Full Article
Transcription
Full Article
J. OF PUBLIC BUDGETING, ACCOUNTING & FINANCIAL MANAGEMENT, 14(2), 159-173 WINTER 2002 AN EXAMINATION OF THE EFFECTS OF BALANCED BUDGET LAWS ON STATE BORROWING COSTS Cynthia Sneed* ABSTRACT. This study investigates the relationship between different levels of state balanced budget laws and state borrowing costs. Using federal guidelines for state balanced budget law classifications, this author inserted dichotomous variables in an empirical model of state borrowing costs. Ordinary Least Squares Regression is utilized to determine which balanced budget laws are recognized in state interest costs. The results indicate a significant relationship between the most restrictive levels of balanced budget laws and state borrowing costs. The strongest balanced budget laws are associated with lower interest costs while the weakest budget laws are associated with higher costs. It appears that taxpayers in states with weaker balanced budget amendments may not be as protected against excessive government growth as those in states with the most stringent balancing requirements. INTRODUCTION Washington is very excited over the apparently generous surpluses that resulted from the bipartisan balanced budget agreement. However, if the experience of state governments is any indicator, a balanced-budget amendment requiring the federal government to balance the budget in the Constitution may reduce future federal budget deficits, but not eliminate them entirely. Some 49 states have balanced budget provisions, yet many of them have significant debt. This debt results from state provisions allowing their individual budgets to be separated and exceptions to be written into their constitutions or statutes. Thus, whether a balanced-budget amendment is truly effective in curbing spending depends on the relative restrictiveness of the agreement. ____________ * Cynthia Sneed, Ph.D. CGFM, is an Associate Professor of Accounting at Jacksonville State University. Her research interests are public sector retirement system issues and governmental bond market issues. Her teaching interests are information systems and e-commerce issues related to accounting. Copyright 8 2002 by PrAcademics Press 160 SNEED It can be quite difficult to determine exactly what a state's constitutional balanced budget agreement requires. Some constitutions only require that the governors submit a balanced budget plan to the state legislature but there is no requirement that the legislature actually pass a balanced budget. Also, some twenty constitutions allow states to carry a deficit over into the next fiscal year, yet these states are classified as "balanced budget" states. According to a recent General Accounting Office survey, budget officers in ten states, most of which are balanced budget states, said they were not subject to requirements of a year-end balance. Budget officers in another eleven states said that they were allowed to carry over deficit financing if necessary. Thus, one can see that simply classifying a state as having a balanced budget law does not mean that the state has no deficit spending. This research investigates governmental bond market recognition of the relative strength of the balanced budget laws when pricing state bonds. The research hypothesizes that bond markets recognize the most restrictive balanced budget laws in borrowing costs because these types of balanced budget laws are harder to circumvent through accounting gimmicks or (1) tricks. Evidence that governmental bond markets recognize the most restrictive form of balanced budget laws (relative to the other more moderate or weaker laws) would suggest that governmental markets may be more efficient than prior research suggests. The research results also would have important implications for fiscal policy. Taxpayers in states with moderate or weak balanced budget laws may believe they are protected from excessive government growth when in fact the states have ample opportunity to increase spending and their borrowing costs are higher. The increased government growth would have important intergenerational consequences for the future citizens the laws were purportedly designed to protect. Finally, if Congress does decide to pass a constitutional amendment to balance the budget, the policymakers may want to look at the most restrictive state balanced budget agreements for guidance in choosing an effective balanced budget law. This paper will describe the state balanced budget rules, discuss the prior literature, describe the research methodology, and discuss the empirical results, followed by a conclusion. EFFECTS OF BALANCED BUDGET LAWS ON STATE BORROWING COSTS 161 OVERVIEW OF BALANCED BUDGET LAWS Those who advocate amending the United States Constitution to require a balanced federal budget enjoy substantial public, if not political, support for their position. However, until the Keynesian school of microeconomic theory during the 1930s and 1940s, the norm of our national political history was to balance the budget. Even without Constitutional requirements to do so, the accepted attitude about how our government should finance its affairs was fiscal conservatism. Buchanan (1995) notes that, prior to World War II, politicians would have considered it to be immoral to spend more than they were willing to generate in taxes except in the case of extreme national emergency. Spending borrowed funds on current expenditures was not acceptable political behavior. There were basic moral constraints in place and no need for an explicit fiscal rule in the written constitution. However, during the Keynesian period, the government budget was seen as a mechanism to set social and economic policy as well as political agenda. Keynesian economists believe that the budget provides an instrument through which economic objectives such as full employment and economic growth can be attained. In order to sell the theory, the public had to believe that government debt and deficits were not only socially acceptable, but also a desirable method to attain socioeconomic objectives. Thus, it became necessary to convince the public that there were no negative consequences to debt financing. Needless to say, this turnabout in public attitudes did not take place overnight. What started as a theoretical movement during the 1930s and 1940s took almost thirty years for fruition. However, by the time the Johnson administration's "Great Society" programs of the 1960s were proposed, the politicians were ready for a way to spend money for their constituents without having to raise immediate tax revenues to pay for the increase in services. Thus, a cycle of substantial deficit financing began that has resulted in a federal debt in excess of five trillion dollars. Summaries of state balanced budget requirements and state budgeting practices can be found in the National Association of State Budget Officers (NASBO 1992) and the U.S. General Accounting Office (1993). Only Vermont does not have a balanced budget law; however, the restrictiveness of the laws varies from state to state. The laws can be classified into three broad categories depending on where in the budget process the balance is required. The least restrictive of the laws require simply that the governor submit a balance budget but says nothing about enforcement of the law. 162 SNEED Forty-four states have this requirement but many also have other restrictions. Thirty-seven states require that the state legislature enact a balance budget but allow for deficits to run in years where actual revenues and expenditures deviate from budget expectations. How long states are allowed to carry these deficits varies from state to state. Some states require a balanced budget by the next fiscal year, others the next biennium. The third and most stringent type combines the requirement to enact a law with no deficit carryforwards. Twenty-four of the thirty-seven states that require a balanced budget have this law. According to Portera (1994), forty-eight of the forty-nine states classified by NASBO have balanced budget laws but the laws apply only to the general fund. In 34 states the rules do not apply to special funds (such as those earmarked for tax receipts of capital spending or trust funds). In some states funds raised by issuing long-term debt for capital projects can be included in the revenues for the budget period. Table 1 provides a summary of the state balanced budget laws including where in the budgeting process the budget is required to be balanced. PRIOR LITERATURE Prior literature on balanced budget laws and state fiscal policy includes recent studies by Alt and Lowry (1994) and Poterba (1994). Alt and Lowry investigate how anti-deficit carrying provisions affect state spending and taxes. Using Census data, they model state revenues and expenditures as a function of current state income, current federal grants, lagged difference between revenues and expenditures and variables for political circumstances. Comparing state fiscal policy reactions to differences between revenues and expenditures, they found that a $1 state deficit in the current year triggers a 77 cent response, through tax increases or spending reductions in Republican-controlled states that restrict deficit carryovers, and a 34 cent response in Democratic states with deficit restrictions. In states that do not restrict carryovers, the adjustments were 31 cents and 40 cents for Republican and Democrat states respectively. Thus, the evidence suggests that state politics is an important influence in controlling state spending. Poterba (1994) investigates how state fiscal policies respond to unexpected deficits or surpluses. His results indicate that states with weak balanced budget laws adjust spending less in response to unexpected EFFECTS OF BALANCED BUDGET LAWS ON STATE BORROWING COSTS TABLE 1 Summary of State Balanced Budget Laws with ACIR Index Classification State (1) (2) (3) (4) (5) (6) Alabama X 10 Alaska X 6 Arizona X X 10 Arkansas X 9 California X 6 Colorado X 10 Connecticut X 5 Delaware X 10 Florida X X 10 Georgia X X 10 Hawaii X 10 Idaho X 10 Illinois X 4 Indiana X X 10 Iowa X X 10 Kansas X 10 Kentucky X 10 Louisiana X 4 Maine X 9 Maryland X 6 Massachusetts X 3 Michigan X 6 Minnesota X 8 Mississippi X 9 Missouri X X 10 Montana X X 10 Nebraska X 10 Nevada X 4 New Hampshire X 2 New Jersey X 10 New Mexico X 10 New York X 3 North Carolina X 10 163 164 SNEED TABLE 1 (Continued) State (1) (2) (3) (4) (5) (6) North Dakota X 8 Ohio X X 10 Oklahoma X 10 Oregon X X 8 Pennsylvania X 6 Rhode Island X 10 South Carolina X 10 South Dakota X 10 Tennessee X 10 Texas X 10 Utah X X 10 Virginia X 8 Washington X 8 West Virginia X 10 Wisconsin X 6 Wyoming X 8 Notes: Column (1) = Governor only has to submit a balanced budget. Column (2) = Legislature Only has to Pass a Balanced Budget Column (3) = May Carry Over a Deficit but Must be Corrected in Next Fiscal year. Column (4) = State Cannot Carry Over a Deficit Into Next Biennium Column (5) = State Cannot Carry Over a Deficit Into Next Fiscal Year Column (6) = Stringency Index. deficits than do states with more restrictive laws. Poterba found that a $100 deficit overrun leads to only a $17 expenditure reduction in a state with a weak balanced budget law, while it leads to a $44 reduction in other states. In addition to how balanced budget laws affect fiscal policy, there also is research on balanced budget rules and state borrowing. Von Hagen (1991) compares the level of state general obligation debt, and per capita state income, in states with and without stringent balanced budget requirements. Von Hagen used the Advisory Council on Intergovernmental Relations 2 (ACIR index 1987) to classify the balanced budget laws. His results EFFECTS OF BALANCED BUDGET LAWS ON STATE BORROWING COSTS 165 indicate that general obligation debt is substantially lower in states with stringent balanced budget amendments than in other states. He also examines the effect of stringent balanced budget rules on the ratio of nongeneral obligation debt to general obligation debt across states. The results indicate that states with more restrictive balanced budget laws, as well as states with lower general obligation debt limits, exhibit higher levels of revenue debt and other non-general obligation debt. These results were consistent with Bunche (1991) and Kiewiet and Szakaly (1992). Bunche found that states with more stringent debt limits or balanced budget laws are more likely to use alternatives to state-approved borrowing to finance projects. Kiewiet and Szakaly examined the effects of antideficit measures and limits on state borrowing and found that a combination of stringent anti-deficit rules and voter-approval of debt was more likely to bring about pressure for tax increases or spending cuts, rather than debt financing. Lowry and Alt (1995) found that the bond market reaction to a state deficit depended on whether the state had a balanced budget requirement. States with balanced budget rules experience small increases in their borrowing costs for a given deficit period. This study extends prior research by examining the bond market reaction to the various levels of restrictiveness of the balanced budget laws. Using ACIR guidelines, the balanced budget rules are classified into three categories (strong, moderate and weak) and the significance of the dichotomous variables representing the balanced budget rules is examined to test the research hypothesis. The following section describes the research hypothesis and methodology. RESEARCH METHODOLOGY This research investigates the following hypotheses: H1: States with strong balanced budget laws have lower borrowing costs then states with moderate or weaker balanced budget laws. H2: States with moderate balanced budget laws have lower borrowing costs than states with weaker balanced budget laws. The research hypotheses are investigated by examining the significance of two dichotomous variables representing three different levels of balanced budget laws in a standard model of state borrowing costs developed in the 166 SNEED literature. The research model, control variables and research variables are discussed below. Research Model and Variable Selection The empirical model is summarized in the following equation. The variable definitions and expected signs of the coefficients are presented below. NIC = B0 + B1(HI) + B2(MED) + B3(INT) + B4(MAT) + B5(GODEBT) + B6(OWNREV) + eit where: NIC = HI = yield to maturity on state bonds; a dichotomous variable representing the balanced budget law. Variable is 1 if the state had the most restrictive law and 0 otherwise (-); MED = a dichotomous variable representing the balanced budget law. Variable is 1 if the state has a moderate BBL and 0 otherwise (); INT = the general level of interest rates on the date the bond was issued (+); MAT = term to maturity, measured in years (+); GODEBT = general obligation debt per capita (+); OWNREV = ratio of general revenues from own sources to total general revenues (-); i = state; t = year; e = random error term. The research model tests whether states with the strongest laws have lower borrowing costs than states with moderate or weaker balanced budget laws. To test the hypothesis the dependent variable, net interest cost (NIC), is operationalized as average yield-to-maturity on new state bond issues. The interest variable (INT) is included to control for the general level of interest rates at the time of the bond issue. The predicted sign of the coefficient is positive as higher levels of general interest rates are expected to raise state borrowing costs. Maturity (MAT) is a control variable included to control for differences in yields across different maturities. Longer maturities are expected to have EFFECTS OF BALANCED BUDGET LAWS ON STATE BORROWING COSTS 167 higher yields because of increased risk. Thus, the anticipated sign of the MAT coefficient is positive. A variable to control for the amount of general obligation debt the state has incurred is included in the model. GODEBT is operationalized as general obligation debt per capita. The predicted sign of the coefficient is positive since states with higher per capita debt are expected to have higher levels of borrowing costs. The last control variable (OWNREV) measures the percentage of a state's general revenues that are provided by the state's own sources. This variable measures a state's selfreliance in generating revenues. The higher the ratio the lower the default risk, which is expected to result in a negative relationship between OWNREV and NIC. The research variables, HI and LOW, are dichotomous variables included in the model to test the research hypothesis. Using ACIR classifications, the balanced budget laws were classified according to level of restrictiveness. The ACIR classified BBL on a scale of 1 to 10, with 10 being the strongest form of BBL. In keeping with ACIR classifications, states with ranking 1-3 are weak, 4-7 are moderate and 8-10 are strong. Table 1 describes the state ACIR classifications. The predicted sign of the HI coefficient is negative. States with strong and moderate balanced budget laws are hypothesized to have lower borrowing costs relative to other states with weak balanced budget laws. Sample Selection and Data Sources Data for the dependent variable, NIC, for all new states bond issues from 1990-1992 were collected from Bloomberg's Financial Markets database. Examining the data indicated that 26 states combined for a total of 78 new bond issues for the sample period. The new bond issues are serial bonds, where a portion of the principal is repaid each year. Data for the balanced budget laws for states over 1991-1992 were (2) obtained from the ACIR classification table. The data sources for other model variable were State Government Finances 1991 and 1992. Ordinary least squares regression was used to estimate the parameters of the regression model. The data were first examined for evidence of the violations of the ordinary least square assumptions that would cause inefficient estimates of the regression coefficients or biased estimates of their variances. Plots of the independent variables against the dependent variable indicate that the data were normally distributed. Additionally, the 168 SNEED data were examined for evidence of correlation among the variables. The correlation matrix in Table 2 did not indicate a multicollinearity problem among the variables. White's adjustment procedure was used to correct heteroscedasticity (Breusch-Pagan Chi-Square 62.7684 with 6 degrees of freedom). TABLE 2 Correlation Matrix Variables in the Regression Model HI MED INT MAT GODEBT OWNREV -----------------------------------------------------------------------------------------HI 1.000 -.290 .257 -.144 -.359 -.027 MED -.290 1.000 -.088 -.015 -.037 .196 INT .257 -.086 1.000 -.037 -.054 .158 MAT -.144 -.015 -.037 1.000 .110 .142 GODBT -.359 -.037 -.054 .110 1.000 .305 OWNRV -.027 .196 .158 .142 .305 1.000 where: NIC = yield to maturity on state bonds; HI = a dichotomous variable representing the balanced budget law. Variable is 1 if the state had the most restrictive law and 0 otherwise (-); MED = a dichotomous variable representing the balanced budget law. Variable is 1 if the state has a moderate BBL and 0 otherwise (-); INT = the general level of interest rates on the date the bond was issued (+); MAT = term to maturity, measured in years (+); GODEBT = general obligation debt per capita (+); OWNREV = ratio of general revenues from own sources to total general revenues (-); i = state; t = year; e = random error term. EFFECTS OF BALANCED BUDGET LAWS ON STATE BORROWING COSTS 169 EMPIRICAL RESULTS The descriptive statistics for the continuous variables in the research model are presented in Table 3 and estimates of the coefficients and pTABLE 3 NIC = B0 + B1(HI) + B2(MED) + B3(INT) + B4(MAT) + B5(GODEBT) + B6(OWNREV) + eit Panel A. Descriptive States for Model Variables Variable Mean Std. Dev. Minimum Maximum ---------------------------------------------------------------------------------------NIC HI MED INT MAT GODEBT OWNREV 5.8923 .9125 .6025 .0512 6.7565 8.9840 $419.5500 .7407 2.3000 .4925 .2200 .3603 2.6265 $436.4300 .0298 10.0000 .0000 .0000 5.8800 2.5000 $12.2000 .6784 1.0000 1.0000 7.5300 15.5000 $2001.0000 .8027 ---------------------------------------------------------------------------------------Panel B. Estimates of the Model Parameters Variable Coefficient T-ratio P-value* ---------------------------------------------------------------------------------------One HI MED INT MAT GODEBT OWNREV R2 = .4558 N = 78 P-values are one tail. where: NIC = HI = MED = -2.0592 - .2717 - .2717 1.3596 .0675 .0358 -2.1504 -1.4890 -1.8870 -1.1520 7.3680 2.7300 2.1360 -1.7110 .0704 .0374 .1230 .0000 .0039 .0015 .0456 yield to maturity on state bonds; a dichotomous variable representing the balanced budget law. Variable is 1 if the state had the most restrictive law and 0 otherwise (-); a dichotomous variable representing the balanced budget law. Variable is 1 if the state has a moderate BBL and 0 otherwise (-); 170 SNEED INT = the general level of interest rates on the date the bond was issued (+); MAT = term to maturity, measured in years (+); GODEBT = general obligation debt per capita (+); OWNREV = ratio of general revenues from own sources to total general revenues (-); i = state; t = year; e = random error term. values of the variables are presented in Table 4. The overall model is 2 statistically significant (p-value .000) with an adjusted R of .4553. All of the control variables are significant and have the predicted sign. INT is highly significant (p-value .0000) and positive, indicating that state borrowing costs are higher as interest rates rise. As expected, MAT is significant and positive (p-value .0039) indicating that borrowing costs are higher for longer maturities. GODEBT is highly significant and positive (pvalue .0015) indicating that higher levels of general obligation debt are associated with higher borrowing costs. Finally, OWNREV is significant (p-value .0456) and has the predicted negative sign indicating that borrowing costs are less for states with a higher ratio of their own revenues to general revenues. As was expected, the research variable, HI is highly significant (pvalue .0374) and negative, indicating that states with the strongest form of balanced budget laws have lower levels of borrowing costs than states with weaker or more moderate laws. The MED variable is not significant in the research model (p-value .1230) but did have the predicted sign. These findings suggest that taxpayers in states with the less restrictive balanced budget laws pay higher interest costs than do taxpayers in other states. The results also suggest that the bond market recognize the less restrictive laws as somewhat less effective than the stronger and acts accordingly. EFFECTS OF BALANCED BUDGET LAWS ON STATE BORROWING COSTS 171 SUMMARY AND CONCLUSIONS This research examines the association between state borrowing costs and different levels of balanced budget laws. States with budget deficits have only three options for dealing with the problem. They can raise taxes, reduce spending or close the deficit through some other method such as deferring payments across the fiscal year or accelerating receipts. The results provide evidence that governmental bond markets do recognize the most restrictive form of balanced budget laws (relative to the weaker laws) and also suggest that governmental markets may be more efficient than previously recognized. In a descriptive sense, the findings indicate that taxpayers in states with strong balanced budget laws pay less in interest costs than other states with balanced budget laws. Taxpayers in states with moderate or weak balanced budget laws may believe they are protected from excessive government growth when in fact the states have ample opportunity to increase spending and their borrowing costs are higher. The increased government growth would have important intergenerational consequences for the future citizens the laws were purportedly designed to protect. Proponents of the balanced budget law often cite states' success in achieving fiscal restraint as the catalyst for the balanced budget movement. But supporters often ignore the fact that the combined debt for all 50 states in 1993 was $387.7 billion, or more than half their combined annual expenditures of $742.9 billion. Estimates are that local governments carry even larger amounts of debt (Hosansky, 1995). States can incur this debt because they are able to split apart their budgets and use loopholes written into their constitutions or statutes. Investigating the effects of balanced budget laws on state borrowing costs is timely given the political debates regarding the disposition of current federal surpluses. If the Congress does decide to require a balanced budget, then policymakers may want to look at the most restrictive state balanced budget agreements for guidance. Foes of a federal balanced budget amendment oftentimes cite the problems with funding capital projects as a reason not to enact a constitutional amendment to balance the federal budget. However, state officials have long argued that relying on debt for capital project expenditures is necessary and fair because it allows them to spread the burden of capital projects over many years. Florida's budget director, Robert Bradley said that asking current-year taxpayers to absorb 172 SNEED all the costs that the future residents are going to cause is unfair and impractical (Hosansky, 1995). The results of this research may provide insight for policymakers into practical methods of writing a balanced budget law that would insure protection of taxpayers against massive government growth while also providing for vital government services such as infrastructure and capital expansion. ACKNOWLEDGMENTS The author would like to thank participants at the 1999 American Accounting Association Midwest Meeting as well as several anonymous reviewers for their insightful and helpful comments. Any errors and omissions are entirely my own. NOTES 1. Examples of actions taken by states to remedy budget deficits include California’s transferring revenues from an oil extraction tax in a trust fund to the general fund and New York’s deferring the 1983 payroll period over into 1984 to balance the budget (GAO 1985). 2. The ACIR Index can be found in the publication "Significant Features of Fiscal Federalism" complied by the Advisory Commission on Intergovernmental Relations. This publication provides information about state statutory or constitutional limits on borrowing, including information about when the balanced budget law was adopted and what type of balanced budget law was adopted. The ACIR Index describes degree of stringency of the balanced budget laws. REFERENCES Advisory Council on Intergovernmental Relations. (1990). Significant Features of Fiscal Federalism, Washington, DC: Author. Alt, J., & Lowry, R. (1994). “Divided Government and Budget Deficits: Evidence from the States.” American Political Review, 88(8), 11-28. Buchanan, J. (1995). “Clarifying Confusion about the Balanced Budget Agreement.” National Tax Journal, 48, 347-355. EFFECTS OF BALANCED BUDGET LAWS ON STATE BORROWING COSTS 173 Bunche, B. (1991). “The Effect of Constitutional Debt Limits on State Governments' Use of Public Authorities.” Public Choice, 68, 57-69 Hosansky, D. (1995, February 4). “Running a Debt, Amendment or No.” Congressional Quarterly Weekly Report, 1. Kiewit, R., & Szakaly, K. (1992). “The Efficacy of Constitutional Restrictions on Borrowing, Taxing and Spending: An Analysis of State Bonded Indebtedness, 1961-1990.” (Unpublished Manuscript), Pasadena, CA: California Institute of Technology. Lowery, R., & Alt, J. (1995). “A Visible Hand? Intemporal Efficiency, Costly Information and Market-Based Enforcement of Balanced Budget Laws” (Mimeo). Cambridge, MA: Harvard University, J. F. Kennedy School of Government. National Association of State Budget Officers. (1992). State Balanced Budget Requirements: Provisions and Practices. Washington, DC: Author. Potera, J. (1994). “State Responses to Fiscal Crisis: The Effects of Budgetary Institutions and Politics.” Journal of Political Economy, 102, 799-821. U.S. General Accounting Office. (1985). Budget Issues: State Balanced Budget Practices (GAO/AFMD-86-22BR). Washington, DC: Author. U.S. General Accounting Office. (1993). Balanced Budget Requirements: State Experiences and Implications for the Federal Government (GOA/AFMD-93-58BR). Washington: Author. Von Hagen, J. (1991). “A Note on the Empirical Effectiveness of Formal Fiscal Restraints.” Journal of Public Economics, 44, 199-210.