During a recent visit to the office of one of the Senate's more thoughtful and influential Democrats, I was told by the senator's legislative director: "The Democrats' first priority in this Congress is a middle-class tax cut. Now let's move on to the second priority: if s a middle-class tax cut. Then there's the third priority: a middle-class tax cut."
For a decade, Democrats have been badgered as the tax-and-spend party that ached to put its hand in your pocket and take some of your hard-earned cash. Many Democrats yearn for the times when they appeared as champions of the average family, as they did during the fall 1990 budget clash when they fought off clumsy White House efforts to shift tax increases from the rich to the middle class.
But when the Democrats finally bring up a "middle-class tax relief bill," what form will it take -- and whom will it actually benefit? Many Democrats cite compelling evidence showing widening income disparities between the wealthy and the rest of the nation. Yet they could end up pushing a bill that widens income disparities further. They could also support a legislative package that ignores the working poor and other lower-income families, denying them the benefits distributed to those in middle- and upper-middle-income brackets. Or Democrats could fashion an exemplary package that helps the middle class while also providing significant assistance to poor children and their families.
The question facing the Democrats is partly a matter of just what the "middle class" means to them. Some legislators believe the hard-pressed middle class in need of tax relief extends to those earning $100,000 a year or more. Mention the fact that median income stood at $34,213 in 1989 (the latest year for which data are available), or that households with incomes above $91,750 were among the top 5 percent of all U.S. households that year, and you may be met with disbelief. One liberal House member told me recently it was critical to provide a tax cut for families at the $90,000 level, but there was no reason to aid those with incomes below $20,000. Thus the rhetorical consensus of the Democrats on aiding the middle class conceals some differences in political understanding and intentions.
The Basic Alternatives
Three major approaches to tax relief are now on the table. One approach would liberalize Individual Retirement Accounts (IRAs), primarily by restoring tax deductibility for IRA contributions to those who lost deducibility in the Tax Reform Act of 1986. A second would boost the personal exemption, now $2,150, to higher levels. The third approach would convert the personal exemption for children into a refundable tax credit. Earlier this year, what might have been considered a fourth alternative -- the proposal to cut Social Security tax rates, introduced by Senator Daniel Patrick Moynihan -- was turned down.
Choices in tax policy often involve complex, overlapping, and sometimes competing concerns: distributing the burdens of government; shaping the effects that taxes have on incentives to work, save, invest, and marry; and rewarding some activities while discouraging others. In the emerging debate over competing measures to aid the middle class, the primary issue is who should pay what, and who should receive a break, particularly in the aftermath of the 1980s. The debate turns on other questions as well, such as how best to strengthen families raising children. But while these can be difficult issues, one fact stands out that should both clarify and simplify the debate: Two of the three principal approaches the Democrats are considering do much less for the families in the middle of the income spectrum than for those at higher income levels.
Liberalizing IRAs
The IRA approach has a legion of supporters on Capitol Hill. IRA legislation introduced earlier this year by Senate Finance Committee Chair Lloyd Bentsen and Republican Senator William Roth has 77 Senate co-sponsors. A companion House bill has more than 250 co-sponsors. Some of its proponents describe this legislation as "restoring IRAs to the middle class."
Far from aiding middle-class families, however, the bill ignores most of them. It confers the bulk of its new tax breaks on individuals in the upper parts of the income scale, thereby widening income disparities further. And it increases the deficit through a sleight-of-hand budgeting gimmick.
The legislation would repeal the IRA provisions of the Tax Reform Act of 1986. That act slashed tax rates for those in the upper parts of the income spectrum and, in turn, closed off various tax deductions, credits, and write-offs for these taxpayers, including tax-deductible contributions to an IRA. Until 1986, any taxpayer could deposit up to $2,000 in an IRA and take a deduction for the full amount deposited. In addition, the interest earned on these accounts was tax-free until the funds were withdrawn after retirement. But under tax reform, couples with adjusted gross incomes over $50,000 became ineligible for IRA deductions, although they can still deposit funds in an IRA and avoid tax on the interest until retirement. Couples with incomes in the $40,000 to $50,000 range are eligible for a deduction of less than $2,000.
The Bentsen-Roth legislation would drop the $50,000 limit and restore universal eligibility for IRA tax deductions. It liberalizes IRAs in other ways as well.
Yet few of the benefits from the legislation would go to those in the middle of the income spectrum. A Joint Committee on Taxation analysis of a 1990 proposal to restore IRA deductibility showed that the top 20 percent of all taxpayers would receive about 95 percent of the tax benefits. This is similar to the percentage of tax benefits that would go to the top 20 percent under a capital gains cut. The bottom four-fifths of taxpayers would receive the remaining 5 percent of the tax benefits.
There are several reasons the tax benefits are so skewed. First, most middle and lower income taxpayers are already eligible for IRA tax deductions. A Joint Tax Committee analysis found that 82 percent of all taxpayers with earnings are eligible for IRA tax deductions in 1991. Some 73 percent are eligible for the maximum deduction of $2,000. If these figures seem surprising, remember that median U.S. family income is only about $35,000, well below the IRA cut-off of $50,000. (Moreover, adjusted gross income as measured for tax purposes excludes some types of income, which thus are not counted against the $50,000 limit.)
Second, middle-income taxpayers do not make much use of IRAs, presumably because they do not have $2,000 a year to spare. In 1986, the last year IRA deductions were available to all taxpayers, only 13 percent of tax filers in the middle third of the income scale made a contribution. By contrast, 66 percent of the filers in the top 4 percent of the income scale did.
Finally, the higher an individual's tax bracket, the greater the tax benefit of an IRA deduction. The deduction is worth twice as much to a taxpayer in the 31 percent bracket, which consists mainly of taxpayers with incomes over $100,000, as it is to a tax filer in the 15 percent bracket, which primarily includes households with incomes below $50,000.
The IRA legislation includes a second major provision, which magnifies its defects. In addition to restoring universal deductibility for IRA contributions, the bill creates a new type of IRA, known as a "back-loaded IRA." Under a back-loaded IRA, a taxpayer depositing up to $2,000 a year would not receive an up-front tax deduction, but all interest earned by the account would be tax-free in perpetuity, so long as the funds were not withdrawn for five years. Both the interest and the principal could later be withdrawn free of any tax whatsoever. Under the Bentsen-Roth bill, a taxpayer could use either a conventional IRA, a back-loaded IRA, or a combination of both.
The back-loaded IRA provision represents public policy at its most cynical. Initially, the proposal causes relatively little revenue loss for the government, because no up-front tax deductions are taken when deposits are made. But because the interest in these accounts is permanently tax-free, the revenue loss grows larger with each passing year. As additional funds are moved into these IRA accounts, the interest on a steadily increasing share of national savings is sheltered from taxation. Analyzing a similar proposal earlier this year, the Joint Committee on Taxation estimated it would cost $355 million in revenue in the first year but $1.8 billion by the fifth year. The Congressional Budget Office noted the proposal eventually could cost as much as $8 billion a year.
The advantage to lawmakers is that federal budget rules adopted last fall require the revenue loss from any tax cut to be offset -- through corresponding tax in- creases or reductions in entitlement programs -- for only the next five years. There is no requirement for offsetting any increased revenue loss beyond that period. Thus, a back-loaded IRA proposal "paid for" by an offsetting tax increase for the next five years will produce a rising deficit after the fifth year, as the revenue loss caused by the proposal swells. Tax experts such as Henry Aaron of the Brookings Institution have denounced back-loaded IRAs, therefore, as devices whose long-term revenue-losing effects are concealed under the current budget rules.
This feature of back-loaded IRAs increases their attraction to some members of Congress and the White House. The Bush administration has included a back-loaded ERA proposal in both budgets it has submitted. But Democrats should shun this approach. It will only boost chances that large deficits will continue to limit their ability to fashion a coherent domestic agenda into the next century.
The likelihood that the IRA bill will generate long-term revenue losses also undercuts the claim that the bill would increase savings for productive investment. The savings available for growth-producing investments are those savings left over after government deficits are financed. IRA proposals, cannot increase savings available for investment unless they stimulate an increase in personal saving that exceeds the government's long-term revenue loss. The evidence to support a belief that a substantial increase in savings would occur is weak, however. Henry Aaron has pointed out that "despite the alleged stimulation of savings from IRAs, personal saving plummeted in the 1980s after IRAs were liberalized and rose following the Tax Reform Act of 1986, which curtailed IRAs. Many other factors influenced savings in addition to these particular tax changes. But those who claim that IRAs boost saving have relied on evidence that is subject, in my view, to devastating criticism." Aaron says that IRA liberalization proposals should be seen as "primarily a give-away to people with enough assets to shift them into sheltered accounts."
The IRA proposals contain another notable feature. They would allow people with IRAs to make early withdrawals without penalty if the funds were used for a first-time home purchase, education expenses, or medical costs exceeding 7.5 percent of income. Some IRA proponents argue these provisions would help the middle class. But such provisions need not depend upon restoring ERA deducibility for those in higher-income brackets. Congress could allow penalty-free IRA withdrawals for the middle- and lower-income people now eligible to use IRAs.
Increasing the Personal Exemption
A second entry in the middle-class tax-cut competition is an increase in the personal exemption, which taxpayers claim for themselves and their dependents. In 1948 the personal exemption stood at $600; now it is $2,150. Had it kept pace with inflation since 1948, it would be approximately $3,400 today. (Some even claim that to match its 1948 value, the personal exemption should now exceed $8,000, but that figure is correct only if the exemption is adjusted for real per capita income growth as wellas for inflation.) Proposals to raise the personal exemption abound on Capitol Hill. The leading proposal, which would raise the exemption to $3,500 in 1992, was introduced by two Republican lawmakers, Rep. Frank Wolf and Senator Dan Coats. The measure has 240 co-sponsors in the House, including more than 75 Democrats, and 13 co-sponsors in the Senate. Its leading Democratic proponent is Rep. Patricia Schroeder, new chair of the House Select Committee on Children, Youth, and Families.
Once again, however, widespread support does not necessarily indicate sound policy. While vastly superior to plans to liberalize IRAs, the personal exemption proposals still have serious drawbacks. The fundamental problem with a simple increase in the personal exemption is that it provides the largest benefits to those in the top income tax bracket. Each $1,000 increase in the personal exemption is worth $310 to a taxpayer in the 31 percent tax bracket but only $150 to a taxpayer in the 15 percent bracket. In 1992 a married couple that has two children and does not itemize deductions remains in the 15 percent bracket until its adjusted gross income exceeds about $51,000. The 31 percent bracket, by contrast, does not start for a married couple with two children until income surpasses about $110,000. Thus, a flat increase in the personal exemption confers more than twice as large a tax benefit on a taxpayer making $125,000 a year as on one earning $25,000.
Low- and moderate-income taxpayers would gain little or nothing at all. Americans whose incomes are too low to owe federal income tax are not touched by a personal exemption increase. In 1992 families of four with incomes below $15,250 will fall into this category. And moderate-income working families of four in the $15,000 to $20,000 range gain only a small benefit.
Consider the impact of raising the personal exemption from its scheduled level of $2,300 in 1992 to $3,500, as the Wolf-Coats bill would do. The following table shows the tax benefits for a two-parent family of four at different income levels.
A tax relief bill with this effect would be a strange response by Democrats to widening income disparities, especially since the increased disparities were partly caused by tax cuts skewed to the wealthy.
In an attempt to address this problem, two recent Democratic proposals tie increases in the personal exemption to a household's income. Separate bills introduced by Senator Christopher Dodd and Representative Rosa DeLauro, both of Connecticut, would raise the personal exemption by 50 percent for taxpayers in the 15 percent tax bracket and by 25 percent for those in the 28 percent tax bracket. Taxpayers in the 31 percent bracket would not receive a personal exemption increase.
This approach is superior to a flat personal exemption boost. But is it the best Congress can do? This approach yields about the same average dollar gain to hard-pressed middle- and lower-middle-income taxpayers, such as four-person families in the $20,000 to $50,000 range, as it would to upper-middle-income households. A 50 percent increase in the personal exemption is worth $172.50 per exemption to taxpayers in the 15 percent bracket. A 25 percent increase in the exemption is worth $161 to the average taxpayer in the 28 percent bracket. In addition, poor and near-poor households, including families with an earner working full-time for low wages, would receive little or no benefit at all.
Yet a new Congressional Budget Office (CBO) analysis shows that the average after-tax incomes of the 100 million Americans who make up the bottom two-fifths of the income spectrum -- and who would either be in the 15 percent tax bracket or owe no income tax -- have fallen since 1977, after adjustment for inflation. The average income of the middle fifth of households, who generally would also be in the 15 percent bracket, has edged up 2 to 4 percent. But the average after-tax income of upper-middle income households likely to be in the 28 percent bracket has climbed more than 10 percent. Among the wealthiest 5 percent of taxpayers, most of whom would be in the 31 percent bracket, after-tax income rose more than 60 percent during this period.
The Refundable Children's Tax Credit
While the DeLauro-Dodd approach has some attractions, they pale in comparison to the third major approach, converting the personal exemption for children to a refundable children's tax credit. This idea received national attention when proposed earlier this year by the National Commission on Children; it is also the central feature of legislation introduced by Representative Tom Downey and Senator Albert Gore. The tax credit provides greater tax benefits for the core of the middle class (those in the 15 percent tax bracket), the lower-middle class, and the working poor than for upper-middle and upper-income taxpayers.
Why a credit has this effect is easy to see. Suppose the personal exemption for each child is replaced by an $800 credit, as Downey and Gore propose. Taxpayers in the 15 percent bracket lose a personal exemption worth $345 for each child in 1992 (15 percent of $2,300), while they gain an $800 tax reduction. Their net benefit is $455 per child.
For those in the 28-percent bracket, the personal exemption will be worth $644 next year. So they would gain $154 per child. And for taxpayers in the 31 percent bracket, the net benefit would be $87 per child, or about one-fourth what those in the 15 percent bracket gain. Thus, the distributional effect of a flat personal exemption increase is neatly reversed. And because the credit concentrates its benefits on families with children in the 15 percent tax bracket, it provides more relief to a typical middle-income family with children than do alternative tax cut approaches that cost the same amount.
The progressivity of the credit is not its only attraction. While providing larger benefits than any other approach to typical middle-income families with children, the credit is also the one approach that assists poor families. Since the credit would be refundable -- meaning that families would receive a check for the amount by which the credit exceeded their income tax liabilities -- it would reach poor families that owe no income tax as well as those at greater income levels. After two decades in which child poverty has been rising as wages from low-paid work declined and public assistance benefits eroded, establishing a refundable tax credit for children would be a major step. Indeed, it would represent one of the more important antipoverty policies of recent years.
This measure would effectively create a children's allowance in this country. The United States stands virtually alone among Western industrialized countries in lacking some form of children's allowance or similar benefit for families raising children, a point many poverty analysts and children's advocates have long decried. The National Commission on Children, despite the ideological diversity of its thirty-four members, unanimously recommended that the personal exemption for children be replaced with a $l,000-per-child credit.
One reason that some conservatives as well as liberals support the idea of a credit is that, unlike welfare, the credit is less vulnerable to the charge that it discourages work and marriage. Welfare benefits can drop as much as one dollar for each additional dollar in earnings and can disappear altogether if a single parent marries. But taking a job or getting married would not diminish a child tax credit.
Given these advantages, you might think Democrats would flock to the tax credit approach. For several reasons, however, it remains unclear what course they will pursue.
First, to some policy makers, the tax credit concept is synonymous with the Downey-Gore bill, which is one of the few proposals that raises other taxes to offset the revenue losses it generates. While neither the Bentsen-Roth ERA bill nor the Wolf-Coats personal exemption bill include mechanisms to "pay for" their tax cuts, the Downey-Gore measure seeks to cover its costs by raising taxes on high-income individuals. It splits the current 31 percent top tax bracket into two brackets pegged at 32 percent and 35 percent and imposes an 11 percent tax surcharge on those with incomes above $250,000. It also stiffens the alternative minimum tax, which is designed to catch high-income filers who take so many write-offs they would otherwise pay little tax. Overall, Downey-Gore would lower taxes for 134 million Americans while raising them for just 15 million. Nevertheless, the inclusion of these tax-raising provisions has led some Democrats, including some lawmakers who otherwise favor the credit approach, to keep their distance. But since any tax relief measure will eventually have to include offsetting tax increases, this issue should not disadvantage one approach over another.
The other arguments against a tax credit approach pose more serious obstacles. Some object that it misses middle-income taxpayers without dependent children, a group that would benefit if the personal exemption were raised. But as Urban Institute researchers Eugene Steuerle and Jason Juffras argue, families with children are precisely the group to whom tax relief should be targeted. "Over the past four decades, gradual changes in the tax code have shifted more of the burden of taxation to families with children," Steuerle and Juffras note. "...At each income level, households with dependent children now pay a larger share of the total tax burden and households without dependent children pay a smaller share of the total tax burden" than they did in the late 1940s.
Moreover, although federal tax burdens are supposed to be related to "ability to pay," the tax code's current adjustments for family size are much too small to cover the costs of raising a child. The current structure reduces a middle-income family's taxes only $345 for each child in the family. From child care to higher education to housing, families with children face the tightest financial squeeze. Yet if the personal exemption approach were adopted instead of the credit approach, so many benefits would be provided to upper-income households and to those without children that benefits for typical middle-income families with children would have to be set at significantly lower levels.
A second argument is that the credit aids the poor as well as the middle class -- and a Democratic tax relief bill should be restricted to the middle class. Some espousing this view argue that Democrats must show they are squarely on the side of the middle-class voter and the way to do this is by fashioning a tax cut directed solely at middle-income households. That way, Democrats will prove their plan is not another welfare bill. Some supporting this view also contend that "the poor got theirs" in 1990 when Congress raised the earned income tax credit, a refundable tax credit for working families with incomes below $21,000.
At best, these arguments betray confusion on both substantive and political grounds. At worst, they suggest a loss of connection with long-standing Democratic ideals and constituencies.
The confusion stems in part from a view that proposals containing benefits for the poor will be regarded as welfare bills. Historically, however, measures that encompass the poor and the non-poor alike have not been viewed in this light, even when (as in the case of Social Security) the benefit structure is tilted toward those at the lower end. Moreover, various means-tested programs that include near-poor and lower-middle-income working families along with the poor, such as the earned income credit (with an income limit of $21,000) and last year's child care bill (with income limits of up to about $30,000), have also been free of a welfare taint.
The argument that the "poor got theirs last year" is also questionable. A recent analysis by the Congressional Budget Office projects that the average after-tax income of the poorest fifth of households will decline 10 percent from 1977 to 1992, after adjusting for inflation. Meanwhile, the average income of the middle fifth of households will edge up 2 percent, that of the top fifth will rise 34 percent, and for the richest 1 percent of the population, average income will more than double. If widening income disparities are a reason to tax the rich more and the middle class less, an even stronger case applies to assisting those closer to the bottom.
Other data indicate that for working poor families, cuts in public assistance benefits in recent years have far outweighed the increases in programs like the earned income tax credit. Calculations by House Ways and Means Committee staff show that the disposable income of a mother of two who earns wages equal to 75 percent of the poverty line (about what full-time work at the minimum wage now pays) was $3,100 lower in 1990 than in 1972, even though the earned income credit did not exist in 1972. Moreover, when fully phased in, the liberalization of the earned income credit enacted last fall will compensate for only one-fourth of this loss.
Recent international comparisons on child poverty also are notable. The Luxembourg Income Study, a comparative investigation of income levels, poverty rates, and government policies in a number of Western industrialized nations, recently found the U.S. child poverty rate in the mid-1980s was 20 percent, while Canada's was below 10 percent. The average rate in the other countries studied was 5 percent. The researchers also found that the poverty rate just among white children in the United States was higher than the poverty rate among all children in all other countries examined except Australia. The study found one key reason for the large gap in child poverty rates: the other countries took much stronger government action than did the United States to reduce poverty.
Still another argument raised against the refundable tax credit, primarily from the right, is that a flat credit such as that proposed by the National Commission on Children would unwisely benefit poor families in which a parent does not work. Income support not conditioned on changes in behavior, according to this view, will only encourage indolence.
The notion that a child tax credit not conditioned on work would be counterproductive is not borne out, however, by recent Canadian experience. Research by Maria Hanratty and Rebecca Blank shows that in 1979 the poverty rate among single-parent families was similar in the United States and Canada. By 1986, however, the U.S. rate was almost 15 percentage points higher. Hanratty and Blank found that nearly 90 percent of the sharp divergence that occurred during this period was due to benefit expansions in Canada and benefit cutbacks in the United States; one of the principal Canadian expansions was establishment of a refundable child tax credit in 1979. Moreover, these increases in benefits did not lead to sharp reductions in work among single-parents in Canada. On the contrary, the number of weeks worked per year by female single parents increased in Canada and also rose relative to the number of weeks worked by U.S. single parents.
Furthermore, even if one believes a flat refundable child tax credit would reduce work significantly, the principal credit proposal on the table -- the Downey-Gore bill -- addresses this issue by tying the amount of the credit a low-income family may receive to the family's work effort. For a family with no earnings, the credit would be limited to $400 per family. The credit would then rise with earnings, equalling 20 percent of family earnings until it reached the full $800 per child level2 Thus, far from discouraging work among the poor, Downey-Gore would promote it. For a family with two children, the credit would be four times larger if the family earns at least $8,000 than if the family has no income from employment.
A final argument some Democrats make against the credit is that it is too difficult for voters to understand. All voters will understand what an increase in the personal exemption means for them, the argument goes, but few will grasp the benefit they would receive from conversion of the personal exemption to a credit.
This argument sells the public short. It is reminiscent of claims some lawmakers made in the summer of 1990 that tables showing the distribution of tax burdens among different income groups were too complicated to use as part of the budget debate. The events of October 1990 proved this pessimistic view wrong. With a vigorous media effort, the Democrats could communicate the advantages of a tax credit to families in the $20,000 to $50,000 range.
One last problem facing the credit approach could ultimately prove its most serious hurdle. This is its large cost. While the revenue loss from a personal exemption increase can be controlled by reducing the amount the exemption is raised, this is not as true of the credit.
To be politically viable, the credit probably must be set high enough to provide a modest tax cut to taxpayers in the 28 percent bracket as well as to those in lower brackets. Since the personal exemption will be worth $644 next year to those in the 28 percent bracket, the credit probably cannot be reduced below $700 or $750.
At that level, the Downey-Gore approach would likely cost close to $100 billion over five years. A strong case can be made for raising that amount from other revenue sources, but it may be difficult to sell that argument in 1992.
If raising such an amount proves prohibitive next year, however, the credit should not be abandoned in favor of the flawed personal exemption strategy or the IRA proposal. On the contrary, the preferred approach should remain the refundable children's credit. If necessary, its costs can be reduced by limiting the credit to children below a certain age -- for example, those twelve or younger. Whatever Congress is willing to spend on tax relief, the credit approach will do more for typical middle- and low-income families with children than any of the competing approaches. And it will do more to honor the Democratic Party's desire to serve the middle class, without betraying its historic concern for those struggling to make it to the middle class.
