For other versions of this document, see http://wikileaks.org/wiki/CRS-RL33285 ------------------------------------------------------------------------------ Order Code RL33285 Tax Reform and Distributional Issues Updated September 7, 2007 Jane G. Gravelle Senior Specialist in Economic Policy Government and Finance Division Tax Reform and Distributional Issues Summary Tax reform proposals may change the tax base as well as the rate structure. Both affect the distribution of the tax burden across individuals and across income classes. While relief for lower-income individuals can be achieved by exemptions and credits, the choice of tax base has consequences for the distribution of the tax burden between middle- and upper-income taxpayers, even for cases where rates are graduated. The most significant tax reform in recent history, the Tax Reform Act of 1986, moved in the direction of expanding the individual income tax base, in part by reducing preferences for capital income. Many of these preferences have since been restored and, in some cases, expanded, moving further towards a wage base, and there are proposals for further expansion. Some current proposals, including congressional proposals, would move instead to a consumption base: H.R. 25, H.R. 1040, S. 1025, and S. 1040, and S. 1081. S. 1111, however, would move towards a broader income base. The President's Advisory Panel on Tax Reform presented two proposals, one moving toward a consumption tax base and the other reducing the coverage of capital income (and moving towards a wage base). The consequences of choice of tax base for distribution depend on the source of income in different income classes. According to tax data, over 80% of income in the middle class came from wage income, while only a third to a half of income in the very high income levels was derived from wages. Some forms of income do not fit clearly into the wage or capital income categories. When adjustments are made to allocate these forms of income, such as pension income and proprietorship income, the concentration of labor income increases in the middle classes. When all income is assigned to either capital income or labor income, over 90% of income earned in the middle classes derives from labor, while half or less of income in the highest brackets derives from labor. The more heavily that tax revision moves towards a wage base, which also occurs with consumption taxation in the long run, the more difficult it is to avoid shifting the tax from higher-income families to the middle class. This report will not be updated. Contents Sources of Income: Wages and Capital Income . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Implications for the Distributional Effects of Tax Reform Proposals . . . . . . . . . . 9 List of Tables Table 1. Major Sources of Income by Adjusted Gross Income Class, in Percentages, 2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Table 2. Major Sources of Income by Income, Ordered by Population Share, in Percentages, 2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Table 3. Self-Employment, Rent, and other Income by Income Class, as a Percentage of Total Income, 2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Table 4. Labor Share of Individual Income, by Adjusted Gross Income Class . . 7 Table 5. Interest, Dividends, and Capital Gains by Adjusted Gross Income Class, as a Percentage of Total Income, 2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Table 6. Labor Share of Individual Income, by Adjusted Gross Income Class, Including Corporate Source Income, 2003 . . . . . . . . . . . . . . . . . . . . . 9 Tax Reform and Distributional Issues Tax reform proposals may change the tax base as well as the rate structure. Both affect the distribution of the tax burden across individuals and across income classes. While relief for lower-income individuals can be achieved by exemptions and credits, the choice of tax base has consequences for the distribution of the burden between middle-and upper-income taxpayers, even for cases where rates are graduated. The most significant tax reform in recent history, the Tax Reform Act of 1986, moved in the direction of expanding the individual income tax base, in part by reducing preferences for capital income. Many of these preferences have since been restored and, in some cases, expanded. Among other individual provisions, the 1986 law taxed capital gains at ordinary rates, restricted fully tax-favored individual retirement accounts to those not covered by employer pension plans, and provided significant restrictions on tax shelter operations. In 1997 capital gains rates were reduced and the general availability of fully tax-favored IRAs restored, but with income limits. Dollar limits on contributions to IRAs were increased in 2001 and rates were lowered further on capital gains and on dividends in 2003. The 2001 and 2003 changes, absent new legislation, expire after 2010 and 2008 respectively. There are proposals to maintain or extend the recently enacted tax cuts, including the benefit for IRAs and dividends and capital gains. The President has proposed greatly expanding preferences for individual savings accounts as well. Some current proposals, including congressional proposals, propose a fundamental change by moving to a consumption base.1 These consumption base proposals include S. 1040 (Shelby) and S. 1081 (Specter), which propose a flat tax (a tax on wages at the individual level and cash flow of businesses). H.R. 25 (Linder) and S.1025 (Chambliss) propose a national retail sales tax. H.R. 1040 (Burgess) would allow the option of a flat tax. The President's Advisory Panel on Federal Tax Reform presented two proposals, one of which was a proposal similar to a flat tax in its base, but allowing a graduated rate on wages (a type referred to as an x tax).2 The plan also included an additional tax on financial income. The plan, however, allows greatly expanded tax exempt savings plan limits without income limits and with much higher 1 See CRS Report RL33443, Flat Tax Proposals and Fundamental Tax Reform, by James Bickley, for an overview of proposals. 2 The President's Advisory Panel on Federal Tax Reform, Simple, Fair, and Pro-Growth: Proposals to Fix America's Tax System, November 2005. CRS-2 contribution limits, which would exclude much income from this tax. The panel discussed, but did not recommend, a value-added tax (VAT). There are also proposals that would retain an income base. S. 1111 (Wyden) would broaden the income base, in part by reducing preferences for capital income. The President's advisory panel also has an income base reform plan, which they were instructed to provide as at least one option. The proposal, however, reduces the coverage of capital income by exempting dividends and capital gains, as well as expanding tax-exempt savings plans. The advisory panel proposal, therefore, moves towards a wage tax base, whereas S. 1111 moves towards a broader income tax base. The consumption tax is equivalent to a tax on wages and a tax on old capital. The flat tax approach collects the tax on old capital at the business level and the tax on wages directly from individuals, while the retail sales tax collects all taxes at the retail stage. (A VAT collects the tax at each stage of production.) Thus while current owners of capital would pay the tax on consumption out of existing savings, in the long run the consumption tax is essentially a tax on wages. The economic standards for evaluating tax reform proposals usually include effects on economic efficiency, administrative simplicity, and distribution. The first and second are often stressed with respect to proposals for consumption or wage tax bases, although the efficiency and simplicity benefits are not without dispute.3 Distributional issues are less easy to evaluate because there are subjective judgements about what type of system is fair. However, an issue that can be addressed is the consequences of tax base choice for the distribution of the tax burden. This report examines the shares of income attributable to capital and labor income across the income brackets and discusses the implications of choice of tax base on the distribution of alternative tax systems. Sources of Income: Wages and Capital Income Income falls into two broad types, capital income and labor income. Table 1 uses data from the Internal Revenue Service Statistics of Income to examine the sources of income by income brackets.4 This table reports the four major sources of income. The first three are wages, financial capital income (interest, dividends and capital gains), and pensions and distributions from individual retirement accounts (IRAs). The remainder of the income ("other") is derived largely from unincorporated business activities, including 3 See CRS Report RL32603, The Flat Tax, Value-Added Tax and National Retail Sales Tax: An Overview of the Issues, by Jane G. Gravelle and Gregg A. Esenwein. For discussions of the effects on saving see CRS Report RS22367, Federal Tax Reform and Its Potential Effects on Savings, by Gregg Esenwein, along with the discussion of savings effects from reducing capital income taxes (and the consequences for distribution) in CRS Report RL32517, Distributional Effects of Taxes on Corporate Profits, Investment Income, and Estates, by Jane G. Gravelle. 4 Found at [http://www.irs.ustreas.gov/taxstats/indtaxstats/article/0,,id=96981,00.html]. CRS-3 proprietorships, partnerships, Subchapter S (small corporations taxed as partnerships), rents, and royalties, estates and trusts, and minor items such as alimony and state tax refunds. Table 1. Major Sources of Income by Adjusted Gross Income Class, in Percentages, 2003 Adjusted Gross Interest, Wage Pensions and Income Class Dividends, Other Income IRAs ($ thousands) Capital Gains Under $15 75% 5% 11% 8% $15-$30 80 3 11 5 $30-$50 82 3 9 6 $50-$75 81 3 9 6 $75-$100 82 4 8 4 $100-$200 78 6 7 8 $200-$500 64 13 4 18 $500-$1000 52 20 2 23 Above $1000 32 39 1 24 Overall 75 9 7 9 Source: CRS calculations based on the Internal Revenue Service 2003 Individual Statistics of Income. Excludes returns with negative adjusted gross income. Note: Rows may not add to 100% because of rounding. Table 2 arrays the same information by population share. Both tables show that wage income accounts for the bulk of income in the middle income classes, but that share declines substantially for higher income classes. In the middle three fifths of the population, and for those earning under $100,000, wage income accounts for close to 80%, or more, of total adjusted gross income. For those in the top 1% of the population, wage income is less than half of income, and for those with income of $1 million or more, wage income is only a third of income. The importance of interest, dividends, and capital gains is reversed. For those with incomes under $100,000 and for the middle three fifths of the distribution, this income accounts for less than 5% of total income. For those with income over $1 million, interest, dividends, and capital gains accounts for almost 40% of the total and for the top 1%, almost 30%. CRS-4 Table 2. Major Sources of Income by Income, Ordered by Population Share, in Percentages, 2003 Financial Population Pensions and Wage Income Capital Other Share IRAs Income Bottom 20% 78% 6% 8% 8% Second 20% 75 4 13 7 Third 20% 82 3 9 6 Fourth 20% 82 3 9 6 Top 20% 69 12 5 14 Top 10% 64 15 4 19 Top 5% 57 20 3 20 Top 2% 50 25 2 21 Top 1% 44 29 1 26 Source: CRS calculations based on the Internal Revenue Service 2003 Individual Statistics of Income. Excludes returns with negative adjusted gross income. Note: Rows may not add to 100% because of rounding. The relative concentration of wage income as a source of income in the middle brackets and reliance on capital income at the higher levels is increased if we consider pensions and IRAs to be forms of deferred wage income. The tax treatment of these items allows the deduction (or exclusion) of contributions, with tax due on benefits. This treatment effectively exempts income from capital from tax and thus is a capital income preference, not a wage income preference.5 We would obtain the same economic result if contributions were not deductible but earnings were not taxed (as is the case for Roth IRAs). If that method of providing a benefit were used, wage income would be larger because contributions would not be excluded from income (and pensions would not be included). Depending on the growth rate versus the rate of return of these provisions, the pension and IRA income may overstate or understate the value of wage income. In a steady state, with the real rate of return 5% and the growth rate 3%, typical of the past economy, and assuming an average lag of 15 years between contribution and benefit, wage inclusions would be about 75% of benefits.6 Assigning most of the pension and IRA benefits to wage income would move the share of wage income to 5 For a mathematical demonstration, see CRS Report RL30255, Individual Retirement Accounts: Issues and Proposed Expansion, by Thomas L. Hungerford and Jane G. Gravelle. 6 The formula for constant growth is e(g-r)t, where e is a natural constant, approximately 2.718, g is the growth rate, r is the real rate of return (i.e., nominal return minus inflation rate), and t is the time period. CRS-5 close to 90% for middle income groups while having little effect on higher-income individuals. Most of the remainder of the income is from self-employment, and some income that is largely capital income (rents, royalties, and estate and trust income). Self- employment income is a combination of the return to labor and the return to capital, and, particularly in the lower and middle classes, much of it is labor income. Unincorporated businesses are divided into proprietorship (or earnings from a business or profession) and the combination of partnership and Subchapter S income. Table 3 shows the distribution of this unincorporated business activity by income class. (Note that the sums do not strictly match because of loss deduction restrictions.) Proprietorship income dominates at the lower income levels, while partnership and Subchapter S income is more important at higher levels. A typical rule of thumb in the economy is that capital income accounts for around 25% of total income. But this share can vary across types of businesses, and in particular, proprietorships tend to be smaller and the owners' labor income likely to be larger relative to capital income. Allocating labor income for unincorporated businesses is hampered by lack of data on capital assets for proprietorships. Accordingly, we used data on depreciation deductions, based on the assumption that the real rate of return is about 5%. Then using estimates of the distribution of asset types and the average depreciation rates of equipment and structures,7 we estimated the depreciation rate; this rate actually turned out to be almost identical. This suggests that depreciation deductions as a percentage of net income from trade or business is a good proxy for the share of capital income. These data, taken respectively from the Statistics of Income data for proprietorship returns (2003), partnership returns (2003), and Subchapter S returns (2002), indicate that capital income is about 18% for proprietors and 53% for partnerships and Subchapter S firms. 7 The distribution of assets was taken from Jane G. Gravelle, The Economic Effects of Taxing Capital Income, Cambridge, MIT Press, 1994, p. 300. CRS-6 Table 3. Self-Employment, Rent, and other Income by Income Class, as a Percentage of Total Income, 2003 Partnership Rents, Adjusted Gross Proprietor- and Royalties, Income Class Total ship Subchapter Estate and ($ thousands) S* Income Trust Under $15 10.7% -0.3% 1.1% 11.5% $15-$30 4.1 0.1 0.7 4.9 $30-$50 2.8 0.4 0.6 3.8 $50-$75 2.6 0.8 0.7 4.1 $75-$100 2.9 0.9 0.7 4.6 $100-$200 4.3 3.1 1.2 8.6 $200-$500 6.4 11.4 2.2 20.0 $500-$1000 4.3 19.4 2.9 26.6 Above $1000 1.5 22.6 2.4 26.7 Overall 4.7 4.1 0.2 8.0 Source: CRS calculations based on the Internal Revenue Service 2003 Individual Statistics of Income. Excludes returns with negative adjusted gross income. * The allocation rule for assigning income to labor and capital is based on 2002 data, but the aggregate amounts of Subchapter S income are from 2003. Using these data, we then summarize the labor income share in Table 4, which involves reducing the denominator by 25% of pension and IRA income, and counting as labor income wages, 75% of pensions and IRA distributions, 82% of proprietorship income, and 47% of income of partnerships and sole proprietorships. We also report in that table the overall share of returns that fall into each income class. CRS-7 Table 4. Labor Share of Individual Income, by Adjusted Gross Income Class Adjusted Gross Income Share of Tax Returns Labor Share of Income Class ($ thousands) (%) (%) Under $15 28.1% 94% $15-$30 23.0 94 $30-$50 19.0 94 $50-$75 13.5 93 $75-$100 7.4 93 $100-$200 6.9 90 $200-$500 1.6 78 $500-$1000 0.3 66 Above $1000 0.1 45 Overall 100.0 87 Source: CRS calculations, see text. As this table indicates, for individuals earning under $100,000, the labor share of income, in general, is in excess of 90%. For very high income levels the share drops, and is only 45% for those earning more than a million. Overall, the labor share of income is 87%. These calculations do not take account of the fact that some income is subject to tax at the corporate level, which means that labor income is smaller than that reported in Table 4. There are two types of income reported on tax returns that reflect corporate income, dividends and capital gains.8 As shown in Table 5, which reports the distribution of financial income by type of income, capital gains are much more concentrated in higher income levels than are dividends. For example, although capital gains in the aggregate is 3.6 times dividends, in the $1 million income class, gains is 7.6 times dividends. That greater concentration is likely due to two effects, the first being that some capital gains are earned on assets other than corporate stock, and the second that higher-income individuals are more likely to hold riskier stocks and stocks that tend to pay out less income in dividends. 8 Pension and IRA distribution also reflect corporate source income, but do not separate interest income from dividends and capital gains, or easily permit a separation of principal from return. CRS-8 Table 5. Interest, Dividends, and Capital Gains by Adjusted Gross Income Class, as a Percentage of Total Income, 2003 Adjusted Gross Total Income Class Interest Dividends Capital Gains Financial ($ thousands) Income Under $15 4.7% 0.9% -0.5% 5.2% $15-$30 2.6 0.6 0.2 3.4 $30-$50 1.8 0.5 0.3 2.6 $50-$75 1.8 0.6 0.6 3.0 $75-$100 1.7 0.8 1.1 3.5 $100-$200 2.0 1.2 2.7 5.9 $200-$500 2.9 2.1 8.2 13.2 $500-$1000 3.5 3.0 13.4 19.9 Above $1000 5.0 4.0 30.3 39.4 Overall 2.5 1.3 4.7 8.6 Source: CRS calculations based on the Internal Revenue Service 2003 Individual Statistics of Income. Excludes returns with negative adjusted gross income. Table 6 reports labor income shares by income class using the dividend distribution and the capital gains distribution to allocate corporate source income, with the true share likely somewhere in between. Accounting for corporate source income decreases the share of labor income, but does so notably for the higher income classes, and this is particularly the case when the capital gains allocation is used. CRS-9 Table 6. Labor Share of Individual Income, by Adjusted Gross Income Class, Including Corporate Source Income, 2003 Labor Share Labor Share Adjusted Gross Income of Income, of Income, Share of Returns Class ($ thousands) Dividend Capital Gains Allocation Allocation Under $15 28.1% 90% 94% $15-$30 23.0 92 94 $30-$50 19.0 92 94 $50-$75 13.5 91 92 $75-$100 7.4 90 92 $100-$200 6.9 85 87 $200-$500 1.6 72 71 $500-$1000 0.3 59 58 Above $1000 0.1 39 34 Overall 100.0 83 87 Source: CRS calculations, see text. The calculation accounting for corporate income indicates that the middle class receives over 90% of their income from labor income. Overall, labor income accounts for less than 85%. However, for those with incomes over $1 million, labor income is only about a third of income, and for the $500,000 to $1 million class, about 60%. Implications for the Distributional Effects of Tax Reform Proposals There are important implications in the choice of tax base in tax reform proposals for the distribution of the tax burden across income classes. For example, under a tax base that is confined to wage income, only a significant increase in the progressivity of the rate structure would prevent the burden in a revenue neutral change from being shifted away from high-income individuals. While low-income individuals can be protected through tax relief provisions, absent a change in the rate structure, the middle class would pay higher taxes. Keeping roughly the same relative progressivity of the current rate structure, moving towards a wage base would shift the burden from high- to middle-income taxpayers, whereas moving towards an income base (that includes a more comprehensive measure of capital income) would shift the burden from middle- to high-income taxpayers. CRS-10 Proposals to reduce taxes on capital income through, for example, reducing or eliminating taxes on dividends, capital gains, and interest income (as in the case in one of the tax options by the President's advisory commission) would likely shift the burden, other things equal, away from high-income individuals to the middle class. The commission report shows a distributionally neutral system,9 but it is likely that is a temporary artifact of the shift into back-loaded savings accounts (which raise revenue on owners of assets in the short run but lower it dramatically in the long run).10 Proposals such as S. 1111, which expand the base towards an income base, would shift the burden away from the middle class and towards high-income individuals. As noted earlier, a consumption tax base is, in the long run, roughly a tax on wages. (For those individuals who leave significant bequests, some part of wage income is also exempted.) Consumption taxes are more likely to shift the burden in the long run away from high-income individuals, not only because high-income individuals over their lifetime tend to consume less of their earnings but also because some forms of consumption taxes necessarily have a flat rate (e.g., national sales tax and the flat tax). The President's advisory panel had a partially graduated consumption tax (x tax) with a supplementary lower-rate tax on passive financial income, and reported distributional neutrality for that proposal as well.11 That outcome, however, is due to two reasons. The first is the effect of savings accounts that also contributed to higher rates on owners of capital in the short run, while allowing increased sheltering of passive financial income in the long run. A second reason is the method of distributing a tax. A consumption tax is a tax on wage income and a lump sum tax on old capital that is effectively collected over time as the assets are consumed. For very-high-income individuals who indefinitely pass on assets in estates, that consumption may never occur. If one distributed the tax on the basis of consumption, the tax would decline as income rises despite the rate structure. The tax was, however, distributed as if it were an income tax and thus the cash flow tax at the firm level (which is really a lump sum tax on old capital that may or may not be translated into an effective tax on consumption) is treated as if it is a tax on income. To illustrate the importance of these approaches, consider a recent study that compared the distributional effects of an x tax with a 15% and 30% rate and a demogrant (rebate to lower-income individuals to offset the tax) under both 9 The President's Advisory Panel on Federal Tax Reform, Simple, Fair, and Pro-Growth: Proposals to Fix America's Tax System, November 2005, p. 136. This discrepancy also means that the proposed tax reform is not fully revenue neutral in the long run. 10 See CRS Report RL32228, Proposed Savings Accounts: Economic and Budgetary Effects, by Jane G. Gravelle and Maxim Shvedov for an explanation of this budget effect. 11 The President's Advisory Panel on Federal Tax Reform, Simple, Fair, and Pro-Growth: Proposals to Fix America's Tax System, op cit., p. 175. CRS-11 approaches.12 This plan is similar in many respects to the panel's proposal. If distributed according to consumption, the middle quintile has a tax rate of 23.3%, the top quintile a tax rate of 12.1% and the top 1% a tax rate of 6.1%. If distributed according to income, the tax rate is 11.4% for the middle quintile, 22.5% for the top quintile and 22.0% for the top 1%. Distributing a consumption-based tax in the short run is tricky, and there is no perfect answer because the cash flow tax is a tax that causes asset values (or their purchasing power) to fall but does not burden new saving which can be purchased at a discount. However, in the long run the consumption tax base tends to be similar to a wage tax base, except that it also favors higher-income people, even in the long run, because they are less likely to consume all of their lifetime income 12 See Leonard Burman, Jane Gravelle, and Jeff Rohaly, "Towards a More Consistent Distributional Analysis," Proceedings of the 2005 Conference, National Tax Association, Washington. D.C. 2006. ------------------------------------------------------------------------------ For other versions of this document, see http://wikileaks.org/wiki/CRS-RL33285