An Analysis of Budget Offset Requirements for Tax Expenditures


Budget rules enacted by Congress in the 1990’s (and currently still in force) established pay-as- you-go (PAYGO) rules requiring that the revenue loss from any new or expanded tax expenditure provision be offset in the same fiscal year by legislation raising an equivalent amount of extra revenue. This note extends Dharmapala’s (1999) model of the political economy of tax expenditures to incorporate the effects of these budget offset requirements (and associated political constraints on taxation). It characterizes how the policy outcomes in Dharmapala (1999) are affected by the PAYGO provisions, and analyzes the welfare consequences of budget offset requirements. It is shown that, while the PAYGO rules lead to reduced levels of tax expenditures, they also induce the tax committee to shift the financing of tax expenditures from (relatively less distortionary) general revenues to (relatively more distortionary) sector-specific taxes or reductions in existing tax expenditures. Hence, the welfare consequences of budget offset requirements are ambiguous.

Keywords: Tax Expenditures; Interest Groups; Budget Rules

JEL Classification: D72; H20

1. Introduction

In addition to establishing caps on discretionary spending, the Budget Enforcement Act of 1990 (hereafter, BEA90) extended the scope of Congressional budgetary rules to encompass decisionmaking about tax expenditures (Keith, 1997; Garrett, 1998a).[1] The pay-as-you-go (hereafter, PAYGO) rules of the act require that, if a new tax expenditure provision is enacted (or an existing one expanded), the resulting revenue loss must be offset in the same fiscal year by another piece of legislation raising an equivalent amount of extra revenue. These ‘budget offset requirements’ are, moreover, enforced by automatic sequestration. BEA90 was originally scheduled to expire in FY1995, but was extended until FY1998 by Congress in 1993. Subsequently, the Budget Enforcement Act of 1997 (hereafter, BEA97) further extended these provisions until the end of FY2002.[2] While there has been some commentary by legal scholars (e.g. Garrett, 1998a, b), there has been little formal economic analysis of the effects of these budget rules on policy outcomes and social welfare. Moreover, these issues are now particularly topical in view of the impending expiration of the BEA97 provisions and the debates concerning their renewal.

Dharmapala (1999) develops a framework for the formal analysis of the political economy of tax expenditures, combining an endogenous commodity taxation model (e.g. Dixit, 1996; Dixit, Grossman and Helpman, 1997) with a model of legislative institutions. However, it does not fully take into account the PAYGO provisions of BEA90. The aim of this note is to extend the model in Dharmapala (1999) to incorporate the effects of these budget rules on the budget constraint facing legislative decisionmakers who enact tax expenditure provisions. In particular, it considers how the policy outcomes characterized in Dharmapala (1999) are affected by the PAYGO provisions, and analyzes the welfare consequences of budget offset requirements.

Briefly, Dharmapala (1999) models a tax-writing committee (such as House Ways and Means) that chooses sector-specific tax expenditures under the influence of a subset of sectors organized into lobby groups.[3] These tax expenditures can be financed either through uniform lump sum taxes, or through sector-specific (distortionary) taxes on unorganized sectors. In this note, the PAYGO rules (and associated informal political constraints on raising general taxation levels) are captured by restricting the tax committee’s access to lump sum taxes (analogous to relatively less distortionary general revenues) in the financing of tax expenditure provisions. This makes the enactment of tax expenditures more costly (in terms of social welfare, which enters the tax committee’s objective function). As expected, this results in lower levels of tax expenditures. However, from a public finance standpoint, a paradoxical feature of the PAYGO rules is that they tend to shift the financing of tax expenditures from (relatively less distortionary) general revenues to (relatively more distortionary) sector-specific taxes or reductions in existing tax expenditures.[4] It is shown that, hence, the welfare consequences of budget offset requirements are ambiguous. Moreover, for the budget rule to achieve an unambiguous welfare improvement, it would have to be sufficiently detailed to virtually specify the policy outcomes, particularly in the case where the socially optimal levels of some of the tax expenditures are positive.

Section 2 discusses the background to the adoption of offset requirements. The model and results are presented in Sections 3 and 4. Section 5 considers the implications of these results.

2. The Background

The development of budgetary politics and institutions in Congress since the 1980’s has been recounted, for instance, in Auerbach (1994, pp. 155ff) and Garrett (1998a, pp. 502ff). In 1985, the Balanced Budget and Emergency Control Act (also known as Gramm-Rudman- Hollings or GRH) was enacted. While it applied only to appropriations, it sought to foster an informal norm of revenue-neutrality. BEA90 formalized this revenue-neutrality norm by introducing the PAYGO provisions. Henceforth, the revenue loss from any new or expanded tax expenditure provision would have to be offset by revenue-increasing measures enacted in the same fiscal year. In theory, an increase in tax expenditures could be financed by an increase in general taxation, or a reduction in entitlement spending. However, Garrett (1998a, pp. 515-516) claims that these options were, in political terms, prohibitively costly; in practice, the only option involves financing new or expanded tax expenditures by reducing or eliminating existing tax expenditures. She argues that, hence, the PAYGO provisions raise the cost of enacting or expanding tax expenditures. Moreover, budget offset requirements force interest groups seeking new or expanded tax expenditures to become ‘predators’, diverting funding directly from other groups (p. 515), and impose greater post-enactment costs on successful interest groups, requiring them to defend their gains from future predators. Thus, these rules lead to greater scrutiny of existing tax expenditures. Garrett concludes that offset requirements represent a mechanism for harnessing interest group competition to limit Federal spending (see also Garrett, 1998b).

There appears to have been little empirical analysis of the impact of PAYGO provisions on the enactment of tax expenditures. Table 1 presents data on revenue losses from tax expenditures as a percentage of GNP for 1975-97.

A preliminary examination suggests that there has been a slight decrease in the level of tax expenditures, from 7.3% before BEA90 to 6.6% over 1991-97. However, there are any number of confounding factors that may affect this figure. The model in this note provides a theoretical foundation for interpreting the data, and for analyzing the welfare consequences of budget offset requirements.

3. The Model



In Dharmapala (1999), the tax committee can finance the tax expenditures it enacts for the organized groups either using uniform lump sum taxes (which can be interpreted as approximating the use of general revenues) or using sector-specific taxes on unorganized groups.[5] The budget constraint is satisfied through a uniform per-head lump-sum transfer of r,


and G is the (exogenous) revenue requirement. The formulation in Dharmapala (1999) conforms to the letter of BEA90 (in the sense that the PAYGO rules permit the use of general revenues to finance expansions in tax expenditures); however, it fails to incorporate the political constraint

highlighted by Garrett (1998a) on raising general taxation levels. In the context of the model,

Garrett’s point may be understood as follows: the notion of the tax committee ‘choosing’ p is essentially a simplification of the idea that the tax committee has a substantial degree of agenda- setting power (in relation to the floor of Congress) over tax policy. However, it may be the case, as Garrett argues, that the political climate is such that a majority of legislators on the floor are committed to opposing any increase in general taxation.

This additional constraint can be captured by limiting the extent to which the tax committee can use lump sum taxes. Suppose that the maximum amount of lump sum tax revenue


4. Results

Dharmapala (1999), following the earlier common agency literature, solves the model sketched above using the notion of a ‘truthful’ Nash equilibrium (TNE), in which each group’s marginal contribution equals its marginal benefit. This approach is equivalent to the tax committee simply maximizing its joint surplus with the organized groups:










On the whole, then, it would seem that Eq. (5) is a more realistic characterization of the practical consequences of the PAYGO rules than is Eq. (16). In these circumstances, it not clear that budget offset requirements will necessarily raise social welfare. The more general lesson for institutional design in legislatures is very simple but powerful: constraining political actors by forcing them to adopt more socially costly means to achieve their aims need not reduce the total social cost that they create in their pursuit of these aims.


Auerbach, A. J. (1994) “The U.S. Fiscal Problem: Where We Are, How We Got Here, and Where We’re Going”, in S. Fischer and J. J. Rotemberg (eds.) NBER Macroeconomics Annual 1994, Cambridge and London: MIT Press, 141-175.

Brennan, H. G. and J. M. Buchanan (1980) The Power to Tax: Analytical Foundations of a

Fiscal Constitution, Cambridge University Press: Cambridge, UK.

Dharmapala, D. (1999) “Comparing Tax Expenditures and Direct Subsidies: The Role of

Legislative Committee Structure”, Journal of Public Economics, 72, 421-454.

Dixit, A. K. (1996) “Special-Interest Lobbying and Endogenous Commodity Taxation”,

Eastern Economic Journal, 22, 375-388.

Dixit, A. K., G. M. Grossman and E. Helpman (1997) “Common Agency and Coordination: General Theory and Application to Government Policy Making”, Journal of Political Economy, 105, 752-769.

Garrett, E. (1998a) “Harnessing Politics: The Dynamics of Offset Requirements in the Tax

Legislative Process”, University of Chicago Law Review, 65, 501-569.

Garrett, E. (1998b) “Rethinking the Structures of Decisionmaking in the Federal Budget

Process”, Harvard Journal on Legislation, 35, 387-445.

Grossman, G. M. and E. Helpman (1994) “Protection for Sale”, American Economic

Review, 84, 833-850.

Keith, R. (1997) “The Budget Enforcement Act of 1997”, Congressional Research Service

Report 97-930: Washington, D.C.

Pechman, J. A. (1987) Federal Tax Policy, 5th ed., Brookings Institution: Washington, D.C. Saturno, J. V. (1998) “The Budget Enforcement Act: Its Operation Under a Budget Surplus”,

Congressional Research Service Report 98-97: Washington, D.C.

Table 1: Tax Expenditures as a Percentage of GNP 1975-1997

Fiscal Year Value of Tax


(Billions of $)

Corresponding GNP

(Billions of $)

Tax Expenditures as

a % of GNP

1975 92.86 1474.3 6.3
1976 97.36 1599.1 6.1
1977 113.46 1785.5 6.4
1978 123.47 1994.6 6.2
1979 149.82 2254.5 6.6
1980 181.48 2520.8 7.2
1981 228.60 2742.1 8.3
1982 253.50 3063.8 8.3
1983 286.70 3179.8 9.0
1984 317.02 3434.4 9.2
1985 347.62 3801.5 9.1
1986 380.73 4053.6 9.4
1987 288.97 4277.7 6.8
1988 278.85 4544.5 6.1
1989 284.00 4908.2 5.8
1990 276.37 5200.8 5.3
Average 1975-1990 7.3
1991 340.02 5764.9 5.9
1992 431.46 5932.4 7.3
1993 455.59 6255.5 7.3
1994 480.28 6563.5 7.3
1995 426.12 6931.9 6.1
1996 450.44 7246.7 6.1
1997 470.72 7567.1 6.1
Average 1991-1997 6.6

Source: Adapted from Dharmapala (1999, p. 450, Table 1), where the sources and methodology

are explained in more detail. Note that the estimates of the value (i.e. revenue loss due to) tax expenditures pertain to the fiscal year, and (following the Congressional Budget Office’s procedure) the GNP for the previous year is used as the ‘corresponding’ GNP.

[1 ]Tax expenditures can be defined as “special provisions of the income tax laws that reduce the tax liability of those who make payments or receive income in certain designated forms” (Pechman, 1987, p. 355).

[2] It should be emphasized that the PAYGO provisions remain applicable, even in the altered fiscal environment resulting from the recent budget surpluses – see Saturno (1998) for a discussion of this point.

[3] The main focus in Dharmapala (1999) is on comparing the tax committee’s policies with those resulting from a decentralized committee structure; but it is only the former that are relevant for this extension.

[4]In this sense, there is an analogy here to the ‘Leviathan’ model of Brennan and Buchanan (1980). There, citizens choose fiscal constitutions that impose relatively distortionary tax instruments in order to constrain a revenue- maximizing government.

[5] It may seem unrealistic to assume that sector-specific taxes can be imposed; however, there exist examples of such taxes (e.g. the gasoline tax and various other federal excise taxes). Moreover, it is possible to interpret these ‘taxes’ as reductions in the tax expenditures granted to the unorganized groups – see Section 5 for a further discussion of this point.

[6] Note that, if (5) is binding, r is now – (T + G)/N.

Previously published by the Australian National University- Economics Program, Research School of Social Sciences