David Weisbach and Jacob Nussim’s 2004 paper “The Integration of Tax and Spending Programs” marked an important law & economics contribution to the tax-policy literature. The paper argues that, given a particular policy, the part of the government best suited to implement that policy should be tasked with doing so.
Government Program Implementation Choices
Weisbach and Nussim focus their analysis on whether government spending should be enacted as a tax expenditure through the tax code or as a direct expenditure through an appropriations bill. An agricultural subsidy, for example, could be administered either by the U.S. Treasury Department (specifically, the Internal Revenue Service) or the U.S. Agriculture Department. The key to deciding which should be assigned the task is to determine which would do a better job. This is, in essence, an application of cost-benefit analysis, as both departments will have advantages and disadvantages.
Weisbach and Nussim note that government departments are not all the same. They may, for example, be staffed with people who have different competencies, incentives, and preferences. And different departments may have different institutional designs. As a result, departments may differ in:
- How costly it would be for them to implement a policy;
- How likely they are to make errors implementing a policy; and
- How well they are able to modify a policy, should circumstances change quickly.
These types of departmental differences should be the deciding factors in where policies ought to be implemented.
Alternative Views of Program Implementation
It may seem obvious to use cost-benefit analysis to determine which parts of the government should undertake which activities, but there are a few related strands within the tax-policy literature that take a different approach, perhaps less informed by cost-benefit analysis. The first argues for a comprehensive tax base. In other words, an income tax should tax all income and nothing else. The second argues against tax expenditures. In other words, the tax code should not be used to achieve means that could otherwise be achieved with direct expenditures. The third argues in favor of tax simplicity. It follows that government objectives should not be imported into the tax system because doing so would increase the tax system’s complexity.
To be fair, someone applying cost-benefit analysis might find in favor of a comprehensive tax base, against tax expenditures, or in favor of tax simplicity. The issue that Weisbach and Nussim point out is that these arguments sometimes rely on implicit value judgments that give rise to a notion of tax exceptionalism. If the government intends to implement a given program, it would seem to defy cost-benefit analysis to house the program outside the tax system solely to achieve a comprehensive tax base, to avoid tax expenditures, or to maintain tax simplicity. If the government were committed to achieving a particular objective, implementing it as a direct expenditure and not through the tax system would avoid an increase in tax complexity but would increase complexity somewhere else. A fixation on decreasing tax complexity might cause one to overlook this tradeoff.
Weisbach and Nussim’s approach, moreover, avoids analysis that may not have much policy relevance. A tax expenditure is defined as a deviation from an income tax. But it turns out there is no universal definition of an income tax; there are many different plausible views. There is therefore no consensus on what constitutes a tax expenditure, meaning that ink is spilled on definitional disagreements that do not necessarily have a direct impact on optimal policy.
A somewhat similar problem arises in the literature with regard to how to determine what a comprehensive income tax base is. Here, Weisbach and Nussim go too far when they write that differing views on the comprehensive tax base “depend on each advocate’s taste for purity over administrative complexity.” While there is indeed a tradeoff between the “purity” of an income tax and its administrative complexity, this tradeoff need not be made based on one’s taste for purity.
In “How Tax Complexity and Enforcement Affect the Equity and Efficiency of the Income Tax,” Louis Kaplow uses an economics approach to explore this tradeoff. There are, for example, many taxpayer costs that could plausibly be either personal or business in nature (or potentially a little of both). Under an income tax, business expenses should be deductible; personal expenses should not be. Cars have both personal and business uses. A “pure” income tax would thus have to allocate the cost of a taxpayer’s car between its personal and business uses. To do so perfectly (were that even possible) would impose substantial costs on both taxpayers and administrators.
Kaplow’s approach trades off these taxpayer and administrator costs against the harm of deviating from a pure income tax. This harm includes an equity and an efficiency cost, both of which arise when taxpayers who have the same true income are treated differently for tax purposes. For example, some taxpayers may be able to deduct personal expenses, and others may not be able to deduct business expenses. Using a social-welfare approach, these deviations from a pure income tax will reduce social welfare for two reasons. First, the marginal social welfare of all taxpayers will not be the same (the equity cost). Second, taxpayers will have an incentive to prefer expenditures (both personal and business) that are more likely to result in a tax deduction, which may result in economic distortions (the efficiency cost).
After specifying a social-welfare function, one can trade off the costs and benefits of income tax purity. One can, for example, determine how much additional budget should be allocated to tax enforcement to determine how taxpayers use their vehicles. In Kaplow’s setting, policy preferences do arise. But these are preferences over social-welfare functions that are arguably unavoidable, and not merely “tastes for administrative purity.”
Extensions and Avenues for Further Research
Weisbach and Nussim make a compelling argument but, for the most part, assume both a fixed institutional setting and a fixed policy objective. Further research might explore when a government ought to make changes to its departments. Governments decide not only which departments ought to implement particular policies but also how to constrain and support different departments, and even which departments exist. The two-stage problem of how to set up various departments and then how to assign policy objectives across those departments would be a worthwhile extension of Weisbach and Nussim’s work.
Assuming a fixed policy objective (e.g., an agricultural subsidy) simplifies the problem but misses some important political-economy issues. Allowing governments the choice to implement policies either as direct expenditures or as tax expenditures may itself change how many and which policies a government chooses to implement. Voters may, for example, support a policy when presented as a tax expenditure but not as a direct expenditure. Political actors may take advantage of these plausibly irrational voter preferences.
Weisbach and Nussim do grapple with this issue but perhaps understate its importance. In “Perceptions of Taxing and Spending: A Survey Experiment,” published in 2015, Conor Clarke and Edward Fox present evidence that people are not equally likely to support a policy if it is presented as a tax expenditure, rather than a direct expenditure, even if there are no other differences between the policies.
Conclusion
Weisbach and Nussim’s paper is a good reminder that policy analysis may be erroneous if it looks only at one part of a larger system. This is particularly true when different parts of the system are interconnected and there are ways to substitute activities between those parts. The question of whether a program belongs in the tax code cannot be fully answered without considering what the program would look like in an appropriations bill, and vice versa. Policies are carried out by real institutions with strengths, weaknesses, and incentives of their own. Taking account of those realities, the familiar debates over tax expenditures, tax-base purity, and tax simplicity start to look less like first principles and more like incomplete heuristics.
For Further Reading
- Louis Kaplow, “How Tax Complexity and Enforcement Affect the Equity and Efficiency of the Income Tax,” National Tax Journal, Vol. 49, No. 1 (1996).
- Conor Clarke & Edward G. Fox, “Perceptions of Taxing and Spending: A Survey Experiment,” Yale Law Journal, Vol. 124 (2015).
- Boris I. Bittker, “The Value of a Comprehensive Tax Base as a Tax Reform Goal,” Harvard Law Review, Vol. 80 (1967).
- Stanley S. Surrey & Paul R. McDaniel, “Tax Expenditures,” Harvard University Press (1985).
