March 28, 2007
STATE EARNED INCOME TAX CREDITS AND THE "OVERPAYMENTS" ISSUE
By Nicholas Johnson, John Wancheck, Robert Greenstein
The federal Earned Income Tax Credit is a tax credit for low- and moderate-income working families. The EITC is widely hailed for its success in subsidizing work, reducing poverty and making the tax code fairer. For these reasons, 20 states have enacted EITCs that piggyback on the federal credit and thereby further its successes. Other states are considering similar measures.
The federal EITC, however, is sometimes is criticized for having a high rate of overpayments. Based on these reports of overpayment rates, some state policymakers have questioned the wisdom of enacting a state version of the federal EITC.
In fact, overpayment reports tend to exaggerate the actual level of overpayments. In part this is because of methodological problems in the IRS studies of 1999 and 2001 upon which the current overpayment rate estimates are based. In addition, significantly improved enforcement efforts by the IRS and legislative changes enacted since 2001, the year on which current IRS estimates of overpayments are based, should have resulted in a decline in overpayments that is not reflected in the error rate estimates. Finally, as explained below, IRS’ efforts to reduce EITC overpayments are growing more effective each year, and in February the Administration submitted EITC simplification proposals to Congress that would shrink overpayments further.
EITCs at both the federal and state level function quite well despite the overpayment problem. Administrative costs are very low (with most of the administrative work being carried out at the federal level), and the rate of participation among eligible families is quite high. Numerous tax policy experts have voiced strong support for the credit. Current IRS Commissioner Mark Everson, for instance, calls the EITC “one of the government’s most successful anti-poverty programs” even as he has expressed a strong desire to reduce overpayments and has intensified IRS efforts to do so.
States also can take their own actions to reduce errors in the EITC. Specifically, they may be able to make use of a promising new IRS initiative under which the IRS and state tax departments are beginning to collaborate on a variety of EITC compliance measures.
Why Do EITC Overpayments Occur?
The Internal Revenue Service considers an EITC to have been overpaid if, upon questioning, a taxpayer is not able to substantiate his or her EITC claim, or if the IRS specifically determines that the filer was ineligible or received too large of a credit. An overpayment of the federal EITC may occur as a result of honest error or deliberate fraud.
Although deliberate cheating certainly occurs, just as it occurs in other areas of the tax system, it accounts for only a fraction of the overpayments. The majority of the errors appear to result from the EITC’s complexity, especially with regard to families whose composition changes over the course of the year. IRS Commissioner Everson has said that studies have identified a minority of EITC overpayments as resulting from intentional action by tax filers.
The complexity of the EITC results in significant part from efforts by Congress to target the EITC carefully to intended recipients in order to minimize the budgetary cost and increase the effectiveness of the program. Overpayments often result from the interaction between the complexity of the EITC rules and the complexity of some families’ lives.
Consider, for example, a case where parents are divorced or separated. The EITC can be claimed only by the parent who has custody of the child for more than half the year, assuming that the parent is eligible. Sometimes, however, a non-custodial parent may erroneously claim the EITC related to that child. This could occur if the non-custodial parent pays child support and consequently has a perception of being eligible for the credit. This perception can be strengthened by the fact that a non-custodial parent who pays child support may be entitled, under the terms of a divorce agreement, to claim the child for the personal exemption and the Child Tax Credit. Under such a circumstance, the non-custodial parent may understandably assume he can claim the child for the EITC, as he does for the other tax benefits.
Or consider a case of married but recently separated parents. Normally, married couples must file joint returns to claim the EITC. For obvious reasons, however, separated parents often file their own returns. In such a case, a complex rule governs whether the custodial parent may claim the EITC if she files a separate return. For her to do so, she and her spouse must have lived apart for more than six months of the tax year, and she must have lived with the qualifying child for more than six months of the year. This part of the rule is straight-forward. But she also must be able to claim head-of-household filing status, and to do so, a parent must meet the “household maintenance” test. That test is complicated, hard to apply, and not well understood. As a result, numerous errors result, with low-income working mothers who are separated but not yet divorced mistakenly claiming head-of-household status. (As described below, one of the principal EITC simplification proposals that the Administration has submitted to Congress this year would greatly simplify this rule, and thereby significantly reduce EITC errors.)
It may be noted that the EITC is rarely claimed for children who do not exist. For a number of years, the IRS has used a database of Social Security numbers to screen and automatically reject tax returns with improper Social Security numbers. Rather, EITC overpayments typically result from misunderstanding of how to apply the EITC rules regarding who may claim a child, especially in complex or changing family situations, or from incorrect information about family composition or income levels.
Note also that there is no evidence that the refundable nature of the EITC leads to overpayments. A 2000 study by Treasury economist Janet McCubbin found that the error rate among tax filers who were receiving the EITC as a refund was less than one-third the error rate among tax filers for whom the EITC acted to reduce income tax liability. After conducting a sophisticated statistical analysis that took into account a variety of other factors, McCubbin found no evidence that the credit’s refundable nature was related to the error rate.
The Actual Level of Overpayments Is in Doubt
The IRS estimates that EITC overpayments for tax year 2004 represented approximately 23 percent to 28 percent of the EITC amounts paid, but these figures are in doubt. These figures were based on a sample of 2001 federal income tax returns that was analyzed as part of the IRS’ National Research Program, which sought to assess the overall “tax gap,” or level of tax noncompliance. The IRS then adjusted the error rate estimate from that sample in order to reflect the Treasury Department’s estimate of the effects that a package of EITC changes enacted in 2001 have had on the error rate. The changes enacted in 2001 included simplification and related measures designed to reduce EITC overpayments.
If the 23 percent to 28 percent estimate were accurate (as explained below, there is reason to believe it overstates the EITC error rate), it still would represent a much lower rate of noncompliance than exists in other portions of the tax code. For example, the 2001 IRS “tax gap” study, which assessed noncompliance throughout the tax code, found that 51 percent of rent and royalty income, 57 percent of small business income and 72 percent for farm income went unreported. The total cost to the federal government from such misreporting of business income — estimated at $109 billion — is at least ten to twelve times the size of total EITC overpayments.
But various experts who have examined the IRS error rate estimate — including the IRS’ National Taxpayer Advocate, who has studied this matter extensively — believe the IRS estimate is likely not to be accurate, because it is seriously compromised by major methodological problems that likely cause the estimate to overstate the actual error rate. The 23 percent-to-28 percent estimate has a number of significant weaknesses, including the following.
The IRS study on which the estimate is based assumed that any EITC claims not documented to an IRS official’s satisfaction were erroneous; other evidence indicates that such an approach significantly overstates the error rate. The study that the IRS conducted in 2001, like the similar study it conducted in 1999, was based on random audits of EITC claimants. Most EITC recipients cannot afford to hire lawyers or accountants to help them navigate an audit. If, in the course of the audit, the EITC claimant was unable to document his or her EITC claim, the claim was considered an overpayment.
That this approach almost certainly results in an overstatement of errors has been shown persuasively by Nina Olson, the IRS’ National Taxpayer Advocate. Olson has reported that her office has regularly found that when the IRS asks for documentation to prove the validity of a filer’s EITC claim, an extensive process involving numerous contacts with the filer is often necessary to show that a claim that IRS officials initially classify as erroneous — because it has not been documented to their satisfaction — is, in fact, valid. Olson has explained that in the studies the IRS has undertaken to estimate EITC overpayments, a sufficient process of this nature was not provided for the claimants whose EITCs were being examined and that this likely resulted in incorrect judgments in a number of cases that EITC payments were made in error.
Specifically, in a report to Congress, Olson pointed out that in EITC audits and other IRS actions where tax filers’ EITC claims are questioned, “Taxpayers often obtain completely different [i.e., more favorable] results” than those the IRS initially made when the taxpayers receive assistance from IRS’ Taxpayer Advocate Service (which the National Taxpayer Advocate oversees) in understanding what documentation the IRS is seeking and in assembling the appropriate documents. Olson reported that in nearly half of the cases where the IRS auditors classified an EITC claim as invalid and the filer requested and received assistance from the Taxpayer Advocate Service, the denial of the EITC was subsequently reversed. 
Since the IRS studies used to estimate error rates do not allow for an adequate process of this nature, Olson has explained, the IRS overpayment estimates are likely to reflect premature (and incorrect) judgments in a number of cases that EITC payments were made in error. In a report to Congress in 2003, she stated that for these reasons, she has concluded that the IRS study of EITC error rates in 1999 “overstates the overclaim rate.” (The IRS study of 2001 error rates, which uses a similar approach, had not been issued at the time she issued this report.)
The National Taxpayer Advocate has not been alone in questioning the validity of the IRS error rate estimate. A 2001 report by the Treasury Department’s Inspector General for Tax Administration reported “significant weaknesses in its [the IRS’] methodology” for estimating EITC error rates in its 1999 study. Similarly, at a Brookings Institution conference on tax administration in 2002, C. Eugene Steuerle, a Senior Fellow at the Urban Institute and Deputy Assistant Treasury Secretary for the Tax Analysis in the mid-1980s, raised major methodological concerns about the study.
The error rate estimate does not account for offsetting underpayments. Another reason the IRS studies on which the 23 percent-to-28 percent estimate is based are likely to overstate the level of erroneous payments is that the studies failed to take into account that in some cases where an EITC is wrongly claimed, another taxpayer may have been rightfully eligible for EITC but not have claimed it. Consider the case of a non-custodial parent who pays child support and mistakenly claims an EITC. If the custodial parent was eligible for the EITC but did not claim it because the non-custodial parent did, the IRS study would simply count the EITC that the non-custodial parent received as an overpayment, without netting out the EITC that the custodial parent should have claimed.
The error-rate estimate does not reflect the effects of various actions taken since 2001 to lower overpayments. IRS’ estimate of the EITC error rate takes into account the effects of statutory changes in the EITC enacted in 2001. But it does not take into account a statutory change enacted after that which is likely to be lowering error rates, nor does it take into account a series of important administrative actions that the IRS has instituted since 2001 to reduce overpayments. These changes are not reflected in the IRS’ error rate estimate because the IRS and Treasury do not yet have the data needed to quantify the magnitude of the error-reduction effects of these actions. The next section of this analysis describes these recent actions in more detail.
Actions Since 2001 to Reduce EITC Overpayments
In 2004, Congress enacted the Uniform Definition of Qualifying Child as part of the Working Families Tax Relief Act. That legislation much more closely aligned the definitions of “child” used in the EITC, the Child Tax Credit, and the dependent exemption.
As a result of this legislation, to claim a child as a dependent or to claim the child tax credit, a parent now typically must live with a child for more than half of the year. This is the same residency test as is used for the EITC. (As described above, there remains an exception for divorced couples whose child-support agreements allow the non-custodial parent to claim the child for the dependent exemption and the child tax credit.)
This represents a significant change. It has substantially reduced the number of cases in which a parent who cannot claim a child for the EITC can claim the child for the other two child tax benefits. As a consequence, this change should be reducing the EITC errors that result when a parent who properly claims a child for the other child tax benefits mistakenly claims the child for the EITC. The Uniform Definition of Qualifying Child also resolved inconsistencies in the rules for claiming foster children for various child tax benefits, another change that should be reducing EITC errors.
IRS Administrative Improvements
In addition, the IRS has taken a number of administrative actions since 2001 to reduce EITC errors. The IRS increasingly relies on automatic computer screens to identify questionable EITC claims (75 percent of EITC claims are now filed electronically), has intensified the examination of questionable EITC claims before paying them, and is undertaking more vigorous action with regard to commercial tax preparers that have high rates of erroneous EITC claim submission. Several of these IRS actions are worthy of particular note.
For example, the IRS has developed a powerful database to help it identify cases in which an EITC claimant may be wrongly claiming a child. The IRS uses this system to identify 500,000 EITC claims a year that it examines before making payment; many of these claims are denied or reduced as a result. This database, which was not well developed in 2001, combines data from the following:
The Federal Case Registry, a national database built in recent years that contains extensive information on custodial and non-custodial parent-child relationships. These data, based on child support cases, help the IRS identify cases in which a non-custodial parent is wrongfully claiming a child for the EITC;
Kidlink, a database of Social Security Administration records that contains the Social Security numbers of both children born since 1998 and these children’s parents. This database enables the IRS to ascertain whether an EITC claimant is the parent of the child he or she is claiming;
DM-1, another database of Social Security numbers and Individual Taxpayer Identification Numbers, and;
Numdent, which consists of Social Security Administration data that provide information from birth certificates, including parents’ names.
The IRS is now using this combined database extensively. Furthermore, the IRS refines its methods for using the database each year, based on experience, to improve its ability to identify EITC claims that are likely to be erroneous. As a result, the IRS is now able to identify and prevent payment on a growing number of inaccurate EITC claims.
Since 2001, the IRS also has improved its methods for identifying EITC filers who appear to be under-reporting their income. The IRS has recently altered the methodology by which it identifies problematic EITC claims when it compares the income data on the tax return with data reported by employers on information documents such as the W-2 form. A test run of this approach conducted in 2005, based on tax year 2003 returns, proved quite successful, and the IRS has now begun utilizing this approach on an ongoing basis to identify EITC cases for closer scrutiny.
The IRS also has launched an initiative to identify commercial tax preparers who have submitted a significant volume of erroneous EITC claims and to target the preparers for more intensive enforcement and education activities. (The IRS’ National Taxpayer Advocate has proposed an additional step to Congress: mandatory registration and competency requirements for all tax preparers. Currently, there is no requirement that a tax preparer have any training or demonstrate any knowledge of tax rules.)
The effectiveness of these initiatives in combating error appears to be growing year by year. For example, the Kidlink database covers children born since 1998. In 2001, it only covered children up to age 3. Today, it covers children up to age 8. Over time, it will grow to cover virtually all children born in the United States and their parents. With each passing year, the IRS’ error-rate estimate is likely to be increasingly out of date.
Continued Reduction in EITC Overpayments Is a Priority for the IRS and Federal Policymakers
Compliance with the EITC is likely to continue to improve over the next several years, not only because the IRS is continuing to test and implement more effective approaches to improving compliance, but also because the Administration and Congress are considering ways to simplify the EITC legislatively and make it less error-prone. In the budget it submitted to Congress in February 2007, the Administration included the following proposals.
Simplifying the rule governing how married parents who are separated can claim the EITC. For the reasons described on page 2, low-income working mothers who are separated but not yet divorced may mistakenly use the head-of-household filing status. When they do, EITC errors result. And because low-income parents who separate may take a long time to obtain divorces or legal separations if they cannot afford the legal expenses, such errors can continue for a number of years. One Administration proposal would greatly simplify these requirements by permitting a separated parent who lives with her child for more than six months of the year, and also lives apart from her spouse for at least the final six months of the year, to claim the EITC without having to meet the complex head-of-household filing test. Enactment of this proposal would cause an immediate, substantial reduction in EITC errors.
- Allowing filers who live with a qualifying child but do not claim the child for the EITC to claim the smaller EITC for workers not raising a child. Sometimes either of two adults who live in the same household with a child may qualify to claim the child for the EITC. This can occur, for example, when a mother, aunt, and child live together; the aunt may be eligible to claim the child for the EITC if the mother does not. If the mother does claim the EITC for the child, however, current rules prohibit the aunt from claiming the smaller EITC for very poor workers who are not raising a child, even if she otherwise qualifies. The Treasury proposal would address this inequity by removing the disqualification that applies to the aunt in such a case. This proposal would reduce errors. Under current rules, the aunt may mistakenly assume she can claim the EITC for childless workers, since she is not claiming a qualifying child for the EITC. If she claims the childless workers’ EITC, however, an error occurs.
- Increasing funds for low-income filing assistance. The Administration requested an additional $5 million in the fiscal year 2008 IRS budget proposal to expand the Volunteer Income Tax Assistance program. This expansion would focus on taxpayers who live in communities where it has proven more difficult to establish free tax filing programs, such as rural areas. By using IRS-trained volunteers to help taxpayers fill out their returns, this proposal could have a modest error-reduction effect.
Although a major focus of the IRS has been the prevention of EITC overpayments, there also are significant penalties for taxpayers and tax preparers who are found to have wrongly claimed the EITC, just as there are for other taxpayers who violate the tax laws. These can include civil and, in some cases, criminal penalties. Invalidated EITCs as well as monetary penalties can be withheld from wages, Social Security payments, and future tax refunds. In addition, taxpayers found to have wrongly claimed the EITC are blocked from claiming the EITC in the future unless they provide advance documentation of their eligibility. They must meet this requirement even after they have repaid the EITC amounts they wrongly received and also paid any required penalties.
From an Administrative Perspective, the EITC Has Strengths that Outweigh the Problem of Overpayments
The EITC is widely considered to be an effective and efficient program. Even before the IRS stepped up efforts to reduce overpayments, policymakers and tax administrators saluted the EITC for its effectiveness in reducing poverty, inducing low-income parents to work, and making the tax code fairer. As Treasury Secretary Henry M. Paulson said recently:
“The Earned Income Tax Credit helps Americans who work hard but need extra support to make ends meet – people who are often on the first step of the economic ladder, gaining the experience and skills to land a better job and earn a higher income in the future. ... Our goal is not just to help people get by. Our goal is to help people get ahead.” 
Participation in the EITC among eligible taxpayers is generally quite high; studies indicate that the participation rate is probably in the range of 80 percent to 85 percent. This is much higher than participation in other forms of means-tested assistance to families; for example, only about 50 percent of low-income working families eligible for food stamps receive them. Participation is high because the EITC is reasonably easy to apply for as part of filing an annual income tax return. In other words, particularly for low-income working families, the EITC is much more likely to reach its targets than other forms of assistance. This is part of the reason why the EITC is considered such a successful anti-poverty program.
In addition, administrative costs are very low. At the federal level, despite a substantially increased enforcement effort, EITC administrative costs remain less than 1 percent of total program expenditures. At the state level, because most state EITCs piggyback on the federal EITC, the percentage of state EITC costs going for administration should be even lower. The percentage of EITC costs that go for administration is much lower than the percentage for most or all other benefit programs.
Implications for States
For states, the administrative costs of an EITC are particularly low, in part because the computation of a state EITC is very simple. It requires only that a taxpayer multiply his or her federal credit by a specific percentage. An EITC typically adds only one line to a state tax form. Because it is so simple, there is little room for error in the taxpayer’s calculation of a state EITC based on the federal EITC. Since the IRS routinely shares computer files with state tax departments, states can check easily whether a taxpayer has made this calculation correctly.
This direct piggybacking of the state credit on the federal credit also has other advantages. When an EITC claimant (or, for that matter, any other tax filer) files a tax return electronically — as three-fourths of EITC claimants now do — the return will automatically be rejected by the IRS at the time of filing if it contains invalid Social Security numbers, incorrect calculations, or certain other flaws. The IRS also rejects EITC claims from people who have been found to claim the EITC erroneously in the past unless they submit specific documents confirming their eligibility for the credit. In most states, such a rejection is likely to lead to automatic rejection of the state return as well, as a result of joint federal/state filing arrangements that 38 of the 42 states with an income tax have with the IRS. (The exceptions are California, Maine, Massachusetts, and Minnesota.)
A number of other IRS efforts to minimize EITC errors at the federal level also reduce state errors. The IRS audits a large number of EITC recipients. (Some 44 percent of all IRS individual income tax audits now are audits of filers who claim the EITC, due in part to Congressional pressure to lower EITC overpayment rates.) Information about any adjustments that the IRS makes as a result of an audit is shared with state revenue departments. In addition, as noted above, the IRS is focusing education and enforcement efforts on tax preparers who have been found to submit a significant volume of incorrect EITC claims. Such efforts should lead to greater compliance with state EITCs as well.
In addition to benefiting automatically from federal administrative efforts on the EITC, states can take proactive steps of their own.
- States can undertake their own screening efforts to help identify potentially ineligible EITC claimants. States have access to Social Security numbers, wage records, and other databases that allow them to screen potentially erroneous EITC claims. The New York State Department of Taxation and Finance, for instance, uses computer modeling, based on several different databases of tax information and other information, to identify potentially erroneous EITC claims. This approach also helps tax department staff determine which potentially erroneous claims may be most worth pursuing.
- In a new development, the IRS has begun an effort over the past two years to work jointly with state revenue departments to improve compliance with both federal and state EITCs. These efforts include sharing data on a more timely basis, helping states develop compliance strategies, and avoiding duplication in efforts to educate taxpayers and tax preparers about the EITC. The IRS has indicated willingness to work with additional states to help them maximize EITC compliance. This approach appears to have the potential to reduce error rates to a greater extent than either the IRS or state revenue departments could achieve on their own.
- In addition to boosting enforcement after tax returns are filed, states also can expand assistance to taxpayers as they fill out their tax returns. For example, states can support the work of the Volunteer Income Tax Assistance (VITA) program, through which IRS-trained volunteers at local nonprofit organizations fill out tax returns free for low-income families. If EITC claimants use such services, they may not need to rely as much on sometimes inadequately trained commercial preparers or to try to fill out the complex tax forms themselves. This can reduce the number of erroneously claimed EITCs.
States can support the VITA program in a number of ways. They can refer participants in various government assistance programs to VITA sites or publicize the sites on state tax booklets, web sites, and the like. Several states provide direct grants to help nonprofit organizations operate such sites.
In short, a state EITC — like the federal EITC — has substantial strengths that are likely to substantially outweigh any drawbacks. C. Eugene Steuerle, who was Deputy Assistant Secretary for Tax Analysis at the Treasury Department in the Reagan Administration, and Leonard E. Burman, who held the same post during the Clinton Administration, recently wrote: “Experience with the federal EITC should be encouraging to states considering adopting their own. The federal EITC is an effective subsidy program that helps 22 million working families and has been an important part of a longer-term strategy to move poor households from welfare to work. … Those arguments apply with even more force, not less, at the state level.”
 Testimony before the Senate Homeland Security and Governmental Affairs Committee, March 9, 2006.
 Among other things, the household maintenance test requires filers to show that they pay more than half of household expenses from their own income, and requires them to subtract (or not to count) child support payments and public program benefits from that income even though such payments or benefits are commonly used for household expenses.
 Janet McCubbin, “EITC Noncompliance: The Determinants of the Misreporting of Children,” National Tax Journal, Vol. 53, No. 4 Part 2 (December 2000), pp. 1135-1164.
 IRS Earned Income Tax Credit Initiative, Report on Fiscal Year 2005 Tests, January 2007, p.14. The 23 percent to 28 percent range reflects the fact that a significant fraction of families selected for the 2001 study did not respond to the IRS when contacted. This is a typical occurrence when studies of this nature are conducted in means-tested programs, and past evaluations in other means-tested programs have found that many people who do not respond in such circumstances actually are eligible for program benefits but are intimidated by the request, do not understand the request (due to literacy problems or lack of fluency in English), or simply wish to avoid the hassle. The 23 percent EITC overpayment rate estimate — the low end of the range and in the IRS study — assumes that overpayments among the non-responders occurred at the same rate as among those who did respond. The 28 percent overpayment rate assumes that all of the non-responders were ineligible for every dollar of EITC benefits they received. The 28 percent figure, which represents the upper bound of the range, clearly overstates the true error rate.
 These figures represents the estimated impact of business underreporting in the personal income tax; they do not include underreporting or other sources of error in the corporate income tax. Internal Revenue Service, “Individual Income Tax Underreporting Gap Estimates, Tax Year 2001,” February 2006.
 Taxpayer Advocate Service, Internal Revenue Service, “The National Taxpayer Advocate’s Report to Congress: Fiscal Year 2004 Objectives,” June 30, 2003, pp. 20-21.
 Taxpayer Advocate Service, Internal Revenue Service, “The National Taxpayer Advocate’s Report to Congress: Fiscal Year 2004 objectives,” June 30, 2003, pp 20-21.
 Treasury Inspector General for Tax Administration, “There are Significant Weaknesses in the Internal Revenue Service’s Efforts to Measure Earned Income Tax Credit Compliance,” December 2001.
 IRS Earned Income Tax Credit (EITC) Initiative, Report on Fiscal Year 2005 Tests, January 2007, p. 85-86.
 Taxpayers may be banned entirely from claiming the EITC for a period of two years in the case of intentional disregard of eligibility rules or ten years in the case of fraud. In such cases, recertification of eligibility during the banned period is not possible.
 “Paulson, IRS Launch Campaign to Help Low-Income Taxpayers Take Advantage of Tax Credit, Free Tax Help,” IRS News Release, IR-2007-24, February 1, 2007.
 Mark Everson, testimony before the Senate Homeland Security and Governmental Affairs Committee, March 9, 2006.
 A few states have EITCs that are slightly more complicated. For instance, Maryland and Rhode Island have different rates for refundable and nonrefundable EITCs, Wisconsin has different rates depending on the number of qualifying children, New Jersey limits eligibility to families with incomes below $20,000, and Minnesota has its own “lookup table” for EITC amounts. In the majority of states with EITCs, however, the state credit is a flat percentage of the federal credit.
 Nina Olson, testimony before U.S. House of Representatives Appropriations Sub-committee on Financial Serivces and General Government, March 5, 2007. By way of comparison, about 17 percent of all tax returns include an EITC claim.
 Tom Heinz, Director of Audits, New York State Department of Taxation and Finance, “Predictive Modeling and Earned Income Tax Credit,” presentation to Federation of Tax Administrators, June 7, 2004. available online at http://www.taxadmin.org/FTA/meet/04am_pres/heinz.pdf.
 States that support free tax assistance projects realize other benefits as well. For one, they may be able to increase the extent to which families that are eligible for the federal EITC actually receive it. For another, they may be able to reduce the extent to which EITC claimants lose a portion of their credit to high-cost tax preparers and “refund anticipation loans” that can reduce the value of the credit by several hundred dollars. See National Governors Association Center for Best Practices, State Efforts to Support Low-Income Families and Communities Through the Earned Income Tax Credit, 2006.
 “The EITC Attempts to Subsidize Work, Rather Than Provide Welfare,” State Tax Notes, February 27, 2007.