Tax Expenditures: Trends and Critiques
Tax Expenditures: Trends and Critiques
Updated February 4, 2008
Thomas L. Hungerford
Specialist in Public Finance
Government and Finance Division
Tax Expenditures: Trends and Critiques
Special deductions, exclusions, and exemptions (sometimes characterized as
“loopholes”) have always been in the tax code. In the mid-1960s, the Department of
the Treasury became interested in tracking and accounting for these tax subsidies.
Indeed, the term “tax expenditures” was first used at this time. Both the Department
of the Treasury and the Joint Committee on Taxation prepare annual lists and
estimates of tax expenditures, and Treasury’s list is included in the President’s annual
budget submission. For FY2007, the Joint Committee on Taxation (JCT) estimates
that tax expenditures will amount to $1,034 billion — over six times the FY2007
federal budget deficit. Despite their widespread use, several concerns about tax
expenditures have been voiced over the past 30 years.
One key “flaw” in the budget process identified by critics of tax expenditures
is the divided treatment of tax expenditures and direct expenditure programs. In the
executive branch, the Treasury draws up the tax expenditure budget for review by the
Office of Management and Budget (OMB), and other administrative agencies prepare
the direct expenditure budgets. In Congress, the Budget and Appropriations
Committees fix the discretionary spending levels for each functional budget category,
while the tax-writing committees (the Ways and Means Committee in the House, and
the Senate Finance Committee) are given a revenue floor. Within this revenue level,
the tax-writing committees can trade off tax rate changes with tax expenditures, and
the appropriations committees can trade-off one direct expenditure program for
another. Typically replacing a tax expenditure with a direct expenditure would
involve moving a bill through multiple committees. Consequently, tax expenditures
often either overlap or conflict with direct expenditures because they have not been
integrated into the budget process for appropriations.
Tax expenditures are often alternatives to other policy instruments such as
grants. Consequently, national social and economic goals are sometimes met through
the tax code rather than through direct expenditures. Regardless, many experts argue
tax expenditures are often less efficient than direct expenditure programs in
promoting these important economic and social goals. At the same time, still other
tax expenditures appear not to meet any social or economic goal. In general, tax
expenditures tend to reduce the progressivity of the income tax system, and add to
the complexity of the tax system from the taxpayer’s point of view. Furthermore,
unlike direct expenditures, the benefits of much of the tax expenditures go to
taxpayers in the upper part of the income distribution, and they often subsidize an
activity for which the taxpayer receives a benefit.
This report will be updated as legislative developments warrant.
Estimates and Trends...............................................2
Revenue Losses, 1974-2007.....................................4
A Word of Caution About Aggregate Revenues Losses................6
Criticisms of Tax Expenditures.......................................8
Tax Expenditures in the Budget Process............................8
Effectiveness of Tax Expenditures...............................11
Fairness and Complexity of the Income Tax........................12
The Changing Composition of Direct and Tax Expenditures...............14
Social Welfare Expenditures........................................16
Social Welfare Direct Expenditures...............................17
Social Welfare Tax Expenditures................................19
Data and Methods Appendix........................................23
2002 Public Use Tax File.......................................23
March 2003 Current Population Survey...........................23
Suits Progressivity Index.......................................23
List of Figures
Figure 1. Tax Expenditures. Tax Revenues, and Outlays as a
Percentage of GDP, FY1974-2007................................5
Figure 2. Composition of Government Support in Five Budget Categories....15
Figure 3. Lorenz Curves............................................24
List of Tables
Table 1. Largest Tax Expenditures for Individuals, FY2007.................4
Table 2. Estimated Average Preparation Times of Tax Forms,
2007 Tax Year...............................................14
Table 3. Social Welfare Direct Expenditures, FY2007....................18
Table 4. Social Welfare Tax Expenditure Estimates, FY2007..............20
Table 5. Pension Coverage and Health Insurance Coverage by Income.......21
Tax Expenditures: Trends and Critiques
Special deductions, exclusions, and exemptions (sometimes characterized as
“loopholes”) have been in the tax code since the passage of the progressive income1
tax in 1913. Since then, over 100 special deductions, exemptions, and credits have
been added to the tax code. Of course, for the first 25 years of the income tax, tax
expenditures were relatively unimportant. Prior to World War II, the federal income
tax was of little economic importance — individual and corporate income tax
receipts amounted to less than 2% of gross domestic product (GDP). By 1945,
however, income tax receipts accounted for over 15% of GDP. And as the income
tax became more economically important, so did the tax subsidies from the special
deductions, exemptions, exclusions, and credits. The Joint Committee on Taxation
(JCT) estimates that these tax subsidies, also known as tax expenditures, will amount
to $1,034 billion in 2007 — over six times the projected 2007 federal budget deficit.2
In the mid-1960s, the Department of the Treasury became interested in tracking
and accounting for these tax subsidies. Indeed, the term “tax expenditures” was first
used at this time.3 With the enactment of the Congressional Budget and
Impoundment Act of 1974 (P.L. 93-344), tax expenditures were officially defined as
“those revenue losses attributable to provisions of the Federal tax laws which allow
a special exclusion, exemption, or deduction from gross income or which provide a
special credit, a preferential rate of tax, or a deferral of tax liability.”4 Both the
Department of the Treasury and the JCT prepare annual lists and estimates of tax
expenditures. Further, the Treasury’s list is included in the President’s annual budget
Tax expenditures are often alternatives to other policy instruments such as
grants. Consequently, national social and economic goals are sometimes met through
1 The United States had an income tax in the years during and immediately following the
Civil War. The income tax was repealed in 1872 because the revenue was no longer needed.
An income tax was re-established in 1894, but was declared unconstitutional by the Supremeth
Court in 1895. After the 16 Amendment was ratified on Feb. 3, 1913, which permitted
Congress to tax income, the graduated income tax was established on Oct. 3, 1913. See Roy
G. Blakey and Gladys C. Blakey, The Federal Income Tax (New York: Longmans, Green
and Co., 1940) for a history of early years of the income tax.
2 Eliminating all of the tax expenditures could result in more or in less than $1,034 billion
in additional federal revenue. This issue of uncertainty in the valuation of aggregate tax
expenditures is discussed in detail below.
3 See Stanley S. Surrey and Paul R. McDaniel, Tax Expenditures (Cambridge, MA: Harvard
University Press, 1985). The first author was the Assistant Secretary for Tax Analysis at
4 2 U.S.C. § 622.
the tax code rather through direct expenditures. Examples of tax expenditures range
from the earned income credit (EIC), which raises the after-tax income of low- and
moderate-income families, to the expensing of multiperiod timber-growing costs,
which lowers the effective tax rate on timber growing.
Several concerns about tax expenditures have been voiced over the past 30
years. First, tax expenditures represent a substantial commitment of support in terms
of forgone revenues on the part of the federal government.5 Second, tax expenditures
are another form of entitlement spending, since they are not examined in the annual
budget process.6 Third, more well-off taxpayers benefit disproportionately from tax
expenditures because of the progressive nature of the income tax system.7 This
report examines the trends in tax expenditures, the arguments for and against tax
exemptions, the composition of tax expenditures, and who benefits from selected tax
Estimates and Trends
Tax expenditures include reductions in tax liability resulting from special tax
provisions. To determine if a provision is a tax expenditure, the JCT defines a
baseline or reference income tax structure referred to as the “normal income tax law,”8
which has a broader concept of income than under U.S. tax law. The committee
staff uses its judgement to distinguish between what are normal income tax
provisions and what are special provisions.
The Department of the Treasury uses a similar procedure to identify tax
expenditures, but uses two baseline income tax structures: the normal income tax
baseline and the reference income tax baseline. The reference tax baseline is closer
to existing tax law and, consequently, identifies fewer tax expenditures. This
baseline has been used by the Treasury since 1982. Prior to 1982, there were few
differences between the tax expenditures lists of the JCT and the Department of the
Treasury, since both used the same baseline. After 1982, the differences between the
two lists have grown. The JCT has used a consistent methodology to define and
estimate tax expenditures over time, whereas the Department of the Treasury’s
methodology has changed from administration to administration.9
5 See, for example, U.S. Government Accountability Office, Government Performance and
Accountability: Tax Expenditures Represent a Substantial Federal Commitment and Need
to Be Reexamined, GAO-05-690, Sept. 2005.
6 See, for example, Paul R. McDaniel and Stanley S. Surrey, “Tax Expenditures: How to
Identify Them; How to Control Them,” Tax Notes, May 24, 1982, pp. 595-625.
7 See, for example, John F. Witte, The Politics and Development of the Federal Income Tax
(Madison, WI: the University of Wisconsin Press, 1985).
8 U.S. Congress, Joint Committee on Taxation, Estimates of Federal Tax Expenditures for
Fiscal Years 2007-2011, joint committee print, 110th Cong., 1st sess., JCS-3-07 (Washington:
9 The current Bush administration, which has questioned the whole concept of tax
Other baseline tax law structures have been proposed.10 One commonly
proposed baseline is the consumption tax baseline. This baseline has fewer built-in
distortions than the income tax. For example, the choice between future
consumption (that is, saving) and current consumption is not biased as it is when
income is taxed. Many of the largest tax expenditures would no longer be defined
as tax expenditures using this baseline concept, and an alternative set of tax
expenditures would be produced. Neither the JCT nor the Department of the
Treasury, however, use this baseline. Furthermore, many observers believe it is
unlikely that the U.S. income tax will be replaced by a consumption tax in the
The value of tax expenditures can be estimated in two ways. The JCT estimates
tax expenditures in terms of revenues lost to the U.S. Treasury. The revenue loss of
a tax expenditure is a straightforward and easily understood concept — it is simply
taxes not paid and represents revenues forgone by the government. In addition to the
revenue loss estimates, the Department of the Treasury also estimates the “outlay
equivalent” of tax expenditures. This concept is not as intuitive as the revenue loss
concept and assumes that the tax expenditures could be replaced by a direct
government outlay.11 The JCT’s revenue loss estimates are used throughout this
The top 16 tax expenditures in terms of revenue losses for FY2007 are listed in
Table 1. These selected tax expenditures account for $844.0 billion of the total
$1,034.1 billion in tax expenditures. By far the largest tax expenditure is for
retirement saving incentives (net exclusion of pension contributions and earnings).
Many of these tax expenditures are exclusions or deductions of employer fringe
benefits and expenses associated with owner-occupied housing.
expenditures, stated in the FY2002 budget: “Because of the breadth of this arbitrary tax
base, the Administration believes that the concept “tax expenditure” is of questionable
analytic value.” See “Tax Expenditures,” ch. 15 in OMB, Analytical Perspectives, Budget
of the United States Government, Fiscal Year 2002, Feb. 2001, p. 61.
10 The choice of the baseline tax law system is probably the most controversial aspect in
defining and measuring tax expenditures. See Douglas A. Kahn and Jeffrey S. Lehman,
“Tax Expenditure Budgets: A Critical View,” Tax Notes, Mar. 30, 1992, pp. 1661-1665, for
a critical review of various baseline tax law structures.
11 See Adam Carasso and C. Eugene Steuerle, “Tax Expenditures: Revenue Loss Versus
Outlay Equivalents,” Tax Notes, Oct. 13, 2003, p. 287.
Table 1. Largest Tax Expenditures for Individuals, FY2007
Dollars P e rcentage
Net exclusion of pension contributions and earnings132.90.97
Reduced tax rates on dividends and long-term capital127.10.93
Exclusion of employer contributions for health care105.70.77
Deduction for mortgage interest73.70.54
Exclusion of capital gains at death51.90.38
Tax credit for children under age 1745.00.33
Earned income credit (EIC)44.70.33
Exclusion of Medicare benefits40.90.30
Deduction of state and local taxes33.90.25
Exclusion of benefits provided under cafeteria plans30.00.22
Exclusion of investment income in life insurance, annuity28.60.21
Exclusion of capital gains on sales of principal residences28.50.21
Exclusion of untaxed Social Security benefits22.40.16
Exclusion of interest on public purpose state and local20.00.15
Deduction for property taxes on owner-occupied16.80.12
Source: Author’s calculations of Joint Committee on Taxation data.
Revenue Losses, 1974-2007
The 30-year trend in aggregate tax expenditure estimates, income tax revenues
(individual and corporate), and federal outlays are expressed as a percentage of GDP
and are displayed in Figure 1. Federal outlays (the dashed line) follow a cyclical
pattern — rising during recessions and falling during the subsequent economic
recoveries. Income tax revenues (the dotted line) tend to follow both changes in tax
laws and the cyclical behavior of the economy. The Economic Recovery Tax Act of
12 See U.S. Congress, Joint Committee on Taxation, General Explanation of the Economic
Partly because of the deep recession in the early 1980s, income tax revenues fell in
relation to GDP. The economic expansion in the mid- to late-1980s did not increase
income tax revenues relative to GDP to any appreciable extent because of the 1981
tax cuts. After the 1990 recession ended, the strong economic expansion in the mid-
to late-1990s increased income tax revenues. Income tax revenues fell after 2001
because of the 2001 recession, and the 2001 and 2003 tax cuts. Income tax revenues
began to pick up in 2004 due to the strengthening economic recovery.
Figure 1. Tax Expenditures. Tax Revenues, and Outlays as a
Percentage of GDP, FY1974-2007
taIncome Tax Revenues
197 4 197 7 1980 1 983 198 6 1989 1 992 1 995 1 998 2 001 2 004 20 07
Source: Author’s calculations of OMB and JCT data.
The trend in aggregate revenue losses from tax expenditures (the solid line) is
displayed in Figure 1. Aggregate revenue losses from tax expenditures depend on
the number of special tax provisions, tax rates, and the level of income. Marginal tax
rates affect revenue losses of tax expenditures by changing the amount of taxes that
would be paid in the absence of tax expenditures. For example, the government
forgoes $50 in tax revenue for each $100 in shielded income if the tax rate is 50%,
but forgoes only $30 if the tax rate is 30%; so even if shielded income remains
constant, the revenue loss can change by changing the tax rate. As the level of
income rises, individuals will be pushed into higher tax brackets with higher
marginal tax rates. Consequently, the revenue loss from shielded income will
increase, holding the amount of shielded income constant. In addition, rising income
Recovery Tax Act of 1981, 97th Cong., 1st sess., Dec. 29, 1981, JCS-71-81 (Washington:
GPO, 1981) for an explanation of the act.
levels probably increase the amount of income shielded from taxation as individuals
increase charitable contributions or take on greater mortgage debt.
The number of tax expenditures increased from 71 in 1974 to 133 by 1987 —
the period when aggregate revenue losses of tax expenditures increased from less
than 6% of GDP to almost 10% of GDP (see Figure 1). In addition to the increase
in the number of tax expenditures, some existing tax expenditures were expanded
such as allowing non-itemizers to deduct charitable contributions.13 The Tax Reform
Act of 1986 (P.L. 99-514) reduced tax rates, eliminated or scaled back various tax
expenditures, and broadened the tax base.14 Consequently, the aggregate revenue
losses of tax expenditures fell dramatically from 9.7% of GDP in 1987 to 5.4% of
GDP just two years later. Essentially the tax rate reductions were paid for by the
elimination or reduction of various tax expenditures. In addition, the tax rate
reductions reduced the revenue losses of the remaining tax expenditures. Throughout
the 1990s, the revenue losses of tax expenditures slowly increased as (1) income
increased during the economic expansion, (2) new tax expenditures were enacted,
and (3) several existing tax expenditures were expanded. By 2007, the aggregate
revenue losses of tax expenditures were estimated to be 7.6% of GDP — about 6.3
times the federal budget deficit in that year.
A Word of Caution About Aggregate Revenues Losses
The revenue loss of each tax expenditure is estimated separately; that is, it is an
estimate of the gain in income tax revenue from the repeal of only the provision
under consideration, assuming no changes in other provisions or in taxpayer behavior
(many tax expenditures were in fact adopted to affect taxpayer behavior). It is,
therefore, unlikely that the aggregate estimated revenue losses of all tax expenditures
as reported in Figure 1 would be equal to the actual gain in income tax revenue from
the repeal of all tax expenditure provisions. There are a variety of reasons the two
would not be equal.
First, taxpayers itemize only if their deductions are greater than the standard
deduction or zero bracket amount. For example, suppose two itemized deductions
are individually less than the standard deduction, but together are greater.
Eliminating one deduction would increase the amount of tax paid by taxpayers
affected because they now take the standard deduction. Elimination of the other
deduction now would have no effect on tax liability. In this case, the sum of the
individual tax expenditure estimates would be greater than the gain in revenue from
elimination of both tax expenditures. In a simulation, elimination of 12 selected tax
expenditures led to the sum of the individual tax expenditures estimates being 17.5%
13 Most of the new and expanded tax expenditures were enacted in the Economic Recovery
Tax Act of 1981 (P.L. 97-34).
14 See U.S. Congress, Joint Committee on Taxation, General Explanation of the Tax Reform
Act of 1986, 100th Cong., 1st sess., May 4, 1987, JCS-10-87 (Washington: GPO, 1987) for
an explanation of the act.
greater than the revenue gain from elimination of all 12 tax expenditures.15 This
simulation ignores, however, unmeasurable changes such as the increase in interest
rates and thus interest income, if tax-exempt bonds were no longer tax-exempt. The
tax-exempt interest reported on tax returns likely understates the additional taxable
income taxpayers would have if this exclusion were eliminated.
Second, eliminating exclusions and deductions could push a taxpayer into a
higher tax bracket. In this case, the sum of the individual tax expenditure estimates
could be less than the revenue gain from elimination of all the tax expenditures. For
example, the elimination of two large tax expenditures — the exclusion of pension
earnings and contributions, and employer contributions for health insurance — would
push many middle-income taxpayers into higher tax brackets, whereas the
elimination of only one may not.
Third, taxpayer behavior may change in response to the elimination of special
tax provisions. Taxpayers would probably look for alternative ways to shield income
from taxation. Consequently, revenue gains from the elimination of tax expenditures
may be less than estimates that incorporate the above interaction effects but ignore
Because of these three effects, simply summing the tax expenditure revenue loss
estimates may not provide an accurate estimate of the effect that elimination of tax
expenditures would have on federal tax revenues. It is unknown if the sum of
individual tax expenditure revenue losses overstates or understates the actual effect
of the simultaneous elimination of all tax expenditures. Nevertheless, examining the
sum of tax expenditure revenue loss estimates over time probably provides a good
approximation of the general trend in the effect of tax expenditures on income tax
revenue. Given the relatively long life of many tax expenditures, it is unlikely that
the bias from simply summing tax expenditures estimates changes much from year
15 This is based on the author’s analysis of the IRS’s 2002 Public Use Tax File. The 12 tax
expenditures are the mortgage interest deduction, the property tax deduction, the deduction
for state and local income taxes, the charitable contribution deduction, the IRA contribution
deduction, the student loan interest deduction, the exclusion of untaxed social security
benefits, the exclusion of tax-exempt interest, the tuition tax credit, the saver’s tax credit,
the child tax credit, and the earned income credit. These tax expenditures were chosen
because they are reported on tax returns and are available in the data file. See the appendix
for a description of this data file.
16 The same logic applies to the budgetary savings from eliminating or scaling back a
program that provides transfer payments to individuals. As the benefits are reduced or
eliminated, individuals may become eligible for other similar government programs or
change their behavior to qualify for these programs. Consequently, the budgetary savings
could be less than expected.
Criticisms of Tax Expenditures
Criticisms of tax expenditures appeared when the term was first coined. Stanley
Surrey, the Assistant Secretary for Tax Policy in the Johnson Administration, argued
that tax expenditures were inferior to direct expenditures in achieving various social
and economic goals, and were a barrier to tax reform that would restore fairness to
the tax system.17 Recent criticisms of tax expenditures often fall into one of three
categories. First, there are those who identify the budget process as the source of the
growth in the use of tax expenditures. Second, many analysts continue to argue that
tax expenditures are less effective than direct expenditures in achieving social and
economic goals.18 Third, some argue that tax expenditures increase the complexity
and reduce the fairness of the income tax, thus undermining public support for the
tax system. These criticisms are examined in turn.
Tax Expenditures in the Budget Process
The “flaw” in the budget process identified by critics of tax expenditures is the
divided treatment of tax expenditures and direct expenditure programs.19 In the
executive branch, the Department of the Treasury draws up the tax expenditure
budget for review by the Office of Management and Budget (OMB). Other
administrative agencies prepare the direct expenditure budgets. For example, the
Department of Health and Human Services prepares the budget for several public
assistance programs for the poor (such as Temporary Assistance for Needy Families
and Medicaid) but has no responsibility for the Earned Income Credit (EIC), a tax
credit, which also benefits low-income families.
The Government Performance and Results Act of 1993 (GPRA, P.L. 103-62)
requires annual evaluations of tax expenditures by the Administration. OMB circular
A-11 directs agencies to consult with the Office of Tax Analysis in the Department
of the Treasury on tax proposals and “be prepared to submit justifications for
continuing or reenacting existing taxes and tax expenditures in the program areas for
which you have primary responsibility.”20 An official in the Clinton Administration,
however, stated that President Clinton’s Department of the Treasury was
“unenthusiastic about performing these evaluations.”21 The Bush Administration also
17 See Stanley S. Surrey, Pathways to Tax Reform (Cambridge, MA: Harvard University
18 The term direct expenditures should not be confused with direct or mandatory spending.
Direct expenditures are simply budgetary spending, which then can be classified as either
discretionary or mandatory spending.
19 See, for example, Victor Thuronyi, “Tax Expenditures: A Reassessment,” Duke Law
Journal, vol. 1988 (1988), pp. 1155-1206.
20 OMB, Circular No. A-11, Preparation, Submission, and Execution of the Budget, Jun.
21 Leonard E. Burman, “Is the Tax Expenditure Concept Still Relevant?” National Tax
Journal, vol. 56, no. 3 (Sept. 2003), p. 624.
has not issued any evaluations of tax expenditures as part of GPRA, and,
furthermore, has questioned the whole concept of tax expenditures.22
The congressional budget process also divides the spending and revenue
functions.23 The Budget and Appropriations Committees fix the discretionary
spending levels for each functional budget category. The tax-writing committees (the
Ways and Means Committee in the House, and the Senate Finance Committee) are
given a revenue floor. Within this revenue level, the tax-writing committees can
trade off tax rate changes with tax expenditures. The Appropriations Committees can
trade off one direct expenditure program for another, but no committee can trade off
a tax expenditure for a direct expenditure in a particular category. The two tax-
writing committees could, however, replace tax expenditures with mandatory
spending programs within their jurisdiction. But typically replacing a tax expenditure
with a direct expenditure would involve moving a bill through multiple committees.
The Budget Enforcement Act of 1990 (BEA) indirectly reduced some of the
advantages that tax expenditures had in the budget process. The BEA not only set
discretionary spending caps, but also set a pay-as-you-go (PAYGO) requirement for
mandatory spending and tax changes. PAYGO required that any new increase in
mandatory spending or decrease in tax revenues could not add to the deficit. The
BEA was revised and extended throughout the 1990s, but most of its provisions
expired at the end of FY2002.
Critics of tax expenditures argue that the two tax-writing committees and the
Department of the Treasury (especially the Internal Revenue Service) “often lack the
expertise, background, and staff resources” to design effective tax expenditures to
meet desired economic and social goals.24 Consequently, these critics suggest closer
coordination between the Department of the Treasury and the other agencies, and
between the tax-writing committees and other committees to make better use of
program area expertise. But they note that “efforts at coordination between
government agencies often produce delay and confusion rather than cooperative
In contrast to these critics, there are those who argue that the Treasury and the
congressional tax-writing committees are less prone to interest group capture than the
other executive agencies and congressional committees. One analyst concludes that
22 See, for example, U.S. Government Accountability Office, Government Performance and
Accountability: Tax Expenditures Represent a Substantial Federal Commitment and Need
to Be Reexamined, GAO-05-690, Sept. 2005, in which they recommended that tax
expenditures be subjected to systematic reviews and performance evaluations. The Bush
Administration appears to regard the elimination of tax expenditures as tax increases rather
than as the elimination of tax loopholes, special government grants, and special subsidies.
See the Bush Administration’s response to GAO’s report in appendix II of the GAO report.
23 See CRS Report 98-721 GOV, Introduction to the Federal Budget Process, by Robert
Keith and Allen Schick, for a review of the budget process.
24 See, for example, Stanley S. Surrey and Paul R. McDaniel, Tax Expenditures (Cambridge,
MA: Harvard University Press, 1985), pp. 106-107.
“the committees and agencies that design and administer tax subsidies are less prone
to capture by clientele groups ... and are better positioned to make decisions informed
by expertise than their direct expenditures counterparts.”26
A recent nonpartisan expert panel has questioned the extensive use of tax
expenditures.27 The panel recommended that any formal justification for new tax
expenditures should answer the following questions:28
!Why is a government program necessary at all?
!What objectives is the tax break meant to accomplish, and how will
success or failure be measured?
!What evidence can be cited that suggests the tax break will
accomplish these objectives at an acceptable cost?
!Why is a tax break better than a direct spending program for
accomplishing this purpose?
The expert panel noted that this is similar to the OMB circular A-11 requirements
(which were initially considered important but have fallen into disuse over the years).
The various analyses of tax expenditures suggest that neither the executive branch
nor the Congress appear to have much enthusiasm for a comprehensive examination
of tax expenditures.
Several analyst have observed that tax expenditures resemble entitlement
spending in many respects.29 First, everyone who qualifies receives the benefits of
tax expenditures. Second, jurisdiction for tax expenditures and many entitlement
programs is with the two congressional tax-writing committees. From time to time,
commissions have been set up to examine entitlement spending and recommend
reforms to reduce the long-term costs of these programs. Yet tax expenditures are
rarely discussed within the context of entitlement reform. In 1994, however, some
members of the Bipartisan Commission on Entitlement and Tax Reform thought that
entitlement reform should also include reform of tax expenditures.30 But the
commission was unable to agree on a set of specific proposals to reduce the long-
term budget deficit problems associated with entitlements.
26 Edward A. Zelinsky, “James Madison and Public Choice at Gucci Gulch: A Procedural
Defense of Tax Expenditures and Tax Institutions,” Yale Law Journal, vol. 102, no. 5 (Mar.
27 The Century Foundation Working Group on Tax Expenditures, Bad Breaks All Around
(New York: The Century Foundation Press, 2002).
28 Ibid., pp. 28-29.
29 See, for example, Eric J. Toder, “Tax Cuts or Spending — Does it Make a Difference?”
National Tax Journal, vol. 53, no. 2 part 1 (Sept. 2000), pp. 361-372; and Christopher
Howard, The Hidden Welfare State (Princeton, NJ: Princeton University Press, 1997).
30 Bipartisan Commission on Entitlement and Tax Reform, Final Report to the President
(Washington, DC: Government Printing Office, Jan. 1995). The commission was composed
of 22 Members of Congress and 10 members from the private sector.
Effectiveness of Tax Expenditures
Many tax expenditures were designed to meet important economic and social
goals, while others appear to lack clear economic or social justification.31 An expert
panel suggests that tax expenditures are justified when they (1) correct a market
failure, (2) are appropriately targeted, (3) do not unduly compromise the progressivity
of the income tax, (4) do not add excessively to the complexity of the income tax, (5)
avoid economic disruptions, and (6) are more cost-effective than a direct expenditure
program. This panel proposed a “dirty dozen” list of tax expenditures that could
easily be eliminated, ones which neither correct market failures nor help
disadvantaged groups such as timber subsidies and special rules for employee stock
ownership plans.32 The panel further argued that many tax expenditures are often less
efficient than direct expenditure programs in promoting important economic and
Tax expenditures often either overlap or conflict with direct expenditures
because they have not been adequately integrated into the budget process for
appropriations.33 Some current housing tax expenditures overlap with Department
of Housing and Urban Development direct expenditures to encourage building low-
income housing. One analyst notes a situation during the 1980s when some tax
expenditures subsidized dairy production and encouraged dairy herd expansion, while
the Department of Agriculture paid dairy farmers to reduce production and slaughter
Large tax expenditures range from the earned income credit (EIC) to saving
incentives. The EIC, which is targeted to low-income working families, is a
refundable tax credit, with FY2007 estimated tax expenditures of $44.7 billion and
direct expenditures of $38.3 billion.35 Most analysts agree that the EIC has been
notably successful in encouraging work among disadvantaged families and in lifting
the after-tax income of working families.36 Total tax expenditures for saving
incentives to boost retirement savings are estimated to total $132.9 billion for
FY2007. In contrast to the cost to the government, personal saving for FY2007 was
31 See The Century Foundation Working Group on Tax Expenditures, Bad Breaks All
Around (New York: The Century Foundation Press, 2002).
32 Ibid., p. 23. The expert panel also listed a group of tax expenditures called the
“Troublesome Ten,” which they note are problematic but may have some economic
33 Zhicheng Li Swift, Managing the Effects of Tax Expenditures on National Budgets, World
Bank Policy Research Working Paper no. 3927, May 2006.
34 Victor Thuronyi, “Tax Expenditures: A Reassessment,” Duke Law Journal, vol. 1988
(1988), p. 1161.
35 Tax credits directly offset tax liability. In the case of a refundable tax credit, if the tax
credit is greater than total tax liability, then the government sends a check to the tax filer.
36 See, for example, V. Joseph Hotz, Charles H. Mullin, and John Karl Scholz, Examining
the Effect of the Earned Income Tax Credit on the Labor Market Participation of Families
on Welfare, National Bureau of Economic Research Working Paper no. 11968, Jan. 2006.
less than $60 billion.37 Research has shown that personal saving has been fairly
unresponsive to tax incentives, and such incentives may substantially decrease public
saving (that is, increase the budget deficit): the long-term net effect on national
saving is likely negative.38
Early theoretical and empirical research suggested that tax expenditures for
charitable contributions would be more effective than a comparable direct
expenditure program.39 Subsequent research challenged the theoretical justification
for the charitable contribution tax expenditure.40 Furthermore, recent empirical
research indicates that a taxpayer’s charitable contributions may not be as sensitive
to the tax system as originally estimated.41
Fairness and Complexity of the Income Tax
Ideally, commentators suggest that tax policy should be structured to meet
several basic principles or goals.42 Most importantly, a tax system should raise
adequate revenue to run the government and meet the needs of the governed.
Second, the tax system should be fairly simple and comprehensible. Third, a basic
sense of fairness suggests that those with equal ability to pay taxes should pay equal
taxes. This is sometimes referred to as horizontal equity. Lastly, many argue that the
principle of vertical equity or progressivity — those with greater ability to pay taxes
should pay a greater proportion of their income in taxes — is important.43 This last
principle, however, is somewhat controversial, even though about two-thirds of
37 Federal Reserve Board, Flow of Funds, table F.8, Dec. 6, 2007.
38 See CRS Report RL33482, Saving Incentives: What May Work, What May Not, by
Thomas L. Hungerford.
39 See Martin Feldstein, “A Contribution to the Theory of Tax Expenditures: The Case of
Charitable Giving,” in Henry J. Aaron and Michael J. Boskin, eds., The Economics of
Taxation (Washington, DC: Brookings Institution, 1980), pp. 99-121.
40 Patrick A. Driessen, “A Qualification Concerning the Efficiency of Tax Expenditures,”
Journal of Public Economics, vol. 33 (1987), pp. 125-131; and Martin Feldstein, “The
Efficiency of Tax Expenditures: Reply,” Journal of Public Economics, vol. 33 (1987), pp.
41 See William C. Randolph, “Dynamic Income, Progressive Taxes, and the Timing of
Charitable Contributions,” Journal of Political Economy, vol. 103, no. 4 (1995), pp. 709-
738; and Kevin Staton Barrett, Anya M. McGuirk, and Richard Steinberg, “Further
Evidence on the Dynamic Impact of Taxes on Charitable Giving,” National Tax Journal,
vol. 50, no. 2 (June 1997), pp. 321-334. However, Gerald E. Auten, Holger Sieg, and
Charles T. Clotfelter, “Charitable Giving, Income, and Taxes: An Analysis of Panel Data,”
American Economic Review, vol. 92, no. 1 (Mar. 2002), pp. 371-382 present evidence that
the sensitivity estimates may be larger than reported in these two articles but still lower than
42 See C. Eugene Steuerle, Contemporary U.S. Tax Policy (Washington: The Urban Insitute
Press, 2004) for a more extensive discussion of the principles.
43 Another principle frequently cited is minimizing distortions of taxpayer behavior.
Without a doubt, taxes can distort market behavior, but they can also be used to correct
distortions due to market failures.
Americans believe “that people with high incomes should pay a larger share of their
income in taxes than those with lower income.”44
Tax expenditures or the special exclusions, exemptions, and deductions from
income often conflict with one or more of these tax policy principles. One tax
analyst, for example, simply states “tax expenditures make the tax system worse
according to the goals of fairness, efficiency, and simplicity.”45 In addition, a Wall
Street Journal columnist noted that tax breaks generate “a lot of hassle and
complexity that chews up time and money. And, I suspect, the government creates
tax breaks that are claimed far more often by sophisticated, upper-income taxpayers
than by others.”46 This suggests that tax expenditures tend to reduce the progressivity
of the income tax system.
As an example of the complexity that tax breaks or loopholes add to the tax
system, consider the individual income tax forms. Taxpayers must file one of three
income tax forms: form 1040EZ, form 1040A, or form 1040. The easiest form to fill
out is the form 1040EZ; it is for taxpayers who claim no dependents, do not itemize
deductions, do not claim adjustments to income, and receive only selected types of
income. Next, form 1040A is for taxpayers who can take advantage of some tax
breaks but do not itemize deductions. Last, form 1040 is for taxpayers who receive
income from a variety of sources, can take advantage of any tax break, or itemize
deductions. Table 2 shows the estimated average preparation times for the three
income tax forms for the 2007 tax year. The amount of time that must be devoted to
filling out each of the forms increases dramatically with the number of tax breaks the
taxpayer can take advantage of. Clearly, tax expenditures add to the complexity of
the tax system from the taxpayer’s point of view.
44 CRS tabulation of the General Social Survey, 1972-2004, National Opinion Research
Center. In 2000, 65% of the survey respondents said higher-income people should pay a
much larger share or a larger share of income in taxes than lower-income people. Similar
survey results were reported in 1987.
45 Eric J. Toder, “Tax Cuts or Spending — Does It Make a Difference?” National Tax
Journal, vol. 53, no. 3, part 1 (Sept. 2000), p. 362.
46 David Wessel, “Tons of Tiny Tax Breaks Prove to Be Addictive and ...Taxing,” Wall
Street Journal, July 13, 2006, p. A2.
Table 2. Estimated Average Preparation Times of Tax Forms,
2007 Tax Year
Form 1040, Schedules A and D, and other forms and33.5
Form 1040A, schedules and worksheets, and Form10.4
Tax expenditures often conflict with the goal of horizontal equity since the
deductions, exemptions, exclusions, and tax credits are typically targeted or claimed
by a subset of taxpayers. One analysis shows that the EIC and the child tax credit
provide much more favorable treatment for low-income families with children than
for other families with similar income.47 The same analysis suggests that phasing out
itemized deductions, personal exemptions, and the child tax credit could increase
horizontal equity at the upper end of the income distribution. With the demonstrated
inequities in the U.S. income tax system, it is hardly surprising that almost half of the
respondents in a recent survey said the federal income tax is either not too fair or not
fair at all.48
The Changing Composition of Direct
and Tax Expenditures
Tax expenditures represent an important share of total government support for
many economic and social goals. Over time, national priorities change as do the
favored policy instruments to address new and existing priorities. Consequently, the
importance of tax expenditures in providing government support in meeting various
goals will likely change as well. For example, one analyst has documented that tax
expenditures have been increasingly used to promote social policy goals instead of49
business investment. But how does the changing composition of tax expenditures
interact with direct expenditures? Have tax expenditures become more important
than direct expenditures in providing government support?
The first set of two bars in Figure 2 shows the proportion of total government
support accounted for by direct expenditures (the upper gray bar) and tax
expenditures (the lower black bar). In 1980, tax expenditures accounted for 23% of
47 Jane Gravelle and Jennifer Gravelle, “Horizontal Equity and Family Tax Treatment: The
Orphan Child of Tax Policy,” National Tax Journal, vol. 59, no. 3 (Sept. 2006), pp. 631-
48 The Survey conducted by the Gallup Organization, April 1-April 16, 2006.
49 Eric J. Toder, “The Changing Composition of Tax Incentives: 1980-1999,” in Proceedings
of the National Tax Association 91st Annual Conference, 1998.
total government support, and the percentage increased to 27% by 2007.50 This
relatively small change in the composition of government support, however, masks
marked changes in how the government offers support to meet various goals. Over
the past 27 years, there have been important changes in the composition of
government support in various functional categories.51
Figure 2. Composition of Government Support in Five Budget
19 80 2 00 7 19 80 20 07 1 98 0 2 00 7 19 80 20 07 1 98 0 2 00 7 19 80 2 00 7
Tota l Co mmer c e Income Educa ti on, Hea l t h Ge n e r a l
and Sec urit y Tr ai ni ng, G o ver nm ent
Source: CRS calculation of OMB and JCT data.
Figure 2 also shows the composition of direct and tax expenditures in five
functional categories; tax expenditures in these functional categories currently52
account for almost 90% of total tax expenditures. The largest functional category
of tax expenditures is commerce and housing, which included 37% of total tax
expenditures in 2007. Tax expenditures in this function include the mortgage interest
deduction, the deduction for property taxes on owner-occupied residences, and the
exclusion of investment income on life insurance and annuity contracts, among
others. Most government support in this function comes through tax expenditures:
50 All years refer to fiscal years.
51 Federal government outlays are divided into 18 broad functions or categories, which
provide a comprehensive basis for arranging the federal budget.
52 Direct expenditures in these five categories, on the other hand, account for 27% of total
The next largest category is income security which includes saving incentives
(such as the exclusion of pension contributions and earnings) as tax expenditures, and
many public assistance programs as direct expenditures. Tax expenditures as a
percentage of government income security support increased from 22% in 1980 to
35% by 2007. In the education, training, employment, and social services function
(the fourth set of bars in Figure 2), the importance of tax expenditures in government
support doubled from 28% in 1980 to 61% by 2007, which represents a major change
in how government support is offered in this functional category.
The fifth set of bars displays the composition of government health support.53
Health tax expenditures include the exclusion of employer contributions for health
care and insurance, while direct expenditures include spending for the Medicaid
program. The percentage of government support from tax expenditures fell from
44% in 1980 to about 34% in 2007. This is probably due to the dramatic growth in
Medicaid outlays over this 27-year period.
The final functional category with a significant proportion of tax expenditures
is general government and fiscal assistance. This category includes, for example, the
deduction of state and local government income taxes as a tax expenditure, and
funding for legislative functions as a direct expenditure. The importance of tax
expenditures in this functional category increased from 62% in 1980 to over 75% in
Overall, the proportion of government support offered through tax expenditures
increased by only three percentage points over the past 27 years. But almost 90% of
tax expenditures occur in only 5 of 18 broad functional categories. In these five
functions, the importance of tax expenditures changed significantly between 1980
and 2007. In four of the five, tax expenditures became a more important source of
government support. Furthermore, almost all of the government support for
commerce and housing is offered through the tax system.
Social Welfare Expenditures
A large proportion of tax expenditures are directed to addressing various social
goals and could be considered a form of government social welfare expenditures.
Social welfare policy can be broadly defined as government policies intended to
improve the well-being of individuals and families. These policies can include
government programs that promote education, health, housing, and income security.
Often social welfare direct expenditures (i.e., entitlements) are targeted to lower-
income individuals and families. But social welfare tax expenditures are frequently
criticized as having an “upside down” or regressive subsidy feature.54 In addition, the
53 Medicare is not included in this functional category.
54 See, for example, Stanley S. Surrey, “Tax Incentives as a Device for Implementing
Government Policy: A Comparison with Direct Government Expenditures,” Harvard Law
Review, vol. 83, no. 4 (Feb. 1970), pp. 705-738; and Eric J. Toder, “Tax Cuts or Spending
— Does it Make a Difference?” National Tax Journal, vol. 53, no. 3 part 1 (Sept. 2000), pp.
beneficiaries of social welfare tax expenditures often receive a benefit from the
activity being subsidized. For example, state and local taxes can be deducted from
income, but state and local taxes are used to pay for services that benefit the taxpayer
such as police and fire protection. Both social welfare direct and tax expenditures
are examined in this section.
Social Welfare Direct Expenditures
The 15 largest social welfare direct expenditures are listed in Table 3 along with
FY2007 federal outlays. Two of these (the earned income and the child tax credits)
are refundable tax credits in that if the tax credit exceeds the taxpayer’s tax liability,55
the government sends the taxpayer a check for the difference. Consequently, these
two tax credits are both a tax expenditure and a direct expenditure. The federal
outlays shown in Table 3 are the direct expenditure portion of these tax credits.
The total federal outlays for these 15 direct expenditure programs were $1,397.7
billion in FY2007. The Social Security program (both retirement and disability) is
the largest with direct expenditures of $584.3 billion. The next two largest social
welfare direct expenditure programs are the two health programs — Medicare and
Medicaid. These three programs account for over 80% of the social welfare direct
expenditures. The remaining programs are considerably smaller (each has less than
$40 billion in federal outlays).
These social welfare programs can be divided into public assistance and social
insurance programs. Public assistance programs are generally means-tested and
directed to the poorest individuals and families in society. They have been described
as redistributing income from the “haves” (taxpayers) to the “have nots.”56 Social
insurance programs usually are an earned right based on work histories and have been
described as redistributing income from good times (employment) to bad times57
(unemployment, disability, injury, or retirement).
55 The child care tax credit is only refundable for taxpayers with at least one qualifying
56 See Thomas L. Hungerford, “The Distribution and Anti-Poverty Effectiveness of U.S.
Transfers, 1992,” Journal of Human Resources, vol. 31, no. 1 (1996), p. 264.
Table 3. Social Welfare Direct Expenditures, FY2007
(billions of dollars)
Expendi t ures Index
Medicare 380.1 0.56
Medicaid 181.5 0.60
Earned Income Tax Credit38.3 —
Supplemental Security Income (SSI)37.20.79
Temporary Assistance for Needy Families (TANF)17.10.86
Child tax credit16.2 —
Higher Education financial assistance15.5 —
Training and Employment6.9 —
Women, Infants, and Children food program5.2 —
Child support and family support programs4.4 —
Source: OMB and author’s analysis of March 2003 supplement of the Current Population Survey,
Bureau of the Census.
*Includes only federal expenditures for special workers’ compensation program and federal employees
workers’ compensation. Does not include funds provided by state governments.
The last column of Table 3 reports a measure of progressivity for selected direct
expenditure programs. The Suits progressivity index varies between -1 (completely
regressive) to +1 (completely progressive).58 The Suits index is negative if the
benefits from the program are received predominantly by families in the upper part
of the income distribution. It is positive if the benefits are received predominantly
by lower-income families. Lastly, it is zero if the benefits are proportionally
distributed throughout the income distribution. In the nine cases where the Suits
index could be calculated, it is positive and often closer to +1 than to zero, suggesting
that the benefits from these programs are largely received by those in the lower part
of the income distribution — they are fairly well targeted to lower-income
58 See the appendix for a description of the data used for these calculations and a description
of the Suits progressivity index.
individuals and families. These results for 2002 are in accord with results from an
earlier study for 1992.59
Social Welfare Tax Expenditures
Social welfare tax expenditures have been referred to as a hidden welfare state
and rarely show up in the accounts of the nation.60 The largest social welfare tax
expenditures are listed in Table 4. The aggregate estimate of these 19 tax
expenditures amounted to $631.8 billion in FY2007 — about 61% of the total for all
tax expenditures in that year. Many of these tax expenditures include incentives to
encourage retirement saving, charitable contributions, higher education, and home
ownership. Others are the result of the exclusion from taxable income of benefits
from some of the social welfare direct expenditure programs. Also on this list are the
two refundable tax credits — the earned income and child tax credits. The revenue
losses of these two tax credits are greater than their direct expenditures.
The final column of Table 4 reports the Suits progressivity index for selected61
tax expenditures. Unlike social welfare direct expenditures, the Suits index for
these tax expenditures is not uniformly positive. For example, the incentives to
encourage charitable contributions and home ownership are received largely by
taxpayers in the upper parts of the income distribution and have an upside down
59 Thomas L. Hungerford, “The Distribution and Anti-Poverty Effectiveness of U.S.
Transfers, 1992,” Journal of Human Resources, vol. 31, no. 1 (1996), pp. 255-273.
60 Jeffrey P. Owen, “Tax Expenditures and Direct Expenditures as Instruments of Social
Policy,” in Sijbren Cnossen, ed., Comparative Tax Studies (Amsterdam: North-Holland
Publishing Co., 1983), pp. 171-197.
61 See the appendix for a description of the data used for the calculations.
Table 4. Social Welfare Tax Expenditure Estimates, FY2007
(billions of dollars)
Expendi t ure Index
Net exclusion of pension earnings and contributions132.9 —
Exclusion of employer contributions for health insurance105.7 —
Deductibility of mortgage interest on owner-occupied73.7-0.19
Child Tax Credit45.00.31
Earned Income Tax Credit44.70.91
Exclusion of untaxed Medicare benefits40.9 —
Deductibility of non-business state & local taxes33.9-0.41
Exclusion of capital gains on sales of principal28.5 —
Exclusion of Social Security benefits22.40.40
Exclusion of interest on state & local public purpose20.0-0.65
Deduction of state & local property taxes on owner-16.8-0.22
Deduction for medical expenses8.4 —
Exclusion of workers’ compensation benefits7.5 —
Tuition tax credit3.10.44
Exclusion of public assistance benefits2.9 —
Additional standard deduction for blind and elderly1.7 —
Savers’ Tax Credit0.90.68
Deduction of interest on student loans0.90.29
Source: JCT, and author’s analysis of SOI 2002 Public Use File.
The Suits index could not be calculated for the two largest social welfare tax
expenditures. Data are available, however, to determine which income classes have
pension coverage or employer-provided health insurance and are likely to benefit
from these exclusions. The first column of Table 5 shows the income ranges for the
five income classes; each income class contains about 20% of the families in the
Table 5. Pension Coverage and Health Insurance
Coverage by Income
Percentage ofAverage Tax
Percentage ofFamilies withEmployer-Saving fromExclusion of
Income RangeFamilies withprovided HealthEmployer-
Pension CoverageInsuranceprovided Health
Less than $16,2004.615.5$7.19
$16,200-$30,999 21.4 46.1 $100.06
$31,000-$50,219 45.3 69.8 $317.27
$50,220-$81,513 64.4 81.5 $658.55
$81,514 or more74.386.1$1,482.17
Source: CRS analysis of March 2003 CPS.
The next column of Table 5 reports the percentage of families in each income
class with pension coverage. The proportion with pension coverage increases in
moving from the poorest to the richest income class. Therefore, upper-income
families are much more likely to benefit from the exclusion of pension earnings and
contributions than lower-income families. The next column shows the percentage
of families in each income class with employer-provided health insurance. Again,
those families in the upper part of the income distribution are more likely to benefit
from the exclusion of employer contributions for health insurance. The last column
reports an estimate of the average tax savings for families in each income class from
this exclusion.62 The average tax savings increases in going from the poorest to the
richest income class.63 The progressivity results are consistent with results from an
earlier study for 1977.64
62 The estimate is calculated by multiplying the employer contribution for health insurance
reported in the data file by the reported marginal tax rate. If the employer contribution were
included as taxable income, some taxpayers would be pushed into a higher tax bracket.
Consequently, the estimates reported in Table 5 likely underestimate the tax saving.
63 Broadly similar results are obtained when the sample of all families is restricted only to
families receiving no retirement income and pay taxes.
64 John F. Witte, The Politics and Development of the Federal Income Tax (Madison, WI:
The University of Wisconsin Press, 1985).
The revenue losses from tax expenditures are not trivial. For FY2007, they are
estimated to total $1,034 billion — equivalent to 7.6% of GDP and 40.1% of
estimated federal revenues. In comparison, the Congressional Budget Office projects
that the FY2007 budget deficit was $162 billion — about 16% of the estimated
revenue losses from tax expenditures. In recent years, both the dollar amount and the
number of tax expenditures have been growing.
Tax expenditures, however, are an important source of government support
designated to achieve a variety of economic and social objectives. Of the total
government support directed to achieving social policy goals, tax expenditures
account for almost a third of this support. But unlike social welfare direct
expenditures, the benefits of much of the tax expenditures go to taxpayers in the
upper part of the income distribution, and they often subsidize an activity for which
the taxpayer receives a benefit. Furthermore, research has questioned the
effectiveness of some tax expenditures in achieving the stated social objectives.
Over the years, a number of analysts, academics, and policy makers have voiced
their concern about the growing importance of tax expenditures and their effect on
long-term fiscal problems. Some members of the 1994 Bipartisan Commission on
Entitlement and Tax Reform thought that tax expenditures should be included as part
of reforms to rein in entitlement spending. More recently, the Century Foundation
Working Group on Tax Expenditures recommended that the Administration and the
Congress consider scaling back or eliminating many existing tax expenditures and
exercising restraint in proposing new ones. And the Government Accountability
Office recommended that tax expenditures be subjected to systematic reviews and
performance evaluations. But various observers suggest that neither the executive
branch nor Congress appear to have much enthusiasm for such a comprehensive
The research evidence suggests that a comprehensive examination of tax
expenditures would likely conclude that
!some tax expenditures make use of the strengths of the tax system
and have been successful in achieving their economic or social
!some tax expenditures address valid economic and social goals, but
have been ineffective — direct expenditures may be a more effective
means for achieving the objectives; and
!some tax expenditures do not appear justified on either economic or
Data and Methods Appendix
Two nationally representative data files are used in the analysis on the
progressivity of social welfare expenditures.
2002 Public Use Tax File
The Public Use Tax File is a nationally representative sample of tax returns files
for the 2002 tax year. This is the most recent year of tax return data available to the
public from the Internal Revenue Service (IRS). To protect the identity of individual
taxpayers while preserving the character of the data, the IRS made some changes to
the data. Consequently, while reliable aggregate information can be obtained,
individual taxpayer records in the data file may or may not contain information from
just one tax return. The unit of analysis is the tax return for a taxpayer and IRS-
provided sample weights are used throughout the analysis.
March 2003 Current Population Survey
The Current Population Survey (CPS) is a monthly nationally representative
sample of the noninstitutionalized population. In March of each year, the Census
Bureau includes the Annual Social and Economics supplement with the CPS. The
March 2003 supplement includes detailed information about demographic
characteristics, income, and income sources for the 2002 calendar year. The year was
chosen to match the most recent year of the IRS Public Use Tax File. The unit of
observation is the family, and Census Bureau provided sample weights are used in
Suits Progressivity Index
The progressivity measure used in the analysis is a modified Suits index. The
Suits index was originally developed to measure tax progressivity.65 Since transfers
— either direct transfers or through the tax system — can be considered negative
taxes, the modified Suits index used is the negative of the original index developed
Figure 3 illustrates the principle behind the Suits index. The horizontal axis
measures the cumulative percent of income from poorest to richest. Essentially the
income of taxpayers is ranked from poorest to richest and then summed. The first
of the taxpayers in the sample. The top 40% of cumulative income (from 60% to
100% in the figure) is the income reported by the richest 10% of taxpayers in the
sample. The vertical axis measures the cumulative percent of benefits.
65 See Daniel B. Suits, “Measurement of Tax Progressivity,” American Economic Review,
vol. 67, no. 4 (Sept. 1977), pp. 747-752.
Figure 3. Lorenz Curves
80e fi t
40 Mor tgageInte re st
a t B
Cumulative Percent of Income
The curves shown in Figure 3 are Lorenz curves. The diagonal dashed line in
the figure is the Lorenz curve for a proportionally distributed benefit. It does not
mean that all taxpayers receive the same benefit level; it means that all taxpayers
receive the same proportion of their income in benefits. A progressive benefit (for
example, the EIC) will have its Lorenz curve above the diagonal dashed line (see the
thin line in the figure). This means that those with lower incomes receive benefits
that are a larger proportion of their income than higher income taxpayers.
Conversely, a regressive benefit (for example, the mortgage interest deduction) will
have a Lorenz curve below the diagonal dashed line (see the heavy curve in the
The modified Suits index is calculated from the information contained in the
Lorenz curve diagram. It is equal to negative one plus the ratio of the area under the
Lorenz curve to the area under the diagonal dashed line. For the two tax expenditure
benefits shown in Figure 3, the Suits indices are:
a rea A a rea B area C++
S EIC =− +1
S MID =− +1.
The Suits index varies between -1 (a completely regressive benefit) to +1 (a
completely progressive benefit). A Suits index of zero is for a proportional benefit.