The Hidden Paradox of Welfare Reform

When Bill Clinton first sought the presidency, he promised to "end welfare as we know it." Instead of letting single mothers stay home until their children were fully grown, he argued that mothers who sought government help should go to work within two years. Polls showed overwhelming popular support for this change, but there were two big problems. First, some welfare recipients are only marginally employable. Second, welfare mothers who find jobs mostly earn between $5 and $7 an hour. Since that is not enough to support a family, they still need help paying their bills if they are to keep their families together.

To solve these problems, Clinton made three proposals. First, he pushed through a big increase in the earned income tax credit (EITC). If a single mother with two children earns $5 an hour and works 35 hours a week throughout the year, she now gets a refundable tax credit of $3,556—an extra $1.95 for every hour she has worked. Next, Clinton asked Congress to make all employers offer their workers health insurance. Congress rejected that proposal. Finally, Clinton proposed that the government should serve as an employer of last resort for mothers who could not find any other job after two years on welfare. That proposal died after the Republicans won control of Congress.

The Personal Responsibility and Work Opportunity Reconciliation Act, which the Republican Congress passed and Clinton signed last summer, assumes that raising the minimum wage to $5.15 and expanding the EITC will suffice to ensure that all working mothers can pay their bills. The law deals with mothers whom nobody wants to hire by allowing states to exempt up to 20 percent of their caseload from its five-year lifetime limit on welfare receipt. These solutions reflect the usual triumph of hope over experience.

The first thing to recognize about the new law is that its most important feature is not the time limits that have gotten the bulk of media attention but the transfer of power from Washington to the states. The new law replaces Aid to Families with Dependent Children (AFDC) with a program called Temporary Assistance for Needy Families (TANF). Under AFDC, every dollar that a state appropriated was matched by $1 to $4 of federal money. (The matching formula was more generous for poorer states.) Under TANF, states get a federal block grant whose size is essentially fixed. If a state wants to spend more than it has in the past, either because the cost of living has risen, because it has more single mothers who need help, or because mothers who have hit their time limit cannot find work, it will now have to pay the full cost from its own funds. Conversely, if a state limits eligibility for TANF or cuts benefits, it can (within very broad limits) keep every dollar that it saves. Instead of paying 20 to 50 cents for a dollar's worth of charity, states now have to pay a full dollar.

Raising the cost of altruism almost always reduces its frequency. Once legislators digest the fact that cutting TANF outlays by a dollar will add a full dollar to the funds available for more popular programs like schools, convention centers, or tax cuts, single mothers' share of their state's budget will almost inevitably shrink. Since the federal block grant will not rise with inflation, its real value will also shrink. To get out of the resulting fiscal bind, states will have to tighten their time limits, let benefits lag behind inflation, or both. More than a third of the states have already adopted time limits more restrictive than the ones Congress mandated. As the fiscal crunch intensifies, more and more states will presumably follow this path.

Subscribe to The American Prospect


Falling real benefits and tighter eligibility standards will obviously force more single mothers to seek work. No one knows how many will find it. In the past, most welfare mothers found work once their children no longer required constant supervision. (Benefits end when a mother's youngest child turns eighteen. Most mothers left the AFDC rolls before that happened.) TANF will force mothers to begin working sooner than they otherwise would have. But once the labor market adjusts to the influx of relatively unskilled women, most former recipients will probably find some kind of work.

Nonetheless, two problems are certain to arise. First, there will be a significant minority of mothers whom nobody wants to hire, even at the minimum wage. That has always been true among men, and among women without children. It will be true among single mothers too.

The second problem is that unskilled work is usually unstable work. A cost-conscious employer who can train people to fill a position in a few days has no incentive to keep them on the payroll when demand is slack, when they frequently miss work because of a sick child, or when their supervisor gets mad at them. Unskilled adults tend to be unemployed at least 10 percent of the time. During recessions, the figure is much higher. In poor inner-city neighborhoods and depressed rural areas, it is higher yet. As a result, a lot of unskilled mothers who work whenever they can will nonetheless be out of work fairly often. Some of these women are likely to hit their TANF time limit before their children are grown, especially if they live in a state that sets its lifetime limit at less than five years.

When that happens, welfare workers and legislators who care about the poor will probably push for more "flexible" time limits. That could be a political disaster. There is no reliable way to distinguish those who cannot find work from those who merely prefer not to work. As a result, "flexible" time limits will tend to make work requirements meaningless. Without work requirements, support for single mothers will dry up. Offering long-term support to single mothers who did not work made political sense in 1935, when Congress created AFDC, because married mothers seldom worked then. But that era is long gone. Now that most married mothers work, single mothers must do the same if they are to have a strong claim on public sympathy and assistance.

In theory, welfare mothers who are truly unemployable ought to qualify for federal disability benefits. In practice, many do not. There is no objective way of deciding whether applicants for disability benefits are capable of working. A "compassionate" system whose primary goal is to make sure that the unemployable always get benefits inevitably allows a lot of malingerers on the rolls, enraging their neighbors. A "tough-minded" system that excludes everyone who might be a malingerer inevitably excludes many other people who really cannot work. Since 1981, Washington has chosen to err on the side of tough-mindedness rather than compassion.

Given all these difficulties, liberals' best response to the fact that some mothers cannot find work is not to push for more flexible time limits but to revive the principle that the government should serve as an employer of last resort when mothers cannot find any other job. This is particularly important during recessions. Without a backup system of public employment, states will either have to fudge their time limits or let a lot of destitute families break up.



Even when unskilled single mothers find jobs, they seldom earn enough to support a family. Since a working mother earns far more than she gets on welfare, this may seem puzzling. In 1996, the average state gave a single mother with two children $4,668 in cash for the year. If she took a job paying $5 an hour and worked 35 hours a week, she could earn $9,100. She could also get $3,556 from the EITC, bringing her total income to $12,656. If a mother can survive on $4,668, how could she be worse off on $12,656?

One answer is that welfare mothers do not survive on $4,656 a year. A new book by Kathryn Edin and Laura Lein, called Making Ends Meet, provides the best currently available evidence about how working affects an unskilled single mother's standard of living. Edin and Lein studied mothers in four metropolitan areas (Boston, Chicago, San Antonio, and Charleston, South Carolina). Most of the interviews were done between 1990 and 1992. Roughly half the mothers received AFDC. The other half held low-wage jobs, most of which paid between $5 and $7 an hour. [Edin and I originally reported similar findings in the first issue of this magazine: Christopher Jencks and Kathryn Edin, "The Real Welfare Reform Problem," TAP, Spring 1990.]

Neither welfare nor low-wage work paid enough to cover these mothers' bills. Every mother had to supplement her income from these sources in some way. Some got extra money from relatives, boyfriends, or the absent father of their children. Some worked off the books. Some did both. Those on welfare had more "irregular" income than those who held regular jobs, but because low-wage work paid far more than welfare, the working mothers' total income was still considerably higher than the welfare mothers' income.

Nonetheless, the working mothers reported more material hardship than the welfare mothers. How can this be? The answer is simple. A single mother's expenses rise sharply when she takes a job:






  • Newly employed mothers usually have to pay for child care. (Most mothers who can get free child care are already working, either formally or informally.)
  • Most newly employed mothers have to pay for transportation to work. They also need to get their children to and from child care. So they usually need an automobile.
  • Because newly employed mothers have more cash income, their food stamp allotment is cut. After various deductions, each $100 of extra earnings brings a $24 reduction in food stamps.
  • If newly employed mothers have a federal housing subsidy, every extra $100 in cash income also brings a $30 rent increase.
  • In most cases taking a job means that a mother eventually loses her Medicaid benefits. In some cases her children also lose their coverage. Some newly employed mothers get health insurance at work, but many do not. Working mothers therefore have more out-of-pocket medical expenses.

Federal poverty statistics take no account of such changes. The Census Bureau just compares a family's cash income to a threshold based on the family's size and its members' ages. If a single mother with two children had a cash income of $12,514 in 1996, she was poor, even if she had subsidized housing, free medical care, and food stamps.

If her income was more than $12,514 she was not poor, no matter what her expenses were. It follows that moving a single mother from welfare to work usually cuts the poverty count. Whether it reduces material hardship is far less certain. The working mothers whom TANF will push into the labor force will differ in a number of ways from those whom Edin and Lein studied. In the early 1990s, when Edin and Lein did most of their interviewing, mothers who could not earn significantly more than the minimum wage, who could not find low-cost child care, or who had substantial medical bills almost always stayed on welfare. When TANF pushes such mothers into the labor force, they will earn less than the mothers whom Edin and Lein studied, and they will have substantially higher expenses. These disadvantages will be partially offset by the fact that the EITC is now considerably more generous than it was when Edin and Lein did their field work. But no matter how you do the arithmetic, the increase in the EITC is not going to be enough to solve working mothers' budgetary problems. Once TANF pushes the least resourceful and most troubled single mothers into the labor force, they will have even more difficulty paying their bills than they had when they were on welfare.

Judging by Edin and Lein's budget studies, a working mother who paid market prices for everything her family consumed would need at least $18,000 and probably closer to $20,000 a year to pay for what most Americans regard as basic necessities: food, rent, utilities, telephone service, second-hand furniture, clothing, medical care, and transportation to work and to a babysitter or child care center. Food stamps, housing subsidies, child care subsidies, and health care subsidies can, of course, reduce a working mother's need for cash, but they cannot reduce the cost of what she and her children consume. Assuming that former welfare recipients find jobs paying an average of $6 an hour, work 35 hours a week, and are unemployed 10 percent of the time—all fairly optimistic assumptions in my view—they will earn an average of $10,000 a year. The EITC will add another $3,600. That will leave a gap of something like $5,000.

Some mothers will fill this gap with money from the father of their children, their current boyfriend, or their relatives. But not all mothers have such resources. More diligent child support enforcement might help a few mothers bridge the gap, but most of the single mothers whom TANF will push into low-wage jobs have had children with men almost as poor as themselves. Pursuing these men may be good public policy, because it "sends a message" that men cannot avoid taking responsibility for their children simply by leaving home. But pursuing these men will not raise much money for the poorest single mothers. Indeed, it will make some of these mothers significantly worse off. Poor mothers often break up with the father of their children because he is physically abusive. Once the break comes, they want him out of their life. If a state agency forces an angry, abusive man to start paying child support, he may reassert his parental rights and begin harassing the mother again. Fanning these embers may not, in fact, be such a good policy.



If we want to make low-wage work a viable option for all single mothers, we need a governmental support system that gives working mothers enough help so they can pay their bills whenever they work full-time. The key elements of such a system have been around for a long time. All we need to do is make them more generous.

The Minimum Wage. The minimum wage will rise to $5.15 an hour in September 1997. That will put it at about 40 percent of what the average blue-collar worker earns in manufacturing. An easy way to reduce TANF's adverse impact on single mothers would be to push the minimum up to $6.50 an hour. That would be about 50 percent of the blue-collar average in manufacturing, which was the usual level of the minimum wage during the 1950s and 1960s. Such a floor would eliminate some jobs, but past research suggests that number would be quite small. The payoff for single mothers—who will constitute a much larger fraction of all minimum-wage workers once TANF's time limits begin to bite—would be quite large.

Child Care Subsidies. For unskilled single mothers with young children, the cost of child care is the most obvious obstacle to working. While some of these mothers get child care subsidies, not nearly enough money is available to subsidize all the mothers who will need help once TANF is fully effective. Expanding the number of subsidized slots would be expensive, but so is any scheme aimed at making work a viable option for unskilled mothers. The most cost-effective use of an unskilled mother's time is usually caring for her children, not serving burgers while someone else cares for her children. The rationale for putting these mothers to work is political and cultural, not economic. That means somebody has to pay the cost. The question is whether it will be the government or the poor.

Until the Gingrich revolution, even congressional Republicans understood this dilemma. Their alternative to Clinton's original welfare reform bill had tougher work requirements, but it also cost more. Since 1994, both Congress and the White House have been trying to tell themselves that if the government does nothing, the free market will generate affordable child care. This is wishful thinking. Once that fact becomes obvious, child care subsidies may turn out to be the most politically palatable way of ensuring that unskilled mothers can both work and pay their bills. Unlike other ways of subsidizing work, child care subsidies can be limited to parents with young children.

Federal Housing Subsidies. The single mothers who experience the greatest economic hardship when they take low-wage jobs are those in high-rent areas like New York, Boston, and parts of California. One way to help these mothers would be to alter the way we distribute federal housing subsidies. While almost all low-income families are theoretically eligible for a housing subsidy, only a lucky minority get one. Housing subsidies are distributed by local housing authorities, which screen applicants and put them on waiting lists. The length of the waiting list varies dramatically from one community to the next, depending on local housing costs and the number of low-income families. When a family gets a subsidy, its rent is set at 30 percent of its income. The government then makes up the difference between what the family pays and what its housing actually costs. The poorer the tenant, the more the government pays. Nearly half of all subsidies go to tenants in privately owned apartments. The rest go to tenants in public housing, some of which is awful and some of which is quite good.

If families that included a full-time worker went to the head of the waiting list for a subsidy, and if Congress set these families' rent at 25 rather than 30 percent of their income, single mothers would have more incentive to take low-wage jobs and would need far less help from other sources when they did so. Increasing the proportion of working adults in public housing would also reinforce the work ethic in the projects and might reduce the level of social pathology. And even if working mothers paid only 25 percent of their income, the cost to the government would be less than it is for a welfare recipient, so we could afford to subsidize more families.

Health Insurance. Allowing all low-income families to buy into Medicaid for, say, 10 percent of their earnings would go a long way toward making low-wage work an economically viable option. Better yet, we could revive the idea of universal health insurance.

Many different combinations of health care subsidies, housing subsidies, child care subsidies, and minimum wages would suffice to ensure that every single mother who found a job could make ends meet. We would not need to do all these things to make such a support system work. Legislators' choices among these options can and should be driven by political expediency. But we have to do something. The market cannot solve these mothers' problem.



Some conservatives oppose all efforts to help single mothers balance their budgets, on the grounds that making life easier for such women will make them more numerous. Most liberals scoff at this argument, but few really doubt that big changes in the economic consequences of single motherhood would affect its frequency. Imagine a society in which unmarried women knew that if they had a baby out of wedlock their family would turn them out, the baby's father would never contribute to its support, no employer would hire them, and the government would give them no help. Hardly anyone, liberal or conservative, doubts that unwed motherhood would be rarer in such a society than it is in the United States today.

Those who want to discourage unwed motherhood could not create such a world, even if they wanted to. They cannot prevent parents from helping out daughters who become single mothers, keep employers from hiring such women, or stop men from marrying them. They can mount "just say no" campaigns to discourage unwed motherhood, but there is not much evidence that these campaigns help. Legislators who want to discourage single motherhood therefore tend to focus on the one policy lever they really control: government assistance for single mothers and their children. The purists want to eliminate such programs. The pragmatists hope to have an effect simply by making such programs stingier.

Economists have tried to estimate the impact of this strategy by comparing the frequency of single motherhood in states with high and low AFDC benefits. At first glance, this strategy looks promising. In 1996, cash benefits for a family of three ranged from $650 a month in Vermont to $120 in Mississippi. If abolishing all cash benefits were going to have a major impact on the number of single mothers, they should already be scarce in Mississippi, since $120 a month does not buy much even in the nation's poorest state. In generous states like Vermont, in contrast, single mothers should be far more numerous.

Unfortunately for social science, state-to-state differences in welfare benefit levels are far smaller than they seem. To begin with, food stamp benefits fall as welfare benefits rise. Rents also tend to be higher in high-benefit states. A new book by Susan Mayer, What Money Can't Buy (Harvard University Press), reports that if two states' nominal benefits differed by $100 per month in 1990, welfare recipients' disposable income (after paying for food and rent) typically differed by only $30. The true difference in single mothers' living standards is thus relatively small. Edin and Lein's research shows the same thing. They found more hardship in low-benefit states, but not a lot more. Welfare recipients everywhere had to hustle to pay their bills, and the task was only marginally more difficult in low-benefit states. Thus one might not expect to find a lot more single mothers in generous states than in stingy ones.

In any event, state-to-state differences in AFDC benefits do not seem to have much impact on women's decisions about whether to have a child out of wedlock. Benefit levels appear to have somewhat more influence on the percentage of married couples who divorce and on the percentage of unmarried mothers who eventually marry. But even these effects are modest. The creation of food stamps and Medicaid may have played a role in the surge of single motherhood between 1965 and 1975, but so much else was happening at the time that there is no way to be sure about this.

The economic consequences of single motherhood vary more from country to country than from state to state. International comparisons therefore provide a useful check on the theory that the economic cost of single motherhood has little impact on its frequency. Using data from the Luxembourg Income Study (LIS) on 14 rich countries, sociologist Lee Rainwater provided me with estimates of mean income for different kinds of households with children under the age of 18. The estimates treat cohabiting couples as if they were married and adjust each household's income to take account of its size. The data were collected between 1989 and 1994.

To get a rough idea of how costly it was to become a single mother in each country, I compared the mean household income of households headed by a married (or cohabiting) couple to the mean for households headed by a single mother. This ratio overstates the true "marriage bonus," because it is not adjusted to take account of the fact that most married mothers have social and economic advantages that would enhance their income even if they were not married. But if my estimate of the marriage bonus is much bigger in, say, English-speaking countries than in Latin countries, it would be surprising if the true difference were zero.

The graph above ("Money, Marriage, and Motherhood") shows that among rich countries the marriage bonus is highest in Britain's former colonies: the United States, Australia, and Canada. It is lowest in Spain and Italy. If women's decisions about their living arrangements were shaped mainly by economic considerations, single motherhood should be relatively common in Spain and Italy, somewhat less common in the rest of Europe, and least common in the English-speaking world. Yet the graph shows precisely the opposite pattern. Single motherhood is more common in English-speaking countries, where its cost is relatively high. It is least common in countries where its cost is low.

The graph does not prove that economic incentives have a perverse influence on single motherhood. Rather, it suggests that both the frequency of single motherhood and its economic consequences depend on political, social, and cultural factors. The fact that the big English-speaking countries all penalize single motherhood but still have a lot of it cannot be pure coincidence. In Europe, the frequency of single motherhood seems to depend more on a country's distance from the Mediterranean than on the size of the marriage bonus. Both domestic and international evidence suggests, then, that the United States could do substantially more than it now does to help single mothers without appreciably increasing the percentage of women raising children alone. But evidence of this kind is unlikely to influence many legislators' votes. For conservatives, single motherhood is a moral rather than a practical problem. They do not want the government to reward what they see as bad behavior, because they think this undermines society's moral norms. Whether punishment actually deters bad behavior is largely irrelevant, just as it is to those who want to punish crime more severely.

Some American liberals and progressives are equally passionate in their defense of single motherhood. Many American feminists see divorce and unwed motherhood as sensible responses to the defects of American men. From their vantage point, having to depend on a potentially abusive or unreliable man looks even worse than having to depend on the government. Race also complicates the American debate. Marriage rates among African Americans are extremely low. Partly for that reason, some blacks now see marriage as a "white thing" that should not be imposed on people of color. Many white multiculturalists echo this view.

The "culture wars" between liberals and conservatives have probably left American children worse off than they would be if either liberalism or conservatism were hegemonic. Because such a large minority of Americans see single motherhood as a legitimate and sometimes even the most prudent available choice, our children are unusually likely to live in households without a male breadwinner. Because an equally vociferous (and somewhat larger) minority regard single motherhood as a menace, the American government does very little to help single parents. What really puts American children at risk thus seems to be our much-vaunted diversity. What children in every country need is consensus—any kind of consensus will do—about how they should be supported. But consensus is America's scarcest commodity.

If this argument is correct, the United States is unlikely to do much to improve single mothers' economic status in the foreseeable future. But while we may not do much, we may still do something. If TANF leads to a significant increase in material hardship among single mothers, and if this increase is not obscured by misleading statistics about income gains (or declines in the official poverty rate), public awareness that working mothers need more help will begin to grow. If these mothers are, in fact, doing everything they can to help themselves, they will probably inspire far more sympathy than today's welfare recipients do.

Constructing a support system for working mothers will take decades, not years. That is partly because the effort will have to begin at the state rather than the national level. Washington faces a financial crunch that is bound to get worse in the years ahead. The federal government's share of gross domestic product has been roughly constant since 1950, and there is no sign that tax increases will win significant support in the years ahead. With baby boomers retiring and new medical technology driving up the cost of health care, and with legislators unwilling to cross the elderly, Social Security and Medicare seem almost certain to claim a growing share of federal revenue. That means other federal programs will have to be cut, not expanded. Conservatives will concentrate on cutting programs for the poor, while liberals will try to cut defense. But no matter which party is in power, proposals for big new federal programs will get a frosty reception.

In the long run, all programs that involve significant redistribution from rich to poor have to be financed nationally. No state can afford to be substantially more generous that its neighbors, lest it become a magnet for the poor while driving away the affluent. But states can make a beginning, and once a few relatively liberal states have shown what might be done, ideas are much easier to sell in Washington. Moving in this direction depends on convincing middle-of-the road state legislators that working mothers need more help. That will not happen until legislators have seen a lot of stories about working mothers who cannot pay their rent and end up in shelters, or cannot afford child care and have to send their children to live with far-away relatives.

This is a grisly way of rebuilding support for compassion. Liberals helped erode that support by clinging to AFDC long after it had lost public legitimacy. Conservatives saw that this liberal folly offered them a chance to discredit liberalism and dismantle a big piece of the safety net. Having swept the 1994 elections and cowed Bill Clinton, the right carried the day. Now single mothers will have to pay for everyone else's sins.

You may also like