Journal Issue: The Next Generation of Antipoverty Policies Volume 17 Number 2 Fall 2007
Gordon L. Berlin
A Counterintuitive Response with a Radical Twist in Tax Policy
A bold and equitable strategy for reducing poverty would tackle the interrelated problems of men’s low wages, single parenthood, and a tax and transfer system that, by giving primacy to families with children, has unintentionally distorted incentives on the margin to work, marry, and have children. An earnings supplement for individuals could accomplish these goals by providing all low-wage workers aged twenty-one to fifty-four who work full time (thirty hours a week) with a payment approaching that of the current EITC but with a crucial, if radical, twist: payment would be based on an individual’s personal income, not joint or family income, and singles would be eligible for the supplement whether or not they have children and whether or not they marry, as would second earners in a married family receiving the existing family EITC.
The credit should fall somewhere between 50 and 100 percent of the current EITC payment for families with one child.46 For equity and simplicity, to illustrate the new credit’s properties let’s assume that the maximum credit is about 75 percent of the current onechild EITC. Thus, instead of the tiny $400 current annual maximum EITC payment for singles, every dollar of earnings under the new plan would be supplemented by 25 cents until earnings reached $7,800 (for a maximum credit amount of $1,950), after which the supplement would remain level until earnings reached $14,400, and it would then fall by 16 cents for every dollar of earnings until it was phased out entirely at $26,587. In effect, this policy would turn a $6 an hour job into a $7.50 an hour job. For someone who worked 1,500 hours over the course of a year, it would turn a below-poverty annual salary of $9,000 into an above-poverty salary of $11,250. Figure 4 offers another example of how this new EITC would benefit a single man working full time in a job that put him in the bottom tenth of earners. The average man in the bottom tenth of the earnings distribution earned just over $22,000 in 1979 but only around $18,000 in 2004; the new EITC would raise this to nearly $20,000.
To avoid penalizing children or couples who marry, the existing benefit structure and governing rules for the one- and two-child EITC would remain in place for the primary earner in a family with children. The second earner in a married-couple family would now qualify for the new individual’s credit without regard to the primary worker’s earnings, essentially giving married couples the option of filing taxes individually, at least with regard to the EITC. In addition, with income treated separately, some married families with joint income that exceeds current cutoffs would now qualify for both the family and this new individual EITC. Emulating the approach to taxation used in Canada and many European countries, this strategy eliminates the marriage penalties now embedded in tax policy (at least for low-income couples claiming these EITC benefits), while also reestablishing social policy parity between men and women, and between parents and childless individuals, by bolstering the earnings position of low-wage individuals.47
The benefits to second earners in married or cohabiting couples would be substantial. Couples with joint earnings would each be able to receive an EITC payment, with the principal earner qualifying for the current child benefit and the second earner qualifying for this new individual EITC payment (possibly subject to an income cap of 250 percent of poverty). As figure 5 illustrates, for a two-child family in which each parent earned $14,000 ($28,000 together), the total maximum EITC payment could provide an additional $6,350—$4,400 for the existing two-child credit due one spouse plus $1,950 for the other spouse’s individual credit. Note that in this example, without an income cap or a lower overall subsidy, the earnings supplement would continue, although at progressively lower amounts, until it was zeroed out when combined income reached $62,000.
At first blush, a subsidy that reached this far into the middle of the family income distribution would likely be politically unacceptable. But unless the earnings of families above the poverty line are subsidized to some degree, it will be very difficult to resolve the work and marriage disincentives imposed by current policy on families with similar earned income in the $20,000-and-up range that result from high—and behavior-distorting—cumulative marginal tax rates. Moreover, as the example makes clear, adopting this policy would require coordination with, and possibly rethinking, the $1,000 per child tax credit, which is now partly refundable if income exceeds $11,000 a year and is not phased out until income reaches $75,000 for a single parent and $110,000 for a married couple.48
Adult men and women who are not caring for children and who work full time but are still poor—a group that has been largely ignored despite their substantial work effort—would also benefit. Using data from the National Survey of America’s Families, Stephen Bell and Jerome Gallagher paint a portrait of twenty-five- to forty-nine-year-olds with incomes below 200 percent of the poverty line who are not disabled and who do not have children of their own living with them. More than three-quarters work, mostly full time, but their wages are low (90 percent make less than $10.30 an hour) and one-third have incomes below the poverty line for their family size, while another 27 percent are near poor. A quarter are noncustodial parents and a quarter live with a spouse. More than half are white men with a high school diploma or less. They receive little in the way of government assistance— only 6 percent get food stamps and only 5 percent, unemployment insurance benefits— yet they too have been hit hard by labor market changes over the past thirty years.49
The impact on poverty of an individually based EITC of the size proposed here would be certain, large, and immediate. Individuals who now work more than thirty hours a week and earn less than the threshold amount annually (whether married, cohabiting, or unattached) would receive an immediate supplement to help to restore earnings to pre-1973 levels, when the average high school graduate— or even a dropout—could support a family above the poverty line. Those working less than thirty hours a week, including second earners in two-parent households, would have an incentive to increase their work hours, further boosting income, promoting self-sufficiency, and reducing poverty. Finally, those not in the labor force would have added incentive to find a full-time job, which would substantially boost total income.
Reliable experimental evidence indicates that employment effects could be significant. Economists estimate that increasing the hourly wage of a low-income worker by 10 percent would boost employment between 2 and 10 percent.50 Under this proposal, a potential minimum-wage worker, who could earn about $875 for 170 hours of full-time work per month, would be eligible for a monthly supplement of about $200. This 20 percent increase in income would be expected to increase employment rates by 4 to 20 percent.51 Adding credence to these estimates, the make-work-pay experiments described above had similar employment, earnings, and income effects, albeit for a population of mostly single mothers. Data from all four years of follow-up from each of these programs (including the period after the supplement programs ended) indicate that employment increases ranged from 8 percent in New Hope to 19 percent among long-term welfare recipients in the Canadian Self-Sufficiency Project (SSP).52 Intriguingly, employment effects for the most disadvantaged (those with limited education, little previous work history, and long prior spells of welfare dependency) were very large—up 56 percent in the SSP, 48 percent in Minnesota, and 25 percent in New Hope.
Less reliable observational evidence suggests that an earnings supplement could also have small but significant secondary beneficial effects on crime and marriage. Men’s involvement in criminal activity might reasonably be expected to decline as their earnings rise and the opportunity cost of crime goes up. Similarly, higher earnings, together with the elimination of EITC-related tax and transfer penalties on marriage, might also lead to more co-parenting, cohabitation, and marriage. Although these secondary effects are somewhat speculative, an individually based EITC payment at the very least creates the necessary, if not sufficient, conditions to make an increase in marriage feasible when combined with a direct intervention to promote marriage of the kind proposed by Paul Amato and Rebecca Maynard in their article in this volume.53
Finally, by supplementing the earnings of single men in low-wage jobs and increasing their income, this plan would encourage more “on the books” work, while helping men meet their child support obligations. As in current law, single people who are parents and owe child support would have their EITC payment attached to pay their child support obligations.
Anticipating Unintended Consequences
Even as it addresses important gaps in current policy, this plan, like any tax and transfer policy reform, opens up the possibility of creating unintended consequences. Five such consequences are anticipated below.
What about the principle of “horizontal equity”? An important and long-standing principle of tax policy is horizontal equity: the tax system should treat all married couples with the same total family income similarly, regardless of the source of their income. For some families, this plan violates that principle: a family in which one spouse accounts for most of the household’s earnings would receive substantially less in benefits than a family in which the earnings of both spouses are similar. Ironically, current tax policy has the opposite effect, penalizing couples when both parents’ earnings are similar. The problem of unequal earners could be partly remedied by following the Canadian example, which allows the high-earning spouse to claim a credit for the not-working or lowearning spouse, a credit that declines in value as the spouse’s earnings rise (see note 47).
Would there be an incentive for some people to reduce the number of hours they work? In the EITC and Minnesota programs, second earners in families were likely to reduce their work hours, especially when their earnings were in the phase-down range. Both the New Hope program and the Canadian Self- Sufficiency Project substantially alleviated this problem by conditioning the payment of earnings supplements on full-time work of at least thirty hours a week. As findings from the two projects demonstrate, a full-time work requirement has important advantages: it lowers total costs, moves people closer to self-sufficiency, and limits reductions in hours worked among those already working. However, applying a full-time “hours rule” to second earners in two-parent households, especially those with young children, would effectively penalize some for making the legitimate decision to spend more time with their children. Several modifications might help alleviate these concerns, including establishing a twenty-hour work minimum for the second earner only, creating a fifty-hour combined minimum per week for both earners, or lowering the minimum hours rule for parents with children under age three. There is also some risk of a cutback in work effort among primary earners receiving the existing EITC for families, which does not include an hours requirement. That risk is minimal, however, and could be reduced further by imposing a combined hours requirement or by making the individual EITC less generous.
Would it be feasible to administer an hours requirement? To facilitate administration of an average thirty-hour-a-week work requirement (using a monthly, quarterly, or annual accounting period), employers would have to report hours worked in at least one of two ways. The simplest approach would be to require employers to report monthly or quarterly hours worked on the end-of-the-year W-2 forms now given to all employees for tax-filing purposes. Then, individuals would claim the EITC as they do now, when filing their annual tax returns. Alternatively, employers could report hours worked along with the quarterly earnings report they now submit for unemployment insurance purposes on every paid employee. Several states, including Minnesota, New Jersey, Oregon, Washington, and Wyoming, already require employers to report hours worked. The unemployment insurance system would then have to share these employer reports electronically with the IRS to ensure accurate calculation of the EITC. Given that it now takes employers about five months to file wage reports, this process would have to be speeded up. Requiring employers to report hours worked would also make it possible for the government to pay the EITC quarterly instead of annually at tax time and would enable states to use hours rather than earnings to more fairly and accurately determine eligibility for unemployment insurance, a factor in the system’s currently low coverage rates.
Wouldn’t private employers be less likely to increase wages? An increase in public sector wage subsidies would likely make private employers less likely to raise wages. To avoid this problem, policymakers should consider indexing the minimum wage to inflation, so that it provides a floor below which wages cannot fall as the EITC expands. The higher one sets the minimum wage, the lower the cost of an EITC expansion. However, the politics of the minimum wage and its targeting inefficiencies (particularly the problem of teenage workers in affluent households) may make policymakers reluctant to expand the minimum wage much beyond $7.25 an hour nationally (although California’s minimum wage will reach $8 an hour by 2008). A minimum wage increase could substantially boost wages at the bottom, but it would not address the problems of persistent poverty or of high marginal tax rates that undermine decisions about parenting and marriage. In addition, the political unpredictability of the minimum wage makes it an unreliable policy lever for supporting low-wage workers.
Wouldn’t young people be lured out of school? Young people between eighteen and twenty-five have to make a range of decisions about the present value of work as measured against the future value of additional schooling. The opportunity to get a generous earnings supplement might change that calculus somewhat. Starting the supplement at age twenty-one, however, reinforces the value of postsecondary education, and the thirty-hour work requirement limits the likelihood of making payments to teens and college students who often have other means of support, although it might require the use of a four-month accounting period to avoid paying supplements to college students in the summer. Waiting until age twenty-five, as the current EITC for singles does, may unnecessarily penalize young, non-college-bound workers who are supporting families.
Eligibility and Costs
The plan affects three distinct populations. The first group is single men and women (some have children who do not live with them and others may be cohabiting). The second group is second earners in married two-parent households. The third is married couples without children (some whose children are over age eighteen and others who are just starting life as a couple). According to the eligibility criteria—age twenty-one to fifty-four, thirty hours or more of work a week for at least twenty-six weeks in the past year (a proxy for a quarterly accounting period), individual income below $31,030— roughly 35 million additional people would be eligible to receive EITC payments (44 million who meet these criteria, minus the 9 million who already receive EITC payments, according to the Current Population Survey). 54 About 15 million of these newly eligible recipients would be unmarried singles; 11 million, married individuals with children under eighteen (7 million second-earner spouses and 3.8 million primary earners); and 9 million, married individuals without children under eighteen. Another 3 million, mostly single, individuals (some supporting children and some not) who now receive small EITC supplements would be eligible for a threefold increase in their EITC payment. 55 More than half of the newly eligible group is married. And somewhere between one-third and one-half are supporting children.
Two key questions remain. First, how much would this plan cost? Assuming no change in the work behavior of recipients, a rough cost estimate is $29 billion a year to supplement the earnings of existing full-time workers.56 Today 35 million people are not eligible for the existing child-based EITC but work thirty hours a week or more, are struggling economically, and have earnings low enough to make them eligible for this new credit. The expected average earnings supplement of $1,000 would substantially raise their income and reduce the number of people living in poor families. What about additional costs for people moving from unemployment or parttime work to full-time jobs? Roughly 14 million potentially eligible people now work less than thirty hours a week for twenty-six weeks, and 22 million more are not working at all. Assuming a 15 percent increase in the share of part-time or part-year workers moving to full-time work, and a 10 percent increase in the number of nonworkers who take a fulltime job because of the supplement, benefits would be paid to an additional 4 million people. 57 With an average supplement of $1,000, these increases would bring total costs to roughly $33 billion annually.
Second, can we afford it? Based on the size of the federal budget deficit and the enormity of the Social Security and Medicare obligations that the nation faces, the obvious answer would seem to be no. But it is also true that if the economy had continued growing at 1960s-like rates and if the distribution of income between rich and poor had been the same in 2006 as it was in 1975, then the poverty rate would be about half what it is today.58 Instead, growth was decidedly slower during the 1970s and early 1980s and when it began accelerating significantly thereafter, the main beneficiaries were the top 10 percent of the income distribution—especially the top 0.1 percent59—who are accumulating unprecedented wealth (roughly an extra $750 billion annually)60 that would, in 1975, have gone disproportionately to the bottom of the income distribution. Assuming, as most economists would, that there is nothing inherently superior, from a national investment and savings perspective, about the distribution of income in 2006 relative to 1975, a $29 billion to $33 billion tax on these very high earners that was used to supplement the income of low earners would make no discernible difference in the economic position of the United States. To put this number into perspective, it represents about one-third of the annual tax reduction for the top 1 percent of the income distribution as a result of the Jobs and Growth Tax Relief Reconciliation Act of 2003. In that case, the answer is yes, we can afford it. Still, $29 billion to $33 billion a year is a lot of money. There are less expensive alternatives. Total costs could be halved, to $15 billion, if the subsidy rate were reduced from 25 percent to 20 percent, and further reduced to $4 billion if the rate were reduced to 15 percent. But at these lower rates, average per-person subsidy amounts fall from $1,000 per year (at the 25 percent rate) to $720 (20 percent) and $437 (15 percent), significantly reducing the plan’s antipoverty benefits.61