March 11, 1998
New Research Findings on the Effects of the
Earned Income Tax Credit
by Robert Greenstein and Isaac Shapiro
It has been widely recognized that one of the not-so-bright spots in today's glowing economy is the lack of significant economic progress for low- and moderate-income working families. It is somewhat surprising therefore that until recently, one of the nation's principal programs to assist these families the federal Earned Income Tax Credit received relatively little attention among researchers. This now appears to be changing. In recent years, as the EITC has expanded substantially, researchers have conducted an increasing amount of work on the EITC's effects.
The EITC was enacted in 1975. Substantial expansions were adopted in 1986, 1990, and 1993. The credit goes only to households with earnings, with the size of the credit initially rising as earnings increase. EITC benefits both offset taxes and, frequently, provide a wage supplement. More than 19 million low- and moderate-income households the vast majority of whom are families with children received the credit in 1996.
This analysis examines recent research, data, and findings on the EITC. The principal theme that emerges from this analysis is an encouraging one in a world where academic research often finds social programs to have modest effects at best in increasing work effort and employment, the emerging research literature on the EITC is strikingly positive. The research findings indicate the EITC increases work effort substantially among single mothers. The research also finds that the EITC moderates the widening of income gaps between the wealthy and the working poor and is now lifting 4.6 million people in working families out of poverty, including 2.4 million children.
- A series of studies have been conducted in the past few years on the EITC's effects on work behavior. These studies have consistently found that the EITC has substantial positive effects in inducing single parents to go to work. One of the most important of these studies finds that the proportion of single mothers who are in the labor force rose sharply between 1984 and 1996 and that the EITC expansions instituted during this period are responsible for more than half of this increase.
- Another study finds that the EITC offsets between one-fourth and one-third of the decline during the past 20 years in the share of national income received by the poorest fifth of households with children.
- Recent Census data show that among working families, the EITC lifts substantially more children out of poverty than any other government program or category of programs. The EITC is particularly important in reducing poverty among children in the South, where working families tend to have lower wages and consequently are more likely to qualify for the EITC. In addition, a just-released study by Columbia University's National Center for Children in Poverty found that the EITC reduces poverty among young children by nearly one fourth.
As the EITC has expanded, it also has become more politically controversial. In recent years, the EITC has begun to draw criticism in some political quarters on two grounds that its costs are escalating rapidly, making it one of the fastest growing federal programs, and that it is error-ridden. New data and research shed light on these issues as well.
- EITC costs rose significantly between 1986 and 1996 as a result of the EITC expansions enacted in 1986, 1990, and 1993, each of which phased in over several years. New cost estimates issued by the Congressional Budget Office, however, indicate that with the 1993 EITC expansion having phased in fully, significant EITC growth has ended. In fact, the EITC has now become one of the slowest growing of all entitlement programs between 1998 and 2003, the EITC is projected to grow more slowly than 13 of the 14 other largest entitlements. EITC growth is projected to be so modest that, when adjusted for inflation and population, EITC costs will decline between 1998 and 2003.
- The EITC error rate remains high, although it has fallen considerably over the past decade as a result of a number of legislative changes and tightened administrative procedures. Further error rate reductions are likely in the years ahead, moreover, as a result of a battery of significant legislative measures that either have very recently taken effect or will take effect in coming years, including a series of measures incorporated into the Taxpayer Relief Act of 1997 that were specifically designed to curb the types of EITC errors identified by IRS studies.
- Data from the Internal Revenue Service and the General Accounting Office show that EITC errors account for a relatively small share of total losses due to errors in the income tax. The IRS and GAO data indicate the amount lost to EITC errors accounts for less than five percent of the total amount the Treasury loses due to income tax errors. IRS and GAO data also show that a number of other provisions of the tax code have higher error rates than the EITC does and lose larger amounts of revenue. The errors in these other provisions, however, have not been the subject of much legislative scrutiny.
The EITC, whose early supporters include Ronald Reagan (who championed a similar proposal as early as 1972 while governor of California), has received bipartisan support throughout most of its history. This bipartisan support has eroded in Congress in the past few years, as the program has become increasingly associated with President Clinton, who proposed the major EITC expansion enacted in 1993. This erosion has occurred despite the fact that Presidents Reagan and Bush were heavily involved in the large EITC expansions enacted in 1986 and 1990. But outside the national political arena, the EITC's unique features have continued to generate strong support across the political spectrum. Several states have established state earned income tax credits in recent years with bipartisan support, and the EITC has drawn praise from some leading conservative thinkers, as well as from liberals and moderates.
For example, in a 1996 Wall Street Journal column, conservative Harvard economist and Journal contributing editor Robert J. Barro observed: "...There exists a serious program in the form of the earned income tax credit that actually helps the working poor in a way that promotes work and discourages welfare. The EITC was originally a Republican idea started by the Ford administration in 1975 and expanded by the Reagan administration during the glorious 1980s and the Bush administration in 1990....Mr. Clinton's support is not sufficient reason to regard the program as mistaken. In fact, it has a well conceived structure that ought to be retained and perhaps expanded in a comprehensive welfare reform package."(1)
Similarly, in a 1996 Business Week article, the well-known conservative economist and Nobel laureate Gary S. Becker praised the EITC for aiding poor families without reducing employment, discouraging work, or increasing reliance on public assistance. Becker wrote that the EITC "rewards rather than penalizes poor families with working members....Empirical studies confirm the prediction of economic theory that the EITC increases the labor force participation and employment of people with low wages because they need to work in order to receive this credit." Becker also applauded the EITC for being "fully available to families with both parents present, even where only one works and the other cares for their children...."(2) [i.e., for being available to low-income working families with stay-at-home mothers].
Background on the EITC
Although the benefits of economic growth have been shared across the income spectrum in the last couple of years, that has not been true for most of the past two decades. Income disparities the gaps between high-income families and those in the low or middle parts of the income range have grown considerably and remain close to their post-World War II highs.
- Census data show that both the poorest fifth of families and the next-to-the-poorest fifth had lower average incomes in 1996 than in either 1989 or 1979, comparable years of the business cycle to 1996. For example, after adjusting for inflation, the average income of the poorest fifth of families in 1996 was more than 10 percent lower than in 1979.
- In contrast, higher-income families experienced large income gains. In 1996, the average income of the top fifth of families was 28 percent higher than in 1979.(3)
These trends reflect the growing disparities between wages and salaries paid for highly compensated work and those paid for low-paying work.(4) Those at the top of the wage scale have secured large wage and salary gains. Less-educated workers have fared poorly. An analysis of Census Bureau data by the Economic Policy Institute found that from 1979 to 1996, the average hourly wage of workers with less than a high school education fell 26 percent, after adjusting for inflation. High school graduates experienced a wage decline of 13 percent.
A study by economist Rebecca Blank found the same pattern of erosion in the wages of less-educated workers between 1979 and 1993. Blank also found that fringe benefits deteriorated for low-wage workers, with the proportions of low-wage workers that receive employer-based health insurance and pension benefits dropping significantly between 1979 and 1993.(5)
These wage and compensation trends among less-educated workers are of particular note because these are precisely the groups of workers on which the EITC is targeted. Jeffrey Liebman, a Harvard economist who has conducted a series of studies of the EITC, found that in 1990, some 78 percent of EITC beneficiaries had no more than a high school education. Some 41 percent did not have a high school diploma.(6)
How the EITC Works
The Earned Income Tax Credit was established in 1975 to offset the adverse effects of Social Security and Medicare payroll taxes on working poor families and to strengthen work incentives by increasing the renumeration from low-paid work.(7) Administered through the federal income tax system, the EITC is a refundable tax credit. If the amount of a family's credit exceeds its income tax liability, the family receives a refund check for the difference from the Internal Revenue Service.
The EITC is targeted on low- and moderate-income workers. In 1997, approximately 16 million working families with children benefited from the credit.
The maximum benefits for tax year 1997 are $2,210 for families with one child and $3,656 for families with two or more children. These maximum amounts go to families with one child that have earnings of at least $6,500 and adjusted gross income of no more than $11,930, and to families with two or more children that have earnings of at least $9,140 and adjusted gross income of no more than $11,930. Benefits phase down gradually once income exceeds $11,930. (See Figure 1.)
Unlike public assistance programs in some states, the EITC does not treat two-parent families less well than single-parent families by restricting the eligibility of two-parent families. Two-parent and single-parent working families at the same income level receive the same EITC benefit.(8)
Similarly, two-parent families receive the same EITC benefit amount if they have the same level of income, regardless of whether both parents or only one parent in the family works. In the past, the EITC has been strongly praised by some conservatives for the support it provides to low-income working families with stay-at-home mothers.
Only families with children may receive the family EITC. Parents who do not live with their children do not qualify for the family credit, although they may be eligible for the small EITC for workers without a child in the home if their income is below $9,770. About three million very low-wage workers not raising minor children receive this small EITC, for which the maximum credit is $332 (and the average credit is $186). Because this credit is so small and the income limits for it are so low, it accounts for less than three percent of total EITC benefits.
Finally, only those who work can qualify for the EITC, and for families with very low earnings, the value of the credit increases as earnings rise. This is the opposite of how most means-tested benefit programs work; under most other programs, benefits fall as earnings rise. This aspect of the EITC operates effectively as a work incentive for those with no or very low earnings.
Academic Research on the Effects of the EITC on the Employment of Single Parents
As the size and prominence of the EITC has grown in recent years, so has the volume of academic research conducted on it. Perhaps the most striking research findings are those relating to the EITC's effects in increasing the workforce participation of single mothers.
Northwestern's Bruce Meyer and Dan Rosenbaum as well as Harvard's Liebman have noted that the percentage of single women with children who work has risen dramatically since the mid-1980s. In 1984, some 72.7 percent of single women with children worked during the year. In 1996, some 82.1 percent did. This increase has been most pronounced among women with less education.(9)
During this same period, there was no increase in work participation among single women without children. This raises a question why did the proportion of single women with children who work rise dramatically while the proportion of single women without children who work remained unchanged? Recent studies indicate the large expansions since the mid-1980s in the EITC for families with children have played a substantial role.
Studies by Liebman and University of California economist Nada Eissa, which have been published by the National Bureau of Economic Research and in the Quarterly Journal of Economics, find the EITC expansion contained in the Tax Reform Act of 1986 had a significant effect in inducing more single women with children to go to work.(10)
Of particular interest is a new study by Northwestern University economists Bruce Meyer and Dan Rosenbaum. Meyer and Rosenbaum find that a large share of the increase in employment of single mothers since the mid-1980s can be attributed to expansions of the EITC. They find these expansions explain more than half of the substantial increase in employment rates among single mothers over the 1984 to 1996 period; that is, the EITC expansions had a larger effect during this period than all other factors combined in spurring more single mothers to go to work. They also find that over just the 1992 to 1996 period, when many states instituted major welfare policy changes, the EITC expansions accounted for somewhat less than half of the increase in employment rates among these mothers.(11)
These findings are consistent both with the Eissa and Liebman study and with a study by Stacy Dickert, Scott Hauser, and John Karl Scholz of the University of Wisconsin. Dickert, Hauser, and Scholz projected that the EITC expansions in the 1993 budget law would induce a sizeable number of non-working single parents to seek employment and, in so doing, generate a reduction in welfare receipt. The Dickert, Hauser, Scholz study, published in a book edited by MIT economist James Poterba, estimated that the 1993 EITC expansions would induce approximately 500,000 families to move from welfare to the workforce.(12) The study suggests that a movement of 500,000 families from welfare to work means a reduction in spending on these families of an estimated $2 billion a year net of the increase in EITC payments to these families.
Recent academic research also contains encouraging findings regarding another issue related to the EITC whether the EITC reduces work effort among parents who already are working and are in the income range over which EITC benefits phase down as earnings rise. Some analysts have theorized that particularly among two-earner families in these income ranges, the EITC may cause some reductions in hours worked. The largest effects have been assumed to occur among mothers in two-earner families who take advantage of the income the EITC provides to work less and spend more time with their children.
The research conducted to date on this issue, however, does not bear out these theories. As Liebman noted in a recent paper: "While there is only limited evidence so far, the evidence that does exist suggests that the phase-out of the EITC has little or no impact on hours of work." He explains that he and Eissa examined taxpayers who were already in the labor force when the number of workers in the EITC phase-out range was increased substantially as a result of the EITC expansions enacted in 1986. He notes that "We observed no decline in hours [of work]...."(13)
Effects on Poverty and Income Disparities
Recent research also finds that the EITC reduces poverty significantly, especially among children, and moderates the degree to which income gaps have widened between rich and poor.
The EITC has a strong impact in reducing poverty both because it is a refundable tax credit and because of its targeting on working families with low incomes. If a family's EITC is larger than the income taxes the family owes, the family receives a check from the Internal Revenue Service for the difference; this is why it is termed a "refundable" tax credit. This feature is essential to the working poor since they typically do not owe federal income taxes. (For the working poor, payroll taxes constitute their largest tax burden.)
The targeting of the EITC, along with the fact that the largest EITC benefits go to working families below the poverty line, make it more likely to lift families out of poverty than programs that provide their largest benefits to families with no or little earnings. The EITC can provide the boost needed to lift low-income working families somewhat below the poverty line to above the poverty line. The EITC also is somewhat unique in that it provides substantial assistance, in the form of offsetting taxes and a wage supplement, to adults who find jobs and leave the welfare rolls but continue to be poor or to have modest incomes.
As the EITC has expanded in recent years, its impact in raising families out of poverty has grown much stronger. This effect does not show up in the Census Bureau's official poverty figures because those figures neither subtract income and payroll taxes from family income nor count the EITC as part of family income. Most economists who have analyzed these issues, however, believe that employment-related taxes and the EITC, as well as certain non-cash benefits such as food stamps, should be counted in measuring poverty. A National Academy of Sciences panel recommended such a change in the measurement of poverty in 1995.
The Census Bureau issues data and alternative poverty measures that include such effects. When income and payroll taxes, the EITC, and non-cash benefits are taken into account in measuring poverty, the effects of the EITC are seen to be large and to have grown noticeably over time.
- Census data show that in 1989, prior to the EITC expansions enacted in 1990 and 1993, federal income and payroll taxes pushed more families with children into poverty than the EITC lifted out of poverty. The net effect of federal income and payroll taxes and the EITC was to push 400,000 children from low-income working families into poverty in 1989.(14)
- By 1993, the impact of the EITC was large enough to offset fully the extent to which income and payroll taxes increase poverty among working families with children. The net effect of these policies was to move a very small number of children out of poverty.
- In 1994, the final phase of the EITC expansion enacted in 1990 took effect, and the expansion approved in 1993 began to phase in. With a larger EITC, the net effect of federal income and payroll taxes and the EITC was to lift 600,000 children out of poverty.
- The number of children removed from poverty due to the combined effect of federal income and payroll taxes and the EITC increased to 1.2 million in 1996, the year the EITC expansions enacted in 1993 phased in fully.
To see the full effects of the EITC in reducing poverty, we need to separate its impact from that of the rest of the federal tax system. A new Center report, Strengths of the Safety Net: How the EITC, Social Security, and Other Government Programs Affect Poverty, uses Census data to compare the number of people poor after all government programs and taxes are counted, but before the EITC is taken into account, with the number of people poor after counting programs and taxes including the EITC. The difference is the number of people lifted out of poverty by the EITC. When this is done, the EITC is shown to lift more children out of poverty than any other government benefit program.
- Data from the Census Bureau's most recent Current Population Survey show that in 1996, the EITC lifted out of poverty 4.6 million people in low-income working families that would have been poor without it.
- More than half of those whom the EITC lifted out of poverty 2.4 million people were children.
- In 1996, more than one in every seven children who would have been poor in the absence of any government benefits 14.5 percent were lifted from poverty by the EITC.
- Various other government programs, from Social Security and unemployment insurance to cash welfare assistance, food stamps, and housing assistance, also lift families out of poverty. Of all children moved out of poverty by government programs in 1996, some 37 percent nearly two in every five were lifted from poverty by the EITC. No other government benefit program or category of programs lifted as many children out of poverty. (This does not mean the EITC provides more support for poor families with children than any other program. The TANF cash assistance program provides more aid to poor families with children than the EITC program does. But since a much larger proportion of TANF families than of families receiving the EITC have incomes far below the poverty line and since the size of TANF benefits is reduced as a family's earnings increase TANF grants are much less likely than the EITC to lift families out of poverty. The same is true of the food stamp program.)
Not surprisingly, among children in working families that would be poor without the EITC, the effects are even larger. Some 30 percent of all children in families that otherwise would be poor despite significant earnings are lifted from poverty by the EITC.(15) This constitutes about half of all children in working families moved out of poverty by all government benefit programs combined. The EITC provides more support to working families with children than any other government program.(16)
The EITC is especially effective in reducing poverty among certain groups of children and in certain regions of the country.
- The EITC lifted 800,000 Hispanic children out of poverty in 1996. This was nearly one-fifth 18 percent of all Hispanic children who were poor before counting government benefits. Nearly half of all Hispanic children moved from poverty by government benefit programs are lifted out by the EITC.
- Similarly, the EITC lifted out of poverty 18 percent of all children in the South who were poor before government benefits were counted, a higher proportion than in any other region. Here, also, close to half of all children lifted from poverty by government benefit programs are removed from poverty by the EITC.
The EITC has its most pronounced effect in the South because the South has a larger proportion of families in which a parent works but remains poor than the other regions due largely to a lower wage structure. Similarly, the EITC has a particularly large effect in reducing poverty among Hispanic children because the proportion of poor children living in families with a full-time worker is considerably greater among Hispanics than among either whites or blacks.(17)
The effect of the EITC on families with young children also is notable. A March 1998 study by the Columbia University's National Center for Children in Poverty found that in 1996 the EITC reduced the poverty rate among children under age six from 23 percent to 18.7 percent. This amounts to a reduction in poverty among young children of nearly one quarter.(32)
In addition to reducing poverty, the EITC has moderated the degree to which income gaps have widened between rich and poor. During the past 20 years, the share of national income that the poorest fifth of households with children receive has declined, while the share the top fifth receives has risen markedly. Liebman's work shows the EITC offsets 29 percent of the decline that occurred between 1976 and 1996 in the share of income that the poorest fifth of households with children receives.(18)
How Fast is the EITC Growing?
In recent years, some EITC critics have characterized the Earned Income Tax Credit as one of the nation's fastest-growing entitlement programs. Although once true, this characterization is no longer accurate. The EITC grew rapidly for a number of years as EITC expansions approved by Congress were phasing in. Now that the last expansion, enacted in 1993, has phased in fully, significant EITC growth has ended.
Congressional Budget Office projections show that from fiscal year 1998 to fiscal 2003, the EITC is expected to grow at a much slower rate than the economy and to decline in cost as a share of the Gross Domestic Product. The projections also show that the EITC is now one of the slowest growing entitlements. Of the 15 largest entitlement programs, the EITC is growing more slowly than all others except farm price supports.(19) The Office of Management and Budget's cost projections are very similar.
Projected Growth in EITC Costs, 1998-2003
Average annual growth rate Growth rate adjusted for inflation and population The economy (GDP) 4.5% 1.2% EITC 2.5% -0.7%
CBO projects that from 1998 to 2003, the EITC will grow at a 2.5 percent annual rate, well below the projected annual rate of 5.3 percent for overall mandatory spending and just more than half the 4.5 percent average nominal growth rate projected for the economy. In fact, after adjustment for inflation and population growth, EITC costs are projected to decline at a rate of 0.7 percent per year.
The EITC Error Rate
The major criticism that continues to be levied against the EITC is that its error rate is too high. House Speaker Newt Gingrich has cited savings from eliminating EITC errors as one of the principal ways to finance a tax cut this year.(20) Several other Republican Congressional leaders have made similar comments, and House Republican conference chairman John Boehner has argued that increased highway spending could be financed by reductions in EITC errors.
The EITC error rate is high. An IRS study released in April 1997 estimated that 20.7 percent of EITC benefits are paid in error.(21) This is the figure Speaker Gingrich cited.
Nevertheless, a 20.7 percent error rate represents substantial progress in reducing EITC error rates since the late 1980s, when IRS studies estimated the error rate at approximately 35 percent. Tougher compliance measures and simplification of EITC rules have helped to lower the error rate appreciably.
It should be noted that the IRS study from which the 20.7 percent figure is taken is a study of tax returns for 1994 and that this figure appears to overstate the current error rate. More important, further significant reductions in EITC errors are expected in the years ahead as a result of substantial legislative measures enacted in the past few years, especially in 1996 and 1997.
The Limitations of the IRS Study
The 20.7 percent figure appears to overstate the current error rate for two reasons. First, the study does not reflect the effects of several legislative changes now in effect that were not in place during the period the study covered. These changes, summarized in the Appendix, are enabling the IRS to make much more effective use of Social Security numbers in verifying EITC eligibility and preventing payment of erroneous claims.
Second, due to limitations in the data that the IRS study collected, the study overstated the size of the errors in a number of cases it examined. For example, the study found nearly one-fifth of the errors occurred when a parent claimed a child for EITC purposes but, under the EITC's complex rules, another relative in the household such as a grandparent should have claimed the child instead. The loss to the Treasury in these cases is the amount by which the EITC amount paid to the parent exceeds the amount that should have been paid to the grandparent. The IRS study, however, was not designed to collect the data needed to determine the amount the grandparent should have received in such cases; as a result, the study simply classified the entire amount paid to the parent in these cases as being in error. The Treasury Department has noted that the study overstates the magnitude of the error in such cases.
In addition, the 20.7 percent figure does not adjust for underpayments. The IRS study indicates the EITC underpayment rate is at least 1.7 percent. The net loss to the Treasury is thus something less than 19 percent.(22)
Recent Legislative and Administrative Actions to Curb Error
The Taxpayer Relief Act of 1997 and the Balanced Budget Act of 1997 contain a series of major error-reduction measures. These new measures are based on the IRS study and were designed to reduce errors the study found. The measures, summarized in the Appendix and described in more detail in a Center analysis, The Earned Income Tax Credit and Error Rates, hold promise of reducing EITC errors significantly in coming years.
These measures include new mechanisms through which the IRS will be able to access the Federal Child Support Case Registry, which the welfare law established, to ascertain if a parent claiming a child for the EITC is a non-custodial parent and thus ineligible for the EITC. The 1997 law also requires that starting in 1998, whenever a child receives a Social Security number, the numbers of the child's parents be listed. This will link a child's Social Security numbers to those of his or her parents in the Social Security database, enabling the IRS to verify, by accessing the Social Security database, whether EITC filers presenting themselves as parents of children they claim for the EITC actually are the parents. In addition, the 1997 legislation effectively provides the IRS $716 million over five years for substantially intensified efforts to curb EITC errors. The IRS has received the first installment of these funds and instituted an aggressive error-reduction program that includes screening approximately four million returns claiming the EITC this year.
A substantial portion of the EITC errors did not reflect deliberate misrepresentation on the part of the tax filers. The IRS study found relatively few EITC errors to have resulted from actions by families to conceal incomes, a more common source of error in some other means-tested programs. Much larger shares of EITC errors were related to the varied and often-changing living arrangements of families and the complex and sometimes contradictory rules the tax code establishes with regard to family relationships, tax filing status, and the tax treatment of children in divorced, separated, and multi-generational families and families where the caretaker is someone other than the parent. This is true both for most inadvertent errors and for most errors that appear intentional. The new error-reduction measures, such as the expansion of the child support registry and granting IRS access to it, are designed to respond to these findings (although the findings indicate that simplification of complex tax rules relating to the EITC also could lower error rates).
Finally, the study contains one other significant finding it indicates EITC errors are not due primarily to the EITC's refundable nature. The study found the error rate among working families that have incomes too low to have a pre-EITC income tax liability (and consequently receive their full EITC payment in the form of a refund check) was only one-third as high as the error rate among families that do have an income tax liability before the EITC is applied.
Is Reducing EITC Error Rates a Realistic Way to Raise Revenue?
Statements that large, additional EITC error-reduction savings can be secured this year to help finance a new round of tax cuts are very questionable. There presently are no legislative options that would yield billions of dollars in savings from EITC error rate reduction. All known viable options that yield significant savings have been enacted, and Congressional leaders who have spoken about eliminating or reducing EITC errors to finance tax cuts or highway spending increases appear not to have any specific EITC error-reductions proposals in mind that would yield such savings. (The Administration's budget contains several useful legislative proposals to reduce EITC errors, but these provisions are not large enough to finance significant tax cuts; the Treasury Department estimates they will save approximately $400 million over the next five years.)(23)
Recent history is instructive here. Although a number of members of Congress cited the EITC error rate during the 1995 budget debate as justification for alterations in the EITC, the budget bill Congress passed that year (and President Clinton later vetoed) contained no error-reduction measures except for those the Treasury Department had proposed. Cost estimates issued by the Joint Tax Committee on Taxation, Congress' official scorekeeper, showed that approximately 95 percent of the large EITC savings in the 1995 budget reconciliation bill came from cuts in EITC benefits which would primarily have affected low-income working families that completed their tax returns accurately not from error-reduction measures.
Furthermore, during the final stages of the 1997 budget negotiations, some Congressional negotiators sought EITC savings in return for agreeing to an Administration request to allow more low- to moderate-income families to qualify for the new child tax credit. Here, too, Congressional leaders were unable to come up with EITC error-reduction proposals. All of the EITC error-curtailment measures ultimately included in the new tax law came from the Treasury Department.
There are no remaining Treasury legislative proposals that would produce sizeable EITC savings. In addition, as in previous years, Congressional leaders have not advanced specific error-reduction proposals. It is unclear how new EITC legislation could produce large savings to finance substantial tax cuts this year unless the savings came from sizeable EITC benefit reductions that penalized honest low-wage working families and bore little relationship to error reduction.
Are EITC Errors a Principal Cause of Income Tax Error and Fraud?
The EITC error rate exceeds the average error rate for the income tax as a whole, which is approximately 15 percent. Many elements of the income tax, however, have higher error rates than the EITC does.
For example, the Internal Revenue Service has reported that 29 percent to 30 percent of business income is not reported on tax returns. The IRS also estimates that sole proprietors who formally operate businesses other than farms fail to report 31 percent to 32 percent of their business income. Similarly, some 31 percent to 32 percent of farm income and 27 percent to 28 percent of income from the sales of business property go unreported.(24)
The IRS and the General Accounting Office have estimated that the amount lost due to errors and fraud in the individual and corporate income tax is well over $100 billion a year. Data from the GAO and the IRS indicate that EITC losses account for less than five percent of this amount.(25) Unreported business income alone results in $40 billion in lost individual income tax collections, substantially more than the EITC's entire cost.(26)
Statements from Congressional leaders that tax cuts can be financed partly by reducing errors have focused solely on EITC errors. No mention has been made of reducing errors in other parts of the tax code, such as errors and abuses in tax credits for corporations or for individuals at higher income levels.
In addition, some experts have warned that pending legislation to restructure the Internal Revenue Service could lead to an increase in undetected tax fraud in areas other than the EITC. The IRS legislation the House of Representatives approved last fall limits the use of "economic reality audits." As the New York Times has reported, "in an economic reality audit, IRS agents try to show that an individual cannot support his way of life with the income shown on his tax return for example, a taxpayer with a mansion and yacht whose tax return lists a modest income." The Times cited tax experts such as Donald C. Alexander, IRS Commissioner during the Nixon administration, as expressing concern that this provision would inhibit aggressive enforcement against tax evaders.(27)
EITC Continues to Draw Support From Across the Ideological Spectrum
Throughout most of its history, the EITC has received support across the political spectrum. Presidents Reagan, Bush and Clinton all proposed EITC expansions and signed expansions into law. In addition, in recent years, EITCs have been established or expanded in some states with bipartisan support; in 1997, Republican governors and Democratic legislatures worked together to establish a state EITC in Massachusetts and to expand a state EITC in Minnesota. In Oregon, a state EITC was established by a Republican legislature and a Democratic governor.
Similarly, the EITC has frequently drawn support from unusually broad coalitions. In March 1998, for example, a coalition of business, labor, and anti-poverty organizations jointly called for expanding New York's state EITC. Those endorsing the expansion include what the New York Times described as the state's two "foremost" groups representing large and small businesses, the Business Council of New York State and the National Federation of Independent Business.(28)
To some degree, however, bipartisan support in the U.S. Congress has eroded since the expansion sponsored by President Clinton was enacted in 1993. Even so, the EITC continues to draw support from various conservative policy experts.
For example, in a Wall Street Journal column published in 1996, conservative economist Robert J. Barro a Harvard economics professor who is a fellow of the conservative Hoover Institution and a Wall Street Journal contributing editor praised the EITC highly. Barro noted the bipartisan roots of the EITC and its positive work incentive effects. Barro wrote:(29)
"...there exists a serious program in the form of the earned-income tax credit that actually helps the working poor in a way that promotes work and discourages welfare. The EITC was originally a Republican idea started by the Ford administration in 1975 and expanded by the Reagan administration during the glorious 1980s and the Bush administration in 1990. However, the program has been identified with Democrats since the increase in benefits in 1993 (arguably the one sensible feature of the 1993 tax law).
"In 1996, a family with two or more children receives a 40% wage subsidy on earnings up to $8,890 (to yield a maximum credit of $3,556). This aspect of the program motivates labor-force participation, particularly by offsetting the work disincentives from the Social Security payroll tax and lost welfare payments....
"....Mr. Clinton's support is not sufficient reason to regard the program as mistaken. In fact, it has a well conceived structure that ought to be retained and perhaps expanded in a comprehensive welfare reform package.
"Nada Eissa of the University of California at Berkeley and Jeff Liebman of Harvard University have examined the effect of the EITC on labor supply....Their study compares work experience in 1984-86 before the increase of the credit in the 1986 tax reform with that in 1988-90. The key finding is that labor-force participation of people likely to be eligible for the credit notably single women with children rose by two to three percentage points relative to those people unlikely to be eligible. Moreover, since the EITC was made far more generous in the 1990 and 1993 tax laws, the positive effects of the present system on labor-force participation are likely to be greater....
"Opposition to the EITC focuses on two issues the large budget outlays and the high incidence of fraudulent claims. Two offsets to the extra spending are the reduced welfare dependency and the benefits from distributing income to a group that much of society regards as deserving the working poor with children. This seems to be a case (admittedly rare) where an excessive focus on budget numbers would curtail an effective government program....
"Two mitigating points about the fraud problem can be raised. First, transferring money to the working poor without EITC eligible children is not entirely bad.(30) Second, the IRS can and apparently now is doing better at policing the EITC.
"It makes sense that a comprehensive welfare program would include EITC-type payments to the working poor...."
Two weeks after Barro's column appeared, a column praising the merits of the EITC on similar grounds appeared in Business Week. This column was authored by noted conservative economist and Nobel laureate Gary S. Becker. Becker declared that the EITC helps the working poor without raising employer costs or encouraging dependence on welfare, increases work effort among those with no or very low wages, helps single-earner as well as two-earner families, and is well targeted.(31)
The ongoing support for the EITC across the ideological spectrum constitutes a relatively rare case of agreement about a method to combat poverty. The new evidence emerging from the academic literature about the positive effects of the EITC on workforce participation, as well as the evidence concerning the EITC's powerful effects in reducing poverty among the working poor and moderating income disparities, is likely to strengthen this support. Moreover, past concerns about the program's escalating costs should be allayed by the new projections indicating that significant program growth has halted and that costs are projected to decline modestly over the next several years, after adjusting for inflation and population growth.
The high rates of error in the program remain a concern. This concern has led to important legislative changes and intensified administrative efforts, many of which have either been implemented recently or are slated to take effect in the next few years. EITC error rates have already fallen substantially from their mid-1980 levels, and as a result of the new initiatives should fall further in coming years.
A paramount economic challenge facing the United States is to how to help ensure that the benefits of economic growth are shared broadly among workers at all income levels without interfering with the basic workings of the marketplace. In the face of the long-term trend of eroding wage levels for workers on the lower rungs of the earnings scale, there is growing evidence that the EITC is an important and effective policy tool to help address this challenge.
New Measures Should Further Reduce EITC Error Rates
The IRS study of error rates in the EITC does not reflect the effects of several significant anti-error measures that were enacted prior to 1997 and are now in effect, but were not in effect in tax year 1994, the period the study covered.
- In filing tax returns for 1994, filers did not have to provide Social Security numbers for children under age one. This created a potential for abuse; fictitious children could be claimed without much risk of detection if they were listed as being less than one year of age. Accordingly, legislation was enacted requiring taxpayers to provide Social Security numbers for all children, including those under age one. The requirement to provide Social Security numbers for infants took full effect in tax year 1997.
- In 1994, the IRS could hold up an EITC claim if it found what appeared to be a discrepancy in the Social Security number provided for a tax filer or his or her spouse. But the IRS ultimately paid most of these claims. It lacked the legal authority to deny such claims unless it could prove the Social Security number was not valid, and the IRS generally lacked the resources to secure such proof. As a result of legislation enacted in 1996, the IRS now simply denies these claims, with the burden of proof being shifted to the tax filer to demonstrate he or she has provided a valid Social Security number.
Last year's budget agreement contains measures that hold considerable promise for reducing EITC errors in coming years. These measures were specifically designed to address some of the principal causes of error the IRS study identified.
- The study found that errors occurring when a parent claims a child who did not reside with the parent more than six months of the year accounted for nearly two-fifths of all erroneous EITC claims. In the past, it was difficult for the IRS to verify the residency of children claimed for the EITC without corresponding with the taxpayer. To help address this problem, the 1997 act expands the data to be included in the Federal Child Support Case Registry and gives the IRS access to this important new database, which was established by the 1996 welfare law. In processing an EITC claim from an individual who files as a single individual or head of household, the IRS will be able when this database is fully developed in coming years to key into the database and ascertain whether an individual claiming a child for the EITC is classified as the child's custodial parent. If so, the IRS will be able to withhold the EITC payment and release it only if the claimant can show he lived with the child for more than half of the tax year.
- Until now, the IRS also has had no good way of determining whether an adult claiming a child for EITC purposes and presenting himself or herself as the child's parent actually is the parent. The IRS study found more than $500 million of EITC overclaims resulted from taxpayers claiming children who did not meet the EITC's relationship test. Accordingly, the 1997 Tax Act requires that starting in 1998, whenever a parent seeks a Social Security number for a child, he or she must provide the Social Security numbers of the child's parents. Social Security datafiles will link information on children's Social Security numbers to their parents' numbers. This will enable the IRS to verify parental status by keying into the Social Security datafiles.
Last year's budget agreement also takes one other important step. It effectively provides $716 million over the next five years to the IRS to beef up EITC enforcement activities. These funds should enable the IRS to institute the reforms described here and also conduct much more substantial efforts to identify questionable EITC claims and follow up on such claims. The IRS received the first installment of these funds in fiscal year 1998 and has begun to institute an aggressive error reduction program that includes the following components:
- The IRS is stopping returns with missing or invalid Social Security numbers; no EITC payment is being made on these claims unless filers provide valid Social Security numbers. The IRS projects these activities will result in 830,000 returns claiming the EITC being stopped this year, saving some hundreds of millions of dollars.
- The IRS has substantially enhanced the resources for the "Questionable Refund Detection Teams" it has established in its service centers. The additional resources will enable the teams to screen more returns and identify more questionable claims and questionable preparers. The IRS expects to screen four million EITC returns this year, some of which will be selected for more intensive examination. The IRS also is hiring and training 40 new special agents to investigate preparers whom the Teams identify as having submitted questionable EITC claims.
- The IRS sent out warning notices earlier this year to all taxpayers who last year claimed the EITC or the dependent exemption for a child also claimed on another tax return. Such action has never been taken before; it should have a strong deterrent effect. In addition, the IRS plans this year to examine 750,000 EITC claims, including claims submitted by tax filers who claimed the EITC last year for a child also claimed by another tax filer and claims identified for more intensive scrutiny by the Questionable Refund Detection Teams.
The 1997 legislation also toughens penalties for EITC fraud, enhances the IRS' ability to recoup EITC overpayments, and stiffens penalties on commercial preparers who fail to exercise adequate diligence in handling EITC claims. (The IRS study found the error rate to be 50 percent higher among claims filed by preparers who were not attorneys, CPAs, enrolled agents, or representatives of a nationally known firm.) This provision is designed to induce preparers to become better trained in EITC rules and more careful in handling EITC claims.
1. Robert J. Barro, "Workfare Still Beats Welfare," Wall Street Journal, May 21, 1996.
2. Gary S. Becker, "How to End Welfare 'As We Know it' Fast," Business Week, June 3, 1996.
3. The families included in these figures include those without children. If just families with children are examined, the poorest fifth of these families experienced on average a drop in income of 21 percent from 1979 to 1996, after adjusting for inflation. The most affluent fifth of families with children experienced an average income gain of 34 percent over this period.
4. Unpublished tables from the Economic Policy Institute indicate that wage inequality has grown in 15 of the 17 years since 1979. The Institute measured this trend by examining the "Gini coefficient" for wages, a technical measure of inequality generally regarded by economists as the best available measure.
5. Rebecca M. Blank, It Takes a Nation: A New Agenda for Fighting Poverty, Russell Sage Foundation and Princeton University Press, 1997.
6. Jeffrey B. Liebman, "The Impact of the Earned Income Tax Credit on Incentives and Income Distribution," included in Tax Policy and the Economy, National Bureau of Economic Research, edited by James M. Poterba, forthcoming.
7. The employee share of payroll taxes consumes 7.65 percent of the earnings of these families, with employers paying an equal amount. Most economists believe that the employer's share of the payroll tax effectively comes out of workers' paychecks as well that is, wages would be higher if employers didn't have to pay this tax.
8. Some policymakers have expressed concern that the structure of the EITC creates "marriage penalties" because some individuals would receive a smaller EITC if they married than if they remained single. In actuality, the impact of the EITC on individuals who marry is mixed, with some families facing an increase in EITC benefits if they marry and others facing a reduced EITC benefit. Of particular note, the EITC provides a "marriage bonus" to unemployed single mothers and low-wage male workers; if they marry, they generally qualify for a substantially larger EITC.
9. Jeffrey B. Liebman, op. cit.
10. Nada Eissa and Jeffrey B. Liebman, "Labor Supply Response to the Earned Income Tax Credit," Quarterly Journal of Economics, May 1996, 112(2), pp. 605-637 and Liebman, op. cit.
11. Bruce D. Meyer and Dan T. Rosenbaum, "Welfare, The Earned Income Tax Credit, and the Labor Supply of Single Mothers," March 7, 1998.
12. Stacy Dickert, Scott Hauser, and John Karl Scholz, "The Earned Income Tax Credit and Transfer Programs: A Study of Labor Market and Program Participation," in James M. Poterba, ed., Tax Policy and the Economy, Vol. 9., MIT Press, 1995. Scholz is now deputy assistant secretary for tax analysis at the Treasury Department.
13. Liebman, op. cit. Liebman noted that the confidence intervals in the analysis were such that he and Eissa could not rule out a small decline in hours of work.
14. All data cited in this section of the report are from the Census Bureau. The Census Bureau makes data available each year on poverty rates both before and after government benefits are counted as income. These data also show poverty rates after only social insurance benefits are counted, after all cash benefits (including means-tested benefits) are counted, after both cash and non-cash benefits such as food stamps, housing assistance, and school lunch benefits are counted, and after all such benefits are counted and the effects of federal income and payroll taxes including the EITC are taken into account. A new Center on Budget and Policy Priorities report, Strengths of the Safety Net: How the EITC, Social Security, and Other Government Programs Affect Poverty, examines in detail these Census data and what they show about the effects of various government benefit programs on poverty.
15. Poor families with significant earnings are defined here as families with earnings at least equal to current half-time minimum wage earnings throughout the year (i.e., earnings of at least $5,150, which equals 20 hours a week for 50 weeks at the $5.15-an-hour minimum wage).
16. This comparison does not include Medicaid, which is an insurance program rather than a program providing assistance families can use to meet ongoing expenses.
17. Census data show that in 1996, some 39 percent of Hispanic children classified as poor under the official poverty measure lived in families where someone worked full time throughout the year. By comparison, 32 percent of poor white children and 18 percent of poor black children lived in families with a full-time worker.
18. Liebman, op. cit.
19. In alphabetical order, the 13 mandatory programs expected to grow at a faster rate than the EITC from 1998 to 2003 are: child nutrition, civil service retirement, family support, food stamps, foster care, Medicaid, Medicare, military retirement, Social Security, student loans, Supplemental Security Income, unemployment compensation, and veterans compensation. Only farm price supports are projected to grow more slowly.
20. Speech before Cobb County Chamber of Commerce, Marietta, Georgia, January 5, 1998. Mr. Gingrich repeated this statement on February 11, 1998; see Congress Daily/A.M., February 12, 1998, p. 7.
21. Internal Revenue Service, "Study of EITC Filers for Tax Year 1994," April 1997. The IRS study estimated that 25.8 percent of EITC benefits claimed in 1995 (for tax year 1994) were overclaimed erroneously. The 25.8 percent figure refers to the EITC amounts erroneously claimed on tax returns not to the amounts actually paid out erroneously; the IRS catches some erroneous claims before payment is made, so the actual error rate is lower than the rate of erroneous claims. The IRS study estimated the rate of erroneous payments at 20.7 percent.
22. The IRS study estimated the EITC underpayment rate at 1.7 percent, but this figure is a low estimate because it reflects only cases in which a tax filer eligible for the EITC filed for the EITC but claimed too small an EITC benefit. This figure does not include cases where individuals eligible for the EITC filed for the EITC but claimed too small an EITC benefit. This figure does not include cases where individuals eligible for the EITC filed a tax return but did not claim the EITC on their returns. Had such cases been included to provide a full accounting of underclaims by eligible tax filers, the underpayment rate would have been greater than 1.7 percent.
23. Simplification measures could reduce the EITC error rates but generally would not save significant amounts that could be used to finance tax cuts and in some cases could increase costs. See Robert Greenstein, The Earned Income Tax Credit and Error Rates, February 1998.
24. Internal Revenue Service, Federal Tax Compliance Research: Individual Income Tax Gap Estimates for 1985, 1988, and 1992, April 1996. The figures cited here are for tax year 1992, the latest year for which these data are available.
25. GAO and IRS reports estimate that individual and corporate income tax losses from error and fraud totaled $128.4 billion for tax year 1992. The IRS estimates its enforcement actions reduced these losses by about one-sixth. If net revenue losses due to income tax errors and fraud are at the same dollar level today as in 1992, then losses due to erroneous EITC payments using the 20.7 percent figure in the IRS report and current EITC cost levels constitute less than five percent of these losses. (In fact, net revenue losses in the tax code as a whole are almost certain to be significantly higher today than in 1992, because total tax liabilities are larger. IRS studies show that the dollar amounts lost to error and fraud in the tax code as a whole increase as the total amount of income tax owed rises.) See General Accounting Office, "Earned Income Credit: Noncompliance Relative to Other Components of the Income Tax Gap," June 13, 1997, and Internal Revenue Service, op. cit.
26. Internal Revenue Service, op. cit., and General Accounting Office, op. cit.
27. The Times quoted Lee A. Sheppard, a tax expert who critiques tax law for the highly regarded publication Tax Notes, as stating that the provision to limit the IRS' ability to use economic reality audits to detect tax evasion "should be called the mobsters and drug dealers tax relief act of 1997." David Cay Johnston, "Helping a Few Taxpayers At the Expense of Many?," New York Times, November 2, 1997.
28. Richard Perez-Pena, "Tax Credit Rise Urged for Poor in New York," New York Times, March 2, 1998.
29. Robert J. Barro, op.cit.
30. Liebman has found that most tax filers who receive the EITC erroneously have low incomes. Liebman found that in 1988, six of every seven ineligible EITC recipients had income below the EITC income limits. These recipients were ineligible for the EITC for other reasons, such as lack of a qualifying child.
31. Gary Becker, "How to End Welfare 'As We Know It' Fast". Business Week, June 3, 1996.
32. National Center for Children in Poverty, Columbia School of Public Health, Young Children in Poverty: A Statistical Update, March 1998.