The Earned Income Tax Credit (EITC) is the nation’s second largest means-tested cash welfare program. Its major function is to provide “refundable” tax credits to low-income individuals. The Additional Child Tax Credit (ACTC) is a second refundable tax credit, available only to families with children. Most families with children that receive the EITC also receive the ACTC.

While the EITC plays an important role in the means-tested welfare safety net, it is rife with problems that include fraud and erroneous payments due to false reports of earnings and false residence claims, benefits intended for working parents going to nonparents, very high multi-tier benefits, and discrimination against married couples. The first three problems are also prevalent in the ACTC program.

To help restore the integrity and mission of the EITC and ACTC programs, the following reforms should be implemented:

• Require the IRS to fully verify reported income before any refundable EITC payment is made;
• Allow only those with formal legal custody of a child to claim the EITC and ACTC for that child;
• Prohibit families with dependent children from receiving EITC and ACTC benefits if they are also receiving public housing assistance;
• Eliminate EITC marriage penalties by allowing only the parent with sole legal custody to be able to claim a child (among unmarried couples); eliminating EITC benefits for adults without dependents; and allowing married couples filing jointly the option of claiming the EITC on the basis of their joint earnings or on the basis of the earnings of either spouse; and
• Reduce marriage penalties throughout the rest of the means-tested welfare system by increasing EITC benefits for married couples with children.

A sound welfare system should encourage self-sufficiency through work and marriage. The EITC and ACTC are intended to support work and marriage but currently have substantial flaws that stand in the way of these aims. The reforms outlined in this paper would promote these principles in these large means-tested cash welfare programs.

The EITC and ACTC Programs

Although the EITC in some cases reduces federal income taxes owed, its major function is to provide “refundable” tax credits to low-income individuals. A “refundable tax credit” is simply a cash welfare grant to individuals who have no federal income tax liability.

About 13 percent of the total expense of the EITC goes to tax reduction, while 87 percent goes to refundable cash credits.[1] In 2013, 20.1 million families with children and 5.1 million individuals without children received refundable EITC payments at a cost of $59.1 billion.[2] The 20.1 million families with children receiving the EITC represent 43 percent of all tax filers with children.[3] Historically, around two-thirds of the families with children receiving the refundable EITC are single parents, and one-third are married couples.[4] The average value of a refundable EITC payment to a family with children was $2,919.[5]

The Additional Child Tax Credit is a second refundable tax credit that in most cases is added on top of the EITC. It is available only to families with children. In 2013, there were 19.9 million tax filers who received the refundable ACTC at a cost to taxpayers of $26.7 billion.[6] Most families with children who receive the EITC are also eligible for the ACTC. The cost of the refundable EITC and ACTC combined was $85.8 billion in 2013. The maximum annual benefit for the two credits combined for a family with two children was $7,548 per year in 2015.

While the earned income tax credit is an important tool in the welfare safety net, it is plagued by a number of significant problems.

1. Fraud and Erroneous Payments. Erroneous overclaims are at least one-quarter of the $59 billion in annual EITC spending. Some 43 percent to 50 percent of EITC tax returns claim illegitimate excess benefits. The most important causes of erroneous overpayments are false reports of earnings and false residence claims by adult claimants.
2. EITC Benefits to Nonparents. The EITC is designed to encourage increased work by parents, but the EITC law permits persons other than the parent to receive benefits on behalf of children. Grandparents, aunts, uncles, and older siblings and stepsiblings can often claim EITC cash bonuses for children. This option leads to “benefit shopping,” arbitrarily assigning children for EITC purposes to relatives whose earnings will elicit the highest EITC payment. Benefit shopping increases costs to taxpayers while undermining the core principle of promoting parental work; it is also a major factor in residence fraud.
3. Very High Multi-tier Benefits. Many families receiving the EITC also receive benefits from other welfare programs such as food stamps, Medicaid, and subsidized housing. When these benefits are piggy-backed on top of each other, the aggregate benefits received can be quite high; this is particularly the case with families that receive both the EITC and subsidized housing.
4. Discrimination Against Married Couples. In most cases, the EITC benefits received by unmarried parents who cohabit are significantly higher than those received by similar couples who are married. A government policy that explicitly rewards parents for remaining unmarried and cohabiting while financially punishing those who do marry is unwise.

The first three problems also appear in the ACTC program. Overall, the four problems are inter-related. Policies that will reduce EITC fraud and excessive multi-tier benefits will also reduce the anti-marriage penalties built into the welfare system.

 

How the EITC and ACTC Work

In the typical means-tested welfare program, maximum benefits are given when the recipient has zero earned income; benefits are phased down as earnings rise. The EITC deliberately differs from this pattern. As Chart 1 shows, EITC benefits are shaped like a trapezoid. As an incentive to work, EITC benefits increase as earnings increase at least through low-income levels.

The EITC has different benefit maximums and phase-in and phase-out rates depending on the number of children and whether the tax filer is single or married. There is a small EITC with a maximum value of $503 for childless adults. In an effort to slightly mitigate the marriage penalties built into the credit, the phase-down rate for married couples with children begins at a higher earnings level than it does for single parents.[7] (The parameters of EITC benefits in 2015 are shown in Budgetary Appendix Table 1.)

In contrast to other welfare programs, the EITC is designed to give incentives to work, at least at lower earnings levels. Below the $13,870 level, benefits rise as earnings increase. For each added $1.00 of earnings, the recipient typically receives an extra $0.40 in benefits. Although the pro-work impact of the EITC is often exaggerated, it is true that the EITC and its companion ACTC are the only welfare programs that require an individual to work to receive aid.[8] Other programs such as food stamps and housing begin by assisting those who do not work and then allow recipients who volunteer to work to retain partial benefits. Even Temporary Assistance for Needy Families (TANF), which is supposed to be work-based, has half of its able-bodied caseload idle.[9] Ideally, the EITC should operate more as a wage supplement than as a conventional welfare program. Although it has deep flaws, the EITC is more supportive of low-income work than any other welfare program.

 

The Additional Child Tax Credit is the refundable component of the Child Tax Credit (CTC). Like the CTC, the ACTC has a maximum value of $1,000 per child; there are no limits on the number of dependent children a family can claim. In the low-income range, the ACTC has pro-employment features that, like the EITC, link payments to employment. The ACTC has a bottom refundability threshold of $3,000; families with earnings below that level receive no benefits. As yearly earnings rise above $3,000, the credit equals 15 percent of any earnings above $3,000 up to a maximum credit of $1,000 per child.[10] For a family with two children, the ACTC reaches a maximum value of $1,000 per child when earnings reach $16,334.

As earnings rise above $20,000, a family’s pre-credit federal income tax liability begins to increase. The value of the nonrefundable CTC begins to increase incrementally to offset this tax liability; the ACTC is decreased in direct proportion to the CTC increase. (The sum of the ACTC and CTC can never exceed $1,000 per child.) The ACTC typically reaches zero value when family incomes are around $40,000; at the point the ACTC reaches zero, the CTC will equal $1,000 per child.

Chart 2 shows the combined EITC, ACTC, and CTC benefits received by a single parent with two children at various income levels.[11] The maximum combined value of the EITC and ACTC for a family with two children is $7,548. This maximum occurs in the $16,334 to $18,110 income range.

The chart clearly shows that as family income rises above $20,000, the refundable ACTC is incrementally replaced by the nonrefundable CTC, which reduces the taxes the family would otherwise owe. (As noted, above $16,334, the sum of the ACTC and CTC is always $2,000.) The ACTC reaches zero value around $38,000. The EITC reaches zero value around $44,453.

Erroneous and Fraudulent EITC Claims

Erroneous payments are pervasive in the EITC. An Internal Revenue Service (IRS) compliance estimate audit conducted in 2006–2008 found that between 43 percent and 50 percent of tax returns claiming the EITC involved erroneous overclaims.[12] Half of EITC errors appear to be due to fraud and half to unintentional errors.[13] The 2006–2008 EITC compliance audit consisted of careful review of a representative sample of 7,635 tax returns that claimed refundable EITC benefits. In most cases, the audit included a face-to-face interview between an auditor and the tax filer. The audit is a slow, careful process. The IRS states:

Because the NPR audits are oriented toward generating high quality, accurate data, examiners are trained to make every accommodation to meet with taxpayers to educate them about the necessary documentation for substantiating EITC eligibility, and to give them sufficient opportunity to obtain and supply the necessary information.[14]

Whenever necessary, multiple meetings with the tax filers are held to allow the filer sufficient opportunity to provide necessary evidence.

One difficulty in the audit process is that 14 percent of the EITC tax filers selected for auditing failed to respond to the IRS audit request despite the fact that this nonresponse, in most cases, would lead to the EITC benefit’s being revoked.[15] To address this nonresponse rate, the IRS makes two separate error estimates: The high estimate assumes that all nonresponding tax filers who fail to participate in the audit have erroneous or fraudulent claims. The low estimate assumes that the error rate among the nonresponders is the same as the rate among responders.

The low estimates found that 43 percent of all claims were erroneous and that erroneous overclaims accounted for 28.5 percent of the dollar value of all EITC claims.[16] The high estimate found that 50 percent of all tax returns claiming the EITC during the 2006–2008 period contained erroneous overclaims and that erroneous overclaims accounted for 39.1 percent of the dollar value of all EITC claims made during the period.[17]

For purposes of analysis, this paper will use the low estimates: 43 percent of EITC returns involve an erroneous overclaim, and the dollar value of overclaims equals 28.5 percent of all EITC claims. In 2013, that would mean there were an estimated 12.4 million tax returns with erroneous EITC overclaims. The dollar value of EITC overclaims in 2015 would have been $18.7 billion.[18] (For a detailed explanation of the calculations used in this paper, see the Budgetary Appendix.)

The overclaims found in the IRS audit were not minor filing errors. The overwhelming majority of individuals making overclaims were not eligible for the credit at all. The IRS audit “found that between 79% and 85% of EITC dollars claimed incorrectly were claimed by tax filers ineligible for the credit.”[19]

Improper Payment Audits. In addition to compliance audits, the IRS conducts separate annual “improper payment” audits.[20] According to the IRS, improper payments of the EITC were $15.6 billion, or roughly 23.8 percent of all EITC payments in 2015.[21] Improper EITC payments cost the taxpayer $124 billion to $148 billion between 2003 and 2013.[22]

The EITC overclaim and improper payment figures are quite similar; the main difference between them is that overclaim figures ignore current IRS enforcement actions. Improper payments, on the other hand, reflect the impact of these enforcement actions. Improper payments equal overclaims minus the amount saved by current enforcement.

The initial portion of this paper will use overclaim data from the 2006–2008 compliance audit, because that audit provides much more detailed information on sources of error. The final budget estimates in the paper, however, use current improper payment estimates, which means that the budget savings figures are net of current enforcement activities.

 

Similarly high improper payment rates appear to exist in the refundable Additional Child Tax Credit program. EITC and ACTC claims come from the same income tax returns, and eligibility rules for the two credits are similar; therefore, there should be a large overlap in errors between the two programs. The 2015 report of the Treasury Inspector General for Tax Administration (TIGTA), states that:

[O]ur review of the IRS’s own enforcement data indicates that the ACTC improper payment rate is similar to that of the EITC. TIGTA estimates that the ACTC improper payment rate for Fiscal Year 2013 is between 25.2 percent and 30.5 percent, with potential ACTC improper payments totaling between $5.9 billion and $7.1 billion.[23]

Sources of Erroneous EITC Claims

There were 28.8 million tax returns with EITC refundable payments or tax reductions in 2013.[24] Based on the results of the 2006–2008 compliance audit, we can estimate that at least 43 percent (or 12.4 million returns) contained an overclaim in 2013.

Table 1 shows EITC overclaims by the type of error according to the low estimates from the IRS 2006–2008 compliance audit.[25] The first column shows, as noted, that 43 percent of all EITC claims had an overclaim. Some 24.9 percent had an overclaim involving income misreporting. Another 9 percent had a qualifying child (QC) error; in most of these cases, a tax filer claimed the EITC for a child but was not in fact eligible to claim that credit. Some 3.9 percent of EITC returns had both an income misreporting and a qualifying child error. Around 3.6 percent had “filing status” errors in which married couples filed separate returns to evade the EITC marriage penalties. A final 1.6 percent of EITC returns had other errors.

The second column shows the dollar value of the overclaims. Again using the low estimate, some 28.5 percent of dollars claimed on EITC returns were overclaimed benefits. Ten percent of all EITC costs were erroneous overclaims due to income misreporting; 10.8 percent involved qualifying child errors; and 4.3 percent involved both income misreporting and qualifying child errors. Overclaims due to filing status errors represented 2.7 percent of all EITC costs.[26] The first and second columns differ because qualifying child errors are generally more costly on a per-case basis than are income misreporting cases.

Erroneous Payments Due to Income Misreporting.27 The EITC differs favorably from other means-tested aid programs: Individuals must report earned income to receive cash aid. Thus, in certain income ranges, the EITC can encourage apparent work effort. However, millions of individuals each year report nonexistent or nonverified income to obtain EITC cash bonuses. Many of these individuals invent fictitious earnings (or less frequently underreport real earnings) to maximize their EITC welfare payments.

According to IRS compliance audit data, at least 29 percent of all EITC claims are based on false or erroneous income claims.[28] In the 2006–2008 audit, 50 percent of the total EITC overclaim dollars involved false income reporting.[29] If the income misreporting ratios from the 2006–2008 audit are applied to 2013 EITC payments, at least 8.3 million households would have engaged in false income reporting in that year. In 2015, income misreporting was involved in roughly $9.4 billion in erroneous overclaimed EITC credits.[30]

Erroneous Payments Due to Qualifying Child Errors and False Residence Claims. Qualifying child (QC) errors occur when the adult claimant has no right to claim a child as a dependent. Some 30 percent of erroneous EITC tax returns have QC errors,[31] and returns with these errors account for around half of all erroneous overclaim dollars. There are three main types of qualifying child errors: errors concerning the age of the child, errors concerning the relationship of the adult to the child, and false residency claims by the adult claimant.

The most important of these categories is false residency claims. An adult EITC claimant is not required to financially support the child he claims.[32] However, in order to claim EITC benefits for a child, the adult claimant must reside with the child for at least half of the year.[33] A false residence claim occurs when the tax filer did not actually reside with the child either for the specified period or at all.

False residency claims account for three fourths of qualifying child errors.[34] Roughly one in 10 total EITC claims, or some 2.8 million claims per year, are based on false claims of residence by absent parents or relatives.[35] On a per-case basis, false residence overclaim cases are more expensive than false income cases. Returns with false residence claims account for roughly $7.4 billion in erroneous EITC claims each year.[36]

Other Qualified Child Errors. Some 20 percent of qualifying child errors occur because the claimant does not have the legal relationship to the dependent child that is claimed.[37] These errors represent about 3 percent of all EITC returns, or 750,000 returns in 2013. These errors cost the taxpayer around $2.0 billion per year.[38] Another common error is over-age children; about 1.5 percent of all children claimed on EITC returns are ineligible because they are over the age limit.

Errors Due to Marriage Penalty Avoidance. The 2006–2008 audit found 1 million tax returns with “filing status” errors.[39] These involved married couples who erroneously filed two separate income tax returns as heads of households or as single persons in order to avoid the EITC marriage penalty. One survey of beliefs about welfare and marriage in low-income communities found that many low-income mothers understood that getting married could jeopardize their refundable tax credits but incorrectly believed that they could lawfully circumvent this problem by having both spouses file as separate heads of households.[40] This error caused about 9.2 percent of the dollar cost of all EITC overclaims net of overlapping errors.[41] In 2015, this cost would be around $1.7 billion per year.[42]

Chart 3 displays overclaim errors in a slightly different format.[43] Income misreporting errors (excluding returns with other errors) comprised 35 percent of total overclaim costs. Qualifying child errors including returns with other errors comprised 53 percent of total overclaim costs. QC errors can be further subdivided into residency errors at 40 percent of total costs and other QC errors at 13 percent of total costs. Filing status errors to avoid EITC marriage penalties were 9.2 percent of total overclaim costs, and other errors comprised 2.8 percent.

 

Why Does the EITC Appear to Have More Fraud than Other Welfare Programs?

Each year, the IRS releases EITC audits that show improper payments due to error and fraud equaling about 25 percent of total benefit costs. Other welfare programs such as food stamps and public housing do not appear to have similar problems. What causes the difference? There are three reasons.

First, the IRS audits the EITC with remarkable accuracy and thoroughness. No other welfare program is audited in this fashion.

Second, the EITC is saddled with arbitrary bureaucratic rules that make fraud inevitable. The most significant counterproductive rule is that the IRS is required to make tens of millions of very large cash payments each year before income can be corroborated. No other welfare program operates under this type of restriction. Fortunately, this and other fraud-prone rules for the EITC can easily be fixed.

Third (and most important), the design of the EITC makes fraud easier to detect. In most welfare programs, the lower the earnings, the higher the benefits. A recipient who wishes to fraudulently obtain higher benefits must conceal earnings. To do this, the recipient may work “off the books” or may have working individuals in the household whose presence and earnings are not reported to the welfare agency. By contrast, to obtain the EITC, the recipient must have earnings. The most common form of EITC fraud is to report earnings that do not exist. It is much easier for an audit to find that claimed earnings are fictional than it is to discover hidden off-the-books income or unreported household members.

This asymmetry in detection means that EITC audits historically reveal far higher error and fraud rates than other welfare programs. In fact, the food stamp and public housing programs make no serious efforts to detect hidden earnings or hidden workers in the household; this allows them to claim low but illusory error rates. In reality, public housing and food stamps are unlikely to have lower fraud rates than the EITC has; they simply have lower rates of fraud detection. Moreover, the policies proposed in this paper could eliminate most EITC/ACTC fraud and errors. A reformed EITC/ACTC system would almost certainly have lower de facto fraud and error rates than other welfare programs have.

Are EITC Errors Due to the Technical Complexity of the Law?

There is a common argument that the EITC’s high error rate is due to the EITC law’s complexity. While some aspects of EITC eligibility are complex, the major errors shown in Table 1 are straightforward and rarely the result of technical complexity. For example, income misreporting occurs in 58 percent of overclaims, but income reporting in most cases is comparatively simple. According to the IRS:

[The existing data] provide good reason to believe that the income misreporting errors [found in the audit]…generally do not stem from complexity created by the EITC…. [F]or 96 percent of EITC claimants the correct amount of earned income can be determined by combining at most 4 line items from the front of the Form 1040…. Thus, for most taxpayers, the steps required to calculate income concepts for the EITC are fairly simple.[44]

Similarly, 40 percent of erroneous overclaim dollars involve “residency errors.”[45] Again, the law is not complex. As the EITC instructions clearly explain, the adult claimant must reside with the child for more than half of the year; if the he does not, the claim is invalid.[46] The EITC instructions issued by the IRS explain this simple point clearly, yet more than one in 10 EITC filers violates the rule.

Some 2.6 percent of all EITC returns are erroneous because the adult does not have a legal relationship to the claimed child. Again, the rules are not complex and are clearly explained in the EITC instructions. The adult claiming the EITC can be any of the following: a parent, stepparent, foster parent, grandparent, aunt, uncle, older adult sibling, or older adult stepsibling; if he is none of these, the claim is invalid.[47]

Two-thirds of EITC claimants use a paid tax preparer,[48] but the error rates of self-prepared returns are actually slightly lower than the error rates for returns from a paid preparer.[49] This underscores the point that knowledge of EITC technicalities does not seem to be a principal factor in the high error rate.

The EITC code admittedly contains abstruse legal complexities. Particularly complex are the “tiebreaker” rules that govern situations in which more than one parent or relative can claim the EITC for a given child. But most of the tiebreaker rules were eliminated from the law during the 2006–2008 period covering the data in Table 1. Tiebreaker errors were responsible for only 0.8 percent of the overclaims in the 2006–2008 compliance audit.[50] At least 42 percent of all EITC claims in that audit had overclaims without any tiebreaker complexity. Finally, the reforms presented in this paper would greatly simplify EITC eligibility, removing tiebreaker issues and other complexities from the law.

Preventing Erroneous Payments Due to Income Misreporting

As noted, the EITC differs favorably from other means-tested aid programs: Individuals must report earned income to receive cash aid. Thus, in certain income ranges, the EITC can encourage apparent work effort. However, millions of individuals each year report nonexistent, nonverified, or inaccurate earnings to obtain EITC cash bonuses. Many of these individuals invent fictitious earnings (or less frequently underreport real earnings) to maximize their EITC welfare payments.

According to the IRS, at least 29 percent of all tax returns claiming the EITC during the 2006–2008 period contained false or erroneous earnings claims.[51] The IRS audit process gives filers ample opportunity to demonstrate the validity of their income claims. The fact that they cannot do so indicates that the reported income was either fictitious or entirely off the books.[52] If the income misreporting ratios from the 2006–2008 audit are applied to 2013 EITC rolls, an estimated 8.3 million households would have engaged in erroneous income reporting in that year. In 2015, income misreporting probably generated $9.4 billion in overclaimed refundable EITC credits.[53] (This includes claims with multiple errors.)

Overreporting of self-employment income is particularly prevalent; two-thirds of overclaims due to income misreporting involve erroneous claims of self-employment income. Analysis of EITC claimants with dependent children shows that false claims of self-employment income cluster dramatically around the first kink or plateau point in the EITC benefit schedule. These claims clearly involve individuals with low work levels claiming just enough alleged earnings to obtain the maximum EITC benefit ($5,548 for a family with two children).[54]

Tax returns claiming the EITC with false earnings reports are very likely also to have erroneous claims for the refundable ACTC.[55] Both the EITC and the ACTC are claimed on the same income tax return; the reported earnings used to claim the EITC are also used to claim the ACTC. If the earnings used to claim the EITC are false, the earnings used to claim the ACTC must also be false. This means that erroneous EITC payments are often matched by erroneous ACTC payments and erroneous tax reductions through the CTC. For example, an individual who used fictitious earnings to fraudulently claim the maximum EITC for two children ($5,548) would also receive around $2,000 in erroneous ACTC payments. At higher earnings levels, these individuals may claim a combination of the refundable ACTC and the nonrefundable CTC.

Each year, there are some 8.3 million tax returns with children that contain EITC overclaims based on false earnings data. These returns are likely to contain a related $7 billion in erroneous ACTC claims.[56]

Why False Earning Claims Are Permitted

The IRS has an effective capability to detect false income claims but typically scrutinizes claims only in a small number of cases. Audits occur long after the EITC cash payments have already been made. Current legal restrictions on the IRS require that it make EITC refunds prior to proper income verification. The Department of the Treasury Agency Financial Report for FY 2015 states:

The Internal Revenue Code (IRC) requires the IRS to process tax returns and pay any related tax refunds within 45 days of receipt of the tax return or the tax return due date, whichever is later. Beyond this 45 day period interest begins accruing on the refund amount, which must be paid along with the refund to the taxpayer. For the majority of EITC returns, which tend to be filed early in the filing season, the IRS does not receive information to verify income reporting accuracy or other data needed to validate these EITC claims at the time of filing or within a reasonable period thereafter in which a taxpayer would expect their refund.[57]

The Treasury Department’s Agency Financial Report for FY 2013 further states, “Income reported through information returns such as Forms W-2, Forms 1099, etc., which can be used for verification of some income, becomes available only after tax returns are processed.”[58] In other words, the information to stop the erroneous payments exists but is not used. The ensuing fraud and erroneous payments are not inherent in the program; they are the result of inept administration.

The IRS contends that it is legally required to send out the EITC and ACTC benefit checks before the corroborative information can be examined , but members of the House Ways and Means Committee recently sent a letter to the IRS contending that the IRS has legal authority to withhold EITC checks until corroborating income information is obtained through its “Do Not Pay (DNP) portal.”[59] The letter reveals a growing understanding that delaying the release of EITC checks until after corroborating income information has been obtained can prevent a substantial share of EITC improper payments.

Paradoxically, there is no logic to the current policy of making EITC payments in a hurried manner. Some welfare programs such as food stamps are designed so that they can quickly address temporary shortfalls in monthly income. The EITC, however, is not designed to address short-term drops in monthly income. The EITC is an annual subsidy based not on a family’s monthly income in February or March, but on its total income during the 12 months of the prior year. There is no evidence whatsoever that EITC recipients face greater financial pressures in the early spring compared to other times of the year. Indeed, it is often argued that poor families experience greater distress in summer when school-age children no longer receive free lunches and breakfasts at school.

Most EITC recipients use the program not to deal with short-term cash shortages, but as a system of long-term forced savings. The EITC benefit, when it arrives, is generally used for big-ticket purchases.[60] Since the EITC is not designed to meet short-term emergency need, there is no need to make payments early in the year. It is true that many families have become accustomed to receiving EITC checks in February and March. Delaying EITC payments until income claims were verified would cause some initial inconvenience, but the modified annual schedule would soon become customary.

Requiring the IRS to make EITC payments before corroborating evidence becomes available serves no rational policy purpose. It does, however, waste funds and undermine support for the program by promoting widespread fraud.

Recommended Policies to Reduce Erroneous Payments in the EITC and ACTC Through False Income Claims

EITC overclaims and overpayments due to false income reporting can be significantly reduced by requiring the IRS to fully verify reported income before any refundable EITC payment is made. Specifically, policymakers should change the law to:

• Require the IRS to fully verify income through a review of Form W-2, Form 1099, business licensing or registration, and relevant invoices.
• Require that the IRS not issue any refundable EITC or ACTC payment until all corroborating income information is received and checked.
• Require individuals claiming self-employment or small-business income to:
  1. Provide a Form 1099 documenting the income; or
  2. Be a registered or licensed small business and provide invoices of payments received including date of service and identifying contact information from customers; and
  3. Have paid self-employment taxes regularly during the year.
• Require the IRS to provide a summary denial of any EITC claim that does not provide the information listed above. The IRS should then issue a letter to the tax filer explaining the denial and offering an opportunity to respond to the ruling by providing additional information.
• Provide a $2,000 penalty for any tax filing involving an erroneous claim for a refundable tax credit that is based on substantial misreported income.

These changes would go a long way toward substantially reducing or eliminating overclaims and overpayments due to false income reporting. If implemented in 2015, these policies might have blocked or reduced $6.6 billion in erroneous EITC benefits and $4.5 billion in erroneous ACTC payments. The savings would be even larger if returns with multiple errors were included.[61]

Limiting Residency Fraud

As noted, roughly one in 10 EITC claims, or some 2.8 million claims per year, are based on false claims of residence by absent parents or relatives.[62] These false residence claims may result in as much as $7.4 billion in erroneous EITC payments each year. Many individuals who make false residence claims to receive the EITC will also receive the refundable ACTC for the erroneously claimed child; at higher earnings levels, they may benefit from the nonrefundable CTC. The cost of the erroneous ACTC claims may be $2.5 billion.

The EITC program permits married couples, single parents with legal custody, parents without legal custody, and stepparents, as well as adult relatives such as a child’s grandparent, aunt, uncle, and/ or older siblings and stepsiblings, to file for EITC benefits. The same rule applies to the refundable ACTC.[63] An adult claiming the EITC or ACTC is required to reside with the child for at least half of the year; there is no requirement that the adult claimant financially support the child.[64] Residency fraud occurs when the adult claiming the child did not reside with the child for half of the year or perhaps at all.

A demographic breakout of false residence claims is not available , but it is reasonable to assume that relatively few single biological parents with legal custody fail to reside with their children but still claim the EITC. Similarly, it is unlikely that large numbers of married couples claim the EITC when neither parent resides with the child. A reasonable assumption, therefore, is that most residency fraud involves claims by noncustodial unmarried parents (generally absent fathers) and adult relatives who do not actually reside with the child. One study of EITC recipients in Wisconsin found that among cases in which the EITC was claimed by someone other than the lawful custodial parent, about half were claims by the noncustodial parent and half were claimed by other unidentified individuals such as relatives.[65]

In such cases, an absent parent or nonresident relative can gain greater EITC and ACTC benefits by claiming a child than the resident parent or caregiver can. The resident parent or caregiver therefore permits the absent parent or nonresident relative to claim the child as a dependent on his or her income tax and receive the relevant EITC and ACTC payment. The EITC and ACTC payments are then presumably shared between the resident parent and the claimant. If there are more than two children in the family, the parent may seek to maximize EITC and ACTC benefits by claiming some of the children on his or her own tax return and allowing absent parents or relatives to claim the remaining children.[66]

EITC recipients making false residence claims generally have no children in their homes. One study found that one-tenth of all adults claiming the EITC for dependent children actually had no children in their households during the previous two years.[67] One-third of males filing as separate heads of households fell into this category.

With over 28 million EITC claims each year, it would be impossible to audit claims manually to verify residence claims. Moreover, many low-income individuals have fluid and changing residence patterns over a year. Verifying actual residences throughout the year would be a laborious and difficult task even with vast manual audits.

The answer is to reexamine the lax EITC eligibility rules that promote widespread fraud. For example, there appear to be nearly three relatives who obtain the EITC through residence fraud for every relative who gets the credit legally.[68] Most residency fraud can be prevented by straightforward changes in the eligibility standards. In the future, only parents with formal legal custody of a child should be permitted to claim the EITC and ACTC for that child. If no biological parent has legal custody, then an adoptive parent, legal guardian, or foster parent with court-assigned custody could claim the child.[69] Adult relatives, stepsiblings and noncustodial unmarried parents would no longer be eligible to receive the EITC or ACTC.[70]

These changes would not only eliminate most residence fraud, but also provide overall improvements in the program. Restricting the EITC and ACTC to parents rather than relatives would strengthen the incentives for parental work that are at the philosophical heart of the program. Barring unmarried, noncustodial parents from receiving the EITC would also reduce the marriage penalties imbedded in the EITC and the overall welfare state.

Net Savings from Preventing Residence Errors. If EITC payments based on false residence claims were barred, the custodial parent would still be eligible to receive the EITC and ACTC, although in most cases, the benefit would be lower. The “replacement rate” measures the ratio between the EITC benefit that would be received by the custodial parent and the benefit currently received by the nonresident claimant.

As noted, detailed research on EITC payments in Wisconsin found that among those cases in which the EITC was claimed by someone other than the lawful custodial parent, about half were claims by the noncustodial parent and half were by other unidentified individuals. The study found that the replacement rate for erroneous payments to nonresident parents (usually absent fathers) was 56 percent. The replacement rate for erroneous payments to relatives and other nonqualified individuals was 39 percent.[71]

Total EITC overpayments due to false residence claims are around $7.4 billion per year.[72] The net savings from eliminating false residence claims would equal the cost of such overpayments minus the cost of the replacement benefits. Assuming that the replacement rate is 50 percent, the net savings from eliminating $7.4 billion in EITC overpayments due to false residence claims would be half of that amount, or $3.7 billion. The ACTC would be around 50 percent of $2.5 billion, or $1.25 billion.[73]

Limiting Errors Due to the Lack of a Legal Link Between Filer and Child. Some 2.6 percent of all erroneous EITC returns occur because the claimant does not have a legal relationship to the dependent child who is claimed. Around 750,000 EITC tax returns have this type of error; these erroneous claims cost the taxpayers around $2 billion per year.[74]

Limiting EITC adult claimants to parents with legal custody, legal guardians, and formal foster parents would vastly simplify verification of adult/child relationships. Tax forms for filers seeking refundable benefits should require the filer to specify his exact legal relationship to the claimed child. The form should clearly specify that penalties may be assessed for false information. The IRS has access to a number of automated databases that can be used to verify the relationship of the child to the adult claimant. These include Kidlink which links the Social Security numbers of children to one or more of their parents; Numident, which provides data from birth certificates, including parents’ names; and the Federal Case Registry (FCR) of child support orders.[75] These databases should be accessed and the relationship of the claimant to the child should be verified before any EITC or refundable ACTC payment is issued.

Verifying the legal relationship between claimants and children before benefits are paid should eliminate some $2 billion in erroneous EITC overpayments and another $750 million in ACTC payments each year. With a replacement rate of 50 percent, net savings would by about $1.4 billion.

Limiting Age-of-Child Errors. To be claimed as a dependent for purposes of the EITC, a child must be under age 19 at the end of the calendar year. Two exceptions are provided: A child who is a full-time student for five months out of the year remains an eligible dependent up to age 24, and a permanently disabled dependent is not subject to age limits.

The 2006–2008 compliance audit found that 1.5 percent of all dependents claimed on EITC returns were over age and not subject to exemption.[76] (The audit process provides ample opportunity for the filer to demonstrate that the over-age dependent was actually disabled or in school.) In 2015, the cost of age-of-dependent errors was about $1 billion per year.[77] Filers who erroneously claim the EITC for over-age children may also erroneously claim the ACTC; potential ACTC overpayments could be $300 million per year.

Any dependent Social Security number and age should be automatically checked against the Social Security Administration’s database before any refundable EITC payment is made. Filers who wish to claim an exemption from the EITC age rules should be required to declare explicitly the nature of the exemption on the return and provide evidence of school attendance or disability with the original return. If the child is 19 or older, this evidence should be reviewed before EITC payments are made.

Multi-Tier Welfare Benefits

Most lower-income families receiving the refundable EITC and ACTC also receive benefits from other welfare programs. When these benefits are piggybacked on top of each other, the cost of the total benefit package can be quite large. For example, in 2015, a single mother with two school-aged children who worked full-time at the federal minimum wage through the year would receive $13,853 in post-tax earnings. In most cases,, this mother would also receive $5,548 in the refundable EITC, $1,800 in the refundable ACTC, $3,974 in food stamp benefits, and $1,269 in school lunch and school breakfast benefits. In most states, the whole family would also be eligible for Medicaid coverage, which would cost the taxpayer $10,005 per year.[78] The total value of post-tax earnings plus benefits would come to $36,448—almost twice the federal poverty level for a family of three.[79]

If Section 8 housing or other subsidized housing is added, the benefit stack becomes much higher. If the parent received a housing voucher for a two-bedroom apartment, the combined post-tax earnings and benefits would reach, on average, $43,395 per year.[80] As Chart 4 shows, if the parent received a voucher for a three-bedroom dwelling, combined earnings and benefits would reach $47,385 per year.[81] This computes to an effective hourly wage rate of $22.78 per hour.

 

The benefit package for a family with housing is overly generous. Excluding medical care, the per capita cash and benefits of the minimum-wage single mother with a three-bedroom housing voucher equal the per capita post-tax income of a married couple with two children earning $53,000 per year.

U.S. Department of Housing and Urban Development (HUD) rental assistance programs provide disproportionately expensive aid to a small portion of low-income families with children. Most low-income working families do not receive housing benefits. In 2015, there were approximately 9 million families with children that had non-welfare cash incomes below 125 percent of the federal poverty level (FPL),[82] but only 1.8 million families with children (or one-fifth of these low-income families) received HUD housing aid.[83] Housing aid programs generate a substantial problem of “horizontal inequity”: unequal treatment of individuals with the same incomes.

Moreover, HUD aid programs appear to discriminate heavily against married couples with children. Nine out of 10 families with children receiving housing benefits are headed by single parents.[84] Roughly one-quarter of poor and near-poor single mothers receive HUD rent subsidies, compared to only 6 percent of married couples with children at similar income levels.[85]

Clearly, multi-tier stacking of welfare benefits can result in very high aggregate benefit levels. Families with children receiving housing aid receive very high aggregate benefits both in an absolute sense and in relation to most other low-income working families. Excessive benefits should therefore be reduced by limiting the piggybacking of multiple benefits. Specifically, families with dependent children that receive subsidized housing benefits during a year should not be eligible for refundable EITC and ACTC benefits during the same year.

As noted, 1.8 million families with children receive HUD housing subsidies from public housing, Section 8, or other HUD programs. Some 64.8 percent of families with children under 18 years old who report receiving housing subsidies also receive refundable credits, with an average of $4,340 ($1,057 ACTC and $3,283 EITC) per recipient family.[86] The total annual cost of refundable credits to these families is around $5.1 billion per year; the proposed policy would reduce EITC and ACTC payments by approximately that amount.[87]

Benefit Gaming: Using Relative Filers to Maximize EITC Benefits

As noted, the EITC law permits grandparents, aunts, uncles, and older siblings and stepsiblings to claim the EITC cash refund for children. To be eligible for the EITC, the relative is not required to contribute to the financial support of the child claimed, but the adult relative’s earned income must be greater than that of the child’s parent. Surveys show that low-income parents are very familiar with the technique of maximizing refundable credits by allocating their children among resident and nonresident relatives and boyfriends.[88]

Overall, approximately a half-million households engage in relative “benefit gaming” inside the household.[89] In these households, a parent has allowed another adult relative residing within the household to claim a child or children on his or her tax return, in nearly all cases to maximize EITC benefits.[90] These households represent about 1 percent of all households with children in the nation. Most commonly, these households consist of a parent, children, and a grandparent, although many other combinations occur.[91]

In addition, some 2.8 million EITC claimants each year violate the EITC residency rule with de facto impunity. These claimants will typically be absent parents or relatives. Overall, it seems that millions of parents permit absent fathers, grandparents, aunts, and uncles to claim their children for purposes of maximizing EITC payments, irrespective of their actual residence.

Allowing a relative, inside or outside the home, to claim a child for purposes of EITC benefits can be quite lucrative. Consider two hypothetical examples:

• A single mother has two children and has earned $10,000 in a year. Ordinarily, she would receive $5,060 in refundable tax credits ($4,010 from the EITC and $1,050 from the refundable ACTC). However, if she resides with her children’s grandparent (or aunt or uncle) who earned, say, $11,000 per year, then she can assign one or both of the children to the relative for purposes of claiming refundable credits. If she split the two children with the grandparent, the combined refundable tax credits would increase to $8,718 ($6,718 from the EITC and $2,000 from the ACTC). Having the relative claim one of the children increases EITC and ACTC benefits by $3,658 for the household.
• Another single mother has three children and earns only $3,000 during the year. She resides with her own mother, who has earned $15,000 in the year; she also has a sister who resides elsewhere and has earned $16,000. If the single mother filed for the EITC for all three of her own children, she would receive only $1,361 in EITC payments and no ACTC payments. However, under the law, the grandmother can claim the EITC for any or all of the children. The aunt can also claim the EITC for any child if she asserts that the child resided with her for half of the year. If the grandmother claimed two of the children and the aunt claimed the third, combined EITC and ACTC refundable credits would equal $11,907.[92] Refundable credits would be increased by over $10,000 per year by parceling eligibility among relatives. The incentive for the mother to work is diminished because she can share large EITC payments with little or no employment.

Moreover, there is no coordination between the EITC and ACTC and other welfare programs. In practical terms, one adult can often claim children for purposes of EITC and ACTC benefits while another adult claims them for purposes of Medicaid, food stamp, and TANF benefits.

EITC claims for eligible children are rarely, if ever, cross-checked against claims made for other benefits. As a consequence, there are few administrative barriers to a single mother with low or zero earnings claiming children for purposes of obtaining food stamp, TANF, and Medicaid benefits while allowing an employed relative to claim the children for purposes of obtaining EITC and ACTC benefits. For example, in the second hypothetical scenario, the mother may claim the children for purposes of food stamp, Medicaid, and TANF benefits while the relatives claim the children for purposes of EITC and ACTC benefits.[93]

Overall, a parent who works little or not at all can receive large EITC payments indirectly by claiming that her children reside with employed relatives in other locations. Moreover, even if the relative does reside with the child, it is unfair to the taxpayer to permit a parent to maximize EITC benefits simply by assigning children to other adults. A major rationale for the EITC is that it incentivizes increased work by parents; this principle is undermined if a parent who works little can simply transfer EITC eligibility to a relative who works more.

Recommended Policies to Limit Benefit Gaming Through Relatives

As noted, current law permits a parent to allocate some or all EITC eligibility for her children to a grandparent, uncle, aunt, older sibling, or step sibling as long as the relative ostensibly resided with the child for more than half of the year. If the relative resides in the same home with the parent, he or she must have higher earnings than the parent to qualify.

The law should be changed to eliminate relatives’ filing for the EITC and ACTC. In the future, only parents with formal legal custody, formal foster parents, and legal guardians should be permitted to file for EITC and ACTC credits.[94] Most of the savings from this policy change have been counted under the foregoing section on preventing residency errors. However, net of replacement costs, another $1 billion might be saved by preventing EITC and ACTC benefits from going to relatives rather than to parents inside the home.[95]

Eliminating EITC Benefits for Adults Without Dependents

Some 5.1 million adults without dependents received refundable EITC payments in 2013 at a cost of $1.5 billion.[96] An adult who works full-time, even at the minimum wage, will not fall below the poverty line; therefore, there is no need to provide welfare benefits to boost the income of these able-bodied adults without dependents. Nearly all of the beneficiaries work part-time or less than a full year. In fact, the structure of the credit actually rewards and encourages part-time rather than full-time work because the credit phases to zero at $14,820, which is roughly the amount of earnings of an employee working full-time/full-year at the federal minimum wage. Moreover, the credit for adults without dependent children contributes to the EITC marriage penalties described below because a father cohabiting with a mother can receive it, but a married husband with children cannot. The credit is wasteful and has harmful incentives. It should be eliminated.

Rewards for Cohabitation and Penalties for Marriage in the EITC

According to the well-known theory proposed by Nobel Prize–winning economist Gary Becker, individuals will choose to marry if their utility within marriage exceeds their utility outside of marriage.[97] A corollary of this principle is that couples are less likely to marry if marriage causes a decrease in their joint economic resources. In particular, if cohabiting couples lose welfare benefits upon marrying, the probability of marriage will be decreased.

 

Single parenthood and cohabitation have increased greatly over the past half-century. There were 2.5 million cohabiting couples with children in the U.S. in 2007. Roughly half of these, or 1.2 million, involved nonmarried mothers and fathers who jointly shared one or more biological children.[98]

All means-tested welfare programs provide penalties against marriage because their benefits fall as family income rises. When a couple marries, their joint income rises, and in most cases, welfare benefits will be cut.[99] The Earned Income Tax Credit often applies significant financial penalties against lower-income biological parents who marry while giving substantial cash bonuses to parents who cohabit without marriage.[100] The bias in favor of cohabitation and against marriage by parents comes from three factors.

• Cohabiting parents file for income tax separately; therefore, their earnings will be counted separately for purposes of calculating the EITC and ACTC. Married couples will file jointly, and the earnings of both spouses are added together for purposes of calculating benefits. For married couples, the value of the EITC decreases as joint earnings rise above $23,630 per year. For the overwhelming majority of dual-earner households, the EITC benefit would be higher if earnings were counted separately rather than jointly.
• In addition to counting earnings separately, cohabiting parents can assign children to either partner to maximize benefits; this option is not available to married couples.
• Cohabiting parents can assign all of their qualified children to one parent or split the children between the parents.[101] As long as there is more than one child in the home, cohabiting parents are given the option of making two separate returns with EITC claims for dependent children.[102] The EITC benefit per child decreases as the number of claimed children increases.[103] This means that splitting children between separate returns will generally yield a higher EITC benefit for dual-earner couples. Unlike cohabiters, married couples cannot take advantage of this double filing. The rewards of filing two separate returns increase as the number of children in the family increases. For example, the maximum EITC benefit for a married couple with four children is $6,242; for a cohabiting couple who each claim two children on separate returns, the maximum combined credit would be $11,196.

To understand the marriage penalty/cohabitation bonus provided by the EITC, consider a father and mother who each earn $20,000 per year and have two children.[104] The impact of marriage on this couple is shown in Table 2. If the couple cohabits, each parent will file a separate income tax return. Income will be maximized if each parent claims one child for tax purposes. The cohabiting couple will pay no income tax but will receive a combined $7,269 in refundable tax credits through the EITC and ACTC.

If they marry, the couple will still owe no federal income tax, but their refundable tax credits will be cut from $7,269 to $2,960, a loss of $4,308. The act of marriage requires this couple to sacrifice roughly one-tenth of their income. The cohabiting couple is also likely to receive other benefits such as food stamps that are not available to the married couple; however, those benefits do not appear in the table.[105]

Tables 3 and 4 show the penalties or gains from marriage for biological parents with different earnings levels and different numbers of children. The tables describe the combined effect of the EITC, ACTC, and federal income tax if a given couple cohabits rather than marries.

Table 3 expresses the annual dollar loss (for negative numbers) or dollar gain (for positive numbers) if cohabiting parents marry. Table 4 shows this loss or gain as a percentage of the couple’s combined earnings.[106] For example, if a couple has three children and the father earns $30,000 while the mother earns $10,000, Table 3 shows that the couple will lose $3,473 in income if they marry; Table 4 shows that this loss would be 8.7 percent of the couple’s combined income.

If a couple has only one earner, marriage will generate an income increase. (However, the one-earner couple may still face an overall marriage penalty if other benefits such as food stamps and subsidized housing are counted. Those benefits were not included in the analysis in Tables 3 and 4.) If the couple has two earners, marriage will result in an income loss from the combined effects of the EITC, ACTC, and federal income tax. In most cases, the income loss due to marriage will be between 3 percent and 15 percent of combined earnings.

The income loss due to marriage increases as the number of children in the family increases. For example, if a family has one child and both the mother and father earn $20,000, transitioning from cohabitation to marriage will cause a loss of $2,585, or 6.5 percent of joint earnings. If the same couple has two children, marriage will cause a loss of $4,308, or 10.8 percent of joint earnings. With three children, the loss would be $5,707, or 14.3 percent of earnings; for four children, the loss would be $7,675, or 19.2 percent of joint earnings. This pattern occurs because, as noted, under the law, cohabiting parents can apply for two separate EITC benefits and parcel their children between the two claims. The more children a couple have, the more advantageous this procedure is.

To file for the EITC, the unmarried father is supposed to reside with the child for at least half of the year. As discussed previously, this provision is widely violated. Some 2.8 million EITC claims (roughly one in 10) are based on false residency assertions by the tax filer. In the preceding examples, even if the father does not reside in the home, the unmarried parents may well file for the EITC as if they cohabited. As a result, they will gain the welfare benefits of cohabitation while residing separately. The higher EITC benefits would be shared between the absent father and the mother. Because the EITC residency rule is not and probably cannot be realistically enforced, in practical terms, the EITC program preferentially rewards not just cohabitation, but single mothers and absent fathers living separately as well.

 

Recommended Policies to Eliminate Marriage Penalties in the EITC

The law determining eligibility to claim the refundable portion of the EITC and the refundable ACTC should be changed to:

• Permit married couples filing jointly with dependent children, government-authorized foster parent or parents, and legal guardians of dependent children to claim the EITC and ACTC.
• Allow, among unmarried parents, only the parent with sole legal custody to claim a child for purposes of obtaining the EITC or ACTC.[107] The second cohabiting parent and absent parents would not be eligible to claim the credit. This change would restrict the EITC to the parent who is actually providing the primary support to a child. It would eliminate the current advantage given to cohabiting couples by permitting them to claim two separate EITC payments, an option that is not available to married couples. This change would also significantly reduce residency fraud; since absent parents could not claim the credit, they would have no incentive to make false residence claims.
• Allow married couples filing jointly with combined annual earnings of less than $50,000 to have the option of claiming the EITC on the basis of their joint earnings or on the basis of the earnings of either spouse considered separately. The couple would still file a single joint tax return and would be permitted to make only one EITC claim covering all of their dependent children.[108]
• Eliminate the EITC for adults without dependents. This credit is available to an absent parent or to one partner among unmarried cohabiting couples with children but is not available to married couples with children. The credit therefore increases the penalty against marriage. The credit is also wasteful. Compared to a family with children, an adult without dependent children has much less financial need. As noted, any adult without dependents who works full-time even at the minimum wage will have an income well above the poverty level. Giving cash welfare to such an individual is unnecessary.

Under current law, in most cases, an employed single mother who marries the father of her children will experience a significant drop in EITC benefits. The foregoing changes would remove the EITC’s penalties against marriage.

Cost of Eliminating the EITC Marriage Penalty

In 2014, there were 2 million dual-earner married-couple families with children under 18 and a joint income below $50,000 per year.[109] Most of these families already received EITC benefits; most would receive higher benefits if the income-splitting provisions described above became law. The precise cost of the pro-marriage benefit is difficult to calculate, but as a general rule of thumb, one could assume that policy would double EITC payments to eligible dual-earner families. Currently, the average refundable EITC payment to families with children is $2,900. Income splitting for dual-earner families might add another $2,500 in benefits per family. If so, this would increase EITC benefits by $5 billion per year (2 million families times $2,500).

However, the 2006–2008 compliance audit found that 1 million tax filers claiming the EITC had “filing status” errors.[110] These were married couples who filed two separate returns as heads of households and claimed the EITC. (Married couples filing separately are not eligible for the EITC.) The erroneous overclaims from this error totaled approximately $1.7 billion in 2015.[111] If the marriage penalty in the EITC were eliminated, these couples would have no incentive to file separate erroneous returns. Part of the cost of eliminating the EITC marriage penalty would be offset by the erroneous EITC payments currently made to these couples.

Thus, the cost of eliminating the marriage penalty in the EITC, net of the current cost of “filing status” errors, would be around $3.3 billion. This added cost would be more than offset by the savings from elimination of the current waste, fraud, and abuse in EITC discussed elsewhere in this paper.

Effects of Eliminating the Marriage Penalty

Research consistently demonstrates that married adults are physically and emotionally healthier than nonmarried adults; they are also more economically secure.[112] A healthy marriage is one of the two most important factors contributing to personal happiness.[113] Family structure is also the most important factor in predicting upward social mobility of children, and children with two parents do best.[114] But since the beginning of the War on Poverty, marriage has been seriously undermined in many low-income communities.[115] Replacing husbands with welfare checks has degraded personal well-being for men, women, and children in low-income neighborhoods.

Census data and the Fragile Families survey show that marriage can be extremely effective in reducing child poverty,[116] but the positive effects of marriage are not limited to income alone. When compared to children in intact married homes, children raised by single parents are more likely to have emotional and behavioral problems; be physically abused; smoke, drink, and use drugs; be aggressive; engage in violent, delinquent, and criminal behavior; have poor school performance; be expelled from school; and drop out of high school. Many of these negative outcomes are associated with the higher poverty rates of single mothers. In many cases, however, the improvements in child well-being that are associated with marriage persist even after adjusting for differences in family income. This indicates that the father brings more to his home than just a paycheck.

The effect of married fathers on child outcomes can be quite pronounced. For example, examination of families with the same race and same parental education shows that, when compared to intact married families, children from single-parent homes are:

• More than twice as likely to be arrested for a juvenile crime,[117]
• Twice as likely to be treated for emotional and behavioral problems,[118]
• Roughly twice as likely to be suspended or expelled from school,[119] and
• A third more likely to drop out before completing high school.[120]

The effects of being raised in a single-parent home continue into adulthood. Comparing families of the same race and similar incomes, children from divorced and single-parent homes are three times more likely to end up in jail by the time they reach age 30 than are children raised in intact married families.[121] Compared to girls raised in similar married families, girls from single-parent homes are more than twice as likely to have a child without being married, thereby repeating the negative cycle for another generation.[122]

Finally, the decline of marriage generates poverty in future generations. Children living in single-parent homes are 50 percent more likely to experience poverty as adults when compared to children from intact married homes. This intergenerational poverty effect persists even after adjusting for the original differences in family income and poverty during childhood.[123]

Further Steps to Reduce Welfare’s Bias Against Marriage

As noted, the welfare anti-marriage penalties are not limited to the EITC. Substantial marriage penalties also exist in the food stamp, Temporary Assistance for Needy Families, Supplemental Security Income (SSI), and housing programs.[124] For example, a single mother with $15,000 in annual earnings will typically receive $4,636 in annual food stamp benefits. If she married a father with similar earnings, the couple’s benefits would be cut to $2,357. When the marriage penalties of different means-tested aid programs are added on top of each other, the combined disincentives to marriage become notably significant.

One way to reduce the combined marriage penalties in the welfare state would be to reduce the marriage penalties in each individual welfare program. Regrettably, food stamps, housing programs, and the SSI program for children do not have work requirements for parents. Even in the TANF programs, work requirements are lax; on any given day, half of able-bodied TANF parents are completely idle.

Overall, these programs have strong work disincentives compared to the EITC, especially compared to a reformed EITC purged of the current fraud and error problems. Therefore, it would be a mistake to attempt to reduce welfare’s aggregate anti-marriage penalties by enrolling more married couples and raising benefits in these deeply flawed programs. Instead, the anti-marriage penalties in these flawed programs could be offset by increasing the EITC benefits for married couples. This approach would reduce the welfare state’s overall anti-marriage bias without notably increasing anti-work incentives.[125]

Currently, the maximum EITC credit varies by the number of children in a family, but the maximum credit for married-couple families and single-parent families is the same despite the fact that the married-couple family has one additional person.[126] To mitigate the anti-marriage bias in the overall means-tested welfare state, the maximum EITC benefit for married couples with children should be raised by 20 percent. This would raise the maximum benefit for a married couple with one child from $3,359 to $4,031 and the maximum credit for a married couple with two children from $5,548 to $6,657.[127] The phase-down rates of the credits would remain unchanged.

 

The cost of increasing the maximum credit for married couples in this manner would be around $4.5 billion per year.[128] To pay for this EITC expansion, funds should be diverted from the TANF program. When TANF was created in 1996, two of its four main goals involved strengthening marriage and two-parent families. Yet despite some $600 billion spent on TANF over the past two decades, state governments have spent only a pittance on these objectives. Currently, only 0.8 percent of TANF funds is ostensibly spent to promote two-parent families,[129] and even this small sum is usually spent on activities unrelated to increasing marriage.[130]

Since state governments have failed for two decades to act on TANF’s explicit pro-marriage goals, future federal TANF funding should be cut by $4.5 billion per year, and funds should be redirected to pay for the pro-marriage EITC expansion; this sum represents around 15 percent of current federal and state TANF expenditures. There is ample surplus within the TANF program: Currently, approximately 75 percent of TANF spending goes to the core program tasks of providing cash benefits, promoting work, and providing daycare to support work. The remaining $7 billion to $8 billion per year goes to tasks largely unrelated to the goals of the program.

It seems reasonable to redirect a portion of this random spending to the original goal of marriage promotion. The $4.5 billion allocated to reduce marriage penalties constitutes about 1 percent of total means-tested aid spent on families with children each year.

Impact of Reducing Marriage Penalties

Research indicates that marriage penalties in the tax code, and particularly in the EITC, have significant effects on marriage and cohabitation rates. One longitudinal study examined unmarried women with a high school degree or less, aged 18 to 50, over a period of roughly 48 months.[131] The study found that over this period, a potential EITC anti-marriage penalty of $1,000 reduced the probability that a woman would marry by 10 percent and raised the probability of cohabitation by a similar rate.[132] Women residing in states with their own EITC programs showed even further reductions of 20 percent to 30 percent in marriage rates.[133] The study found that dollar for dollar, EITC marriage penalties had a much stronger impact on the probability of marriage than did the earnings of a potential spouse.[134]

Other research by Hayley Fisher of the University of Sydney has examined the effect of marriage penalties in the U.S. tax code, including the EITC, on marriage rates. Fisher found an elasticity of –1.1 between marriage penalties and the probability of marriage. An increase in a couple’s marriage penalty equal to 1 percent of their combined income would lower their probability of marriage by 1.1 percent. The effect was strongest among less educated individuals.[135]

Budgetary Impact of the Proposed Policies[136]

Improper payments equal benefit overclaims minus the savings from current IRS enforcement actions. Improper payments in general appear to be around 15 percent less than overclaims. In recent years, according to the Treasury Department, improper payments in the EITC and ACTC programs have totaled $22.1 billion.[137] As Table 5 shows, improper payments on average are 24.8 percent of total benefits in the two programs. Nearly all improper payments are overpayments.[138]

Table 6 shows the budgetary impact of the policies presented in this paper.[139] The first line of the table shows the impact of policies to reduce improper overpayments due to income misreporting, residence errors, and other qualifying child errors. (Filing status error due to filers attempting to avoid marriage penalties are dealt with later in the table.) Total EITC and ACTC improper overpayments due to income misreporting, residence errors, and other qualifying child errors have been approximately $20.4 billion per year in recent years. The analysis assumes that the proposed policies, if fully implemented, would eliminate 95 percent or $19.4 billion per year in these improper payments. (Reducing the error rate is made easier by the fact that many erroneous returns have multiple errors. On average, there are 1.3 errors for each erroneous EITC return, so mechanisms for stopping overpayments are numerous.)

 

However, if a tax filer is found ineligible to receive a credit for a given child, the primary parent may still be eligible for alternative replacement benefits. As noted, replacement benefits for returns with residence errors and other qualifying child errors are around 50 percent of the current erroneous benefits. The majority of filers with misreported income are not eligible for any benefit, but around a third appear to be eligible for reduced benefits. Overall, the replacement rate for misreporting income returns is estimated at 15 percent of current benefits. Net of replacement costs, all of the proposed error-reduction policies would generate an estimated $12.8 billion in savings each year.

The second line on the table shows that the policies to limit multi-tier benefits, restrict eligibility for relative filers, and eliminate benefits for adults without dependent children would generate $7.3 billion in annual savings. Relative filers in the home would generally be replaced by primary parents, producing $1 billion in replacement costs. EITC and ACTC benefit reductions net of replacement costs would be $6.3 billion.

Eliminating the Marriage Penalty in the EITC. In 2014, there were 2 million dual-earner married-couple families with children under 18 and a joint income below $50,000 per year. Most of these families already received EITC benefits; most would receive higher benefits if the proposed income-splitting policies became law. Eliminating the EITC marriage penalty for these families would cost an estimated $5 billion per year.

However, the 2006–2008 compliance audit found 1 million tax filers claiming the EITC with “filing status” errors.[140] These were married couples who filed two separate returns as heads of households and claimed the EITC separately to avoid the marriage penalty; such separate filing is illegal. The erroneous overclaims from this error were 9.3 percent of total overclaims.[141] In 2015, the cost of these errors was around $1.7 billion.

If the marriage penalty in the EITC were eliminated, these couples would have no incentive to file separate erroneous returns. Part of the cost of eliminating the EITC marriage penalty would thus be offset by the erroneous EITC payments currently made to these couples. Since the couples already evade the EITC marriage penalty, eliminating the marriage penalty for them would produce no added cost. The real cost of eliminating the marriage penalty in the EITC, net of the current cost of “filing status” errors, would thus be $5 billion less $1.7 billion, or $3.3 billion.

The second proposed policy to reduce the welfare state’s anti-marriage bias is to raise the maximum EITC benefit for married couples with children by 20 percent. The cost of this increase would be around $4.5 billion per year. Federal funding for the TANF program would be reduced by this amount to offset the added EITC costs.

The policies outlined would eliminate nearly all of the improper EITC and ACTC overpayments discussed in this paper. Improper payments due to income misreporting, residence errors, and qualifying child errors would be almost eliminated. Illegal filing to avoid marriage penalties (or filing status errors) would be greatly reduced because the incentive to misfile would be removed. Even if couples continued to file incorrectly, no financial gain and no benefit overpayment would occur.

Altogether, the policy would produce a net yearly reduction of $19.1 billion in fraudulent, erroneous, and wasteful benefits. Eliminating the marriage penalty in the EITC would increase costs by $3.3 billion, yielding net savings of $15.8 billion per year.

Conclusion

The foundational principle of a sound welfare system is that benefits should complement and reinforce efforts toward self-support through work and marriage rather than weakening or displacing those efforts. A welfare system that melds assistance with an individual’s self-help efforts is more efficient, is more humane, and confers greater dignity on beneficiaries.

On the surface, the EITC and ACTC seem well suited to this task, but both have many significant problems. The EITC actively discriminates against and penalizes marriage. Extensive fraud wastes taxpayer funds and undermines the pro-work incentives of the EITC and ACTC. Excessive piggy-backing of benefits further wastes funds.

The EITC and ACTC cost the taxpayers close to $90 billion per year. Some 90 percent of this cost takes the form of refundable cash payments to low-income individuals. Viewed together, the EITC and ACTC are the second most expensive component of the means-tested welfare system, exceeded in cost only by Medicaid.

One-quarter of total EITC and ACTC benefits—or $22 billion per year—are erroneous or fraudulent payments. Each year, at least 8.4 million families with children receive erroneous or fraudulent cash aid from these programs; most are completely ineligible for aid. Annual erroneous cash payments through the EITC and ACTC programs now exceed the cost of yearly cash benefits paid by the Temporary Assistance for Needy Families program. Five times as many families with children receive erroneous overpayments from EITC and ACTC as receive any cash aid from TANF.

Providing massive cash aid to ineligible persons is not just inefficient; it erodes the very purpose of the credits in the first place. The EITC and ACTC are designed to provide income support while simultaneously increasing incentives for work and earnings, but fraud undermines these incentives. For example, allowing an individual to receive EITC benefits on the basis of fictitious earnings claims undermines the incentive to obtain real earnings. Similarly, allowing a custodial parent who works little to obtain EITC benefits based on the earnings of absent parents or relatives clearly dilutes work incentives for the custodial parent.

It is often argued that the high error rate in the EITC and ACTC is due to the complexity of these programs. This is not true. Many also suggest that removing the fraud from these programs would be difficult and expensive. This also is untrue: Simple changes in eligibility and administration could eliminate most of the fraud. For example, simply delaying payment of the credit until incomes and identities have been verified would save around $10 billion each year. In addition, restricting receipt of benefits to parents and guardians with legal custody of the children would close the remaining loopholes through which the bulk of the remaining fraud occurs.

Other forms of excess in the EITC and ACTC should be reduced as well. The current welfare systems allow individuals to piggy-back benefits for many different welfare programs on top of each other. In some cases, piggy-backed benefits can become excessive, allowing a single mother with two children earning the minimum wage to obtain as much as $47,000 per year in combined benefits and earnings. Excessive piggy-backing of benefits should be limited.

Finally, the EITC and the welfare system in general discriminate against and penalize low-income parents who marry. The current marriage penalties within the EITC should be eliminated. In addition, the basic EITC benefit for married couples with children should be increased to partially offset the marriage penalties embedded in other welfare programs such as food stamps and public housing. Modest changes that reduce marriage penalties in welfare could significantly increase marriage in low-income communities. This, in turn, would substantially increase personal well-being and upward social mobility.

To accomplish their intended mission, the current EITC and ACTC programs must be dramatically reformed to eliminate marriage penalties, wasteful spending, and extensive fraud. The changes proposed in this paper should produce $19.1 billion per year in savings. Reducing marriage penalties would cost some $3.3 billion. Overall net savings would therefore come to $15.8 billion per year.

—Robert Rector is a Senior Research Fellow in the Institute for Family, Community, and Opportunity at The Heritage Foundation. Jamie Bryan Hall is Senior Policy Analyst in the Center for Data Analysis, of the Institute for Economic Freedom, at The Heritage Foundation.

Budgetary Appendix

 

 

EITC and ACTC Estimates

Earned Income Tax Credit Estimates for 2013. In 2013, if both returns with EITC cash payments and returns with tax reduction due to the EITC were counted, some 28.8 million returns used the EITC.[142] Of those returns, 21.6 million had dependent children and 7.2 million had no children.[143] The total value of the EITC in the 28.8 million returns was $68.1 billion.[144]

In 2013, refundable credits equaled 87 percent of the total value of the EITC. Nonrefundable credits that reduced family income tax payments that otherwise would have been made comprised 13 percent of the total value of the EITC.[145] IRS analyses of EITC overclaims and improper payments use the total value of the EITC, including both refundable and nonrefundable portions, as the basis for analysis.[146]

If only returns with refundable payment are counted, 20.1 million families with children and 5.1 million individuals without children received refundable EITC payments in 2013 at a cost of $59.1 billion.[147] The 20 million tax returns with dependent children that claim the refundable EITC had, on average, 1.7 children per return.

EITC Estimates for 2013. EITC error rates do not seem to have changed notably over time. For example, the IRS estimated the improper payment rate at 25.5 percent in 2009 and 23.8 percent in 2015.[148] For illustrative purposes, Appendix Table 3 applies the ratios for total overclaims and overclaim error types from the 2006–2008 compliance audit to the 2013 EITC totals. The table assumes that the overclaim ratio and the distribution of types of errors were roughly the same in 2013 as in 2006–2008; however, the number of EITC returns has been raised to 2013 levels.[149] According to the low estimates from the IRS 2006–2008 compliance review, 43 percent of EITC returns were overclaims; this rate would mean 12.4 million overclaim returns in 2013.[150]

EITC Estimates for 2015. According to the Department of the Treasury, the estimated value for total EITC claims in 2015 was $65.6 billion.[151] Data on the number of returns are not given. Appendix Table 4 applies the overclaim and error ratios from the 2006–2008 audit to the EITC cost data for 2015. The 2006–2008 overclaim ratio of 28.5 percent would yield $18.7 billion in overclaims in 2015. In the 2006–2008 compliance reviews, roughly 97 percent of the dollar value of EITC overclaims involved returns with dependent children.[152]

Additional Child Tax Credit Estimates. In 2013, 19.9 million tax filers received the refundable ACTC at a cost of $26.7 billion.[153] The average value of the ACTC was $1,330 per return. The 2015 report of the Treasury Inspector General for Tax Administration (TIGTA) stated:

[O]ur review of the IRS’s own enforcement data indicates that the ACTC improper payment rate is similar to that of the EITC. TIGTA estimates that the ACTC improper payment rate for Fiscal Year 2013 is between 25.2 percent and 30.5 percent, with potential ACTC improper payments totaling between $5.9 billion and $7.1 billion.[154]

The estimates in this paper assume that in most cases, tax returns with children with erroneous ETIC overclaims also have erroneous ACTC payments. Cost data from 2013 are used because later data are not available. The erroneous ACTC payments are estimated at $900 per return. This leads to an overall ACTC improper payment sum of $6.5 billion in 2103, which matches the midpoint of the TIGTA improper payment estimates.[155]

 

 

 

 

Budgetary Analysis: Key Assumptions

The cost savings derived from error and fraud reduction appearing in Appendix Table 5 and the paper in general are based on six assumptions.

• The Treasury midpoint estimate of $15.6 billion in EITC improper payments in 2015 is assumed to be correct.[156] Cost savings estimates are derived directly from this total improper payment estimate. The overclaim figures that appear on lines 1–3 of Appendix Table 5 are used to estimate types of errors, not total improper payments or overall savings.
• The proposed policy changes are assumed to eliminate 95 percent of improper payments. This assumption is used because the policies address 99 percent of past errors. Moreover, many erroneous claims have multiple errors; there are, on average, around 1.3 separate errors for each erroneous return. The presence of overlapping errors increases the probability of nearly eliminating erroneous returns and their cost.
• Overclaim estimates appear on lines 1–5 in Appendix Table 5; these overclaim estimates do not affect the principal improper payment total on line 6. The overclaim data simply provide estimates of the distribution of types of errors. The distribution of dollar costs due to income misreporting errors, qualifying child errors, filing status errors, and other errors is assumed to be the same in 2015 as it was in the 2006–2008 compliance audit. (Changes in the proportionate distribution of errors will have modest effects on the savings presented in the paper. For example, if the share of errors due to income misreporting increased, potential cost savings for a given rate of improper payments would also increase because the replacement rate for income misreporting appears to be lower.)
• The replacement rate for returns with misreported income is assumed to be 15 percent; for qualifying child errors, it is assumed to be 50 percent. These assumptions are explained below.
• The cost of ACTC errors is assumed to equal the midpoint improper payment estimate provided by the Treasury Inspector General for Tax Administration for 2013, or $6.5 billion.[157]
• The average ACTC improper payment is assumed to be $900 per return. This is less than the average ACTC payment per return but matches well with an overall ACTC improper payment rate of $6.5 billion.

 

Detailed Budgetary Estimates

A detailed budgetary analysis is presented in Appendix Table 5. The following discussion explains this table on a line-by-line basis.

1. Policies to Reduce Fraud and Erroneous Payments

Overclaims Due to Income Misreporting (line 1). As noted, according to IRS compliance audit data, 43 percent of EITC returns had an overclaim; in 2013, that would be 12.4 million returns. A quarter of all EITC returns had overclaims due solely to income misreporting in 2006–2008.[158] In 2013, that ratio would result in 7.2 million erroneous returns.[159] Income misreporting (without other errors) produced dollar overclaims equaling 10 percent of all EITC costs in 2006–2008.[160] In 2015, one-tenth of EITC costs would be $6.6 billion.

Some EITC overclaim returns due to income misreporting are zero-child returns. There were 7.2 million zero-child returns in 2013, and the IRS finds that about 30 percent of zero-child returns are erroneous.[161] This would be some 2.2 million erroneous zero-child returns. Therefore, the maximum number of zero-child return with income misreporting errors would be around 2.2 million. The remaining 5 million returns with income misreporting would be filers with children. Most of these would also receive the ACTC. ACTC overclaims by these 5 million filers are assumed to equal $4.5 billion (5 million times $900 per return).

The IRS compliance audit finds that in that 80 percent of returns with erroneous overclaims, the filer is completely ineligible to receive the credit.[162] Some 10 million overclaim returns would fall in that category in 2013. In another 20 percent of overclaims, the filer is eligible to receive a benefit of lower value; that would be an estimated 2.2 million filers in 2013. Since returns with qualifying child errors would be for the most part completely ineligible for the credit, the returns with eligibility for lower payment would be predominantly among the 7.2 million overclaim returns involving income misreporting and no other errors.

It seems reasonable to assume that roughly a third of these cases with only income reporting errors (2.4 million out of 7.2 million) were eligible for a reduced EITC payment and two thirds were eligible for no payment. If the value of the reduced credit is assumed to equal to half the value of the erroneous credit, then the partial credits altogether would be worth 15 percent of the total value of overclaims due to income misreporting. That figure is used in the “replacement rate” column of line 1. The budget savings net of replacement costs come to $9.4 billion.

False Residency Claims (line 2). Some 13 percent of EITC returns contained a qualifying child (QC) error.[163] Some 75 percent of QC errors were residence errors,[164] so roughly one in 10 total EITC claims (or 2.8 million claims in 2013) are based on false claims of residence by absent parents or relatives.[165]

Since false residency claims account for 75 percent of all returns with qualified child errors, the present analysis assumes that 75 percent of the dollar value of EITC overclaims due to qualifying child errors is also due to false residency errors.[166] The total estimated value of EITC overclaims due to qualified child errors in 2015 was is $9.9 billion.[167] Overclaims due to false residency claims are assumed to be 75 percent of that amount, or around $7.4 billion.

Most individuals who make false residence claims to receive the EITC will also be eligible to receive the refundable ACTC for the erroneously claimed child; at higher earnings levels, they may benefit from the nonrefundable CTC. The cost of the erroneous ACTC claims linked to erroneous residency claims is estimated to be 2.8 million returns at $900 per return, or $2.5 billion.[168]

Research on EITC payments in Wisconsin indicates that the replacement rate for erroneous payments to nonresident parents (usually absent fathers) is 56 percent. The replacement rate for erroneous payments to relatives and other nonqualified individuals is 39 percent.[169] The present analysis assumes a replacement rate of 50 percent on all returns with false residence claims. The net savings from eliminating false residence claims would equal the cost of such overpayments ($7.4 billion) minus the cost of the replacement benefits. Fifty percent of $7.4 billion is $3.7 billion. ACTC savings would be around 50 percent of $2.5 billion, or $1.25 billion.

Other Qualified Child Errors (line 3). In addition to residency errors, the remaining four major types of qualifying child errors are lack of legal relationship, over-age children, lack of a valid Social Security number for the child, and tiebreaker errors. Since residency errors are assumed to equal 75 percent of the error costs due to qualifying child errors, the remaining qualifying child error will be 25 percent of qualifying child costs. In 2015, that would be 25 percent of $9.9 billion, or $2.5 billion.[170] There were an estimated 3.7 million EITC returns with qualifying child errors in 2013; 25 percent of these comprise about 0.9 million returns with other QC errors.[171] The associated ACTC costs would be around $0.8 billion.[172] These figures appear on line 3 of Appendix Table 5.[173]

The rest of this section discusses various types of qualifying child errors separately; these separate figures do not appear in Appendix Table 5. Around 3 percent of all EITC returns are erroneous because the filer does not have a legal relationship to the claimed dependent child.[174] These errors account for around 20 percent of all qualifying child errors at a probable cost of $2 billion in 2015 (20 percent of $9.9 billion).[175] In 2013, about 750,000 EITC tax returns probably had this type of error.[176] ACTC costs associated with these errors are estimated at around $0.7 billion.[177] If the replacement rate on these erroneous EITC and ACTC claims is assumed to be 50 percent, the net savings would be around $1.3 billion.[178]

The 2006–2008 compliance audit found that 1.5 percent of all dependents claimed on EITC returns were overage and not subject to exemption.[179] (The audit process provides ample opportunity for the filer to demonstrate that the over-age dependent was actually disabled or in school.) Around a tenth of qualifying child errors were age-of-child errors.[180] The cost of all qualifying child errors in 2015 was $9.9 billion; this implies that age-of-child errors cost about $1 billion. There would be no replacement factor with these errors. Filers who erroneously claim the EITC for overage children may also erroneously claim the ACTC; potential ACTC overpayments could total $300 million per year.

Claims with invalid Social Security numbers represent 8 percent of qualifying child errors. These claims are probably blocked by current IRS enforcement procedures.[181] Tiebreaker errors represent 7 percent of qualifying child errors. These errors would be eliminated by the proposed changes in EITC eligibility.

Some returns have multiple qualifying child errors. Added together, residency errors, lack of legal relationship, age errors, invalid Social Security numbers, and tiebreaker errors amount to about 1.5 QC errors for each return with any QC error.[182] More than a quarter of returns with qualifying child errors also have income misreporting errors. It seems likely that the multiple reforms proposed in this paper combined with existing enforcement procedures would be sufficient to eliminate nearly all qualifying child improper payments.

Reduced Overclaims (line 4). These figures equal the sum of lines 1–3. The cost of filing status errors (on line 15) is not included in this sum. Total estimated EITC overclaims appear on line 22 and equal $18.7 billion; this figure equals the sum of line 4 plus filing status errors on line 15 and $0.5 billion in nonlisted erroneous claims.

Existing IRS Enforcement Actions (line 5). The IRS does take modest enforcement actions to block payments of erroneous overclaims or, more typically, to recover erroneous payments after they are made. The IRS states that savings from enforcement actions in 2011 equaled $2.5 billion.[183] The report of the Treasury Inspector General for Tax Administration states that the IRS prevented or recovered $2.2 billion in erroneous EITC payments in 2014.[184]

The analysis in Appendix Table 5 calculates current Treasury enforcement savings on line 5 as a residual. The enforcement savings equal total estimated overclaims on line 4 minus the improper payment totals on line 6. The improper payment totals are taken directly from the midpoint most recent Treasury improper payment estimates. (See the discussion on line 23.)

It is possible that the savings from current enforcement actions are more or less than the estimates on line 5. However, the present analysis assumes that the Treasury Department’s improper payment estimates for 2015 are correct. Given that, raising the enforcement savings estimate would necessarily lead to higher overclaim estimates in lines 1–3. (This occurs since improper payments plus enforcement savings equals overclaims.)

Therefore, raising or lowering the savings estimate from current IRS enforcement actions (on line 5) would not affect the savings estimates in this paper since those estimates are based on current IRS improper payment estimates (on lines 6, 15, and 23) that are net of current enforcement actions. Similarly, raising or lowering the overclaim numbers on lines 1–3 would not affect the overall budget savings figures in this analysis since those savings are not tied to overclaims, but rather are based on the improper payment figures in lines 6 and 15 (which are summed to equal total $15.6 billion in improper payments on line 23).

Subtotal Improper Payments (line 6). The figure of $13.9 billion in improper payments equals the total improper payments figure of $15.6 billion on line 23 less the filing status error figure of $1.7 billion on line 15. The $15.6 billion improper payment total on line 23 equals the midpoint Treasury Department estimate of improper EITC payments for 2015.[185] Line 6 also equals the reduced overclaim estimate on line 4 minus current IRS enforcement actions on line 5.

Reduced Improper Payments (line 7). Overall, the analysis assumes that the proposed error and fraud reduction reforms would eliminate 95 percent of current improper payments. The savings of $13.2 billion on line 7 equals 95 percent of $13.9 billion, the improper payment figure on line 6.

2. Additional Policies to Reduce Waste

Eliminating Benefits for Relative Filers (line 8). Multi-family households with children that contain at least two adult biological relatives filing separate tax returns are about 11 percent of the 48 million households with children. (A spouse or cohabiting partner is not considered a relative for purposes of this calculation.) Approximately one-tenth of these households engage in “sorting,” or gaming behavior in which one or more children are assigned to a relative other than the parent or closest relative in order to obtain tax benefits.[186] This means that roughly 530,000 households with children are engaged in benefit gaming within the households. (This implies that the number of relatives receiving EITC benefits as a result of false residence claims is likely to greatly exceed the number of relative filers who actually reside with mother and child.) EITC returns by relatives are assumed to have the average EITC refundable benefit cost for a return with children in 2015 of $2,919 and an average ACTC cost of $900.[187] The gross cost of EITC and ACTC benefits would be around $2 billion. Replacement costs are assumed to be 50 percent. Thus, the net savings from the policy would be 50 percent of $2 billion, or roughly $1 billion.

Eliminating EITC Benefits for Adults Without Dependent Children (line 9). Some 5.1 million adults without dependent children received refundable EITC benefits in 2013 at a total cost of $1.5 billion.[188] Eliminating eligibility for this group could therefore reduce EITC costs by $1.5 billion.

Eliminating Multi-tier Benefits for Recipients of Subsidized Housing (line 10). Some 1.8 million families with children receive HUD housing subsidies from public housing, Section 8, or other HUD programs.[189] Some 65 percent of families with children under 18 years old who report receiving housing subsidies also receive refundable credits.[190] Thus, approximately 1.2 million families with children receive both the housing subsidies and the EITC and ACTC. The average cost is $1,057 in ACTC benefits and $3,283 in EITC benefits per recipient family.[191] The total annual cost of refundable credits to these families is around $5.1 billion.[192] The proposed policy change would therefore reduce ETIC and ACTC payments by an estimated $5.1 billion each year.

Reduction in Savings Due to Overlap Between Policies (line 11). Eliminating income misreporting errors, residence errors, and other qualified child errors would reduce total EITC and ACTC refundable benefits by 14 percent after accounting for replacement costs. (This ratio is calculated by the total savings on line 7 of $12.8 billion divided by the total combined cost of the EITC and ACTC of $88.9 billion in 2015.) These changes should reduce benefits for relative filers, households in subsidized housing, and adults without dependents by a similar amount. As a result, the net savings from eliminating EITC and ACTC benefits on lines 7, 8, and 9 would probably be 14 percent lower. The reduced savings due to policy overlap is shown on line 11.

Subtotal for Policies to Reduce Waste (line 12). This line equals the sum of lines 8 through 11.

Subtotal for Additional Reduction Policies (line 13). This equals the sum of savings from fraud and error reduction on line 7 and the additional policies to reduce waste on line 12.

3. Policies to Strengthen Marriage

Eliminating the Marriage Penalty in the EITC (line 14). In 2014, there were 2 million dual-earner married couple families with children under 18 and a joint income below $50,000 per year.[193] Most of these families already received EITC benefits; most would receive higher benefits if the income splitting provisions proposed in this paper became law. The precise cost of the pro-marriage benefit is difficult to calculate, but as a general rule of thumb one could assume that policy would nearly double EITC payments to eligible dual-earner families. (Note that benefits under the proposed policy would be less than the current benefits for cohabiting couples because married couples under the proposals would still file a single return and could not benefit by parceling children among separate returns.)

Currently the average refundable EITC payment to families with children is $2,919.[194] Income splitting for dual-earner families might add another $2,500 in benefits per family. If so this would increase EITC benefits by $5 billion per year (two million families times $2,500).

Cost of filing status errors (line 15). The 2006–2008 compliance audit found 1 million tax filers claiming the EITC with “filing status” errors.[195] These were married couples who filed two separate returns as heads of households and claimed the EITC separately; such separate filing is illegal. The audit found that erroneous overclaims from these errors cost $1.3 billion. These errors were 9.3 percent of total overclaims.[196] In 2015, 9.3 percent of overclaims would equal $1.7 billion.[197] If the marriage penalty in the EITC were eliminated, these couples would have no incentive to file separate erroneous returns. Part of the cost of eliminating the EITC marriage penalty would be offset by the erroneous EITC payments currently made to these couples. Since these couples already evade the EITC marriage penalty, eliminating the marriage penalty in those cases would produce no added cost.

Cost of Eliminating the Marriage Penalty Net of Current Status Filing Errors (line 16). The real cost of eliminating the marriage penalty in the EITC net of the current cost of “filing status” errors would thus be $5 billion less $1.7 billion, or $3.3 billion.

Expanding the Maximum EITC Benefit for Married Couples (line 17). Currently, married couples with children receive around 30 percent of all EITC benefits; 30 percent of $65.6 billion equals roughly $20 billion in benefits each year.[198] Increasing the maximum benefit by 20 percent would add around $4 billion in costs.

The phase-out rate of the EITC for married couples is 15 percent for couples with one child and 21 percent for two or more children. Raising the maximum benefit without altering the phase-out rates would raise the break-even point at which the benefit is phased down to zero. For example, the break-even point for a married couple with two children would increase from $49,974 to around $55,000. This increase would result in roughly 1 million married couples being eligible for the credit at an average cost of around $500 per family.[199] The cost of shifting the break-even earnings point upward by $5,000 would be around $500 million (1 million families times $500). Thus, the overall cost of the expansion would be around $4.5 billion.

Transfer of Funds from TANF (line 18). Federal TANF funding would be reduced by $4.5 billion per year to pay for the cost of increasing the maximum EITC benefit for married couples with children.

Net Cost of Marriage Policy (line 19). The gross cost of the proposed marriage policies would be $9.4 billion (the sum of lines 14 and 17). Offsetting savings of $6.2 billion are found on lines 15 and 18. The cost net of savings would be $3.3 billion.

Overall Budgetary Impact of Policy Changes (line 20). The policies outlined on lines 1–12 eliminate an estimated $26.7 billion in erroneous, fraudulent, and wasteful expenditures in the EITC and ACTC programs. (See line 13.) However, part of these savings would be offset if legitimate filers claimed the children in future years. With a replacement rate of 50 percent for qualified child errors and 15 percent for income reporting errors, the net savings would come to $19.1 billion (on line 13). Eliminating the EITC marriage penalty would have a net cost of around $3.3 billion (on line 19), so the net savings would come to $15.8 billion per year.

Total Costs (addendum line 21). According to Treasury figures, total refundable payments and tax reduction from the EITC in 2015 was $65.6 billion.[200] Total refundable ACTC payments were $23.3 billion.[201] Combined costs were $88.9 billion.

Total Overclaims (addendum line 22). Based on the 2006–2008 compliance audit, overclaims in 2015 were estimated to be 28.5 percent of total EITC costs in line 21, or $18.7 billion.[202] This figure also equals the sum of the overclaim errors on lines 1, 2, and 3 ($16.5 billion) and filing status overclaims on line 15 ($1.7 billion) plus $0.5 billion in other overclaims not listed in this paper.

Total Improper Payments (addendum line 23). The total EITC improper payment estimate of $15.6 billion equals the midpoint IRS estimate for 2015.[203] The ACTC overpayment figure equals the midpoint TIGTA estimate of $6.5 billion for 2013.[204] More recent estimates are not available. The figures on line 23 equal the sum of lines 6 and 15.