‘What Happened to Welfare?’

To the Editors:

The stakes are too high to let Christopher Jencks’s optimistic assessment of the 1996 welfare legislation go unchallenged [“What Happened to Welfare?” NYR, December 15, 2005]. In February of this year, building on popular perception that “welfare reform” has been successful, the Republican Congress narrowly passed a harsh new measure that will most certainly hurt the 1.9 million families left on the welfare rolls.

Continuing confusion about the consequences of the legislation is itself a result of political calculations. Back in 1996, the late Senator Paul Wellstone pushed for an amendment requiring states to keep track of “welfare leavers” to facilitate assessment of this social experiment. The amendment was rejected and social scientists have been forced ever since to play detective, trying to piece together what happened from indirect evidence. Unfortunately, Jencks misses one big clue and misinterprets another.

The big clue is that actual transfers to poor families have dropped very sharply between 1994 and 2003. Adding together cash benefits provided by AFDC and the successor program—TANF—with food stamps and the Earned Income Tax Credit, total transfers in constant dollars fell from $85 billion to $67.1 billion. This 20 percent decline in anti-poverty effort is substantial, and certainly has had an impact on low-income minority communities where welfare reliance has been high. In Louisiana, for example, AFDC/TANF benefits declined from $168 million for 245,000 recipients in 1994 to $57 million for 42,000 recipients in 2004. This withdrawal of assistance is implicated in the harsh poverty in New Orleans that was uncovered by Hurricane Katrina.

To his credit, Jencks recognizes that many of those welfare leavers who have not found work have fared poorly. But the big question is whether their losses have been offset by income gains among those who successfully moved from welfare to work. Most who have traded welfare for work are just about in the same place that they were before welfare reform; income from low-wage jobs just offsets declining benefits and the increased costs associated with work. Jencks’s positive assessment rests on the presumed existence of a sizable group of former welfare recipients whose earnings have allowed them to move above the poverty line. It is here that he misreads one of the clues.

Jencks focuses on the total population of mothers heading households with children under eighteen. He finds that while 44 percent of this population lived in poverty before “welfare reform,” that number dropped to 36 percent in 2004. But this population is a poor proxy for past and present welfare recipients.

The reason is that there has been a dramatic change in nonmarital births in the US. While the birth rate among teens has been dropping since 1994, white women in their twenties and thirties account for many more out-of-wedlock births. These are mostly not college-educated women, but they are employed in jobs that put them above the poverty line. Hence, the improvement for the population of single mothers that Jencks reports results from decisions by nonpoor women to have children without a partner. The nonpoor single-mother households rose by 28.5 percent between 1994 and 2004, while those in poverty fell by only 8.6 percent or 327,000 households. As a rough approximation, this suggests that maybe 9 percent of adults who were welfare recipients in 1994 are better off, while the rest are either the same or worse off. And Jencks agrees that there is no evidence of gains for the children.

But the changing composition of female-headed households has even larger implications. Analysts have long been divided between those who emphasize the causal role of structural factors such as resource shortages in perpetuating poverty and those who attribute causal importance to behaviors such as out-of-wedlock births and insufficient work discipline. The 1996 legislation was designed by those who believed that changing behavior is central to poverty reduction.

Hence, the actual story is deeply ironic. Starting well before the new legislation was implemented in 1997, teenagers changed their behavior resulting in significant declines in their childbearing. (This was probably the result of heightened HIV/AIDS awareness.) The 45 percent decline in the rate of teen childbearing among African-American women since 1991 is particularly impressive. Had teen childbearing remained at its 1994 level, there would have been an additional one million births to teen mothers during the decade, many born to minority mothers and at high risk for poverty. Meanwhile, the single parenthood of women in their twenties and thirties has had no impact on poverty rates.

Yet despite this major shift in behavior by teens, which began well before the new policies, the number of impoverished single mother families fell by only 8 percent. As people on the resource side of the debate have long emphasized, structural forces will push people into poverty regardless of behavior. Two conclusions seem inescapable. The right’s “war on bad behavior” is fundamentally misguided and the negative consequences of “welfare reform” on family well-being have outweighed any positive effects.

Fred Block

University of California, DavisThe Longview InstituteBerkeley, California

Christopher Jencks replies:

Fred Block is right about the changes in federal welfare rules that Congress passed last February. They are indeed harsh, and they are likely to make life worse for a lot of single mothers and children. There is a clear risk that cutting the welfare rolls, like deinstitutionalizing the mentally ill, will be pushed much too far. The fact that some mothers can get along without welfare does not mean everyone can, any more than the fact that some people who are mentally ill can get along without constant supervision meant that everyone who was mentally ill could do so.

Block is also right that declining poverty among single mothers is partly due to changes in the kinds of people who become single mothers. The official poverty rates for single mothers that I reported in my review of Jason DeParle’s book have numerous problems: they overestimate inflation since the poverty line was established; they ignore noncash benefits like food stamps and housing subsidies; they ignore taxes and the Earned Income Tax Credit; and they ignore the income of men who live in a mother’s household but are not married to her.

Joe Swingle, a sociologist at Wellesley College, has constructed a poverty measure that corrects these problems. These adjustments cut the official poverty rate among single mothers by roughly half (from 41.9 to 20.4 percent in 1996 and from (35.5 to 16.0 percent in 2003). Of the 4.4 point decline between 1996 and 2003, roughly a quarter is attributable to the fact that single mothers were older and whiter in 2003 than in 1996. Another quarter is attributable to the fact that single mothers were better educated. So Block is half right in his explanation for the declining poverty rate. But another quarter of the decline is due to the fact that single mothers were more likely to live with men and these men had more reported income. The last quarter is due to the fact that single mothers have more earnings.

His other points are less persuasive. The fact that transfers to the poor fell, for example, is hardly surprising given that poor people had more income from other sources. That was what was supposed to happen. The question is whether smaller transfers to the poor were offset by increases in their income from other sources. In some cases they were, and in some cases they were not. On balance, there was a small gain, as DeParle suggests.

I also agree with Block that the people who designed the 1996 legislation believed that changing behavior was central to reducing poverty. That strategy has clear limits, but it can be effective. Indeed, Block’s alternative explanation for the decline in poverty among single mothers also involves behavioral rather than the “structural” change to which he refers. According to Block, unprotected sex became less common among black teenagers because they did not want to die of AIDS. As a result births to black teenagers fell. Block may well be right about this.

The behavior that welfare reform was meant to encourage, however, was not chastity, contraception, or abortion but work. Work did increase, and this increase did play some role in the decline of poverty among single mothers. Employment among single mothers did not increase as much as the welfare rolls fell, and it didn’t raise incomes as much as some optimists expected. But the point of requiring mothers to work was never really to raise single mothers’ incomes, since raising their incomes would presumably have made single motherhood even more common. The point of welfare reform was to force single mothers to work for a living, just like other able-bodied adults. The country has been moving in that direction, and state legislatures have become more willing to give help to single mothers who work. If we ever get a more generous Congress and a more compassionate president, the federal government may also move in that direction.