I am pleased to have the opportunity to discuss the Administration's proposals to improve the earned income tax credit (EITC) and look forward to working with the Committee on this issue.

  The Administration is strongly committed to the goals of the EITC and will oppose any proposals which reduce the EITC and raise taxes on millions of working families who play by the rules. The goals of the EITC are to make work pay and to lift workers out of poverty in the most efficient and administrable manner possible. With its message of "work pays," the EITC helps reduce dependency on welfare and increase reliance on jobs.

ECONOMIC CONDITIONS AMONG LOW-WAGE WORKERS

  To understand the role of the EITC, a couple of facts about the labor market for low-skilled workers in the United States are useful.

  There has been a striking drop in real wages for unskilled workers, beginning in the 1970s and accelerating over the 1980s. Between 1979 and 1992, the earnings of full-time male workers who had not graduated from high school declined by more than 23 per cent in real terms. Among full-time male workers with a high school diploma, real earnings fell by 17 per cent over the same period.

  This decline in the real wage for many unskilled workers has serious implications. In the United States, it is still possible for a family, containing a worker, to live in poverty. According to the Census Department, there were 2.4 million persons, over the age of 16, who lived in poverty and had worked year-round at full-time jobs in 1995.

EFFECTIVENESS OF EITC IN MAKING WORK PAY AND REDUCING POVERTY

  The ETC makes work pay in two ways. Unlike many assistance programs for low-income families, the EITC is limited to working families. Moreover, the credit amount initially increases - rather than decreases - for each additional dollar of earnings. As a consequence, the EITC is different from many low-income assistance programs that are characterized by a reduction in benefits for each additional dollar of earnings. In my work prior to coming to Treasury, I - together with Stacy Dickert-Conlin and Scott Houser - examined the net impact of the OBRA 1993 expansion of the EITC on labor supply. We found that the EITC has a modest, positive effect on labour supply by encouraging individuals to enter the workforce. The EITC also directly increases the disposable income of working families. According to the most recent Census data, the EITC lifted 3.7 million persons out of poverty during 1995.

  By making work pay, the EITC increases the probability that some parents may enter the workforce and perhaps leave the welfare rolls. The EITC, then, plays a key role in our efforts to reform welfare.

ADMINISTERING THE EITC THROUGH THE TAX SYSTEM

  The EITC achieves the goals of making work pay and relieving poverty by reducing the tax liabilities of low and moderate-income families. Thus, it is improper to characterize the EITC, as some have done recently, as a "non-tax function" of the IRS. The EITC was created and expanded to offset the overall tax burden of low and moderate-income families and should not simply be measured as an offset to income and SECA taxes. About 85 per cent of EITC costs will offset the combined Federal tax burden of families receiving the credit in 1998.

  As these numbers suggest, EITC claimants are taxpayers. If the EITC did not exist, almost all EITC filers would still file an individual income tax return (in addition to paying payroll and excise taxes), and the IRS would still have to process their returns and verify much of the same information regarding their filing status, number of children, and income. In 1998, about 69 per cent of EITC claimants will be required to file a tax return because they have an individual income tax liability (before the EITC), owe special taxes, have self-employment income in excess of $400, or their gross income will exceed the filing threshold. In addition, over 25 per cent of EITC claimants will file a tax return in order to obtain a refund for overwithheld taxes paid throughout the year.

  Because most EITC claimants would be filing a tax return even if the credit did not exist, the direct budgetary costs of administering the EITC are significantly lower than if the credit were provided through another means. The IRS cannot easily disentangle the costs of administering one line on the Form 1040 from other lines on the tax return, and we thus do not have estimates of the costs of administering this particular tax provision through the tax system. We can safely say, however, that the costs are lower than those associated with certain government expenditure programs. For example, in FY 1995, the food stamp program cost $3.7 billion to administer, while AFDC administrative costs were an additional $3.5 billion - nearly 14 per cent of the combined costs of these two programs. For these administrative costs, the AFDC program served, on average, about 4.9 million families in a given month, while over 10 million households received food stamps. By way of comparison, the entire IRS budget in FY 1995 was $7.6 billion, and the IRS served over 116 million individual taxpayers and 15 million corporations.

  Taxpayers also benefit from obtaining the EITC through the tax system. Many low-income workers learn about the EITC when they file a tax return to obtain a refund. By claiming the credit on tax returns, EITC claimants do not have to take time off from work to apply for the credit at a government office.

  Not surprisingly, then, participation in the EITC tends to be higher than many other assistance programs targeted to low-income families. In my research prior to joining Treasury, I found that 80 to 86 per cent of those eligible received the credit in 1990. This high participation rate is striking when compared to the AFDC participation rate of 62 to 72 per cent and the food stamp participation rate of 54 to 66 per cent. International comparisons also confirm this finding. The United Kingdom has an EITC-like program called the Family Credit. It is administered through the transfer system and directed toward families with children. Official estimates place the participation rate of the Family Credit at around 50 per cent. Thus, both compared to cash and in-kind transfers in the United States and comparable work-related benefits in the United Kingdom, the EITC is much better at reaching those who are eligible for the credit.

  Notwithstanding these benefits, there are costs associated with operating the EITC, as with other tax provisions, through the tax system. A system based largely on self-assessment will have lower administrative costs than a more bureaucratic approach, but it will also lead to higher noncompliance. Many of us were very concerned when EITC compliance data, from the 1980's, first became available. The Taxpayer Compliance Measurement Program (TCMP), last conducted in 1988, showed that 35.4 per cent of the EITC claimed ($2 billion) exceeded the amounts to which taxpayers were eligible.

  But the same TCMP also places the problems of the EITC in perspective. Last April, the IRS released a study, based on the 1988 TCMP, showing that the gross individual income tax gap in 1992 was between $93.2 and $95.2 billion. The IRS estimated that the total "true" individual income tax liability was between $550.2 and $552.3 billion for tax year 1992. Over 40 per cent ($39.1 to $39.9 billion) of the gross tax gap for 1992 was attributable to the underreporting of business income (including self-employment income, partnership income and rents and royalties). About 20 per cent ($18.1 to $18.7 billion) of the gross tax gap was due to the underreporting of non-business income. Over 14 per cent ($13.5 to $13.8 billion) of the gross tax gap was due to persons who failed to file tax returns. These problems exceed any noncompliance problems associated with the EITC.

  Nonetheless, the Administration and Congress have recognized that the EITC can best meet its goals - of making work pay and lifting families out of poverty - by ensuring that only those who are eligible and deserving receive the credit. Congress took a first step in this direction during the consideration of OBRA 1990, when data from the 1985 TCMP became available. The TCMP data suggested that EITC errors were linked to complicated and unverifiable support and household maintenance tests. OBRA 1990 replaced the support and household maintenance rules for EITC eligibility with simpler age, residency, and relationship tests, lowered the age requirement for reporting a taxpayer identification number for EITC qualifying children, and created a separate schedule to claim the EITC.

  This Administration, with the support of Congress, has taken 17 additional legislative and administrative actions to further improve the targeting and operations of the credit. First, Congress has enacted stricter reporting requirements proposed by the Clinton Administration, and the IRS has tightened enforcement of these requirements. Since 1995, the IRS has transcribed the social security numbers of all EITC qualifying children and most dependents, and it has intensified its examination of returns with missing social security numbers. The Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (the welfare reform act) contains a Clinton Administration proposal which will enable the IRS to use the simpler and more cost-efficient mathematical error procedures to deny both the EITC and dependent exemptions to taxpayers who fail to provide valid social security numbers. As a consequence of the Uruguay Round Agreement Act of 1994, taxpayers will also be required to provide social security numbers for all dependents and EITC qualifying children without regard to their age on their 1997 tax returns.

  Other reporting requirements have also been strengthened. The Uruguay Round Agreement requires the Department of Defense to report to both the IRS and military personnel nontaxable earned income used in the computation of the EITC. The 1996 welfare reform act also authorizes the IRS to treat the omission of self-employment taxes as a mathematical error, if the taxpayer claims eligibility for the EITC on the basis of self-employment income.

  The IRS, with the support of Congress, has also intensified scrutiny of "questionable" EITC claims and preparers. For the last several years, the IRS has conducted studies of EITC compliance and has used this information to better identify questionable returns. In addition, the IRS increased scrutiny of electronic return originators (EROs), instituted fingerprint and credit checks on certain new ERO applicants, and eliminated the direct deposit indicator.

  Finally, the Administration has consistently supported provisions that would simplify the EITC, opposed provisions that would add significant complexity to the EITC, and has striven to ensure that EITC reforms can be administered. In 1993, the Administration proposed the repeal of two supplemental credits (for children under the age of one and for the purchase of health insurance for qualifying children), arguing that the IRS could not enforce the eligibility criteria for them, and these supplemental credits were subsequently repealed. In 1995, the Administration opposed, on administrative grounds, proposals to base EITC eligibility on child support payments and hours of work. The Administration's proposal to deny the EITC to undocumented workers, included in the welfare reform Act, was also designed in a manner which could be administered by the IRS.

ANALYSIS OF EITC COMPLIANCE STUDY FOR TAX YEAR 1994

  The combined effects of these efforts cannot be fully measured at this time, since several key steps did not take effect until the 1997 filing season and another step - the requirement that all children, regardless of their age, have a social security number - will not be fully implemented until the 1998 filing season. Today's hearing, nonetheless, has been called in response to the recent release of new IRS data on EITC noncompliance for tax year 1994.

  The Criminal Investigation (CI) Division of the IRS conducted this study of compliance among 2,046 taxpayers who claimed the EITC on tax returns filed and accepted by the IRS between January 15 and April 21, 1995. CI Special Agents visited a random sample of EITC claimants, shortly after they filed their paper or electronic tax returns. Taxpayers (and often their employers, tax return preparers, family members, and neighbours) were interviewed at length and asked to produce verification that they met the EITC eligibility criteria. While the Special Agents made initial judgments about the legitimacy of the EITC claim, these judgments were reviewed - and sometimes changed - in subsequent review by Examination staff who had access to other sources of independent information (such as the Forms W-2 and 1099 sent by employers and other payers).

  The study found that of the $17.2 billion claimed in EITC between January and April 1995, $4.4 billion, or 25.8 per cent of total EITC claimed, exceeded the amount to which taxpayers were eligible. The overclaim rate among EITC claimants was slightly higher among paper filers (26.1 per cent) than for electronic returns accepted by the IRS (25.3 per cent). Noncompliance was found to be much higher among filers who claim EITC qualifying children than for those EITC claimants without qualifying children. Among those who claimed EITC qualifying children, the overclaim rate was 26.1 per cent, while the overclaim rate was 15.7 per cent for those who did not reside with a qualifying child. IRS enforcement practices, in place during the 1995 filing season, reduced the estimated net overclaim rate from 25.8 per cent to 23.5 per cent If the IRS had been able to treat a taxpayer's failure to provide valid social security numbers for EITC qualifying children over the age of one as a mathematical error on 1994 tax returns, the net overclaim rate would have been reduced further, to an estimated 20.7 per cent.

  While EITC noncompliance remains at unacceptably high levels, the study's results do show significant improvement since the late 1980s, the last time that the IRS examined a comparable group of taxpayers as part of the TCMP. The improvement in EITC compliance since 1988 reflects the implementation of many, but not all, of the steps described earlier.

  To better understand the remaining sources of noncompliance, we have conducted an analysis of the data. We have found that the most common EITC error is caused by taxpayers claiming qualifying children who do not reside with them for over half the year. Among taxpayers with children, such errors account for about 39 per cent of overclaimed EITC amounts. Under current law, taxpayers are required to reside with their qualifying children for at least six months or a full year, depending on the relationship of the child. Taxpayers fail the residency test for many different types of reasons. For example, divorced parents who share the custody of their children might both claim the EITC because they both feel the child lived with them for over half the year. At the other extreme, a taxpayer may claim a child with whom he or she has never resided.

  A second common error is due to misreporting of filing status among married taxpayers. Filing status errors account for about 31 per cent of overclaimed EITC amounts among taxpayers with children.[38] Sometimes, separated couples do not understand that they must still file as married persons if they have not yet obtained a legal separation. In other cases, married couples, who are still living together, do not file either a joint return or a "married filing separate" return.

  The third most common error results from complicated living arrangements. In such situations, a child lives with more than one adult who appears qualified to claim him or her for EITC purposes. However, about 18 per cent of overclaimed EITC amounts result when, in such households, the caregiver with the lower AGI claims the child. In some cases (although it is difficult to quantify), the other caregiver was, in fact, qualified to claim the EITC but did not. The study does not account for the offsetting errors which occur because the taxpayer's relative, with the higher AGI, did not claim the EITC when he or she was eligible.

  Even among EITC claimants without qualifying children, many errors are caused by the misreporting of family structure. Among these taxpayers, about 40 per cent of overclaims are attributable to the misreporting of filing status among married taxpayers. However, most errors among EITC claimants without qualifying children are due to the misreporting of income.

  While we can identify the sources of EITC errors in this study, we do not know from the study the extent to which the EITC, itself, is the root cause of the noncompliance on the part of the taxpayers. By misreporting filing status, child dependents, and income, taxpayers may be able to reduce their tax liability through other provisions in addition to the EITC. Because this study focused only on EITC claimants, it does not isolate the effect of the EITC on noncompliance, or the extent to which higher income taxpayers are benefiting from misreporting their income or family circumstances.

  The study does provide evidence that the refundable nature of the credit does not induce ineligible individuals to enter the tax system simply to claim the credit. As I have discussed, 95 per cent of EITC claimants have a reason other than the EITC to file a return. The overclaim rate among those with a positive pre-EITC tax liability is nearly three times larger than the rate among those who did not have a tax liability. The data thus suggest that noncompliant EITC claimants do not enter the tax system merely to claim the credit.

  While the results of this study are not fully applicable to the current EITC, the study does point to the need for new approaches. Many types of EITC errors are difficult to detect with the current IRS enforcement tools, such as matching of information reports and Social Security Administration records to tax returns. Our proposals are designed to provide the IRS with new tools to identify erroneous EITC claims while minimising additional administrative costs to the Federal government.

LEGISLATIVE AND ADMINISTRATIVE PROPOSALS

  The Treasury Department's eight-point plan contains six legislative proposals and two administrative actions. These proposals will help reduce EITC errors by increasing IRS's ability to detect errors before EITC refunds are paid out, by imposing new, more effective penalties on EITC claimants, and by reducing the risk of unintentional errors by law-abiding taxpayers.

PROPOSALS TO IMPROVE THE FLOW OF INFORMATION PRIOR TO RELEASE OF EITC CLAIMS

  Due diligence requirements for preparers - About half of earned income tax credit (EITC) claimants use a paid preparer to complete their income tax returns. As a consequence, tax preparers can play a key role in helping working families file accurate tax returns. While there is little significant difference among returns prepared by the taxpayer and those prepared by a paid preparer, the error rate does differ depending on the type of preparer consulted by the taxpayer. Noncompliance was much lower among taxpayers who went to a preparer who was either a certified public accountant, lawyer, enrolled agent, or a representative of one of the large nationally-recognised organisations. It was higher among those who sought other types of preparers.

  Under our proposal, the responsibilities of paid preparers, with respect to potential EITC claimants, would be clarified. Preparers who do not fulfil certain due diligence requirements would be subject to cash penalties ranging from $50 to the full amount of an EITC overclaim. The proposed penalties would be in addition to the penalties imposed on preparers and taxpayers under current law. The proposal would be effective for taxable years beginning after 31 December 1997.

  Recertification - When questions arise about EITC claims, the IRS generally must follow deficiency procedures to determine the accuracy of the taxpayer's return. While deficiency procedures protect taxpayers' rights, they can be time-consuming and relatively expensive when compared to the amount of tax at issue.

  Under the proposal, a taxpayer who has been denied the EITC as a result of deficiency procedures would be ineligible to claim the credit in subsequent years unless he or she provides evidence of his or her eligibility for the credit. To demonstrate current eligibility, the taxpayer would be required to meet evidentiary requirements established by the Secretary of the Treasury. Failure to provide this information when claiming the EITC would be treated as a mathematical or clerical error. If a taxpayer is recertified as eligible for the credit, he or she would not be required to provide this information in the future unless the IRS again denies the EITC as a result of a deficiency procedure. Ineligibility for the EITC under the proposal would be subject to review by the courts. The proposal would be effective for taxable years beginning after 31 December 1997.

  Demonstration Projects - The Treasury Department is seeking legislation permitting it to select four states to experiment with alternative ways of providing the EITC throughout the year. Under the proposal, the four states could provide advance payments of the EITC to wage earners through state agencies rather than employers for a three year period. States would be required to verify eligibility for the EITC before paying out the credit. Effects on advance payment participation and compliance would be studied by Treasury. Applications would be submitted by the states to the Treasury Department during 1998 for demonstration projects to begin in January, 1999.

  Earmarking of IRS Resources - Using information from the EITC compliance studies and other ongoing pilot projects, the IRS will continue to develop and use profiles of potentially erroneous EITC claimants. These profiles will be used to identify questionable EITC claims during the 1998 filing season. The IRS will expand the number of questionable EITC claims that it investigates during the 1998 filing season. Refunds associated with these claims will be delayed until the investigation is complete. Out of its current appropriations request, the IRS is earmarking 550 full time equivalent staff persons for this intensified effort during the 1998 filing season.

INCREASING THE PENALTIES FOR INTENTIONAL NONCOMPLIANCE

  New Penalties for Intentional and Fraudulent Errors - Existing civil penalties have a limited deterrence effect against ineligible taxpayers repeatedly claiming the EITC. Denying subsequent eligibility to claim the EITC to taxpayers who have recklessly, intentionally, or fraudulently claimed the EITC in the past should help ensure that only those who are eligible for the credit receive it.

  Under the proposal, any person who fraudulently claims the EITC would be ineligible to claim the EITC for a subsequent period of ten years. In addition, any person who erroneously claims the credit and such error is due to the reckless or intentional disregard of rules or regulations would be denied eligibility for the EITC for two subsequent years. The sanction under the proposal would be in addition to civil and criminal penalties imposed under current law. In addition, the sanction would be subject to review by the courts. The proposal would be effective for taxable years beginning after 31 December 1997.

  Continuing Levy - The IRS does not generally find it cost-effective to recoup overpayments of the earned income tax credit (EITC) or impose monetary penalties on noncompliant claimants. To some extent, these efforts are hindered by the exemption from levy of certain types of income prevalent among EITC claimants. By removing these exemptions, this proposal would make it more likely that the IRS would recapture overpayments.

  In our FY 1998 budget, the Administration proposed that certain exemptions be partially lifted from the levy. Under the budget proposal, Federal workers' compensation payments, annuity or pension payments under the Railroad Retirement Act, and benefits under the Railroad Unemployment Insurance Act would no longer be fully exempted from levy. The proposal would change the exempt amount of Federal wages, salaries, and other income to a flat 85 per cent exemption. The proposal would provide for "continuous" levy on non-means tested, recurring Federal payments.

  Under the EITC initiative, unemployment benefits and means-tested public assistance would no longer be fully exempted from levy for any purpose. Up to 15 per cent of these benefits would be subject to levy. The proposal would also provide for the option of a "continuous" levy on these payments. Treasury would work with affected Departments and state agencies to design the mechanisms appropriate for each program. If necessary, conforming changes would be made to the laws and regulations governing public assistance to ensure that there would not be offsetting changes in these benefits to compensate for the levy. The proposal would apply to levies issued after 31 December 1997.

  As under current law, taxpayers would be allowed to apply for relief from a levy if they can demonstrate that they are suffering significant hardship as a consequence.

REDUCE UNINTENTIONAL ERRORS

  Simplification of Foster Child Rule - Under current law, a taxpayer is eligible to claim the earned income tax credit (EITC) if he or she resides with a son, daughter, or grandchild for over half the year. EITC qualifying children also include individuals who reside with taxpayers for a full year and for whom the taxpayers "care for as the taxpayers' own children". All EITC qualifying children (including foster children) must either be under the age of 19 (24 if a full-time student) or permanently and totally disabled.

  The foster child rule is confusing to both taxpayers and the IRS. Clarifying the definition would eliminate unintentional errors by taxpayers and provide better guidance to the IRS. In addition, the definition of a foster child for EITC purposes would be conformed to the dependency exemption definition proposed as part of the Administration's simplification package.

  Under the proposal, a foster child would be defined as a child who (i) is under the age of 19 (24 if a full-time student), (ii) is cared for by the taxpayer as if he or she were the taxpayer's own child, and (iii) either is the taxpayer's niece, nephew, or sibling or was placed in the taxpayer's home by an agency of a state or one of its political subdivisions or a tax-exempt child placement agency licensed by a state. The proposal would be effective for taxable years beginning after 31 December 1997.

  Improve Access to Taxpayer Assistance - In 1996, 1.9 million low-income taxpayers received assistance preparing their tax returns from over 47,000 volunteers in IRS-sponsored VITA (Volunteer Income Tax Assistance) facilities. The IRS provides training materials and tax forms to 8,300 sites. The IRS also provides software for electronic filing and lends computer hardware to selected sites. These VITA efforts will be continued and strengthened as part of the Administration's commitment to volunteerism. The Treasury Department is contacting businesses and tax professional organisations to make sure that they are aware of the need for VITA volunteers, computers, facility sites, and outreach assistance. By improving access to free taxpayer assistance and electronic filing, these efforts will help reduce the risk of unintentional errors.

  This concludes my remarks. We look forward to working with you toward the enactment of these provisions. Thank you once again for providing me with the opportunity to testify. I would be pleased to answer any questions that the Committee may have.

8 May 1997