VoxEU Column Taxation

Tax expenditures: The big government behind the curtain

As the US tries to cut back its debt, the battle lines are already being drawn. Republicans are in favour of spending cuts; Democrats in favour of tax rises. Putting political ideology to one side, this column asks what objective economics has to say.

As the Congressional Super Committee grapples with painful options to cut the federal debt, taxes have become the sticking point. Democrats have insisted that tax increases be part of any agreement while Republicans counter that all the cuts must come on the spending side.

Absent ideological objections, there really should be no conflict. Public finance economists and academic tax lawyers have long recognised that there are a number of spending programmes run through the tax code.

  • The biggest “tax expenditures” – the tax exclusion for health insurance and the mortgage interest deduction – are just subsidy programmes run through the income tax rather than a programme agency. (See Table 1.)
  • Repealing or curtailing some of those subsidies would simultaneously increase tax revenues and cut spending.

It could, in principle, be a solution to the logjam over the debt.

Table 1. Largest tax expenditures in FY 2011, in billions of dollars





Exclusion for employer-sponsored health insurance



Mortgage interest deduction



401(k) plans



Deduction for state and local taxes other than property taxes



Step-up basis of capital gains at death



Lower rate on capital gains



Charitable deduction (other than education and health)



Pensions (defined benefit)



Exclusion of net imputed rental income



Capital gains exclusion on home sales


Note:  Provisions are ranked based on five-year total cost, FY 2011–15.

Source:  US Budget, Analytical Perspectives, FY2011

Tax expenditure concept

The tax expenditure concept dates back to 1967, when Treasury Assistant Secretary Stanley Surrey directed his staff to compile lists of “government spending for favored activities or groups, effected through the tax system rather than through direct grants, loans, or other forms of government assistance” (Surrey and McDaniel 1985).

Some critics object to the notion that letting taxpayers keep more of their own money could be construed as spending. But most economists can readily see the duality between tax expenditures and traditional spending programmes in the sense that they have nearly identical effects on the budget, resource allocation, relative prices, and the distribution of income. The only difference, typically, is in who administers the programme.

Cutting taxes by buying weapons?

The late economist David Bradford (2003) famously illustrated this point by proposing, with tongue firmly in cheek, a Weapons Supply Tax Credit, which would allow arms manufacturers to sell their ordinance to the Pentagon in exchange for tax credits rather than cash. Instantly, the Defence Department’s budget would decline by the amount of transformed spending. By some measure it would look like a tax cut. Tax revenues would fall by a similar amount (or more, if weapons suppliers demanded a premium on account of the complexities and uncertainties associated with the tax credit mechanism). But government would be doing exactly the same thing. Only the accounting would change.

A more substantive debate relates to the baseline against which tax expenditures are measured. (Donald Marron (2011) has an especially lucid discussion of the baseline and measurement issues.) Surrey and McDaniel thought a very comprehensive income tax should be the baseline, but others have pointed out that, against that yardstick, tax incentives for saving and lower tax rates on dividends and capital gains are counted as tax expenditures when those provisions would be the norm under a consumption-based tax system. Since the US income tax is really a hybrid combining aspects of income and consumption taxes – and many economists favour a consumption tax on efficiency grounds – it is not clear which baseline is most appropriate. Donald Marron and Eric Toder (2011), however, have estimated that about 70% of tax expenditures would be treated as such against either baseline. At 2011 levels that would amount to about $800 billion of spending that most economists would agree should be subject to scrutiny.

While Surrey thought that the sole function of the tax code was to raise revenue to finance the government, there may be good reasons to run some programmes through the IRS. For example, when information on eligibility is already reported on tax returns or easily obtainable by the tax authorities, a tax expenditure might be easier to administer and comply with than a traditional spending programme. But tax expenditures receive preferential treatment in the US political and budgeting process.

The basic problem is that tax expenditures are mostly hidden from public view. Political scientist Chris Howard (1997) aptly named them “The Hidden Welfare State”. Another political scientist, Suzanne Mettler (2011), referred to “The Submerged State”. Mettler contends that the relative invisibility of tax expenditures undermines democracy because their relative obscurity makes it more difficult for citizens to understand how government programmes affect them. Lobbyists can sneak expensive ineffective subsidies into the tax code that would never pass muster as direct spending programmes—think ethanol tax credits. Moreover, even relatively worthwhile programmes (Mettler cites the Affordable Care Act) may be misunderstood when important provisions are run through the tax code.

In recent research (Burman and Phaup 2011), we outline another concern. Voters may not fully perceive the cost of tax expenditures, resulting in a government that is larger and less efficient than would prevail if citizens had full information. Tax expenditures have a privileged status in the budget process. A new tax credit or deduction is considered a ‘tax cut,’ and thus relatively immune from the ‘tax and spend’ critique that would apply to a similar spending programme. Tax expenditures are scored as reductions in revenues rather as a new spending programme. As a result, both spending and taxes are understated. In a political context where both are considered bad, this clearly creates a bias in favour of tax expenditures over traditional spending.

We develop a simple model to illustrate the distortions that could be created from current treatment of tax expenditures. In this model voters value both tax expenditures and cash outlays, but they underestimate the cost of the former. This ‘fiscal illusion’ lowers the relative price of tax expenditures compared with cash outlays, resulting in over-provision of tax subsidies and a higher overall level of government spending than would occur with an accurate perception of price. Taxes, either now or in the future, also are higher than they would be in the absence of this distortion. Indeed, if tax expenditures and cash outlays are complements, even traditional spending could rise. And there is a production inefficiency as more and more resources are diverted from cash outlays into increasingly inefficient tax expenditures.

It is not clear how to test this hypothesis empirically. As noted above, tax expenditures are quantitatively large – $1.2 trillion by the Treasury Department’s broad definition, and $0.7-0.8 trillion under the narrower baselines suggested by Marron (2011). By comparison, individual income tax revenues in 2011 are an estimated $1.1 trillion. But the size of tax expenditures has not varied greatly relative to GDP since the Tax Reform Act of 1986 eliminated a number of them. (See Figure 1.) The largest determinant of the value of tax expenditures turns out to be marginal income tax rates – since deductions and exclusions are worth more at higher tax rates than lower ones – but that makes interpreting trends difficult.

Figure 1. Number and value (as percent of GDP) of tax expenditure provisions, 1983–2009

Source: For tax expenditures as percent of GDP, GAO analysis of OMB, Analytical Perspectives, Budget of the United States Government, Fiscal Years 1985-2011; for count of provisions, Joint Committee on Taxation (annual tax expenditure compilations back to 1985), and author's calculations.

Perhaps a better metric is the number of tax expenditures, which has increased sharply since 1986. The Joint Committee on Taxation (2011) estimated that there were 202 tax expenditures in 2007, a 50% increase from 1986, when there were 135. Some of this increase was due to a change in the way the JCT compiled tax expenditures (Buckley 2011), so the specific estimates should be taken with a large grain of salt. But there's no doubt that the number has increased dramatically. However, Buckley also points out that many of the largest tax expenditures have been in the tax code for a very long time and survived the massive tax reform enacted in 1986, suggesting that there is little political will to revise these programmes.

So far, we’ve emphasised that controlling tax expenditures is necessary to make government smaller and more efficient, an argument that would seem to appeal to Republicans, but there is also a ‘Democratic’ argument for subjecting tax expenditures to scrutiny. With the exception of refundable tax credits, which are available even to taxpayers with little or no income tax liability, tax subsidies are most valuable to people with higher incomes. (See Table 1.) Burman et al (2008) estimate that income tax expenditures reduced tax liability by 13.5% of income for taxpayers in the top 1% of incomes, compared with less than 7% for households with low or moderate incomes. Thus, if safety net programmes like food stamps and Medicaid are on the chopping block, upper middle class tax entitlements like the mortgage interest deduction and tax-free health insurance should also be scrutinised.

Table 2. Tax expenditures as a percentage of after-tax income, selected quintiles, 2007











Top 1%








  Above-line deductions






  Capital gains, dividends






  Itemised deductions






  Non-refundable credits






  Refundable credits












All provisions












Source: Burman, Geissler, and Toder (2008)



Finally, the most successful budget plan will be one that reduces the debt without impairing economic growth. Raising tax rates would entail significant economic costs while base-broadening reduces the cost of taxation (Saez et al 2009). According to James Poterba (2011), “because [tax expenditures] distort behaviour relative to a neutral tax code, it is possible that eliminating some or all of them could simultaneously raise revenue and reduce tax-induced distortions of economic activity.” That is, curtailing tax expenditures can be part of much needed tax reform that could simultaneously make the tax system simpler, fairer, and more efficient.

The bipartisan National Commission on Fiscal Responsibility and Reform (2010), also known as the Bowles-Simpson Commission, and Bipartisan Policy Center (2010) followed this model – cutting tax expenditures while reducing marginal tax rates and raising net tax revenues. The super committee should take note.


Bipartisan Policy Center (2010), “Restoring America’s Future:  Reviving the Economy, Cutting Debt, and Creating a Simple, Pro-Growth Tax System”.

Bradford, David (2003), “Reforming Budgetary Language”, in Sijbren Cnossen and Hans-Werner Sinn (eds.), Public Finance and Public Policy in the New Century, MIT Press, 93-116.

Buckey, John L (2011), “Tax Expenditure Reform: Some Common Misconceptions”, Tax Notes, 18 July.

Burman, Leonard E, Christopher Geissler and Eric J Toder (2008), “How Big Are Total Individual Income Tax Expenditures, and Who Benefits from Them?”, The American Economic Review,98(2):79-83.

Burman, Leonard E and Marvin Phaup (2011), “Tax Expenditures, the Size and Efficiency of Government, and Implications for Budget Reform”, NBER Working Paper 17268.

Howard, Christopher (1997), The Hidden Welfare State:  Tax Expenditures and Social Policy in the US, Princeton University Press.

Joint Committee on Taxation (2011), Background Information on Tax Expenditure Analysis and Historical Survey of Tax Expenditure Estimates(JCX-15-11). 28 February.

Marron, Donald B (2011), “Spending in Disguise”, National Affairs, 20-34.

Marron, Donald and Eric Toder (2011), “Measuring Leviathan: How Big is the Federal Government?” Tax Policy Center, Presentation at “Starving the Hidden Beast: New Approaches to Tax Expenditure Reform”, Loyola Law School, Los Angeles, 14 January.

Mettler, Suzanne (2011), The Submerged State, The University of Chicago Press.

National Commission on Fiscal Responsibility and Reform (2010), The Moment of Truth.

Poterba, James M (2011), “Introduction: Economic Analysis of Tax Expenditures”, National Tax Journal, 64(2, Part 2):451-458.

Saez, Emmanuel, Joel B Slemrod, and Seth H Giertz (2009), “The Elasticity of Taxable Income with Respect to Marginal Tax Rates:  A Critical Review”, NBER Working Paper 15012.

Surrey, Stanley S, and Paul R McDaniel (1985), Tax Expenditures,Harvard University Press.