to lower costs of tax
administration and compliance. Some differences in approach may be defensible,
however, to tailor application to the paradigm tax-avoidance transactions that
otherwise attempt to manipulate the private-law concept of share ownership in the
assignment of tax attributes.
The second claim is that a required level of risk exposure could be specified explicitly as a percentage amount across all of the risk-based overrides in an attempt
to improve their target-effectiveness. An important constraint on realization of this
desirable policy feature is a lack of any systematic empirical evidence regarding the
boundary between tax-driven and non-tax-driven positions in shares that is defined
by risk-based overrides. Given this lack of systematic evidence, the policy maker is
left with only unsatisfactory anecdotal evidence that may indicate the general direction in which targeting could be improved. Precise specification also requires the
use of a targeted purpose-based standard to suppress problematic discontinuities
where small changes in the level of risk exposure could otherwise result in disproportionate changes in tax treatment.
The third claim is that the measurement of risk exposure could be based on the
financial concept of delta. Delta is a measurement tool used by options traders to
calculate the rate of incremental price change between a derivative instrument and
an underlying asset. Although noted in some of the literature, the possible use of
delta for tax purposes tends to be rejected as too administrative- and complianceintensive.24 Yet delta is explicitly required to be used to measure risk exposure for
24 See, for example, Miller and Milet, supra note 8, at 10:27-28. See also David M. Schizer,
“Scrubbing the Wash Sale Rules” (2004) 82:3 Taxes: The Tax Magazine 67-88, at 78 (rejecting a
delta-based measure of risk exposure for the purpose of the wash sale rules as too sophisticated
for many taxpayers); David M. Schizer, “Frictions as a Constraint on Tax Planning” (2001)
101:6 Columbia Law Review 1312-1409, at 1364-67 (rejecting the use of delta as the basis for a
partial disposition constructive sale rule); and New York State Bar Association, Tax Section,
Comments on “Short-Against-the-Box” Proposal, report no. 868 (Albany, NY: NYSBA, March 1,
1996), at 20 (observing that delta analysis is based on mathematical models and may be too
esoteric to be administrable). But see also Thomas J. Brennan, “Law and Finance: The Case of
Constructive Sales” (2013) 5 Annual Review of Financial Economics 259-76 (arguing that the
addition of a delta-based test would improve the targeting of the constructive sale rule in IRC
section 1259).
(2015) 63:2
the purpose of the Australian anti-dividend-credit-trading rules25 and in the us
proposed regulations under the dividend equivalent payment rules.26 This limited
use of delta may indicate that any administrative and compliance burden is at least
not prohibitive for those risk-based overrides that can be seen to require a more
precise specification of risk exposure as a targeting mechanism than can otherwise
be realized using standard verbal formulas. Because of limited availability of price
data, however, the use of delta may not be feasible for positions in non-traded shares
as well as property other than shares that is subject to certain of the risk-based overrides. A legislative alternative to precise specification as a percentage amount using
a delta-based measure would combine an expression of risk exposure using an imprecise verbal formula with a purpose-based inquiry.
This article is limited to consideration of the specification of risk exposure as a
targeting mechanism under risk-based specific anti-avoidance rules that override
share ownership in the assignment of tax attributes. Although there are other features of risk-based overrides that also have a targeting effect, the specification of a
level of risk exposure is the most significant targeting feature common to all such
overrides in a tax-avoidance context and warrants an extended policy analysis not
found in the existing literature. Abstracting from the other features of these riskbased overrides allows for this more policy-intensive treatment, although there is
some consideration of these features targeting overlays of the specification of risk
exposure. Moreover, because the targeting role is quite different in a non-avoidance
context, the article does not address the specification of risk exposure for other
overrides, such as the sla legislation, which are intended to clarify the treatment of
transactions in shares that are not tax-driven; nor is there any discussion of the policy case for repeal of any of the existing legislative overrides, or any discussion of the
policy case for the addition of other overrides, such as a character conversion rule
that would address the use of long derivative positions in shares to avoid non-resident
25 Australia, Income Tax Assessment Act 1936 (herein referred to as “ITAA 1936”), former
sections 160APHO-160APHU. As part of a comprehensive rewrite of the Australian income
tax legislation, the dividend imputation rules in the ITAA 1936 were repealed and incorporated
in the Income Tax Assessment Act 1997 (herein referred to as “ITAA 1997”). The anti-dividendcredit-trading rules were not, however, incorporated in the ITAA 1997. Instead, section 207-145
of the ITAA 1997 provides that taxpayers are not entitled to dividend credit relief unless they
are within the definition of a “qualified person” under division 1A of former part IIIAA of the
ITAA 1936 (hold period requirement and related payment rules). The use of delta for the
purposes of the Australian rules is discussed in detail in the text below, at note 198 and following.
26 IRC section 871(m) and prop. Treas. reg. section 1.871-15, REG-120282-10, 2013-52 IRB
837. The proposed regulations are issued under the rule-making authority to (1) specify when a
notional principal contract (NPC) “is of a type which does not have the potential for tax
avoidance” (IRC section 871(m)(3)(B)); and (2) specify other payments as dividend equivalents
because they are substantially similar to specified NPC payments and substitute payments
under SLAs (IRC section 871(m)(2)(C)). For a detailed discussion of the use of delta as a
measure of risk exposure under the US rules, see the text below at note 198 and following.
withholding tax generally.27 In short, the existing set of risk-based overrides in the
Act that is intended to address discrete avoidance transactions is taken as a given;
that being the case, the suggested legislative template could be used in specifying
risk exposure for the purpose of any other similar overrides that might be adopted.
The discussion that follows is divided into five parts. Part one provides some
necessary background and is broadly descriptive of current law governing the assignment of tax attributes associated with shares. The possible policy rationale for
the legislative pattern is discussed, along with the rationale for the use of detailed
rules rather than standards in legislatively overriding the core assignment principle
based on share ownership. There is also some discussion of certain general types of
legislative overrides, which include risk-based anti-avoidance rules for hedged and
derivative positions in shares, and possible reasons for differences between these
types based on the nature of the tax attribute to be assigned. Parts two, three, and
four articulate the details of a legislative template that could be used in the specification of the set of risk-based anti-avoidance rules in the Act that override share
ownership in the assignment of tax attributes. The discussion in each of these parts
explores the three claims noted above regarding desirable design features. Part five
provides concluding comments.
L e g i s lat i v e Pat t e r n A s s i g n i n g Ta x
At t r i b u t e s t o P o s i t i o n s i n S h a r e s
The term “tax attributes,” as used in this article, applies to features or characteristics
of property that have income tax consequences, including liability to tax on income
or gain, the availability of expense or loss recognition, and the timing of the recognition of these same amounts.28 In the context of us income tax law, Thompson and
Weisbach describe the legislative pattern assigning tax attributes of shares as well as
securities more generally.29 They characterize us law as assigning tax attributes on
the basis of ownership as a matter of private law, subject to the application of targeted legislative overrides as well as an indeterminate judicial overlay. This same
27 See generally, Edgar, supra note 7, at 359-61. The need for a character conversion rule
addressing the use of long derivative positions in shares to avoid non-resident withholding tax
was the subject of considerable debate in the United States before the enactment of IRC
section 871(m). See, in this respect, United States Senate Committee on Homeland Security
and Governmental Affairs, Permanent Subcommittee on Investigations, Staff Report, Dividend
Tax Abuse: How Offshore Entities Dodge Taxes on U.S. Stock Dividends, 110th Cong., 2d sess.
(Washington, DC: Permanent Subcommittee on Investigations, September 2008). With the
enactment of IRC section 871(m), this avoidance business has migrated to the London-based
operations of many US financial institutions where long derivative positions in shares are used
to avoid non-resident withholding taxes in those countries without a comparable character
conversion rule. See Jenny Strasburg, “Fed Questions Bank Maneuver To Reduce Hedge
Funds’ Dividend Taxes,” Wall Street Journal, September 29, 2014.
28 See Thompson and Weisbach, supra note 18, at 250 (defining “tax attributes” as “the rules that
govern deviations from a pure Haig-Simons base”).
29 Ibid., at 254-69.
(2015) 63:2
pattern is broadly characteristic of Canadian income tax law, but in contrast to us
courts, Canadian courts have been much less willing to articulate an inquiry into
ownership that deviates significantly from the private-law concept.
In a tax-avoidance context, legislative overrides of ownership in the Act focus on
transactions that can be seen to manipulate the private-law concept to assign, or to
avoid the assignment of, tax attributes. Although they are associated with distinct
transactional fact patterns and different tax attributes, the legislative overrides have
two general design features in common: (1) identification of prohibited transactions
primarily without requiring an inquiry into taxpayer purpose, and (2) legislative
expression in the form of detailed rules rather than a generally expressed standard.
As a subset of legislative overrides, risk-based rules can be seen to use risk exposure,
either explicitly or implicitly, as a proxy for taxpayer purpose.
Assignment of Tax Attributes Based on Share Ownership
and the Policy Significance of Risk
Shares have a number of attributes that are tax attributes in the sense that they are
relevant to the determination of the income tax consequences of holding shares or
otherwise transacting in shares. For example, dividend distributions are taxable with
relief for corporate income tax assumed to be paid on the underlying income. For
resident individuals, the relief takes the form of the dividend gross-up and credit
mechanism,30 while an offsetting taxable income deduction is provided for taxable
Canadian corporations.31 For non-residents, part xiii withholding tax (reduced by
treaty where applicable) applies to dividends paid by Canadian-resident corporations.32 “Adjusted cost base”33 and “proceeds of disposition”34 determine capital gain
or loss realized on a “disposition”35 of property, including shares.36 “Paid-up capital”37
determines the character of distributions on certain transactions in shares as dividends
rather than a distribution of capital or a realized capital gain or loss. Because of the
judicial interpretation of the concept of corporate control, voting power is relevant for
a range of purposes,38 including the status of two or more corporations as “associated
30 Paragraphs 82(1)(a) and (b) and section 121.
31 Subsection 112(1).
33 Section 54, definition of “adjusted cost base.”
34 Section 54, definition of “proceeds of disposition.”
35 Subsection 248(1), definition of “disposition.”
36 Paragraphs 40(1)(a) and (b).
37 Subsection 89(1), definition of “paid up capital.”
38 But see Robert Couzin, “Some Reflections on Corporate Control” (2005) 53:2 Canadian Tax
Journal 305-32 (arguing that the concept of “control” articulated in the case law as residing in
ownership of shares with a majority of the votes in an election of the board of directors—
referred to as de jure control—remains uncertain in a range of fact patterns).
corporations,”39 “related corporations,”40 or “affiliated corporations,”41 as well as
the status of a non-resident corporation as a “controlled foreign affiliate”42 and a
resident closely held corporation as a “Canadian-controlled private corporation.”43
The income tax consequences that follow on an acquisition of corporate control44
similarly turn on the location of shareholder voting power. Other provisions, such
as status as a “specified non-resident” for the purpose of the thin capitalization
rules, 45 “connected” corporate status for part iv tax purposes,46 and a “substantial
interest” for the purpose of part vi.1 tax on dividends on taxable preferred shares,47
depend (in part) on the status of shares as voting shares.
In the first instance, these various tax attributes are assigned to the taxpayer who
owns or disposes of the shares as a matter of private law. Although this is admittedly
an oversimplification, the approach to the characterization of share ownership
under the Act has been, for the most part, formalistic, with the registered owner (or
in the case of certificated securities, the holder) being treated as the owner for income tax purposes.48 The fact that the definition of a disposition is inclusive only,
with the relevant jurisprudence supporting a broad interpretation that extends beyond a sale, has not led Canadian courts and the Canada Revenue Agency (cra) to
undertake any sort of wide-ranging assessment of the risk and return characteristics
of a particular arrangement and to articulate a concept of share ownership that is
unique to the tax law.49 In the absence of specific legislation, this approach has been
40 Paragraph 251(2)(c). The concept of corporate control is also relevant in determining when a
person and a corporation are related: paragraph 251(2)(b).
41 Paragraph 251.1(1)(c). The concept of corporate control is also relevant in determining when a
person and a corporation are affiliated: paragraph 251.1(1)(b).
42 Subsection 95(1), definition of “controlled foreign affiliate.”
43 Subsection 125(7), definition of “Canadian-controlled private corporation.”
44 Subsections 111(4) through (5.5). See also subsection 67(11), paragraph 87(2.1)(b), subsection
110.1(12), and subsection 111(12).
45 Subsection 18(5), definitions of “specified shareholder” and “specified non-resident shareholder.”
46 Paragraph 186(4)(b).
47 Subsection 191(2). See also paragraph 187.1(b), definition of “excepted dividend” for the purpose
of part IV.1 tax payable in respect of taxable dividends received on taxable preferred shares.
48 A formalistic approach to the concept of share ownership is apparent, for example, in the
characterization of SLAs in the absence of section 260 and of “repo” financing transactions.
See, for example, Jonathan W. Wilson, “Securities Lending: An Overview and Update for
Domestic and Cross-Border Transactions,” in the 2009 Conference Report, supra note 8,
9:1-36, at 9:5-6, reviewing the disposition issue, including relevant statements of the Canada
Revenue Agency’s (CRA’s) administrative position, with respect to SLAs and repos.
49 See, for example, Wortsman et al., supra note 8, at 8:20 (“There is no discernible policy
suggesting that a taxpayer should be considered to have realized capital gains on a property
solely because the taxpayer has transferred some of the economic risk in respect of the property”).
But see Hayward, supra note 9, at 1791-96 (arguing that the non-exhaustive definition of a
(2015) 63:2
maintained even though case law in the context of non-financial assets suggests that
it is an amalgam of title, possession, use, and risk that constitutes ownership for
income tax purposes.50 Moreover, a transfer of title (or a transfer of possession in
the case of securities held in certificated form) has been accepted as the occasion
supporting characterization as a disposition.51 This approach has been qualified in
some limited instances in which a change in the legal rights associated with shares,
as well as debt instruments, is considered to result in a disposition52 or share value
is shifted and effectively transferred between taxpayers.53 Although a concept of
beneficial ownership, in apparent contrast to legal ownership, is considered relevant
to the determination of ownership for income tax purposes, its content remains
unclear. With the possible exception of its use in certain tax treaty articles,54 the
concept has arguably not been invoked as the basis for a specification of ownership
that is unique to the tax system, requiring an inquiry into perceptions of economic
substance entirely independent of any private-law meaning.55
As described by Thompson and Weisbach, reliance on the private-law concept
of ownership as the basis for the assignment of tax attributes is similarly characteristic of us income tax law.56 They explore in some detail the private-law concept in
the context of publicly traded securities that are held indirectly through brokers,
clearinghouses, custodians, and nominees,57 and propose that ownership for us
income tax purposes should coincide with a security entitlement in the case of publicly traded securities held indirectly.58 In the case of shares held directly, they
disposition and the acceptance by Canadian courts that the concept can be given a broad
interpretation mean, in principle at least, that equity monetization transactions could be
treated as dispositions independent of the private-law characterization).
50 See Wortsman et al., supra note 8, at 8:9-11 (reviewing case law and CRA statements of
administrative practice regarding the incidents of ownership and its transfer sufficient to attract
disposition treatment). See also Hayward, supra note 9, at 1791-96.
51 See Wortsman et al., supra note 8, at 8:9-10.
52 See, for example, Interpretation Bulletin IT-448 (Archived), “Dispositions—Changes in Terms
of Securities,” June 6, 1980.
53 The case law is not entirely clear on the disposition issue. See, for example, The Queen v.
Kieboom, 92 DTC 6382 (FCA) (no disposition without a transfer but possible taxation as a
benefit conferred on the new shareholder).
54 See Velcro Canada Inc. v. The Queen, 2012 TCC 57; and Canada v. Prévost Car Inc., 2009 FCA 57
(considering the meaning of beneficial ownership in the tax treaty context).
55 Wortsman et al., supra note 8, at 8:9 (an acquisition of property occurs when the acquiror holds
either legal or beneficial ownership). For detailed discussions of the concept of beneficial
ownership under the Act, see Mark D. Brender, “Symposium: Beneficial Ownership in Canadian
Tax Law: Required Reform and Impact on Harmonization of Quebec Civil Law and Federal
Legislation” (2003) 51:1 Canadian Tax Journal 311-54; and Catherine Brown, “Symposium:
Beneficial Ownership and the Income Tax Act” (2003) 51:1 Canadian Tax Journal 401-53.
56 Thompson and Weisbach, supra note 18, at 254-69.
57 Ibid., at 281-84.
58 Ibid., at 284.
conclude that ownership should follow current us law and be equated with title.59
Thompson and Weisbach suggest, however, that ownership functions as what they
refer to as a “default rule” for the assignment of tax characteristics or attributes to
positions in securities.60 They argue that reliance on the private-law concept of
ownership as the default rule for assigning tax attributes to shares is desirable because it provides a simple basis for doing so, it “gets the assignment correct for the
overwhelming majority of cases,”61 and therefore, for the vast majority of cases, it
minimizes the costs of assigning attributes without loss of correctness.62 For those
cases in which the default rule is seen to realize an inappropriate assignment,
Thompson and Weisbach prefer the use of legislative overrides that can be tailored to
particular transactions in light of the underlying rationale for the relevant tax attribute.63 They emphasize that legislative overrides operate most clearly and effectively
where they do so on the basis of a “known-background rule”;64 and they critique the
decisions in Anschutz Co. v. Commissioner 65 and Calloway v. Commissioner 66 for the confusion that these decisions create by approaching the assignment of tax attributes
using a concept of ownership that is modified for income tax purposes by an indeterminate “substance-over-form” or “benefits and burdens” analysis, in an attempt to
address the separation of ownership as a matter of private law and the associated
economics of the positions in shares implemented by the taxpayers in these cases.67
Although Thompson and Weisbach assert that reliance on the private-law concept of share ownership gets the assignment of tax attributes correct for the vast
majority of cases, they do not suggest what criterion or criteria they have in mind
for the assessment of the correctness or incorrectness of their proposed default rule.
59 Ibid., at 284. Where shares are held in registered form, title would reside with the registered
owner. Where shares are held in certificated form, title would reside with the bearer.
60 Ibid., at 251.
61 Ibid., at 281. See also Edward D. Kleinbard, “Risky and Riskless Positions in Securities” (1993)
71:12 Taxes: The Tax Magazine 783-99; and Alex Raskolnikov, “Contextual Analysis of Tax
Ownership” (2005) 85:2 Boston University Law Review 431-516 (emphasizing the role of the
concept of share ownership for private-law purposes as determinative of share ownership for US
income tax purposes). Thompson and Weisbach, supra note 18, at 281, acknowledge that their
proposed approach is similar to that of Kleinbard and Raskolnikov but “with the strong emphasis
not found in those articles, that ownership is just a default rule for assigning attributes.”
62 Thompson and Weisbach, supra note 18, at 281.
63 Ibid.
64 Ibid.
65 135 TC 78 (2010); aff ’d. 644 F. 3d 313 (10th Cir. 2011).
66 135 TC 26 (2010).
67 Thompson and Weisbach, supra note 18, at 269-80 (also discussing the decision in Samueli v.
Commissioner, 132 TC 37 (2009), particularly with respect to its strained interpretation of the
assignment of tax attributes under the securities lending rules in IRC section 1058 rather than
the application of a judicially articulated override of the private-law concept of ownership in the
assignment of tax attributes).
(2015) 63:2
Their unstated assumption may be that reliance on the private-law concept of share
ownership to assign tax attributes is somehow correct because the owner bears all
risk of loss and opportunity for gain or profit.68 The correctness of such an assumption is not, however, obvious on its face and requires some unpacking. Descriptively
at least, the legislative concept of a disposition—and, by implication, the concept of
ownership—can be seen to be based on three assumptions centred on risk: first, that
the essence of a disposition is the transfer of risk associated with an asset; second,
that transfer of the ownership of an asset as a matter of private law is an accurate
proxy for changes in risk; and third, that changes in risk appropriately signal a
change in the assignment of tax attributes associated with share ownership.69 In fact,
as Thompson and Weisbach emphasize, ownership has no income-measurement
role to play under an ideal income tax that brings into account unrealized gains and
losses annually on an accrual basis. Once an income tax system moves away from that
ideal to a realization-based approach for apparent practical reasons, such as liquidity
and valuation concerns, there are no obvious principles that suggest an appropriate
boundary between a disposition transaction and the retention of ownership. Instead
of an assessment of risk unique to the tax law as the basis for the application of this
tax distinction, the private-law concept of ownership is accepted as a proxy. Hedged
and derivative positions in shares challenge this reliance on the private law as an
accurate proxy and lead to responses based on perceptions of risk associated with
offsetting long and short positions in an asset.70
Yet these observations only invite the normative question: “Why should risk
have any significance in the assignment of tax attributes?” Risk is a measure of the
variability of cash flows that, in turn, affect the pricing and value of financial and
non-financial assets. It is suggested here that the normative significance of risk—with
both a reliance on ownership as a matter of private law as the basic principle for the
assignment of tax attributes and the adoption of risk-based legislative overrides of
the core principle—is its use as a proxy for taxpayer purpose in the identification
of tax-avoidance transactions.71 Risk exposure may defensibly be used as a proxy if
68 See, in this respect, 2015 Budget Plan, supra note 3, at 463 (“From a tax policy perspective, a
case can be made that a shareholder should always be required to bear the risk of loss and enjoy
the opportunity for gain or profit on a Canadian share in order to take advantage of the
inter-corporate dividend deduction on dividends received on that share”).
69 See, for example, Daniel Shaviro, “Risk-Based Rules and the Taxation of Capital Income” (1995)
50:4 Tax Law Review 643-724 (arguing that the source of different tax treatments and associated
avoidance opportunities is a reliance on perceptions regarding risk and the identification of
differences in legal rights and obligations as accurate indicators of changes in risk).
70 See Kleinbard, supra note 61, at 786-87 (noting that the fungibility of publicly traded securities
means that there can be many long and short positions but only one owner).
71 Edgar, supra note 7, at 348-51. See also, for example, the preamble to prop. Treas. reg. section
1.871-15, supra note 26, at 841 (“When dividends paid on physical securities are subject to
tax while dividend equivalents with respect to economically comparable derivatives are not,
those derivatives have a potential for tax avoidance regardless of whether a long party is using
the derivative in a particular case to avoid tax. Accordingly, the Treasury Department and the
it is considered intolerably costly, in terms of both compliance and administrative
burden, to inquire directly into a taxpayer’s purpose with respect to a position in
shares, given the frequency of hedged and derivative positions. As a proxy, risk exposure is assumed to be an accurate indicator of purpose that can be observed at a
tolerable cost. As an initial proposition, the taxpayer who owns shares as a matter of
private law is assigned the associated tax attributes because it is assumed that the
taxpayer bears all risk exposure, and this assumption supports the conclusion that
non-tax factors drive the decision to retain the shares or dispose of them. Risk-based
rules override this basic assignment principle, because it is suspected that a range of
hedge and derivative strategies are driven by the attempt to separate risk and
ownership in order either to assign tax attributes to a taxpayer other than the owner
or to retain ownership without attracting a disposition for income tax purposes.72
In the use of a proxy for taxpayer purpose, the set of risk-based rules in the Act is,
in fact, part of a family of specific anti-avoidance rules overriding share ownership
as the basis for the assignment of tax attributes. This family of overrides is required
because an unqualified reliance on the private-law concept of ownership as the basis
for the assignment of tax attributes is readily open to manipulation. The general
transactional pattern of the relevant tax-planning techniques is familiar to tax practitioners, policy makers, and tax administrators. Moreover, implementing any of
these arrangements is relatively costless, and it is therefore unsurprising that specific legislative responses are used to override the assignment of tax attributes that
would otherwise follow ownership based on the private-law concept. It is also unsurprising that these legislative overrides have developed on a piecemeal basis in
response to particular transactions that are seen to be tax-driven in the sense of a
perceived manipulation of ownership as the basis for the assignment of tax attributes.
In addition to the series of risk-based rules, which override ownership to assign or
deny the availability of a tax attribute on the basis of perceptions of exposure to
risk,73 other examples of overrides affecting share ownership include
n
the anti-income-splitting rules that apply to certain transfers of property, to
assign income from the property to the taxpayer who transferred it to the
owner;74
IRS [Internal Revenue Service] favor a delta approach that objectively identifies transactions in
which the long party is able to sufficiently approximate the economic return associated with an
underlying security”).
72 See, for example, the preamble to prop. Treas. reg. section 1.871-15, ibid., at 842 (suggesting
that a long derivative position in shares with a delta of 0.7 or greater sufficiently approximates
the economic returns associated with ownership of shares to be subject to non-resident
withholding tax on cross-border dividend equivalent payments).
73 See the discussion below under the heading “Specifying Risk Exposure Differently for
Different Overrides.”
74 Sections 74.1 through 74.5. See also subsections 56(2) and (4) through (5), and sections 75 and
120.4.
n
n
n
(2015) 63:2
the constructive share ownership rules that assign voting power to a taxpayer
other than the owner;75
the anti-surplus-stripping rules that override the paid-up capital result that
would otherwise occur on a sale of shares between certain persons that do not
deal at arm’s length;76 and
the back-to-back loan provision in the thin capitalization rules that overrides
the interposition of an arm’s-length intermediary.77
Thompson and Weisbach do not address the design details of the various legislative overrides in us tax law that are intended to alter the assignment of tax attributes
otherwise made on the basis of the private-law concept of ownership as the default
rule.78 They assume that the different policy contexts implicated by these legislative
overrides dictate that there will be no commonality in design. As a minimal proposition, their observation regarding the inevitable heterogeneity of legislative overrides
is evident in the different fact patterns that are the subject of the categories of such
overrides in the Act noted above. As suggested here, however, the set of rules within
each category reflects a similar approach to the execution of the identification function in a tax-avoidance context: that is, these rules avoid an inquiry into taxpayer
purpose as the primary mechanism for identifying prohibited transactions and instead use proxies for such purpose as refined by limited exclusions for circumstances
that can be assumed to be non-tax-driven.79 This legislative pattern reflects a more
generalized policy choice regarding the expression of legislative overrides of ownership as detailed rules rather than standards.
Form of Legislative Overrides: Rules Versus Standards
The dfa and the sda legislation, as well as the sea proposals, reflect a rules-based
approach characteristic of specific anti-avoidance rules. As Finance appeared to
suggest in the 2013 budget and again in the 2015 budget, reliance on the standard
75 Paragraph 251(5)(b) and subsection 256(1.4). See also subsection 256(5.1) (extending the
concept of corporate control to include “control in fact” [de facto control] where the phrase
“controlled directly or indirectly in any manner whatever” is used in the Act). See Couzin,
supra note 38, at 328-32 (contrasting de jure and de facto control); and Brian M. Studniberg,
“The Concept of De Facto Control in Canadian Tax Law: Taber Solids and Beyond” (2013)
54:1 Canadian Business Law Journal 17-37 (reviewing case law and CRA administrative positions
considering when influence is sufficient to constitute control in fact).
76 Sections 84.1 and 212.1.
78 Thompson and Weisbach, supra note 18, at 290-94, provide some detailed discussion of the
possibility of a unified legislative regime overriding ownership with respect to securities
lending and repo transactions.
79 The anti-income-splitting rules and the surplus-stripping rules applicable on a non-arm’slength sale of shares both use the relationship between the transferor and the transferee as a
proxy for taxpayer purpose. The constructive share ownership rules use the description of
arrangements with the potential to separate share ownership and voting power as a proxy for
expressed in gaar could, in principle at least, realize the same result. In fact, the
rationale for the choice to use a detailed rule rather than a generalized standard to
override ownership in the assignment of disposition treatment, gain/loss characterization, and the provision of dividend tax relief may be found in the wider debate in
the legal literature over the choice of rules versus standards as expressions of the
law.80 Although the arguments emphasized in this broader debate are not articulated
in any detail in the 2013 and 2015 budgets, they arguably underlie Finance’s stated
preference for the adoption of the dfa and sda legislation and the sea proposals
consistent with the use of targeted risk-based and other legislative overrides of
ownership to assign particular tax attributes.81
A rules-based approach to the articulation of the content of the law attempts to
specify in advance the legal consequences of behaviour or transactions; it is an approach to law that tries “to make most or nearly all legal judgments under the
governing legal provision in advance of actual cases.”82 A standards-based approach
involves the expression of a broad principle, with the legal consequences of behaviour or transactions being determined after the fact; it leaves the resolution of legal
consequences to an ad hoc, case-by-case analysis that is based on the application
of the broad principle or principles. As Kaplow argues,83 detailed legislative rules
are the preferred means, from an efficiency perspective, to describe the legal consequences of such transactions, because the one-time promulgation costs are likely to
be less than the costs associated with the process of defining, after the fact, the details of the application of broad standards to a range of such transactions. And as
Gergen notes,84 standards are to be preferred over detailed rules as an effective
means to address areas of the law that are in a “state of flux” since, in these circumstances, legislators do not have sufficient foresight to anticipate all conceivable cases
that must be addressed by the system of rules.85 Unanticipated transactions are
taxpayer purpose. These legislative overrides also refine the proxies through the exclusion of
specified circumstances in which it is assumed that tax consequences are not a motivating factor.
80 See Louis Kaplow, “Rules Versus Standards: An Economic Analysis” (1992) 42:3 Duke Law
Journal 557-629.
81 2013 Budget Plan, supra note 1, at 341 and 343.
82 Cass R. Sunstein, “Problems with Rules” (1995) 83:4 California Law Review 953-1026, at 961,
quoted in Mark P. Gergen, “Afterword: Apocalypse Not?” (1995) 50:4 Tax Law Review 833-59,
at 856.
83 Kaplow, supra note 80.
84 Gergen, supra note 82, at 856-57.
85 See John Avery Jones, “Tax Law: Rules or Principles?” (1996) 17:3 Fiscal Studies 63-89, at 64-65
(arguing that more detail does not necessarily provide certainty because it is not possible to
anticipate all circumstances that may need to be addressed by legislated rules). See also John
Prebble, “Should Tax Legislation Be Written from a Principles and Purpose Point of View or
a Precise and Detailed Point of View?” [1998] no. 2 British Tax Review 112-23; and David P.
Hariton, “The Tax Treatment of Hedged Positions in Stock: What Hath Technical Analysis
Wrought?” (1995) 50:4 Tax Law Review 803-27.
(2015) 63:2
more efficiently addressed ex post as they arise, with their resolution being an outcome of reasoning by analogy to the supposed paradigm transactions that can be
described in general principles-based legislation.86
The wider debate in the legal literature over the choice of rules or standards as
expressions of the content of the law is reflected in a longstanding theme in the tax
literature concerning the form of legislation as detailed rules or general principles.
In this literature, tax law is often criticized for an overemphasis on rules and the
attendant compliance and administrative costs. By comparison, the expression of tax
legislation in the form of general principles tends to be seen favourably,87 with desirable simplicity features that reduce compliance and administrative costs. This tax
literature often misses the point, however, in the choice between rules and standards,
which is not a choice over the content of the tax law but rather a choice regarding
the process for the articulation of the content of the tax law.88 Where legislation is
expressed at a level of general principle, the detailed content of the law must be
worked out either in specific regulations or through administrative practice and the
decisions of the courts. The use of detailed legislation or general principles is ultimately a debate over matters of process, which involves a judgment regarding the
relative advantages and disadvantages of an apparent certainty of application and
complexity of expression associated with the former approach and the apparent
simplicity of expression and uncertainty of application associated with the latter
approach.
Where a range of common transactions are affected, the expression of the tax
law as a standard is costly for two principal reasons that are attributable to the required articulation of the transactional content of the tax law ex post.89 First, the
judicial process does not lend itself easily to the efficient resolution of a large volume of specific cases. In particular, it takes considerable time for an issue to work its
86 See, in this respect, Alice G. Abreu and Richard K. Greenstein, “The Rule of Law as a Law of
Standards: Interpreting the Internal Revenue Code,” Duke Law Journal Online (forthcoming)
(arguing that the dominance of a rules-based expression of the tax law does not require an
approach to interpretation of all provisions as rules, and that many IRC provisions should be
interpreted as standards to allow the IRS and the courts sufficient flexibility to respond to
“unforeseen cases as they arise”).
87 But see Lawrence Zelenak, “Custom and the Rule of Law in the Administration of the Income
Tax” (2012) 62:3 Duke Law Journal 829-55 (arguing that a rules-based system and a limited
interpretive approach that regards provisions as self-contained rules most effectively promote
rule-of-law values).
88 Kaplow, supra note 80 (emphasizing that the content of the law remains the same, with the
important policy choice being centred on matters of process—that is, whether to use detailed
legislative rules that determine legal results ex ante or to use general standards that require a
determination of the details of the law after actions are taken).
89 See, for example, Wei Cui, “Taxing Indirect Transfers: Improving an Instrument for Stemming
Tax and Legal Base Erosion” (2014) 33:4 Virginia Tax Review 653-700, at 672-81 (discussing
the use of a rules-based versus a standards-based response to ensure source-country taxation of
capital gains realized by non-residents through indirect share transfers).
way through the audit, assessment, and appeals processes, which must be followed
before reaching formalistic and time-consuming court procedures. Without some
adjustment of these aspects of dispute resolution, it is costly to rely on the courts for
the development of a framework that can be designed to address the relevant policy
issues. In the absence of detailed rules, the judicial process would likely have to be
supplemented, in practice, with an even greater reliance on administrative rulings90
and the judgment of tax practitioners to resolve specific cases in a timely manner.91
A second source of costliness of application is expectations of judicial performance,
particularly where the range of common transactions involves tax-avoidance transactions. The past performance of the courts in Canada (and many other countries)
makes it difficult to have confidence in the judicial ability to produce consistent
results. The source of much of this inconsistency, at least in an avoidance context,
is the factual determination of taxpayer purpose where a purpose-based standard is
specified to identify prohibited transactions. The inconsistency is compounded
where the standard also involves an inquiry into legislative purpose as an overlay of
taxpayer purpose such as that found in subsection 245(4). The focus on statutory
interpretation as an identification tool has arguably made gaar underinclusive (or
at least somewhat inconsistent in its application). The underinclusiveness (or inconsistency of application) arises because of the perception that there is a range of
acceptable tax-avoidance behaviour (other than that expressly provided for in taxexpenditure programs), and this behaviour can be identified by examining the wording
of the relevant legislation. Although unclear, this premise may be the reason why
Canadian courts, in particular, are inclined to determine tax consequences on the
basis of private-law relationships irrespective of perceptions of the associated economic substance, with the frequent result that particular rules trump the expression
of an anti-avoidance provision in the form of a standard. Given this judicial environment, it is not especially difficult to see why Finance would prefer enactment of specific
legislation targeted to common transactions when the range of such transactions
includes tax-driven and non-tax-driven transactions that must be distinguished in
order to apply suitably different tax consequences. Under these conditions, there is
no reason to wait for lengthy administrative and judicial resolution of matters that
can be resolved effectively and immediately through detailed legislation.
90 See Gergen, supra note 82, at 857 (suggesting in the US context that a reliance on the rulings
process as the principal resolution mechanism to resolve the treatment of new financial
instruments generally would require significant changes, including the use of retroactive
rulings and a caveat on all rulings that they are “conditional on an instrument not later
exhibiting undesirable properties”).
91 Ibid., at 857-58. See also United States, Government Accountability Office, Financial
Derivatives: Disparate Tax Treatment and Information Gaps Create Uncertainty and Potential Abuse,
GAO-11-750 (Washington, DC: GAO, September 2011) (herein referred to as “US GAO
report”), at 35 (noting that, in the absence of guidance from the IRS, taxpayers may attempt to
take positions that may be abusive).
(2015) 63:2
In fact, the structure of gaar, with its purpose-based inquiry overlaid with an
exercise in statutory interpretation, incorporates two general types of costly standards that should be invoked solely as a response to uncommon or unanticipated
transactions, the tax treatment of which cannot, by definition, be specified ex ante.92
One type of standard is comparable to that in the non-tax law; it involves the interpretation and application of general wording to particular fact patterns most often,
but not always, in a non-avoidance context. For example, the concepts of residence,
employment, and an adventure or concern in the nature of trade are not articulated
legislatively in any comprehensive detail, leaving the content of these concepts, with
attendant tax consequences, to be determined ex post.93 The inquiry in these instances is one of determining the meaning of the general expression of the relevant
standard in the context of particular transactions or behaviour; it is characterized by
legal uncertainty ex ante that is clarified ex post. The other type of standard is used
in a tax-avoidance context characterized by the motivation for a transaction. In
identifying prohibited tax-avoidance transactions, that motivation is commonly
expressed legislatively as a “purpose requirement”: that is, a potentially prohibited
transaction is one that is undertaken primarily to access a tax benefit. Determining
the transactional content of the tax law in these instances requires an inquiry into the
motivation of a particular transaction; it is characterized by factual uncertainty that
is clarified ex post. Costliness of the identification function is associated equally,
however, with both of these types of standards, because they both leave the articulation of their transactional content to the tax administration and the courts. An
element of costliness may even be unavoidable, since the choice to use rules in the
tax law is rarely one that involves a complete rejection of the use of a standard and
is probably more accurately framed as a question of the extent to which the latter is
incorporated in a rule to most effectively execute the underlying policy decision.
Where the tax consequences for a range of common transactions in a nonavoidance context must be specified, costliness in the application of the tax law is at
least mitigated by the expression of the law as detailed rules that describe the relevant set of common transactions and thereby minimize the interpretive exercise in
determining the transactional content of the law. In an avoidance context, costliness
of the identification of prohibited transactions can be mitigated further if observable proxies for taxpayer purpose can be identified.94 In other words, proxies must
92 See, in this respect, David A. Weisbach, “Formalism in the Tax Law” (1999) 66:3 University of
Chicago Law Review 860-86 (emphasizing that a strictly rules-based system would be more
complex and costly than a system that combined rules and standards).
93 The fundamental concept of income is itself defined at the margins as a standard. See Alice G.
Abreu and Richard K. Greenstein, “It’s Not a Rule: A Better Way To Understand the
Definition of Income” (2012) 13:3 Florida Tax Review 101-32; and Alice G. Abreu and Richard
K. Greenstein, “Defining Income” (2011) 11:5 Florida Tax Review 295-348.
94 See, for example, Canada, Department of Finance, Explanatory Notes Relating to the Income Tax
Act, the Excise Tax Act and the Income Tax Regulations (Ottawa: Department of Finance, September
2013), at 62 (stating that the synthetic disposition rules are based on the effects of an arrangement,
be easily observable and credibly accurate for a range of transactions that are common in the sense that they can be anticipated by the policy maker as avoidance
techniques. As in the non-avoidance context, the set of anticipated transactions
should be specified in as much detail as possible to avoid the interpretive exercise
that a more general expression of the tax law requires. Under these conditions, the
anti-avoidance rule can ideally be applied ex ante at lower cost and without affecting
the set of common transactions that are non-tax-driven.95
It is argued in the next part of the article that a weakness of both the dfa and the
sda legislation, as well as the sea proposals, is the use of a standard to specify the level
of risk exposure that will attract their application. This is a weakness because it means
that the legislation suffers from the same problems as a more generalized standard
in the tax law; yet the reasons for choosing a standard over a rule do not necessarily
support expression of the level of risk exposure as a standard. By providing a wider
choice of levels of risk exposure, a more precise specification as a percentage amount
expands the potential to more finely distinguish between hedged and derivative positions in shares that can be considered tax-driven and those that can be considered
non-tax-driven. This particular proposition can also be extended to other risk-based
rules in the Act that override the assignment of tax attributes on the basis of the
private-law concept of ownership, with the idea being that the content of these
other rules can be expressed with much more commonality in an effort to realize a
greater level of target-effectiveness at tolerable compliance and administrative cost.
Sp e c i f i c at i o n o f R i s k E x p o s u r e
This part of the article, together with the next two parts, articulates a legislative
template for specific anti-avoidance rules that are risk-based overrides of ownership
as the basis for assigning tax attributes to hedged and derivative positions in shares.
The legislative template consists of the following three elements:
with the relevance of taxpayer purpose being limited to the determination of whether an
offsetting position entered into by a non-arm’s-length person should be accounted for as part
of a hedged position of a taxpayer). See also the preamble to prop. Treas. reg. section 1.871-1,
supra note 26, at 841 (“[T]he Treasury Department and the IRS believe that the delta-based
standard of the 2013 proposed regulations provides a simpler and more administrable framework
than the seven-factor test of the 2012 proposed regulations”); and Andrew Walker, “The Most
Recent Proposed Regulations Under Section 871(m): The Perfect Is the Enemy of the Good,”
Tax Matters feature (2014) 5:2 Columbia Journal of Tax Law 19-22, at 20 (suggesting that the
use of a bright-line approach based on the specification of delta to identify positions that
economically resemble share ownership “has the advantage of being more objective than a test
based on factual indicia of abuse and therefore is potentially more administrable”).
95 Recent contentiousness over the form of an anti-treaty-shopping provision as either an inquiry
into taxpayer purpose or a US-style limitation-on-benefits (LOB) provision emphasizes these
lines of argumentation for a purpose-based standard or a detailed rule with the use of proxies
for taxpayer purpose. See Canada, Department of Finance, Treaty Shopping—The Problem and
Possible Solutions (Ottawa: Department of Finance, August 2013); and Organisation for
Economic Co-operation and Development, BEPS Action 6: Preventing the Granting of Treaty
Benefits in Inappropriate Circumstances (Paris: OECD, 2014).
(2015) 63:2
1. explicit specification as a percentage amount of a level of risk exposure that
either attracts the application or avoids the application of the legislative
override;
2. use of an explicit purpose-based inquiry to support the application of the
legislative override by suppressing discontinuities created by the sharp
boundaries attributable to specification of a level of risk exposure; and
3. use of the financial concept of delta to measure the level of risk exposure associated with hedged or derivative positions in shares.
This part of the article considers the first element. The second and third elements
are considered in the next two parts, respectively.
The policy maker can tailor each of these three elements to best fit the income tax
consequences associated with the assignment of tax attributes, given the assumed
policy underlying the attribute. This kind of tailoring is illustrated using the examples of the dfa and sda legislation, as well as the dividend stop-loss rules, the
anti-foreign-tax-credit-trading rule applicable to short-term securities acquisitions,
the dra rules and the sea proposals, and the wash sale rules in the Act. It is suggested
how and why each of these types of rules can defensibly differ in their specification
of a level of risk exposure. It should be evident that use of the legislative template,
even with tailoring to fit different policy purposes, would result in a much greater
level of commonality of legislative design than is currently the case and might realize an element of economies of scale in tax administration and compliance.
Two policy criteria inform an assessment of the design details that implement the
above three elements of the suggested legislative template:
1. target-effectiveness in the use of risk as a proxy for a tax-motivated position
in shares; and
2. suppression of wasteful and costly discontinuities whereby small changes in
risk exposure result in a disproportionately different assignment of a tax
attribute.
Line Drawing and the Legislative Expression
of a Level of Risk Exposure
Specification of a level of risk exposure that attracts or avoids the application of a
legislative override of an assignment of tax attributes based on ownership performs
a targeting function that is most problematic with long positions in shares that are
partially hedged or long derivative positions that partially replicate an open position.
At one extreme, there is an unhedged or long position in shares that entails full exposure to risk for the owner. In these circumstances, tax attributes of the shares are
assigned on the basis of ownership on the apparent assumption that such exposure
precludes any attempt to separate a tax attribute from ownership. At the other extreme is a long position in shares that is fully hedged or a long derivative position that
perfectly replicates an unhedged position. In these circumstances, it can be assumed
that the hedged position or the long derivative position represents an attempt to
separate a tax attribute from ownership or, alternatively, that the position should be
treated consistently with non-ownership or ownership, respectively, because it so
closely approximates either of those positions.96 Between the two extremes of unhedged positions and fully hedged, as well as fully exposed, derivative positions are
long positions in shares that are partially hedged and long derivative positions that
partially replicate an open position in the underlying. Because these partial positions
leave the taxpayer partially exposed to the risk associated with an open position,
they require a distinction to be drawn between, on the one hand, a tax-driven attempt to manipulate tax-attribute assignment based on ownership and, on the other
hand, hedged or derivative positions that are entered into for non-tax reasons.
In the context of hedged and derivative positions in shares, the specification of a
level of risk exposure as a proxy for taxpayer purpose targets risk-based overrides of
ownership and should ideally have two properties in performing this function:
1. The level of risk exposure should be specified so as to affect only a limited
range of positions that can be assumed with some confidence to be taxmotivated. The narrow targeting of legislative overrides can address, at least
in part, the problem of revenue loss without imposing unnecessary costs on
commercially driven hedging and trading strategies.97 Any other approach
that applies beyond instances of substantial risk reduction or replication potentially imposes tax and compliance costs that can undermine the supposed
benefits attributable to non-tax-driven risk-reduction and trading strategies.
In that case, it becomes difficult to determine whether the revenue concerns
associated with the prevention of the tax-driven manipulation of ownership
justify the potentially adverse effect on non-tax-driven positions in shares.
Moreover, the practical audit difficulty associated with the application of
risk-based legislative overrides to a broad range of common transactions is
much less problematic when these overrides are limited to transactions involving offsetting positions that eliminate a substantial amount of the risk
exposure associated with a long position or that substantially replicate that
same risk exposure.
2. The level of risk exposure should be specified as a precise percentage amount
rather than by less precise verbal formulas such as “substantially,” “all or
96 See Miller and Milet, supra note 8, at 10:43 (emphasizing consistency of treatment of
disposition transactions and hedged positions within the standard under the SDA legislation as
providing greater coherence in the taxation of certain financial transactions).
97 See, for example, Alan Macnaughton and Amin Mawani, “Contributions of Employee Stock
Options to RRSPs and TFSAs: Valuation Issues and Policy Anomalies” (2008) 56:4 Canadian
Tax Journal 893-921, at 901, note 13 (noting the use of equity collars and swaps as hedging
techniques for employees with significant equity holdings in their employers). See also Alan D.
Jagolinzer, Steven R. Metsunaga, and P. Eric Yeung, “An Analysis of Insiders’ Use of Prepaid
Variable Forward Transactions” (2007) 45:5 Journal of Accounting Research 1655-79 (concluding
that evidence suggests the use of variable prepaid forward contracts [VPFCs] to, at least in
part, diversify holdings).
(2015) 63:2
substantially all,” “materially,” “principally,” “primarily,” or “a majority of.”98
Provided that the specified percentage is drawn at a level that ensures the
limited application of a risk-based override, precise specification as a percentage amount provides a legislative bright line that avoids the costs that
taxpayers might otherwise incur to ensure that non-tax-driven partial hedges
of, and derivative positions in, shares are unaffected.
As legislative standards, the verbal formulas that are commonly used to specify a
level of risk exposure are usually found in a range of provisions other than riskbased legislative overrides of ownership. As a result, administrative practice and
judicial interpretation in these other contexts will often have resulted in quantification as a percentage amount.99 The content of the standards expressed in these
verbal formulas is therefore usually well known and will have been converted to a
known quantity for the application of risk-based overrides. Nonetheless, an element
of imprecision may persist with such verbal formulas,100 and there will remain some
legal uncertainty, at least along the immediate edges of the relevant boundary that
they execute legislatively. For some of the less precise verbal formulas, such as “material” or “in a material respect,” the conversion to a percentage amount through
the interpretive process may remain ambiguous, with resulting legal uncertainty as
to where exactly the boundary lies, rather than just along the immediate edges of
a boundary.101 Resources must then be devoted to the discovery of the boundary
between partial hedges and derivative positions that are considered tax-driven manipulation of share ownership and positions that are considered non-tax-driven
strategies. This can only be done ex post as particular hedged and derivative positions
98 The explanatory notes to the definition of “derivative forward agreement” in subsection 248(1)
use “primarily” as a synonym for the “majority of ” standard relevant for sales agreements.
Explanatory Notes, supra note 94, at 58.
99 See, for example, Brian R. Carr and Duane R. Milot, “Copthorne: Series of Transactions
Revisited,” Corporate Tax Planning feature (2008) 56:1 Canadian Tax Journal 243-68, at 263
(citing case law standing for the proposition that the same term should be given the same
meaning under different provisions of the Act). See also Tamara Larre, “Misguided Inferences?
The Use of Expressio Unius To Interpret Tax Law” (2014) 51:3 Alberta Law Review 497-524, at
519-23 (reviewing case law in which the failure to use parallel drafting has not led to the
inference that different meanings were intended).
100 Miller and Milet, supra note 8, at 10:25-26 (arguing that Canadian courts have adopted a
qualitative and contextual approach to the interpretation of an “all or substantially all” standard,
but observing that the CRA’s administrative position has tended to reduce the standard to a
quantitative 90 percent or more bright line, which can be assumed to be incorporated for the
purpose of the definition of an SDA, given Finance’s presumed knowledge of that position).
101 See, for example, Wilson, supra note 48, at 9:13 (observing that the threshold standard of
“materiality” in paragraph (d) of the definition of “securities lending arrangement” in
subsection 260(1), which requires that “the lender’s risk of loss or opportunity for gain or profit
with respect to the security is not changed in any material respect,” is undefined, and that
counterparties should be careful when entering into non-standard SLAs whereby less than all
of the accrued gain or loss is retained by the lender).
challenge that boundary and as it is determined, through administrative practice and
ultimately judicial interpretation, where exactly the boundary lies in any particular
fact pattern. Perhaps more importantly, taxpayers may forgo, or alter, commercial
hedging and trading strategies to avoid the imprecision along the edges of the
boundary executed by a verbal formula. In effect, taxpayers may alter their preferred
risk exposure simply to ensure that a position is outside a particular legislative override, with resulting non-tax costs in the form of a loss of preferred exposure.
A precise specification of a level of risk exposure avoids these potentially adverse
consequences of legal uncertainty and also allows for different, and more nuanced,
tailoring of different risk-based overrides of ownership as the basis for the assignment
of tax attributes. Use of any of the limited set of familiar verbal formulas to specify
a level of risk exposure under a risk-based override of ownership means that the
boundary between partially exposed positions that are tax-driven and those that are
non-tax-driven is drawn at one of the limited set of possibilities that the interpretive
process has provided in other legislative contexts using the standard verbal formulas.
For example, “substantially” or “all or substantially all” have tended to be equated
with a quantity of 90 percent or more.102 “Principally,” “primarily,” and “majority”
have tended to be equated with a quantity in excess of 50 percent.103 Specification of
a level of risk exposure using these verbal formulas thus limits the choice of boundary between tax-driven and non-tax-driven positions in shares to two possibilities:
(1) risk exposures of 90 percent or more and risk exposures of less than 90 percent;
and (2) risk exposures in excess of 50 percent or risk exposures of 50 percent or less.
Any specification between these two levels, with a boundary somewhere between
90 percent and 50 percent, is ruled out by the interpretive process in other legislative contexts using the same verbal formulas. The content of alternative verbal
formulas, such as “material” or “material respect,” may remain ambiguous and suggestive of anything between, for example, 50 percent and 90 percent and even well
below 50 percent.104 Given its ambiguous content, this alternative may be rejected
102 See Miller and Milet, supra note 8, at 10:25-26.
103 See, for example, Miller and Milet, ibid., at 10:8 (stating that “majority” is generally assumed to
mean more than 50 percent). See also Richard Marcovitz and Chris Van Loan, “Amendments
to the Rules Governing Securities Held by Financial Institutions and Other Recent
Developments,” in the 2009 Conference Report, supra note 8, 10:1-40, at 10:11 (noting that
the CRA has taken the administrative position that “primarily” means more than 50 percent,
and this meaning presumably extends to the definition of “tracking property” subject to
mark-to-market treatment for a financial institution where the particular property derives its
value primarily from underlying property that would be subject to mark-to-market recognition
if held directly).
104 Apparently, any agreement that the shares are to be transferred back by the recipient at their
value at the time of the initial transfer is regarded as a sufficient indication of risk retention by
someone else. See Canada, Department of Finance, “Draft Legislation Concerning Securities
Lending and Dividend Rental Arrangements Released,” News Release no. 89-042, April 26,
1989 and accompanying “Securities Lending: Explanatory Notes.” See also CRA document
no. 9511155, October 18, 1995.
(2015) 63:2
because of the potentially adverse effect on non-tax-driven positions in shares. If the
policy maker decides that a point somewhere in between 50 percent and 90 percent
would, in fact, draw the boundary between tax-driven and non-tax-driven positions
in the most target-effective manner for the assignment of a particular tax attribute,
the most cost-effective means to implement this policy choice is by specifying the
level of risk exposure as a percentage amount.
In assigning tax attributes, risk-based overrides of share ownership elevate risk
exposure to an independent normative significance that, in fact, determines the
point at which significantly different tax treatment occurs. Yet as a proxy for a taxavoidance purpose motivating a position in shares, there is no obvious point at
which a boundary should be drawn in the absence of systematic empirical evidence.
Drawing the line relatively tightly to require a substantial reduction or assumption
of risk exposure as a condition of application for a risk-based override attempts to
avoid any adverse effect on commercially driven hedging and trading strategies, but
this policy prescription is somewhat coarse-grained and does not suggest in any
obvious manner where exactly a boundary should be drawn. Indeed, the lack of
any obvious point at which risk exposure serves as an accurate proxy for taxpayer
purpose means that the line-drawing exercise with risk-based overrides suffers from
an inevitable element of arbitrariness, even where a more precise specification of risk
exposure is otherwise desirable because the policy maker has some rough empirical
confidence that verbal formulas are intolerably overinclusive or underinclusive. For
example, there does not appear to be anything normatively compelling that would
suggest drawing a boundary at a level of risk exposure of 90 percent versus 95 percent or, alternatively, 90 percent versus 85 percent, let alone deciding between 1 or
2 percentage point differences in such specification. On the other hand, an element
of arbitrariness at this level is arguably not a sufficient reason to reject a precise
specification of risk exposure as a percentage amount where an alternative verbal
formula leaves a wider range of tax-driven positions unaffected or affects a wide
range of non-tax-driven positions. Under these circumstances, the policy maker
may defensibly specify a level of risk exposure somewhere between the two coarsegrained alternatives provided by the verbal formulas on offer to improve the targeting
of a risk-based override in a broad directional sense, even though the choice, at the
margin, will have an inevitable arbitrariness.
Specifying Risk Exposure Differently for Different Overrides
The legislative overrides of ownership in the Act affecting the assignment of tax
attributes to hedged and derivative positions in shares are somewhat eclectic in their
design details. Most importantly for the purpose of this article, they do not all incorporate risk exposure explicitly as a proxy for targeted tax-avoidance transactions,
and even where they do so, they differ in how that proxy is specified.105 It is suggested
105 See, in this respect, Alex Raskolnikov, “Relational Tax Planning Under Risk-Based Rules”
(2008) 156:5 University of Pennsylvania Law Review 1181-1262 (distinguishing risk-based rules
below that these risk-based legislative overrides of ownership may be designed with
broadly similar patterns in the specification of levels of risk exposure, although they
can also be tailored differently in their targeting. What follows are some suggested
specifications of risk exposure; they are not offered as hard policy prescriptions but
rather as illustrative possibilities of the direction in which targeting might be improved by precisely specifying a level of risk exposure as a percentage amount for
some of the different categories of legislative overrides in the Act. The illustrative
specifications also highlight why the specified level of risk exposure might be different for the different overrides, and how those overrides would differ from their
existing legislative form in the Act.
Synthetic Dispositions
The sda legislation is explicitly risk-based through the definition of a “synthetic
disposition arrangement,” which requires the elimination of all or substantially all
of the taxpayer’s risk of loss and opportunity for gain or profit in respect of a particular property.106 In the expression of a level of risk reduction that is substantial,
the legislation is broadly consistent with the us constructive sale rule107 and the
comparable anti-synthetic rule proposed in Australia.108 This level of risk reduction
is strongly suggestive of a tax-avoidance motivation for a hedged position in an
appreciated asset and is deemed to constitute a disposition. In other words, risk
reduction of this level is unlikely to be undertaken primarily for non-tax reasons,
that are explicit in specifying a required level of exposure to risk and risk-based rules that are
implicit in requiring a level of exposure to market risk in the form of asset value changes or
business risk in the form of a level of ownership).
106 Paragraph (b), definition of “synthetic disposition arrangement” in subsection 248(1).
Paragraph (c) extends the risk-reduction inquiry to include any arrangement entered into by
related parties where the purpose of the arrangement is the requisite reduction associated with
an otherwise open position.
107 IRC section 1259. Under the original legislative proposal, a constructive sale was defined
generally as any transaction that “substantially eliminates the risk of loss and opportunity for
gain” on appreciated shares, debt, or a partnership interest held by a taxpayer. This general
definition was replaced in the final legislation with a set of four enumerated transactions
consisting of a long position in an appreciated asset and an offsetting position or positions. See
the discussion in the text below at notes 168-178.
108 ITAA 1997, proposed section 230-348C(1) (defining “effective disposal arrangements” as one
or more arrangements that have the effect of (1) transferring to another person, or relieving
the taxpayer of, the risk of all or a substantial proportion of the negative aspects of the subject
rights and/or obligations; and (2) transferring to another person, or depriving the taxpayer of,
the opportunity to obtain the benefit of all or a substantial proportion of the positive aspects
of the subject rights and/or obligations). ITAA 1997, proposed section 230-348A (providing
that the object of subdivision 230-J is to minimize tax deferral and arbitrage that would
otherwise occur through arrangements that would allow circumstances to be brought into
existence that would have substantially the same effect as, but do not take the form of,
circumstances that give rise to assessable gains from financial arrangements).
(2015) 63:2
and it can be seen to serve as a credibly accurate proxy for a tax-avoidance purpose
in the vast majority of cases; alternatively, even if non-tax-motivated, risk reduction
to this extent can be considered sufficiently close to a disposition in its non-tax
features that it should be treated consistently with a transaction that is considered
to be a disposition under accepted notions of ownership.
By limiting the application of a risk-based override to a narrow range of partially
hedged positions in shares, an “all or substantially all” standard is, nonetheless,
potentially underinclusive. An uncertain range of partial hedges providing risk reduction that is significant, but not necessarily within the standard of “all or substantially
all,” are unaffected, even though a not insignificant number of these hedged positions may be tax-driven. The policy maker could use a more nuanced approach to
specification of the level of risk exposure than a coarse-grained “all or substantially
all” standard to address potential underinclusiveness while still avoiding problems
of overinclusiveness. In particular, a problematic range of partial hedges could be
isolated by providing taxpayers with a safe harbour beginning at a lower bound of
risk reduction. Only risk-reduction levels below this lower bound would categorically avoid disposition treatment, on the assumption that the vast majority of partial
hedges at these lower levels can credibly be considered non-tax-driven and should be
unaffected. In this respect, some illustrative guidance may be found in the Australian
anti-dividend-credit-trading legislation, which draws the boundary between taxdriven and non-tax-driven hedges at a risk-reduction level of 70 percent.109 Although
designed for the purpose of assigning a different tax attribute, this 70 percent specification could be incorporated under a synthetic disposition rule to provide that
partial hedges reducing risk exposure by 70 percent or less would not attract treatment as a disposition for what would seem to be the same rationale: it may be the
case that a broad range of partial hedges at this level of risk reduction are undertaken primarily for non-tax reasons and should be unaffected by these different
risk-based overrides.
With synthetic dispositions, a problematic range of partial hedges could also be
defined as those that reduce risk exposure in an amount greater than 70 percent
and less than 90 percent, assuming that partial hedges reducing risk exposure by
90 percent or more should categorically be subject to disposition treatment. A not
insubstantial number of partial hedges reducing risk exposure to shares within this
problematic band may be motivated by tax considerations; yet, there may be a not
insubstantial number of such hedges within the same band of risk reduction that
might be undertaken primarily for non-tax reasons. Partially hedged positions in
shares within this problematic range could be sorted into tax-driven and non-taxdriven characterizations by an inquiry into taxpayer purpose. The inquiry could be
unspecified in terms of what factors are considered relevant, or it could be directed
by the policy maker through the creation of a rebuttable presumption that a partially
109 See infra notes 203-207 and the accompanying text.
hedged position within the problematic range is either tax-driven or non-tax-driven.110
With the former, relevant evidence would be required to rebut the presumption and
recharacterize the hedge as non-tax-driven; with the latter, relevant evidence would
be required to rebut the presumption and recharacterize the transaction as taxdriven. The choice of presumption would be a function of an empirical judgment
that the majority of partially hedged positions within the problematic range are
tax-driven or are not.
An inquiry into taxpayer purpose as an additional targeting feature of a synthetic
disposition rule would increase costliness of application and should probably be
tolerated only if the policy maker is confident in an empirical judgment that there
is a not insubstantial range of partially hedged positions that otherwise would be
mischaracterized using a stand-alone “all or substantially all”—that is, 90 percent or
more—risk-reduction requirement. To mitigate the administrative and compliance
costliness of the additional identification exercise, the policy maker could indicate
legislatively a set of factors that may be considered relevant.111 The lack of systematic empirical evidence of the motivation for hedged positions within a defined
problematic range means, however, that the policy maker must rely on unsatisfactory anecdotal evidence as the basis for this important targeting refinement.
Dividend Stop-Loss Rules and Foreign Tax Credit Trading
The dividend stop-loss rules address the familiar avoidance opportunity that arises
where a corporate taxpayer acquires shares shortly before or after the dividend date
and sells them shortly thereafter.112 In those circumstances, the amount paid for the
shares includes the amount of the declared or expected dividend, and the subsequent receipt of the dividend that compensates the corporate shareholder for a
portion of the cost of the shares. Although the dividend effectively represents a recovery of a portion of the cost, a tax benefit arises if the corporate shareholder can
claim an intercorporate dividend deduction for the amount of the cost recovery as
well as a separate loss on the sale of the shares equal to the amount of the dividend.
The anti-foreign-tax-credit-trading rule applicable to short-term acquisitions of
110 See, for example, in the US context, Mike Farber, “Section 871(m) and Delta: When Should a
Dividend Equivalent Be Treated Like a Dividend?” Tax Matters feature (2014) 5:2 Columbia
Journal of Tax Law 12-14, at 14 (suggesting that the proposed dividend equivalent payment
regulations should establish a presumption that a long derivative position with “high-delta
exposure to dividend-paying stock” is subject to non-resident withholding tax on dividend
equivalent payments, but the presumption could be rebutted “by showing either material
transactions costs or that the counterparty does not own the underlying in connection with
the transaction”).
111 For example, in the Australian context, ITAA 1997, proposed sections 230-348D(5), 230348Q(5), and 230-348X(3) (listing factors to be considered in determining taxpayer purpose).
See also the anti-treaty-shopping proposal in the 2014 federal budget: Canada, Department of
Finance, 2014 Budget, Budget Plan, February 11, 2014, at 351-52.
112 Subsections 112(3) through (5.6).
(2015) 63:2
securities113 addresses transactions in which foreign tax credits are effectively traded
between taxpayers by transferring ownership of shares (or other securities) as the basis
for the assignment of credit entitlement. Before the enactment of the sda legislation,
neither of these anti-avoidance rules was explicitly risk-based. Instead, both relied
on a required one-year hold period as a proxy for a tax-avoidance purpose, on the
apparent assumption that the associated risk exposure credibly indicates a non-tax
motivation for an acquisition. This assumption could be based, in turn, on an empirical assumption that the transaction costs of implementing and maintaining a hedge
over the hold period in order to separate ownership and economic exposure in a
tax-driven context would be prohibitive.114 With the enactment of the sda legislation,
consequential amendments to the dividend stop-loss rules and the anti-foreign-taxcredit-trading rule incorporate an additional at-risk requirement by excluding from
the hold period those days during which a synthetic disposition was in place.115
Incorporation of the at-risk standard in the sda legislation arguably improves
the targeting of these anti-avoidance rules, at least to the extent that hedge transaction costs do not in fact constrain entirely all tax-motivated share acquisitions.
However, by affecting only a narrow range of partially hedged positions that are
likely tax-driven (or are sufficiently close to non-ownership that ownership should
be ignored in assigning the relevant tax attributes), the improvement in targeting
may be slight. A wide range of partially hedged positions remain unaffected, even
though some may credibly be assumed to be associated with share acquisitions intended to generate a loss using the intercorporate dividend deduction or to transfer
entitlement to a foreign tax credit. At the other extreme of overinclusiveness, the us
legislation comparable to the dividend stop-loss rules denies the intercorporate
dividend deduction unless the relevant shares are held “naked”116 for at least 46
days.117 More particularly, share ownership for the purposes of the hold period is
ignored for any day during which the corporate shareholder (1) has an option to sell
(“covered put”), (2) is under a contractual obligation to sell (“covered call”), or
(3) has made (and not closed) a short sale of substantially identical shares or securities.118 The apparent empirical assumption underlying this specification is that any
level of risk reduction is tax-driven. The Australian anti-dividend-credit-trading
113 Subsection 126(4.2). Foreign tax credit trading first surfaced in the United States and prompted
a joint response by Treasury and the IRS in Notice 98-5, 1998-3 IRB 49. The Department of
Finance followed with a comparable response in Canada, Department of Finance, 1998 Budget,
Budget Plan, February 24, 1998, at 41-45, and resolutions 46 and 47 of the accompanying
Notice of Ways and Means Motion To Amend the Income Tax Act.
114 The significance of transaction costs and other non-tax constraints on tax-motivated hedged
and derivative positions in shares for the design of risk-based overrides more generally is
discussed further in the text below at notes 152-163.
115 Subsections 112(8) and 126(4.5).
116 Hariton, supra note 85, at 813.
117 IRC section 246(c).
118 IRC section 246(c)(1)(A).
rules lie in between the two extreme specifications of the us rule on the one hand
and the Canadian dividend stop-loss rules and anti-foreign-tax-credit trading rule
on the other, by specifying a level of acceptable risk reduction of 70 percent or
less.119 A comparable compromise specification of a permitted level of risk reduction
would obviously expand the application of these two risk-based overrides in the Act.
Although this more expansive override for the purpose of the dividend stop-loss
rules and the anti-foreign-tax-credit trading rule applicable to short-term securities
acquisitions would almost certainly be overinclusive in its characterization of a not
insubstantial number of partially hedged positions that may in fact be non-taxdriven, overinclusiveness may be tolerated because the tax consequences of the
classification of an expansive range of partial hedges as tax-driven would be mitigated by the hold period requirement. In other words, risk reduction in excess of
70 percent would not necessarily disentitle an owner of shares to dividend tax relief
or foreign tax credits but would result in the exclusion of those days during which
the relevant position was hedged from satisfaction of the hold period requirement.
A synthetic disposition rule, in contrast, would result in the deemed disposition of
an appreciated asset. Because of this all-or-nothing tax consequence, an additional
inquiry into the motivation of those partial hedges, outside a specified safe-harbour
level of risk reduction but less than that which would be categorically considered
tax-driven, may warrant the additional administrative and compliance costs associated with an inquiry into taxpayer purpose.120 Costliness of the same inquiry across
the entire range of problematic partially hedged positions may not necessarily be
warranted for the purpose of those overrides imposing an at-risk requirement for
shareholding purposes. In fact, incorporation of an at-risk requirement for hold period purposes could permit a reduction in the length of the period used as a proxy for
tax-driven hedging.121 Where no such requirement is incorporated, or it is narrowly
119 As originally introduced, the Australian anti-dividend-credit-trading proposals provided that
any reduction of the risk of loss below 70 percent of the value of an exposed long position in
the relevant shares would result in tolling of the hold period for the number of days during
which this level of risk reduction was exceeded. Following significant criticism emphasizing the
adverse effect of the proposals on share-hedging strategies, the Australian Treasury refined the
specification of risk exposure to require maintenance of the current minimum net equity
exposure (that is, both the opportunity for profit and risk of loss) of 30 percent. See Australia,
Department of the Treasury, “Measures To Prevent Trading in Franking Credits,” Treasurer’s
Press Release no. 47, May 13, 1997; and “Budget Measures To Prevent Trading in Franking
Credits: Outcome of Consultations,” Treasurer’s Press Release no. 89, August 8, 1997.
120 See the discussion in the text above under the heading “Synthetic Dispositions.”
121 See, for example, Australia, Review of Business Taxation, A Tax System Redesigned—More
Certain, Equitable and Durable (Canberra: Review of Business Taxation, July 1999), at 227-50
(recommending that the 45-day rule be replaced with a more limited 15-day rule, and that an
exemption threshold for small transactions be increased from A $2,000 to A $5,000). The
incorporation of an at-risk requirement to enhance the target effectiveness of a hold period
requirement as a proxy for taxpayer purpose would not necessarily affect other targeting
features, such as the shareholding condition in the dividend stop-loss rules of 5 percent or less
of the affected class of shares, which overlay risk exposure.
(2015) 63:2
focused, hold periods may be extended in length as a coarse-grained means to further screen out days during which a position in shares is hedged, but with associated
administrative and compliance costs. The policy maker can defensibly trade off
expansiveness of an at-risk requirement and the length of a hold period in refining
the targeting of the relevant risk-based override, while attempting to keep administrative and compliance costs roughly constant.
Dividend Rental Arrangements and
Synthetic Equity Arrangements
The dra legislation, denying either an intercorporate dividend deduction122 or a
dividend tax credit123 in respect of a dividend received as part of a “dividend rental
arrangement,” incorporates the combination of a purpose-based standard and a
risk-exposure requirement. As currently conceived, the definition of a dra is expressly
purpose-based in the sense that the main reason for entering into the arrangement
must be the receipt of a dividend in respect of the relevant shares.124 The arrangement must also be structured such that someone other than the recipient bears the
risk of loss or enjoys the opportunity for gain or profit with respect to the shares “in
any material respect.”125 By incorporating both of these standards, the dra legislation also incorporates maximum costliness of application ex post. The requirement
for a purpose-based inquiry in addition to a level of risk reduction is presumably
included because of the especially vague expression of the latter, which could potentially affect non-tax-driven hedging.126 As well, potential overinclusiveness may be
122 Subsection 112(2.3).
123 Subparagraphs 82(1)(a)(i) and (a.1)(i), and paragraphs 82(1)(b) and (c).
124 Subsection 248(1), definition of “dividend rental arrangement.” One interpretive position in
support of the inapplicability of the DRA rules to the dividend equivalent payment transactions
that are the target of the SEA proposals characterizes the requisite arrangement as the entering
into of an offsetting short derivative position by a bank owning the shares, in which case either
(1) the prior acquisition of the shares is not part of the arrangement or (2) the arrangement is
not for the main purpose of receiving dividends on the shares.
125 Ibid., at subparagraph (a)(ii). See also supra note 104. Another interpretive position in support
of the inapplicability of the DRA rules characterizes the equity receiver as not enjoying
opportunity for gain or bearing risk of loss in respect of the shares held by the equity payer
where the equity derivative is cash-settled rather than physically settled.
126 The proposed anti-synthetic rules in Australia similarly combine the specification of a level of
risk exposure and an inquiry into taxpayer purpose. See supra note 14, “Exposure Draft.” The
need for the latter requirement seems more debatable in the Australian context than it does
under the DRA rules, since the required level of risk exposure under the Australian proposals is
specified using verbal formulas that are arguably less ambiguous than a “material” standard and
that suggest a level affecting a narrow range of hedged or long derivative positions, which could
be assumed to exclude the vast majority of non-tax-driven positions. Moreover, the proposed
rules provide that a tax-driven purpose need not be the dominant purpose of a partially hedged
or derivative position and thus limit even further the category of excluded transactions.
attributable to a specification of risk exposure that is altered in terms of either the
downside or the upside rather than a material reduction of both.
The 2015 budget proposals extend the definition of a dra to include a “synthetic
equity arrangement,” which is defined in part to mean one or more agreements or
other arrangements that have the effect of providing to a counterparty all or substantially all of the risk of loss and opportunity for gain or profit in respect of a share.127
The proposed definition incorporates the same expression of risk reduction as that
used in the definition of an sda and raises the same issues already noted in relation
to that articulation of a required level of risk exposure.128 In addition, the proposed
definition of an sea attempts to refine the use of risk as a proxy for taxpayer purpose
with a set of exceptions from sea characterization even in the presence of what is a
substantial level of risk reduction. In particular, the effect of sea characterization is
avoided where a taxpayer’s counterparty is not a tax-exempt or non-resident person
(referred to as a “tax-indifferent investor”) or where the counterparty otherwise
hedges its position; moreover, characterization as an sea is avoided if the synthetic
position is constructed through a “recognized derivatives exchange.”129 In each of
these fact patterns, in contrast to the exclusively risk-based targeting of the sda
legislation, the fact that the specified level of risk reduction is inconsistent with an
assumption of risk exposure and ownership is seen to be insufficient to attract the
denial of dividend tax relief in favour of a more thoroughgoing identification of
taxpayer purpose through these specified additional proxies.130
With the elaborate definition of a dra, and particularly the proposed addition of
the definition of an sea, costliness of application is unavoidable—a concern that is
acknowledged in the budget documents and in the stated intention of the Department of Finance to engage in consultation with a view to determining whether a
simpler, yet still target-effective, response can be designed.131 One possibility is to
follow the approach suggested here and redesign the definition of a dra to incorporate an exclusively risk-based approach. More particularly, the existing risk-based
standard, which requires that someone other than the shareholder bear the risk of
loss or opportunity for gain or profit in “any material respect,” as well as the “all or
127 Notice of ways and means motion, supra note 3, at resolution 32(1), amending the definition of
“dividend rental arrangement” in subsection 248(1), and resolution 32(2), adding to that
subsection the definition of “synthetic equity arrangement.”
128 See supra notes 106-111 and the accompanying text.
129 Notice of ways and means motion, supra note 3, at resolutions 32(1) and (2).
130 2015 Budget Plan, supra note 3, at 463. See, for example, Joint Committee on Taxation of the
Canadian Bar Association and Chartered Professional Accountants of Canada, “2015 Federal
Budget—Synthetic Equity Arrangements: Submission by the Joint Committee on Taxation,”
submission to the Department of Finance, May 26, 2015 (highlighting the potential
overinclusiveness of the SEA proposals in the context of commercial share acquisitions
incorporating provision for pre-sale dividends).
131 Ibid.
(2015) 63:2
substantially all” standard incorporated in the proposed definition of an sea, could
be replaced with the precise specification of a level of risk exposure as a percentage
amount comparable to that suggested above for the dividend stop-loss rules and the
anti-foreign-tax-credit-trading rule applicable to short-term securities acquisitions.
Risk exposure should also be specified fully in terms of both the downside and the
upside associated with a position as proposed in the definition of an sea. The additional condition in the existing dra legislation that the main purpose for entering
into the particular arrangement was to enable receipt of a dividend could be eliminated and replaced with an at-risk hold period requirement. Indeed, use of a specified
level of risk reduction and a hold period requirement in the dra legislation would
provide clear rules for a set of reasonably foreseeable share acquisitions that would not
necessarily be any less target-effective in distinguishing tax-driven from non-taxdriven share acquisitions. Moreover, by framing the required inquiries as an exercise
in parallel drafting, each of the risk-based legislative overrides stipulating minimum
at-risk hold periods could be applied in much the same way, thereby realizing economies of scale in the tax administrative function and perhaps even the tax compliance
function.
Character Conversion Transactions—
Derivative Forward Agreements
The dfa legislation is conceptually related to certain other, more narrowly focused,
overrides of ownership that address the separation of ownership and risk exposure
to avoid a specified character of property that is assigned as an attribute of ownership. For example, the concept of “tracking property” addresses the attempt to
avoid mark-to-market treatment for certain shares and specified debt obligations
held by financial institutions using long derivative positions, or other intermediate
entities, not otherwise subject to mark-to-market recognition.132 A similar tracking
approach is used to address the attempt to avoid foreign accrual property income
(fapi) status on the insurance of Canadian risks by entering into swap arrangements
132 Subsection 142.2(1), definition of “tracking property.” See Canada, Department of Finance,
Legislative Proposals and Explanatory Notes Relating to the Income Tax Act, the Excise Tax Act, 2001
and the Excise Tax Act (Ottawa: Department of Finance, July 2008), at clause 26 (stating that the
definition of “tracking property” is intended to “ensure that a financial institution will not be
able to avoid mark-to-market treatment on properties by investing through an intermediary or
through the use of another financial instrument (such as a derivative)”). The proposed, but
never enacted, rules requiring inclusion of investment income from foreign investment entities
(FIEs) included a similar concept of a “tracking entity,” which would not otherwise be a FIE
but which derived entitlement to payment primarily from enumerated criteria with respect to a
participating interest in a FIE. See Canada, Department of Finance, “Revised Legislative
Proposals Released Relating to the Taxation of Non-Resident Trusts and Foreign Investment
Entities and Other Technical Amendments to the Income Tax Act,” News Release no. 2005-049,
July 18, 2005, proposed subsection 94.2(1), definition of “tracking entity,” and proposed
subsection 94.2(9).
of a foreign policy pool for a pool of policies that includes the insurance of Canadian risks.133 These overrides have presumably been seen to be necessary in light of
the longstanding cra administrative position that long derivative positions in foreign property constitute property separate from the underlying risk exposure and
would not be considered foreign property for the purpose of the former restrictions
on foreign property holdings by registered plans and certain other taxpayers if the
counterparty was a Canadian-resident issuer.134
The dfa legislation is a more generalized override of the character of gain (or
loss) that would otherwise be associated with ownership of capital property, including shares, treating such amounts realized on a disposition as non-capital amounts
where the gain or loss is linked through derivative instruments to amounts, such as
interest, that would be considered non-capital.135 For example, in the absence of the
dfa legislation, a taxpayer with a hedged position in shares (or other property)
could take the position that the amount of gain locked in by the hedge is on capital
account when realized on a disposition, even though the gain could be considered
equivalent to interest or other non-capital amount. Similarly, a taxpayer with a long
derivative position in shares or other capital property could take the position that
an acquisition and subsequent sale of a variable amount of shares or such property
is on capital account, even though the purchase price (and the number) of the shares
or other capital property is determined by reference to changes in value of an
underlying that would be on non-capital account if owned by the taxpayer.
Hedged positions in shares (or other capital property) are potentially “sales agreements” within the definition of a dfa. This element of the definition is explicitly
risk-based, requiring that the agreement eliminate a majority of the taxpayer’s risk of
loss and opportunity for gain or profit in respect of the particular property.136 Expressed using a verbal formula, this level of risk reduction is readily converted to a
precise specification of more than 50 percent, much like the accepted interpretation
of the “all or substantially all” standard in the sda legislation (90 percent or more).
As it applies to hedged positions, the dfa legislation is notable for its incorporation
of this relatively low level of risk reduction in combination with the determination of
the sale price of the particular capital property with reference to an underlying
interest. Other character conversion rules, such as the conversion transaction rule in
133 See 2013 Budget Plan, supra note 1, at 341-43, and resolution 35 of the accompanying notice
of ways and means motion.
134 See Miller and Milet, supra note 8, at 10:3 (observing that forward agreements similar to those
used in DFAs were used to avoid the former foreign property holding restrictions and were
approved as doing so by the CRA); and Marcovitz and Van Loan, supra note 103, at 10:10
(observing that the CRA has indicated that equity derivatives are considered property for the
purpose of the definition of “tracking property” in subsection 142.2(1), consistent with its
administrative position under the former foreign property holding restrictions).
135 Paragraphs 12(1)(z.7) and 20(1)(xx).
136 Paragraph (c), definition of “derivative forward agreement” in subsection 248(1).
(2015) 63:2
the United States,137 as well as a proposed anti-synthetic rule for interest equivalents
in Australia,138 require a much higher level of risk reduction through the requirement
that a long position in an asset be hedged such that any realized gain substantially
replicates a return equivalent to interest.
As a targeting mechanism, a relatively high level of required risk reduction for
the application of a character conversion rule is consistent with the attempt to avoid
adversely affecting non-tax-driven hedging. The dfa level of 50 percent or greater
shifts the boundary between tax-driven and non-tax-driven hedges so as to ensure
that the vast majority of hedges in the former category are subject to the legislative
override but potentially affects a significant range of hedges in the latter category.
Shifting the boundary in this way to avoid underinclusiveness, while tolerating
overinclusiveness, may reflect an intolerance for character conversion as a manipulation of ownership in the assignment of the character of gain or loss and may be a
function of a muted policy enthusiasm for maintenance of the capital/income distinction when faced with the structured transactions that motivated the dfa legislation
(mutual funds using the capital gain election in subsection 39(4) for their hedged
positions in shares). Alternatively, Finance may have assumed that an acceptable
level of risk reduction must be specified using the limited set of available verbal
formulas. Constraining specification in this way arguably leaves a choice between a
“majority” standard and an “all or substantially all” standard, since a standard of
“materiality,” such as that under the dra rules, may be ruled out as being too uncertain from a legal perspective. An “all or substantially all” standard would be
consistent with the specification of the required risk reduction under the sda legislation but might be considered too permissive, and therefore underinclusive, in the
case of character conversion using sales agreements. A mixed characterization rule
that would apportion capital and non-capital treatment based on relative exposures
to capital and non-capital assets can presumably be rejected (or perhaps more likely,
not even considered) as being too administrative- and compliance-intensive.
Long derivative positions in shares (or other capital property) are potentially
“purchase agreements” within the definition of a dfa. Unlike “sales agreements,”
this element of the definition is not risk-based, either explicitly or implicitly; instead,
the difference between the fair market value of property delivered on settlement of
137 IRC sections 1258(c)(1) and (c)(2)(A) (defining a “conversion transaction” to include an
acquisition of property covered by a substantially contemporaneous contract to sell the same or
a substantially identical property where the covered position results in substantially all of the
expected return being attributable to the time value of the taxpayer’s net investment).
138 ITAA 1997, proposed section 230-348W(1) (defining an “effective gain arrangement” to
include one or more arrangements that provide a gain that is sufficiently certain); and proposed
section 230-348T (providing that the object of subdivision 230-JB is to minimize tax deferral
and arbitrage that would otherwise occur through arrangements that would allow circumstances
to be brought into existence that would have substantially the same effect as, but do not take
the form of, circumstances that give rise to gains that are sufficiently certain and to which the
accruals method would apply).
the agreement and the amount paid for the property attracts the application of the
legislation if the difference is attributable to an underlying interest.139 The apparent
assumption is that all long derivative positions in shares (or other capital property)
are tax-driven if structured such that the amount to be delivered is variable. The
failure to specify any risk-exposure requirement as a targeting mechanism in the case
of purchase agreements under the dfa legislation may again reflect a dissatisfaction
with the limited set of alternative specifications using standard verbal formulas and
a confidence that the legislative description of these agreements without a specified
level of risk exposure is suitably target-effective. In this respect, the approach to the
characterization of a long derivative position as a “purchase agreement” differs from
the more restrictive character conversion rule in the United States applicable to
derivative positions in hedge funds, which applies to recharacterize gain or loss on
a position as a non-capital amount if substantially all of risk of loss and opportunity
for gain is attributable to returns of the fund.140 The approach to long derivative
positions under the dfa legislation also differs from the override for the purpose of
mark-to-market treatment, which applies to require mark-to-market recognition if
tracking property derives its fair market value primarily from a set of enumerated
criteria that includes factors relevant to risk of loss and opportunity for gain or profit,
as well as extending to revenue, income, cash flow, and any other similar criteria.141
A similar approach to the characterization of a long derivative position in shares
(or other capital property) as a “purchase agreement” subject to gain/loss recharacterization would include only those positions that provide risk exposure that is
substantially invariant to the underlying non-capital property. Alternatively, a long
derivative position in shares (or other capital property) could be excluded from recharacterization if, for example, a “majority” of the risk or “substantially all” of the
risk was attributable to changes in the value of the shares. However, a level of risk
exposure to the subject shares or other capital property consistent with the “majority” standard used in the case of “sales agreements” under the dfa legislation could
be considered too permissive, and therefore underinclusive, in allowing relatively
easy avoidance of the character conversion rule. Inclusion of only those long derivative positions that closely track a fixed non-capital return described by an “all or
substantially all” standard would be potentially even more underinclusive, such that
relatively modest alteration of risk exposure to the underlying could avoid recharacterization. On the other hand, an “all or substantially all” level of risk exposure to
the underlying shares may be seen to be so narrow as an exclusion that the character
conversion rule remains overinclusive, with the narrowly drawn exclusion adding
very little of substance as a targeting mechanism. Possible overinclusiveness resulting
139 Paragraph (b), definition of “derivative forward agreement” in subsection 248(1).
140 IRC section 1260. As with the constructive sale rule in IRC section 1259, an “all or
substantially all” standard is not explicitly specified but is instead implied by the description of
the enumerated derivative positions. See the discussion in the text below at notes 168-170.
141 Subsection 142.2(1), definition of “tracking property.”
(2015) 63:2
from a rejection of these verbal formulas to define a range of excluded long derivative positions is presumably tolerated to avoid underinclusiveness.142
The uniform approach to risk-based overrides of share ownership suggested
here could be used as a more fine-grained and potentially more effective targeting
of the dfa legislation applicable to both hedged positions (“sales agreements”) and
long derivative positions (“purchase agreements”). It could draw the boundary between tax-driven and non-tax-driven hedging in a character conversion context at a
higher level of risk reduction—say, the same illustrative 70 percent incorporated in
other overrides—than the majority standard used in the case of sales agreements.
Similarly, a long derivative position in shares could be excluded from gain/loss recharacterization if, following the example of the proposed dividend equivalent payment
regulations in the United States, the risk exposure is 70 percent or greater than that
associated with ownership of the shares.143 Moreover, as in the sda legislation, the
all-or-nothing consequences of the application of the dfa legislation could be
muted, to some extent, using precise specification of risk exposure as a percentage
amount to create three categories of hedged positions (sales agreements) and long
derivative positions (purchase agreements) measured in terms of risk exposure to
shares or other capital property:
1. less than 30 percent exposure (“sales agreements”)/less than 70 percent exposure (“purchase agreements”)—subject to gain/loss recharacterization;
2. greater than 50 percent exposure (“sales agreements”)/greater than 90 percent
exposure (“purchase agreements”)—not subject to gain/loss recharacterization; and
142 See, in this respect, New York State Bar Association, Tax Section, Report on Proposed Regulations
Under Section 871(m), report no. 1306 (Albany, NY: NYSBA, May 20, 2014), at 25 (noting that
most US practitioners use a delta standard of 0.8 in applying the constructive ownership rule in
IRC section 1260 to determine whether all of the economic return with respect to a financial
asset has been conveyed through a forward contract).
143 Prop. Treas. reg. sections 1.871-15(d)(2) and (e). See, in this respect, NYSBA report no. 1306,
supra note 142, at 4 (suggesting that a delta threshold of 0.7 is too low and a threshold of at
least 0.8 should be used); and Walker, supra note 94, at 20-21 (acknowledging that derivative
positions that are imperfectly correlated with a long position in shares should be subject to the
proposed dividend equivalent payment regulations but suggesting that a delta threshold of 0.8
would strike a better balance between tax-driven and non-tax-driven derivative positions). See
also the preamble to prop. Treas. reg. section 1.871-15, supra note 26, at 839 (noting comments
requesting that final regulations limit the scope of dividend equivalent treatment to derivative
positions that provide delta 1 or near delta 1 exposure to a long position in shares, but also
noting comments arguing that such an approach would provide non-delta 1 derivative positions
with a competitive advantage over delta 1 positions because of inconsistent tax treatment); and
New York State Bar Association, Tax Section, Report on Proposed and Temporary Regulations
Under Section 871(m), report no. 1264 (Albany, NY: NYSBA, April 25, 2012), at 8-9 (observing
that a majority of the members of the Tax Section believe that the dividend equivalent payment
regulations should be limited to arrangements that are substantially equivalent to delta 1
exposure or are part of larger arrangements that provide that exposure).
3. exposure between 30 percent and 50 percent (“sales agreements”)/exposure
between 70 percent and 90 percent (“purchase agreements”)—subject to a
further inquiry into taxpayer purpose as the basis for recharacterization.
Here again, the policy maker would need to determine that precise specification
improves the target-effectiveness of the risk-based override, at least in a general
directional sense; moreover, it would need to be determined that the additional
costliness associated with a purpose-based inquiry for hedged and long derivative
positions within a problematic range of risk exposure is similarly warranted by improved target-effectiveness.
Wash Sales
In addition to the dividend stop-loss rules, a generalized set of stop-loss rules in the
Act applies to “wash sales.”144 These rules reflect a similar legislative template and
apply to deny the recognition of a capital or a non-capital loss where a taxpayer
disposes of property, including shares, but maintains the same risk exposure to the
property through an acquisition of the same (or an identical) property by the taxpayer (or an “affiliated” person) within a 61-day window around the date of the
disposition (30 days before and 30 days after).145 Although not explicitly risk-based,
this set of overrides of ownership in assigning loss recognition is implicitly riskbased as a proxy for taxpayer purpose. In particular, specification of a 61-day window
as a condition for denial of loss recognition appears to be based on an empirical
assumption that asset price volatility suggests that risk exposure associated with an
acquisition outside the window is sufficiently different from the extreme case of a
contemporaneous disposition and acquisition that the transactions can be characterized as non-tax-driven. A contemporaneous disposition and acquisition is the extreme
case in the sense that there is a disposition as a matter of private law, but the identical risk exposure otherwise associated with ownership is maintained. In the absence
of a legislative override, a loss would be recognized for income tax purposes based
on ownership as a matter of private law as the basis for loss recognition. The 61-day
window under the set of generalized stop-loss rules can thus be seen to serve as a
proxy for risk exposure that serves as a proxy for taxpayer purpose, much like a hold
144 The term “wash sale” is used to describe transactions that are subject to the generalized set of
stop-loss rules in the Act, which in turn are referred to as “the wash sale rules” to distinguish
them from the dividend stop-loss rules. These terms are more commonly used in the US
context to refer to the comparable rules in IRC section 1091 and the transactions subject to
these US stop-loss rules for capital losses.
145 With the exception of capital losses realized by an individual, the denied loss is recognized by
the taxpayer subsequently when the property is disposed of to a person who is not affiliated
with the taxpayer. “Superficial losses” are added to the adjusted cost base of the transferred
property and recognized subsequently by the transferee on a disposition to a person who is not
affiliated with the transferee. See paragraph 53(1)(f ).
(2015) 63:2
period requirement under the dividend stop-loss rules and the anti-foreign-taxcredit-trading rules applicable to short-term securities acquisitions.
The wash sale rules in the Act are potentially both underinclusive and overinclusive. They are underinclusive to the extent that the price of an asset, particularly a
non-traded capital asset, is not especially volatile, so that an acquisition outside the
61-day window may still closely approximate the risk exposure of a contemporaneous sale and acquisition, but is treated as if it does not. The rules are overinclusive
to the extent that the price of an asset, such as a publicly traded share, is especially
volatile, so that an acquisition within the 61-day window may not approximate the
risk exposure of a contemporaneous sale and acquisition, but is treated as if it does.
Underinclusiveness may also be attributable to a failure to extend the rules to the
entire range of long derivative positions that can be used to maintain the risk exposure associated with ownership of an asset within the 61-day window.146 In this respect,
the wash sale rules extend to a “right to acquire property,” which is deemed to constitute an identical property.147 This extension appears to include simple derivative
positions such as physically settled call options and forward contracts; it is much less
clear whether it can be interpreted to extend to other long derivative positions that
are cash settled, including a total-return swap.148
Recasting the wash sale rules as an override of ownership that is explicitly riskbased could potentially improve their target-effectiveness. For example, the rules
could be recast along the lines suggested immediately above for “purchase agreements” under the definition of a dfa, to override ownership and deny the recognition
of a loss wherever
n
n
the taxpayer (or an “affiliated person”) enters into a long derivative position
in respect of a disposed of asset at, or just before or after, the time of disposition;149 and
the risk exposure of the long derivative position was, say, 70 percent or
greater than the risk exposure associated with ownership of the asset.
146 See, for example, Schizer, “Scrubbing the Wash Sale Rules,” supra note 24, at 71-76
(describing strategies to avoid the application of IRC section 1091).
147 Clause 13(21.2)(e)(iii)(A); subsection 14(13); subsection 18(16); paragraph 40(3.5)(a); and
paragraph (i), definition of “superficial loss” in section 54.
148 Schizer, “Scrubbing the Wash Sale Rules,” supra note 24, at 72-73 (discussing the possible
failure of IRC section 1091 to apply to a swap as a replacement property providing the same
risk exposure as a long position in property that is sold).
149 In the Australian context, see, for example, ITAA 1997, proposed section 230-348P (defining
“effective retention arrangement” but proposing a risk-exposure standard of “all, or a substantial
proportion”). In the US context, see Schizer, “Scrubbing the Wash Sale Rules,” supra note 24,
at 78 (arguing that IRC section 1091 should be extended to dispositions of call options replaced
by a long position in the underlying shares and short positions in shares replaced by put
options written on the shares).
As a proxy for taxpayer purpose, targeting of the override would depend entirely
on the specification of a level of risk exposure that could be seen to be sufficiently
close to a contemporaneous sale and acquisition to warrant loss denial.150 However,
legal uncertainty would be associated with a vaguely specified temporal connection
between a disposition and entering into a long derivative position.151 An alternative
would be to maintain a specified temporal window, as in the existing wash sale rules,
accepting the target-ineffectiveness that is attributable to it. As with the illustrative
risk-based specification of both the sda and the dfa legislation, the targeting of the
wash sale rules could be refined further by creating a range of problematic risk exposures (between 70 percent and 90 percent) that would warrant an inquiry into taxpayer
purpose along with a safe harbour (less than 70 percent risk exposure) and categorical
loss denial (90 percent or greater risk exposure) for those positions that are not considered to be as problematic in their characterization as non-tax-driven or tax-driven.
S u pp r e s s i n g D i s co n t i n u i t i e s U s i n g
a P u r p o s e - Ba s e d S ta n d a r d
Precise specification of a level of risk exposure as the basis for the application of an
override of ownership in the assignment of tax attributes provides bright-line
boundaries that result in sharp discontinuities whereby a small change in the risk
exposure associated with a hedged or long derivative position in shares can avoid
application of the risk-based override. This part of the article develops more fully
the basis for the proposition suggested above that a direct inquiry into taxpayer
purpose mandated by a targeted purpose-based standard is the preferable instrument to suppress discontinuities along the boundary between what are considered
tax-driven and non-tax-driven positions. In particular, a purpose-based standard is
preferable to the use of legal uncertainty to serve this function, while costliness of
application is potentially mitigated somewhat by a focus on a limited set of positions
that are clustered around a particular boundary.
Discontinuities as a Focus of Tax-Driven Substitution
Discontinuities are focal points for planning throughout the tax law and are a familiar problem for tax policy makers and tax administrators;152 they are associated
150 But see Schizer, “Scrubbing the Wash Sale Rules,” supra note 24, at 76-79 (arguing for
symmetrical treatment of positions under the constructive sale rule in IRC section 1259 and
the wash sale rules in IRC section 1091, on the basis that the former permits relatively easy
avoidance and deferral of gain recognition while the latter should eliminate loss-recognition
strategies rather than induce the adoption of more sophisticated and wasteful planning).
151 Australia’s proposed stop-loss rules impose a purpose requirement that presumably is intended to
screen out those acquisitions that may be sufficiently close temporally to a disposition but that are
not motivated by the recognition of a loss for income tax purposes while maintaining the risk
exposure associated with ownership of the asset. See ITAA 1997, proposed section 230-348P(e).
152 David A. Weisbach, “Line Drawing, Doctrine, and Efficiency in the Tax Law” (1999) 84:6
Cornell Law Review 1627-81 (arguing that the lack of a clear normative basis for the drawing of
(2015) 63:2
with the drawing of boundaries where small changes in transactional form can
produce disproportionate differences in tax treatment defined by the particular
boundary. As a focus of tax planning, discontinuities effectively provide taxpayers
with a choice of transactional form. On one side of a boundary is a higher-taxed
transaction; on the other side is a lower-taxed transaction with similar non-tax features. Substitution of the lower-taxed transaction for the higher-taxed transaction
occurs where the cost of sacrificing desirable non-tax features of the higher-taxed
transaction is less than the tax savings provided by the substitution.153 Because it
involves a sacrifice of desirable non-tax features associated with the higher-taxed
transaction, substitution of the lower-taxed transaction is imperfect; it results in
efficiency loss equal to the value of the desirable non-tax features that must be sacrificed to access the tax saving associated with the lower-taxed transaction. Tax
planning focused on discontinuities is thus problematic not only for the revenue
loss; it is also wasteful because it entails efficiency loss in addition to the transaction
costs of implementing a lower-taxed substitute. In the absence of constraints on
tax-driven substitution along a boundary in the tax law, the lower-taxed transaction
will be substituted for the higher-taxed transaction until, at the margin, the loss in
desirable non-tax features equals the available tax saving.154
The risk-based overrides of share ownership that are the focus of this article attempt to address the substitution of a transaction by which share ownership as a
matter of private law and risk exposure otherwise associated with ownership are
separated in order to
n
n
n
n
avoid treatment as a disposition,
recognize a loss,
transfer dividend tax relief or foreign tax credits, or
convert the character of a gain from non-capital to capital.
These tax-driven transactions are themselves lower-taxed substitutes for highertaxed transactions (or a non-transaction in the case of a wash sale) that otherwise
leave ownership and full exposure to risk with the same taxpayer (without loss recognition in the case of a wash sale) or eliminate such exposure on a disposition. In
drawing a distinction between what are assumed to be tax-driven and non-taxdriven positions in shares, this category of legislative overrides of ownership draws
boundaries at any particular point means that they should ideally be drawn on the basis of
efficiency considerations, with a transaction being taxed consistently with its closest substitute,
but a transaction should not be taxed too severely if there are other substitutes that taxpayers can
access with associated efficiency costs). See also Jeff Strnad, “Taxing New Financial Products:
A Conceptual Framework” (1994) 46:3 Stanford Law Review 569-605 (emphasizing the
significance of discontinuities under a non-linear system of taxation of financial instruments).
153 See Myron S. Scholes, Mark A. Wolfson, Merle Erickson, Michelle Hanlon, Edward L.
Maydew, and Terry Shevlin, Taxes and Business Strategy: A Planning Approach, 5th ed. (Upper
Saddle River, NJ: Prentice Hall, 2014).
154 Tim Edgar, “Building a Better GAAR” (2008) 27:4 Virginia Tax Review 833-905.
a line in the tax law that itself becomes the focus of tax planning, rather than the
transactional choice that would otherwise be available in the absence of the override. In effect, the boundary in the tax law becomes different, but the dynamic is the
same. To realize the same desired tax treatment, taxpayers are required by a riskbased override to retain a specified level of risk exposure or, in the case of a long
derivative position, to avoid a specified level of risk exposure.
One possible policy option to suppress discontinuities is elimination of a particular difference in tax treatment and its defining boundary that gives rise to tax-driven
substitution. With the assignment of tax attributes to hedged and derivative positions in shares, this result could be realized by partial allocation based on relative
risk exposure. Such an approach would be continuous in the sense that it would
eliminate disproportionate “jumps” in tax treatment attributable to small changes in
risk exposure. However, perceived administrative and compliance intensity arguably
precludes this policy option, which necessitates the use of risk-based overrides of
ownership and their associated discontinuities attributable to an all-or-nothing allocation of tax attributes contingent on the effective characterization of a hedged or
derivative position as tax-driven or non-tax-driven.155 On the assumption that elimination of these relevant boundaries is not an option, the policy maker is left with a
choice among a limited set of instruments to suppress discontinuities created by
risk-based overrides. Not surprisingly, given the pervasiveness of boundaries in the
tax law, the availability of these instruments is not unique to this type of legislation.
A potentially effective alternative approach to elimination of discontinuities is
the definition of a boundary at a point where the policy maker can have confidence
that differences in non-tax features completely constrain substitution of a lowertaxed transaction for a higher-taxed transaction. Referred to in the literature as
“frictions,”156 these non-tax constraints can be found in various sources, including
non-tax regulatory restrictions and financial accounting treatment; they also include
constraints that are market-based in the sense that a transactional form required to
155 But see Schizer, “Frictions as a Constraint,” supra note 24, at 1364-68 (discussing a partial
recognition approach to gain realization under a constructive sale rule applied to partially
hedged positions in shares, but rejecting it as requiring a delta-based measure that is too
administrative- and compliance-intensive and that can realize results that are inconsistent with
an equivalent partial disposition of a block of shares). But see also United States Senate
Committee on Finance, Revenue Reconciliation Act of 1997 (as Reported by the Committee on Finance),
S rep. no. 105-33, 105th Cong., 1st sess. (1997), at 124 (indicating a preference for a pro rata
approach to constructive sale treatment for partially hedged positions in appreciated assets).
156 See Gergen, supra note 82, at 834 (emphasizing the role of transaction costs, credit risk, and
legal risk as constraints on tax-driven financial contracting); and Schizer, “Frictions as a
Constraint,” supra note 24, at 1326-34 (reviewing business preferences, state of technology and
markets, agency costs, credit risk, financial accounting treatment, and regulatory requirements
as constraints on tax planning). See also Victor Fleischer, “Regulatory Arbitrage” (2010) 89:2
Texas Law Review 227-89 (extending the concept of transactional substitution to non-tax areas
of law in order to avoid regulatory constraints and discussing legal, business, professional,
ethical, and political factors that can constrain substitution).
(2015) 63:2
access a tax treatment is not supplied by the market, or the cost of its supply is
prohibitive.157 Whatever the source of a particular friction, the result is the same:
although the substitution of a lower-taxed for a higher-taxed transactional form
along a particular boundary is available in principle, it cannot be implemented in
practice because of the binding nature of the non-tax constraint. A number of considerations, however, undermine a reliance on non-tax factors to constrain tax-driven
substitution. For example, informational barriers can undermine the ability of the
policy maker to identify non-tax constraints.158 Even where non-tax constraints can
be identified, it is often difficult to know ex ante the extent to which any constraint
is binding159 when particular taxpayers have differing degrees of tolerance for an
identified constraint. Moreover, the binding nature of identified constraints must be
monitored, since regulatory or financial accounting practice can change, while a
previous failure of the market to supply a lower-taxed substitute can change with
changing conditions, including the availability of tax-avoidance opportunities. Most
importantly perhaps, unless an identified constraint prevents the entire range of
possible substitutions along a boundary in the tax law, substitution will occur to the
point at which the constraint is binding. The result is a different equilibrium than
that which would prevail under an environment of unconstrained substitution; yet
there is absolutely nothing to suggest that the revenue and efficiency loss that occurs
up to that equilibrium point is a desirable policy result.160
When non-tax factors serve as weak constraints or are unavailable, the policy
maker must consider other instruments to suppress the discontinuities that are the
source of tax-driven substitution. In this respect, there does not appear to be much in
the way of binding non-tax constraints that could confidently be relied on to maintain
a distinction between tax-driven and non-tax-driven positions in shares under overrides of ownership that use risk exposure as a proxy for taxpayer purpose. Indeed, the
presence of relatively deep and liquid markets for traded shares and derivative instruments, in particular, as well as the availability of over-the-counter products, suggests
157 Raskolnikov, supra note 105.
158 See, in this respect, US GAO report, supra note 91, at 37-41 (emphasizing the need for more
sharing of information between the IRS and other governmental agencies, such as the
Securities and Exchange Commission and the Commodity Futures Trading Commission, as
being critical to the identification of new financial products and emerging tax schemes
requiring a response). See also, for example, Schizer, “Frictions as a Constraint,” supra note 24,
at 1335-36 (observing that a reliance on frictions to constrain tax planning requires tax policy
makers to learn a wide range of institutional detail, but noting that such detail tends to be
acquired and applied on an ad hoc basis).
159 See, for example, Schizer, “Frictions as a Constraint,” supra note 24, at 1368 (observing that
US policy makers did not anticipate that the ability of taxpayers to avoid the application of the
constructive ownership rule in IRC section 1260 would be constrained by the unavailability of
the necessary derivative instruments).
160 Schizer, ibid., at 1337 (noting that a reliance on frictions to constrain tax planning “can
redistribute tax burdens in random or undesirable ways”).
that there is very little in the way of market-based non-tax constraints, other than
transaction costs, on the creation of hedged and derivative positions in shares.161
This kind of market is especially conducive to counterparties serving as accommodation parties, since they are able to hedge out their own exposure at low cost and
earn riskless fees for their services.162 As described in the next section, experience
with the us constructive sale rule supports the proposition that the design of
hedged positions in traded shares to manipulate boundaries under risk-based rules
is, in fact, largely frictionless, at least in the absence of any obvious regulatory constraint or financial accounting constraint. In this market environment, the policy
maker may choose among a menu of legal instruments, as an alternative to a reliance on non-tax constraints, to suppress discontinuities along the boundary between
tax-driven and non-tax-driven hedged and derivative positions in shares under riskbased overrides of ownership.163
A Purpose-Based Standard as the Preferable Policy
Instrument To Suppress Discontinuities
The following three legal instruments can be used to address tax-driven substitution
along boundaries in the tax law generally, including the boundary implicated by
risk-based overrides of ownership:
1. a lack of clarity in drawing a boundary,
2. a purpose-based standard, and/or
3.a gaar and/or generalized judicial anti-avoidance doctrines.
These three instruments are not mutually exclusive and can be employed together. It is suggested here, however, that a purpose-based standard is the preferable
legal instrument to maintain the boundary between tax-driven and non-tax-driven
161 But see Farber, supra note 110, at 13-14 (suggesting that the proposed dividend equivalent
payment regulations should focus on whether a counterparty can cheaply hedge, otherwise
than by holding the underlying shares, and observing that a hedge becomes more expensive as
the delta of a derivative position falls below 1). See also NYSBA report no. 1306, supra note 142,
at 25 (arguing that parties attempting to earn dividend equivalents will not find it efficient to
do so at deltas as low as 0.7 because such transactions would not sufficiently approximate
economic returns associated with the underlying shares and would have materially higher
transaction costs than high-delta transactions).
162 See, for example, Farber, supra note 110, at 14 (speculating on whether the use of delta under
the proposed dividend equivalent payment regulations in the United States might be a proxy
identifying those circumstances in which a counterparty might be expected to hold the
underlying shares and the “expected avoidance value might be expected to exceed transaction
costs”). See also Schizer, “Frictions as a Constraint,” supra note 24 (arguing that the
constructive sale rule in IRC section 1259 and the constructive ownership rule in IRC section
1260 are both vulnerable to avoidance through relatively modest changes in economic return,
but that avoidance of the latter is uncommon because of the inability of securities dealers to
supply the necessary derivative instrument).
163 Edgar, supra note 7, at 348-50.
(2015) 63:2
hedged and derivative positions in shares under risk-based overrides of ownership.164 Lack of clarity in defining the boundary using imprecise verbal formulas
tends to be eroded through the tax administration process, undermining the effectiveness of legal uncertainty as a constraint on tax-driven substitution. Moreover, as
suggested earlier, inconsistency of application by the judiciary is a defining feature
of a gaar that is characterized by an exercise in statutory interpretation as an identification tool; it is therefore a potentially weak constraint on tax-driven substitution,
and its use should be limited to transactions that cannot be anticipated. Because
tax-driven adjustment of risk exposure to avoid the application of a risk-based override can be anticipated, reliance on a gaar to suppress such substitution should be
rejected, and an exclusively purpose-based standard should be used instead as a
potentially robust constraint.
As an initial proposition, legal uncertainty is the principal property of the expression of a level of risk exposure as a standard under risk-based overrides of
ownership,165 and the policy maker may use this imprecision as an instrument to
suppress discontinuities along the boundary between tax-driven and non-tax-driven
positions under risk-based overrides.166 In effect, the legal uncertainty attributable
to an imprecisely specified standard may be relied on to constrain tax-driven hedged
or derivative positions, since taxpayers must be willing to accept an element of legal
164 See, for example, Weisbach, supra note 92 (arguing that purpose-based anti-avoidance rules are
an effective means to suppress discontinuities that are the focus of tax-driven transactions).
165 The explanatory notes to the definition of “synthetic disposition arrangement” in subsection
248(1) state only that “whether all or substantially all of a taxpayer’s risk of loss and opportunity
for gain or profit in respect of property has been eliminated is highly factual”: Explanatory
Notes, supra note 94, at 62. To emphasize this point, the explanatory notes then suggest that
risk of loss or opportunity for gain or profit may be considered to be eliminated even though
there is some exposure to dividends, creditworthiness of a counterparty, or variability in
interest rates or currency exchange rates, depending on the circumstances. This proposition
and the enumerated factors are found in the proposed constructive sale rule in Australia: ITAA
1997, proposed section 230-348D(2).
166 The examples in the explanatory notes accompanying the definitions of “derivative forward
agreement” and “synthetic disposition arrangement” in subsection 248(1) provide little in the
way of informational content that would reduce legal uncertainty for partially hedged positions
in shares with a level of risk exposure lying close to the boundary between tax-driven and
non-tax-driven hedging specified by the verbal formulas in the legislation. The risk exposures
in these examples are either clearly within or clearly outside the relevant boundary. See
Explanatory Notes, supra note 94, at 34, 59-60, and 63-65. Possibly instructive examples are two
different equity collars in the explanatory notes accompanying the SDA definition (ibid., at 63).
But the retention of the opportunity for gain of 2 percent in one of the examples (example 2—
two-year put-call arrangement) is clearly trivial, assuming that price volatility of the underlying
is non-trivial. The other example (example 3—five-year put-call arrangement) involves retention
of 50 percent downside and upside risk, which is clearly outside the “all or substantially all”
risk-reduction standard, assuming that the price volatility of the underlying is substantial. The
importance of the assumptions regarding price volatility is discussed in the text below at notes
218-224. For a thorough review of the interpretive significance of the explanatory notes for the
SDA legislation, see Miller and Milet, supra note 8, at 10:28-33.
risk to implement a position that is tax-motivated but that may or may not be classified as non-tax-motivated.167 The empirical assumption underlying reliance on
legal uncertainty and the associated legal risk to constrain tax-driven substitution is
that such risk adversely affects the pricing of those transactions that are intended to
transfer tax attributes by separating ownership from exposure to risk or, alternatively, that legal risk adversely affects the value of the available tax benefit for a
taxpayer implementing a strategy to attract a tax attribute or to avoid a tax attribute.
To avoid these adverse effects, a taxpayer must adjust risk exposure sufficiently to
have some confidence that it will, in fact, be classified as non-tax-driven under a
risk-based override of ownership. However, legal uncertainty can serve as a binding
constraint on tax-driven substitution only if it can be assumed empirically with
some confidence that affected taxpayers are unwilling to assume the risk exposure
associated with the required adjustment of their hedged or derivative positions in
shares. Moreover, as emphasized earlier, in the discussion of the choice between detailed rules and standards in the tax law, the binding constraint that would otherwise
be created by legal uncertainty is invariably undermined by the tax administration
process, with the determination of the content of a standard unfolding ex post as
transactions are undertaken and classified on one side or the other of a boundary.
The experience with the constructive sale rule in the United States provides an
instructive illustration of this policy dynamic that can arise with an imprecisely expressed level of risk exposure.
Since 1997, the us Internal Revenue Code (irc) has included a provision addressing constructive sales.168 Unlike the sda legislation, the us constructive sale
rule does not explicitly indicate a required level of risk reduction;169 it is implied
167 See Mark P. Gergen and Paula Schmitz, “The Influence of Tax Law on Securities Innovation in
the United States: 1981-1997” (1997) 52:2 Tax Law Review 119-97 (emphasizing the significance
of tax-law uncertainty as a friction that constrains tax-driven innovation of publicly traded
securities because of the pricing effect for tax clienteles, but also observing that the effect is
weakened by administrative guidance intended to clarify tax treatment). But see also Schizer,
“Frictions as a Constraint,” supra note 24, at 1317, note 10 (noting that legal uncertainty does
not always serve as an effective constraint on tax-driven substitution and citing the issuance of
debt exchangeable into common shares [DECS] as an example).
168 IRC section 1259.
169 A short sale against the box is included only if the offsetting positions are in the same or a
substantially identical property (IRC section 1259(c)(1)(A)). An offsetting NPC (notional
principal contract) is included only if it pays all, or substantially all, of the investment yield on a
long position and provides for reimbursement of all decreases in value (IRC section 1259(c)(1)(B)).
A forward contract must cover a long position by providing for delivery of a substantially fixed
amount of property for a substantially fixed price (IRC section 1259(c)(1)(C)). A long purchase
must cover an appreciated short position in the same or a substantially identical property (IRC
section 1259(c)(1)(D)). IRC section 1259 also includes transactions that have substantially the
same effect as any of the enumerated transactions to the extent that they are prescribed by
the secretary of the Treasury (IRC section 1259(c)(1)(E)). This regulatory power has not been
exercised.
(2015) 63:2
instead in the requirement that each of the enumerated transactions subject to the
rule consist of substantially offsetting positions. This requirement has been understood by tax practitioners as implying that significant variation, taken to mean
20 percent or more, of the exposure associated with an underlying and the exposure
associated with a hedge is sufficient to avoid application of the rule. The accepted
interpretation was developed in the context of two different sets of transactions
designed to closely approximate the elimination of risk exposure associated with a
sale while avoiding such treatment.170 One set of transactions arose in the corporate
sector and involves tailoring of a hedged position through the issuance of contingent payment debt intended to monetize an appreciated position in the shares of an
unrelated corporation held by the issuer.171 Two notable examples are debt exchangeable into common shares (decs)172 and premium equity participating securities
170 See Hayward, supra note 9, at 1773-80 (reviewing different examples of monetization strategies).
171 Schizer, “Frictions as a Constraint,” supra note 24, at 1318 (suggesting that the ease with which
IRC section 1259 can be avoided was understood by government officials when the constructive
sale rule was enacted and was “tolerated for reasons of politics and administrability”). See also
Hayward, supra note 9, at 1799-1801 (noting that the original proposal for a constructive sale
rule in the United States would have resulted in disposition treatment where there was
elimination of substantially all of the risk of loss and opportunity for gain, but the replacement
of this standard with the enumerated set of transactions and the extension of regulatory authority
to other transactions with substantially the same effect have resulted in underinclusiveness);
and NYSBA report no. 1306, supra note 142, at 25-26 (observing that forward contracts
providing for delivery of an amount of cash or shares subject to significant variability are not
forward contracts for the purpose of IRC section 1259, which requires delivery of a
substantially fixed amount of property or cash for a fixed price).
172 As the equivalent of a fixed-payment debt instrument and an embedded on-market forward
contract written on a long position in shares of a corporation other than the issuer,
exchangeable debt normally eliminates all of the risk associated with that position. DECS are a
variation of exchangeable debt and are structured such that the issuer retains sufficient upside
appreciation to avoid the application of the constructive sale rule. See William M. Gentry and
David M. Schizer, “Frictions and Tax-Motivated Hedging: An Empirical Exploration of
Publicly-Traded Exchangeable Securities” (2003) 56:1, part 2 National Tax Journal 167-95, at
174 (characterizing the tax bar as being very confident that the constructive sale rule would not
apply to a partial hedge by which the taxpayer retains the first 20 percent of appreciation in the
underlying property [as in a DECS transaction], and suggesting that it is doubtful that many of
these transactions have been deterred by legal uncertainty). See also Edward D. Kleinbard and
Erika W. Nijenhuis, “Everything I Know About Financial Products I Learned from DECS,” in
Tax Law and Practice Course Handbook Series (New York: Practicing Law Institute, 2001),
1183-1234. Another version of exchangeable debt, referred to as “PHONES,” is equivalent to
the issuance of a long-term put option on referenced shares held by the issuer. In a typical
issue, the holder of a PHONES is entitled to a below-market rate of interest plus dividends
paid on the referenced shares and, on maturity, may receive the greater of the issue price of the
PHONES (equal to the spot price of the referenced shares on issue) and the spot price of the
referenced shares on maturity. Because the issuer retains the risk of loss on the referenced
shares, the constructive sale rule may not apply. See Lee Sheppard, “Rethinking DECS, and
New Ways To Carve Out Debt” (1999) 83:3 Tax Notes 347-52, at 349-50 (arguing that the
retention of the risk of loss by the issuer of PHONES is insignificant in light of their 30-year
(peps).173 Another set of comparable transactions arose with high-wealth individuals
holding appreciated positions in shares and involves the use of either zero-cost
collars174 or variable prepaid forward contracts (vpfcs).175
In Revenue ruling 2003-7,176 the Internal Revenue Service (irs) confirmed that
a vpfc with a kinked payout structure would not be subject to the constructive sale
rule because of the “significant variation” in the taxpayer’s delivery obligation. The
to 60-year terms); and Gentry and Schizer, supra, at 174 (characterizing PHONES as a more
aggressive variation of DECS and the subject of legal uncertainty, and noting that taxpayers
who are less tolerant of legal risk can use the safer DECS structure).
173 PEPS are, in form, unsecured debt instruments, issued for a price equal to the spot price of
referenced shares held by the issuer. PEPS have a limited term (commonly five years) and pay
interest at a below-market rate. On maturity, the holder is entitled to cash equal to the spot
price of the referenced shares. PEPS also provide the issuer with a redemption option during
the latter portion of the term of the security. The option is equivalent to a call option with
payment terms that permit the issuer to retain a portion of any appreciation in the referenced
shares. Like DECS, PEPS are arguably outside the constructive sale rule, because the issuer
retains a portion of the equity upside sufficient to support the conclusion that all or
substantially all of the risk exposure associated with a long position in the referenced shares has
not been eliminated. See Lee A. Sheppard, “Adding PEP to the Constructive Sale Debate”
(1996) 70:13 Tax Notes 1592-96.
174 Under a zero-cost collar, a taxpayer with a long position in appreciated shares issues a call and
purchases a put written on the shares for equal premiums (hence the “zero-cost” label). In
contrast to a married put and call, with a zero-cost collar the exercise prices of the options
differ and provide a price range within which the taxpayer is again exposed to sufficient risk to
avoid the application of the constructive sale rule. The long position in the shares can be
monetized by acquiring low-cost financing secured against the collar. See, in this respect,
Schizer, “Scrubbing the Wash Sale Rules,” supra note 24, at 77-78 (suggesting that a call
spread of 20 percent is accepted as sufficient risk exposure for a partially hedged position to
avoid the application of IRC section 1259 and arguing that comparable risk exposure should be
used as the basis for determination of a replacement position under the wash sale rules).
175VPFCs are structured to provide payouts comparable to those of zero-cost collars. VPFCs
differ from standard forward contracts in that an upfront payment is received by the taxpayer
writing the forward contract over the share position. On maturity of the contract, the taxpayer
receives no additional payment and must deliver a variable number of shares, depending on
their value at that time, or pay an amount equal to that value. The payout under the VPFC is
structured such that the taxpayer retains the value of the shares within a specified range, which
provides a kink in the payout profile (hence the label “kinked forwards” or “kinky forwards”).
176 Rev. rul. 2003-7, 2003-1 CB 363. See also Internal Revenue Service, Variable Prepaid Forward
Contracts Incorporating Share Lending Arrangements, Coordinated Issues Paper LMSB-04-1207
(Washington, DC: IRS, February 6, 2008); Internal Revenue Service, Legal Advice
Memorandum LAM 2007-004, January 24, 2007; and Internal Revenue Service, Technical
Advice Memorandum 200604033, October 20, 2005. For a discussion of the IRS’s
administrative position, see Jasper J. Nzedu, “Sale or No Sale: The Taxation of Variable
Prepaid Forward Lending Contracts Involving Share Lending Agreements” (2008) 7:2 Journal
of Taxation of Financial Products 51-55; Steven M. Rosenthal and Liz Dyor, “Prepaid Forward
Contracts and Equity Collars: Tax Traps and Opportunities” (2001) 2:1 Journal of Taxation of
Financial Products 35-44; and Lee Sheppard, “Should Share Lending Affect a Prepaid Forward
Contract?” (2006) 110:1 Tax Notes 12-19.
(2015) 63:2
taxpayer in the ruling owned 100 shares with a value of $20 per share on entering
into the vpfc, which provided that the taxpayer was to deliver on maturity, in kind
or in cash, (1) all 100 shares if the per share value was $20 or less; (2) the number of
shares having a value of $2,000 if the per share value was between $20 and $25; or
(3) 80 shares if the per share value was greater than $25. The taxpayer was thereby
protected from any subsequent decline in value below the market price of $20 per
share, retained all possible gain between $20 and $25 per share, and retained
20 percent of any gain in excess of $25 per share. The taxpayer also retained all dividend rights. In concluding that the constructive sale rule did not apply, the ruling
focuses on the extreme potential values under the delivery obligation of either 100
or 80 shares as the basis for a characterization of the vpfc as a forward contract that
leaves the taxpayer with sufficient risk exposure. The same characterization of vpfcs
is reflected in the subsequent decision in Anschutz.177 The vpfcs considered in that
case protected the taxpayer from any decline in value below the initial value, while
the taxpayer retained all gain in excess of 150 percent of that value and gave up all
gain between 100 percent and 150 percent. The taxpayer also retained all dividend
payments on the underlying shares if the share price was less than the initial value
or greater than 150 percent of that value. Where the share price was between these
two amounts, the taxpayer retained a portion of the dividend payments. Consistent
with Revenue ruling 2003-7, the court focused on the extreme potential values (in
kind or in cash) under the delivery obligation for the taxpayer of either (1) all shares
(where share value is less than the initial value) or (2) 66.66 percent of all shares (where
share value is 150 percent of the initial value).178 It was accepted that the constructive sale rule did not apply because the “ultimate delivery obligation may vary by as
much as 33.3 percent,” with the taxpayer thereby retaining sufficient risk exposure.
Also consistent with Revenue ruling 2003-7, it was held that the presence of a securities lending agreement between the taxpayer and the counterparty financial
institution, which facilitated a short sale of the shares as a hedge of the long forward
position, resulted in a sale under an indeterminate “benefits and burdens” approach
to the characterization of ownership. In the absence of a “benefits and burdens”
gloss on the private-law concept of share ownership, the cra would presumably
have to rely on gaar in section 245 to address the same kinds of transactions that
are designed to avoid application of the sda legislation.
More generally, the experience with the us constructive sale rule illustrates the
inevitable erosion of legal uncertainty and the associated legal risk that can occur with
the imprecise expression of risk exposure using verbal formulas under risk-based
overrides. The erosion occurs as the content of the standard expressed by the formula is articulated ex post in the context of particular transactions, and the resulting
177 Supra note 65.
178 For share value between 100 and 150 percent of the initial value, as well as share value in excess
of 150 percent, the number of shares under the delivery obligation was a variable amount
between the extremes of 66.67 percent and 100 percent.
clarity of the boundary between tax-driven and non-tax-driven hedging exposes the
standard to tax-driven substitution. Verbal formulas such as “all or substantially all” of
risk exposure in the sda legislation and the sea proposals, or a “majority” of risk exposure in the dfa legislation in the context of sale agreements, are especially weak as
constraints on tax-driven substitution, because they come prepackaged with content
that has been articulated in the context of other provisions. Expressing the required
level of risk reduction using a verbal formula like the “materiality” standard in the
dra rules constitutes a difference in degree, and not one of kind. Although a standard such as “materiality” does not necessarily come with the same level of existing
content as the verbal formulas “all or substantially all” and “majority,” the content of
the standard will also inevitably be revealed ex post through the tax administration
process. Indeed, as emphasized above in respect of the different types of overrides
in the Act, the principal difference between a precise specification of a required
level of risk exposure and its expression as an imprecise verbal formula remains a
matter of legal process, with specification being revealed ex post through the tax
administration process rather than ex ante as a rule.
An ultimate similarity of result under a precise specification of risk exposure as a
percentage amount and an imprecise specification using a verbal formula means that
both approaches suffer from the same vulnerability, and both approaches require a
purpose-based standard to most effectively suppress discontinuities and the taxdriven substitution that they engender along the boundary between tax-driven and
non-tax-driven positions in shares.179 A purpose-based standard is a specific antiavoidance rule that requires a direct inquiry into taxpayer purpose in determining
the application of the rule to deny the tax treatment otherwise available under the
relevant provision for a transaction, or to subject such transaction to a tax treatment
that it was intended to avoid.180 A purpose-based standard supersedes gaar and
substitutes a legally determinate, but factually uncertain, inquiry into taxpayer purpose as the exclusive means to determine the application of a particular risk-based
179 See, in this respect, Patricia A. Brown, “Policy Forum: What Makes a Dutch Company Dutch?
The Evolution of US Limitation-on-Benefits Provisions” (2014) 62:3 Canadian Tax Journal
741-52, at 742 (noting that inappropriate use of tax treaties is addressed in the United States by
a combination of LOB treaty provisions and anti-abuse doctrines developed more generally
under income tax law).
180 See, for example, 2015 Budget Plan, supra note 3, at 463, and resolution 32(2) of the
accompanying notice of ways and means motion (proposing that a series of transactions will be
deemed to be a DRA where the transactions have the effect of eliminating all or substantially
all of the taxpayer’s risk of loss and opportunity for gain or profit in respect of a share if the
purpose is to avoid characterization as a DRA). In the US context, see prop. Treas. reg. section
1.871-15(n) (permitting the commissioner to treat any payment with respect to a transaction as
a dividend equivalent payment to the extent necessary to prevent avoidance of the application of
the proposed regulations where a taxpayer, either directly or through a related person, acquires
a transaction or transactions with the principal purpose of avoiding the application of the
proposed regulations).
(2015) 63:2
override to transactions that might otherwise avoid its application.181 The factual
uncertainty concerns the determination of the purpose of a taxpayer in structuring a
transaction that accesses a tax treatment or avoids a tax treatment under the relevant
provision. Yet in contrast to the legal uncertainty associated with other imprecisely
specified standards, an inquiry into taxpayer purpose does not necessarily reveal
anything about the content of the standard as applied to another taxpayer and can
remain robust as a constraint on tax-driven substitution.182 Although there may be
some inconsistency in the articulation of this inquiry in particular fact patterns, the
inconsistency would almost certainly be much less pronounced than that under
gaar, which involves the same inquiry into taxpayer purpose as well as a misuse or
abuse analysis as an exercise in statutory interpretation. The binding constraint of a
purpose-based standard can also be enhanced by specifying that the requisite purpose is “one of the main purposes” of a transaction,183 thereby perhaps avoiding
arguments over the characterization of the relative strengths of the tax-avoidance
purpose and possible non-tax reasons for the adoption of a particular position in
shares.184 Costliness of the execution of a purpose-based standard may also be tolerable because of a limited focus on the subset of partially hedged positions in shares
and long derivative positions that provide partial risk exposure clustered at or near
the specified boundary between tax-driven and non-tax-driven positions.185
181 But see NYSBA report no. 1306, supra note 142, at 54-55 (recommending that the anti-abuse
rule under the proposed dividend equivalent payment regulations be limited by the requirements
that the relevant transaction or transactions provide the substantial economic equivalent of an
investment in the stock of a US corporation and that the transaction or transactions was or were
structured with a principal purpose of avoiding the application of the proposed regulations).
182 See generally US GAO report, supra note 91, at 29 (recognizing that the issuance of guidance
by Treasury or the IRS can provide new opportunities for tax abuse, particularly when
derivative instruments can be altered easily to achieve a desired tax effect).
183 See Nathan Boidman, “ ‘One of the Main Purposes’ Test” (2014) 22:5 Canadian Tax Highlights
9-10 (noting that the phrase “one of the main purposes” or an equivalent phrase—“one of the
main reasons,” “one of the main objectives,” or “one of the principal purposes”—appears in
the Act at least 79 times). In the Australian context, see ITAA 1997, proposed sections
230-348D(4), 230-348Q(4), and 230-348X(2) (a tax-avoidance purpose under the proposed set
of anti-synthetic rules need not be the dominant purpose of a hedged or derivative position,
but it must not be a merely incidental purpose).
184 See, for example, Hayward, supra note 9, at 1769 and 1780-81 (noting that synthetic dispositions
may be implemented to avoid price pressure from a block sale or to avoid contractual or
regulatory restrictions on a sale transaction). See also Wortsman et al., supra note 8, at 8:3-4 and
8:21-22 (suggesting that synthetic dispositions also may be undertaken to diversify a taxpayer’s
portfolio while retaining voting rights, and character conversion transactions may be undertaken
to allow retail investors indirect access to investments not otherwise available directly).
185 See Boidman, supra note 183, at 9-10 (arguing that a “one of the main purposes” test is
conceptually incoherent because there cannot be more than one main purpose and reviewing
the limited case law considering the phrase, which has ignored the incoherence and has
reduced the requirement in applying the test to “one of the purposes” or “one of the important
purposes”). See also Joint Committee on Taxation of the Canadian Bar Association and
Application of a Purpose-Based Standard for
Different Specifications of Risk Exposure
The need for a purpose-based standard to suppress discontinuities along the boundary between tax-driven and non-tax-driven positions in shares is greatest where the
particular boundary under a risk-based override is drawn relatively tightly to avoid
any adverse effect on positions that are not tax-driven. Drawing the boundary in
this manner is especially vulnerable to alteration of a position in order to fall on the
side of the boundary that is effectively assumed to be non-tax-driven. As the us experience with synthetic dispositions illustrates, a lack of non-tax constraints on this
behaviour, other than a willingness to assume some additional market risk and transaction costs, means that taxpayers will readily engage in alteration of their positions
to avoid an undesirable tax treatment that would otherwise result from the application of a risk-based override of ownership in the assignment of a tax attribute.
As described in part two, a precise specification of risk exposure as a percentage
amount allows for the articulation of a safe harbour, as well as a category of positions in shares that is considered more problematic in their motivation. Although
possibly improving the target-effectiveness of a risk-based override, this feature
creates multiple boundaries as a focus of tax planning. For example, it was suggested
that a safe harbour under a constructive sale rule could be drawn at a level of risk
reduction of 70 percent or less on the empirical assumption that partially hedged
positions with risk exposure of 30 percent or more indicate a dominant non-tax
purpose. Risk reduction of 90 percent or more could attract treatment as a disposition,
with a range of problematic partial hedges being defined as those with risk-reduction
levels greater than 70 percent and less than 90 percent. For these partial hedges, an
inquiry into taxpayer purpose in establishing the hedge would be required to characterize it as tax-driven or non-tax-driven. However, this inquiry would likely have
to be extended to partial hedges with risk-reduction levels just below 70 percent if
it were suspected that a hedge could be adjusted to move to a position in shares
within the safe harbour and avoid an inquiry into taxpayer purpose as the basis for
constructive sale treatment. Again, there seems to be very little of significance in
terms of taxpayer tolerance for risk exposure that would suggest that a 70 percent
level is a binding constraint on such behaviour. Unless transaction costs can confidently be relied on as a constraint, safe-harbour characterization can be framed as a
rebuttable presumption and contingent on an inquiry into taxpayer purpose.
A character conversion rule similarly appears to suffer from the same vulnerability as a constructive sale rule to tax-driven alteration of hedged and derivative positions. And this vulnerability seems to be present when the required level of risk
reduction for a hedged position in shares is drawn at any point in excess of 50 percent. For example, if this level were drawn at 70 percent, as suggested for illustrative
Chartered Professional Accountants of Canada, “Consultations on Treaty Shopping: Submission
by the Joint Committee on Taxation,” submission to the Department of Finance, December 11,
2013, at 13-14 (criticizing the use of a “one of the main purposes” test as an approach to treaty
shopping).
(2015) 63:2
purposes in part two, any level of risk reduction below that level would result in
characterization as a non-tax-driven hedge, and the risk-based override of share
ownership would not apply to recharacterize gain as a non-capital amount. The
70 percent level would thus provide the functional equivalent of a safe harbour,
much like that under a constructive sale rule, and would arguably require a purposebased standard to suppress the resulting discontinuity for partially hedged positions
clustered around the boundary. Lowering the level of required risk reduction to a
“majority” of the risk of loss and opportunity for gain/profit, as it is specified for
sale agreements under the dfa legislation, may do little to alleviate this vulnerability.
In effect, a lowering of the level of risk reduction in this way represents a difference
in degree rather than one of kind if there is nothing in the way of binding non-tax
constraints on tax-driven alteration of partial hedges that can be attributed to the
particular lower level. In the absence of evidence of a jump in transaction costs along
the specified boundary between tax-driven and non-tax-driven hedging, a purposebased standard should therefore probably be used to constrain tax-driven alteration
of positions clustered around the boundary under a character conversion rule by
taxpayers who are not otherwise constrained by intolerance for additional risk exposure. This policy prescription would seem to hold equally in the event that a level
of risk exposure were specified for long derivative positions in shares within the
definition of a purchase agreement under the dfa legislation and for the purpose
of the wash sale rules. It would also seem to hold for the purpose of the dividend
stop-loss rules, the anti-foreign-tax-credit-trading rule, and the dra rules and sea
proposals where these rules incorporate an at-risk requirement in the application of
a hold period requirement.
One possible difference related to the architecture of these risk-based overrides
that could negate the need for a purpose-based standard concerns a plausible empirical assumption concerning short-term share acquisitions around a dividend payment
date that is the occasion for the availability of dividend tax relief or a foreign tax
credit. Because of the short-term nature of these transactions, it might be the case
that taxpayer intolerance for risk is greater than it is with synthetic dispositions or
character conversion transactions. If that is the case, the level of risk reduction that
would attract application of any of these risk-based overrides could be specified at a
relatively high level that would define the category of affected partially hedged positions narrowly. However, the policy maker would have to have confidence that the
resulting boundary was drawn at a point where, because of a heightened intolerance
for additional risk exposure, taxpayers would be constrained in making small changes
to their partially hedged positions in order to avoid application of the relevant override and thereby maintain entitlement to dividend relief or foreign tax credits.
Me a surement of Risk E xposure
The discussion to this point has proceeded on the premise that risk exposure can be
measured with sufficient precision that risk-based overrides of ownership are credibly administrable. In fact, this measurement exercise is required whether the level
of risk exposure is expressed as a verbal formula or more precisely as a percentage
amount. It may therefore be somewhat surprising that, in the context of risk-based
overrides that explicitly use risk exposure as a proxy for taxpayer purpose, the measurement of risk exposure appears to be conducted on an ad hoc basis with no clear
guidance, even in a broad sense, as to what methodology or methodologies may be
used.186 It may be the case that a combination of the expression of levels of risk exposure using imprecise verbal formulas and the lack of any measurement guidance
is seen to provide suitable flexibility in determining risk exposure, and this flexibility
is considered a desirable policy property given an evident mistrust of the administrative integrity of formal measurement models.
It is suggested in this part of the article that the financial concept of delta could
be used to measure risk exposure with positions in traded shares, primarily because
it provides policy makers with greater flexibility in drawing a boundary between
tax-driven and non-tax-driven positions by specifying a level of risk exposure as a
percentage amount. Delta may be intractable, however, as a measurement tool for
positions in non-traded shares and other types of capital property that are subject
to risk-based overrides of ownership. In principle at least, different approaches
could be used for traded shares and non-traded shares, as well as capital property
more generally. For example, non-traded shares and other capital property could be
subjected to overrides that combine the expression of a level of risk exposure using
an imprecise verbal formula with the factual uncertainty that follows from a failure
to specify any measurement methodology. Traded shares could be subjected to overrides that combine the precise specification of a level of risk exposure as a percentage
amount with a required delta-based measure. If the policy flexibility provided by a
precise specification is not seen to warrant separate risk-based overrides, the alternative legislative pattern is continued use of coarse-grained expressions of risk exposure
through a limited set of verbal formulas, although nothing prevents the use of delta,
where practical, through administrative practice under the same specification.187
Methodological Indeterminacy as an
Approach to Risk Measurement
Where risk exposure is used as a proxy for taxpayer purpose, the characterization of
a particular position in shares (or other property) requires measurement and comparison of the exposure associated with an open position and the exposure associated
with a hedged or derivative position. This measurement exercise, in turn, requires
186 See Miller and Milet, supra note 8, at 10:26-28 (citing the US legislative history with IRC
section 1259 and suggesting that the similar standard in the definition of “synthetic disposition
arrangement” in subsection 248(1) of the Act results in methodological indeterminacy).
187 See Miller and Milet, supra note 8, at 10:27 (suggesting that a delta-based measure of risk
exposure would be attractive to those seeking a measurement tool in applying a 90 percent or
more risk-reduction test for the purpose of the SDA legislation).
(2015) 63:2
some type of accounting for asset price volatility and the volatility of dividend distributions.188 Consider, for example, a simple married put and call.189 Assume that a
block of shares is currently trading at $100 and the prevailing riskless interest rate
is 5 percent for the relevant time horizon. At one extreme, an open position exposes
the holder to a maximum loss of $100, while any appreciation in value is captured
as gain with, in theory, no limitation. At the other extreme, a fully hedged married
put and call position eliminates all exposure to loss (other than that attributable to
counterparty risk) and any chance of gain in excess of the riskless interest rate. The
latter result can be realized by (1) purchasing an at-the-money put option with an
exercise price of $100; and (2) selling an out-of-the-money call option with an exercise price of $105. In between these two extremes, however, is a range of partially
hedged positions. One possible point in this range consists of (1) the purchase of
an out-of-the-money put option with an exercise price of $90 and (2) the sale of an
out-of-the-money call option with an exercise price of $110. The long position in
the shares is partially hedged because the purchased put option protects the holder
from loss if the price of the shares moves below $90. The position is also partially
hedged on the upside because the sold call option requires the holder to pay out any
appreciation above $110. This partially hedged position leaves the holder with a
maximum risk of loss of $10 and a maximum gain of $10 ($5 expected gain equal to
the riskless interest rate + $5 unexpected gain attributable to changes in the price
of the shares). The fully hedged position leaves the holder with no risk of loss and a
maximum gain of $5 equal to interest at the riskless rate (plus or minus any gain/loss
attributable to differences in the amount of the option premiums).190
Even with this simple example, a comparison of the risk exposure associated with
an open position and the risk exposure associated with the partially hedged position
188 NYSBA report no. 868, supra note 24, at 11 (noting that the expected yield and price volatility of
a position in an underlying, as well as the terms of a hedge, are all relevant to a determination
of whether a taxpayer has substantially eliminated risk of loss and opportunity for gain, and
doubting that there is a single bright-line methodology for all transactions). See also S rep. no.
105-33, supra note 155, at 126-27 (stating that, in the exercise of the legislative power to write
regulations specifying transactions under IRC section 1259 with substantially the same effect
as the four enumerated transactions, it is anticipated that Treasury will address equity collars
and, in doing so, will take into account various factors with respect to an appreciated financial
position, including the yield and price volatility of shares and the period and other terms of the
options).
189 See NYSBA report no. 868, supra note 24, at 12-13. See also Explanatory Notes, supra note 94,
at 59 (example—put/call).
190 The locked-in return of $5 could be subject to a character conversion rule such as the DFA
legislation. Any appreciation on the underlying shares on entering the hedge would be subject
to a constructive sale rule such as the SDA legislation, while the locked-in gain should not be
accounted for in determining satisfaction of the relevant risk exposure. See Explanatory Notes,
supra note 94, at 33-34 (example—non-recourse loan) (referred to in Miller and Milet, supra
note 8, at 10:29, note 57). See also NYSBA report no. 868, supra note 24, at 11 (noting that
“locked-in” amounts that do not depend on changes in the value of an underlying should
generally be ignored in measuring opportunity for gain).
can only be executed with any factual integrity by accounting for the probability
that the price of the shares, on implementation of the partially hedged position, will
fall to any point from $90 to zero or will rise above any point from $110 to infinity.
If there is a trivial, or only a minimal, chance that the share value will move outside
a price range of $90 to $110, it may be confidently concluded that the partially
hedged position is sufficiently close to an open position to justify consistency of tax
treatment. On the other hand, if there is a likely chance that the price will move
below $90 or above $110, it may be confidently concluded that the level of risk exposure associated with the partially hedged position is not sufficiently close to an
open position to justify consistency of tax treatment. Confident conclusions about
the character of the partially hedged position can be drawn only if the price volatility
of the shares in this simplest of examples is accounted for in measuring the associated risk exposure.191
The explicitly risk-based overrides in the Act do not specify any particular
method of measurement of risk exposure. This lack of measurement specification
appears to be closely associated with the expression of a level of risk exposure using
an imprecise verbal formula rather than a precise percentage amount. The resulting
legislative pattern is characterized by both legal uncertainty and factual uncertainty.
As emphasized in parts two and three, legal uncertainty is attributable to the quantitatively imprecise specification of a level of risk exposure using verbal formulas.
Tax administrators and taxpayers are unclear as to the precise point at which a
boundary between tax-driven and non-tax-driven hedging is drawn, although the
tax administration process tends to bring some clarity to the boundary as the content
of the verbal formula is developed ex post in the context of hedged and derivative
positions. Factual uncertainty is attributable to the lack of any stipulated measurement of risk exposure, but unlike legal uncertainty, it is not necessarily clarified by
the tax administration process. The fact that a particular approach to the measurement of risk exposure is used in the context of a particular position does not mean
that it translates to another position.192
In fact, the tax administration process, including limited case law, under various
risk-based overrides in Canada and other countries reveals little, if anything, that is
instructive about the approach to the measurement of risk exposure with hedged
and derivative positions. In the absence of any specification of a methodological
approach to the measurement of risk exposure that would account for asset price
volatility, as well as dividend volatility in the case of shares, an apparently plausible
approach is to simply ignore volatility and determine the level of risk exposure using
191 See Brennan, supra note 24, at 267-68 (criticizing the holding in Rev. rul. 2003-7 and the
decision in Anschutz for their failure to account for asset and price volatility in considering the
application of IRC section 1259 to a VPFC). See the discussion in the text above at note 176
and following.
192 See Explanatory Notes, supra note 94, at 62 (characterizing the determination of whether all or
substantially all of the risk of loss and opportunity for gain or profit has been eliminated as
“highly factual”).
(2015) 63:2
extreme values as realistic possibilities.193 Because of an obvious inability to provide
any kind of weighting to upside or downside risk, this simplified approach presumably
also dictates that risk of loss and opportunity for gain or profit be assessed independently rather than as a whole in determining the risk exposure of a position.194
It may not be an overstatement, therefore, to suggest that the lack of any indication
as to what methodologies are acceptable, and in what circumstances they are acceptable, leaves the determination of risk exposure for the purpose of risk-based overrides
as an ad hoc “back-of-the-envelope” exercise, with assessments of volatility being
made in a casual and informal manner rather than by the use of a systematic or
standardized approach. The factual uncertainty is that much greater when volatility
of dividend distributions, as well as asset price volatility, must be accounted for with
positions in shares.195 Failure to adequately account for volatility also undermines the
target-effectiveness of risk-based overrides, with underinclusiveness being the more
likely feature, since taxpayers are free to elect alteration of their positions on either
the downside or the upside as the basis for avoidance of the stipulated standard.196
The lack of guidance, whether legislative or administrative, regarding the measurement of risk exposure is curious, at least with hedged and derivative positions in
shares, given that the financial concept of delta, in particular, measures directly what
the tax law requires to be measured under risk-based overrides of share ownership.197
That is, risk exposure associated with a hedged or derivative position in shares is
quantified by taking into account both asset price and dividend volatility, while
implicitly assigning weights to risk of loss and opportunity for gain or profit through
193 See, for example, supra notes 168-178 and the accompanying text regarding the experience
with IRC section 1259. All of the examples in the explanatory notes to the DFA and SDA
legislation reflect this approach, and there is no mention of how volatility is to be assessed. See
Explanatory Notes, supra note 94, at 34, 59-60, and 63-65. The importance of volatility to an
assessment of risk exposure is acknowledged, however, in the explanatory notes where it is
stated that an equity swap limited to dividend equivalent payments would not be considered to
be an SDA for the equity payer with a long position in shares where the price of the shares is
volatile and therefore most of the equity payer’s risk of loss and opportunity for gain or profit is
attributable to changes in the value of the shares: ibid., at 65 (example 2—swap).
194 See Miller and Milet, supra note 8, at 10:28 (noting that the expression of risk exposure as “risk
of loss and opportunity for gain or profit” could be interpreted as requiring either a combined
assessment of downside and upside risk or an independent assessment, but observing that
certain examples in the explanatory notes to the SDA legislation indicate that Finance accepts
the latter interpretation).
195 See, for example, Rev. rul. 2003-7, supra note 176 (concluding only that the taxpayer’s delivery
obligation under the VPFC entails “significant variation”).
196 See, for example, NYSBA report no. 868, supra note 24, at 16 (suggesting the use of a
rebuttable presumption rather than a safe harbour if risk of loss and opportunity for gain are
assessed independently for the purpose of applying a constructive sale rule).
197 See Brennan, supra note 24, at 273-74 (arguing that the use of delta to measure risk exposure
would improve the target-effectiveness of IRC section 1259 by avoiding discontinuities under a
“benefits and burdens” analysis of share ownership whereby small changes in some features can
result in different tax treatment for positions with the same delta values).
an assessment of volatility. In this respect, Australian and us policy makers have
deviated from standard legislative practice by explicitly mandating the use of delta
as a measurement of risk exposure for certain risk-based overrides of share ownership.
This limited use of delta is arguably suggestive of those circumstances in which the
measurement tool may be most suitable, with the policy gains from the enhanced
targeting of risk-based overrides outweighing the administrative and compliance
costs. In particular, a delta-based measure may be especially well suited to risk-based
overrides that employ a hold period requirement as a targeting mechanism or to
risk-based overrides in the context of character conversion transactions using long
derivative positions in shares.
Use of a Delta-Based Measure as an Exception
to Methodological Indeterminacy
As noted in part two, the anti-dividend-credit-trading rules in Australia198 and the
proposed dividend equivalent payment regulations in the United States199 specify a
level of required risk exposure precisely as a percentage amount. More particularly,
they stipulate that risk exposure is to be measured using the financial concept delta,
which is one of a set of risk measures used by options traders.200 Delta measures the
rate of incremental price change between a derivative instrument and an underlying
position;201 in other words, it provides the ratio of the price change of a derivative
instrument and an underlying asset, such as a share, and is most commonly associated
with dynamic hedging strategies.202 Both the Australian rules and the us proposed
regulations incorporate the concept of “position delta,” which uses delta to determine the required hedge ratio between a derivative instrument and a long or short
position in an asset.
Stated very generally, the Australian rules require a taxpayer to calculate a net
position for the period during which they hold a share each time that a position is
entered into by the taxpayer. A position in respect of a share is anything that has a
delta in relation to the share.203 A taxpayer with a net position that has a delta of 0.3
or less is considered to have “materially diminished risk of loss and opportunity for
198 See supra note 25.
200 See John C. Hull, Options, Futures, and Other Derivatives, 9th ed. (Englewood Cliffs, NJ: Prentice
Hall, 2014), chapter 19. Other measures are (1) vega (rate of change of the delta of an option
with respect to a change in volatility), (2) theta (rate of change of the value of a portfolio of
options with respect to the passage of time), (3) gamma (rate of change of the delta of an option
with respect to the price of the underlying asset), and (4) rho (rate of change of the value of a
position with respect to the interest rate).
201 Hull, supra note 200, at 285 and 402-3.
202 Ibid., at 403 (the goal of dynamic hedging as an arbitrage trading strategy is the construction
and maintenance of a delta-neutral position, which means that the delta of a share position is
offset by the delta of an option position and results in a net position delta of zero).
203 ITAA 1936, former section 160APHJ(2).
(2015) 63:2
gain,”204 with the result that days on which the delta of the net position is less than
0.3 are not counted in determining satisfaction of the hold period requirement for
dividend credit entitlement.205 For this purpose, a share is allocated a delta of positive 1,206 and the deltas of derivatives held or issued in combination with the share
are added (long positions) or subtracted (short positions) in determining a net position.207 The proposed us regulations specify that dividend compensation payments
under a “specified notional principal contract”208 (npc) or a “specified equity linked
instrument”209 (eli) with respect to shares of a us corporation will be treated as dividends to a non-resident recipient for withholding tax purposes if the delta of the long
derivative position under the npc or the eli is 0.7 or greater than that associated
with ownership of the shares.210 Status as a specified npc or eli is tested by reference to the delta of the relevant derivative position when the long party acquires the
transaction.211 As under the Australian rules, the deltas of multiple transactions are
aggregated in determining the delta of the composite long position.212
The combination of the precise specification of risk exposure as a percentage
amount and the specification of delta as the basis for measuring the exposure of a
hedged or derivative position in shares provides a risk-based override of ownership
with the properties of legal certainty and a potentially enhanced factual certainty
attributable to specification of a measurement methodology. There remains factual
uncertainty surrounding the calculation of delta, which is closely related to option
204 ITAA 1936, former section 160APHM.
205 ITAA 1936, former section 160APHO(3).
208 Prop. Treas. reg. section 1.871-15(d)(2). A “notional principal contract” is a notional principal
contract as defined in Treas. reg. section 1.446-3(a) (that is, a swap). Prop. Treas. reg. section
1.871-15(a)(5).
209 Prop. Treas. reg. section 1.871-15(e). An “equity linked instrument” is a financial transaction,
other than a securities lending or sale-repurchase transaction or an NPC, that references the
value of one or more underlying securities. Prop. Treas. reg. section 1.871-15(a)(4). Specific
examples include a futures contract, forward contract, option, debt instrument, or other
contractual arrangement that references the value of one or more underlying securities.
210 Prop. Treas. reg. section 1.871-15(g)(1). “Delta” is defined as the ratio of the change in the fair
market value of an NPC or ELI to the change in the fair market value of the property
referenced by the NPC or ELI.
211 Prop. Treas. reg. section 1.871-15(a)(7). “Long party” is defined as the party to a potential
section 871(m) transaction with respect to an underlying security that is entitled to a dividend
equivalent payment. Prop. Treas. reg. section 1.871-15(a)(7)(i). “Short party” is defined as the
party to a potential section 871(m) transaction with respect to an underlying security that is
liable for a dividend equivalent payment. Prop. Treas. reg. section 1.871-15(a)(7)(ii).
212 Two or more transactions entered into by a long party (or a related person) are treated as a
single transaction if they reference the same underlying security and are entered into in
connection with each other. Prop. Treas. reg. section 1.871-15(l).
pricing models 213 and requires the application of suitably complex mathematical
formulas.214 Even so, the calculation of delta for simple positions in options is reasonably straightforward,215 while on-market forward contracts have a delta of 1.216
More complex calculations are required for more complex derivative positions,
which generally must be decomposed into a known set of basic instruments.217 In all
instances, however, the more problematic inputs are asset price volatility and dividend
volatility (as well as the nature of dividends as fixed or proportionate to value). The
specification of these critical parameters requires the application of both mathematical expertise and an exercise in judgment;218 but the factual uncertainty is at least
confined for both the tax administration and taxpayers as compared with risk-based
overrides that do not specify a measure of risk exposure. Despite these desirable
properties, the use of delta tends to be viewed skeptically in the literature as too
compliance- and administrative-intensive.219 This characterization presumably follows
213 Hull, supra note 200, at 403 (“Delta is closely related to Black-Scholes-Merton analysis”).
214 Ibid., at 402-24, describing delta hedging.
215 Ibid., at 404-5 (calculation of delta of European share options). See, in this respect, NYSBA
report no. 1306, supra note 142, at 32 (suggesting that the calculation of delta can be made
simpler where a broker or dealer is not involved, by permitting the long party to use commonly
available online tools but subject to a condition that the necessary inputs, such as the interest
rate curve and volatility numbers, are within a range of commercially acceptable variation). See
also Schizer, “Frictions as a Constraint,” supra note 24, at 1365, note 188 (referring to the
availability of online tools for the computation of delta for simple derivative instruments and
suggesting that the government could offer presumptions based on typical facts to spare
taxpayers from the need to do computations, although the inaccuracy of such presumptions
could undermine the rationale for the use of delta).
216 Hull, supra note 200, at 420-21.
217 See, for example, Farber, supra note 110, at 14 (characterizing delta as “an unwieldy and
imprecise concept” whose functionality is debatable because it cannot be applied coherently in
not uncommon fact patterns, such as digital options that provide a fixed payout once a specified
threshold price is exceeded); and Walker, supra note 94, at 20 (noting that delta is difficult to
apply to derivatives with more complex economics than a simple total return swap referencing
an equity security). An example of the complexity of calculation that is required for more
complex derivative positions is provided by Brennan, supra note 24, at 264-74 (calculating delta
for the VPFC that is the subject of Rev. rul. 2003-7, as well as the VPFCs at issue in Anschutz).
See also NYSBA report no. 1306, supra note 142, at 26-31 (describing a series of examples
requiring a disaggregation approach to the calculation of delta and noting the difficulty with
digital options because there is no objectively determinable number of shares).
218 See Nassim Nicholas Taleb, Dynamic Hedging: Managing Vanilla and Exotic Options (New York:
Wiley, 1996), at 88-95 (calculating price volatility directly from past price movements or
indirectly from the price of traded options on shares). See also Hull, supra note 200, at 425
(noting that delta hedging is based on the assumption that asset price volatility remains constant);
and Miller and Milet, supra note 8, at 10:28, note 54 (emphasizing that some measures of risk
exposure are based on an assumed invariant level of distributions that focuses the measure on
asset price, which is not the standard under the SDA legislation, while asset price volatility tends
to be a trailing measure and is not a direct measure of market expectations of such volatility).
219 See, for example, the sources cited in note 24, supra.
(2015) 63:2
from the fact that there is no particular pricing model that fits all circumstances;
instead, the computation methodology for the determination of delta requires an
exercise in judgment that commonly involves a choice as to the use of option pricing models. Both taxpayers and the tax administration are thus required to obtain
suitable financial market expertise with a delta-based measurement of risk exposure
as an element of the targeting of risk-based overrides of share ownership.
Yet there is nothing unique about the need for expertise with the use of a deltabased measurement of risk exposure.220 Any number of factual determinations
under the Act, such as the determination of fair market value or an acceptable transfer price, requires the employment of expert evidence, and this requirement arguably
remains tolerable for compliance and administrative purposes. As with both of these
other factual determinations, the necessary data and calculations are normally generated for non-tax purposes,221 which in the case of delta will commonly be housed
with a broker-dealer who can be required to provide the relevant information to the
tax administration on behalf of a taxpayer with a hedged or derivative position in
shares.222 Moreover, the administrative and compliance burden assumed to be associated with the use of delta is overstated if there is no accounting for the administrative
and compliance burden associated with the standard legislative practice of failing to
specify a measurement of risk exposure under risk-based overrides. Less rigorous
220 But see Miller and Milet, supra note 8, at 10:28 (emphasizing that formal risk-measurement
methodologies may not be available to, or capable of being applied by, all taxpayers and CRA
auditors).
221 See, in this respect, prop. Treas. reg. section 1.871-15(g)(1) (“For purposes of this section, the
delta of an NPC or ELI must be determined in a commercially reasonable manner. If a taxpayer
calculates delta for non-tax purposes, that delta ordinarily is the delta used for purposes of this
section”). See also Mark J. Laurie, Liam Collins, and John Murton, “The 45 Day Holding
Period Rule—The Ultimate Walnut Crusher” (1999) 2:3 Journal of Australian Taxation 122-48,
at 128-29 (observing that the term “delta” is undefined legislatively on the apparent assumption
that its common usage in the finance industry will prevail but suggesting that the value of delta
may be calculated differently depending on the taxpayer’s perception of the relationship
between a derivative and an underlying share, while difficulties are created for taxpayers who
enter into transactions for which there is “no readily ascertainable delta”).
222 For example, prop. Treas. reg. section 1.871-15(o) requires brokers and dealers that are parties
to a potential section 871(m) transaction to provide information on request to the counterparty,
and where there is no broker or dealer, the short party must provide the information. See also
Alison Noble, “The Holding Period and Related Payment Rules: Are You Qualified for
Franking Credits?” August 10, 2010 (Deloitte Touche Tohmatsu, Sydney), 1-23, at 11 (“There
may be a number of ways to calculate the delta, and the tax legislation does not specify how a
delta is to be calculated. Deltas are frequently calculated by option traders. The delta may not,
however, be known by a taxpayer if shares and positions are held directly, rather than through a
broker or managed account, or if the position is something for which a delta is not usually
calculated by participants in the financial market. This presents some practical difficulties for
taxpayers in calculating their net position for shares”); and Walker, supra note 94, at 20 (noting
that derivatives dealers typically compute delta to establish an appropriate hedge of their
positions and should be in a position to provide this information to counterparties).
assessments of risk exposure are probably less administrative- and complianceintensive, but the reduction in cost comes at the price of a loss of precision. In short,
casual or informal assessments of risk exposure are coarse-grained, although they
probably arrive at a plausible result in extreme cases that are clearly within or outside the level of risk exposure described under a risk-based override of ownership.223
But the extreme cases are not the problematic ones along the boundary between
tax-driven and non-tax-driven hedged or derivative positions in shares defined by
risk-based overrides. With the problematic cases of positions with risk exposures that
lie along the boundary, it is not as clear that the tax administration or taxpayers can
be confident in a characterization that is based on an informal assessment of risk
exposure, and a more systematic accounting of price and dividend volatility for such
positions in shares should be required. In fact, failure to specify a measure of risk
exposure does not preclude the use of delta alongside other approaches that may
vary in their analytical rigour.224
Where the policy maker specifies the use of delta, it may even be the case that a
compliance and administrative dynamic will result that is comparable to that when
a methodological approach to the measurement of risk exposure remains unspecified. For those positions that are clearly within or outside the level of risk exposure
under a risk-based override, taxpayers and the tax administration presumably do not
have to expend significant (if any) compliance and administrative resources. By far the
greatest expenditure of these resources would be reserved for problematic positions
that lie close to the boundary and that require some precision of measurement, given
the associated tax consequences. This tax administration dynamic is again comparable to valuation and transfer-pricing inquiries where factual uncertainty allows for
a range of acceptable determinations and methodologies that are scrutinized, with
attendant administrative and compliance costs, only where the outer bounds of the
223 See, for example, IRC v. Scottish Provident, [2005] STC 15 (HL) (concluding that a sale and
purchase of call options on government bonds issued together with the same counterparty, to
take advantage of a gap in the transition to the application of capital gains tax to such options,
did not involve exposure to risk because the value of the underlying government bonds was
unlikely to fall within a range that would result in an unexpected gain or loss).
224 See NYSBA report no. 1306, supra note 142, at 31-32 (suggesting the use of surrogates for
delta where obtaining information for the calculation of delta is not practical because, for
example, the relevant positions are exchange-traded and do not involve a broker or dealer,
while the short party may not have sufficient expertise to generate the required delta
calculation). See also NYSBA report no. 868, supra note 24, at 13-17 (discussing an “option
pricing” approach to the measurement of risk exposure whereby the price of put and/or call
options written on shares would be compared with the price of comparable options that
eliminate all risk exposure, but acknowledging that such an approach is workable primarily
with liquid options markets and otherwise requires the use of option pricing models that entail
some subjectivity in the assessment of share price volatility, not unlike a delta-based measure);
and S rep. no. 105-33, supra note 155, at 127 (suggesting that regulations under IRC section
1259 consider the use of option prices and option pricing models in the assessment of risk
exposure associated with a partially hedged position).
(2015) 63:2
range are at issue. The unique issue under risk-based overrides, emphasized in part
three, is the discontinuity that arises from the drawing of a boundary between taxdriven and non-tax-driven positions, which makes these overrides vulnerable to
alteration of a hedged or derivative position to ensure inapplicability. As Brennan
illustrates in the context of the us constructive sale rule,225 the assumptions that
define the parameters of a pricing model used to calculate delta as a measure of risk
exposure can lead to discontinuous tax treatment attributable to small differences in
delta that cause characterization of a hedged or derivative position in shares to fall
on one side or the other of the boundary between tax-driven and non-tax-driven
positions. As with a fair market value or transfer-pricing determination, differences
in critical assumptions can be entirely plausible factually, particularly as they relate
to future asset price and dividend volatility; but the use of delta as a measure of risk
exposure under a risk-based override of ownership can have discontinuous tax consequences for plausible assumptions, and this is not necessarily the case with a
transfer-pricing or valuation inquiry, which tends to lack the same bright-line
boundaries as legislative parameters under risk-based overrides even where the level
of risk reduction is expressed using imprecise verbal formulas. The kind of purposebased standard suggested in part three can be adopted to constrain alteration of
hedged or derivative positions in shares, as well as the tailoring of pricing models
and resultant delta, where either is done in an attempt to avoid application of a riskbased override.226 The standard can also backstop legislative specification, if any, of
broad parameters that must be incorporated in modelling techniques.227 Indeed,
tailoring of modelling parameters would presumably be closely associated with the
alteration of a position that lies along the boundary between what are considered
tax-driven and non-tax-driven positions under a risk-based override.228
Although perceived administrative and compliance costs may be overstated, the
use of delta to measure risk exposure under risk-based overrides should probably
remain limited. Two generalized features of delta as a measurement tool tend to
suggest why its unqualified use is problematic: (1) the need for asset price history;
and (2) a focus on incremental price changes. With respect to the first feature, it is
notable that Australian policy makers have not proposed the use of delta as a measure
225 Brennan, supra note 24, at 264-74.
226 Walker, supra note 94, at 20-21 (characterizing delta as an imperfect proxy for dividend
equivalence that requires the support of specific anti-abuse rules and citing as an example a
long derivative position in shares that provides dividend equivalent payments but no exposure
to changes in value of the underlying shares).
227 See, in this respect, Macnaughton and Mawani, supra note 97, at 917-21 (recommending
legislative prescription of valuation techniques for employee stock options for the purpose of
determining the amount of a contribution to a registered retirement savings plan or a tax-free
savings account).
228 See, for example, prop. Treas. reg. section 1.871-15(n) (permitting the commissioner to adjust
the delta of a transaction where the anti-abuse rule applies).
of risk exposure in the set of proposed anti-synthetic rules as part of the larger legislative project on the taxation of financial arrangements. These proposals, which
include a constructive sale, a character conversion, and a wash sale rule,229 follow instead the standard legislative pattern of specification of a level of risk exposure using
a verbal formula while failing to specify any particular approach to measurement.
Similarly, the preamble to the proposed dividend equivalent payment regulations in
the United States provides that the use of delta is not to be interpreted as an approval of its use in other legislative contexts.230 The resultant limited use of delta to
measure risk exposure in the context of hedged and derivative positions in shares for
the purpose of the Australian anti-dividend-credit-trading rules and the us proposed
regulations for dividend equivalent payments may be attributable to the fact that the
pricing methodologies underlying delta are tractable for positions in shares but not
necessarily other types of property that tend to be thinly traded. Most importantly,
the integrity of the pricing models on which the calculation of delta is invariably
based requires sufficient asset price history as the basis of volatility parameters.231
Because delta measures the change in price of a derivative instrument relative to
small incremental changes in the price of an underlying asset,232 it is also probably
best suited as a measurement tool for the purpose of those risk-based overrides that
incorporate hold period requirements.233 This feature of delta means that it changes
as the passage of time reveals changes in the price of the underlying asset that require rebalancing of a hedged position with a dynamic hedge strategy.234 For the
purpose of a hold period requirement, an approach based on a taxpayer’s net position in shares, such as that reflected in the Australian anti-dividend-credit-trading
rules, fits readily with a delta-based measure of risk exposure. Where a net position
changes as a hedge is rebalanced, delta can be recalculated with respect to the rebalanced position to test for the required level of risk exposure for each day within the
230 Preamble to prop. Treas. reg. section 1.871-15, supra note 26, at 841 (“This notice of proposed
rulemaking should not be construed as providing guidance with respect to any other section of
the Code. For example, this notice should not be used as a basis for applying the delta standard
to interpret other Code sections”).
231 See Brennan, supra note 24, at 275 (noting that there would be difficulties in computing delta
for illiquid assets under a delta-based measure for constructive sale purposes). See also Miller
and Milet, supra note 8, at 10:28 (observing that the precision provided by “formal mathematical
methodologies” to measure risk exposure is limited to assets trading in liquid markets with
sufficient price history to derive volatility estimates); and NYSBA report no. 868, supra note 24,
at 18 (suggesting that evidence regarding the fundamentals of an underlying business could be
relevant in establishing a range of expected future share values that could bear on an assessment
of risk exposure of a partially hedged position in shares).
232 Hull, supra note 200, chapter 19.
233 See NYSBA report no. 868, supra note 24, at 20 (noting that a delta-based measure of risk
exposure avoids the issue of establishing an appropriate time horizon, which is problematic
with an option pricing approach).
234 Hull, supra note 200, at 405-8.
(2015) 63:2
specified hold period. Without some qualification on its use, delta is not necessarily
well suited, however, for those risk-based rules, such as a constructive sale rule, that
apply to deem a disposition or other tax consequence to arise once a specified level
of risk exposure occurs. For example, use of delta to measure risk exposure can result in a deemed disposition of an asset where the required level of risk reduction is
considered to occur simply because of a change in the price of the underlying, necessitating a recalculation of delta, even though the net position is a “hedge-and-forget”
strategy235 with no rebalancing of the position. A deemed disposition could occur
despite the fact that the hedge was constructed as a partial hedge outside the level
of risk reduction specified under a constructive sale rule at the time the taxpayer
entered into the position and was not subsequently altered.236
Irrespective of perceived administrative and compliance costs, there is therefore
a defensible case for the cautious use of delta to measure risk exposure for the purpose of risk-based overrides of share ownership. Indeed, the lack of any particular
measurement tool that can be used with confidence for property other than shares,
as well as thinly traded shares, undermines greatly the precision realized by specification of a level of risk exposure as a percentage amount. In effect, the ability to choose
any point within a range of risk exposure as the boundary between tax-driven and
non-tax-driven positions is largely meaningless if the risk exposure of any particular
hedged or derivative position cannot be measured with equivalent precision. Policy
makers may be left with no choice of instrument other than a choice among a limited set of points within a range of risk exposure that can be described using verbal
formulas such as “all or substantially all,” “majority,” or even “material.” It is possible, of course, to specify the same risk-based rules differently for shares and for
other property (and even thinly traded shares), but this legislative alternative could
result in additional compliance and administrative costs as compared with a single
rule applicable to both shares and other property.237 It is unclear empirically whether any gain in precision of specification for hedged and derivative positions in shares
would warrant the additional costs.
If the gain in target-effectiveness from precise specification of risk exposure is in
fact seen to be warranted, positional delta can, in principle, be used for both riskbased overrides of share ownership that incorporate a hold period requirement and
those overrides that create tax consequences when, at any time, the stipulated level
of risk exposure is not maintained. In effect, stipulating that delta is to be calculated
only when a net position in shares is entered into, or is altered, ensures that a change
235 Ibid., at 403 (referring to the implementation of a hedge without subsequent adjustment as
“static hedging” or “hedge-and-forget”).
236 See NYSBA report no. 868, supra note 24, at 20-21.
237 A differently specified rule is provided for foreign tax credit trading transactions in non-capital
property and short-term securities acquisitions held as non-capital property subject to the hold
period requirement in subsection 126(4.2). See subsection 126(4.1) (identifying foreign tax
credit trading transactions as those that lack economic profit).
in delta without any change in a net position does not give rise to the application of
tax consequences under a risk-based override. However, the use of positional delta
leaves open the application of an override to a dynamically hedged portfolio with
constant rebalancing of a net position. The rebalancing, which is invariably nontax-driven, can result in tax consequences under risk-based overrides, even though
such overrides are targeted at what is considered to be tax-driven hedged or derivative positions, which tend to be associated with hedge-and-forget strategies that are
commonly structured transactions. For traders or dealers who mark their portfolios
to market on income account for income tax purposes, the need for an exception to
a risk-based override is limited to those overrides that incorporate a hold period
requirement. For institutional investors, such as mutual funds, that are subject to
realization-based recognition of gain or loss, which may be on capital account, both
types of risk-based overrides are implicated and arguably require an exception to
refine the targeting otherwise provided through the specification of a level of risk
exposure. One possibility is a “portfolio election,” such as that under the Australian
anti-dividend-credit-trading rules, which permits qualifying taxpayers to claim a
maximum amount of credits on share portfolios without application of the hold
period requirement.238 Designing the boundaries of a suitable exception is nonetheless a difficult proposition, given the policy imperative to exclude hedge-and-forget
strategies from the exception in order to ensure that it is not abused.239
In the absence of a broadly focused dividend equivalent payment rule in the Act
focused on tax-avoidance transactions,240 the dra rules, as modified by the proposed
incorporation of the sea provisions, appear to be the most obvious target for legislative revision using a delta-based measurement of risk exposure. The combination
of an imprecisely specified level of risk exposure and a required inquiry into taxpayer purpose means that the dra rules are characterized by both legal uncertainty
238 ITAA 1936, former section 160APHR and former section 160AQZE. See, in this respect,
Laurie et al., supra note 221, at 141 (characterizing the formula-based ceiling election as a
concession to institutional investors and large retail investment managers that are considered
to be low-revenue risk and otherwise have considerable difficulty in meeting the hold period
requirement because of substantial share turnover). For administrative and compliance cost
reasons, a de minimis exemption from the hold period requirement is also available for
individuals whose dividend credit claim for a taxation year does not exceed A $5,000. See ITAA
1936, former section 160APHT.
239 NYSBA report no. 868, supra note 24, at 21-22 (observing that it would be a mistake to
exclude dynamic hedging from a constructive sale rule if such an exclusion would lead to
substantial abuse, but suggesting that traders and non-dealers not subject to mark-to-market
taxation do not present sufficient concern to justify extension of constructive sale treatment to
dynamic hedging strategies).
240 The SLA rules address the treatment of dividend equivalent, as well as interest equivalent,
payments: subsections 260(5) and (8). The Act does not, however, include an anti-avoidance
rule addressing dividend equivalent payments associated with long derivative positions
generally in shares.
(2015) 63:2
and factual uncertainty, with the kinds of attendant costs emphasized in this article.
Both legal and factual uncertainty could be moderated by providing instead for
n
n
n
n
a hold period requirement,
a requirement that shares be held at risk for the hold period,
specification of a required level of risk exposure as a percentage amount, and
measurement of risk exposure using delta.
The other obvious possibilities for similar revision are the anti-foreign-taxcredit-trading rule applicable to short-term securities acquisitions and the dividend
stop-loss rules. These overrides already have hold period requirements, but could
also be recast to replace the weak at-risk standard of the sda legislation with a
higher level of required risk exposure and a requirement that delta be used to measure risk exposure. Given a lack of price data necessary for modelling purposes, it is
not clear, however, how delta could be used as a measurement tool for non-traded
shares any more effectively than it could for property other than shares. Although
market unavailability of suitable hedging instruments may constrain a range of taxdriven positions, the need to apply each of these risk-based overrides to non-traded
shares seems clear on the likely empirical assumption that such positions are not
completely constrained by the unavailability of accommodation parties and their
ability to hedge out their risk exposure.241 The policy imperative to address partially
hedged positions in both traded and non-traded shares means that the principal
legislative reform options to the status quo are
n
n
differently specified dra rules, dividend stop-loss rules, and anti-foreign-taxcredit-trading rules for traded and non-traded shares; and
the same precisely specified rules for both traded and non-traded shares.
The attractiveness of the first alternative again depends on an assessment that
any possible gain in target-effectiveness from a precisely specified rule for traded
shares warrants the additional administrative and compliance cost of two separate
sets of risk-based rules to address the same policy issue. The attractiveness of the
second alternative depends on an assessment that the limitations of the use of delta
in the context of non-traded shares are insufficient to outweigh the gains captured
by the legal and factual certainty characteristic of a precisely specified rule. In this
respect, the Australian anti-dividend-credit-trading rules and the proposed us dividend equivalent payment regulations both extend to positions in traded and nontraded shares.
241 See, for example, Gentry and Schizer, supra note 172, at 178-89 (analyzing 61 transactions
between 1993 and 2001, and finding that an offering of exchangeable securities as a hedging
technique to avoid IRC section 1259 was associated with a decline of almost 4 percent in the
underlying share price before implementation of the hedge).
Co n c l u s i o n
This article has conceptualized a set of specific anti-avoidance rules in the Act as
risk-based overrides of share ownership in the assignment of tax attributes. It suggests that these rules, which include the recently enacted dfa and sda legislation
and the recently proposed sea provisions, could be made more target-effective if
they were recast to specify risk exposure, as a proxy for taxpayer purpose, explicitly
as a percentage amount. Two important constraints, however, undermine realization of this policy goal. One constraint is the lack of systematic empirical evidence
of the motivation for particular positions in shares that would suggest where precisely
the boundary between tax-driven and non-tax-driven positions should be drawn
under each of the risk-based overrides. This lack of empirical evidence means that
precise specification of risk exposure as a percentage amount remains a necessarily
blunt instrument that can potentially improve target-effectiveness in a general directional sense only, with the relevant direction suggested by a casual empiricism
grounded in anecdotal evidence. The other constraint is the need for a systematic
and factually credible approach to measurement of risk exposure that can be applied
across risk-based overrides. A delta-based measure, which is suggested and rejected
in some of the literature as too administrative- and compliance-intensive, could in
fact prove feasible for traded shares, with some legislative guidance being provided
by its use under the Australian anti-dividend-credit-trading rules and the proposed
us dividend equivalent payment regulations.
Yet, beyond simple derivative positions, a delta-based measure is not entirely
tractable for property other than shares and is much less tractable in the case of
non-traded shares than it is with publicly traded shares. In the absence of any other
obvious method of risk measurement, the policy maker is left to choose among policy options that are all suboptimal to some extent. One policy option is maintenance
of the status quo, which is characterized by an eclectic set of explicit and implicit
specifications of risk exposure as proxies for taxpayer purpose. Another choice is a
more limited application of this status quo to property other than shares under
those overrides that apply to both shares and other capital property. Risk-based
overrides could be recast for positions in shares using precise specification of risk
exposure as a percentage amount and a delta-based measure of risk exposure. A third
option is comprehensive application across all risk-based overrides to all types of
property using an imprecise verbal formula to express the requisite level of risk exposure while also requiring an inquiry into taxpayer purpose. This particular menu
of policy options is problematic largely because the choice must be made in a state
of empirical ignorance. It is unclear whether the second and third policy options
provide an improvement in target-effectiveness that would warrant the associated
administrative and compliance intensity. In this kind of policy environment, maintenance of the status quo tends to be an irresistible policy choice. Nonetheless, there
is a much clearer case for the use of a purpose-based standard to backstop the application of risk-based overrides, whether cast as imprecise expressions of risk exposure
using standard verbal formulas or more precisely as a percentage amount.
Policy Forum: Editor’s Introduction—
Resource Taxation
The taxation of natural resource industries requires consideration of the proper
taxation of economic rents. Consider a situation in which the fair value of a resource
under the ground is $100, the full cost of extracting and marketing the resource (including compensation for all economic costs, including risk and return to capital) is
$20, and the revenue from sales is $120. If a firm could acquire the right to extract
at no charge, it would earn an economic profit (revenue minus all costs) of $100. On
the other hand, if the firm had to pay $100 for the right to extract, its total costs and
its revenue would be equal ($120), and it would earn only the normal rate of return.
Under standard assumptions, the firm would go ahead with the project whether the
price for extraction rights were zero or $100, because any revenue in excess of costs
represents “economic rent”—a return that exceeds what is necessary to keep economic decisions unchanged. Whether and how much to tax this economic rent
presents a vexing challenge for governments of economies with significant resource
industries.
The challenges arise when moving from the abstract to the actual. Some of the
challenges are political: any change from the status quo necessarily produces winners
and losers, and the losers may find politically costly ways to express their objections
to the change. But there are also conceptual and measurement issues surrounding
the accounting of costs: what are the costs to be considered, and how are they to be
measured? In this policy forum, two articles explore different aspects of the taxation
of resource industries.
In the first article, Wayne Mayo explores the general case for resource rent taxation and describes various frameworks that may be used to tax resource rents. In the
course of his discussion, he refers to recent experience in Australia, where a federal
resource rent tax on mining was introduced in 2012, and then repealed in 2014.
In the second article, Brian Conger and Bev Dahlby look at municipal taxation
of machinery and equipment and linear property (melp) in Alberta. Taxes of this
sort are featured in all four western Canadian provinces, and combine elements of
rent taxation and benefit taxation. Conger and Dahlby explore some policy options,
including the use of technology to shift melp taxes closer to a pure benefit basis and
the imposition of a cap to limit excessive reliance on melp taxes.
Together, these two articles highlight some of the difficulties in imposing a pure
rent tax on resource industries. Do the municipal melp taxes constitute a fee-forservice benefit tax, or do they extract some of the rent? This question must be resolved
in order to calculate the base of any resource rent tax, yet it is clearly a difficult one
 467
(2015) 63:2
to answer. It is clear that any attempt to implement an Australian-style resource rent
tax would, at a minimum, have to grapple with these questions of measurement in
order to ensure successful implementation.
Kevin Milligan
Editor
Policy Forum: Resource Rent Taxation—
Experiences from Australia
Wayne Mayo*
Précis
La propriété collective des ressources minérales de l’Australie et les redevances
traditionnelles fondées sur la valeur ou le volume imposées par les gouvernements
régionaux dans le système fédéral de l’Australie ont entraîné des propositions de
remplacement de ces redevances traditionnelles par un mode de perception bien conçu
fondé sur les bénéfices. Un tel remplacement permettrait une meilleure attribution des
fonds d’investissement dans le secteur des ressources minérales de l’Australie, y
compris de l’investissement dans des opérations marginales qui seraient peu lucratives
en vertu des redevances traditionnelles, et la pleine récupération économique des
minerais. Un gouvernement australien récent a proposé un impôt sur les excédents de
trésorerie de conception audacieuse incorporant la compensation retardée de la totalité
des pertes (supprimant le risque, inhérent aux impôts traditionnels sur le loyer des
ressources, de perdre la valeur des déductions des dépenses), ainsi que des dispositions
annulant l’effet des redevances traditionnelles des gouvernements régionaux. Dans ce cas,
cette conception audacieuse n’a pas été mise en œuvre et une idée connexe est que la
compensation de la totalité des pertes, considérée par certains comme une caractéristique
de conception cruciale pour l’obtention des résultats escomptés est un gaspillage de
fonds publics pour d’autres. Le remplacement des redevances traditionnelles exigera un
modèle fiscal solide en plus de consultations méticuleuses et judicieuses.
Abstract
Australia’s community ownership of mineral resources and traditional value- or volumebased royalties imposed by regional governments in Australia’s federal system set the
scene for proposals to replace these traditional royalties with well-designed profit-based
imposts. Such replacement offers better allocation of investment funds within Australia’s
mineral resources sector, including investment in marginal operations that would be
unprofitable under traditional royalties, and full economic recovery of minerals. A recent
Australian government proposed a cash flow tax of bold design incorporating delayed
full loss offset (removing the risk, inherent in traditional resource rent taxes, of losing
the value of expenditure deductions), as well as arrangements nullifying the effect of
traditional regional government royalties. In the event, this bold design was not
* Formerly of the Australian Treasury (e-mail: [email protected]).
 469
(2015) 63:2
implemented, and one related insight is that full loss offset, viewed by one person as a
crucial design feature for efficient outcomes, is a waste of taxpayer funds to another.
Replacement of traditional royalties would require solid tax design coupled with
meticulous and judicious consultation.
Keywords: Resources n rents n investments n neutrality n income taxation n Australia
Contents
Introduction
Existing Projects
Budget Management Risk
RSPT and Budget Management Risks
RSPT and Income Taxation
Tax Revenue Risk
Regional Government Royalties
Future Policy Direction
470
473
475
475
477
480
483
485
Introduc tion
Why is introducing resource rent taxation by government often considered problematic? After all, to tax above-normal profit, or economic rent, all that is needed is
a cash flow tax (cft). Negative cash flow would attract a government cash rebate of
cash outflow multiplied by the cft rate. Positive cash flow would require a payment
to government of cash inflow multiplied by the cft rate.
Thus, under a 50 percent cft, $1,000 of cash outflow in year 1 from capital expenditure (no non-cash items involved here) attracts a $500 cash rebate. Net receipts
in year 2 from the extraction of mineral resources (possibly petroleum) require a
cft payment of $600. (See table 1.)1
Investors assessing the prospective project in table 1 see it offering a 20 percent
(internal rate of ) return both before and after cft. Where is the economic rent to
be taxed? This depends on the potential investor. An investor using a 20 percent
risk-weighted discount rate sees the project as a marginal prospect with zero net
present value (npv) both before tax (−$1,000 + $1,200/1.2) and after (with both
−$1,000 and $1,200 being cut by 50 percent). For this investor, there is no economic rent and no tax on it: in discounted terms, the tax payment in year 2 matches
the cash rebate in year 1.
In contrast, an investor using a 10 percent risk-weighted discount rate views the
project as a profitable prospect. This investor puts the project’s npv and its economic
rent—or the value of its mining lease just before the $1,000 of expenditure in year 1—
at $90.91 (−$1,000 + $1,200/1.1) and sees half of that npv paid out in tax (again
with both −$1,000 and $1,200 being cut by 50 percent). Yet another investor using
1 The tables and figure in this article reflect material in Wayne Mayo, Taxing Resource Rent:
Concepts, Misconceptions and Practical Design (Weston Creek, ACT: Kyscope, 2013).
policy forum: resource rent taxation—experiences from australia  n  471
Table 1 Two-Year Project, 50 Percent Cash Flow Tax
Pre-tax
Tax
dollars
Year 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . −1,000
Year 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1,200
−500
600
NPV @ 20% . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
Post-tax
0
−500
600
0
IRR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20%20%20%
IRR = internal rate of return.
NPV = net present value.
a 25 percent discount rate would view the project as unviable before tax, with negative $40 value and economic rent (−$1,000 + $1,200/1.25). Post-tax value would be
negative $20 (again, pre-tax value being cut by 50 percent), with the upfront rebate
viewed as exceeding the discounted year 2 tax payment (showing how the tax gives
back to unprofitable outcomes while taking from profitable outcomes).
Thus, economic rent is in the eye of the investor, and cft design accommodates
that personal view of economic rent.
With cft applied to just one sector in one country, npvs of prospective ventures
in that sector would be reduced in proportion to the cft rate, affecting the ranking
of projects in that sector relative to untaxed operations elsewhere in that country or
in other countries. But profitable ventures in the sector would remain profitable,
and, in contrast to the situation with imposts based on value or volume of production, marginal ventures should remain so after cft.
Conceptually, cft could apply before or after income tax, but practicalities point
to application before income tax; in particular, difficulties may arise in allocating the
right amount of taxable income of diverse organizations to their cash flows subject
to cft. Where cft is applied before income tax, ideally income tax should be applied to cash flows after cft (as in table 2), not to pre-cft flows.
Table 2 shows income tax applied to post-cft cash flows of our two-year project,
which is being assessed and the value of which is set by investors using a 20 percent
discount rate. As before, such investors view the project as marginal, with zero npv
both before and after 50 percent cft. When income tax is applied to post-cft cash
flow, $30 of income tax is paid on taxable income of $100 in year 2 ($600 cash flow
less the $500 reduction in value or, in practice, writeoff of $500 of year 1 post-cft
capital expenditure), resulting in cash flow of $570 in year 2 after all taxes. For these
investors, the discount rate after income tax is reduced in proportion to their
30 percent income tax rate to 14 percent. Thus, npv after all taxes is zero, the same
as npv before and after cft, because both taxable income and the investors’ discount rate are reduced in the same proportion (by the 30 percent income tax rate).
Such a tax-neutral result would not be achieved under a design that included the
$500 cft cash rebate in year 1 assessable income and in year 2 included $1,200 net
receipts and allowed deductions for changes in pre-tax value (a $1,000 reduction in
this case) and $600 cft payment. While under such design the same $30 net amount
(2015) 63:2
Table 2 Two-Year Project, 50 Percent Cash Flow Tax, 30 Percent Income Tax
CFT
Post-
Taxable Income tax After
Pre-tax (50%)CFTValueaincomeb
(30%)
all taxes
dollars
Year 1 . . . . . . . −1,000
−500
−500500
0
0 −500
Year 2 . . . . . . . 1,2006006000100 30 570
NPV c . . . . . . . 0 @ 20%
0 @ 20%
0 @ 20%
IRR . . . . . . . . . 20% 20% 20%
0 @ 14%
14%
CFT = cash flow tax.
a Value, after cash flow arising in the year, from discounting future post-CFT cash flow at 20%
per annum. Value in year 1 matching the cash outflow in that year shows the marginal nature
of the project at a 20% discount.
b Cash flow plus change in value. In year 1, $500 negative cash flow from capital expenditure is
offset by the resulting $500 value. In year 2, the loss in value is $500 since the project has zero
value when it ends.
c For an investor applying a 20% discount before 30% income tax, the discount rate after income
tax becomes 14% (that is, the post-income tax comparative return of a comprehensively taxed
financial asset with a pre-tax return of 20% reflecting the same risk as the two-year project).
Should a 10% discount rate be used by investors to assess the project and set project value,
year 1 post-CFT value would be $545.45, and in year 2 taxable income would be $54.55,
income tax would be $16.37, and cash flow after all taxes would be $583.63. With income tax
staying at zero in year 1, assuming that the difference between upfront value and required
expenditure is not taxed, NPV after all taxes with discounting at 7% [10 × (1 − 0.3)%] would
be $45.45, the same as post-CFT NPV.
of income tax would be paid, payment would be heavily front-loaded ($150 in year 1
and −$120 in year 2), resulting in post-tax npv at 14 percent of −$18.40.
So, even after factoring in income tax treatment, why is a cft on mineral resources
problematic for government? The answer lies in the tax revenue and budgetary risks
associated with a cft. Here is a tax that, when established as a stable regime and
when combined correctly with income taxation, should have little impact on investment decisions in a sector with potentially sizable economic rent driven by a limited
supply of quality mineral deposits. Nevertheless, the tax-neutral properties of a cft
derive from its taking from profitable outcomes while also giving back to unprofitable
outcomes. This means that government is relying on the commercial nous of investors to achieve net tax from profitable ventures that more than offsets net cash rebates
paid out to failed operations. In addition, a cft poses particular risks to budget
planning from unexpected annual mismatches between cash rebates and tax receipts
and unexpected changes in profitability from commodity price movements.
In contrast, traditional value- or volume-based royalties not only do not give back
to unprofitable outcomes; they also offer the prospect of taking from unprofitable
outcomes. Thus, traditional royalties interfere with investment decision making,
even with the best intentions of a government that is prepared to adjust royalty
levels with changing circumstances. Nevertheless, traditional royalties are easy to
collect and difficult to avoid, and are less sensitive to variations in profitability of
mining operations than is a cft.
Deeper understanding of why governments may balk at the introduction of a cft
may be aided by looking beneath the design of three variants of resource rent taxation either considered or introduced by the Australian government in recent years:
1. a petroleum resource rent tax (prrt), effective from July 1, 1984, initially
applying only to offshore petroleum projects;
2. a resource super profits tax (rspt) proposed in May 2010 to apply to all mining and petroleum projects with provision for projects subject to the prrt to
transition to the rspt, but ultimately not implemented; and
3. a minerals resource rent tax (mrrt), which applied from July 1, 2010 to coal
and iron ore projects, together with an extension of the prrt to onshore
petroleum projects.
The mrrt was repealed in late 2014, leaving the extended prrt in place.
Differences in the design of the prrt, the rspt, and the mrrt help to illustrate
key issues that confront governments interested in considering the introduction of
resource rent taxation.
E xisting Projec ts
First, there is the issue of existing projects.
For example, consider the two-year project illustrated in table 1. If the project is
undertaken after the introduction of a 50 percent cft, post-tax outcomes are as
shown in the table. If, however, the project’s $1,000 year 1 expenditure is made just
before the unexpected introduction of the tax, post-tax cash flows and profitability of
the project look very different. In these circumstances (shown in table 3), the government is not sharing in the $1,000 of year 1 capital expenditure (by providing a
50 percent cash rebate) but is essentially just appropriating 50 percent of year 2 net
receipts. This might look like the impact of an ad valorem royalty, and existing owners of the project might well continue on and produce the $1,200 of net receipts
having sunk the upfront costs. Moreover, the government may wish the cft to take
from existing, as well as new, projects. Nevertheless, the project’s owners no doubt
would be less than happy at having a 50 percent capital levy suddenly imposed on
them (with the value of the project instantly falling from $1,000 to $500 post-cft).
Moreover, the sovereign risk arising from a worry that such government action
might be repeated in future could adversely affect future investment decisions.
If, however, the design of the unexpected cft included a “starting base” year 1
allowance of $1,000 for this existing project, the project would once again attract a
$500 cash rebate in year 1, compensating for the $500 capital levy. Project cash
flows would again look like those in table 1. The owners of this project using a
20 percent discount rate to value the project (seen as marginal to them) would view
(2015) 63:2
Table 3 Two-Year Project, 50 Percent Cash Flow Tax Starting in Year 2
Pre-tax
Tax
Post-tax
dollars
Year 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . −1,0000
−1,000
Year 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1,200
600
600
NPV @ 20% . . . . . . . . . . . . . . . . . . . . . . . . . . . 0 500 −500
IRR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20%
−40%
the $1,000 starting base as matching both book value (expenditure undertaken) and
market value. In contrast, if the owners were using a 10 percent discount rate to
value the project, while the $1,000 starting base would still match year 1 book value,
the associated $500 cash rebate would not match the $545.45 capital levy (the
$1,090.91 upfront market value of the project to the owners cut in half ). For these
owners, a starting base of $1,090.91 would be required to compensate fully for the
capital levy.
From government’s point of view, allowing market value starting bases for existing projects turns them all into marginal projects from a tax revenue perspective.
Differences in discount rates aside (since governments, in any case, often put a premium on getting budget impact in a particular year), the upfront value of expected
net tax revenue from existing projects would be zero. Moreover, under pure cft
design, that zero expectation would be accompanied by potentially large upfront
cash rebates for existing projects. And government would face the risk that future
tax receipts may not justify the size of upfront rebates—particularly if market values
were set during a boom period of high commodity prices. Practical difficulties
would also be involved in setting market value for each existing project, a process
potentially involving case-by-case negotiations.
In Australia, the initial prrt design avoided these issues because, after preliminary consultation on the application of the tax to all petroleum mining onshore and
offshore in Australia, the scope of the tax was narrowed to new offshore projects.
Existing offshore projects, including Bass Strait and the fledgling North West Shelf,
were excluded (though Bass Strait was later included in the prrt net).
In contrast, the proposed rspt was to apply to all existing and new mining and
petroleum operations (except those already subject to the prrt, which could elect
into the rspt). The starting base for existing projects was to be book value rather
than market value, a feature that no doubt left some companies with existing operations somewhat disgruntled.
The mrrt, which superseded the rspt and applied to new and existing iron ore
and coal projects in Australia, provided for either market value or book value starting
base with differing delayed writeoff provisions—pushing toward an effect similar to
the original prrt design that applied only to new projects. Market values likely
reflected high commodity prices at the time. Similar starting base arrangements
apply to existing petroleum projects subject to extension of the prrt, which accompanied the mrrt, to include new and existing onshore petroleum projects.
Budget Management Risk
The potential for large upfront cft cash rebates for existing projects underlines the
more general issue for government of a timing mismatch between immediate cash
rebates and subsequent tax receipts under a cft. This mismatch, posing difficulties
for budget planning and risks for budget outcomes, is exacerbated by unexpected
large projects, or mining booms with extended periods of large negative cash flows
that are often difficult to predict. Beyond budget management, government faces
the risk of providing cash rebates without being sure that future tax receipts will
more than compensate for those payments.
In Australia, bold design of the rspt accepted the risks around tax revenue security but sought to address budget management stress.
RSPT and Budget Management Risks
cft design allows immediate writeoff of expenditures, so that negative cash flow
translates into immediate cft losses, and then provides immediate cash rebates
equal to the tax rate multiplied by any unutilized losses—losses that cannot be written off against cft profit (positive cash flow). In other words, immediate “full loss
offset” is provided.
The rspt sought to achieve financial equivalence to a pure cft while addressing
budget management stress by a design that incorporated depreciation of capital expenditure and loss carryforward. Capital expenditure was to be written off over a
number of years (the number being subject to consultation) with an interest (uplift)
rate to apply to both depreciated value and rspt losses carried forward. Annual
uplift would reduce rspt payments or increase the size of losses carried forward, or
both. Because uplifted unutilized losses ultimately (timing being subject to consultation) were to attract government-guaranteed cash rebates, the uplift rate was set
at the long-term government bond rate (ltbr), incorporating systemic risk of government default and not any risk relating to project characteristics. Thus, the rspt
offered delayed full loss offset that was financially equivalent to immediate full loss
offset under a pure cft.
Table 4 presents a stylized illustration of the rspt applied to our two-year project in table 1. The $1,000 year 1 expenditure is written off for rspt purposes over
the two years of the project. The tax rate is set at 50 percent (although the proposed
rspt was to apply at 40 percent). The written-down value (tax value) of capital expenditure and annual rspt losses both attract an annual uplift at 5 percent ltbr.
Consequently, instead of a $500 year 1 cash rebate followed by a $600 tax payment
under the cft in table 1, the project pays rspt of $75 in year 2.
To better appreciate the operation of the rspt and to properly assess project
viability, the resulting post-rspt cash flow of the project is split in the last two columns of table 4 between
(2015) 63:2
Table 4 Two-Year Project, Stylized 50 Percent RSPT Design
(CFT with Delayed Loss Offset)
Pre-tax
Tax
RSPT
value
writeoff uplifta
Loss
upliftb RSPTc
PostRSPTd
Post-CFT Cash flow
(risky)e (risk-free)f
dollars
Year 1 . . . . . . . .
Year 2 . . . . . . . .
NPV @ 20% . . . .
NPV @ 5% . . . . .
IRR . . . . . . . . . .
−1,000.00 500.00
1,200.00 500.00
0
142.86
20%
25.00
525.00
75.00
−1,000.00 −500.00
1,125.00
600.00
−500.00
525.00
0
71.43
−62.50
0
−62.50
71.43
12.5%
20%
5%
LTBR = long-term government bond rate.
RSPT = resource super profits tax.
a Prior-year written-down value × 5% LTBR.
b Prior-year tax loss ($500 in year 1) uplifted by 5% LTBR.
c 50% tax rate × $1,200 net receipts − $1,050 ($500 annual writeoff + $25 tax value uplift + $525
uplifted prior loss), which discounts back to year 1 to $1,000 (year 1 deduction under a pure CFT)
with discounting at 5% LTBR, reflecting government guarantee of a cash rebate if necessary.
d Pre-tax cash flow minus RSPT payments.
e Cash flow if a pure 50% CFT is applied (50% × pre-tax cash flow).
f Post-RSPT minus post-CFT. An intermediated financial instrument offering a future governmentguaranteed payment of the audited year 1 loss × the tax rate, uplifted annually at LTBR, could
provide the project with a $500 year 1 payment.
the project’s risky cash flow (the same as in table 1) that would have resulted
under a pure cft (50 percent of annual pre-tax cash flow); and
n the remainder (the last column), an asset of the project that is “risk-free”
(subject only to systemic government risk) and equivalent to an implicit loan
to government of the shortfall in the upfront cft cash rebate, with that shortfall guaranteed to be paid with interest by government via later reduced tax
payments or cash rebates if necessary.
n
Through tax design, government has engineered the project’s provision of a loan to
it rather than issuing government bonds to pay for an upfront full loss offset.
The last column of table 4 (risk-free asset) shows that the cash flow deficiency
relative to a cft of $500 in year 1 is fully recovered via rspt payments in year 2 that
are lower by $525 ($75 versus $600 under a pure cft). The project has to wait a
year to recover the deficiency, but the wait is compensated for by interest at 5 percent ltbr. The 5 percent ltbr appropriately reflects government’s standing behind
the compensation should the project produce insufficient net receipts to absorb
carried-forward and uplifted losses. At the extreme, in the event that this project
produced no net receipts and closed down in year 2, the project would attract a cash
rebate of $525 (a $1,050 rspt loss multiplied by 50 percent), financially equivalent
to a $500 cash rebate in year 1.
It would be misleading financially for investors to discount the cash flows of the
risk-free asset in table 4 using a risk-weighted discount rate. Such faulty analysis,
using a 20 percent discount rate applied to aggregate post-rspt cash flows of the
project, produces an npv of −$62.50 (and an associated return of 12.5 percent),
wrongly suggesting that the rspt has turned the marginal project into an unviable
one. The last column of table 4 shows that this −$62.50 npv arises solely because
the risk-free component of overall cash flow is discounted at 20 percent rather than
a risk-free 5 percent. Discounting these risk-free flows correctly at 5 percent must
result in zero npv.
The rspt design focused on achieving a tax-neutral impact on investment decisions. The design lent itself to simple explanation in the form of cft outcomes like
that in table 1 (with an accompanying explanation of some fully compensated delays). Nevertheless, as noted above, implementation of the rspt was not pursued.
Sources and forms of opposition to the rspt varied. While the mineral resources
sector had expressed some sympathy toward profit-based imposts over those based on
volume or value, its strong opposition may have resulted in part from surprise at the
proposed 40 percent rate of tax and book value starting base. At the same time, some
with prospective ventures complained that existing projects were to receive preferential treatment. People generally had trouble understanding how the rspt worked
(a problem that was perhaps exacerbated by the number of loose ends in the design
that remained to be resolved and a paucity of clear hypothetical examples), and
some considered it “a bit over-engineered.” Many people erroneously saw the ltbr
as a measure of when the tax kicked in, noting that they themselves could earn a
similar return on their own regular savings. The potential for engineering by the
financial sector of the implicit loans to government (the last column of table 4) into
government-guaranteed financial instruments was portrayed by some as being an
unworkable key design feature, rather than a potentially convenient side effect.
Some pointed to what they saw as the destructive nature of the tax, a view that was
typically based on faulty application of risk-weighted discount rates to aggregate
post-rspt cash flows as shown in table 4. Even though cash rebates are similar to
consumption tax refunds, some questioned whether the government would actually
honour its commitment to provide them. Moreover, the preparedness of the government to sign up to full loss offset was viewed by some, not as the path to tax-neutral
design, but rather as an allocation of their taxes to failed mining operations undertaken by unreliable or risk-preferring operators.
RSPT and Income Taxation
Proposed income tax treatment of the rspt (applied before income tax with cash
rebates assessable and rspt payments deductible for income tax purposes) attracted
little attention. Conceptually, however, the income tax treatment of implicit loans
to government embedded in a cft with delayed full loss offset determines whether
the ltbr uplift should be applied at a before- or after-income-tax level.
Table 5 shows ideal income tax treatment of cash flows of our two-year project
subject to cft with delayed loss offset (matching that in table 4) applied before
(2015) 63:2
income tax. Despite delayed full loss offset, the ideal income tax treatment of the
risky component of project cash flows reflects that in table 2. In practice, this means
that pre-tax capital expenditure would be reduced by the cft rate (as if a pure cft
were operative) before applying income tax depreciation, and pre-tax net receipts
would also be cut by that rate before being included in taxable income (similar to
the situation with value-added consumption taxes). Thus, ideal income tax treatment sees the $1,000 year 1 capital expenditure cut by the 50 percent cft rate and
written off in year 2 against $1,200 of net receipts, also cut by 50 percent, to produce $100 of taxable income on which income tax of $30 is paid.
If income tax assessment stopped there and if $75 of rspt (table 4) and $30 of
income tax were simply taken from pre-tax cash flows to assess project viability, the
project would appear unviable to an investor using a 20 percent pre-tax discount
rate and facing a 30 percent income tax. That investor would compute an npv after
income tax of −$39.47, using a 14 percent discount rate, and a return of 9.5 percent
(compared to the 14 percent return that the investor could earn after income tax
from a financial asset with similar risk to that of the project).
But table 5 shows that this −$39.47 npv comes entirely from the combination
of inappropriate mixing of the risk-free asset (an implicit loan to government) with
overall project cash flows and no income taxation being applied to that risk-free
asset. The implicit loan alone has a pre-tax npv of −$39.47 with discounting at
14 percent. Investors facing a 30 percent rate of income tax, however, would discount
flows associated with implicit loans to government at 3.5 percent [5 × (1 − 0.3)%],
not 14 percent. It is therefore necessary to subject the implicit loan to income tax
(with consequent payment of $7.50 in income tax in year 2) in order for the loan’s
$500 upfront face value to be maintained after income tax with discounting at 3.5 percent. The loan then provides the same 3.5 percent return after income tax that the
investor with a 30 percent income tax rate could make on 5 percent government
bonds.
Splitting the project’s cash flows into risky and risk-free components and subjecting the split flows separately to income tax as in table 5 results in the project’s risky
flows achieving a 14 percent return (remaining marginal) and the risk-free asset a
3.5 percent return.
Should the implicit loan not be subject to income tax, a tax-neutral outcome would
require the uplift rate to be set at post-income-tax ltbr (raising complications regarding differing investor tax rates). Applying a 3.5 percent uplift rate, instead of
5 percent, to our two-year project would see $17.50 tax value uplift and $517.50 loss
uplift in table 4, resulting in rspt payments increasing to $82.50 (a $7.50 increase
matching the loss in income tax from not taxing the risk-free asset) and putting
year 2 post-rspt flow at $1,117.50. While risky project flows would remain as in
table 5 after income tax, the flow of the risk-free asset in year 2 after all taxes would
be $517.50 ($1,117.50 − $600.00 from the computations in table 4). Thus, the riskfree asset viewed separately again has a post-tax return of 3.5 percent, consistent
with what our investor with a 30 percent income tax rate can get from the government bond market.
0
20%
NPV @ 20% . . . . . . . . . . . . . . NPV . . . . . . . . . . . . . . . . . . . IRR . . . . . . . . . . . . . . . . . . . . 30.00
Income
tax a
14%
0 @ 14%
−500.00
570.00
Post-tax a
a From table 2.
b Capital component of implicit loan or its annual change in value.
c Income component of loan.
−500.00
600.00
Year 1 . . . . . . . . . . . . . . . . . . Year 2 . . . . . . . . . . . . . . . . . . Risky
cash flow
Risky project: income taxation
9.5%
−87.50
−39.47 @ 14%
−1,000.00
1,095.00
Flow after
all taxes
5%
−62.50
−39.47 @ 14%
−500.00
525.00
dollars
Risk-free
asset
500.00
Writeoff b
25.00
Taxable
incomec
7.50
Income
tax
Risk-free asset: income taxation
Table 5 Ideal Income Tax Treatment of Two-Year Project Subject to CFT with Delayed Loss Offset
3.5%
0 @ 3.5%
−500.00
517.50
Post-tax
(2015) 63:2
Such considerations—which apply equally to a cft with delayed loss offset applied
after, rather than before, income tax—put an even greater premium on taxpayers’
clear understanding of the design of the tax. Investors would need some convincing
that paying more cft via a lower uplift rate was justified because no income tax
would be applied on implicit loans to the government that the investors were not
asked to provide. Moreover, arguments over the justification for applying income
tax to implicit loans would inevitably be set against the reality that the lack of tax on
implicit loans would not be the only feature of income tax treatment of mining
operations at variance with tax neutrality. It is not hard to imagine cries that investors are not even allowed to earn the government bond rate before cft kicks in.
Ta x R e v e n u e R i s k
A pure cft does not involve implicit loans to government. A cft with delayed, but
guaranteed, full loss offset invokes an implicit loan to government, ideally with an
interest (uplift) rate reflecting the systemic risk of government default and the income tax treatment of the loan.
What happens to this framework in circumstances where government’s concern
about budget-planning risks shifts to concern about potential revenue risk, and that
concern results in government not backing immediate or delayed full loss offset?
Traditionally, in these circumstances (as in the case of a pure cft), immediate writeoff
of all expenditures is allowed for resource rent taxation purposes (so that negative annual cash flow matches the tax base), and (as in the case of a cft with delayed full loss
offset) an interest rate applies to tax losses carried forward. But (in contrast to a cft
with delayed full loss offset) investors face the real possibility that losses uplifted and
carried forward will be lost in the event of insufficient net receipts to absorb them.
An implicit loan to government is still involved, but with the open-ended risk
that the loan will never be repaid. Against such possibility, attempting to isolate cash
flows of such risky implicit loans and to apply income tax to those flows would not
be practicable. The focus of design shifts to the level of the uplift, or “threshold,”
rate to apply to losses being carried forward. Such design might be regarded as a
traditional resource rent tax (rrt) and is the design underpinning the Australian
government’s prrt and mrrt (where the only specific income tax provision applied
is to allow income tax deductions for rrt payments).
Overarching cft concepts make it clear that the level of the threshold rate may
be viewed as a general measure of the risk of not achieving the equivalent of full loss
offset under the cft—or, more simply, the risk of losing the value of available rrt
deductions. Selection of the threshold rate has, however, often been linked more to
the level of project risk typical of the industry or sector subject to rrt than to the
risk of losing rrt deductions. Certainly, where rrt is applied to individual ringfenced projects, the payback of implicit loans can be heavily dependent on the
characteristics of individual projects, particularly marginal ones. The risk of losing
rrt deductions is also likely to decline in concert with project risk as the development of the project progresses. But, even under tight project-based design, with
some highly profitable projects it may be clear from the outset that there is little or
no chance of unutilized rrt losses.
The risk of unutilized rrt losses faced by individual projects and the unique
project risk faced by those projects are two very different things.
Figure 1 indicates the nature and extent of this difference. It shows the spread of
possible npv outcomes of a marginal stand-alone project subject to rrt. Instead
of a risk-weighted discount rate measuring a single project npv, the npv of each
possible outcome of the project (like the one outcome for our two-year project in
table 1) is determined using a risk-free discount rate to avoid double-counting for
risk. Before rrt, the potential investor judging the investment in figure 1 as marginal sees positive expected npv just balancing the spread of possible outcomes, with
that spread including a significant percentage of npv outcomes below zero npv.
If the pre-rrt npv schedule in figure 1 were shifted to the right so that there
were no npv outcomes below zero, project risk (the spread of possible project outcomes) would remain unchanged. Despite unchanged project risk, there would then
be zero risk of losing rrt deductions regardless of the npv outcome (the rrt
threshold set at a risk-free rate then appropriately reflecting the fact that rrt deductions will always be absorbed before the project’s end). Post-rrt outcomes of
the project would be equivalent to those under a cft, in discounted terms. The
dichotomy between project risk and the risk of losing rrt deductions is further
underscored by the fact that rrt design may reduce the risk of losing deductions
while project risk remains unchanged—for example, by applying rrt on a taxpayer
or company basis and thereby allowing rrt losses from one project to be offset
against rrt profits from other projects of the same taxpayer or company.
Return now to unchanged figure 1, which, more realistically, has the left tail of
its pre-rrt schedule going below zero npv. In these circumstances, rrt imposes an
asymmetrical effect on the project’s npv probability distribution. The asymmetry
arises because while profitable outcomes attract cft-equivalent treatment, unprofitable outcomes do not. At the extreme of unprofitable outcomes, when no positive
annual positive cash flow is ever realized, no loss offset at all is provided. Such a
situation would arise where no net receipts at all in year 2 resulted from the $1,000
expenditure outlay in year 1 of our two-year project. Thus, negative post-rrt npvs,
instead of following the dashed schedule in the figure (which would be the case
under a cft), approximate the “a” to “b” part of the pre-rrt schedule where npv is
unchanged by rrt—where no rrt is paid and no recompense given for unutilized
rrt losses. No amount of extra annual uplift applied to those losses will provide any
direct compensation for the unutilized losses (but extra uplift would have an impact
on the upside). Relative to the investment neutrality of cft design, distortions
imposed by the lack of full loss offset under rrt design cannot be simply and neatly
offset by imposing a generally applicable loading above ltbr on the threshold rate
applying to losses carried forward—including a loading based on supposedly typical
project risk and associated risk-adjusted discount rates. Against the illustration in
figure 1, an rrt threshold rate above ltbr may be viewed as an attempt at a secondbest offset to probability distributions skewed by the RRT.
(2015) 63:2
FIGURE 1 NPV Probability Distribution for a
Marginal Project Pre- and Post-RRT
Marginal project
post-RRT
Probability
(%)
Marginal project
pre-RRT
a
b
0
NPV
RRT = resource rent tax.
Both the prrt and the mrrt incorporate additive (rather than multiplicative)
loadings above ltbr. The project-based prrt started out with a general threshold
rate of ltbr plus 15 percentage points (ignoring the implications of future reductions
in ltbr at a time when the bond rate was around 15 percent). Dated exploration
expenditure attracted a threshold rate of ltbr plus a gross domestic product deflator (and from 1990, exploration expenditure losses could be transferred to other
projects). prrt losses created by a project’s eligible closing-down expenses attract a
refund of the project’s prior prrt payments equal to 40 percent of the losses (the
prrt rate), thus removing the risk of losing the value of available deductions for
closing-down expenditure (a provision announced without any associated change to
previously announced prrt threshold rate).
The threshold rate for mrrt was set at ltbr plus 7 percentage points (when
ltbr was about 6 percent). mrrt losses (other than starting base losses) from one
project of a taxpayer could be transferred to mrrt profits of other same-commodity
projects of the taxpayer (subject to a common ownership test).
An imposed uniform loading on ltbr for the rrt threshold rate inevitably influences investment decision making in various ways. A high threshold rate, for example,
provides encouragement for gold-plated expenditure and delayed production to
extend loss carryforward artificially. From July 1990, the prrt’s 15 percentage point
threshold rate loading above ltbr was reduced to 5 percentage points for general
project expenditure.
The potential for distortions can be exacerbated by the positioning of rrt taxing
point along the sequence of extraction through to sale. Expenditures to bring product to taxing point are included in the rrt net, and product is priced at taxing point
for rrt purposes.
Where taxing point is upstream in the production sequence from the point at
which product can first be priced, mechanisms are required to reconstruct product
price from pricing point back to taxing point. Such mechanisms may include the
application of a specified return to capital downstream of taxing point to pricing
point, and perhaps also upstream of taxing point. Administrative and compliance
complexities aside, investment distortion possibilities are expanded by differences
between the risk of losing rrt deductions, the rrt threshold rate, and the specified
return to capital. prrt design avoids these added complexities by seeking to align
taxing point and pricing point. In contrast, mrrt design had taxing point focused
around run-of-mine stockpile, which could often be upstream of the first point of
commercial pricing of product.
R e g i o n a l G o v e r n m e n t R o ya lt i e s
For countries like Australia with a federal structure and where mineral resources are
assumed to be owned by the community, royalties imposed by regional jurisdictions
are a significant issue. Typically, royalties imposed on the extraction of mineral resources by Australia’s regional jurisdictions (states and territories) are based on
volume or value (ad valorem) of production. Such traditional royalties apply regardless of the profitability of particular operations.
Case-by-case negotiations and adjustments of royalty levels for changing circumstances may occur. Nevertheless, beyond sovereign risk effects of such ongoing
government involvement, investment distortions are inevitable. Similarly to a traditional rrt, royalties that are based on value or volume take from the upside without
giving back to the downside. Moreover, absent sensitive case-by-case adjustment,
royalties offer the prospect of also taking from the downside. To illustrate project
assessment using risk-weighted discounting, in table 6 a 10 percent ad valorem royalty is applied by a regional government in a country to the annual net receipts of
our two-year project.
For the investor using a 20 percent risk-weighted discount rate, the royalty makes
the project, marginal before the royalty, unviable with an npv of −$100. That negative impact of the royalty flows through to the investor’s post-income-tax assessment
of the project (even if royalties are deductible for income tax purposes).
If now the central government of the country imposes a pure cft on the same
project but ignores the royalties entirely, the assessment by the investor (now the
meat in the sandwich between two levels of government) will, by chance, stay at
−$100. While the −$1,000 year 1 expenditure is reduced to −$500, year 2 net receipts after royalties and cft become $480, which discounts to $400 in year 1. Not
a great deal would change if, more realistically, royalties were allowed as a deduction for cft purposes, just like other costs of operation. With a 50 percent cft, the
investor would compute a post-royalty cut in npv by 50 percent (both post-royalty
(2015) 63:2
Table 6 Ad Valorem Royalty of 10 Percent on Net Receiptsa
Pre-tax
Royalty
Post-royalty
dollars
Year 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . −1,000−1,000
Year 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 1,200
120
1,080
NPV @ 20% . . . . . . . . . . . . . . . . . . . . . . . 0
−100
IRR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20%
8%
a Net receipts are gross receipts less operating costs. Ad valorem royalties are often based on
gross receipts.
−$1,000 and $1,080 being cut by 50 percent), but the project would still be unviable
at npv of −$50.
In Australia, the design of the original prrt did not have to grapple with the
treatment of regional government royalties because such royalties did not apply to
the offshore petroleum projects subject to prrt.
The rspt, with its proposed broad coverage of onshore and offshore mineral
resources, incorporated a bold design that compensated mineral resource operations
directly for their royalty payments to regional governments. Thus, in table 6, the
regional government would still receive $120 of royalty payments, but the Australian government would fully compensate the project. For existing projects subject to
rspt, this meant that starting bases built around the aim of market value would
have logically been less by an amount equal to the upfront value of the expected
stream of future royalties at the time of the rspt announcement—perhaps providing
some justification for the use of book value rather than market value in the announced rspt design.
With projects fully compensated for royalty payments, project viability would be
unaffected by distortive value- or volume-based royalties. The rspt was to achieve
this compensation via credit for royalties against rspt payments, with immediate
cash refunds for excess annual credits. The royalties would have had nil net effect
for income tax purposes with royalty payments being deductible and their matching
compensation being assessable for income tax purposes. Royalty credits were to be
limited to royalty increases scheduled at the time of the rspt announcement plus
“appropriate” indexation factors (to avoid revenue siphoning by regional government). Nevertheless, regional governments would still have been able to increase
royalties beyond such levels unilaterally, with a negative impact on mineral resource
operations.
In contrast, Australia’s mrrt, which incorporated an operative 22.5 percent tax
rate, required excess credits to be carried forward to be offset against future mrrt
payments. (Excess royalty payments attracted uplift at the mrrt threshold rate via
conversion of royalty credits into deduction equivalents by multiplying the credits
by 100/22.5.) At the extreme, however, there may be no rrt payable at all. Then,
because there were no refunds for excess credits, the full distortive effects of royalties would again be revealed. To illustrate, if rrt with a 20 percent threshold rate
applied to the project in table 6, no rrt would be payable, and project cash flow
after paying royalties and rrt would be as shown in the table. Moreover, mrrt
crediting arrangements applied regardless of the extent to which regional governments later increased royalties.
A low tax rate for a central government’s resource rent taxation coupled with
rrt-style carryforward of excess credits (as with the alternative of deductions for
royalties) can result in domination by regional government royalties and their distortive effects, leaving resource rent taxation and its associated extra tax revenue to
perform a topping-up role.
Future Polic y Direc tion
The design of resource rent taxation by government can follow that of a pure cft,
a cft with delayed full loss offset, or a traditional rrt. The design could also be a
mixture of these, such as a pure cft applying at the exploration and proving-up
stages; a traditional rrt during the development and production stages with a low
threshold rate, given the low risk of losing the value of deductions; and a pure cft
at the closing-down stage. A key aim is to achieve a design that is more efficient (less
distortive of investment and production decisions) than volume- or value-based royalties. Design that is seen to be stable over the long term is an important part of that.
Take a country like Australia, with community ownership of mineral resources,
a federal system, and traditional royalties imposed by regional governments in that
federation. Should conditions for the achievement of desirable design from an economic perspective come together, out of the large range of possible ways to design
resource rent taxation imposed by central government, one design ticks a number of
boxes in terms of efficiency improvement and the chance of being implemented. That
design is a pure cft applied by central government wherever traditional regional
government royalties currently apply, with the cft rate being set at a level expected
to deliver tax revenue similar to the revenue from existing royalties. Existing royalties would be abolished through negotiations between the two levels of government
resulting in regional government being compensated directly by revenue transfer
from central government. Since expected overall tax revenue from existing projects
would be unchanged, there would be no starting bases needed for existing projects.
In addition, no implicit loans to government would be involved, with the associated
complexities regarding income tax treatment of such loans. Greater focus on forecasting budgetary effects from industry investment surveys would seek to minimize
risks around budget planning.
Intergovernment negotiations would ideally include agreement for all exploration permits to be allocated via cash bidding (reflecting expected economic rent),
removing the encouragement given to excessive expenditure from work program
bidding. Net tax revenue beyond cash bids received from future developments
would depend on upside in profitability not expected at exploration time. Government would be seen to be facing both downside and upside risks to cft revenue and
(2015) 63:2
passively relying on the commercial nous of industry players to come out ahead.
The possibility of limited future tax revenue beyond cash bids highlights the need
for simplicity of compliance and administration, which a pure cft offers (particularly if applied on a company rather than a project basis).
The ranking of profitable prospects would likely be little changed (with npv
under the cft matching that of royalties). But, crucially, better allocation of investment funds would be expected, including into marginal operations unprofitable
under royalties, as well as more finely tuned recovery of minerals around the end of
projects’ lives.
A pure cft substituting for royalties lends itself to simple explanation through,
say, the use by government of simple examples such as tables 1 and 2—expanded as
necessary to realistic mineral resource projects. Inevitably, however, there are risks
with respect to the acceptability of design and the stability of design over the long
term. Ironically, one design risk based on Australia’s experience with the rspt is the
concern that taxpayers generally may have with immediate cash rebates provided for
unutilized negative cash flow. A crucial design feature for the achievement of efficiency to one person is a waste of taxpayer funds to another. A risk to design stability
is the likely pressure that would arise for increased cft rate in the event of unexpected boom conditions in the mineral resource sector accompanied by high
commodity prices and increasing profitability.
Policy Forum: Taxation of Machinery and
Equipment and Linear Property in Alberta
Brian Conger and Bev Dahlby*
Précis
Des municipalités de l’Alberta ont amassé 1,75 milliard de dollars de la taxation sur les
machines et le matériel et les biens linéaires (melp) en 2013. Les taxes sur les melp sont
des frais fixes qui réduisent les entrées de fonds et augmentent le coût des nouveaux
investissements dans les projets de sables bitumineux, la mise en valeur conventionnelle
ou non des ressources pétrolières et gazières, et les pipelines. La distribution de l’assiette
fiscale relative aux melp est fortement concentrée parmi quelques arrondissements
municipaux et municipalités spécialisées, les 10 principales représentant 56 pour cent des
113,7 milliards de dollars de cotisations de péréquation en 2013. Quatre municipalités ont
perçu plus de 10 000 $ par habitant en taxes foncières en 2013, dont le Municipal District
of Opportunity, avec une population de 3 061 résidents, a reçu le montant le plus élevé,
soit 21 329 $ par habitant. En plus des disparités des cotisations par habitant, certaines
municipalités ont tiré profit de la présence d’investissements fixes, en un lieu déterminé,
relatifs aux melp pour imposer des taux non résidentiels par mille très élevés, afin de
percevoir des revenus de taxe supplémentaires. Les politiques provinciales concernant la
taxation sur les melp devraient être revues, étant donné la concentration de recettes
fiscales tirées des melp dans quelques municipalités et de l’incidence négative que ces
taxes peuvent avoir sur l’investissement dans l’industrie pétrolière et gazière.
Abstract
Municipalities in Alberta collected $1.75 billion from the taxation of machinery and
equipment and linear property (melp) in 2013. melp taxes are fixed charges that reduce
cash flow and increase the cost of new investments in oil sands projects, unconventional
and conventional oil and gas developments, and pipelines. The distribution of the melp tax
base is highly concentrated among a few municipal districts and specialized municipalities,
with the top 10 municipalities accounting for 56 percent of the $113.7 billion in equalized
melp assessment in 2013. Four municipalities collected more than $10,000 per capita in
municipal property taxes in 2013, with the Municipal District of Opportunity, population
3,061, receiving the largest amount, $21,329 per capita. In addition to the disparities in
the per capita assessments, some municipalities have taken advantage of the presence
* Of the School of Public Policy, University of Calgary (e-mail: [email protected];
[email protected]).
 487
(2015) 63:2
of fixed, location-specific, melp investments to impose very high non-residential mill
rates, so as to collect additional tax revenues. Provincial policies regarding melp taxation
should be reviewed, given the concentration of melp tax revenues in a few municipalities
and the negative impact that these taxes can have on investment in the oil and gas
industry.
Keywords: Tax policy n rural development n property taxes n pipelines n oil industry n
municipal finance
Contents
Introduction
Background
The Relative Importance of MELP to MDs and SMs
Disparities in Property Tax Revenues
MELP and Competition: Rich Municipal Districts, Poor Towns
Aggressive Mill Rate Setting
MELP as Benefit Taxation
Reform Options
488
488
490
493
493
496
496
498
Introduc tion
While public attention on the taxation of the oil and gas industry in Alberta has
focused on the royalty regime, a longstanding issue that is gaining increasing importance is the taxation of machinery and equipment (m & e) and linear property (lp)—
together referred to as melp—by municipalities across the province. Although these
are not industry-specific property taxes, they generally apply to refineries, pipelines,
and associated industries of oil and gas producers. These property taxes have become a source of concern for the industry1 because they add to the costs of production
and reduce competitiveness. They also raise important questions concerning the
role of municipal governments in the taxation of non-residential property.
Ba c kg r o u n d
In Alberta, municipalities include cities, towns, villages, and summer villages; rural
municipal districts (mds); and specialized municipalities (sms) that combine urban
and rural communities under a single municipal council. In total there are 18 cities,
107 towns, 93 villages, 51 summer villages, 64 mds, and 5 sms.2 In addition to these
1 Canadian Association of Petroleum Producers, “Alberta Municipal Competitiveness Report
and Recommendations,” 2014.
2 The 5 SMs are the Municipality of Crowsnest Pass, the Municipality of Jasper, Mackenzie
County, Strathcona County, and the Regional Municipality of Wood Buffalo. The latter 2 are
especially significant because they incorporate the urban service areas of Sherwood Park and
Fort McMurray that would otherwise qualify as cities. Because of their importance in levying
property taxes on M & E and LP, we have lumped the SMs and the MDs together in the
analysis of property taxes in this article.
policy forum: taxation of m & e and lp in alberta  n  489
municipalities, there are several other local authorities including 8 improvement
districts, which are directly administered by the Ministry of Municipal Affairs; 3 special areas, governed by a single Special Areas Board appointed by the lieutenant
governor in council; 8 Metis settlements, governed by the Metis Settlements General Council; and 48 First Nations (Indian) reserves on federal lands. Hamlets and
urban service areas are not considered municipalities, since they are legislatively
subordinate to their respective municipalities.
As stated in Alberta’s Municipal Government Act (“the mga”), a municipality
has three purposes:
(a) to provide good government,
(b) to provide services, facilities or other things that, in the opinion of council, are
necessary or desirable for all or part of the municipality, and
(c) to develop and maintain safe and viable communities.3
Municipalities have the authority to levy various taxes, of which property taxes
constitute the primary source of municipal own-source revenues. The mga sets out
two types of valuation standards for property taxes: a regulated procedure and a
market-value-based procedure. Railways, farmland, m & e, and lp are defined as
“regulated” property and valued using regulated rates provided annually in the municipal affairs minister’s guidelines. For all other types of property (such as residential,
commercial, and industrial properties), market-value standards apply. Municipalities are responsible for preparing assessments for all property, with the exception
of lp assessments, which are the responsibility of Municipal Affairs.
m & e property includes materials, devices, fittings, installations, appliances, and
apparatus that form an integral part of manufacturing, processing, coal and oil sands
excavation and/or transportation (that does not fit within the definition of lp); telecommunications; and electric power systems. m & e is assessed at 77 percent of its
value. lp includes electric power systems whose rates are controlled or set by the
Alberta Utilities Commission; street lighting and telecommunications systems; and
pipelines.
Municipalities set two different mill rates on the assessed value of their property—
a residential rate that applies to residential property and farmland, and a non-residential
rate that applies to m & e, lp, railways, and other non-residential property. All municipalities in Alberta impose a higher mill rate on non-residential property than on
residential property.4 In addition to these municipal property taxes, the provincial
3 Section 3 of the Municipal Government Act, RSA 2000, c. M-26, as amended.
4 This is true in the cities as well as the MDs and SMs. For a discussion of the ratio of nonresidential to residential mill rates in Edmonton and Calgary, see Melville McMillan and Bev
Dahlby, “Do Local Governments Need Alternative Sources of Tax Revenue? An Assessment
of the Options for Alberta Cities” (2014) 7:26 SPP Research Papers 1-32 (http://policyschool
.ucalgary.ca/?q=content/do-local-governments-need-alternate-sources-tax-revenue-assessment
-options-alberta-cities).
(2015) 63:2
government levies an education property tax,5 which the municipalities collect and
“relinquish” to the province. In 2014, the education property tax mill rate was 2.53
for residential and farmland property and 3.72 for non-residential property.
While this article focuses on the taxation of melp in Alberta, assessments are
made on a variety of oil and gas industry properties in the western provinces. In
British Columbia, this includes the assessment of pipelines and industrial improvements related to the production, processing, or refinement of petroleum or natural
gas. In Saskatchewan, regulated property assessment includes, among other things,
resource production equipment, heavy industrial property, and pipelines. In Manitoba, assessable improvements include plants, machinery, equipment and containers
that are used in the retail marketing of oil and oil products, and pipelines. While
m & e is taxed in all of the western provinces, the Canadian Association of Petroleum
Producers claims that m & e assessments in British Columbia and Saskatchewan are
around 25 percent of comparable property assessment in Alberta.6
T h e R e l at i v e I m p o r ta n c e o f
MEL P t o MDs a n d SMs
Table 1 shows the total property taxes collected from the six categories of property by
seven types of local authority in Alberta in 2013.7 Residential property contributed
46.4 percent of total property taxes, with 70 percent of these taxes being collected
by cities. Property taxes levied on melp contributed 22.7 percent of total property
taxes, just less than the total collected from commercial and other industrial property (referred to as “non-residential land and improvements” by Municipal Affairs)
at 29.8 percent. About 90 percent of melp property taxes were collected from the
mds and sms. Overall, mds and sms represented 18.4 percent of the province’s
population and collected 34.5 percent of total property taxes. By contrast, cities
collected 55.8 percent of the property taxes while representing 67.9 percent of the
population.
Table 2 shows the per capita property taxes by type of municipality. These figures
indicate the substantial variation in property tax revenues that accrue to the different
types of municipalities in Alberta. Both the improvement districts and the special
areas collected over $10,000 per capita in 2013. The mds and sms collected the thirdhighest level of property taxes at $3,851 per capita, which was more than twice the
per capita property taxes collected in the cities. Much of this variation occurs because
the melp taxes mainly accrue to mds and sms.
It should not be surprising to learn that the distribution of melp is highly concentrated among a few mds and sms. Of the $113.7 billion in equalized melp assessment
5 This rate is based on uniform or “equalized” property assessments in the municipalities.
6 Supra note 1, at 27.
7 All of the municipal data used in this article were obtained from the Alberta Municipal Affairs
website, “Municipal Financial & Statistical Data” (www.municipalaffairs.alberta.ca/municipal
_financial_statistical_data.cfm).
7,568,167
2,515,801
29.8
17.5
1,256,198
1,221,602
$3,568,560,919
46.4
16.4
Total . . . . . . . . . . . . . . . Percentage of total . . . . MD and SM shares
of taxes (%) . . . . . . . . $60,222,956
843,693,344
12,079,446
1,756,395
72,543
Linear
property
95.5
10.2
87.6
12.5
$787,686,151 $962,830,201
9,084,588
12,496,019
6,874,810 32,509,498
$11,265,642
752,150,278
7,469,897
840,863
73
Machinery
and
equipment
78.7
0.1
$4,710,514
150,429
8,130
$542,036
3,707,604
241,335
60,361
619
Railway
93.4
1.0
34.5
100.0
$75,766,596 $7,689,988,625
0
30,555,401
2,927,263 46,057,104
$1,366,366 $4,289,503,922
70,801,989 2,656,255,938
639,281
608,425,881
31,412
41,707,662
285
17,482,717
Farmland
0.4
0.6
55.8
34.5
7.9
0.5
0.2
Total property
taxes and grants Percentage
in place
of total
M & E = machinery and equipment.
MD = municipal district.
SM = specialized municipality.
a In the Municipal Affairs municipal financial return coding, this is referred to as “Non-Residential Land and Improvements.”
Source: Alberta, Ministry of Alberta Municipal Affairs, “Municipal Financial & Statistical Data” (www.municipalaffairs.alberta.ca/municipal_financial_
statistical_data.cfm).
$2,291,404,182
$1,716,015,895
401,953,963
156,009,357
7,220,760
120,239
$2,499,658,959
585,273,567
432,062,913
31,798,722
17,288,958
Commercial
and industriala
(excluding M & E)
Cities . . . . . . . . . . . . . . . MDs and SMs . . . . . . . . Towns . . . . . . . . . . . . . . Villages . . . . . . . . . . . . . Summer villages . . . . . . Improvement
districts . . . . . . . . . . . . Special areas . . . . . . . . . Type of local authority
Residential
land and
improvements
Table 1 Total Property Tax Revenues Collected by Municipal Governments, Alberta, 2013
981
848
928
815
3,677
585
272
673
583
335
185
26
3,527
559
Commercial
and industriala
(excluding M & E)
4
1,090
16
22
0
4,233
1,528
dollars
Machinery and
equipment
24
1,223
26
45
15
5,823
7,226
Linear
property
0
5
1
2
0
70
2
Railway
1
103
1
1
0
0
651
Farmland
1,683
3,851
1,307
1,069
3,718
14,238
10,237
Total property
taxes and grants
in place
M & E = machinery and equipment.
a In the Municipal Affairs municipal financial return coding, this is referred to as “Non-Residential Land and Improvements.”
Cities . . . . . . . . . . . . . . . . . . . . . . . . . . .
MDs and SMs . . . . . . . . . . . . . . . . . . . .
Towns . . . . . . . . . . . . . . . . . . . . . . . . . .
Villages . . . . . . . . . . . . . . . . . . . . . . . . .
Summer villages . . . . . . . . . . . . . . . . . .
Improvement districts . . . . . . . . . . . . .
Special areas . . . . . . . . . . . . . . . . . . . . .
Type of local authority
Residential land
and
improvements
Table 2 Per Capita Property Tax Revenues Collected by Municipal Governments, Alberta, 2013
(2015) 63:2
in 2013, the top 10 municipalities accounted for 56 percent of the total. Indeed, two
sms in particular had 28 percent of the total: the Regional Municipality of Wood
Buffalo, which includes Fort McMurray with a population of 116,407, had $22.2 billion; and Strathcona County, where Sherwood Park is located with a population of
92,490, had $9.8 billion. Although these two municipalities had the largest concentrations of melp, they were not among the highest in per capita terms.
D i s pa r i t i e s i n P r o p e r t y Ta x R e v e n u e s
Set 1 in table 3 shows the top 10 municipalities in terms of melp equalized assessment
per capita, led by the md of Ranchland (population 104) at $1.8 million in melp per
capita. Not surprisingly, a high per capita melp translates into high per capita municipal tax revenues. Set 2 in table 3 shows the top 10 municipalities in terms of per
capita net municipal taxes. The 6 municipalities that had the highest per capita
melp assessments in 2013 also had the six highest per capita property taxes, although
the rankings are not preserved between these two measures. Here the leader is the
md of Opportunity (population 3,061), which received $21,329 per capita in property taxes in 2013.
There is a clear geographic pattern for the top 10 municipalities in per capita
property tax revenues. Six are located north or northwest of Edmonton and correspond to areas of oil and gas industry activity. Another area of concentration is
south-central Alberta, near the Saskatchewan border. The mds of Provost and
Wainwright, and Flagstaff County are municipalities where interprovincial pipelines are located. The md of Ranchland, located in the southwest corner of the
province on the border with British Columbia, does not follow this geographic
pattern. The aggregate population of these 10 MDs is 29,637, roughly the size of the
City of Spruce Grove (29,526).
Of course, the relationship between melp assessments and total property tax revenues is not a simple mechanical one, because it depends on both the residential and
the non-residential tax rates that the municipalities adopt. Figure 1 shows that there
is a strong relationship between a municipality’s per capita melp assessment and its
per capita tax revenues. While the presence of the large outliers somewhat clouds
the relationship, a simple regression model indicates that an additional $1,000 of
melp assessment translates into an additional $11.51 of property tax revenues.8
MEL P a n d Co m p e t i t i o n : R i c h M u n i c i pa l
Distric ts, Poor Towns
The extraordinary levels of melp in some mds and the relatively low per capita property assessments in the towns that fall within their boundaries have raised concerns
about the towns’ financing of services and facilities that are used by rural populations.9
8 The 95 percent confidence interval is $9.74 to $13.28.
9 Alberta Urban Municipalities Association, “Future of Local Governance Discussion Paper”
(Edmonton: AUMA, 2009), at 16.
787,246
735,072
726,671
668,592
589,036
474,399
404,152
365,887
Opportunity (3,061) . . . . . . . Provost (2,288) . . . . . . . . . . . Saddle Hills (2,288) . . . . . . . Northern Sunrise (2,525) . . . Yellowhead (10,469) . . . . . . . Cypress (7,214) . . . . . . . . . . . Newell (7,138) . . . . . . . . . . . Woodlands (4,306) . . . . . . . . Clear Hills (2,829) . . . . . . . . Big Lakes (3,861) . . . . . . . . . Flagstaff (3,244) . . . . . . . . . . Wainwright (4,138) . . . . . . . .
Provost (2,288) . . . . . . . . . . . Ranchland (104) . . . . . . . . . . Northern Sunrise (2,525) . . . Saddle Hills (2,288) . . . . . . . Greenview (5,299) . . . . . . . . Opportunity (3,061) . . . . . . . Municipality (population)
Set 2
5,359
5,743
5,744
6,698
7,765
9,964
11,048
11,103
11,232
21,329
Per capita net
municipal
property taxes
Wainwright (4,138) . . . . . . .
Bonnyville (10,101) . . . . . . . Jasper (5,236) . . . . . . . . . . . Lamont (3,872) . . . . . . . . . . Westlock (7,644) . . . . . . . . . Two Hills (3,160) . . . . . . . . Ponoka (8,856) . . . . . . . . . . Wetaskiwin (10,866) . . . . . . Opportunity (3,061) . . . . . . Clear Hills (2,829) . . . . . . . Municipality (population)
Set 3
4.99
5.06
5.10
5.67
5.74
6.24
6.40
8.44
8.66
11.25
Ratio of
non-residential to
residential mill rate
MELP = machinery and equipment (M & E) and linear property (LP).
1,842,999
1,167,337
Greenview (5,299) . . . . . . . . Per capita MELP
equalized
assessment
Ranchland (104) . . . . . . . . . . Municipality (population)
Set 1
Table 3 Assessments, Property Taxes, and Mill Rates of Selected MDs and SMs, Alberta, 2013
(2015) 63:2
FIGURE 1 Per Capita Property Taxes Versus Per Capita MELP Assessment
Per capita property taxes (thousands)
25
20
15
10
5
0
0
500
1,000
1,500
2,000
Per capita MELP assessment (thousands)
MELP = machinery and equipment (M & E) and linear property (LP).
Source: Alberta, Ministry of Alberta Municipal Affairs, “Municipal Financial & Statistical
Data” (www.municipalaffairs.alberta.ca/municipal_financial_statistical_data.cfm).
For example, Grande Cache (population 4,319) raised $1,217 per capita and Valley
view (population 1,972) raised $980 per capita, while the surrounding md of
Greenview (population 5,299) raised $11,103 per capita in 2013. Similarly, Manning (population 1,164) raised $1,014 per capita, while the md of Northern Sunrise
(population 2,525) raised $11,048.10
The disparities between the property revenues of some mds, rich in melp, and
the towns that often provide recreation and other services for rural residents have
become a point of contention. For example, in 2012 the disparity in the assessment
bases of the town of Edson and Yellowhead County motivated the town to consider
dissolving into the county and becoming a hamlet, or possibly forming a new sm
with the county, after which the former urban and rural municipalities would be
governed by a single local government. After months of deliberation, the result was
the enactment of revenue-sharing agreements between the county and the towns of
Edson and Hinton, in recognition that
the urban centres (Edson & Hinton) contribute to the quality of life in the region by
providing a service hub for some of the social, recreational and cultural support that is
10 Opportunity and Saddle Hills only contain hamlets, while Ranchland does not contain any
urban settlements. Property tax data for Fox Creek, which is also located in Greenview, were
not available for 2013.
(2015) 63:2
required in the region and is vital to the industrial sector for its continued operations
and growth.11
This agreement ensures that rural residents have access to the towns’ programs and
facilities in exchange for 85 percent of non-residential assessment from the previous
year from a determined catchment area. The agreement with Hinton extends beyond programs and facilities to include connecting parts of the county to the town’s
water and sewer utility service. In 2012, these agreements amounted to $1.7 million
for Hinton and $4.1 million for Edson.
A g g r e s s i v e M i l l Rat e S e t t i n g
Another concern is that the municipalities with substantial melp assessments can
take advantage of the presence of these fixed, location-specific, investments to impose high non-residential mill rates to collect additional tax revenues. The average
ratio of the municipal non-residential mill rate to the residential mill rate in 2013
was 3.31. Set 3 in table 3 shows the 10 municipalities with the highest ratios of nonresidential mill rates to residential mill rates in 2013. The two highest ratios, Clear
Hills County at 11.25 and the md of Opportunity at 8.66, were among the 10 mds
and sms that had the highest per capita property tax revenues. However, the other
8 municipalities with the highest ratios were not among the municipalities with the
highest per capita assessments or the highest per capita property taxes.
To test whether melp assessments have an effect on the tax rate-setting behaviour
of the municipalities, we have regressed the residential and non-residential mill rates
on four measures of the per capita equalized assessments using cross-section data
for 69 municipalities in 2013. These results, which are shown in table 4, indicate that
a higher residential property assessment is associated with a lower residential mill
rate, but it does not have a significant effect on the municipalities’ non-residential
mill rates. Higher m & e assessments are associated with both lower residential and
lower non-residential mill rates. lp and other property assessments do not have a
significant effect on property tax rates. These regression results should be supplemented by a more detailed econometric analysis using panel data, but they indicate
that at least part of the benefit of higher m & e assessments takes the form of lower
residential property taxes.
MEL P a s B e n e f i t Ta x at i o n
The extraordinary amounts of melp tax revenues that accrue to a few municipalities
and the negative impact that these taxes can have on investment in oil and gas production, distribution, and refining means that provincial policies regarding melp
11 Yellowhead County, “Yellowhead County Revenue Sharing,” January 28, 2013, at 1
(www.yellowheadcounty.ab.ca/images/pdfs/Fact_Sheets/Fact_Sheet_Yellowhead_County
_Revenue_Sharing_2013.pdf ).
Table 4 Regressions on Residential and Non-Residential Mill Rates
Residential Non-residential
mill rate
mill rate
Constant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.884
19.188
(11.60)
(15.11)
Per $1,000 of equalized residential
assessment per capita . . . . . . . . . . . . . . . . . . . . . . . . . . . -0.0361
-0.00995
(3.85)
(0.50)
Per $1,000 of equalized linear property
assessment per capita . . . . . . . . . . . . . . . . . . . . . . . . . . . -0.001490.00108
(1.08)
(0.37)
Per $1,000 of equalized machinery and equipment
property assessment per capita . . . . . . . . . . . . . . . . . . . -0.0075
-0.0193
(2.10)
(2.45)
Per $1,000 of equalized other property
assessment per capita . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.0122 -0.0307
(1.56)
(1.83)
R 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.3892
Adjusted R 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.3510
Root mean squared error . . . . . . . . . . . . . . . . . . . . . . . . . 1.8141
0.3801
0.3413
3.8805
Note: Absolute value of t statistics in parentheses.
Source: Alberta, Ministry of Alberta Municipal Affairs, “Municipal Financial & Statistical
Data” (www.municipalaffairs.alberta.ca/municipal_financial_statistical_data.cfm).
taxation should be reviewed. The need to reform melp taxation should be considered in the context of the $7 billion “gap” in the projected finances of the provincial
government in 2015-16 if the West Texas Intermediate price of oil remains at
$50.00 per barrel. Reducing the melp tax burden may help to preserve employment
in the industry because melp taxes do not vary with firms’ profitability. melp taxes
are fixed charges that reduce cash flow and increase the cost of new investments in
oil sands projects, unconventional and conventional oil and gas developments, and
pipelines.
On the other hand, providing direct and tangible benefits from resource development projects to local communities is important in gaining local acceptance or
the “social licence to operate.” While environmental concerns should be addressed
through strictly enforced regulations, resource development projects often impose
costs on mds and sms, in particular in terms of road repairs, and residents may be
disturbed by noise, road congestion, and other nuisances. These may be soft costs,
but they are real. Employment opportunities, contracts for local businesses, and
higher property taxes are local benefits that help to offset the negative impacts of oil
and gas and pipeline projects. The tangible local benefits of oil and gas development
are among the reasons why the industry has been able to flourish in Alberta, while
similar developments in other countries, such as hydraulic fracturing in the United
Kingdom, have encountered significant local resistance. However, the amount of
melp taxes raised in some municipalities goes well beyond what is necessary to
(2015) 63:2
provide adequate local benefits, and some of the costs associated with resource development projects could be addressed in other ways.
R e f o r m Op t i o n s
The first question that should be addressed is, “Should municipal governments
impose taxes on non-residential property?” A standard response is that businesses
benefit from locally provided infrastructure and services and that non-residential
property taxation is consistent with the benefit principle of taxation. However, the
non-residential property tax burden may be much higher than any reasonable estimate of the benefits that accrue to commercial and industrial property,12 and user
charges are feasible alternatives for the use of roads, water and sewer services, and
fire protection.
Modern technology would enable mds and sms to track and charge commercial
vehicles for the use of roads, providing funds for their construction and maintenance.
The provincial government could take the lead in mandating the use of geographic
information system (GIS) technologies by commercial vehicles that would enable
the provincial and municipal governments to charge for use of highways and roads.
The province could also provide supplementary infrastructure funds for rural road
construction in areas of intense resource development.
Excessive taxation of non-residential property, from the perspective of the benefit principle, at the municipal level is likely because the burden is largely borne by
non-residents—firms’ shareholders and customers. Especially in the case of melp,
which are often location-specific investments, property taxes reduce net returns
from all projects and render some projects uneconomic. The resulting reduction in
investment and economic activity also means that some of this burden is likely shifted
to non-resident workers in the industry. As well, non-residential property taxes,
such as those levied on melp, are deductible under federal and provincial corporate
income taxes. This implies that roughly 15 percent of local melp taxation is borne
by all taxpayers across Canada, and 10 percent is borne by all residents of Alberta.
Given the negative fiscal externalities created by municipal taxation of nonresidential property, a number of reform options should be considered. First, a
uniform system of non-residential property assessment could be implemented by
a provincial agency. Currently, each municipality is responsible for assessment of
non-residential property, with the exception of lp, leading to variations in assessment practices and effective tax rates, and high compliance and administration
costs. Furthermore, the provincial government could establish a uniform mill rate
for melp assessments based on a uniform assessment procedure. This would also
reduce variations in effective property tax rates across the province. This reform
could be combined with a system of user charges for municipal roads, so that those
12 See Jack M. Mintz and Tom Roberts, Running on Empty: A Proposal To Improve City Finances,
C.D. Howe Institute Commentary no. 226 (Toronto: C.D. Howe Institute, February 2006).
municipalities with a heavy utilization of rural roads by commercial vehicles would
have a source of funding for maintenance and repair. Since the uniform rate would
be less than the very high mill rates on non-residential property that are imposed in
some municipalities, transition payments to some mds and sms with high melp assessments could be made over a five-year period.
A less radical option would be to place a limit on the mill rate that all municipalities, including cities, towns, and villages, can levy on non-residential property. Some
transition payments could be made to municipalities that currently levy mill rates
above the threshold rate. Over time, more of the municipal own-source revenue
would have to come from residential property, bringing the system of municipal finances more in line with the benefit principle. This could be part of a general reform
of property taxation in Alberta, in order to improve the transparency of municipal
financing and the accountability of municipal government.
International Tax Planning
Co-Editors: Ken Buttenham and Michael Maikawa*
Reinstated Foreign Accrual Tax and
the Multi-Period Perspective
Adam Freiheit**
Regulations 5907(1.5) and (1.6) facilitate certain deductions against inclusions for foreign
accrual property income, based on foreign tax compensation payments made between
companies in a group. This article identifies challenges in interpreting these regulations
and explores the potential significance of their enactment relative to other provisions of
the Income Tax Act and regulations that compute amounts based on foreign tax payments.
Keywords: Foreign accrual property income n foreign taxes n reasonable
* Of PricewaterhouseCoopers LLP, Toronto.
** Of PricewaterhouseCoopers LLP, Toronto. I would like to thank Michael Maikawa, Melanie
Huynh, Eric Lockwood, and Alan Davis of PricewaterhouseCoopers LLP, Toronto for their
comments on earlier drafts of this article.
 521
Personal Tax Planning
Co-Editors: Pearl E. Schusheim* and Gena Katz**
Tuition Expenses and Tutoring Fees
as Medical Expenses
Jamie Golombek, Debbie Pearl-Weinberg, and Tess Francis***
Parents raising a child with a disability, who are sometimes faced with challenges in
finding appropriate schooling that accommodates the child’s unique needs, often turn
to the private school system. The result is that tuition fees and private tutoring can cost
thousands of dollars. Fortunately, Canadian tax policy recognizes the financial burden
faced by these parents and provides some relief for the cost of schooling and tutoring
by means of the medical expense tax credit. In practice, however, it is often difficult to
claim this credit. The authors review the criteria, as determined by case law and the
Canada Revenue Agency’s published technical interpretations, for claiming tuition fees
and the cost of tutoring as a valid medical expense. They also review two other planning
solutions that could provide relief: using the preferred beneficiary election to allocate
trust income to a beneficiary with a disability and using a prescribed rate loan to split
investment income with minor children.
Keywords: Disability tax credit n medical expenses n tutoring n tuition fees n
prescribed rate loan n preferred beneficiary election
* Of Couzin Taylor LLP (allied with Ernst & Young LLP), Toronto.
** Of Ernst & Young LLP, Toronto.
*** Of CIBC Wealth Advisory Services, Toronto.
 543
Planification fiscale personnelle
Co-rédactrices : Pearl E. Schusheim* et Gena Katz**
Dépenses de tutorat et frais de scolarité
admissibles comme frais médicaux
Jamie Golombek, Debbie Pearl-Weinberg et Tess Francis***
Les parents qui élèvent un enfant ayant un handicap, qui ont parfois des défis à relever
quand vient le temps de trouver un établissement d’enseignement approprié qui
puisse répondre aux besoins uniques de l’enfant, se tournent souvent vers les écoles
privées. Le résultat est que les frais de scolarité et le coût de services de tutorat privé
peuvent atteindre des milliers de dollars. Heureusement, la politique fiscale canadienne
reconnaît la lourdeur du fardeau financier qui pèse sur ces parents et leur offre un
certain allégement pour le coût de la scolarité et du tutorat grâce au crédit d’impôt
pour frais médicaux. En pratique, toutefois, il est souvent difficile de demander ce
crédit. Les auteurs passent en revue les critères établis par la jurisprudence et par les
interprétations techniques publiées par l’Agence du revenu du Canada pour déduire les
frais de scolarité et le coût de services de tutorat comme frais médicaux admissibles.
Ils examinent également deux autres stratégies de planification susceptibles d’offrir un
certain allégement : le choix par un bénéficiaire privilégié pour l’attribution du revenu
d’une fiducie à un bénéficiaire handicapé et le prêt à un taux d’intérêt prescrit pour
fractionner le revenu avec des enfants mineurs.
Mots-clés : Crédit d’impôt pour personnes handicapées n frais médicaux n tutorat n
frais de scolarité n prêt à taux d’intérêt prescrit n choix par un bénéficiaire privilégié
* De Couzin Taylor LLP, Toronto (affilié à Ernst & Young LLP).
** De Ernst & Young LLP, Toronto.
*** De L’Équipe Services consultatifs de gestion de patrimoine CIBC, Toronto.
 565
Selected US Tax Developments
Co-Editors: Peter A. Glicklich* and Michael J. Miller**
Classification of Foreign Trusts for
US Tax Purposes: They May Be Called
Trusts, But Don’t Trust the Label
Michael J. Miller**
Keywords: United States n trusts n classifications n corporations
* Of Davies Ward Phillips & Vineberg LLP, New York.
** Of Roberts & Holland LLP, New York and Washington, DC.
 591
Current Tax Reading
Co-Editors: Robin Boadway, Tim Edgar, Jinyan Li, and Alan
Macnaughton*
With this issue, we welcome Robin Boadway as a contributing editor to Current Tax Reading.
Robin is emeritus professor of economics, Queen’s University. He is an officer of the Order
of Canada and a fellow of the Royal Society of Canada. He is a past president of the Canadian
Economics Association and the International Institute of Public Finance, and a past editor
of the Canadian Journal of Economics and the Journal of Public Economics.
Kevin Milligan, Tax Policy for a New Era: Promoting Economic Growth
and Fairness, Benefactors Lecture 2014 (Toronto: C.D. Howe Institute,
November 25, 2014), 43 pages, www.cdhowe.org
Several countries have commissioned major tax reform studies in recent years, including our Anglo-Saxon counterparts in Australia, New Zealand, the United
Kingdom, and the United States, and many have engaged in significant reforms. In
Canada, we have done neither, although there have been many piecemeal reforms
that have changed some key features of the tax system, including the institution of
the goods and services tax (gst) and refundable tax credits, substantial sheltering
of capital income, the flattening of the rate schedule, and recently the introduction
of a plethora of targeted tax expenditures. Yet, tax policy continues to be inspired by
the ideas of the Carter commission, particularly the notion of a single rate structure
applying to all income and the symbiotic relationship between the personal and the
corporate tax.
In the 2014 benefactor’s lecture, Kevin Milligan takes this perspective as a starting
point and argues convincingly that two important factors call into question the Carter
commission’s continuing legitimacy as an inspiration for tax policy. One of these
factors is the growing concentration of earnings at the top of the income distribution; the other is the documented responsiveness of taxable income to the marginal
income tax rate, again especially at the top of the income distribution. He argues that
these two factors starkly illustrate the critical tradeoff that confronts tax reformers
in Canada between efficiency and fairness. Drawing on the evolving literature on
* Robin Boadway is of the Department of Economics, Queen’s University. Tim Edgar is of the
Osgoode Hall Law School, York University. Jinyan Li is of the Osgoode Hall Law School, York
University. Alan Macnaughton is of the School of Accounting and Finance, University of
Waterloo. The initials below each review identify the author of the review.
 603
(2015) 63:2
tax reform, particularly the Mirrlees review in the United Kingdom,1 and on the tax
reform experience in the Nordic countries and elsewhere, he makes a feasible threetiered proposal for Canada that is designed to achieve both greater fairness and
stronger economic growth (at little cost to the treasury).
As background, he recounts the key recommendation of the Carter commission:2
comprehensive income as the basis for personal taxation with the corporate tax as
a necessary backstop. This recommendation was taken to be both fair (“a buck is a
buck”) and efficient because a broad uniform base discourages avoidance. Since
Carter, different approaches have evolved, challenging comprehensive income as
the ideal tax base (irrespective of the administrative difficulties of implementing
such a tax system). A decade after Carter, there was the progressive consumption tax
proposed by the us Treasury and the authors of the Meade report in the United
Kingdom,3 and later by the authors of the Macdonald royal commission in Canada,4
all recommending complete sheltering of capital income and cash flow corporate taxation. More recently, the welfarist approach, based on optimal tax analysis, found its
way into the Mirrlees review in the United Kingdom and the president’s panel in the
United States,5 among other places. The relevant emphasis in the welfarist approach
was on the preferential taxation of capital income, constrained by the efficiency consequences of high marginal tax rates.
The influence of these alternative approaches can be seen in various gradual tax
reforms, especially the sheltering of a significant proportion of capital income, the
proliferation of consumption taxes, such as the gst, and the flattening of the rate
structure. At the same time, lip service continued to be paid to Carter’s principles:
unsheltered capital income continued to be aggregated with earnings and subjected
to a common rate schedule, and the corporate tax continued to be viewed as a backstop to the personal tax. Milligan’s proposals are directed at that legacy.
To put his approach into context, Milligan documents two important economic
features of the Canadian economy. The first feature is growing income inequality. He
emphasizes the increasing concentration of income in the top income groups, especially the top 1 percent, 0.1 percent, and 0.01 percent, as reflected in the fact that their
incomes have grown much more rapidly than the income of lower-income groups
over the past 30 years. Moreover, much of this increased concentration is attributed to earnings rather than capital income. The second feature is recent empirical
1 James A. Mirrlees, Stuart Adam, Timothy Besley, Richard Blundell, Stephen Bond, Robert
Chote, Malcolm Gammie, Paul Johnson, Gareth Myles, and James Poterba, Tax by Design:
The Mirrlees Review (Oxford: Oxford University Press, 2011).
2Canada, Report of the Royal Commission on Taxation (Ottawa: Queen’s Printer, 1966).
3 Institute for Fiscal Studies, The Structure and Reform of Direct Taxation: Report of a Committee
Chaired by Professor J.E. Meade (London: Allen & Unwin, 1978).
4Canada, Royal Commission on the Economic Union and Development Prospects for Canada, Report
(Ottawa: Supply and Services, 1985).
5 President’s Advisory Panel on Federal Tax Reform, Simple, Fair, and Pro-Growth: Proposals To
Fix America’s Tax System (Washington, DC: US Government Printing Office, 2005).
current tax reading  n  605
literature on the responsiveness of reported taxable income to changes in the marginal tax rate, reflected in the so-called elasticity of taxable income. This elasticity
is particularly high for high-income taxpayers, largely owing to their greater ability
to change their taxable income through tax planning, postponement of reporting
income, evasion, and changes in behaviour.
Milligan argues that this situation challenges tax policy. The ability to deal with
a growing concentration of earnings at the top is hindered by the high tax elasticity
of top incomes, and this hindrance is exacerbated because it is impossible to differentiate among sources of income that may have different tax elasticities. Tax
systems that differentiate earnings from capital income, such as the dual income
taxes in the Nordic countries, are better placed to deal with the challenges. A dual
income tax system allows tax policy to address the earnings concentration issue by
adjusting top tax rates for earnings, while avoiding high taxes on capital income,
which is likely to be more responsive to taxes. At the same time, such a policy presupposes a desire to raise top tax rates, and Milligan makes the case that raising
these rates is desirable despite the fact that relatively little revenue would be raised
as a result. The policy also presupposes that it is desirable to have a low uniform
rate on capital income. The case for this is partly based on optimal tax arguments
that capital income should be taxed at lower rates than earnings, partly because the
amount of unsheltered capital income accruing to middle- and low-income taxpayers is minimal, and partly because uniform low taxation of capital income would be
simpler than the current system of preferential treatment arising from the dividend
tax credit and the half inclusion of capital gains. At the same time, one would have
to guard against the incentive for small business owners to report income as capital
income rather than earnings.
Given this context, Milligan makes a three-tiered reform proposal that could be
implemented sequentially. Tier 1 would “clean the tax base” to make it broader and
less susceptible to manipulation. One component of tier 1 is to reform the tax treatment of stock options by treating them as employment earnings as they accrue.
Currently, they are deferred until the option is exercised and the stock is sold, and
then they are often taxed at only 50 percent. The second component of tier 1 involves eliminating unnecessary tax expenditures, including the many “boutique” tax
credits of recent budgets.
Tier 2 would simplify capital income taxation by taxing dividends and capital
gains at common flat rates of 15 to 19 percent, as opposed to the current system of
dividend tax credits and a partial inclusion of capital gains. Although this reform
would simplify the system, it would do so at the expense of undoing the application
of a progressive rate structure. Since most taxable dividends accrue to higher-income
persons, the equity effects would be minimal. Preferential taxation of dividends and
capital gains as a substitute for imputation systems of integration is becoming a
common choice elsewhere in the world. Given the imperfections of the existing
system of integration, this is a reasonable recommendation, although, unlike the
dividend credit system, it would presumably apply to dividends received from nonCanadian corporations.
(2015) 63:2
Tiers 1 and 2 are small and piecemeal reforms. Tier 3 would be more substantial,
moving to a full dual income tax system. This would first involve taxing all unsheltered capital income at a common flat rate, a straightforward extension of tier 2. It
would then change the corporate tax from a tax on shareholder income to a tax on
rents. The most efficient way to do this would be to adopt the allowance for corporate equity (ace) system, which essentially adds a deduction for capital financed by
equity, using a risk-free interest rate. The ace system, which was recommended in
the Mirrlees review and has been adopted in a handful of countries, would eliminate
both the incentive for debt financing in the current system and the disincentive for
investment because normal returns would no longer be taxed. The final step in tier 3
would increase the taxation of the highest earnings by adding another tax bracket
above the current top one.
Milligan discusses some issues that would arise as a result of these reforms, such
as the need to harmonize provincial income tax systems, the need to ensure that
high earners could not masquerade earnings as capital income, and concerns about
the response of high earners to increased marginal rates.
Overall, the reform is feasible and imaginative. At the same time, while it redefines the architecture of income taxation for Canada, it is limited in scope and leaves
room for developments in the future. Examples of further reforms that could be
considered include progressivity at the bottom of the income distribution as well as
the top; the tax treatment of the financial sector, where rents might be sizable but
would go untaxed under a standard ace tax; the interaction between an ace tax and
natural resource taxation, which also applies to rents; the treatment of rents from
innovation, given that the patent system is expressly intended to protect inventors’
rents; the role of other major taxes such as the gst and harmonized sales tax (hst)
and payroll taxes in the post-Carter approach; and, more significantly, the potential
role of inheritance taxation to address the issue of top incomes, a reform that the
Mirrlees review emphasized.
R.B.
Tanina Rostain and Milton C. Regan Jr., Confidence Games:
Lawyers, Accountants, and the Tax Shelter Industry (Cambridge,
MA: MIT Press, 2014), 424 pages, ISBN 978-0262027137
Tax shelters marketed to high-wealth individuals threaten the health of an income
tax system in much the same manner as a virus threatens the health of its human
host. The factors that precipitate the spread of these aggressive tax-avoidance transactions are always present, but they tend to be muted by offsetting factors akin to
natural immunization agents. More often than not, the elements of the relevant substantive law that are the source of particular transactions are awakened by a strained
interpretation or application by the judiciary, which places other constraining factors
under stress, such that they can no longer contain an outbreak of transactions designed to exploit the resulting technical anomaly in the law. The sustainability of an
income tax system based on self-assessment is threatened by these shelters, especially
because of the corrosive effect on taxpayer morale caused by the perception that
high-wealth individuals have collectively renounced the accepted obligations of fiscal
citizenship and left the rest of the taxpaying population to shoulder the burden.
Under these conditions, taxpayers who do not participate in shelter activity tend to
migrate to tax-evasion strategies, and, in the extreme, the income tax based on selfassessment implodes.
The income tax of all countries is littered with these tax-shelter outbreaks. Not
surprisingly, tax policy makers and administrators, like health clinicians, have responded incrementally to each episode with the development of treatments that
build on earlier successful treatments to ultimately control the virus and maintain
the health of the income tax. The most recent and highly publicized tax-shelter
outbreak occurred in the United States from the mid-1990s to the early 2000s, and,
because of the involvement of major international accounting firms and some prominent second-tier national law firms, it was arguably the most serious episode. In
fact, the authors of this book, who are both professors at Georgetown Law School
whose research focuses on professional ethics, characterize this outbreak of aggressive tax shelters as “the most serious episode of lawyer wrongdoing in the history of
the American bar.”6 This characterization is based on the depth of the involvement
of the accounting and legal professions, which were responsible for the design and
promotion of the relevant transactions. The authors contrast this depth of involvement with, for example, the savings and loan crisis in the United States in the 1980s
and the corporate scandals in the 1990s, which were covered up by members of the
accounting and legal professions.
The design and marketing of tax shelters have tended to be driven principally by
a handful of rogue tax practitioners and promoters who operate on the unacceptable
margins of responsible tax practice.7 These marginalized individuals are seduced by
the “scalable returns” provided by tax shelters. As popularized by the author, philosopher, and mathematician Nassim Nicholas Taleb,8 these returns are characterized
by a separation of time and labour on the one hand and compensation on the other.
In effect, returns at the margin are earned with very little, if any, increase in time
and labour and are referred to as “zero-cost marginal” goods and services in the
mainstream economics literature. By comparison, “non-scalable” returns require a
commensurate additional increase in time and labour for the realization of additional returns to these inputs. Scalable returns are commonly associated with certain
services, such as securities and commodities trading, while the accounting and legal
professions are associated with non-scalable returns in the form of the billable hour
and the constraint of a 24-hour day. As Rostain and Regan emphasize, several factors
6 At 4.
7 See, for example, Michael Gillard, In the Name of Charity: The Rossminster Affair (London:
Chatto & Windus, 1987).
8 Nassim Nicholas Taleb, The Black Swan: The Impact of the Highly Improbable (New York:
Random House, 2007).
(2015) 63:2
combined to cause the most recent tax-shelter episode in the United States to migrate
into most of the major international accounting firms. They cite a lax regulatory
environment and increasingly fiercely competitive “winner-take-all” markets for
professional services.9 At the same time, a booming us economy was generating
freshly minted and wealthy entrepreneurs, who were creating and selling their new
companies. The tax bills associated with these sales transactions represented a means
to increase revenues for tax professionals at the major accounting firms and some
corporate law firms by creating and designing transactions intended to shelter the
realized gains. The attraction of scalable returns associated with the tax-shelter service
proved to be too much to resist for a considerable segment of otherwise mainstream
tax accountants and lawyers, who were keen to enjoy their share of all this new
wealth. In addition to these factors, the Internal Revenue Service (irs) enforcement
capabilities had become severely degraded, and the irs was under severe political
pressure, with alleged enforcement abuses being the subject of high-profile congressional hearings in the late 1990s.
In providing a first-rate and compelling narrative of this particular tax-shelter
episode in the United States, Rostain and Regan are careful to emphasize what they
see as the organizational features of participating accounting and law firms that made
them vulnerable to these pressures. The authors prefer to look to organizational
features rather than take a conceptually simpler focus on the behaviour of a handful
of tax professionals within participating firms who went rogue, chasing scalable
returns and rejecting their responsibilities to maintain a sustainable income tax system, while rationalizing the chase as the single-minded pursuit of the perceived
interests of their clients. It is difficult, however, to discern from their narrative
exactly what organizational features in the major accounting firms led some to institutionalize tax-shelter work and others to avoid it. It is relatively easy to see from
their narrative that a loose organizational structure in law firms is open to such institutionalization to a much greater extent than a tight organizational structure.
Even then, the singular major law firm participating in the latest shelter episode as
marketed to high-wealth individuals was a second-tier Dallas-based national firm,
Jenkens & Gilchrist, which aspired to first-tier status and saw the revenue from
shelter design and marketing as a means of realizing it.
The strength of Rostain and Regan’s work is the narrative itself. The authors provide a fascinating and compelling story drawn from an utterly exhaustive search of
the available public records. It is the rogue characters in the participating firms that
draw the reader in. Organizational features of participating firms as a possible explanation for the institutional acceptance of a tax-shelter practice tend to pale in the face of
what seems to be the force of some of the personalities within these firms. It is perhaps
not surprising therefore that the authors are extremely cautious when they posit a set
of possible responses intended to shore up professional responsibilities of tax practitioners with respect to the sustainability of the public good that is a tax system.
9 Robert H. Frank and Philip J. Cook, The Winner-Take-All Society: Why the Few at the Top Get So
Much More Than the Rest of Us (New York: Penguin Books, 1996).
One possible line of inquiry that the authors only hint at late in the book is a
comparison with the corporate shelter market, which developed at approximately
the same time as the shelter market that is their focus. The former is different primarily because of the participation of some first-tier law firms. The authors also
give only cursory treatment to the role of the judiciary as a source of tax shelters
rather than a buttress against them. In this respect, Rostain and Regan are not
alone, however; for some inexplicable reason, the judiciary in all countries seems to
get a free pass on this count. On a much more positive note, the political response
to the spread of the shelter activity is heartening. Many of the same politicians who
led the charge against the irs in the congressional hearings of the late 1990s threw
their support behind the tax administration and its enforcement capabilities, including the pursuit of criminal sanctions, in obliterating the shelter market.
T.E.
Brian J. Arnold and James R. Wilson, “Aggressive International
Tax Planning by Multinational Corporations: The Canadian
Context and Possible Responses” (2014) 7:29 SPP Research
Papers 1-74, www.policyschool.ucalgary.ca
In this research paper, the authors survey recent Canadian legislative initiatives in
the field of international taxation and the ongoing Organisation for Economic Cooperation and Development (oecd) project on base erosion and profit shifting
(beps). The authors caution that they do not engage in any sort of explicitly focused
policy analysis, although they admit that they cannot resist injecting the odd policy
observation here and there throughout the paper. The principal strength of the
paper is nonetheless its account of the recent Canadian legislative initiatives against
the backdrop of general income tax principles, case law, and pre-existing provisions
that affect the treatment of tax-structured cross-border transactions. They note the
Department of Finance’s focus on inbound direct investment, which is apparent in
recent revisions to the thin capitalization rules, the introduction of foreign affiliate
dumping (fad) rules, and the proposed anti-treaty-shopping provisions, and the
lack of any initiatives focused on outbound direct investment. They suggest that
this difference in focus is attributable to a reluctance to undermine the perceived
competitiveness of Canadian-based multinational enterprises (mnes) in the face of
tax-preferred treatment by the United States, the United Kingdom, and other
countries for their domestically based mnes. Given this non-cooperative policy environment, Arnold and Wilson conclude that legislative initiatives affecting outbound
direct investment should be developed by Canada only within the confines of a
multilateral framework, such as that provided by the beps project. Yet, they are
skeptical about the ability of this project to deliver tangible results, citing what they
see as mixed policy results for recent multilateral projects undertaken by the oecd.
They cite, for example, the harmful tax competition project as having been largely
unsuccessful in realizing its goals, while they believe that the information exchange
project has largely been a success, using the same metric.
(2015) 63:2
The authors begin with a very general description of Canada’s matrix of rules
and principles that together make up its international tax system, including the set
of rules affecting both inbound and outbound investment, as well as Canada’s tax
treaty network. This overview is presumably aimed at interested readers who are
unfamiliar with these rules. The overview is followed by a review of the distinction
between tax evasion and tax avoidance, as well as the distinction between acceptable
and prohibited tax avoidance. This review also seems to be aimed at readers who are
interested in recent international tax developments but who are not otherwise tax
practitioners, academics, administrators, or policy makers. The core of the paper
begins with three transactional case studies as examples of common tax-structured
arrangements. The case studies include one on inbound treaty shopping, another
on outbound transfers of intellectual property, and a third on outbound tower financing structures into the United States. Appendixes provide further examples,
including inbound leveraged fad, outbound investment using repurchase agreements (repos), and outbound Luxembourg financing intermediary structures. The
authors carefully review the key aspects of the Canadian income tax system that
facilitate the perceived tax-effectiveness of the transactional examples and recent
legislation intended to address certain of them, particularly in the context of inbound
direct investment. The low-hanging policy fruit in the area of outbound direct investment, which consists primarily of comprehensive interest expense limitations
and the narrowing of the lookthrough source rule in subparagraph 95(2)(a)(ii) of the
Income Tax Act,10 notably remains on the tree and facilitates the perceived taxeffectiveness of certain of the transactions highlighted in the case studies.
T.E.
David MacDonald, Income Splitting in Canada: Inequality by Design
(Halifax: Canadian Centre for Policy Alternatives, January 2014),
25 pages, www.policyalternatives.ca
MacDonald provides revenue cost and distributional impact estimates of income
splitting in Canada in three different situations:
pension income splitting, as introduced and implemented by the federal government in 2006;11
n income splitting for families with children under 18 years, with a limit of
$50,000, as introduced by the federal government on October 30, 2014;12 and
n income splitting for all families.
n
10 RSC 1985, c. 1 (5th Supp.), as amended (herein referred to as “the Act”). Unless otherwise
stated, statutory references in this article are to the Act.
11 Section 60.03 and paragraph 60(c).
12 Canada, Department of Finance, “PM Announces Tax Cuts, Increased Benefits for Families,”
News Release, October 30, 2014 and accompanying Notice of Ways and Means Motion To
Amend the Income Tax Act.
The author estimates that in 2015 pension income splitting will result in lost
revenue of $1.2 billion for the federal government and $500 million for provincial
governments. He also finds that the gains will continue to be captured disproportionately by senior couples in the richest four income deciles.
In the second situation, MacDonald estimates the federal revenue cost to be
$3 billion, which is slightly higher than the recent revenue loss estimates by the Parliamentary Budget Office of $2.2 billion.13 The provincial revenue loss is estimated to
be $1.9 billion annually. The author finds that the bottom 60 percent of households
earning $56,000 or less would receive a paltry $50 on average in the income-splitting
situation that is closest to the one introduced by the federal government, while the
top 5 percent of households with income of more than $147,000 would realize on
average a benefit of $1,100. In this group, 1 in 10 would realize a benefit of more
than $5,000.
Not surprisingly, the provision of income splitting for all families without limit
would result in the greatest revenue loss, with the benefits captured ever more disproportionately by households at the upper end of the income scale. In this situation,
MacDonald estimates that the federal revenue loss would be $7.5 billion and the
provincial revenue loss would be $4.3 billion annually. He finds as well that the top
10 percent of households on the income scale would capture 34 percent of the
benefits.
MacDonald suggests that if the federal government were genuinely concerned
with supporting families with children, a more equitable and effective policy instrument would be a universal child-care program, such as the model in Quebec, which
could reduce the cost of child care from $40 to $50 per day to around $7 per day.
Moreover, this particular policy instrument would increase female participation in
the labour force, as evidenced by the Quebec experience.
T.E.
Wei Cui, “Administrative Decentralization and Tax Compliance:
A Transactional Cost Perspective,” University of Toronto Law
Journal (forthcoming)
Cui makes a novel argument that certainly constitutes an original contribution to
the literature on tax administration in developing countries. He summarizes the
principal message of this existing literature, which is part of a larger literature on
the role of taxation and economic development generally, as concluding that the
constrained capacity of tax administrations determines the range of available policy
instruments. The relevant constraints are invariably noted, and public finance
economists dutifully return to their more familiar territory of analyzing available
policy instruments and, in particular, the need to modify their design in the context
13Tim Scholz and Trevor Shaw, The Family Tax Cut (Ottawa: Office of the Parliamentary Budget
Officer, March 17, 2015).
(2015) 63:2
presented in developing countries. Tax administration is characterized as a “black
box,” the content of which remains unexplored in any policy-relevant detail.
Cui draws on the tax administration experience in China to sketch a detailed
outline of a very different model. He observes that this model, which is characterized by the collection of taxes not according to the rules of law but according to
informal agreements between tax collectors and taxpayers, is a common phenomenon in developing countries and has its source in a decentralized administrative
structure. He defines “administrative decentralization” as the concentration of government functions at the lowest ranks of a geographically dispersed bureaucracy.
It is seen to increase communication costs associated with the implementation of
law while changing the costs of compliance for taxpayers and resulting in “semicompliant” behaviour, whereby tax is collected and remitted despite the ignorance
of taxpayers about the law. This model of administrative decentralization can persist, however, in the absence of corruption and other conventional explanations for
weak tax administration, and therefore it may explain constrained administrative
capacity in developing countries.
T.E.
Henrik Jacobsen Kleven, “How Can Scandinavians Tax So Much?”
(2014) 28:4 Journal of Economic Perspectives 77-98
Gabriel Zucman, “Taxing Across Borders: Tracking Personal Wealth and
Corporate Profits” (2014) 28:4 Journal of Economic Perspectives 121-48
Erzo F.P. Luttmer and Monica Singhal, “Tax Morale” (2014) 28:4
Journal of Economic Perspectives 149-68
Public finance economists are increasingly turning their attention to tax administration issues, including compliance, avoidance, and evasion, as evidence of growing
income and wealth inequality surface and stories of reducing tax liabilities by international tax planning proliferate. Some of the academic work in this area may be
found in a recent symposium on tax enforcement and compliance.
Henrik Kleven tackles the seemingly paradoxical observation that despite high tax
rates and a high share of gross domestic product (gdp) taken in taxes, the Scandinavian
countries are able to perform well in terms of economic activity and tax compliance
compared with other oecd countries. Tax revenues as a percentage of gdp in Denmark, Norway, and Sweden are close to 50 percent, compared with 25 percent in
the United States, while top marginal tax rates are 60 to 74 percent, compared with
43 percent in the United States. Even starker, participation tax rates (the increase
in net taxes on joining the labour force) are 77 to 87 percent in the Scandinavian
countries, compared with just 37 percent in the United States. (Canada’s numbers
lie between these two extremes.) What is it about the Scandinavian countries that
allows them to tax so much and yet achieve outcomes comparable to those in the
United States? Kleven argues that there are three contributing structural factors
and possibly other societal reasons.
The first factor is that in Scandinavia a large proportion of tax reporting involves a
third party in addition to the taxpayer and the revenue agency. Third-party information is obtained especially from employers and from financial institutions. Evidence
from a Danish tax audit field experiment indicates that tax compliance is close to
perfect when third parties are involved in reporting income, but far from perfect on
self-reported income. In Denmark, about 95 percent of income is subject to thirdparty information, and the evasion rate is only 2.2 percent. The evasion rate for those
who self-report is about 50 percent. Tax-compliance rates for those who self-report
and those who rely on third-party information are comparable in the United States,
but third-party reporting is more important in Scandinavia. These compliance rates
are confirmed by cross-country estimates that show a clearly negative relation between the tax-to-gdp ratio and the proportion of persons in “evasive” jobs.
The second factor concerns the structure of income tax bases. An important constraint on increasing tax rates is the so-called elasticity of taxable income, which is
an estimate of how taxable income responds to marginal tax rate changes, including
both behavioural responses and avoidance or tax-planning activities. The broader the
income tax base (the fewer the exemptions, deductions, and special treatments)
the smaller the opportunity to avoid taxes by tax planning, and the lower the elasticity
of taxable income. Kleven presents evidence to show that the elasticity is substantially lower in Denmark than in the United States and attributes this to Denmark’s
broader income tax base.
The third factor is more subtle. The Scandinavian countries impose relatively
high participation taxes, compared with other oecd countries, yet have high employment rates. Countries such as Canada, the United Kingdom, and the United
States have relatively generous employment tax credits (for example, the working
income tax benefit in Canada), yet have lower employment rates than in Scandinavia.
The reason, Kleven argues, is that Scandinavia provides participation incentives
through public services, such as child care, preschool care, and elderly care that are
complementary with work. In effect, there is a choice between small net tax distortions and the limited provision of child care and elderly subsidies on the one hand
and large net tax distortions and the generous provision of child care and elderly
subsidies on the other. Scandinavia has chosen the latter option and has experienced
no detrimental effect on employment.
Finally, Kleven investigates whether there are broad social and cultural influences
that can account for high taxes, good compliance, and strong economic performance
in Scandinavia. He presents evidence to show a positive correlation between the
tax-to-gdp ratio and various indicators of trust and social capital. Scandinavian
countries score highly on both accounts, and this contributes to explaining why they
are able to maintain relatively high tax-to-gdp ratios. At the same time, these correlation results are far from conclusive.
Gabriel Zucman focuses on the growing problem of corporations and wealthy
individuals evading or avoiding taxes in their resident countries by shifting profits
or wealth to low-tax countries (tax havens). He estimates that about 20 percent of
us corporate profits are shifted to tax havens, typically legally, while 8 percent of the
(2015) 63:2
global financial wealth of individuals is held in tax havens to escape taxes illegally.
He argues for a “world financial registry,” which would record information on corporate shareholders by nationality and residence. Before making this proposal, he
considers the problems of corporate profit shifting and individual offshore tax evasion in sequence.
Zucman observes that corporate taxes were introduced in high-income countries
in the early 1900s at the same time as income taxes. They were and still are viewed
as a backstop to the income tax, taxing at source shareholder income that could
otherwise be accumulated tax-free within corporations. Some countries, including
Canada, provide credit to shareholders for corporate taxes paid on their behalf using
an imputation system, whereas others, including the United States, give preferential
tax rates on shareholder income. Complications arise when corporations operate in
different countries, and different countries may have a claim on the same profits.
Three principles, which were established under the auspices of the League of Nations
to avoid double taxation, still apply today. They are that (1) corporate tax should be
paid to the source country, (2) arm’s-length pricing should apply to intrafirm sales,
and (3) bilateral agreements should be used to address international tax issues.
In the era of globalization, each of these three principles faces challenges. First,
bilateral treaties lead to inconsistent rules among countries, prompting corporations
to avoid tax by choosing the location of their affiliates (so-called treaty shopping).
Zucman’s description of Google’s “double Irish Dutch sandwich” is very revealing.
The technique of tax inversion used by us corporations is another example. Second,
arm’s-length pricing is very difficult to enforce because comparable market prices
do not exist. Corporations are able to shift profits to affiliates in low-tax countries
by transfer pricing as well as by intragroup loans. Third, source-based taxation not
only results in wasteful tax-planning expenditures in treaty shopping and transfer
pricing but also results in incentives to relocate activities to low-tax countries. The
resulting tax competition leads to reduced tax rates all around, which creates lower
tax revenues in non-haven countries.
Using national accounts and balance-of-payments statistics for the United States,
Zucman calculates that foreign profits constitute one-third of us corporation profits,
and tax havens claim 55 percent of these profits. The result, as mentioned, is that
tax havens claim 20 percent of total corporate profits, which is 10 times higher than
in 1984. Since these profits go largely untaxed in the United States, profit shifting
reduces us corporate tax revenues by 20 percent. Viewed another way, the effective
us corporate tax rate fell from 30 to 20 percent between 1998 and 2013. The fact
that corporate taxes as a share of national income have not declined reflects the
fact that corporate profits as a share of national income have risen from 9 percent
in the 1980s to 14 percent in 2013.
Various proposals for reform of the corporate tax have been offered (apart from
abolishing it, which the author regards as infeasible). One option is to harmonize
treaty rules, perhaps in accordance with the oecd’s beps action plan. A second option
is to abandon arm’s-length transfer pricing in favour of allocating profits among
countries by formula apportionment (analogous to the method used to allocate
profits among provinces in Canada). This would require significant international
cooperation and would only roughly approximate the source principle. A third option
is to abandon source-based taxation in favour of taxing profits according to where
shareholders live. This too would involve considerable international coordination,
and would neglect the fact that the corporate tax is also a tax on rents obtained by
natural resources and locational advantages of source countries. A fourth option is
for all countries to fully integrate corporate and personal taxes. This would be very
complicated in an international setting, especially because it would require governments to give credits to domestic shareholders for corporate taxes levied in foreign
jurisdictions.
Zucman suggests that progress can be made in addressing the challenges of international corporate taxation by establishing a world financial registry that records
the residence and nationality of shareholders. This idea is based on the presumption
that the objective of corporate tax policy should be to tax corporate profits earned
worldwide that accrue to domestic residents. He argues that a world financial registry would facilitate this objective by allowing countries to compare the distribution
of revenues that they receive with the ideal benchmark.
Zucman’s discussion of offshore tax evasion by wealthy individuals is more concise. He vividly illustrates how offshore banking accounts in tax havens can be used
to evade taxes. Tax evasion is notoriously difficult to measure since it does not appear
in official statistics. He estimates the amount of wealth held in tax havens indirectly
using anomalies between assets and liabilities in global investment statistics. He
found that in 2013 8 percent of global financial wealth was held in tax havens, a
finding that is comparable to other estimates and that is about 28 percent higher
than it was in 2008. Over half of this wealth is held by residents in high-income
countries, but holdings in emerging countries are growing rapidly. He estimates
that about 80 percent of wealth held in tax havens goes unreported. Recently, some
progress has been made by inducing some tax-haven countries to share banking
information with foreign governments. The impetus came from the us Foreign
Account Tax Compliance Act, which requires foreign banks to disclose accounts
held by us taxpayers under threat of sanctions. But coverage remains incomplete.
Many tax havens do not comply, and, even when they do, offshore banking can
move to non-complying jurisdictions.
Zucman maintains that a world financial registry can contribute to progress by
addressing offshore accounts. However, it is more challenging when dealing with
individuals than when dealing with corporations. Although much individual offshore
banking involves evasion, corporate tax shifting is largely legal avoidance. In addition, to be useful in dealing with individual income tax enforcement, the registry
would have to include all types of financial assets, whereas corporate tax enforcement involves only equities.
Erzo Luttmer and Monica Singhal focus on an aspect of tax administration that
is typically neglected in formal approaches to tax evasion: voluntary compliance.
The textbook approach to compliance treats tax evasion like any other risky decision. The taxpayer weighs the advantages of underreporting (lessened tax liability)
(2015) 63:2
with its costs (the anticipated loss if caught, which depends on both the probability
of being audited and the penalty). This approach ignores other motives for voluntary
compliance, including cultural or social norms, obedience of the law, guilt or shame,
peer sanctions, sense of fairness, sense of value for money from taxes, and so on.
Luttmer and Singhal place all of these non-pecuniary motives within the concept of
“tax morale” and study its consequences in relation to understanding compliance
and how to improve it and in relation to tax policy.
Rather than providing a comprehensive review of the literature, they emphasize
some suggestive observations and illustrate specific aspects of tax morale that affect
compliance. Four types of evidence are presented to support the notion that tax
morale is a significant determinant of compliance, along with enforcement activity.
First, survey evidence from the world values survey and the European social survey
reveals that respondents are overwhelmingly of the view that cheating on taxes is
not justified. Of course, this evidence can be no more than suggestive since it does
not indicate how the respondents actually behave (as opposed to how they think
taxpayers in general should behave).
Second, estimates can be made, based on economic models of decision making
under risk, about the level of compliance that one would expect, given audit rates,
penalties, and realistic assumptions about degrees of risk aversion. Even assuming
relatively high levels of risk aversion, the predicted level of compliance is far below
the level that one observes in practice. There are many caveats associated with this
kind of modelling, as well as with the estimated levels of evasion to which the model
results are compared. Nonetheless, the magnitudes of the results are highly suggestive that the standard public finance model of tax evasion is inadequate.
Third, some evidence emanates from environments in which enforcement is
minimal; therefore one would expect compliance to be minimal as well. A striking
example is given of a local church tax in Bavaria, which church members are legally
obliged to pay; payment, however, is not enforced. A significant proportion of
members nevertheless pay the full amount of taxes owing, indicating voluntary
compliance.
Finally, evidence has been obtained from settings in which enforcement is common but tax morale differs among taxpayers. Luttmer and Singhal report on a study
that compares tax compliance by foreign-owned corporations that pay tax in the
United States but that are based in countries with differing levels of corruption.
Corporations from countries in which corruption is more prevalent evade more
taxes, which is taken to be a tax morale effect.
Next, Luttmer and Singhal provide a detailed discussion about various mechanisms of tax morale and single out five. The first mechanism is “intrinsic motivation,”
which refers to matters such as honesty, civic duty, guilt, and so on. Direct evidence
of the effectiveness of this mechanism includes the church tax example cited above.
In addition, there is some evidence that priming intrinsic motivation by means of
letters of exhortation to taxpayers can have some effect. At the same time, increased
enforcement may serve to overshadow intrinsic motivation, a phenomenon that is
alleged to occur in other areas, such as voluntary compliance with environmentally
beneficial actions.
The second mechanism is “reciprocity”—that is, the willingness to pay taxes on
the basis of the perceived benefits of the state and the fairness of the tax system, or
as part of a social contract between taxpayers and the state. The limited evidence
that reciprocity is important comes from surveys of trust in government or satisfaction with government performance; randomized field studies that prime reciprocal
motivation by emphasizing the benefits of government spending of tax revenues; or
appeals to the effects of tax reforms that are perceived to be unfair (such as the poll
tax in the United Kingdom), which reduced compliance. The introduction of the
gst in Canada might provide fodder for such an investigation.
The third mechanism includes “peer effects and social influences,” including the
desire to abide by social norms or to signal one’s behaviour to others. Compliance
based on social norms is potentially volatile and therefore susceptible to influence.
For example, tax authorities may provide information about general levels of compliance to improve compliance, or they may reward compliance by public recognition.
Luttmer and Singhal summarize the mixed evidence that these interventions have
had positive effects on compliance.
The fourth mechanism is “culture”—that is, social attitudes persisting over long
periods of time and influencing compliance. The most compelling studies of the
influence of culture are those that compare compliance by taxpayers from different
cultural backgrounds in the same taxpaying environment. Some of these studies
involve laboratory experiments, while others examine taxpayers who currently reside in the same country but who have different countries of origin. These studies
tend to find greater compliance by taxpayers who come from countries with greater
compliance.
In addition to these non-pecuniary effects, compliance may be influenced by information imperfections and behavioural anomalies. Taxpayers may overestimate or
underestimate the probability of getting caught, which is not surprising since audit
procedures are not publicized. Failure to comply may reflect procrastination or
forgetfulness. Moreover, taxpayer behaviour may deviate from so-called economic
rationality. Taxpayers may attach excessive weight to small probabilities of detection, and they may be highly loss-averse.
Given these myriad ways in which compliance can be explained, apart from the
standard pubic finance account based on penalties and the probability of getting
caught, Luttmer and Singhal conclude by considering what policies might supplement or substitute for audit-penalty enforcement approaches. Examples include
nudges to taxpayers, such as information, reminders, or exhortations; overwithholding to reduce loss aversion; threat-of-audit letters; information about the benefits
of government spending and about tax compliance by peers; publication of cases of
compliance and non-compliance; and so on. Clearly, voluntary tax compliance, like
lawful socially beneficial behaviour in other spheres, is valuable to society as a
whole. How it can be preserved and enhanced remains an open question.
R.B.
(2015) 63:2
Edward D. Kleinbard, We Are Better Than This: How Government
Should Spend Our Money (New York: Oxford University Press,
2015), 544 pages, ISBN 978-0199332243
Joseph Bankman and Daniel Shaviro, “Piketty in America:
A Tale of Two Literatures,” Tax Law Review (forthcoming)
Growing income and wealth inequality over the past 30 years has moved to the top
of the policy and political agenda. As with climate change, there are some who will
deny its existence; however, the discussion concerning income and wealth inequality
appears to be now primarily in a policy discussion phase in which the relevant issues
are what, if anything, should be done to address it. The monumental historical
study of wealth inequality by the French economist Thomas Piketty has arguably
moved this discussion forward, even though the strength of his work is its empirical
dimension, detailing patterns of wealth accumulation over the past 300 years.14
Piketty’s related attempt to advance the discussion about policy instruments to address wealth disparities is less compelling than his related interpretation of the
sources of wealth (and income) disparity.
We Are Better Than This can be viewed against the backdrop of Piketty’s work as a
sort of manifesto for a modest, but arguably politically robust, welfare state. Indeed,
it provides a thoroughgoing and accessible account of the benefits and burdens of
“fiscal citizenship,” which an unchecked growth in wealth and income inequality
undermines. Taxation is, of course, one of the principal elements of fiscal citizenship;
it is therefore central to the discussion about the choice of instruments to address
the inequality that is the inevitable result of market outcomes. Yet, spending policy
is also one of the main elements of fiscal citizenship and is Kleinbard’s principal
point of emphasis. Kleinbard provides an outline of a somewhat modest policy
agenda in the us context that, unlike Piketty’s policy agenda, is not intended to
serve a redistributive function but rather is intended to ensure that the burden of
fiscal citizenship is borne in a manner that is nonetheless normatively defensible.
His criteria for such an assessment are startlingly simple but compelling. Indeed, he
avoids descending too deeply into arguments about moral philosophy, which could
detract from the accessibility of his text without necessarily adding much substance.
Kleinbard does, however, centre much of his normative discussion on Adam Smith’s
Moral Sentiments.15 As Kleinbard illustrates, a close reading of this canonical work
suggests that Smith is undoubtedly a kindred spirit.
Most importantly for both Kleinbard and Smith is the role of randomness or
blind luck in determining market outcomes, which Kleinbard cites in support of the
provision of a set of “social insurance” programs. He contrasts this position with
14Thomas Piketty, Capital in the Twenty-First Century, trans. Arthur Goldhammer (Cambridge,
MA: Belknap Press of Harvard University Press, 2014).
15 Adam Smith, The Theory of the Moral Sentiment, 6th ed. (London: A. Millar, 1790).
that of “market triumphalists,” who see all market outcomes, no matter how unequal, as morally just in the sense that they are the result of hard work and other
personal traits that we tend to see as virtuous. Kleinbard criticizes those on the left
for framing his concept of social insurance as redistributive tax and spending policy.
He argues that conceiving of the same programs as a form of social insurance in the
face of private market failure makes them much more politically palatable. Because
the premiums can only be levied and collected ex post, the tax system is the obvious
policy instrument to effect the relevant spending programs that are conventionally
characterized in the public finance literature as income-transfer programs. In addition
to supporting the provision of social insurance, Kleinbard also supports spending
programs as the basis for the provision of governmental support for business primarily, but not exclusively, in the presence of perceived market failures.
A large portion of We Are Better Than This is devoted to a review of data detailing
us performance measured in terms of various indicators of well-being that are often
the targets of program spending associated with the welfare state, as compared with
what has become an obsession with growth measured in terms of gdp. Kleinbard
draws rigorously and thoroughly on us data as well as on international sources, such
as those provided by the oecd, the International Monetary Fund, and the World
Bank. The review is generally intended to illustrate how defenders of market outcomes
and the associated increase in income and wealth inequality have used the “gdp
Olympics” as the basis for ignoring much of the poor performance of the United
States in respect of many of these measures of well-being. They have also focused on
the efficiency effects on the work /leisure and current consumption (versus deferred
consumption) behavioural margins as impediments to growth. Kleinbard similarly
provides a balanced account of much of the relevant empirical literature that casts
doubt on these effects.
Another portion of the book details the recent us trend to “starve its fiscal soul”
with a downsized government and its associated tax and spending presence, at least
in the form of explicit spending programs, rather than outsized tax expenditure
programs targeted primarily toward the middle class. In the final portion of the
book, he outlines a modest and realistic policy agenda to “reclaim the American
fiscal soul.” The agenda includes increased spending on physical infrastructure and
public education, along with certain social insurance programs. Kleinbard is careful
to avoid any specific recommendations in each of these areas and avoids wading into
health care reform, although he stresses that this reform is necessary in the long
term. Tax reform includes some reform of tax expenditures programs, but Kleinbard’s agenda consists primarily of the simple expedient of a full return to the status
quo pre-Bush tax cuts and, in particular, the prevailing personal progressive rate
structure. It is notable that Kleinbard does not follow many other us tax reformers
in calling for the elimination of many of the personal tax expenditures delivered to
the us middle class, such as the home mortgage interest deduction. Kleinbard’s
impressive career as a tax professional includes time as chief counsel to the Joint Committee on Taxation, where he led an initiative to re-conceptualize the tax expenditure
(2015) 63:2
concept,16 and he has since written extensively on the need for fundamental reform of
tax expenditures in the United States. However, he advocates here only for the
modest conversion of most personal expenditures to tax credits, rather than income
deductions, and even makes the case for the retention of some personal tax expenditures, such as relief for charitable donations and gifts.
Kleinbard emphasizes that his simple proposal to return to the pre-Bush tax cut
environment would lessen the us structural deficit while also providing sufficient
funding for his proposed social-spending increases. There would be no need for any
type of tax reform that could be considered fundamental, which is certainly a unique
feature of a policy agenda intended to realize the ambitious goals that Kleinbard
envisions. He does, however, briefly canvass some elements of what he would believe to be fundamental tax reform to be considered in the long term. Readers might
note especially Kleinbard’s brief account of his business enterprise income tax
(beit),17 which is a corporate income tax reform that is conceptually similar to a
corporate cash flow tax but is compatible with an income tax that taxes normal returns to capital. Kleinbard’s beit is thus somewhat closer in lineage to the Nordic
dual income taxes, but for some reason it has not received nearly the attention that
it deserves from public finance economists.18
Kleinbard’s modest reform agenda and associated rationale articulated in We Are
Better Than This can be viewed against the background of Piketty’s monumental
empirical work and the relationship between the return to capital and gdp growth as
an explanation for increased wealth inequality. The article by Bankman and Shaviro
explicitly attempts to draw policy lessons from a reading of Piketty’s work against
optimal tax theory, to which Piketty, as well as his frequent co-author Emmanuel
Saez, is a contributor. Bankman and Shaviro focus first on the insights that Piketty’s
work can provide for the optimal tax literature. They suggest that although the
optimal tax literature does not ignore distributional consequences of tax rule choice,
this dimension has been de-emphasized to such an extent that the literature has
unwittingly encouraged political rhetoric that links great wealth to moral desert, on
the assumption that the former is derived from human capital. The rhetoric also
16 United States, Joint Committee on Taxation, A Reconsideration of Tax Expenditure Analysis
(Washington, DC: Joint Committee on Taxation, 2008), reviewed in this feature (2008) 56:4
Canadian Tax Journal 1038-57, at 1038-40.
17 Edward D. Kleinbard, Rehabilitating the Business Income Tax, Discussion Paper 2007-09
(Washington, DC: Brookings Institution, 2007), reviewed in this feature (2008) 56:2 Canadian
Tax Journal 571-88, at 581. See also Edward D. Kleinbard, “Designing an Income Tax on
Capital,” in Henry J. Aaron, Leonard E. Burman, and C. Eugene Steurele, eds., Taxing Capital
Income (Washington, DC: Urban Institute Press, 2007), 165-209; Edward D. Kleinbard, “The
Business Enterprise Income Tax: A Prospectus” (2005) 106:97 Tax Notes 97-107; and Edward
D. Kleinbard, “Beyond Good and Evil Debt (and Debt Hedges): A Cost of Capital Allowance
System” (1989) 67:12 Taxes: The Tax Magazine 943-61.
18 For a critique of these rationales and a stated preference for expensing as a simpler approach,
see Alvin C. Warren, “The Business Enterprise Income Tax: A First Appraisal” (2008) 118:9
Tax Notes 921-39, at 927-28.
extends to the proposition that the rest of society benefits from the “human capital
heroes” through job creation. Bankman and Shaviro observe that if Piketty is correct in the explanatory power of r > g (that is, the return to capital is greater than
the rate of gdp growth), the optimal tax literature has erred in emphasizing human
capital as the primary source of rising wealth inequality and de-emphasizing capital
more broadly. They also suggest that the optimal tax literature may ignore the entire
range of negative spillover effects of wealth inequality, which Piketty only alludes to
and is the subject of a growing empirical literature.19 In this respect, what matters is
relative and not absolute levels of wealth, which is a perspective that is universally
and naïvely ignored in the optimal tax literature.
Bankman and Shaviro then canvass what they see as some of the policy lessons
from the optimal tax literature that might inform a policy agenda driven by Piketty’s
work. They make the obvious point that the optimal tax literature has erred in its
strong policy prescription for the non-taxation of savings and that this taxation can be
usefully conceived of as a corrective tax, addressing a range of negative spillover effects attributable to wealth inequality. They also argue, however, that the optimal tax
literature provides an analytical framework that is more fine-grained than Piketty’s,
in which, for example, the taxation of bequests can be treated differently from the
taxation of savings generally. Tax policy as corrective taxation can also attempt to
address the possible effect of the scaling of risky investments, in which taxpayers can
eliminate the tax by scaling up their positions. Piketty’s coarse-grained view of capital cannot account for these second-order design issues as well as other unspecified
issues. Finally, as Piketty acknowledges, the increased significance of human capital
as a source of income and wealth inequality, giving rise to what has been labelled
“the working rich,” complicates the policy analysis and associated policy prescriptions. Bankman and Shaviro suggest that optimal tax theory may offer some insights
in this respect, although they are suggestive only and not entirely clear.
T.E.
Conor Clarke and Edward Fox, “Perceptions of Tax Expenditures and Direct
Spending: A Survey Experiment,” Yale Law Journal (forthcoming)
Ryan S. Keller, “Beyond Homo Economicus: The Prosocial Brain and the
Charitable Tax Deduction,” Virginia Tax Review (forthcoming)
Clarke and Fox present the results of their survey work, which support the proposition that the public prefers the delivery of government benefits in the form of tax
expenditures, rather than direct spending, even when the two are functionally
equivalent. Moreover, the public views tax expenditures as being less costly than
equivalent direct spending. The results are unsurprising but important since they
confirm systematically, and in an empirically rigorous manner, what policy analysts
19 See, for example, Richard Wilkinson and Kate Pickett, The Spirit Level: Why Equality Is Better
for Everyone (London: Penguin Books, 2009).
(2015) 63:2
have always known through a casual empiricism. Although somewhat analytically
soul destroying, the survey results may be framed positively as suggesting a new
reform approach to tax expenditures. Rather than rail against their use and call for
their elimination, the goal of reform might more realistically focus on the need to
alter the design features of existing and proposed tax expenditures to improve target
effectiveness and cost, including delivery cost.
Keller argues that critics and advocates of tax recognition for charitable giving as
a particular tax expenditure have ignored important behavioural characteristics that
explain its appeal. He draws extensively on research in psychology and neuroscience
that explains complex social behaviour in terms of physical processes and mechanisms
in the brain. Keller argues that these mechanisms are the source of the preference
for tax recognition for charitable giving over direct government spending in the
charitable sector. Keller also argues that these mechanisms suggest the realization
of a range of positive spillover benefits associated with charitable giving that have
gone previously unrecognized in the literature. Whether or not one agrees with this
line of argument, it represents the kind of analysis that reformers might begin to
explore in the face of what seems to be a hard-wired public preference for tax expenditures over direct spending.
T.E.
Rachelle Holmes Perkins, “Salience and Sin: Designing Taxes in the New
Sin Era” [2014] no. 1 Brigham Young University Law Review 143-84
The term “tax salience” is used in two different senses in the literature, both of
which focus on the consequences of the choice of tax-inclusive or tax-exclusive pricing of goods and services for sales and/or excise tax purposes. One sense is the
more common usage of the term and emphasizes the effect of the form of pricing
on government accountability. It is the subject of a relatively extensive literature
that links transparency of pricing with political accountability.20 The other sense of
tax salience, which is the subject of this article, emphasizes the effect of the form of
pricing on consumer choice. It is the subject of a handful of empirical studies that
explore the relationship between the transparency of pricing and consumer behaviour. Perkins attempts to draw some prescriptive policy lessons from this small
empirical literature. Her article can be seen as a particular application of the “libertarian paternalism” advocated by Richard Thaler and Cass Sunstein21 as a generalized
approach to regulatory design.
20 See, for example, Richard M. Bird, “Policy Forum: Visibility and Accountability—Is TaxInclusive Pricing a Good Thing?” (2010) 58:1 Canadian Tax Journal 63-76; and W. Jack Millar,
“Policy Forum: The Case for Maintaining Tax-Exclusive Pricing” (2010) 58:1 Canadian Tax
Journal 77-85. See also David M. Sherman, “Policy Forum: Tax-Included Pricing for
HST—Are We There Yet? (2009) 57:4 Canadian Tax Journal 839-56.
21 Richard H. Thaler and Cass R. Sunstein, Nudge: Improving Decisions About Health, Wealth, and
Happiness (New York: Penguin Books, 2008).
Perkins characterizes tax-exclusive pricing as low salience and tax-inclusive pricing
as high salience. She reviews the handful of empirical studies that suggest a statistically significant behavioural response by consumers to high salience, tax-inclusive
pricing and argues that the goal of a particular tax can be better achieved by use of
one form of pricing or the other. In particular, she argues that policy makers should
use tax-exclusive pricing when a tax is intended primarily to raise revenue with minimal behavioural response, and use tax-inclusive pricing when a tax is corrective and
intended primarily to alter behaviour. This policy prescription is more ambiguous
when a tax is intended to serve both a revenue-raising and a corrective function, and
Perkins does not provide anything in the way of clear guidance regarding these
taxes. Moreover, when transparency or political accountability is the focus of the
form of pricing, tax-inclusive pricing is seen to ideally dominate policy choice.
T.E.
Axel Hilling, “Equal Taxation as a Basis for Classifying Financial
Instruments as Debt or Equity—A Swedish Case Study,”
eJournal of Tax Research (forthcoming)
Australia, Board of Taxation, Review of the Debt and Equity Tax Rules:
Discussion Paper (Canberra: Australian Treasury, March 2014),
188 pages, www.taxboard.gov.au
The different treatment of corporate debt and equity and the related need to categor
ize particular financial instruments as one type or the other is one of those timeless
and structural issues that is the subject of a vast literature, much of which fails to
advance our understanding of the subject, either conceptually or empirically. In this
respect, the article by Hilling is unfortunately inaccurate in its title. The more important aspect of the article is the discussion of the boundary in income tax law
between debt and derivative financial instruments. Much of the structured and taxdriven financial innovation focuses on this boundary and not on the debt-equity
distinction, and there is not nearly as much literature considering its maintenance.
Hilling argues that the approach to the classification of a financial instrument as
debt or a derivative instrument has undermined what was intended to be the functionally consistent treatment of returns to capital and returns to labour under the
Swedish dual income tax. Although the statutory rate applied to the capital income
of individuals under the Swedish dual income tax is a flat low rate, compared with the
progressive rates applied to labour income, the former is also subject to corporatelevel tax as equity income under a classical system. The result is an effective tax on real
inflation-adjusted returns that approximates that which is applied to labour income
at the highest rate. Interest expense on corporate debt is deductible and thereby
avoids the corporate-level tax. However, on the empirical assumption that debt returns are predominantly normal returns to waiting and compensation for expected
inflation, the effective tax rate under the Swedish dual income tax on real inflationadjusted returns is actually again close to the highest marginal tax rate on labour
income. Hilling suggests that the approach to classification of financial instruments
(2015) 63:2
as debt or equity reflected in the Swedish case law and administrative practice has
generally maintained this functionally consistent treatment by focusing on the presence of equity exposure that is associated with a particular instrument. In other
words, the presence of significant contingent payments that are attributable to equity
exposure can result in the classification of an instrument as equity, which is subject to
tax under the classical corporate tax and the dual income tax; however, the formally
single level of tax that is applied to debt remains available only for instruments with
little to nothing in the way of contingent payments attributable to equity exposure
and an effective tax rate under the dual income tax on non-contingent returns that
approximates the rate on contingent returns on equity-flavoured instruments,
whether in the form of shares or debt.
Hilling emphasizes that this classification equilibrium has been upset by the
Swedish approach to the maintenance of the boundary between debt and derivative
instruments, where the form of an instrument is largely determinative of its classification, provided that any contingent payments are not linked to equity exposure. This
approach has allowed a significant element of risky returns to be embedded in the
legal form of a debt instrument by embedding a derivative written on commodities
or an index that is not equity-based. Alternatively, time-value returns can be embedded in a derivative instrument and deduction-inclusion treatment can be maintained
under the Swedish dual income tax, provided that the derivative instrument is not
itself equity-based. In the face of this expansion of the concept of debt to include a
range of risky returns, Swedish policy makers have attempted to realize consistency
of treatment of equity and debt by eliminating the formally preferential treatment of
debt and applying a new flat tax and corporate income tax approach to both equity
and debt. The cost of this approach is the abandonment of functionally consistent
treatment of capital and labour income under the Swedish dual income tax.
The report by the Australian Board of Taxation is noteworthy, not for anything
novel about the debt-equity boundary in tax law but rather for its review of the novel
classification approach adopted by the Australian Treasury Department in 2001.22
The resulting Australian debt-equity classification rules are a legislative outlier. If
imitation is the highest form of flattery, the fact that no other country has seen fit
to follow the Australian lead may be instructive. Unlike the Swedish approach to the
classification of financial instruments, which is largely based on case law as in other
countries, the Australian debt-equity classification rules are an attempt to comprehensively classify financial instruments ex ante by legislative prescription. Yet,
broadly similar to the Swedish classification approach, the presence of contingent
and non-contingent payments is the determinative factor. Very generally, an instrument that provides non-contingent payments that are equal to or greater than the
amount advanced on issue is classified as debt, irrespective of its legal form. When
22 Income Tax Assessment Act 1997, as amended, division 974.
the instrument takes the form of a share or otherwise provides equity exposure to the
issuer, it is considered to be equity, unless the non-contingent payment test is satisfied and it is considered to be debt, subject to deduction/inclusion treatment rather
than the Australian dividend imputation system. When an instrument takes the legal
form of a debt instrument but provides for contingent payments that are attributable to an embedded derivative that does not provide equity exposure to the issuer,
it is classified neither as equity nor as debt, if the non-contingent payments are less
than the amount advanced on issue. The treatment in this instance is consistent with
the treatment of a derivative financial instrument in legal form that provides noncontingent payments that are also considered insignificant under the same debt test.
The Australian Board of Taxation was established following the recommendation in a 1999 report on the taxation of business in Australia.23 As requested, the
board reviews and provides advice to the Australian Taxation Office and Treasury
Department on various technically focused policy issues under existing legislation.
The board was asked in May 2013 to undertake a post-implementation review of
the debt-equity classification rules with a view to identifying areas in which the rules
might be improved in both the domestic and cross-border contexts. This discussion
paper is intended to frame the concerns that will be the focus of a final report after
submissions are received on the issues that are raised in the discussion paper or that
are otherwise relevant to the broader terms of reference.
The discussion paper consists of 10 chapters. Chapters 1 to 3 provide background material, while chapter 4 identifies potential problems in the operation of
the existing legislation. Chapter 5 focuses more narrowly on an issue that has arisen
with stapled securities and a specific anti-avoidance rule. Chapters 6 and 7 review
the interaction of the classification rules and other operative provisions of the Australian income tax legislation. Chapter 9 identifies compliance and administrative
issues. Readers will probably be most interested in chapter 10, which focuses on the
operation of the Australian debt-equity classification rules in a cross-border context.
The classification results provided by the uniquely prescriptive Australian legislative
approach have proven to be a fruitful source of inconsistent country classifications,
which, much like the us check-the-box approach to entity classification, are a source
of hybrid mismatches and double non-taxation. This subject is one of the topics
addressed in the beps project24 and was certainly on the oecd policy radar screen
before that.25 Chapter 10 provides an overview of recent developments in this area
23 Australia, Review of Business Taxation, A Tax System Redesigned—More Certain, Equitable and
Durable (Canberra: Review of Business Taxation, July 1999).
24 Organisation for Economic Co-operation and Development, BEPS Action 2: Neutralising the
Effects of Hybrid Mismatch Arrangements, OECD/G20 Base Erosion and Profit Shifting Project
(Paris: OECD, 2014).
25 Organisation for Economic Co-operation and Development, Hybrid Mismatch Arrangements:
Tax Policy and Compliance Issues (Paris: OECD, March 2012), reviewed in this feature (2012)
60:3 Canadian Tax Journal 765-91, at 774-76.
(2015) 63:2
and the functioning of the Australian classification rules in relation to these developments. General policy directions are tentatively suggested, presumably with the
possible results, if any, of the product of this multilateral initiative in mind.
T.E.
Alice G. Abreu and Richard K. Greenstein, “The Rule of Law as
a Law of Standards: Interpreting the Internal Revenue Code”
(2015) 64 Duke Law Journal Online 53-94
This article is the latest in a series of articles by Abreu and Greenstein advocating
the use of a standards-based approach to the articulation of the concept of income
for us income tax purposes and the use of standards rather than prescriptive rules
more generally.26 The authors make the somewhat unremarkable claim that the
existence of rules in tax law does not prevent the interpretation of certain of these
formulations as more open-ended standards and that such an interpretation does
not violate understandings of the rule of law. In this respect, they are responding to
Professor Lawrence Zelenak, who has criticized an earlier paper of one of the authors that advocates the use of a standards-based approach to the concept of income.27
Zelenak’s critique had emphasized that the authors’ approach “stretched beyond the
breaking point” the bounds of accepted interpretation and, in doing so, raised serious rule-of-law considerations.
T.E.
26 Alice G. Abreu and Richard K. Greenstein, “It’s Not a Rule: A Better Way to Understand the
Definition of Income” (2012) 13:3 Florida Tax Review 101-32; and Alice G. Abreu and Richard
K. Greenstein, “Defining Income” (2011) 11:5 Florida Tax Review 295-348.
27 Lawrence A. Zelenak, “Custom and the Rule of Law in the Administration of the Income Tax”
(2012) 62:3 Duke Law Journal 829-55.

Reforming Old Age Security: Effects and Alternatives

Transcription

Similar documents

COMRADERY FROM COAST-TO

List of Canadian bi#161AC38

INTRODUCTION During the Second World War, the role of women

FOL-1-FORMA 2 - Service Alberta

Multi Sports trades actively as Permanis CEO

October 2011 / Issue 26

Sample Pages/Table of Contents

CTC Environmental Policy_April_2015_low-res

Farm News - Bee Heaven Farm

RESUME Ève Brodeur