Tax Analysts®Tax Analysts®

My Subscriptions:

Featured News

December 6, 2012
Tax Expenditure Reform: Some Common Misconceptions

Full Text Published by Tax Analysts®

by John L. Buckley

Originally posted on July 18, 2011.

John L. BuckleyJohn L. Buckley is a visiting professor at Georgetown University Law Center. He would like to thank Jon Talisman, Joe Mikrut, Rob Dietz, and Michael Hauswirth for their comments on drafts of this report; Thanh Nguyen and Nicholas Wexler of Georgetown University Law Center for research assistance; and his daughter, Monica Weaver Buckley, for editing her father's work.

The debate on tax reform has focused almost exclusively on tax expenditures. This report will examine some common misconceptions that have become part of that debate and understate the difficulty of comprehensive tax reform. Most tax expenditures are not special interest loopholes. Many are long-standing provisions that survived the scrutiny of the Tax Reform Act of 1986. The large tax expenditure estimates have created unrealistic expectations concerning potential revenue gains. The distributional effect of tax expenditure reform will be more complex than the simple premise that tax expenditures disproportionately benefit upper-income taxpayers. Finally, the political system will require an examination of the collateral consequences of reform.

Table of Contents

I. Introduction

II. Concept of Tax Expenditures

III. Composition of Tax Expenditures

    A. Overview

    B. Largely Long-Standing Provisions

    C. Small Corporate Share

IV. Potential Revenue Gains

    A. Tax Expenditure Estimates

    B. Transition

    C. Details Matter

    D. Limited to Income Tax

V. Distributional Impact

    A. Overview

    B. Change as a Percentage of After-Tax Income

    C. Revisiting the Concept of Static Estimates

    D. Burden Overstatement

    E. Employee Benefits

VI. Collateral Consequences

    A. Housing

    B. Employer-Provided Health Insurance

VII. Conclusion


I. Introduction

Since the enactment of the Tax Reform Act of 1986, there have been several periods during which tax reform was the subject of serious public debate. In 1995 then-House Ways and Means Committee Chair Bill Archer began a series of tax reform hearings. His expressed goal was to develop legislation that would "rip the income tax out by its roots" and replace it with a direct consumption tax. The effort never resulted in a legislative proposal. In 2005 President Bush's advisory panel on tax reform released a major study.1 It briefly received some attention in the press but was ignored by Congress. In 2007 then-Ways and Means Committee Chair Charles B. Rangel, D-N.Y., introduced a tax reform proposal that included a revenue-neutral corporate reform plan that would have substantially reduced the corporate tax rate.2 The bill was greeted with indifference or opposition from the corporate community and went nowhere. Now we are in the middle of another debate on tax reform. This debate differs from those in the recent past in one respect: Some now see tax reform as the source of additional revenues to reduce the deficit.

The current debate on tax reform has its origins in the report of another presidential commission, the National Commission on Fiscal Responsibility and Reform (the Bowles-Simpson commission). The commission's recommendations were approved by a bipartisan majority consisting of 11 of its 18 members but not by the supermajority that would have guaranteed congressional consideration. Its final report endorsed the concept of raising additional revenues through repeal or curtailment of tax expenditures, but the report did not endorse a specific plan for tax reform. Instead, the commission's report included an "illustrative plan" for reform and deficit reduction. The plan would dramatically broaden the base of the income tax through reform of tax expenditures. The resulting additional revenues would be in part devoted to deficit reduction and in part designed to cover the cost of a new round of rate reductions. The plan suggested a top individual tax rate of 28 percent, which coincidentally was the top rate in TRA 1986. The plan was a political compromise, offering large rate reductions to appeal to conservative members, while resulting in substantial increases in revenue. At least in the short run, it was more successful than many expected.

The recommendations of the Bowles-Simpson commission have received a sympathetic reaction from opinion makers in the press. However, there have been dissenting views expressed concerning the commission's approach. Former Reagan Treasury official Bruce Bartlett criticized its approach as being based on the assumption "that all tax expenditures are equally unjustified tax loopholes and that the statutory tax rate is the only rate that matters for fairness or economic efficiency. In fact, many tax expenditures are necessary adjustments to the tax base, and their elimination would create unfairness and reduce economic growth."3 To some extent, his comments reflect the views of the authors of the concept of tax expenditures. In their classic 1985 book Tax Expenditures, Profs. Stanley Surrey and Paul McDaniel argued that "the classification of an item as a tax expenditure does not in itself make that item either a desirable or undesirable provision."

In my opinion, it is easier to deplore tax expenditures in principle than to actually address them in legislation, particularly when the debate is dominated by some common misconceptions. Too often the term "special interest loopholes" has been used as a synonym for tax expenditures. Yet a review of the largest tax expenditures indicates that few are special interest provisions, and most survived the scrutiny of TRA 1986. The large tax expenditure cost estimates have created unrealistic expectations about the potential revenue that could be raised through tax expenditure reform. The distributional effect of tax expenditure reform is more complex than the simple premise that tax expenditures disproportionately benefit upper-income taxpayers. Changes to long-standing provisions of our tax system will have collateral consequences. Before examining those issues, I will take a brief look at the concept of tax expenditures.

II. Concept of Tax Expenditures

The Congressional Budget and Impoundment Act of 1974 (the Budget Act) requires the Congressional Budget Office and the Treasury Department to publish annually a list of tax expenditures and the forgone revenue associated with each tax expenditure. Since the CBO relies on the staff of the Joint Committee on Taxation for revenue analysis, the JCT staff has been responsible for the annual congressional publication. The Budget Act defines tax expenditures as revenue losses attributable to provisions of the federal tax laws that allow a special exclusion, exemption, or deduction from gross income, or that provide a special credit, a preferential rate of tax, or a deferral of tax.4

Clearly, a definition that broad leaves much discretion to the JCT and Treasury. Not surprisingly, the JCT and Treasury have different approaches in determining what is a tax expenditure and have different cost estimates. In this report, I will use the JCT definition and numbers.5 The JCT approach has the benefit of being consistent with the legislative history of the Budget Act. It is also the approach most familiar to those who will be responsible for developing and enacting any tax reform legislation, namely the members of Congress and their staffs.

As Prof. Martin McMahon wrote in a recent article, the best explanation of the concept of tax expenditures comes in Surrey and McDaniel's Tax Expenditures:

    The tax expenditure concept posits that an income tax is composed of two distinct elements. The first element consists of structural provisions necessary to implement a normal income tax, such as the definition of net income, the specification of accounting rules, the determination of the entities subject to tax, the determination of the rate schedule and exemption levels, and the application of the tax to international transactions. These provisions compose the revenue-raising aspects of the tax. The second element consists of the special preferences found in every income tax. These provisions, often called tax incentives or tax subsidies, are departures from the normal tax structure and are designed to favor a particular industry, activity, or class of persons. They take many forms, such as permanent exclusions from income, deductions, deferrals of tax liabilities, credits against tax, or special rates. Whatever their form, these departures from the normative tax structure represent government spending for favored activities or groups, effected through the tax system rather than through direct grants, loans, or other forms of government assistance.6

The current tax reform debate is unique in one respect: It seems to be almost completely focused on tax expenditures. Congress has not restricted itself to consideration of tax expenditures in previous tax reform efforts. The concept of tax expenditures may be a useful way to account for the erosion of our tax base that is caused by special provisions. However, the designation of an item as a tax expenditure has limited relevance in the congressional decision of whether it is an appropriate feature of our tax law.

The definition is overbroad in some ways. Simple decisions about what should compose the base of our income tax can result in tax expenditures to the extent that those decisions are inconsistent with the comprehensive definition of income that is part of the normal income tax. "Normal income tax" is a somewhat misleading term, because it is doubtful that any country has an income tax with a base as broad as that implied by the concept of a normal income tax.

In other respects, the definition is too narrow, largely because the Budget Act definition is limited to special provisions in the income tax. Explicit subsidies like ethanol benefits are not treated as tax expenditures, because they are delivered through an excise tax benefit, not an income tax benefit. Special benefits in the estate tax, such as the favorable farm valuation rules, also are not considered tax expenditures.

To some extent, determinations of what is a tax expenditure are inconsistent. The foreign tax credit is a feature of our system designed to prevent double taxation of foreign-source income. Appropriately, it is not considered a tax expenditure. But the deduction for state and local income taxes is treated as a tax expenditure even though the legislative history for the Revenue Act of 1964 indicates that the policy rationale for the deduction, like the FTC, is to prevent double taxation:

    For state and local income taxes, continued deductibility represents an important means of accommodation when both the state and local governments on the one hand, and the federal government on other hand, tap the same revenue source -- sometimes to a significant degree. A failure to provide deductions in that case could mean an extremely heavy combined burden of state, local, and federal income taxes.7

Also, tax avoidance transactions are not reflected in tax expenditure estimates unless the avoidance involves manipulation of a provision that is itself a tax expenditure. For example, tax avoidance techniques that result in an overstatement of the research credit would be reflected in the cost estimate for the credit because it is a tax expenditure. However, manipulation of FTC rules will not be reflected in tax expenditure totals unless the manipulation results in an increased deferral of foreign income, which is considered a tax expenditure.

The above discussion is not intended to be critical of how the JCT or Treasury analyze tax expenditures. They are required to be consistent with the Budget Act definition and its legislative history. However, a tax reform effort should not be limited to an examination of tax expenditures. Tax expenditures are reductions in revenue as the result of deliberate policy decisions made by Congress. It is appropriate to reexamine those policy decisions in a tax reform effort. But tax reductions resulting from the ability of creative tax lawyers and others to take advantage of features of the "normal" income tax also should be part of the debate.

III. Composition of Tax Expenditures

A. Overview

There is no better way to dispel the notion that there are hundreds of billions of dollars of special interest tax expenditures than to simply list the major tax expenditures. In this report, I will focus on individual tax expenditures because they account for more than 90 percent of the total. In the type of tax reform envisioned by the Bowles-Simpson commission, only individual tax expenditures are large enough to cover the proposed deficit reduction and rate adjustments.

Table 1 is a list of the top 10 categories of individual tax expenditures. (See the appendix for the components of each category.) These categories account for roughly 95 percent of the total annual individual tax expenditures ($953 billion of the $1.025 trillion).

                                Table 1
                    (numbers are in billions (USD))

 Capital gains and dividends                          $187
 Health (excluding payroll tax effect)                $181.3
 Pension and other retirement savings                 $167.3
 Owner-occupied housing                               $124
 Family-related benefits                              $101.9
 State and local government (non-housing)              $58
 Employee benefits other than health and pensions      $51
 Charitable deduction                                  $44.1
 Exclusion of government transfer payments             $28.5
 Exclusion of military and veterans benefits           $10

Table 1 and the appendix are based on the JCT estimates for 2008 that were contained in the 2007 JCT tax expenditure pamphlet.8 Those estimates do not reflect the effect of the recession and do not include the tax items that were recently enacted as part of economic stimulus legislation. I therefore believe that estimates in the 2007 pamphlet are a better measure of baseline tax expenditures than the estimates in later pamphlets.

At this point, it is appropriate to raise a significant caveat about the tax expenditure numbers. The numbers are a reasonable measure of the magnitude of tax expenditures, but they often bear little relationship to the actual revenue that would result from a tax reform effort focused on repealing or curtailing tax expenditures. Even with that caveat, the tax expenditure estimates offer some interesting insights.

B. Largely Long-Standing Provisions

The total amount of tax expenditures for 2008 was approximately $1.1 trillion, about 80 percent of total individual and corporate income tax receipts at that time ($1.45 trillion). Tax expenditures are substantially higher today than immediately after TRA 1986, both in total size and as a percentage of receipts.9 But contrary to conventional wisdom, the growth in the total cost of tax expenditures since 1986 is not attributable to an explosion in the number of tax expenditures.

A substantial portion of the growth since 1986 is an indirect effect of the tax rate increases since 1986 and changes in the economy, such as the expanding size of the healthcare sector. Congress also contributed to the growth through three major expansions of individual tax expenditures since 1986: lower rates for capital gains and dividends, the increase of tax benefits for families with children, and the liberalization of retirement savings provisions. With those exceptions, all of today's major individual tax expenditures are substantially similar to ones that existed immediately after the passage of TRA 1986.

Since 1986 there have been two major legislative expansions of corporate tax expenditures: the section 199 domestic manufacturing deduction and liberalization of the international tax rules. Aside from those items and several uses of bonus depreciation for economic stimulus, the major corporate tax expenditures are substantially the same today as they were immediately following the passage of TRA 1986.

It is true that the JCT tax expenditure pamphlets in recent years show a sharp increase in the number of provisions characterized as tax expenditures, almost doubling since 1986. But much of the increase in number is attributable to recent changes in method: "notably expanded breakouts of certain provisions formerly listed as single tax expenditure."10 Setting those changes aside, there still is a significant increase in numbers since 1986, largely in energy- and housing-related areas. The cost of those new benefits is a relatively small portion of the $1.1 trillion in tax expenditures that we have today.

There is one large message implicit in the above discussion. Any tax reform effort based on raising significant revenues from the repeal or curtailment of tax expenditures will have to address issues that Congress avoided in its development of the comprehensive and much-praised TRA 1986.

C. Small Corporate Share

Another noteworthy item is the small portion of tax expenditures attributable to taxable corporations. In 2008 approximately 8 percent of tax expenditures benefited corporations even though corporate tax receipts were approximately 25 percent of individual income tax receipts. As a percentage of corporate receipts, corporate tax expenditures have declined since (and immediately following the passage of TRA 1986).11 Also, corporate tax expenditures include items like tax-exempt bonds, charitable deductions, and low-income housing credits of which the corporation claiming those benefits is not the main beneficiary. The corporation is being used as an intermediary to deliver benefits to others. It receives a portion of the benefit in its role as an intermediary, but the bulk of the tax benefit accrues to others.

The small corporate share of tax expenditures is further evidence of why the suggestion that there are hundreds of billions of special interest tax expenditures is not factual.

IV. Potential Revenue Gains

The large tax expenditure estimates have created some expectations about the ability to raise large amounts of revenue from the elimination or reform of tax expenditures. That is unfortunate because once Congress begins serious consideration of tax reform, the tax expenditure estimates have no relevance. The relevant numbers are the JCT revenue estimates for the proposed legislation. In this section of the report, I will examine why those revenue estimates will sharply differ from the corresponding tax expenditure estimates. In sum, I believe there will be a mixture of good and bad news when it comes to the revenue potential. On the individual side, I think the tax expenditure estimates grossly overstate the potential revenue. When it comes to corporate preferences, the tax expenditure estimates may slightly understate the potential revenue.

A. Tax Expenditure Estimates

1. Static nature of tax expenditure estimates. Tax expenditure estimates are static: "Taxpayer behavior is assumed to remain unchanged for tax expenditure estimate purposes."12 In contrast, revenue estimates are a form of microeconomic analysis in which anticipated behavioral responses are taken into account. Since the normal response to a tax increase is an attempt to rearrange one's affairs to avoid or mitigate the tax increase, the cost estimate for many of the large individual tax expenditures will be significantly larger than the revenue estimate for repeal of the provision.

a. Capital gains/dividends. The clearest example of the difference between static tax expenditure estimates and revenue estimates comes in the area of capital gains and dividends. The 2008 tax expenditure estimate for dividends and capital gains totaled $187 billion, consisting of $128 billion attributable to the preferential rates and $59 billion attributable to the step-up in basis at death and the deferral of tax on transfers by gift.

Congress has attempted to limit the benefits of basis step-up on two separate occasions. TRA 1976 contained a version of carryover basis at death that was repealed retroactively several years later. The estate tax repeal in the Economic Growth and Tax Relief Reconciliation Act of 2001 (EGTRRA) was accompanied by another version of carryover basis. Opposition to that version of carryover basis was one reason there was little political support for estate tax repeal last year. That history leaves me reasonably certain that basis step-up and deferral of tax on gift transfers will remain intact even in an aggressive tax reform effort, leaving $128 billion attributable to the preferential rates.

The derivation of that estimate is fairly simple. The JCT took the current level of dividend payments and capital gain realizations and distributed those amounts among income categories. Then it multiplied the amount distributed to any income category by the difference between the ordinary marginal rate for the category and the preferential rate applicable to capital gains and dividends. The tax expenditure cost is the sum of the amounts determined for each income category. It is a static estimate that assumes no change in the level of dividend payments or capital gain realizations.

A revenue estimate for a change in the capital gains rate is far more complicated. There are two large benefits involved in the taxation of capital gains. The preferential rates are the obvious benefit. The other benefit, the realization principle, is the tax deferral allowed until gains are realized. The second benefit creates a lock-in effect that means there are fewer realizations of capital gains than what would occur in the absence of an income tax. If you reduce the capital gains tax rate, the lock-in effect is reduced and some of the cost of the rate reduction is offset by tax on the assumed higher level of realizations. The opposite occurs if the capital gains rate is increased.

Evidence that much of the $128 billion static estimate would disappear in an actual revenue estimate is the result of a controversy over revenue estimates in the early 1990s. In 1990 the administration of President George H.W. Bush proposed a 30 percent reduction in the capital gains tax, from a maximum rate of 28 percent to 20 percent. The JCT estimated that the change would result in a revenue loss of $11 billion over the first five years. In contrast, the Bush Treasury estimated that the change would raise $12 billion over the same five-year period. The Democrats controlled Congress at the time, and there were suggestions that the JCT estimate was influenced by that. In an attempt to dispel the suggestion of political interference in the estimating process, the JCT issued a pamphlet explaining its method for making the estimate.13 In that pamphlet, the JCT said that the difference between its estimate and Treasury's was a result of different assumptions about the lock-in effect. Treasury assumed that the rate reduction would cause a higher level of realizations. Its assumptions implied that a 23 percent capital gains rate would maximize the revenue from the taxation of capital gains. With a lower assumption about induced realizations, the JCT estimated that the revenue-maximizing rate was 28.5 percent.

Much time has passed since those competing 1990 estimates, but it appears the JCT assumptions have not significantly changed.14 Treasury, however, has modified its assumptions so they are now close to the JCT assumptions. Another thing has changed: An additional 3.8 percent tax on capital gains and other investment income has been imposed under the healthcare reform legislation enacted last year. The new tax would suggest that the revenue-maximizing rate for capital gains in the income tax would now be approximately 25 percent.

A recent revenue estimate shows that the one-year cost for extending the lower rates for dividends is approximately $13 billion. That leaves the question of how much of the remaining $115 billion could be raised by increasing the capital gains tax rate by 10 points rather than the 20-point increase implicit in taxing capital gains at the ordinary marginal rates, which was the basis for the tax expenditure estimate. It would be convenient, but not accurate, to assume that the 10-point increase would eliminate half the tax expenditure estimate. Not all capital gains would fall into the top marginal tax bracket of 35 percent. Also, the minimum tax results in a 22 percent maximum capital gains rate for individuals with incomes in the phaseout range for the alternative minimum tax exemption. In 2010 the phaseout range for joint returns was between $150,000 and $440,000. According to the JCT, approximately 70 percent of individuals with incomes between $200,000 and $500,000 in adjusted gross income will be subject to the AMT even with the AMT patch enacted in 2010. Significant amounts of capital gain income in that income category are taxed at a 22 percent effective rate. For those capital gains, the static tax expenditure estimate is based on a 13-point, not a 20-point, increase. However, it is safe to say that a substantial portion of the static $115 billion will disappear with a 10-point increase in the maximum capital gains rate.

The remaining question is what portion of the remaining static gain would actually be collected from a 10-point increase in the rate. A recent study of the JCT estimates of the elasticity of capital gain realizations suggests that the JCT would estimate that 24 percent of the static gain would actually be realized in additional receipts.15

I am confident that a 10-point increase in the maximum capital gains rate will raise annual revenue that will be significant by almost any standard. Only in comparison with the annual $187 billion tax expenditure estimate will the amount seem small.

b. Mortgage interest deduction. Another example of the large differences that will exist between tax expenditure estimates and revenue estimates is the mortgage interest deduction. In this case, the revenue estimate will depend on projections of how many taxpayers can avoid the impact of repeal by divesting other assets and paying down mortgage debt or by merely shifting leverage to other assets.

There have been many academic studies on the level of portfolio adjustments that could result from repeal of the mortgage interest deduction. A recent study by James M. Poterba and Todd M. Sinai contains a good summary of the consensus view:

    In light of the sensitivity of household balance sheets to the tax treatment of mortgage debt, it seems implausible to hold the loan-to-value fixed when examining the revenue effect of eliminating the deductibility of mortgage interest. Without deductibility, households with both financial assets and mortgages would be borrowing at the pretax rate of return but investing at the after-tax rate of return. Those with substantial financial asset holdings might draw down those assets and repay part of their mortgage debt.16

The Poterba-Sinai study identified two factors that would affect the revenue estimate: the level of assets held by individuals with mortgage debt and the rate of return on those assets. The rate of return is relevant because if the assets are divested to pay down mortgage debt, the government loses the tax that would have been imposed on the income from the divested assets.

The Poterba-Sinai study found a remarkable overlap of non-housing assets and mortgage debt. It concluded that divestment of liquid financial assets could replace 16 percent of total mortgage debt, that divestment of all financial assets could replace 31.2 percent of all mortgage debt, and that divestment of all non-housing assets could replace 56 percent of mortgage debt. Not surprisingly, the ability to pay down mortgages rises with age (because of lower levels of mortgage debt) and income. The study strongly suggests that the additional revenues from repeal of the mortgage interest deduction would be paid by younger and less wealthy taxpayers. It concluded: "Under plausible assumptions between one sixth and one third [of the static revenue estimate] would be undone by changes in leverage."

The Poterba-Sinai study referenced another study, conducted by Martin Gervais and Manish Pandey, that concluded that 42 percent of the static revenue estimate would be lost through portfolio adjustments. That study assumed a single rate of return on the divested assets of 7.3 percent. Poterba and Sinai believed that the assumption overstated the reduction in revenue attributable to portfolio changes because they assumed that the divestment of assets would begin with liquid assets that have low rates of return. I agree that the Gervais-Pandey study overstates the effect of portfolio changes if divestment of assets were the only way to avoid the impact of disallowance of the mortgage interest deduction. However, portfolio changes are not the only potential response to the disallowance.

Debt, like money, is an extremely fungible commodity. Unfortunately, our tax laws assume that there are discrete categories of debt. Interest on consumer debt and on debt to acquire tax-exempt bonds is not deductible. Home mortgages and home equity loans result in tax-deductible interest subject to limitations on the size of the debt. Investment interest is deductible to the extent of investment income. Business-related interest is deductible largely without limit. Merely disallowing home mortgage interest without examining the overall deductibility of interest expense could permit individuals with large amounts of non-housing assets to avoid the effect of a mortgage interest disallowance by simply changing the way they borrow. It is worth noting that the disallowance of consumer interest was followed by expanded use of home equity loans.

The large overlap of mortgage debt and non-housing assets indicates that individuals see some leverage in the form of tax-deductible debt as advantageous. My assumption is that they would first attempt to maintain that leverage by shifting it from their home to business or investment assets (where the interest would continue to be deductible). If that is correct, individuals with large amounts of non-housing assets could essentially continue to receive the full benefit of the home mortgage interest deduction. In that case, the Gervais-Pandey study may be close to the mark.

c. Employer-provided health insurance. One of the largest behavioral responses to the repeal of a single tax expenditure would probably occur for repeal of the exclusions for employer-provided health insurance. Currently, 163 million of the 264 million non-elderly individuals in this country receive their health coverage through their employer. According to a recent study by five Treasury economists,17 the repeal of the income and payroll tax exclusions for employer-provided health insurance would result in 23 million individuals losing coverage from their employer (approximately 14 percent of those with that coverage). In this section, I will discuss the revenue consequences; later I will examine the policy ramifications.

Before the enactment of the healthcare reform legislation, this enormous behavioral response was estimated to have little or no revenue effect. Consistent with prevailing economic theory, the revenue estimators assumed that employers dropping coverage would increase cash wages by an amount equal to the value of the health insurance. Generally, it was assumed that the wage adjustments would be uniform for all employees. One of the unique features of the Treasury economists' study was their assumption that the wage replacement would differ "by wage and education group." The unstated inference was that groups with higher wages and greater education would be more successful in being compensated for the loss of coverage. This unique feature of their analysis would not affect the revenue estimate; because they assumed that the total compensation cost to the employer would remain constant, a better deal for one group of employees would result in a corresponding worse deal for other groups. Although not affecting revenues, such an assumption would have a dramatic impact on the question of who bears the burden of the additional taxes.

With the enactment of healthcare reform, the behavioral response will have a significant effect on the revenue estimate. That legislation provided substantial premiums and cost-sharing subsidies for health insurance purchased in the new health insurance markets. The cost of those subsidies in fiscal 2016 is $59 billion.18 The Treasury study did examine the impact of repealing the exclusions for employer-provided health insurance in the context of subsidies and a restructured insurance market similar to the system in the healthcare reform legislation. The economists estimated that the repeal would shift 23 million individuals from employer-provided coverage, with 16 million of those individuals covered by insurance purchased in the new insurance markets. Much of the revenue gained from repeal of the exclusions could be offset by increased subsidies to those shifted into the new insurance market. Unfortunately, the study did not reveal the increase in premium and cost-sharing subsidies that would result.

2. Separate expenditure estimates do not reflect interaction. Each "tax expenditure is estimated separately, under the assumption that all other tax expenditures remain in the tax code."19 In contrast, a JCT revenue table for a comprehensive reform package will reflect the interaction of the various provisions. The impact of this aspect of tax expenditure estimates is less clear than the impact of their static nature.

A recent article said that the revenue estimate for a repeal of all tax expenditures would be approximately 5 percent greater than the sum of all the tax expenditure estimates.20 Yet, that article indicated that the interactive effect depended dramatically on which tax expenditures were actually eliminated in the reform effort.

If the reform effort focused on itemized deductions, the revenue estimate with interactions would be 15 to 24 percent lower than the sum of the tax expenditure estimates. The lower amount is caused by individuals shifting from itemizing to the standard deduction. For example, repealing the mortgage interest deduction by itself would lead many individuals to shift to the standard deduction. That would mean that the repeal of the other itemized deductions would not further increase their tax liability.

However, repeal of tax expenditures that are exclusions would have positive interactions since more income would be subject to the higher marginal rates. Of course, most of those positive interactions could quickly disappear if the reform also includes marginal rate reductions.

It is extremely unlikely Congress would repeal major exclusions, like the exclusions for pensions and other retirement savings and the exclusion for interest on state and local bonds. The only major exclusion it might repeal is the exclusion for employer-provided health insurance, but even that is unlikely. Therefore, if we see a successful assault on individual preferences, it will probably focus on itemized deductions. That is the area in which the interactive effect is negative.

3. Assumption that law always followed normal tax principles. Some of the major tax expenditure estimates are based on the difference between what is claimed under existing law for the current year and what would be claimed for the current year if the normal tax rules had been in effect for all prior years. This assumption means that the tax expenditure estimate for deferral preferences can be dramatically smaller than the revenue estimate for a repeal of the preference.

For example, for accelerated depreciation the benefit of the acceleration is recaptured in later years through lower depreciation deductions. However, the tax expenditure estimate does not have the effect of the recapture. The tax expenditure estimate is the difference between the depreciation deductions claimed in the current year and the deductions that would have been claimed in the current year if the depreciation deductions had always been computed under a non-accelerated method. As a result, absent a large behavioral response, the revenue estimate for elimination of accelerated depreciation would be larger than the tax expenditure estimate. The revenue estimate would reflect the denial of benefits for future investments and the recapture of accelerated depreciation on old investments.

Another example is the last-in, first-out accounting rules.21 The recapture of prior LIFO benefits is a major part of a revenue estimate for repeal of LIFO. In the tax expenditure estimate for LIFO, that recapture element is not present, because the estimate assumes that the taxpayer used first-in, first-out in all prior years.

B. Transition

One of the most important aspects of any tax reform effort is transition. As explained in the 1984 Treasury studies that laid the groundwork for TRA 1986, transition relief is required "in order to avoid windfall gains and losses and economic dislocations." The studies also acknowledged that transition relief is required "to make tax reform a political reality."22

A successful tax reform effort will include transition relief. Indeed, for more robust reform, more extensive transition relief will be required. The tax expenditure estimates reflect immediate repeal with no transition. Certainly, those estimates overstate the revenue gains that you would see in actual legislation.

An obvious example is tax-exempt bonds. The combined individual and corporate tax expenditure estimates for tax-exempt bonds in 2008 was $40 billion. Much of that estimate would simply disappear unless Congress was willing to tax interest earned on previously issued tax-exempt bonds.

Even the simple setting of an effective date can have profound effects on the revenue estimates. For example, the Bowles-Simpson commission recommended converting the charitable deduction to a 12 percent tax credit with a 2 percent AGI floor. I am confident that if that proposal were in legislation signed into law this year, the effective date would be January 1, 2012. Congress would not make the new rules effective retroactively to the beginning of the year, and a midyear effective date could present administrative difficulties. With that effective date, there would be a large incentive to accelerate contributions into this year to receive a tax benefit now that would be unavailable in the future. Devices like private foundations or donor-advised funds would be especially attractive since they permit the donor to retain control of the ultimate disposition of the contributed amounts. In the past, the maxim, "The earlier you give, the greater the tax benefits," has governed tax planning for charitable giving. If Congress were seriously considering the Bowles-Simpson proposal, "now or never" would soon be the advice. The result could be a substantial acceleration of charitable contributions into the current tax year, with a corresponding reduction of contributions in the future and a revenue gain smaller than that implied in the tax expenditure estimate for the charitable deduction.23

C. Details Matter

The legislative process often requires compromise when faced with a difficult political decision. The compromise frequently involves a middle ground in which the revenue potential could be a small portion of the tax expenditure estimate.

The movement to a middle ground has already begun for the mortgage interest deduction. The Bowles-Simpson commission did not recommend outright repeal of the mortgage interest deduction. Instead, it would convert the deduction into a 12 percent nonrefundable credit, lower the current $1 million limit to $500,000, and restrict the benefit to only one residence. The benefits of the new credit would not be limited to itemizers.

Under the current rate schedule, the proposed 12 percent credit would be almost a two-thirds reduction in benefits for individuals in the top tax bracket, with smaller reductions for individuals in lower brackets. Homeowners in the 15 percent bracket would retain 80 percent of their current benefits.

The proposal is not all bad news. Individuals who have a mortgage and do not itemize would receive a tax benefit for mortgage interest. In the past, approximately 90 percent of all mortgage expense showed up on a return as an itemized deduction,24 which suggests that few homeowners would benefit from that aspect of the proposal. However, many current itemizers could receive additional benefits for mortgage interest.

TRA 1986 substantially increased the standard deduction and eliminated or significantly reduced several itemized deductions.25 As a result, there was a large increase in the amount of mortgage interest deductions that are "wasted" in the sense that they do not result in a tax benefit. This occurs when the individual's other itemized deductions are less than the standard deduction. A 1992 study indicated the scope of the wasted deductions. It concluded that as a result of TRA 1986, the average amount of wasted housing deductions (mortgage interest and real property tax deductions) increased from $1,042 in 1985 to $2,699 in 1989.26

Since the 1992 study, there has been a large increase in the standard deduction for married individuals as part of the marriage penalty relief in EGTRRA. Since most homeowners are married, the size of wasted mortgage interest deductions probably has grown significantly since that study.

An indication of the size of the wasted mortgage interest deductions is that the $80 billion tax expenditure estimate for 1988 is approximately 17 percent of the $481 billion of interest deductions claimed on returns for 1988. Quite likely, the average marginal tax rate of homeowners claiming the mortgage interest deduction is greater than 17 percent. One factor causing the low average benefit of 17 percent is the wasted deductions.

The tax expenditure estimate for the mortgage interest deduction is $80 billion annually. As discussed above, portfolio adjustments could reduce the potential revenue from repeal of the deduction by one-sixth to one-third of the static tax expenditure estimate. The substitution of a 12 percent credit for the deduction would reduce the remaining revenue by approximately two-thirds, given the relationship of the credit rate to the average benefit of current law. Extending benefits to current non-itemizers would further reduce the revenue gain. Even before considering transition relief, only a small fraction of the tax expenditure estimate would be actually collected in additional taxes under the Bowles-Simpson proposal.

In 2008 the mortgage interest deduction was the largest itemized deduction, accounting for 37.7 percent of all itemized deductions.27 For individuals with AGIs between $50,000 and $100,000, it was almost 44 percent of total itemized deductions. The substitution of a tax credit for the mortgage interest deduction might not raise much revenue, but it alone would result in a massive shift of individuals to the standard deduction. As a result, the potential revenue from repeal of other itemized deductions would be far less than the amount suggested from the tax expenditure estimates.

D. Limited to Income Tax

Tax expenditure estimates take into account only the income tax effect of the provision and ignore payroll tax effects. In contrast, revenue estimates account for both income and payroll tax effects. There are two tax expenditures with major employment tax consequences: the exclusion of employer-provided health insurance and the exclusion of pension plan contributions and income on pension assets. For healthcare, the employment tax effects are large, approximately an additional $100 billion per year.28 The amounts involved with the pension items also are significant, but it is unlikely those items would be substantially reduced even in an aggressive tax reform proposal. Many believe deficit reduction will require reductions in entitlement programs for the elderly -- hardly the environment in which one would predict substantial reductions in tax incentives for retirement savings.

IV. Distributional Impact

A. Overview

Many analyses of the distributional effect of tax expenditures focus on their absolute size. For example, the JCT annual tax expenditure pamphlets include tables that distribute the benefits of some major individual tax expenditures among income categories. For many tax expenditures, the tables show benefits that sharply increase with growth in income. Viewed from that perspective, many tax expenditures appear to be clearly regressive. Many are deductions or exclusions that are more valuable to upper-income individuals because of higher marginal rates. Rates of homeownership or access to employer-sponsored health insurance (ESI) increase as incomes grow, so upper-income individuals receive a larger share of the tax expenditures for those items. From that perspective, even the exclusion for ESI can be criticized as disproportionately benefiting upper-income individuals.

However, an article in a journal not often cited in tax policy discussions clearly states the argument against that approach:

    Progressivity in taxation (or tax subsidies) is not calculated on the basis of the absolute amount that a government takes from (or gives to) families at different income levels. Rather, a progressive tax takes a larger percentage of income from wealthier families than from poor ones -- the tax rate rises with income.29

Examining the distributional effect of tax expenditures by reference to their relationship to overall income gives a different picture than merely focusing on their absolute size. If an across-the-board equal-percentage cut in marginal rates were financed by an elimination of tax expenditures, "high-income tax units would be net winners and low-income units would be net losers."30 In other words, tax expenditures in the aggregate are progressive when viewed as a percentage of income. Some tax expenditures, like the preferential tax rates for capital gains and dividends, grow as a percentage of income as incomes increase and are clearly regressive in effect. The ability to save also increases with income, so the retirement savings provisions also have a regressive effect. Items like the exclusion for interest on state and local bonds and the charitable deduction also grow as a percentage of income, but as discussed below, the burden of repeal of those benefits will largely be borne by others. Benefits for owner-occupied housing grow as a percentage of income until the top 5 percent, at which point they begin a sharp decline.31 Most of the other tax expenditures are limited in size by law or are for items like healthcare costs, which decline as a percentage of income as incomes increase.

Before discussing how distributional analyses are made, I will make the obvious point that the distributional effect of a tax reform bill also is affected by how the revenues from the reforms are used. An example is the deduction for state and local income taxes. Viewed in isolation, repeal of the deduction would disproportionately affect upper-income taxpayers, largely because they receive most of the income that is subject to the state and local income taxes. However, if you use the revenue to finance a marginal rate reduction, the result could be neutral because you would be substituting an explicit rate reduction for the implicit rate reduction that results from the deduction for state and income taxes. The major effect would be an increase in the tax burden on residents of states with income taxes and a corresponding reduction for residents in states without income taxes, but there may be no change from a distributional basis.

The area of distributional analysis is particularly difficult for tax lawyers and others, like me, who do not have an extensive background in economics. We often look at distribution tables and reach conclusions not supported by the analysis or intended by those making the analysis. One reason distribution tables can cause misleading conclusions is that each of them tends to overstate tax increases imposed on upper-income taxpayers. It is quite possible to have a table showing that a proposal maintains current law progressivity, even though it would result in a large downward shift of actual tax liability. There is little question in my mind that the illustrative plan in the Bowles-Simpson report would result in a downward shift of tax liability even though the report's distribution table seems to indicate that the plan would maintain the progressivity of the current system.

B. Change as a Percentage of After-Tax Income

The federal income tax system is considered progressive because tax liability, as a percentage of pretax income, grows as incomes increase. Unfortunately, both Treasury and the Tax Policy Center (TPC) show changes in tax burden as a percentage of after-tax income. Perhaps this method could be justified if one were attempting to measure the change in economic position resulting from the legislation. But this method is misleading if you are attempting to determine whether the reform would maintain the progressivity of the current system. Ironically, it overstates the percentage changes at the upper incomes because of the progressivity of the current system.

The easiest way to explain the misleading nature of Treasury's and the TPC's method is through an oversimplified example. Assume that a bill imposes a new tax on gross income without any deductions. The rate of the tax starts at 1 percent at the lowest quintile and gradually decreases until it reaches 0.7 percent for the top 1 percent by income. By any standard, this would be considered a regressive tax. But a distribution table showing changes in tax burden as a percentage of after-tax income would show that tax as a flat 1 percent increase on all income categories. That occurs because the effective tax rates are close to zero for the bottom quintile, so their pretax and after-tax incomes are approximately equal. The effective tax rates for the top 1 percent are slightly more than 31 percent,32 so their after-tax income is approximately 70 percent of their pretax income. Therefore, a 0.7 percent tax for the top 1 percent on their pretax income translates into a 1 percent increase on their after-tax income.

If the overall tax reform legislation is revenue neutral, this method does not result in a misleading result. It overstates the impact of both increases and decreases at the upper income levels. However, if the reform would raise significant net revenue like the Bowles-Simpson illustrative plan would, the method could hide the regressive effect of the reform.

C. Revisiting the Concept of Static Estimates

One might expect that a distributional analysis would start with the estimate of the changes in revenue caused by the legislation and then distribute those changes among income categories. I believe that most members of Congress and most of their staffs think that is what a distributional analysis shows. However, that is not how the JCT did distributional analyses before 1994, nor how Treasury and others do them today.

In June 1993 the JCT staff published a pamphlet that is probably the best explanation of the predominant method for distributional analysis.33 The pamphlet explained that the JCT attempted to measure the economic burden of the tax, which "is not the same as the total amount of the tax revenue collected." The JCT used the example of an excise tax increase to show how the burden of a tax differs from the amount collected. The increased tax will cause some consumers to reduce purchases of the good on which the tax is imposed. "Even though such individuals pay no tax, they bear a burden by forgoing consumption of a good they would purchase in the absence of the tax increase."

The pamphlet describes the theoretically correct approach to determine tax burden but says that that approach would be too difficult to conduct. Therefore, the JCT decided to use the static revenue estimate, one that assumes no behavioral response, as a reasonable approximation of the change in economic burden. It did acknowledge that the method "overstates the loss of economic well-being due to a tax increase and understates the improvement in economic well-being due to a tax reduction." The overstatement of tax burden is especially large in circumstances in which large behavioral responses are expected, because that is where the largest differences occur between static estimates and expected revenue increases. Two major items for which there will be large behavioral responses are capital gains and the mortgage interest deduction. Upper-income individuals have both the largest share of capital gains and the greatest ability to defer realizations. Similarly, those individuals have the greatest ability to make portfolio adjustments to avoid the impact of a disallowance of mortgage interest deductions. Therefore, the overstatement of tax burden acknowledged by the JCT is most likely to occur at the upper incomes.

The treatment of capital gains is the starkest example of how that method of distributing tax burden can overstate the increase in tax liability at the top. According to the JCT, approximately 90 percent of the static revenue loss from the preferential capital gains rates is attributable to individuals making more than $200,000. For distributional purposes, those individuals could be treated as having a burden increase equal to 100 percent of the static revenue estimate, even though the JCT may assume that their actual tax liability will increase by only 24 percent of that amount.

Treasury uses roughly the same method for distributional purposes as described in a 1999 paper by its Office of Tax Analysis.34 It attempts to distribute the benefit or burden of the change in law. The paper states: "The burden, or benefit, of a tax proposal may not be properly measured by the change in tax payments; such is the case with a capital gains tax cut." Like the method outlined in the 1993 JCT pamphlet, Treasury, in analyzing the effect of changes in the capital gains rate, looks only to the current level of realizations without taking into account behavioral responses. The paper acknowledged that under the method, "Treasury would underestimate the amount of tax relief arising from a reduction in the capital gains rate." The paper left unsaid the obvious corollary that its method would overstate the tax increase from an increase in the capital gains rate.

The JCT changed its method for distribution analyses in 1994, and it now distributes the estimated changes in revenue.35 Others, including Treasury and the TPC, continue to use static revenue estimates to measure distributional changes.

The overstatement of tax expenditure estimates caused by the use of static estimates may have created unrealistic expectations but ultimately will have no effect on the shape of the legislation. However, the use of static estimates in making distributional analyses has affected the shape of past legislation and could do so again. Most successful tax reform efforts have had both a revenue target and a desired distributional effect. The use of nonstatic revenue estimates to meet the revenue target but static estimates to meet the distributional target can result in legislation that is determined to be neutral on a distributional basis even when there is a downward shift in actual tax collections.

D. Burden Overstatement

Distributional methods do not always assign the benefits or burdens of a change in law to the taxpayer on whom the legal incidence of the tax falls. The legal liability for most federal excise taxes is imposed on the manufacturer or importer, but the consumer is assumed to bear the burden of the tax. Likewise, the employee is assumed to bear the burden of the employer's share of employment taxes through reductions in wage income. However, all changes in the individual income tax are assigned to individual taxpayers, even when the burden of the change is clearly borne by others.

There are many individual tax expenditures for which the intended beneficiary is not the taxpayer who nominally receives the benefit. However, the entire amount of the tax increase from repeal of the benefit is assigned to the taxpayer when making the distributional analysis under both the current JCT and Treasury methods.

The clearest example of such a tax expenditure is the exemption for interest on state and local bonds. An accurate measure of the detriment to the investor on account of the repeal of the tax exemption would involve comparing his marginal tax rate with the difference between taxable and tax-exempt yields. However, for distributional purposes, all the revenue from a repeal of the tax exemption would be assigned to the investor even though state and local governments would bear most of the loss. Because investors in tax-exempt bonds are typically higher-income taxpayers, this is another instance in which distributional methods can make the bill seem more progressive than it actually is.

The charitable deduction is another example of an overstatement of burden at upper-income levels. Charitable deductions do rise with income, so a repeal of the charitable deduction or conversion of the deduction to a credit will show up as a tax increase disproportionately affecting the upper incomes. However, charitable contributions are purely voluntary. If upper-income individuals make even modest reductions in charitable giving on account of a disallowance of the charitable deduction, much of the burden of the disallowance would be borne by the charitable sector and its beneficiaries. Indeed, the total net worth of the upper-income taxpayers could go up if they retain assets rather than contribute them for charitable purposes.

E. Employee Benefits

Increasing taxes on employee benefits, like employer-provided health insurance, is regressive even if the payroll tax effect is not taken into account.36 It becomes sharply regressive if the change also applies for payroll tax purposes because of the wage base limitation that applies to Social Security taxes, which constitute the bulk of payroll taxes.

In distributing the effect of the increased taxes on employee benefits, it is assumed that all benefits dropped by the employer on account of the increased taxes will be 100 percent replaced for each affected worker by an increase in cash wages. It is a simplifying assumption, but it is also an assumption that could dramatically understate the burdens faced by workers on account of the tax increase.

VI. Collateral Consequences

Congress's consideration of tax reform will not be limited to a discussion of tax policy driven by the promise of economic efficiency and growth resulting from lower rates and fewer economic distortions through the elimination of tax expenditures. Politics will intrude -- appropriately in my opinion -- and require an examination of the proposals' real-world consequences. In that regard, Jonathan Talisman, former Treasury assistant secretary for tax policy, offered the following caution in recent testimony before the Senate Finance Committee:

    While an ideal tax system would not include many tax expenditures, we are not starting a tax system from scratch. Many of the largest "tax expenditures" are long-term features of our system embedded in the fabric of our society. . . . To avoid false expectations, we need to be careful in how we talk about base broadening, and consider the practical, economic and social effects of eliminating tax expenditures.37

Despite the rhetoric about the recent explosion of tax expenditures, any future tax reform focused on tax expenditures will by necessity focus on those long-term features of our system simply because only those items offer significant potential for revenue gains.

I will concentrate on two areas, healthcare and housing, to briefly discuss the possible consequences of reform. These two areas have the potential for the largest effects on our society, but that does not mean they are the only areas in which there will be collateral consequences. As a nation, we have a relatively small government sector and rely on private charity to perform services that are part of the government sector in other countries. What would be the effect of lower charitable giving as the result of repeal or curtailment of the charitable deduction? Repeal of the deduction for state and local taxes would increase the burden of those taxes, but all the revenue from the effective increase of those taxes would accrue to the federal government. What would be the effect of that change on the states' ability to address pressing budget problems? Those questions and others will be part of the debate.

A. Housing

Today we have 75 million units of existing owner-occupied housing. New home construction is at depressed levels: an annual rate of less than 500,000 units. However, long-term estimates suggest that between 1 million and 1.5 million new units are needed every year on account of population growth and replacement of old housing. There is little question in my mind that repeal of or limitations on current housing benefits could affect the level of new investment in owner-occupied housing by increasing its after-tax cost. The new housing could be smaller or on smaller lots, and there could be fewer new housing starts as the increased cost of ownership causes more individuals to rent.

That smaller investment in new housing may have a positive effect on the economy, arguably freeing up capital for more productive uses. But that potential gain carries a very large risk. Increasing the after-tax cost of owning a home cannot reduce the current over-investment in housing. It can only reduce its price. Given the fragile nature of our housing market and the possible consequences of further declines, extraordinarily generous transition relief would be in order, at the very least.

Unfortunately, it is hard to contemplate transition relief that could avoid the imposition of large windfall losses on existing homeowners in the form of value reductions. Even the reduction in benefits for future home purchases only could affect current market prices.

B. Employer-Provided Health Insurance

In 2008, 163 million individuals received health insurance through ESI. Only 17 million individuals had individually purchased coverage.38

The income and payroll tax exclusions for ESI are not the only advantages of that coverage:

    In addition to the exclusions, however, there are additional advantages to ESI that make it more attractive than the non-group market. For example, larger risk pools organized around employment rather than health status allow insurers to offer non-underwritten policies that reduce overhead costs. The large employment base allows for economies of scale and bargaining power that can reduce the costs of insurance. From the employee's point of view, the ease of enrolling through an employer and the use of payroll deductions to pay for an employee's share of the premium help encourage the employee to opt for coverage.39

Another recent study describes the role of ESI slightly differently:

    The goal of insurance providers is to create large pools of individuals with predictable distributions of risk. These pools can be created in many different ways, national health insurance in countries like Canada create one national pool. The United States has long relied on the employer as the main pooling device for insurance.40

For years, many healthcare experts have argued that the tax benefits for ESI have contributed to the rapid growth of medical spending by encouraging overconsumption of healthcare. Essentially, they have argued that the income and payroll exclusions have encouraged increases in plan "generosity," such as low co-pays or deductibles. It is that generosity that they contend leads to overconsumption of healthcare. Those arguments have led to proposals to repeal or cap the exclusions.

In any legislative consideration of those proposals, the major question will be their impact on the availability of ESI. The recent Treasury economists' study attempts to answer that question. It is the first study to analyze the issue with and without the recently enacted healthcare reforms. It concluded that repeal of the income and payroll tax exclusions for ESI would reduce ESI coverage by 14 percent, by approximately 23 million individuals. That conclusion was the same with or without healthcare reform. The study assumed that most of the individuals losing ESI would be able to purchase coverage in the individual market. Approximately 6 million individuals would become uninsured.

Although the overall conclusions were the same with or without healthcare reform, there were important differences as to the type of individuals who would leave the ESI system. The study concluded that those leaving the ESI system without reform would generally be young and healthy workers and that the premiums for the remaining workers would increase as a result. With the restructured insurance market and low-income subsidies in the healthcare reform law, the study concluded that the poorer and less healthy workers would leave the ESI system. If that is accurate, it means that much of the revenue raised from repeal of the exclusions could be offset by large increases in low-income subsidies.

A study of the actual effect of removing tax subsidies for ESIs in Canada suggests a more robust impact on coverage.41 Like our Medicare system, the public health insurance system in Canada has gaps in coverage. As a result, about 80 percent of non-elderly, non-poor Canadians have supplementary health insurance, largely through their employer. Employer contributions for the coverage were totally exempt from income and payroll taxes until 1993 when Quebec began to tax those contributions under its provincial income tax. Those contributions continued to be exempt from payroll taxes and federal income taxes. The change reduced the tax subsidy by almost 60 percent. The study found that there was a 20 percent reduction in employer-provided supplementary plans in Quebec, compared with other provinces. Only 10 to 15 percent of those losing the coverage were able to purchase similar coverage in the non-group market. The decrease in coverage was especially large for small employers.

I am not a healthcare expert, and a thorough discussion of the issue is beyond the scope of this report. But here are several observations.

First, the main argument for curtailing tax benefits for ESI is the promise of future cost reductions. Thus, it is ironic that the first effect of repealing those benefits would be an increase in overall costs as individuals shift from the very efficient ESI system to individual policies with far greater administrative costs. Essentially, we could be incurring immediate additional costs in return for possible cost reductions in the future.

Second, the Treasury economists' study hints at the consequences of breaking up the large pooled risks that are part of the ESI system. For example, the study concluded that without the reforms, young and healthy individuals would leave ESI. With reforms, the poor and less healthy would leave ESI. Under either circumstance, you could have a bifurcated market with much higher costs in one segment.

Finally, the Treasury economists' study is consistent with prior studies in one major respect. It assumes that the employer is totally indifferent as to whether to continue its healthcare plan. If the employer drops coverage, it is assumed the employer will increase cash wages dollar for dollar for the value of the coverage. Indeed, like prior studies, it assumes that employer decisions will reflect "the aggregation of worker preferences within the firm."42 The Treasury economists even assume that employers without current healthcare plans would add them in response to employee preferences.

It is that last assumption that I believe substantially understates the number of individuals who could lose coverage through repeal of the exclusions for ESI. Reducing cash wages is difficult, so many employers have focused their cost reduction efforts on fringe benefits like ESI. There is no reason to believe employers would not use repeal of the tax benefits for ESI as an excuse for further cost reductions. Also, repeal of the payroll tax exclusion for ESI would increase the employer's liability for payroll taxes. If the employer retains its healthcare plan, it could shift the additional cost to the employee only by reducing cash wages. It might be easier to drop the coverage.

VII. Conclusion

It is probably apparent from even a cursory reading of this report that I am a skeptic concerning the prospects for tax reform. In my view, current expectations far exceed what is possible from a technical or political perspective. But I acknowledge that it is too early in the process to know for certain. Tax reform efforts can create a momentum of their own. However, that can happen only if general discussions of reform lead to the development of serious and detailed proposals. We will have to wait and see if that occurs.

The interactions of tax expenditures will be one of the most difficult aspects of any reform effort. Elimination of a major itemized deduction, like the mortgage interest deduction or the deduction for state and local taxes, effectively repeals all other itemized deductions for most individuals as they shift to the standard deduction. It would be difficult to develop transition relief even for existing home mortgages if other itemized deductions were eliminated. For itemized deductions, the choice may be all in, or not at all.

Without a reexamination of distributional methods, tax reform could further increase income inequality. The current methods have serious flaws. Those flaws could make a plan appear fair from a distributional standpoint even if the proposal results in a sharp downward shift of actual tax liabilities. The JCT method, in my opinion, is less likely to create the potential for such a shift, but it is not without flaws. In areas like tax-exempt bonds and charitable deductions, the JCT method, like the others, assumes that upper-income individuals will bear the entire burden of the change even though most of the burden of the change will be borne by others.

Finally, it is clear to me that tax reform legislation along the lines proposed in the Bowles-Simpson commission report would be the most consequential tax legislation enacted since the New Deal. It would change long-standing rules affecting major segments of our economy, and it would have major consequences for our economy that cannot accurately be predicted.


(Appendix tables appear on the following pages.)
                 Table 2. Major Components of Individual Share*
                        (Numbers in billions of dollars)

 A. Owner-occupied housing

    Mortgage interest                                                     $80
    Exclusion of gain on sale                                             $30
    Real property tax deduction                                           $14
    Total                                                                $124

 B. Pension and other retirement savings

    Employer pension plans                                               $114
    Inside build-up on life insurance contracts                           $26.8
    IRA and self-employed retirement plans                                $26.5
    Total                                                                $167.3

 C. Health (only income tax cost, not employment tax cost)

    Employer-provided                                                    $116.5
    Exclusion of Medicare benefits (parts A, B, and D)                    $42.8
    Medical expense deduction                                              $9.5
    Exclusion of health benefits under worker's compensation               $8.1
    Self-employed health insurance deduction                               $4.4
    Total                                                                $181.3

 D. Family-related benefits

    Earned income tax credits                                             $46.5
    Per-child credit                                                      $44.8
    Tuition credits, scholarship exclusion, etc.                           $8
    Dependent care credit and exclusion                                    $2.6
    Total                                                                $101.9

 E. Charitable deduction                                                  $44.1

 F. State and local governments

    Deduction for income, sales, and personal property taxes              $30
    Exclusion of interest on G/O bonds                                    $21.5
    Private activity bonds                                                 $6.7
    Total                                                                 $58.2

 G. Capital gains and dividends

    Reduced rates $128
    No tax on unrealized appreciation for transfers at death or by gift   $59
    Total                                                                $187

 H. Exclusion of government transfer payments

    Social Security                                                       $23.1
    Public assistance                                                      $2.7
    Disability and survivor worker's compensation benefits                 $2.7
    Total                                                                 $28.5

 I. Employee benefits other than health or pension

    Cafeteria plans                                                       $33.6
    Miscellaneous fringe benefits                                          $7
    Group life, accident, and disability benefits                          $5.6
    Transportation (parking and transit)                                   $4.8
    Total                                                                 $51

 J. Exclusion of military and veterans benefits                           $10

 Total                                                                   $953.3

 *These 10 categories of tax expenditures account for $953 billion of the total
 $1.025 trillion of individual tax expenditures.

                 Table 3. Major Components of Corporate Share*
                        (Numbers in billions of dollars)

 A. International tax benefits

    Sales source                                                           $6.6
    Deferral                                                               $6.4
    Active financing                                                       $2.6
    Total                                                                 $15.6

 B. Research and development                                               $5

 C. Section 199 manufacturing deduction                                    $5.5

 D. Charitable                                                             $3.6

 E. Tax-exempt interest

    General purpose bonds                                                  $8.4
    Private activity bonds                                                 $1.9
    Total                                                                 $10.3

 F. Credit union exemption                                                 $1.7

 G. Accelerated depreciation for equipment                                 $6.2

 H. Lower corporate rates (below $5 million)                               $3.5

 I. Reserve deductions for insurance companies                             $5.5

 J. Low-income housing credit and NMTC                                     $5.0

 K. Like-kind exchanges                                                    $3.3

 Total                                                                    $65.2

 *It should be noted that these annual numbers may understate the amount of
 business tax expenditures because some of the provisions do not count the full
 year of the provision's effect (i.e., research credit). The list is not
 exhaustive but does illustrate some of the major corporate tax expenditures
 identified by the JCT. These items account for approximately two-thirds of the
 corporate tax expenditures.

1 President's Advisory Panel on Federal Tax Reform, "Simple, Fair, and Pro-Growth: Proposals to Fix America's Tax System" (Nov. 1, 2005), Doc 2005-22112, 2005 TNT 211-14 .

2 H.R. 3970, the Tax Reduction and Reform Act of 2007. The bill also included an individual reform package that would have repealed the alternative minimum tax.

3 Bruce Bartlett, "Misunderstanding Tax Expenditures and Tax Rates," Tax Notes, Nov. 22, 2010, p. 931, Doc 2010-24447 , 2010 TNT 224-9.

4 Congressional Budget and Impoundment Control Act of 1974 (P.L. 93-344), section 3(l)(3).

5 With one exception, the JCT has consistently followed the same method in determining what is a tax expenditure since its first tax expenditure pamphlet published in 1972. The JCT changed its method in 2008, only to revert to its prior practice the following year.

6 Martin J. McMahon Jr., "Taxing Tax Expenditures?" Tax Notes, Feb. 14, 2011, p. 775, Doc 2011-758 , 2011 TNT 31-7 .

7 Ways and Means Committee report accompanying H.R. 8363 (Sept. 13, 1963).

8 JCT, "Estimates of Federal Tax Expenditures for Fiscal Years 2007-2012," JCS-3-07 (Sept. 24, 2007), Doc 2007-21689 , 2007 TNT 186-12 .

9 Tax expenditures in 1988 totaled $321 billion, which was slightly more than 60 percent of total individual and corporate income tax receipts ($496 billion). JCT, "Estimates of Federal Tax Expenditures for Fiscal Years 1988-1992," JCS-3-87 (Feb. 27, 1987).

10 JCT, "Present Law and Historical Overview of the Tax System," JCX-1-11 (Jan. 18, 2011), Figure 5, Doc 2011-1139 , 2011 TNT 12-20 .

11 In 2008 the corporate share of tax expenditures was $86 billion, approximately 28 percent of corporate receipts ($304 billion). In 1988 the corporate share of tax expenditures was $62 billion, approximately two-thirds of corporate receipts ($94 billion).

12 JCT, supra note 8, at 20.

13 JCT, "Explanation of Methodology Used to Estimate Proposals Affecting the Taxation of Income From Capital Gains," JCS-12-90 (Mar. 27, 1990).

14 Jane G. Gravelle, "Capital Gains Tax Options: Behavioral Responses and Revenues," Congressional Research Service, R41364 (Aug. 10, 2010), Doc 2010-18157 , 2010 TNT 157-73.

15 Id.

16 James M. Poterba and Todd M. Sinai, "Income Tax Provisions Affecting Owner-Occupied Housing: Revenue Costs and Incentive Effects," National Bureau of Economic Research Working Paper 14253 (2008).

17 Robert Gillette et al., "The Impact of Repealing the Exclusion for Employer-Sponsored Insurance," 63 Nat'l Tax J. 695 (2010).

18 CBO, "Estimate for the Combination of the Patient Protection and Affordable Care Act and the Health Care and Education Reconciliation Act of 2010" (Mar. 20, 2010).

19 JCT, supra note 8, at 20.

20 Leonard Burman et al., "How Big Are Total Individual Income Tax Expenditures, and Who Benefits From Them?" Tax Policy Center (TPC) Discussion Paper No. 31 (Dec. 2008), Doc 2008-25536 , 2008 TNT 235-21 .

21 The JCT only recently included LIFO in its list of tax expenditures.

22 Treasury Department, 1 "Tax Reform for Fairness, Simplicity and Economic Growth" 20 (1984).

23 You also have to question the wisdom of a 2 percent floor. There are floors on some deductions under current law, such as the medical expense deduction and the casualty loss deduction. However, those deductions generally reflect expenses that are not voluntary in nature or in timing. The charitable deduction is purely voluntary. Again, devices like private foundations or donor-advised funds could be used to avoid some of the impact of the proposed change. Because those devices require substantial financial assets, the major impact of the 2 percent floor would occur in middle-income categories.

24 CBO, "Computing Effective Tax Rates on Capital Income" (Dec. 2006), n.38, Doc 2006-25321 , 2006 TNT 244-11 .

25 TRA 1986 repealed the deduction for consumer interest and state retail sales taxes, imposed a 2 percent floor on miscellaneous deductions, and increased the floor on the medical expense deduction.

26 James R. Follain et al., "Real Estate Markets Since 1980: What Role Have Tax Changes Played?" 45 Nat'l Tax J. 253 (1992).

27 Michael Parisi, "Individual Income Tax Returns, Preliminary Date, 2008," 29 SOI Bulletin 5, 8 (2010).

28 Jonathan Gruber, "The Tax Exclusion for Employer-Sponsored Health Insurance," NBER Working Paper No. 15766 (2010).

29 David Himmelstein and Steffie Woolhander, "The Regressivity of Taxing Employer-Paid Health Insurance," 361 New Eng. J. Med. e101 (2009).

30 Eric J. Toder et al., "Distributional Effects of Tax Expenditures," TPC (July 2009), Doc 2009-16452 , 2009 TNT 138-27 .

31 Id.

32 CBO, "Historical Effective Federal Tax Rates: 1979 to 2005" (Dec. 2007), Doc 2007-27147, 2007 TNT 239-14 .

33 JCT, "Methodology and Issues in Measuring Changes in the Distribution of Tax Burdens," JCS 93-7 (June 14, 1993).

34 Julie-Anne Cronin, "U.S. Treasury Distributional Analysis Methodology," Treasury OTA Paper No. 85 (Sept. 1999).

35 In 1994 I was JCT chief of staff and played a role in the change.

36 Toder et al., supra note 30.

37 Testimony of Jonathan Talisman Before the Senate Finance Committee Hearing, "How Did We Get Here? Changes in the Law and Tax Environment Since the Tax Reform Act of 1986" (Mar. 1, 2011), Doc 2011-4316 , 2011 TNT 41-44 .

38 Census Bureau, "Current Population Survey, Table H105" (2008).

39 Gillette et al., supra note 17.

40 Gruber, supra note 28.

41 Amy Finkelstein, "The Effect of Tax Subsidies to Employer-Provided Supplementary Health Insurance: Evidence From Canada," 84 J. Pub. Econ. 305 (2002).

42 Gruber, supra note 28.


About Tax Analysts

Tax Analysts is an influential provider of tax news and analysis for the global community. Over 150,000 tax professionals in law and accounting firms, corporations, and government agencies rely on Tax Analysts' federal, state, and international content daily. Key products include Tax Notes, Tax Notes Today, State Tax Notes, State Tax Today, Tax Notes International, and Worldwide Tax Daily. Founded in 1970 as a nonprofit organization, Tax Analysts has the industry's largest tax-dedicated correspondent staff, with more than 250 domestic and international correspondents. For more information, visit our home page.

For reprint permission or other information, contact