Welfare Reform's Hidden Ally

The Earned Income Tax Credit (EITC) lifts more children out of poverty than any other social program in America -- indeed, more than all other means-tested benefit programs combined. That's because it provides a substantial tax credit for low-income working parents, which is also refundable. That is, it is paid in full to them whether or not they owe income taxes. Last year, 16 million low-income families with children received $30 billion under this federal tax program.

The amount a family receives depends on its earnings and whether it has one or more children. Currently, a single mother of two who makes the minimum wage receives 40 cents in EITC benefits for every dollar she earns. The more hours she works, the larger the EITC payment, until her wages reach $10,350 (about full time). Only when an income exceeds about $14,000 a year does the EITC payment begin to phase down. The net effect, according to an array of studies by academic researchers, has been a powerful inducement for low-income single mothers to work. Part of the credit often given to welfare reform for getting single mothers into the labor force properly belongs to the EITC.

To be sure, the EITC has problems. Its eligibility and benefit rules can be mind-numbingly complicated, especially for divorced or separated families, three-generation households, and families in which custody arrangements change during the year. Even professional tax preparers can find the rules difficult to apply. Partly as a result, the EITC error rate is high. While Congress last year took several important steps to simplify the rules and reduce errors, more needs to be done.

Unfortunately, a recently released Treasury Department study of EITC overpayments in 1999 has been treated by some policy makers -- and in these pages [See Robert S. McIntyre, "Free Money: Take Some," TAP, April 8, 2002.] -- as evidence that the program is severely flawed. McIntyre wrote that the Treasury study found $11 billion in EITC overpayments and showed nearly half of all EITC families to be cheating, for an annual "rip-off of $10 billion." While I am a long-standing McIntyre fan, these figures do not accurately reflect the study findings.

The $11 billion figure is the Treasury Department's upper-bound estimate of the EITC benefits claimed in error, not its best estimate of the amount actually paid in error after the IRS caught and corrected mistakes. Moreover, the $11 billion figure is based on the unlikely assumption that every one of the low-income families that did not respond to an IRS request to investigate its finances as part of this study was actually ineligible for the EITC it claimed.

The Treasury's "best estimate" is that between $8.5 billion and $9.9 billion in EITC payments should not have been made. This is a large amount and a source of serious concern. But the implication that virtually all EITC overpayments are due to cheating is mistaken and, for those looking to solve the program's problems, misleading. A large share of EITC errors appear to be due to the mismatch between the program's rigid and intricate rules and the complicated living arrangements of many low-income families.

Here's an example. An arcane EITC rule known as the "AGI tiebreaker" states that if a child and his or her parents live in the same residence as other relatives, only the relative with the highest adjusted gross income may claim the child for the EITC. Suppose Mary Smith earned $15,000 and lived with both her child and her mother, who earned $18,000. At tax time, Mary would appropriately claim the personal exemption for her child. She logically might assume that because she is the parent and supporter of her child and can claim the child for the personal exemption, she also can claim the child for the EITC. But she would be wrong. Her mother would be the only eligible claimant.

If Mary claimed the EITC in 1999, she would have received a payment of $1,906. Had her mother filed for the EITC, she would have gotten $1,427. Thus, the overpayment to the family would have been $479. But due to a flaw in the Treasury study's design -- which the department acknowledges -- the entire $1,906 would have been counted as error, even in the Treasury's lowest estimate of overpayments. And McIntyre's article would have counted Mary among his ineligible cheaters.

Is this an extreme example? Hardly. The Treasury reports that $1.6 billion of the overpayments in the study were due to this obscure AGI tiebreaker rule, which Congress, in any case, has now largely abolished. Similarly, at the time the Treasury study was conducted, the EITC used definitions of earned income and adjusted gross income that differed from the definitions used on other tax forms. Simply transferring the figures from the front page of a tax filer's 1040 form to the filer's EITC form could cause an error. Congress fixed this problem, too, effective in 2002.

Other examples abound of how complex EITC rules can trip up honest families. A low-income working mother whose husband deserts her and their children during the year and who is supporting her children by herself may assume that she can use the "head of household" filing status and claim her children for the EITC. But if she didn't secure a legal separation (which few families get these days) or divorce by December 31, or pass a nightmarishly complicated three-part test that few low-income families are likely to understand, she is ineligible for the EITC. (The Treasury proposed simplifying the three-part test in 2000, but Congress has yet to act on this matter.) By last year, the EITC had grown so complicated that the IRS instructions for the EITC were longer than the instructions for the "alternative minimum tax."

There is no denying that cheating does occur, and tough action is needed to address this problem. One such step is contained in last year's tax legislation. A key source of error involves noncustodial fathers who claim their children for the EITC. The IRS is now able to use a new national child-support registry to verify whether single fathers claiming their children for the EITC actually have custody. If the registry shows that they don't, the IRS will -- starting in 2004 -- summarily deny their EITC claims (unless the father can show that he did live with the child for more than half of the tax year).

With the institution of the measures enacted last year, the EITC error rate should come down -- although additional steps will be needed to reduce errors further. The IRS may be able to use computer technology to identify other types of tax filers who appear to be out of compliance with EITC rules, and to target these filers for audit or some other form of closer examination. Further EITC simplification would also help.

Desirable as well would be EITC benefit improvements -- particularly for families with more than two children and for very poor workers without children -- and additional reductions in EITC marriage penalties.

But we should not forget that the EITC remains one of the outstanding features of what increasingly is becoming a work-based safety net in the United States. Warts and all, its successes have been stunning. It would be a tragedy if policy makers lost sight of that fact.