By Lehn M. Benjamin and Paul L. Posner
Tax expenditures provide public subsidies through the income tax code’s treatment of specified sources or uses of income to certain taxpayers, in order to support specific activities and purposes defined by Congress as promoting specified national policy goals. Unlike subsidies recorded as spending in the federal budget, these subsidies are not provided through outlays from Treasury but rather by revenue losses stemming from taxpayer filings. Tax expenditures subsidize both individual and business taxpayers through deductions, credits, exclusions, preferential rates and deferrals.
The Growing Appeal of Tax Expenditures
Over time, tax expenditures have quietly assumed major fiscal significance. The Government Accountability Office estimates that there were 169 tax expenditures in 2015, with revenue losses of $1.23 trillion, an amount exceeding the spending for all discretionary appropriations in the federal budget. Among selected OECD nations, the United States’ tax expenditures are second only to Canada, exceeding 50 percent of income tax revenues collected.  Tax expenditures are relied upon for most of the 19 mission areas defined as functions in the federal budget.
Rapid growth in the use of tax expenditures can be explained by several factors. They deliver subsidies quickly to targeted individuals and businesses, often through an “announcement effect” that changes behavior once policy changes are announced and enacted. They can reach thousands of taxpayers and influence their behavior without the need to gain prior approval from government agencies or other public actors. They can reduce the stigma attached to government programs since obtaining benefits of tax subsidies does not generally require application or prior approval. A good example of this is the policy objective of supporting low-income people through the Earned Income Tax Credit. Such provisions can be an attractive alternative to government spending programs as ways to help to individuals, communities and businesses.
Perhaps a better explanation for the rise of tax expenditures as a federal policy tool is their political appeal. Elected leaders have strong political incentives to resort to tax expenditures rather than direct spending to pursue their policy objectives. Tax expenditures give proponents the ability to create new benefits and seemingly cut taxes at the same time, while not appearing to increase spending: achieving what some have called a fiscal miracle.
Unlike appropriations, tax expenditures are not capped or, in most cases, reviewed annually by Congress and the Administration. Many tax expenditures are comparable to entitlement programs for which spending is determined by rules for eligibility, benefit formulas, and other parameters rather than by Congress appropriating specific dollar amounts each year. With some exceptions, tax expenditures make funds (through reduced taxes) available to all qualified claimants on an open-ended basis.
Perhaps because traditional federal control and oversight are absent, tax expenditures often constitute weak or poorly aimed subsidies, either reducing their impact on intended policy outcomes or distorting efficient markets when they do affect behavior. Many tax subsidies have been found to provide fiscal windfalls to those who would have undertaken the subsidized activity in the absence of the tax expenditure. A good example is the subsidy for Individual Retirement Accounts (IRAs), which have been shown to produce a net of 9 cents in additional savings for each dollar of lost revenue, as most taxpayers use the provision to simply transfer existing savings from taxed to non-taxed IRA vehicles. The research and development tax credit was found by GAO to produce only a modest increase in corporate research and development spending, but with significant windfalls to certain companies.
Other tax expenditures reduce the efficiency of markets by tilting investment toward tax-preferred activities, raising costs in the process. Two prominent examples are housing and health care, where costs are undoubtedly increased by tax subsidies provided through tax deductions and exclusions. For health care, the income tax exclusion of employer contribution to employees’ health insurance premiums is credited with increasing health care coverage for employees, but also with leading employees and firms to obtain more coverage than they would otherwise, thereby increasing demand for and cost of health care,
Policymakers have few opportunities to make explicit comparisons or trade-offs between tax expenditures and federal spending programs. Growing revenues forgone through tax expenditures reduce the resources available to fund other programs or to cut the deficit, requiring higher statutory tax rates to obtain a desired amount of revenue. Since there is no coordination between tax subsidies and related spending programs, it is not surprising that considerable overlap and inconsistencies emerge with related spending programs. Higher education features federal grants, loans, guarantees and a welter of tax credits, interest deductions for loans and tax preferred savings vehicles. GAO has found that these fragmented programs can offset one another in unpredictable ways and confuse families seeking to make the best use of federal assistance. 
Even when price effects are not clear, tax expenditures can be assessed based on their income effects. Overall, tax expenditures save $6,500 per household annually, cutting the average effective tax rate paid from 26 percent to 18 percent of income. The savings, however, are unequally distributed, with two thirds of tax expenditure savings for individuals accruing to the top income quintile.. Deductions and exclusions have been called “upside down” subsidies because those in higher brackets realize greater tax savings than those in lower brackets. Tax credits, on the other hand, reduce taxes dollar for dollar and, thus, have a more neutral effect on distribution of tax burden by income groups. But, those whose income is too low to pay taxes realize no benefit from these provisions unless the tax subsidy is ‘refundable.’ Refundable tax credits, like the Earned Income Tax Credit are available to the lowest income workers, who have no tax liability, through direct payment by the Internal Revenue Service (IRS).
The shortfalls discussed above prompted Edward Kleinbard, the former Chief of Staff of the powerful Joint Tax Committee, to conclude that tax expenditures unquestionably constitute our most wasteful government spending. If tax expenditures suffer from shortfalls in efficiency, effectiveness and equity, the question then becomes whether policymakers and managers can do more to ensure that these popular tools better achieve their stated policy objectives? Can they also increase accountability for those who benefit from this tool?
Tax expenditures represent the most non-centralized and disaggregated implementation regime of any federal tool of government. In his 2002 book on governmental tools, Salamon ranks tax expenditures as among the least coercive and direct of the tools in the federal arsenal. 
First, unlike other tools, their use requires no formal approval by federal officials prior to the claiming of federal resources by taxpayers, nor is there prior federal review of their planned use. Second, taxpayers select themselves to participate rather than waiting for a federal sanction. This is enabled by the open-ended nature of the federal tax subsidy commitment, so that the ultimate size of the federal commitment is in the hands of taxpayers, not federal officials. As a result, government officials are often left guessing about the magnitude and distribution of tax expenditures in coming years. Often tax expenditures can grow through drift rather than conscious choice, as the individualized behavior of taxpayers collectively determines aggregate revenue losses.
Limited monitoring by the IRS undermines accountability for tax subsidies. Tax audits have been viewed as the bedrock of tax oversight, but as IRS budgets have declined so has the coverage of audits. The average number of taxpayers audited has declined to less than 1 percent of all filers in recent years.
Ensuring that taxpayers appropriately use deductions, exemptions and exclusions in ways that further social policy objectives is not a task well aligned with the core mission or operations of the IRS, which is to ensure the efficient collection of revenue. Given the commitment to enforcing tax law compliance, the proliferation of tax expenditures complicates the IRS’s job and calls on skills and background that the average tax examiner does not possess. As Christopher Howard says, IRS and Treasury have a professional bias against the use of tax expenditures. It is no surprise therefore that the IRS confines itself largely to monitoring for taxpayer compliance and does not collect information to enable it to make broader evaluations about the effectiveness and social utility of such tax expenditures as mortgage interest deductions or research and development credits. This is obvious in the failure of the agency to even obtain data from taxpayers on a majority of tax expenditures. In a very telling report issued in 2013, the GAO found that 63 percent of the 167 tax expenditures did not require specific reporting by taxpayers claiming them.  GAO contrasted this with federal spending programs funding activities in the same policy areas; all spending programs required reporting by third parties on financial and programmatic results and accomplishments.
The constraints on IRS monitoring are also reflected in the general lack of coverage of tax expenditures by government-wide performance and financial accountability regimes. Tax expenditures have largely escaped coverage from the growing web of government-wide performance evaluation and reporting provisions, starting with the Government Performance and Results Act (GPRA) of 1993. While this act included Senate Committee report language urging the Administration to develop a framework for periodically assessing the performance of tax expenditures, with the exception of three pilot studies in 1997, this was largely ignored. The Bush Administration’s Program Assessment Rating Tool (PART) also excluded tax expenditures from those targeted reviews. Amendments to GPRA passed in 2010 (GPRAMA) called for agencies to review tax expenditures in their strategic plans, but a recent GAO study showed that only 11 of the 169 tax programs were covered in agency strategic plans (GAO, 2016).
The federal budget process does not integrate tax expenditures with either the presentation of federal budgets or in the review of agency budget requests by OMB. Notwithstanding an earlier GAO recommendation, OMB has not committed to including tax expenditures during its review of agency budget requests during the annual budget cycle.  Similarly, Congress does not include tax expenditures in the budget resolution, the annual spending plan which encompasses spending for discretionary and mandatory entitlement programs. When it comes to deciding which programs need to be trimmed to meet congressional deficit reduction targets, the committees are on their own to determine which programs must face cuts. The revenue committees of both houses must then determine whether resolution targets are to be met through tax expenditures or other spending programs that fall in their jurisdiction.
A Reform Agenda
The substantial accountability and performance issues associated with tax expenditures have prompted organizations such as the GAO and the Urban Institute-Brookings Tax Policy Center to suggest reforms in the design, oversight and fiscal accounting of this tool to promote greater review and attention in the policy process. Such reforms are often benchmarked to the accountability frameworks already applicable to spending programs at the federal level. 
We suggest that the President and Congress consider the following reforms to improve efficient use of and accountability for tax expenditures:
- Apply All Performance Requirements Under the GPRAMA to Tax Expenditures
Federal agencies with primary responsibility for specified budget functions should work with the Treasury Department to articulate programmatic goals for each tax program in that function and should collect data to illuminate the marginal contribution of these subsidies to the purposes specified in statute and to the Department’s strategic objectives as specified in its most recent strategic plan.
- Integrate Tax Expenditure Reviews in the Federal Budgeting Process
Integrate tax expenditure reviews in the federal budget process and featuring reviews of related tax and spending programs by OMB in executive budget formulation and in reviews directed by the budget committees in the congressional budget process. The integration should ensure that related tax and spending programs are considered together by OMB and the budget committees (see portfolio budgeting recommendation in Memo No. 6, on the federal budget process).
- Restructure Tax Expenditure Programs
Restructure selected tax expenditure programs to resemble grant and loan programs, with annual limits on federal revenue losses and competitive awards to recipients by federal, state or local governments. A limited set of tax expenditures already follows this model, most notably the low-income housing tax credit.
- Reclassify the Accounting for Tax Expenditures
Reclassify the accounting for tax expenditures to record them in the budget totals as spending outlays, with matching increase in revenues. The deficit would remain unaffected by this change, but the change in presentation would eliminate the presumptive tilt toward tax expenditures in the current federal budget process. 
No progress has been made on this agenda over the past twenty years. That is not to say that individual tax expenditures have not been revised to improve accountability, or in some cases eliminated. The 1986 Tax Reform Act alone instituted dramatic base broadening by reducing and eliminating a wide range of individual and corporate tax expenditures ranging from the deductibility of consumer interest to passive loss windfalls on real estate. The drumbeat of concern over fraudulent claims under the Earned Income Tax Credit program and now the Affordable Care Act’s new tax credits for health insurance have provoked congressional oversight and additional IRS reviews. However, there has been no systematic reform during this period to the process of formulating, reviewing or overseeing tax expenditures to ensure they achieve their intended objectives. The reforms that were instituted – notably the GPRMA’s requirement for Treasury to institute performance reviews of tax expenditures – have been largely ignored.
Weak accountability for tax expenditures may reflect a fundamental ambivalence that national leaders and the public about tax expenditures as a tool of government. Their non-administered nature gives rise to a conflict of views among governmental actors about whether tax expenditures represent the delegation of authority to agents—in this case taxpayers – for helping achieve specific national policy objectives or whether tax expenditures are a return of limited authority to taxpayers—the ultimate owners of government—to pursue their own objectives.
Weak accountability for tax expenditures may have far-reaching consequences for public trust in government. Tax expenditures have grown in good part because they are ideally tailored to suit the public’s deep-seated ambivalence about government. Their bipartisan popularity also reflects a political leadership drawn to choosing the policy tools that appear least intrusive. While satiating the public’s appetite for limited government, tax expenditures also pose difficult accountability and management challenges. While gaining leaders short-term electoral approbation, their use such can widen the gap between high expectations and implementation realities on the ground, accelerating and fueling the very disaffection with government that prompted their use in the first place.
 A tax expenditure is operationally measured by reference to the receipts that would be forthcoming under the “normal” tax system. See Joint Tax Committee, Estimates of Federal Tax Expenditures for Fiscal Years 2014 – 2018, August, 5, 2014.
 U.S. Government Accountability Office, Tax Expenditures: Opportunities Exist to Use Budgeting and Agency Performance Processes to Increase Oversight (Washington, DC: GAO, July, 2016, GAO-16-622)
 Organization for Economic Development Cooperation, Tax Expenditures in OECD Countries (Paris: OECD, 2010, p 226.
 Leonard Burman and Marvin Phaup, “Tax Expenditures, the Size and Efficiency of Government and Implications for Budget Reform”, Cambridge, Ma: National Bureau of Economic Research , 2011.
 Orazio P. Attanasio and Thomas DeLeire, “The Effect of Individual Retirement Accounts on
Household Consumption and National Saving,” The Economic Journal, Vol. 112, July 2002,
 Government Accountability Office, The Research Credit’s Design and Administration Can Be Improved (Washington, D.C.: GAO 10-136, December, 2009).
 Originally, some of these earlier tax expenditures ere defined as part of the tax base rather than as subsidies departing from that base. Today, however, such provisions are defined today as subsidies departing from the “normal tax system”.
 Bob Lyke, Tax Benefits for Health Insurance and Expenses: Current Legislation.
Congressional Research Service Issue Brief IB98037 (Washington, D.C.: February 2005).
 GAO, Student Aid and Postsecondary Tax Preferences: Limited Research Exists on
Effectiveness of Tools to Assist Students and Families Through Title IV Student Aid and
Tax Preferences, GAO-05-684 (Washington, D.C.: July 29, 2005).
 Daniel Baneman, Joseph Rosenberg, Eric Toder and Roberton Williams, “Curbing Tax Expenditures”, (Washington: Urban Institute/Brookings Tax Policy Center, January, 2012).
 Edward D. Kleinbard, We Are Better Than This: How Government Should Spend Our Money (New York: Oxford University Press, 2015.
 Lester Salamon, The Tools of Government: A Guide to the New Governance (New York, Oxford Press, 2002), pp. 24-32.
 Christopher Howard, “Tax Expenditures”, in Lester Salamon, ed, The Tools of Government, 2002.
 GAO, Tax Expenditures: IRS Data for Evaluation Limited, GAO-13-479, April 2013.
 GAO, Government Performance and Accountability : Tax Expenditures Represent a Substantial Commitment and Need to be Reexamined, GAO 05-690, September, 2005.
 See GAO, Government Performance and Accountability : Tax Expenditures Represent a Substantial Commitment and Need to be Reexamined, GAO 05-690, September, 2005; Eric J. Toder, “ Tax Cuts or spending – Does it make a difference?” National Tax Journal, 53:3 (September, 2000), pp. 361-371.
 Leonard E. Burman and Marvin Phaup, “Tax Expenditures, the Size and Efficiency of Government and Implications for Budget Reform” NBER Working Paper 17268, August 2011.