The first results are just beginning to trickle in from Washington's last venture in welfare reform, the Family Support Act of 1988, which sought to cut the welfare rolls by collecting more child support from absent fathers and giving single mothers more job training. Not surprisingly, evaluators are finding that modest expenditures on job training yield modest increases in welfare recipients' potential earnings. Faced with such unexciting news, and mindful of its many previous unsuccessful efforts to make America's welfare system more acceptable to the public, Congress now shows little enthusiasm for new legislation in this area.
Nonetheless, grass-roots hostility to the system has been driving both state legislators and presidential candidates to propose more drastic changes. Many states have trimmed benefits for parents receiving Aid to Families with Dependent Children (AFDC) for childless adults receiving General Assistance (GA), or both. Michigan has abolished GA altogether. New Jersey has decided not to increase AFDC recipients' benefits if they have another child while on welfare, and Wisconsin has moved in the same direction. Many other states have been considering changes in AFDC rules that they hope will encourage young women to stay in school, postpone motherhood, or get married. But by far the biggest proposed change is Bill Clinton's suggestion that no one be allowed to collect AFDC for more than two years.
If this proposal does nothing else, it should at least give pause to those who see Clinton as a middle-of-the-road Democrat committed to the status quo. A two-year time limit would be the biggest change in AFDC since Congress established the program in 1935. The average recipient now stays on the rolls 6.6 years. A two-year limit would, therefore, cut the number of recipients by more than 70 percent. At the peak of the last business cycle in 1989, something like 3.8 million families were collecting AFDC. With a two-year time limit, the figure would have been about 1.1 million.
A time limit on AFDC would also largely solve the problem that most worries conservatives these days, namely that AFDC makes recipients "dependent." So far as I know, there is no evidence that this is a real problem. Almost all mothers depend on someone for money: their husband, their employer, their parents, their boyfriend, or the government. All are in some sense "dependent" as a result. But it is far from obvious that depending on the government has significantly worse psychological effects than depending on anyone else. The elderly certainly don't complain about it much. It is true that welfare recipients are more likely than most people to be depressed, passive, and irresponsible. But that does not mean AFDC caused their problems. Many had such problems the first day they walked into a welfare office. Depression, passivity, and irresponsibility make it hard to find either a job or a husband. Without one or the other, a mother is pretty much stuck on welfare. Nonetheless, the doctrine that collecting welfare is bad for the recipients is so widely accepted that we have to accept it as a political fact, regardless of whether it is true.
Until now, conservatives who wanted to minimize welfare dependency have tried to keep benefits as low as possible and make collecting them as onerous and humiliating as possible. Clinton's proposal, which he borrowed from David Ellwood, a welfare expert at Harvard's Kennedy School of Government, would achieve this goal in a far more humane way, by putting a time limit on "dependency." If AFDC became a short-term program designed to help women who had just become single mothers achieve self-sufficiency, it might even be possible to set benefits at a realistic level, allowing recipients to support their families without having to cheat.
But what happens at the end of two years? It is tempting to pretend that if we just invested adequate resources, two years of intensive education and job training could make every single mother economically self-sufficient. But while that will surely be true for some, it will never be true for all. Unless we want another round of welfare "reform" that fails, we need to be realistic about the options open to us.
Contrary to what many liberals claim, the big obstacle to making single mothers economically self-sufficient is seldom the shortage of jobs. During recessions, of course, jobs are hard to find. But when the economy is healthy, minimum-wage jobs are relatively easy to find. The problem is that a minimum-wage job will not make a single mother economically self-sufficient.
There is endless controversy about how much money single mothers need to make ends meet. Conservative legislators and absent fathers seem to imagine that these families can live on air. Liberals have been somewhat more realistic, but not much. In large part this is because we take official income statistics too literally.
In 1990, for example, the official poverty line for a family of three was just over $10,000. Yet when the Census Bureau interviewed the heads of three-person families, one out of every ten reported an income below $10,000. Most observers, both liberal and conservative, conclude from this that families can survive on a sub-poverty budget. That conclusion is probably wrong. If we had accurate measures of what these families consumed, I believe we would find that very few got by with goods and services worth less than $10,000. The idea that millions of families somehow survive on minuscule budgets is, I believe, largely a product of poor measurement, not widespread frugality.
If you look at the Census Bureau's income data carefully, you find that nearly half the families of three with incomes below $10,000 in 1990 actually reported incomes of less than $5,000. Some reported no income at all, and some reported net losses (from a family business). Clearly these families were not living on their reported income alone. Some were living off savings. Some were only temporarily poor and were borrowing against future income. Some were not reporting off- the-books income. Some were single mothers whose budget deficit was being made up by a live-in boyfriend whose presence the mother probably did not report and whose income the Census Bureau would not have counted anyway, because he was not part of the mother's family.
Nonetheless, a lot of three-person families undoubtedly spent less than $10,000 in 1990. But when you look at the way such families live, you almost always find that they are also consuming quite a lot of things for which they do not pay cash. Some get free food stamps, housing subsidies, and medical care from the government. Some get free housing because they work as janitors, caretakers, or tenant farmers, or because they live with relatives. Some ride to work with a friend, or get free child care from a relative. No survey provides a full accounting of what poor families consume, but it seems highly unlikely that a family of three could consume less than $5,000 worth of goods and services over the course of a year. Most surely consume more than $10,000 worth.
Kathryn Edin's interviews with Chicago welfare mothers during 1989 and 1990 (see Christopher Jencks and Kathryn Edin, "The Real Welfare Problem," TAP, No. 1, Spring 1990) confirm this judgment. She found that while the typical mother got less than $5,000 a year in cash from AFDC, all supplemented their checks in various ways. Even if we ignore the cost of their free medical care, almost all were consuming goods and services worth at least $10,000 a year, and the average was around $12,000. Even at this level of consumption many families were desperately poor. Some had to skip meals or settle for rice and beans near the end of the month. Others had no heat in the winter or no telephone. If these mothers had been working regularly, their expenses would have been even higher, because they would have needed more and better clothes, more transportation, and more child care. When Edin went looking for single mothers who supported themselves entirely by working, she found few who spent less than $15,000. More recent work shows that the figure is even higher in Boston and only slightly lower in Charleston, South Carolina.
In order to earn $15,000, a woman who works forty hours a week must earn at least $7.50 an hour. The minimum wage is currently $4.25 an hour, and while many liberals want to raise it, there is no serious support for pushing it above $5 an hour. If you work regularly at $5-an-hour, you can earn about $10,000 a year before taxes and de- ductions. But most $5 an hour jobs involve a lot of layoffs and short weeks, so even if an unskilled single mother is trying to work regularly, she cannot count on doing so. We therefore have to assume that many welfare mothers will earn no more than $8,000 or $9,000 in a good year, and even less during recessions.
Census data confirm that few welfare mothers are likely to earn $15,000 a year in the labor market. Most are under the age of thirty-five, and only a few have attended college. Figure 1 shows what women with these characteristics typically earned if they worked full-time throughout 1990. High school graduates over the age of twenty-five averaged $17,000 a year, but nearly half earned less than $15,000. High school dropouts did far worse, as did younger women. Even these averages overestimate welfare mothers' potential earnings, because they cover a relatively select group of women those who were both willing and able to work full-time throughout the year. The women who were collecting welfare during 1990 would surely earn less than those who worked throughout the year, even if they had the same amount of schooling and were the same age.
If AFDC recipients all spent two years getting additional education and training, their potential earnings would rise, but here again we need to be realistic. The chart shows that high school graduates earned 20 to 25 percent more than high school dropouts in 1990, and women with some college eventually earned 20 to 25 percent more than high school graduates. But it does not follow that two years of training could boost AFDC recipients' potential earnings by 20 to 25 percent.
First, even two years of formal education have less impact than the chart implies, because women who get more education have other advantages as well. High school graduates come from more advantaged families than do dropouts, they are better at reading, math, and other subjects when they enter high school, and they are less like- ly to have been in trouble with the school authorities or the police. The same logic applies when we compare students who attended college for a couple of years to those who merely finished high school. While there is no consensus on how much of the apparent effect of schooling is due to this kind of selection, most estimates suggest that the figure is between a fifth and two-fifths.
A second reason why we cannot expect two years of training to boost AFDC recipients' earnings as much as two years of high school or college boosts others' earnings is that single mothers are not in a position to give as much time to their studies as the average high school or college student gives. Realistically, therefore, we should probably not expect two years of training to raise welfare recipients' earning power by more than 10 or 15 percent.
If the Clinton proposal is to be implemented, then, we need to find a way of closing the gap between what single mothers can earn and what they need to support their families. Otherwise we will end up with hundreds of thousands of homeless families instead of the tens of thousands we have now. Even worse, we will end up with more abandoned children, recreating the very problem we invented AFDC to solve in the 1930s. To avoid all this, we will have to provide single mothers who work in minimum-wage jobs with free medical care and at least $5,000 worth of other resources every year to supplement their wages. In years when the labor market is slack the figure will have to be even higher.
The Income Gap
Ideally, absent fathers should come up with this money. Even if the cost of pursuing absent fathers exceeds the amount collected which will certainly be so for "hard" cases harassing them will make other men more cautious about fathering children they cannot support. Until the 1960s a lot of young men used condoms because they knew that an unplanned pregnancy meant a shotgun marriage. Now that parents no longer pressure expectant couples to marry, we need some other way of encouraging men to use birth control. If they know that having a child out of wedlock will either cost them a lot of money or lead to a lot of grief when they cannot pay, condom sales should edge upward.
But the absent fathers of today's welfare mothers do not have anything like enough money to fill the entire gap between what single mothers can earn and what they need to support their children. If a mother of two is earning $9,000 a year and has no other help, an absent father would typically have to come up with at least $6,000 for her to balance her budget, even if she got free medical care. Child support orders typically require absent fathers to pay about 25 percent of their income. That means the father must earn at least $24,000 a year before his payments reach the required level.
Men with children who get AFDC are mostly young and poorly educated. In addition, many are black, and race affects men's earnings far more than it does women's. As Table 1 indicates, a large minority of these men did not work at all during 1990, and the percentage would be even higher if the table included men who were in jail, in mental hospitals, homeless, or missed by Census Bureau surveys.
The table also shows the mean earnings of the groups most likely to have fathered the children who are currently on welfare. In 1990 no group of twenty-five to thirty-four year-olds without some higher education averaged as much as $24,000. The best available data suggest that the typical absent father of a child on welfare would only have been expected to pay about $2,000 as of 1987. This would not be trivial, but it still leaves a $4,000 gap between what welfare mothers might get from work plus child support and what they need to pay their bills.
When single mothers with low-wage jobs do not have enough money to pay their bills, some can move in with their parents, other relatives, or friends. While these arrangements may well be quite good for children, sharing space tends to create a lot of conflict among adults. As a result, only a minority seem able to sustain such relationships over the long run. Instead, many single mothers with budget deficits end up sharing their bed and their bills with a boyfriend. These boyfriends seldom contribute a lot, but they usually contribute something.
Casual boyfriends help single mothers survive with less public support, but it does not follow that public policy should encourage single mothers to depend on them. Such a policy often pushes women into abusive relationships with men on whom nobody should have to depend. Encouraging single mothers to depend on casual boyfriends also leads to more unplanned pregnancies, more abortions, and more unwanted children. Indeed, such arrangements are often so unsatisfactory that single mothers prefer prostitution, which some find less emotionally demanding and which usually brings in more money.
If we want to give single mothers the option of not depending on men, I fear we must turn to the solution conservatives hate the most, namely using tax dollars to make up part of the difference between what unskilled single mothers can earn and what they need to support their families. The simplest way to do this is through the Earned Income Tax Credit (EITC), which gives employed parents with children a refundable credit if their earnings are low enough. At the moment, EITC is popular with legislators, because it only helps those who work. In addition, it lowers tax revenues instead of increasing expenditures, which looks good in the newspapers. But an EITC that could raise a single mother's income from $9,000 to $15,000 would inevitably arouse conservative ire, both because it would give a lot of money to people who sometimes spent it foolishly and because it would weaken recipients' incentive to work more hours.
The best way around this particular political problem is to supplement the EITC with programs that cut single mothers' expenses by paying for their children's medical care, food, and day care. Ideally, such programs should cover all children, regardless of income, so as to ensure that they win broad political support.
Even if we remain unwilling to provide universal health insurance, for example, we should at least make Medicare cover all children whose parents pay Social Security taxes. That would mean raising Social Security taxes, but it would also cut insurance costs for families that now insure their children privately. The only losers would be families with jobs and no children, who also happen to be the group that can best afford a modest tax increase.
We should also eliminate income tax deductions for dependent children and replace them with food stamps for children at all income levels. All families with children could get a year's worth of food stamps when they filed their tax return, minimizing auditing expenses. Such a system would eliminate the social stigma now attached to cashing food stamps, while increasing parents' incentive to work.
Finally, we could provide families in which all adults worked full-time with tax credits for, say, half the cost of enrolling their children in a preschool or after-school program that met certain minimal educational standards.
In other cases it makes more sense to direct programs primarily at low-income children. Historically, for example, federal housing funds went mainly to build large city-owned projects reserved entirely for the poor. Rents in these projects were limited to 30 percent of the tenant's reported income, which was often almost zero. This policy was, in effect, a way of bribing poor parents to live near other poor parents. That was a plus for more affluent parents, but it has had terrible effects on many of the children involved, who grow up in communities where crime is a way of life and no one takes school very seriously. More recently, we have moved away from city-owned projects and have begun subsidizing tenants in private housing, but in practice most of this housing is still located in very poor neighborhoods.
A child-oriented housing policy would concentrate on giving low- income parents strong incentives to live in better neighborhoods, where their children would be safer and would attend better schools. Chicago's Gautraux program, initiated as a result of a court order designed to eliminate racial segregation in the city's housing programs, has been doing this for many years and has yielded impressive benefits for children. The Gautraux program offers black mothers an apartment in a white suburb, and then makes up the difference between the apartment's market rent and what the federal government thinks she can afford to pay, namely 30 percent of her income. It is important to recognize, however, that a program of this kind will not work unless its benefits are restricted to those who are willing to move. If we offer poor black mothers a housing subsidy they can use wherever they want, past experiments with rent vouchers show that most will remain in their old neighborhood.
If we did all these things, any single mother who worked regularly would be able to make ends meet, even if she held only a minimum wage job. If she got child support, she could even take her children to a movie now and then. But these programs do not help single mothers who cannot find steady work. What are we to do for them?
So far as I can see, single mothers who cannot find jobs after their AFDC runs out have no more and no less claim on the public purse than married parents who cannot find work after their unemployment benefits run out. If a jobless single mother is mentally or physically ill, she should get disability benefits just like anyone else. If she cannot find work because the whole economy is in recession, her AFDC benefits should be extended, just as unemployment benefits usually are. If she cannot find work because she lives in a permanently depressed area, she may have to move, just as married couples do. I personally think a strong case could be made for having the government serve as an employer of last resort in certain cases, but I doubt that Congress will do this any time soon.
Reform at What Cost?
Setting up a program that allowed all single mothers to make ends meet would not be cheap. A two-year time limit on AFDC would save perhaps $18 billion a year, but there were 7.7 million families headed by women with children under the age of eighteen in 1990, and 4.3 million of them had incomes below $15,000. Raising all these families' incomes to $15,000 a year would cost about $33 billion, even if we did not give any benefits to anyone else. We should be able to get perhaps $5 billion of this from absent fathers, but even allowing for the $18 billion saved by the two-year limit on welfare, that leaves a $10 billion tab for taxpayers. If we want all these families to get goods and services worth $15,000 while at the same time ensuring that everyone has a significant incentive to take a better job, we also have to give some benefits to people who earn more than $15,000. By the time we are done, such a program could easily raise government expenditures by $30 billion to $50 billion. If we extend benefits to children in two-parent households the bill could be even higher, as it usually is in Europe.
Why should a program that asks more people to work end up demanding more money from taxpayers? We know that several million single mothers are currently getting by on low-wage jobs that pay less than $15,000 a year. Why can't we just insist that welfare mothers do the same thing? The answer is that single mothers with low-wage jobs currently survive by making arrangements that not all mothers are willing or able to make. One lives with her mother. Another has a boyfriend who beats her up but whom she does not throw out because he also helps pay the rent. A third sometimes works as a prostitute at the hotel where she cleans. A fourth leaves her chil- dren home alone after school because she cannot afford paid child care. But we cannot create a system that assumes all single mothers will make such arrangements. If we try, a lot more single mothers will be unable to make ends meet, and we will end up with more families in shelters and more abandoned children in foster care.
Yet as soon as we construct a system that allows a woman with a minimum-wage job to pay her bills without depending on anyone else, a lot more women will choose to exercise this option. The woman who now lives with her mother will move out, and the woman whose boyfriend beats her up will kick him out. The prostitute will turn fewer tricks, and the woman who works until five will get paid child care. Single mothers' lives will be a lot better, but there will probably be more of them, and taxes will certainly have to be slightly higher. In my judgment, a program of this kind would be worth the price, both because it would make children and single mothers better off and because it would do so in a way that is consistent with deeply held American values about work.
My rough guess is that creating such a system would require us to raise the tax burden from 34 to 35 percent of gross domestic product the equivalent of about one year's growth in personal income. Neither Clinton nor anyone else could easily raise such a sum in today's political climate. One must therefore ask whether America could make a more gradual transition from today's welfare system to a new system based on the expectation that women should work, spending a few billion dollars in the first year, slightly more the second year, and so forth. This is certainly feasible, but only if we keep the present welfare system intact until we have created an alternative for mothers who take low-wage jobs.
Indeed, Congress has been moving in this direction since the mid-1980s, gradually expanding benefits for the working poor while doing very little for welfare recipients. If we increase the Earned Income Tax Credit again, if we extend Medicaid to all poor children, if we expand federal support for child care, and if we were to create a new federal housing program especially for working single mothers, we could gradually make it possible to support a family on lower and lower wages. At that point we could impose Clinton's two-year time limit on AFDC.
But until we make it practical to support a family on a minimum-wage job, setting a time limit on AFDC will impose huge costs on many poor children and their mothers. That fact has defeated all previous efforts to abolish or drastically change AFDC, leaving us with a system very similar to the one the New Deal established in the 1930s. With growth rates low and resistance to new taxes high, improving the position of working single mothers is likely to take at least another decade of incremental change, and even that will suffice only if both liberals and conservatives recognize the long-term benefits of such a strategy. That means ma- jor changes in AFDC will also have to wait.
In the absence of radical reform, however, political discussions of welfare are likely to become steadily nastier. Everyone now agrees that we should encourage single mothers to work. Conservatives want to do this by cutting welfare benefits, passing rules, or both. Liberals want to do it by making work more rewarding. The conservative approach is unlikely to work, because middle-of-the-road legislators are seldom prepared to adopt policies that leave a lot of welfare recipients homeless and force a lot of mothers to abandon their children, and without that ultimate sanction, coercion cannot succeed. The liberal approach may work eventually, but it will not work any time soon, because it requires a lot of money that most Americans are currently unwilling to spend.
The effect of this impasse is a lot of grandstanding and name calling. Such partisan maneuvering creates a constant danger that both liberals and conservatives will lose sight of the ultimate goal, which is to create a system that rewards self-help. To achieve that end liberals must join conservatives in trying to make sure that mothers who work do better than mothers who do not. If liberals fail at that, as they have over the past half century, the public will continue to see welfare as a menace, and will continue to punish legislators who appear intent on making it more generous. If we create a system that rewards work, the politics of helping the poor would be completely transformed. Americans love to help people who are trying to help themselves.
Age No High Some High High School Some School School Graduate College 18-24 * $11,033 $13,385 $14,487 25-34 $11,832 $13,825 $17,026 $20,872 (* Sample size too small to yield reliable estimates)
Age and race No High Some High High School School School Graduate Percent of 25 to 34 year old men with no earnings in 1990 White 16.0% 6.8% 3.9% Hispanic 8.9 9.3 6.3 Black 52.1 24.0 9.4 Mean earnings of 25 to 34 year old men who worked White $12,837 $16,108 $22,312 Hispanic 12,344 14,808 17,861 Black -- 10,935 15,888 All 12,689 15,399 21,371 Percent of 25 to 34 year old men who worked full-time, year round White 49.5% 54.5% 72.1% Hispanic 55.9 58.5 66.0 Black -- 40.9 61.8 Mean earnings of 18 to 24 year olds All men who worked $8,800 $6,879 $11,186 All men who worked full-time, throughout 1990 12,710 14,613 15,829
The sample sizes in empty cells are too small to yield reliable estimates. The data do not cover men living in prisons or other institutions, members of the armed forces, the homeless, or men living in conventional households whose presence was not reported by the household head.