The Earned Income Tax Credit (EITC) is the government’s second-largest means-tested cash benefit. In 2017, the program paid $56.7 million in cash benefits to 23 million families. The EITC was originally designed as a program to help low-wage parents support their children through a refundable tax credit.
In 1993, a small benefit was added for adults without dependent children. Progressives argue that the EITC is unfair because a single mother with three children can receive up to $6,557 in EITC benefits while a childless worker can get only $500. This argument ignores the simple fact that the single mother gets $6,000 more because she has to feed, house, and clothe three extra persons.
The median childless EITC recipient earns $10.00 per hour, but these recipients have low annual earnings primarily because they work only around 700 hours per year. If the typical childless recipient worked full-time through the year (2,000 hours), he or she would have a post-tax income of $17,700, or 30 percent above poverty. Poor recipients have low earnings because they work little, but raising their EITC benefits would not increase that work and would not significantly reduce poverty.
Moreover, fraud and erroneous payments are prevalent in the childless worker credit as they are in the EITC in general. Around 30 percent of childless adult recipients are ineligible for the benefits they receive. EITC expansion would only make that problem worse.
Yet progressives still urge that the EITC cash grant for adults without dependents be increased. They contend that this increase would reduce the EITC bias in favor of parents, increase employment among the covered population by as much as 10 percent, reduce crime and recidivism, and increase marriage. They argue that the expansion would particularly benefit the most disadvantaged groups such as low-wage males, noncustodial parents, and former prison inmates. Using these rationales, both President Barack Obama and former Congressman Paul Ryan (R–WI) supported raising the maximum EITC for adults without dependents to $1,000 per year.
Random assignment experiments on the effects of raising the EITC for childless adults have been conducted in New York City and Atlanta. Called Paycheck Plus, these experiments quadrupled the maximum value of the EITC for childless adults from $500 to $2,000 and raised the income point at which the credit phased down to zero from around $15,000 to around $30,000.
The Paycheck Plus experiments refuted most of the optimistic expectations of those who advocate higher EITC benefits for childless adults. Specifically, quadrupling the maximum value of the EITC credit:
- Had no statistically significant increase on employment among the eligible group in Atlanta;
- Increased employment by 1.8 percentage points among the eligible group in New York City;
- Failed to increase earnings in either city;
- Had no effect on the employment or earnings of the most disadvantaged groups (low-wage males, noncustodial parents, and former inmates);
- Did not reduce poverty;
- Weakened family formation;
- Had no impact on crime or recidivism; and
- Had no impact on overall family income, physical health, mental health, happiness, material hardship, or hunger.
If the Paycheck Plus plan were implemented nationwide, it would triple the number of childless adults receiving refundable EITC cash payments. Nonmarried childless recipients would rise from the current level of around 3.5 million to 12.7 million. Roughly one in five nonmarried childless individuals between the ages of 21 and 64 throughout the United States would be eligible to receive refundable cash payments.
The cost of the refundable benefits to taxpayers would increase fourteenfold from around $1.44 billion per year to over $17 billion per year. While the program might increase the number of employed persons by 200,000, that translates into one extra job for each $80,000 in new spending.
Paycheck Plus had the least impact on the most disadvantaged workers. By contrast, an alternative policy called transitional jobs has the greatest impact on the hardest-to-employ workers. Transitional jobs programs are tightly targeted and provide jobs to individuals who have the greatest difficulty obtaining and sustaining employment.
Many transitional jobs programs assist formerly incarcerated persons. The Paycheck Plus program had zero impact on the employment of this group, but transitional jobs programs routinely increase employment of former inmates by 20 to 30 percentage points during the first year of enrollment.
Recidivism is a key concern among former inmates. Paycheck Plus had zero impact on recidivism. By contrast, some transitional jobs programs show considerable promise on this issue, but the evidence is mixed, and most programs have no impact at all. The effectiveness of transitional jobs programs in reducing crime and recidivism can potentially be improved by funding programs for former inmates on a payment-for-outcome basis with providers receiving bonuses if recidivism among clients is actually decreased.
An EITC expansion for childless adults modeled after Paycheck Plus or similar proposals would be a massive policy change involving millions of recipients and tens of billions of dollars. But this huge program expansion would have minimal impact on employment and earnings and would have the least effect on the most disadvantaged workers. More effective and more wisely targeted policies are needed.
The History of the EITC
Originally called “the work bonus plan,” the Earned Income Tax Credit was devised as a mechanism to provide assistance to low-income families with children while at the same time encouraging work. Proponents felt that families could be better aided and poverty more efficiently reduced if taxpayer-funded assistance encouraged and augmented self-support rather than displacing it.
Enacted in 1975, the EITC differed strongly from all traditional welfare programs; those programs gave the greatest benefits to those who did not work and reduced aid as work increased. In contrast, EITC eligibility was limited to parents who worked, and in the lower income range, aid was increased as work and earnings increased. The original EITC was available to low-income parents with children and paid 10 percent of earnings up to a maximum to $400 at $4,000 of earnings. As earnings rose above $4,000, the credit was phased down incrementally, reaching zero for earnings of $8,000 or more per year.
The original EITC was intended to help low-wage parents support their children. Since childless workers did not have dependents to support and could far more readily support themselves, they were not eligible for the credit. However, the Omnibus Budget Reconciliation Act of 1993 for the first time extended a small EITC payment to workers without dependent children. The intent of this change was not to increase employment or reduce poverty, but simply to offset the negative impact of a gasoline tax increase included elsewhere in the legislation on workers with low earnings.
The Current EITC Credit for Persons Without Dependent Children
In 2019, the EITC for a single person without dependent children had a maximum value of $529 per year. Like all current EITC benefits, the benefit structure of the credit for adults without dependents has a trapezoid shape consisting of a phase-up range, a plateau with the maximum benefit, and then a phase-down range. As Chart 1 shows, the value of the credit for each single equals 7.65 percent of earnings up to $6,920 in earnings and a maximum credit of $529. The credit begins to be phased down when earnings reach $8,650 at a rate of 7.65 percent. The credit falls to zero value when earnings hit $15,570.
The EITC has two aspects. The nonrefundable portion of the credit offsets federal income tax liability; it represents the amount of income tax the tax filer would ordinarily owe but does not pay because he is eligible for the EITC. By contrast, the “refundable” portion of the credit is provided when an eligible filer has no further income tax liability; it is a straight cash grant that is mailed once each year from the IRS to the tax filer. Some 69 percent of the EITC benefits provided to adults without dependents is in the form of “refundable” cash grants.
Who Gets the Credit? According to IRS data, some 4.67 million individuals without dependent children received the refundable EITC benefits in 2017 at a total annual cost of $1.44 billion. The average annual benefit was $308.
As Table 1 shows, median earnings among EITC beneficiaries without dependent children was $8,000; the median number of hours worked during the year was around 800. The median wage was around $10.00 per hour. The typical recipient had low earnings because he worked comparatively few hours during the year. Full-time work for a full year equals around 2,000 hours (40 hours per week times 50 weeks), but the childless EITC recipients work about a third of that amount. If the median recipient worked full-time during the year, he or she would have an annual post-tax income of around $17,700, or 33 percent above the poverty level.
A common perception is that the childless adult EITC credit goes to young adults. This is misleading. In fact, the median age of current recipients is 39.
More than 80 percent of EITC recipients without children reside in multi-person households. Only a fifth reside alone as single persons. Some 56 percent reside with nondependent relatives or married partners; around 8 percent cohabit with a domestic partner, and 17 percent live with nonrelative roommates.
Living with others boosts living conditions by, among other things, providing economies of scale. This is particularly true regarding rent; a two-bedroom apartment costs 25 percent more than a one-bedroom apartment. Thus, two persons sharing an apartment can cut housing costs by around 40 percent relative to two living separately.
Eligibility for the EITC for childless persons is determined on separate tax filer income, but childless EITC recipients often reside with relatives who have independent incomes. These relatives are not included in the tax filing/eligibility unit, but living together enables the recipient and relatives to share resources and raise their standard of living. Because family members and relatives living together generally pool income and expenditure, official government poverty statistics are always calculated on a family unit basis rather than on an individual (tax filer) basis.
The Census defines a family as “a group of two people or more…related by birth, marriage, or adoption and residing together.” Thus, a family would include the head of household and any other persons in the residence who are the spouse, parent, grandparent, child, grandchild, sibling, aunt, uncle, niece, nephew, cousin, or step-relative of the household head. By this definition, family units do not include cohabiting partners or nonrelated roommates.
Overall, childless EITC recipients have an average family unit income of $43,000 per year. (This includes single-person units.) The fact that the majority of EITC recipients without children reside with other adult relatives means that, despite their individual low work levels, over half of these recipients are not poor by official government measures. Overall, 55 percent of EITC recipients without dependents live in family units that are not poor, while 45 percent are poor. (This figure includes those that live alone.)
Poverty can also be measured at the “household” level. The Census defines a household as “all the people who occupy a housing unit.” This would include the recipient and any spouse, relative, cohabiting partner or nonrelated roommate residing with the recipient. (A person living alone is considered one household.) The average household income of childless EITC recipients is $57,600. Measured on the basis of household income, only 28 percent of these recipients are poor. The poverty rate for childless EITC recipients living with cohabiting partners is particularly low at 15 percent.
Fairness and the EITC. The EITC was created explicitly to give aid to low-wage parents who would have difficulty supporting their children without assistance. Oddly, this practical fact has now been transmuted into an argument that the EITC is biased in favor of parents and discriminates against workers without dependents to support.
Advocates routinely complain that the EITC is inequitable because it provides a single mother with three children a maximum benefit of $6,557, while a single adult with no children to support receives “only” $527. This argument ignores the obvious fact that the single mother receives roughly $6,000 more because she has three extra persons to feed, house, and clothe. Moreover, the single mother will generally be the sole or predominant economic support for her family, but that is not the case for EITC recipients without dependents.
Many full-time working parents truly need support from the EITC to raise their family’s income above poverty. For example, a single mother working full-time at the federal minimum wage needs support from both the EITC and food stamps to lift her family above the poverty level. This is rarely the case for single workers without dependents. A childless individual without dependents who works full-time at the federal minimum wage has a post-tax income (without the EITC) roughly equal to the federal poverty threshold for a single person living alone. Most low-skilled workers earn more than the federal minimum wage. Few live alone; most live in households with other members with incomes.
Marriage Penalties. The current EITC credit for persons without dependents has an explicit marriage penalty. If a couple is not married, they will be eligible for two credits with a combined maximum value of $1,058, but if they marry, they become eligible only for a single credit with a maximum value of $529.
Moreover, for a married couple, the EITC credit is based on the pair’s combined incomes rather than their separate incomes; typically, when a couple marries, their combined income makes them ineligible for the credit entirely. Although this loss is slightly mitigated by raising the credit phase-down point for married couples, the marriage penalty still remains explicit and obvious. For example, if a nonmarried, childless couple each earn $10,000 per year, they will receive combined EITC credits worth $852; however, if they marry, the credit will be cut to $105. (See Chart 2.)
Marriage penalties between parents are even more severe. In that case, the EITC for single persons is eliminated entirely, and if the couple earns more than $24,820, the overall EITC for family with children is cut as well. Overall, if parents with two children each earn $20,000, the family will lose around $6,000 in EITC and food stamps if they marry. Although liberals sometimes offer lip service in favor of married two-parent families, nearly all proposals to expand the EITC for singles without dependents would increase marriage penalties.
Fraudulent and Erroneous Benefits. Erroneous and fraudulent EITC overpayments are prevalent for all EITC recipients including those without dependent children. IRS audits reveal that between 30 percent and 38 percent of all EITC benefits payments to childless filers are erroneous overpayments. Contrary to common claims, erroneous overpayments in the EITC are not the result of the law’s complexity; instead, they result from tax filers submitting false information concerning residence, amount of earnings, and relationship to a child to obtain higher benefits. Some 85 percent of these overpayments goes to individuals who are completely ineligible for the EITC but obtain benefits by providing erroneous information on their tax form.
Efforts to Expand the EITC for Childless Persons
In recent years, there has been a push for a substantial increase in EITC benefits for persons without dependent children. According to advocates, this policy would increase incomes, reduce inequities between parents and nonparents, reduce poverty, and substantially increase employment. Economist John Scholz, for example, predicted that doubling the value of the EITC for persons without dependents would bring employment to over 700,000 people. Critically, the policy is viewed as especially beneficial to disadvantaged groups such as low-skilled men who have had very low wage growth for decades.
Advocates contend that the policy would bring social as well as economic benefits. The expanded EITC for persons without dependents, they assert, would “help address some of the challenges that less-educated young people (particularly young African American men) face, including low and falling labor-force participation rates, low marriage rates, and high incarceration rates.” As advocates explain, “increasing the return to work for childless workers will lower unemployment rates and achieve the dual social benefits of reducing incarceration rates and increasing marriage rates.” (The claim that the expansion policy would increase marriage in low-income communities is perplexing given the fact that the policy would explicitly increase penalties against marriage within the welfare system.) Finally, it is argued that increasing the EITC would benefit all of society by reducing crime.
In 2014, support for this idea gained momentum when the Obama Administration proposed that the maximum value of the EITC for childless workers be doubled from around $500 to $1,000. The benefit termination point (at which benefits were phased down to zero) would have been raised from $14,790 to over $18,000. If enacted, these changes would have increased the number of persons eligible for the credit from 7.7 million to 13.5 million.
The Administration suggested that the proposal could increase employment among the eligible population by 10 percent. It emphasized that the pro-employment impacts of the expanded credit would be particularly critical for “groups with low or declining labor force participation rates,” especially less educated and African American men. It also argued that the policy would increase marriage despite the fact that it explicitly increased marriage penalties.
In 2014, Congressman Paul Ryan, then Chairman of the House Budget Committee, issued a welfare reform blueprint that endorsed the Obama plan to expand the EITC for childless persons. Echoing Obama’s arguments, Ryan declared that the current EITC was inequitable for persons without dependents. He predicted that expanding the EITC for childless workers would substantially increase employment. Facing considerable criticism from conservatives, especially over the fact that his proposal would have explicitly increased existing penalties against parental marriage, Ryan dropped this idea from subsequent welfare proposals.
Pressure to expand the EITC for persons without dependents continues. The House Ways and Means Committee approved the Economic Mobility Act of 2019 (H.R. 3300), introduced by Congressman Richard Neal (D–MA), on June 20, 2019. The bill would nearly triple the EITC for persons without dependents, raising the maximum payment from the current $529 per year to $1,464.
Paycheck Plus Experiments
Political interest in expansion of the EITC has led to experiments called the Paycheck Plus demonstrations. These are randomized control trial (RCT) evaluations designed specifically to estimate the economic and behavioral effects of expanding the EITC for single adults without dependent children. The Paycheck Plus demonstrations were conducted by the Manpower Demonstration Research Corporation (MDRC) in New York City and Atlanta.
The features of the Paycheck Plus experimental policy are shown in Chart 3. The policy quadrupled the maximum value of the EITC for singles without children from around $500 to $2,000 per year. The range of eligibility was also greatly expanded as the benefit termination point (at which benefits were phased down to zero) was raised from $15,570 to $29,900. The minimum age for eligibility was dropped from 25 to 21. If implemented nationwide, these changes would roughly triple the number of persons eligible for the childless worker refundable benefits.
The New York City Paycheck Plus demonstration was begun in 2014. Some 6,000 individuals were recruited to participate. Individuals were eligible to participate if they were not married, had a valid Social Security number, were not planning to claim a dependent child in the subsequent year, had earnings below $30,000 in the prior year, and were not receiving or applying for disability benefits through Supplemental Security Income (SSI) or Old-Age, Survivors, and Disability Insurance (OASDI). The 6,000 were divided at random into experimental and control groups of approximately equal size.
In 2017, a duplicate evaluation was begun in Atlanta with a sample of 4,000. At present, the three-year evaluation of Paycheck Plus has been completed in New York City, and the first two years have been completed in the Atlanta program.
Impacts of Paycheck Plus
The actual results of the experiments were at best underwhelming. In New York, the MDRC evaluators issued a report stating that the program “modestly increased employment rates.” A smattering of other weak positive findings were reported. Nonetheless, The New York Times trumpeted these meager results in an article titled “How to Cut Poverty Now.”
Straight Talk on Evidence, an organization devoted to reviewing random assignment evaluations, panned MDRC’s report on the New York Paycheck Plus experiment as “Not Accurately Reported” and stated that “[f]ollowing an all-too-common pattern in the evaluation literature, it selectively reports positive study findings while omitting other key findings that cast real doubt on the program’s effectiveness as a poverty-fighting tool.” Straight Talk further reported that the Paycheck Plus authors cherry-picked a tiny number of positive findings, ignoring the much larger number of findings of no impact that in many cases contradicted the modest positive impacts reported. Moreover, the New York Paycheck Plus study reported nearly 300 outcome measurements covering different, often related variables for different sub-groups in different periods. With such a large number of measured outcomes, a substantial number of false positive effects can be generated by random chance.
The interim Atlanta report was even bleaker. Although participants in the experimental group did receive the enhanced EITC payments and had somewhat higher post-benefit incomes as a result, no other outcome was found. The program did not increase employment or earnings.
Specific Outcome Findings from Paycheck Plus
The following specific outcomes were reported from the New York and Atlanta experiments.
Employment. As noted, the Obama Administration suggested that doubling the maximum value of the EITC for childless workers could increase employment in the affected population by 10 percent. But the Paycheck Plus experiments showed that quadrupling the value of the credit produced an employment increase of only 1.8 percentage points in New York City and no statistically significant increase in Atlanta.
As Table 2 shows, over the three years the Paycheck Plus experiment ran in New York, the percentage of persons in the experimental group who had any employment during a year was, on average, 77.3 percent, while the number for those in the control group was 75.4 percent: a statistically significant difference of 1.8 percentage points. In Atlanta, there was no statistically significant difference in employment between the experimental and control groups.
Moreover, while advocates of expanding the childless EITC had focused their concern on lower-wage males as a disadvantaged group that for decades had suffered from flat wages and falling labor force participation, the New York experiment showed no increase in men’s employment. Nearly all of the small employment increase that did occur was among females. In Atlanta, the outcome was worse: Neither women nor men experienced an employment increase. In addition, none of the other disadvantaged target groups (previously incarcerated persons, noncustodial parents, and disadvantaged males) experienced increased employment in either city.
Earnings. A second expectation was that the enhanced EITC would lead to an increase in hours worked and earnings among those who were employed. This expectation was also frustrated. In both New York and Atlanta, the EITC increase generated no significant increase in earnings for any group including all persons, all women, all men, previously incarcerated persons, noncustodial parents, and disadvantaged men. (See Table 3.)
Income Including the Paycheck Plus Bonus. The Paycheck Plus experiment provided greater cash bonuses to employed people in the experimental group. In New York, this led to an increase in combined income (earnings plus bonuses) of $625 per year in the experimental group.
In Atlanta, there was an increase in combined income of $733 per person in the first year but no significant increase in the second year. To the extent that the Atlanta program increased combined income, it seems to have done so in advantaged rather than disadvantaged groups. For example, the program raised combined income for advantaged men but not for disadvantaged men. The Atlanta program failed to increase combined income for noncustodial parents but did increase it for those who were not noncustodial parents.
Poverty. The New York City experiment did not reduce the percentage of participants who were poor. It did, however, reduce the number who were in deep poverty (defined as those whose household incomes were less than 50 percent of the poverty threshold). Relative to the control group, some 3.4 percent of the experimental group were lifted out of deep poverty. For the most part, these individuals were shifted into the more moderate poverty group (those with incomes between 51 percent and 100 percent of poverty). Poverty and deep poverty impacts were not measured in the Atlanta evaluation.
Impacts on Formerly Incarcerated Persons. In New York, the Paycheck Plus program failed to increase employment, earnings, or after-bonus income among formerly incarcerated persons. In Atlanta, the program failed to increase employment and earnings among this group but did increase post-benefit income in one year out of two. In New York, the program did not reduce poverty for this group but did reduce deep poverty substantially.
In New York, the program was found to have zero impact on recidivism. In Atlanta, recidivism was not measured.
Child Support. Paycheck Plus provided weak to nonexistent results concerning the impact of the increased Paycheck Plus benefits on child support. In New York, there was a statistically significant increase in child support payments in one year out of three. In addition, there was a marginally significant effect in two other child support metrics but no effect in nine others. Overall in New York, out of a total of 12 outcome measures, the program had one statistically significant effect and two marginal effects.
In Atlanta, the Paycheck Plus program had no effect on child support, and the insignificant impact that did occur was at times reversed (experimentals paid slightly less than controls).
Other Outcomes. The New York evaluation found that the program had no effect on overall family income, physical health, mental health, happiness, material hardship, or hunger. The program had a slight negative effect on household formation as persons in the experimental group were less likely to live with a partner.
Nationwide Implementation of Paycheck Plus: Estimated Scope and Costs
As noted, the Paycheck Plus policy would quadruple the maximum value of the EITC. It would also raise the upper limit of eligibility from $15,570 to $29,900 and drop the minimum age of eligibility from 25 to 21.
In Table 4, we estimate the fiscal and demographic impact if Paycheck Plus were implemented nationwide. The table presents only the refundable or cash grant component of the program. We have estimated the figures in Table 4 and all the cost and caseload projections in this paper by applying the eligibility rules and benefits from the Paycheck Plus program to national demographic data from the Census Current Population Survey.
It is important to remember that, in contrast to current law, the Paycheck Plus policy was limited to nonmarried workers without dependent children. We follow this policy in Table 4; all married workers without children are excluded. The exclusion of married recipients means that the figures for recipients under current law in Table 4 differ from the figures in Table 1, which included married but childless recipients. For example, the number of current-law recipients of refundable credits in Table 4 is 3.5 million compared to the figure of 4.6 million in Table 1.
The exclusion of all married beneficiaries means that the real costs of implementing Paycheck Plus have been underestimated in Table 4. Real legislative proposals modeled along the lines of Paycheck Plus would almost certainly include a benefit expansion for married as well as nonmarried childless workers. This would raise the actual costs of the policy substantially.
As designed, the Paycheck Plus policy would more than triple the number of nonmarried childless adults receiving refundable EITC cash payments; nonmarried recipients would rise from the current level of around 3.5 million to 12.7 million. Roughly one in five nonmarried individuals without dependents between the ages of 21 and 64 in the whole nation would be eligible to receive the refundable cash payments.
The total cost of refundable EITC benefits to childless, nonmarried recipients would increase sixteenfold from around $1.1 billion per year to $17.4 billion per year. Approximately 30 percent of the added spending would come from raising the benefits of current beneficiaries, and the remainder would come from benefits paid to new recipients added by eligibility expansion.
The median number of hours that recipients worked during the year would rise from the current level of 720 to 1,300. This would occur not because of an actual increase in work performed but simply because expanded eligibility for the credit would cover millions of new individuals who already had greater work and earnings. Overall, the main impact of the credit would still be to compensate individuals for part-time or part-year employment.
The median annual earnings of nonmarried recipients would rise from $7,072 to $13,200. (Again, this is due solely to the expansion of the range of eligibility up the income scale; it does not reflect any rise in earnings for individuals, because Paycheck Plus did not raise earnings.) The average EITC benefit would rise from $315 to $1,365. Nearly all recipients would continue to reside with relatives or other individuals. The residential breakdown of recipients between living alone, with a cohabiting partner, with relatives, or with nonrelative roommates would be largely unchanged.
Under the policy, childless EITC recipients would typically reside in two-person homes. The median income before receipt of the EITC across all recipient households (including single-person households) would be $48,500. This is 231 percent of the poverty level adjusted for household size. The average household income would be much higher at $70,793 (or 317 percent of the poverty level), indicating that many recipients would reside in middle-income households.
Overall, only 21 percent of recipients without dependents would be poor before receipt of the EITC if measured on the official family unit basis. Measured on a household basis, only 13.6 percent of recipients would be poor before receiving the EITC. About one in eight recipients would live with a cohabiting partner; only 8 percent of these couples would be poor before receiving the EITC.
Benefits and Costs. As noted, the New York experiment raised employment by around 2 percent, and the Atlanta experiment had no statistically significant effect. Splitting the difference, the combined effect might be a 1 percent change. If the plan were implemented nationwide, this would imply an increase of around 200,000 jobs. At a net added cost of $16.3 billion in refundable-cash benefits, that translates into one extra job for each $80,000 in new spending.
The New York experiment also shifted 3.4 percent of the eligible population from deep poverty (family income below 50 percent of the poverty threshold) to moderate poverty (income between 51 percent and 100 percent of the relevant poverty threshold). If Paycheck Plus were implemented nationwide, that would be about 680,000 persons. At an added cost of $16.3 billion, that would represent a net cost of around $23,500 for each person shifted from deep poverty to moderate poverty.
Paycheck Plus Combined Refundable and Nonrefundable Credits. Appendix Table A2 shows the combined demographic and fiscal impacts of the refundable (cash grant) and nonrefundable (income tax relief) components of Paycheck Plus if the policy were implemented nationwide. The number of nonmarried beneficiaries rises from 12.7 million (persons receiving refundable benefits in Table 4) to 20.1 million. The annual budgetary impact rises to $29.2 billion, of which $19.6 billion would be cash grants and $9.6 billion would be income tax reduction.
The median wage of recipients remains at $10 per hour, but the median annual hours worked rises to 1,730 because of the expansion in the eligible population. Consequently, annual median earnings rise to $19,000, and median household income rises to $52,000.
Are Projected Costs Understated or Overstated? On the one hand, the cost and caseload projections in Table 4 are clearly understated because, following the Paycheck Plus specifications, they are restricted to nonmarried persons without dependent children. As noted, actual legislation would almost certainly include some expansion of eligibility and increase in benefits for married persons without dependent children as well. Altering Paycheck Plus to cover married couples without children would most likely to expand the program to cover some 5.3 million families, including some 3.3 million that are currently ineligible for the EITC. This would likely increase the total Paycheck Plus refundable costs by at least $5 billion.
On the other hand, some might argue that the increase in costs and beneficiaries shown in Table 4 overstates the effects of implementing Paycheck Plus because the EITC does not have a 100 percent participation rate: Not everyone eligible for the credit actually applies for and receives it. This point has some validity: The best estimates show that only 63 percent of individuals eligible for the EITC for childless persons actually receive it. However, fraudulent and erroneous payments are extensive throughout the EITC program. The best data indicate that for each eligible person who does not receive the childless credit, there is at least one noneligible person who does receive it. The number of nonparticipating eligibles is offset by a similar number of noneligible participants. The number of eligible persons therefore stands as a reasonable proxy for the number of actual recipients despite the apparently low participation rate.
Neither Paycheck Plus nor any similar proposals contain elements to reduce fraudulent and erroneous payments. Therefore, our estimates in Table 4 assume that the nonparticipation ratio (the percentage of eligible persons who receive the credit) and the erroneous payment ratio (the percentage of recipients who are actually ineligible) would continue at their current rates after the expansion. The number of eligible nonparticipants would still be approximately equal to the number of ineligible recipients. Thus, the proportionate expansions in eligibility and benefits embodied in Paycheck Plus can readily be used to project costs and caseloads that would occur under the policy even though the participation rate remained below 100 percent.
In fact, our estimates in Table 4 may well be significantly understated because increasing the dollar value of the EITC benefit may encourage a greater share of eligible people to apply for and receive the EITC, thereby raising the participation rate. If such an increase in the participation rate did occur, the figures in Table 4 would further underestimate recipients and costs under Paycheck Plus.
Labor Responses to EITC Benefit Increases. A number of policy experts have claimed that the EITC is extremely effective in increasing employment and reducing welfare dependence. They argue that during the 1990s, a preponderance of the strong decline in dependence, increase in single-mother employment, and drop in poverty was due to EITC expansion rather than welfare reform. According to this viewpoint, traditional welfare reform measures such as work requirements for able-bodied recipients are antiquated and unnecessary; all that is needed to increase employment, reduce poverty, and improve the lives of the poor is to entice more people into the labor force with higher EITC benefits.
However, a recent and very thorough study by Princeton economist Henrik Kleven makes this viewpoint implausible. Kleven examined the employment impacts of the creation of the EITC in 1975 and all subsequent changes in the federal EITC as well as all state EITCs. He found that nearly all of the employment increases and welfare declines in the 1990s were driven by welfare reform and the economy, not by the EITC:
[T]he EITC has not had any clear effects on labor supply…. Apart from the expansion enacted in 1993, EITC reforms are not associated with increases in the employment of single mothers relative to single women without children. The 1993 reform, on the other hand, is associated with very large employment increases, but these increases align closely with the confounding effects of welfare reform and a booming macroeconomy.
The Paycheck Plus experiments provide additional data on this topic. Experts touting the efficacy of the EITC in increasing employment had estimated that an increase in EITC benefit levels producing a 10 percent increase in net effective wages would increase labor force participation in the affected group by from 6.9 percent to 11.6 percent. The Paycheck Plus experiments suggest otherwise. In those experiments, the added EITC bonus was worth around 9 percent of wages, but this bonus led to an increase in employment of zero percent in Atlanta and only 1.8 percent in New York. The employment elasticity (change in percent employed over a percent change in net income) appears to have been between 0.0 and 0.2. This is roughly one-fifth of the magnitude of the prior estimates.
Of course, the earlier estimates concerned the labor force elasticity of single mothers; it is possible that the labor force elasticity of childless persons is far lower. But the lackluster employment response in Paycheck Plus reinforces Kleven’s argument that the historic employment-generating aspects of the EITC had been greatly overestimated.
Comparison to Guaranteed Minimum Income. Altogether, the evidence from Paycheck Plus and the best econometric studies suggests that the EITC does not substantially increase or decrease employment. In that respect, the EITC is a significant improvement on traditional welfare programs and alternate designs such as the negative income tax (NIT) and guaranteed minimum income. The impact of such programs on work was tested in a series of large-scale random-assignment controlled experiments in the 1970s called the negative income tax experiments. These experiments—which were conducted in Seattle, Washington; Denver, Colorado; Gary, Indiana; New Jersey; Pennsylvania; and rural areas in North Carolina and Iowa—showed that the benefits provided substantially reduced employment and earnings.
Although the goal of the experiments was to raise family income, the income gained through added welfare was offset to a considerable degree by the income lost through reduced employment. Each $1,000 in added benefits was offset by a $660 reduction in earnings. This meant that $3,000 in government benefits was required to cause a net increase of $1,000 in family income. Programs without a work requirement on able-bodied persons are very inefficient even at the limited goal of raising income.
Paycheck Plus and Disadvantaged Groups
A major problem for Paycheck Plus was its complete failure to increase employment among disadvantaged groups. Transitional jobs programs offer an alternative approach for this population. These programs provide temporary subsidized jobs and work-support payments to disadvantaged persons who otherwise have difficulty obtaining and sustaining employment. In complete contrast to Paycheck Plus, transitional jobs programs are most effective among the hardest-to-employ groups.
A recent review of random-assignment evaluations of 13 transitional jobs programs showed that all but one produced dramatic increases in employment during the first year of participation when a high share of enrollees were actively employed in subsidized jobs and receiving wage supports. Typically, the employment rates of individuals in the experimental group were 20 to 35 percentage points higher than those of individuals in the control group during that period. This may be contrasted with Paycheck Plus, which generated no gains in employment for disadvantaged workers, zero gains for workers overall in Atlanta, and a 1.8 percentage point increase for general workers in New York. In other words, transitional jobs programs are more effective in creating jobs for the hardest-to-employ workers than Paycheck Plus was for ordinary lower-wage workers.
Transitional jobs programs are particularly effective in creating employment for former prison inmates, a group that has perhaps the greatest difficulty in obtaining and sustaining employment. In contrast to Paycheck Plus, which had zero employment impact on this group, transitional jobs programs have regularly been shown to produce substantial employment impacts for this difficult-to-serve population.
For example, the RecycleForce program in Indianapolis placed former inmates in jobs in various social enterprises, including “an electronics recycling plant staffed by formerly incarcerated workers, who provided training and supervision to participants and served as their peer mentors.” During the first year of the experiment, 96 percent of persons in the experimental group were employed compared to 62 percent of those in the control group—a gain of 34 percentage points. The RecycleForce program produced a stable long-term increase in employment as well; in the last year of the evaluation, after all subsidized employment had ended, the employment rate for those in the experimentals remained 9 percentage points higher than the employment rate for the controls.
Table 5 compares the main impacts of Paycheck Plus and six transitional jobs programs for former inmates. Five out of six of the transitional jobs programs produced large, statistically significant increases in employment among former inmates, increasing employment by 20 to 34 percent points during the year that the transitional jobs were offered. By contrast, Paycheck Plus did not increase employment for former inmates in either city in any time period.
Typically, transitional jobs programs produced 20 to 30 times more employment among very hard-to-employ individuals than Paycheck Plus did among ordinary lower-wage workers. Although transitional jobs programs cost about five times as much per participant, they are much more cost-effective as employment-creating vehicles because of their higher job creation rates.
Each year, some 700,000 individuals are released from prison or jail. More than two-thirds of these will be reincarcerated within three years. Recidivism has enormous social costs through crime and human costs to the offenders themselves. Despite optimistic rhetoric, Paycheck Plus had zero effect on recidivism.
By contrast, some transitional jobs programs have shown signs of success in reducing recidivism. For example, the RecycleForce program in Indianapolis was found to reduce future arrests, convictions, and incarcerations by nearly a tenth over the 30-month evaluation period. On the other hand, many or most transitional jobs programs have no impact on recidivism. However, these programs could be improved by funding them on a payment-for-outcome basis linked to reducing recidivism.
The Paycheck Plus model shows us that expanding the Earned Income Tax Credit for childless adults would create a very broad, expensive program that would provide cash benefits to 12.7 million people at a cost of $17.4 billion per year. The overwhelming majority of 9.2 million new recipients added by the program would not be particularly disadvantaged, and few would be poor.
The childless adults who are actually poor are poor because they work relatively little during the year. The Paycheck Plus program did not substantially increase work or earnings among these individuals. As a consequence, the program did not reduce poverty at all and caused only a small decrease in deep poverty.
While the program would redistribute income widely, it would have at best a tiny impact on increasing employment and no impact on reducing social ills. In particular, the program would have no effect on the employment, earnings, or other outcomes for the primary disadvantaged target groups: low-wage males including black men, noncustodial parents, and released prison inmates. It would have no current impact on recidivism and no potential for improvements that could lead to future impacts.
If the goal is merely to enlarge the existing $1.1 trillion welfare state and redistribute income without positively changing behaviors, then increasing the EITC for adults without dependents is an appropriate policy. If the goal is to alter those behaviors that generate low incomes and social ills, then other options should be explored. The nation—and those that suffer without jobs like men who leave prison and want to rebuild their lives—deserve better.
Robert Rector is a Senior Research Fellow in Domestic Policy Studies, of the Institute for Family, Community, and Opportunity, at The Heritage Foundation. Jamie Bryan Hall is Research Fellow in Quantitative Analysis in Domestic Policy Studies. Noah Peterson was the Summer 2020 Graduate Fellow in Welfare Studies in Domestic Policy Studies.
Appendix 1: Sources of EITC Data
There are two principal sources for information about the Earned Income Tax Credit. The first is the Statistics of Income (SOI) data series published by the IRS, which includes tables with individual income tax return (Form 1040) statistics. These provide summarized data from the actual 1040 tax returns processed by the IRS. SOI Table 2.5 presents EITC data concerning number of tax filers receiving the EITC and the amount of benefits provided. These data are broken out by adjusted gross income and number of qualifying children. The table also separates EITC funding into refundable and nonrefundable portions of the credit.
While the published SOI data in Table 2.5 are accurate, they lack all data on the tax filer’s family structure, poverty, age, hours worked during the year, and hourly wage rate. Moreover, the published IRS data cannot be used to estimate the number of recipients and value of benefits if the EITC were expanded along the lines of Paycheck Plus or similar proposals.
The second data source is the U.S. Census Bureau’s Current Population Survey (CPS). In the CPS database, the Census estimates a family’s or single individual’s eligibility for the EITC based on family structure, number of children, and earnings and other income. The Census then imputes an EITC dollar benefit based on assessed eligibility. The assumptions behind this method are that all eligible units receive the correct EITC, that no ineligible units receive the credit, and that the earnings reported to the CPS are accurate or at least as accurate as those given to the IRS.
While the CPS does provide data on tax filers, it is based primarily on families and households. Therefore, the CPS can provide information on the types of family units receiving the EITC, the number of persons in the household not included in the tax filing unit, hours worked during the year, hourly wage rates, and age of the recipient. Unlike the published IRS data, the CPS data can be used to estimate the number of recipients and cost of benefits if the EITC were expanded along the lines of Paycheck Plus or similar proposals. For these reasons, most of the EITC data and estimates in this paper are based on CPS data.
There is however, one problem with the CPS data: The estimated number of EITC recipients and the total value of benefits reported in the CPS are far less than the actual number of recipients and benefits reported in the IRS SOI tables. The IRS numbers are based on actual tax returns and are correct; the CPS habitually underestimates overall EITC receipt and cost. EITC budget estimates using CPS data must normally be adjusted upward for the substantial underreporting in the CPS.
Fortunately, for the subset of EITC recipients without dependent children, the IRS and CPS figures are relatively similar. Upward adjustment of the CPS totals is unnecessary. (Further information on matching CPS and IRS data for childless workers is given below.)
A third internal Census Bureau database also exists. This database links IRS 1040 data with CPS demographics at the individual case level. In other words, it provides accurate demographic information of actual tax filers. This integrated database can be used to determine the share of apparently eligible individuals in the CPS who actually receive the EITC as well as the number of persons who received the credit by IRS records but who were ineligible according to the demographic data provided to the CPS. (These ineligible persons are typically individuals who did not reside with the child they claimed.) This internal data set is not widely available to the public.
The latest published information on the EITC using this integrated Census and IRS data set is provided in “The Antipoverty Impact of the EITC: New Estimates from Survey and Administrative Tax Records,” by Maggie R. Jones of the Census Bureau and James Ziliak of the University of Kentucky. Their data show that after linking the 2016 Census data and IRS tax records, the ratio of persons without children who are eligible for the EITC is 99 percent of the number of actual recipients according to IRS tax records. However, the number of persons who are eligible and actually receive the EITC benefit is only 63 percent of the total that are eligible. Thus, the take-up or participation rate for childless adults would be 63 percent. The same figures also indicate that some 37 percent of childless recipients of the EITC are not eligible for the credit. The number of eligible but nonparticipating persons is roughly equal to the number of ineligible persons who do receive the credit. Moreover, a number of adults who are theoretically eligible for the childless credit do not receive it because they have filed for the larger parental credit when in fact they are not eligible to make that claim.
A final data set that is available are IRS tax compliance reports, which use detailed individual audits to assess the number and cost of erroneous or fraudulent EITC payments. These audits show that between 30 percent and 38 percent of all EITC payments going to persons filing as childless adults are erroneous overpayments. Some 85 percent of these overpayment funds goes to individuals who are completely ineligible for the credit. These numbers correspond roughly to the payments to ineligible persons appearing in the Jones and Ziliak paper.
Adjusting CPS figures to Match IRS Totals
In the present study, eligibility and receipt of the EITC is imputed into the Current Population Survey data set through methods that are very similar to the standard Census methods. This imputation process yields an estimate of eligible childless beneficiaries and benefits for 2017 that is 6 percent to 7 percent higher than the number of actual childless beneficiaries and benefits reported in the IRS SOI tax records for that year. As noted, the excess number is due in part to the fact that a significant number of noncustodial parents who do not reside with their children and are not eligible to claim these children for EITC purposes do in fact claim the nonresident children and receive EITC benefits for them. These individuals appear as childless workers in the Census Bureau Current Population Survey but appear as qualified parents in the IRS data.
To avoid a potential overcount of childless workers, the number of childless recipients, the dollar value of benefits, and the average benefit per person figures at the top of Table 1 in this paper have been adjusted downward by 6 percent to 7 percent to match the published IRS figures exactly. These same figures have been adjusted downward in similar proportions for both the current-law and Paycheck Plus columns in Table 4 to minimize possible overcounts. Thus, all the projected Paycheck figures in Table 4 are rooted in the actual IRS numbers for current recipients and benefits. These downward adjustments are presented in Appendix Tables A1 and A2.
Appendix 2: Random Assignment Studies of Paycheck Plus and Transitional Jobs Programs
The following are the sources used for Table 5.
Transitional Jobs Programs
RecycleForce; Ready, Willing & Able; Workforce Solutions of Tarrant County. Danielle Cummings and Dan Bloom, Can Subsidized Employment Programs Help Disadvantaged Job Seekers? A Synthesis of Findings from Evaluations of 13 Programs, U.S. Department of Health and Human Services, Administration for Children and Families, Office of Planning, Research, and Evaluation, OPRE Report No. 2020-23, February 2020, pp. 48, 51, and 52, https://www.acf.hhs.gov/sites/default/files/opre/sted_final_synthesis_report_feb_2020.pdf (accessed October 22, 2020).
Center for Employment Opportunities (CEO). Table 4.3, “Opportunities Impacts on Employment, by Time Between Prison Release and Random Assignment, Center for Employment Opportunities,” in Cindy Redcross, Dan Bloom, Gilda Azurdia, Janine Zweig, and Nancy Pindus, Transitional Jobs for Ex-Prisoners: Implementation, Two-Year Impacts, and Costs of the Center for Employment Opportunities (CEO) Prisoner Reentry Program, Manpower Demonstration Research Corporation, August 2009, p. 57, and Table 4.1, “Impacts on Employment and Earnings, Center for Employment Opportunities,” in ibid., p. 48, https://www.mdrc.org/sites/default/files/full_592.pdf (accessed October 22, 2020).
Cindy Redcross, Megan Millenky, Timothy Rudd, and Valerie Levshin, More than a Job: Final Results from the Evaluation of the Center for Employment Opportunities (CEO) Transitional Jobs Program, U.S. Department of Health and Human Services, Administration for Children and Families, Office of Planning, Research, and Evaluation, OPRE Report No. 2011-18, January 2012, https://www.acf.hhs.gov/sites/default/files/opre/more_than_job.pdf (accessed October 22, 2020).
Transitional Jobs Reentry Demonstration (TJRD). Table 5.1, “One-Year Impacts on Employment and Earnings: Full Sample,” in Cindy Redcross, Dan Bloom, Erin Jacobs, Michelle Manno, Sara Muller-Ravett, Kristin Seefeldt, Jennifer Yahner, Alford A. Young, Jr., and Janine Zweig, Work After Prison: One-Year Findings from the Transitional Jobs Reentry Program, Manpower Demonstration Research Corporation, October 2010, pp. 70–71, https://www.mdrc.org/sites/default/files/full_615.pdf (accessed October 22, 2020).
Erin Jacobs, Returning to Work After Prison: Final Results from the Transitional Jobs Reentry Demonstration, Manpower Demonstration Research Corporation, May 2012, https://www.mdrc.org/sites/default/files/full_626.pdf (accessed October 22, 2020).
Milwaukee Safe Streets Prisoner Release Initiative (PRIP). Philip J. Cook, Songman Kang, Anthony A. Braga, Jens Ludwig, and Mallory E. O’Brien, “An Experimental Evaluation of a Comprehensive Employment-Oriented Prisoner Re-entry Program,” Journal of Quantitative Criminology, Vol. 31, No. 3 (September 2015), pp. 355–382, esp. p. 730.
EITC Expansion: Paycheck Plus
Cynthia Miller, Lawrence F. Katz, Gilda Azurdia, Adam Isen, and Caroline Schultz, Expanding the Earned Income Tax Credit for Workers Without Dependent Children: Interim Findings from the Paycheck Plus Demonstration in New York City, Manpower Demonstration Research Corporation, September 2017, pp. 47–48, https://www.mdrc.org/sites/default/files/PaycheckPlus_FR_2017.pdf (accessed October 21, 2020).
Appendix Table A.5, “Effects by Former Incarceration Status,” in Cynthia Miller, Lawrence F. Katz, Gilda Azurdia, Adam Isen, and Caroline Schultz, Boosting the Earned Income Tax Credit for Singles: Final Impact Findings from the Paycheck Plus Demonstration in New York City, Manpower Demonstration Research Corporation, September 2018, pp. 59–60, https://www.mdrc.org/sites/default/files/PaycheckPlus_FinalReport_0.pdf (accessed October 21, 2020).
Appendix Table A.7, “Effects by Incarceration Status Prior to Study Entry,” in Cynthia Miller Lawrence F. Katz, Edith Yang, Alexandra Bernardi, Adam Isen, and Kali Aloisi, A More Generous Earned Income Tax Credit For Singles: Interim Findings from the Paycheck Plus Demonstration in Atlanta, U.S. Department of Health and Human Services, Administration for Children and Families, Office of Planning, Research, and Evaluation, OPRE Report No. 2020-28, February 2020, p. 38, https://www.acf.hhs.gov/sites/default/files/opre/paycheck_plus_atlanta_interim_508_final.pdf (accessed October 21, 2020).