In a famous exchange between Hemingway and Fitzgerald, Fitzgerald is reputed to have said, "The rich are different from the rest of us," to which Hemingway replied, "Yes, I know, they have more money." Liberals have long contended that Hemingway had it right. There is nothing wrong with the poor that a little more money wouldn't cure. This view is, I believe, profoundly misguided. Money can alleviate the harsh conditions of poverty, but unless it is used to leverage changes in behavior, it will have little lasting effect.
Not only does behavior matter, it matters more than it used to. Growing gaps between rich and poor in recent decades have been exacerbated by a divergence in the behavior of the two groups. No feasible amount of income redistribution can make up for the fact that the rich are working and marrying as much or more than ever while the poor are doing just the reverse. Unless the poor adopt more mainstream behaviors, and public policies are designed to move them in this direction, economic divisions are likely to grow.
A tale of two families
In the 1990s, two journalists independently chronicled the lives of two inner-city families in Washington, D.C. Both journalists would eventually win Pulitzer Prizes for their reporting, but the portraits they painted could not have been more different. One of them, Leon Dash, a reporter with the Washington Post, followed the life history of Rosa Lee Cunningham and her family. At the time, Cunningham was a 52-year-old grandmother who had had her first child at age 14 and dropped out of school. The daughter of North Carolina sharecroppers, she grew up near Capitol Hill, and then supported herself by waiting tables, working as a prostitute, selling drugs, and shoplifting. She became addicted to heroin and spent time in prison for drug trafficking. She had eight children fathered by six different men and all but two of them became, like their mother, involved in drugs, crime, and teenage parenting.
Contrast this with another story of the inner city, told by Ron Suskind, a reporter with the Wall Street Journal. Suskind followed the life of a teenager named Cedric Jennings, who at the time lived with his mother in the same kind of inner-city neighborhood as Cunningham. But Cedric's mother, Barbara, had three children and had worked for 11 years at a five-dollar-an-hour job as a data-input clerk for the Department of Agriculture. She attended church regularly, lived frugally, supervised her children closely, and had instilled in her son a fierce desire to succeed. Cedric not only became an honor student at Ballou High School but eventually gained admittance to Brown University.
As these stories suggest, people living in poverty are a diverse group. Some are poor primarily because, like Cunningham, they persist in perverse and antisocial behavior. Others, like Jennings, have done the best they can with limited resources. Thus the two contending views of what causes poverty—people's own behavior or their adverse circumstances—will have some validity at least some of the time. Most poor people are neither as down and out as Cunningham nor as hard-working and dedicated to their children's success as Jennings. But what more systematic research shows is that behavior matters and must be taken into account if we are to reduce poverty and inequality.
Ideology vs. reality
My own involvement in this debate began in the late 1980s, when the Rockefeller Foundation established a program of research on what it called "the underclass." The underclass was commonly defined as those families living in areas of concentrated poverty, usually in neighborhoods where at least 40 percent of all households were poor. At the time, I was a scholar at the Urban Institute, a Washington think tank, and I suggested an alternative definition, one that was more behaviorally oriented. It was based on the idea that in order to achieve a middle-class life, an individual must do a few specific things: graduate from high school, defer having a baby until marriage, and obtain steady employment. With the help of several colleagues, I estimated the number of people who lived in neighborhoods where the basic norms of middle-class life had eroded to the point where a large fraction of residents had failed to do these three things. Our research showed that the underclass, thus defined, was still quite small, heavily concentrated in large urban areas, disproportionately made up of racial minorities, and, at the time, growing. I was especially concerned about the prospects of the children growing up in these environments where few men were working, most women were unmarried and on welfare, and dropping out of school was commonplace. Not only were most of the children in these neighborhoods living in poverty, but they lacked the kind of role models that would enable them to take advantage of the opportunities that did exist.
The reactions to this study were as interesting as the findings themselves. Some people, including Dash and Suskind, embraced my research for providing the kind of big picture in which their own stories could be embedded. A much more common reaction from scholars, however, was to suggest that talking about the culture of the underclass was tantamount to "blaming the victim." Bad behavior and poor choices, in this view, were an understandable adaptation to poverty and the lack of opportunity in people's lives. Although my research on the underclass was given a polite reception, most of the academic community has coalesced around the view that bad behaviors are a consequence, rather than a cause, of poverty. The result was that scholars continued to define the underclass simply in economic terms.
The reason for this reaction has more to do, I think, with ideology than with reality. Most academics, myself included, feel considerable sympathy for those who are poor or disadvantaged. We understand that none of us is perfect; and that while bad habits and poor discipline are widespread, they are more consequential for those living on the margin, where any slip-up may tumble someone over the edge. Moreover, children's starting points are very uneven. As a result liberals are wary of taking a judgmental stance, and fear that by "blaming the victim" they will undercut the political will to provide more resources to the poor. The problem with this mindset is that it requires avoiding or downplaying some unpleasant facts.
The belief that lack of income is the central problem of entrenched poverty also promotes continued emphasis on cash welfare, a strategy that has not succeeded. The average welfare check of roughly $400 a month (not including food stamps) may prevent destitution but is not enough to remove the typical three-person family from poverty. And even if it achieved this, it might still not improve the life chances of the children in these families, absent other changes. In her book What Money Can't Buy, Susan Mayer argued that the reason children in higher-income families do better in school and experience fewer behavioral problems than those in lower-income families is not because their families have more income but because they have better parents. In other words, adults who are more financially successful tend to have a variety of other characteristics that bestow advantages on their children. The reason this hasn't been widely recognized is because most studies (Mayer's is the exception) have not been able to adjust for all of these parental differences, some of which—like encouraging children to do well in school, providing them with a structured routine, reinforcing good behaviors, or simply having good genes—are very hard to measure.
None of this means that providing lower-income families with more money is necessarily a bad thing. But we should not pretend that money alone is going to change significantly the lives of these families, beyond easing a few hardships. The challenge is to find ways of providing generous support to the poor without disregarding the unpleasant facts about their behavior. Ideally, we need to nudge them toward a different set of behaviors by linking generous governmental assistance to staying in school, delaying childbearing, getting married, and working full-time.
The importance of work
What areas of behavior are we talking about? As I have suggested, three are critical. The first is education; the second is family formation; and the third is work. These have always been the sources of upward mobility in advanced democracies. Those who graduate from high school, wait until marriage to have children, limit the size of their families, and work full-time will not be poor.
This last statement may be surprising, but it is true. Every year the Census Bureau collects detailed information on a large and representative sample of all households in the United States. From these data, the government calculates the poverty rate for all households. In the year 2001, 12 percent of all households were found to live in poverty. The government's official poverty line in that year was $9,214 for a single individual living alone and $17,960 for a family of four. However, the poverty rate for those households where the primary wage-earner had finished high school, was married, had no more than two children, and worked full-time (2,000 hours a year or more) was trivially small—1 percent.
Of these three things, the most important is full-time work. To understand the importance of work alone in reducing poverty, perform the following thought-experiment. First take every family where there is at least one adult who is not too old or sick to work. Next assume that that adult is employed full-time at a wage commensurate with his or her education and experience. Under this assumption, almost half of those who are currently below the government's official poverty line would not be poor. The poverty rate would fall by a full five percentage points. Contrast this scenario with another, in which we double the amount of welfare benefits provided to America's poor. One might imagine that this would make them far better off. In fact, it does not. It reduces the poverty rate by only one percentage point.
Poverty in America is overwhelmingly associated with the failure to work on a full-time basis. Many immigrant families do well in the United States despite their lack of education because they tend to form stable families and work harder than many similarly disadvantaged native-born Americans. Yet these mobility-enhancing behaviors, and the attitudes that foster them, often disappear within a generation or two, suggesting that it may be aspects of American culture rather than economic stresses alone that hinder further progress. The more general conclusion is that attitudes and behavior matter. If you stay in school, work hard, marry, and have a reasonable number of children, you may still struggle financially, but you will certainly not be destitute.
Choice or opportunity?
This general conclusion is undermined, of course, to the extent that these behaviors are less a matter of choice than of opportunity. Not all parents encourage their children to do well in school; some adults can't find steady jobs; and some men and women have difficulty finding people to marry or end up in troubled marriages through no fault of their own. Even efforts to limit the size of one's family may fail-although repeated failures that lead to a much larger family than one can support are hard to excuse, since birth control is now widely available and highly effective.
The problem is that there are many gray areas where it's difficult to discern what circumstances are chosen and which are the result of a lack of opportunity. Take the issue of work. Liberals contend that people are unemployed or work too few hours because work isn't available. The fact that employment rates among the least skilled rise when overall economic conditions improve suggests that there is some merit to this argument. Yet the much more limited work effort on the part of the poor persists in good times as well as bad. Moreover, surveys of the jobless suggest that a lack of jobs is not the primary reason for their failure to find and maintain employment. In 1999, when the Census Bureau interviewed the heads of poor families about their failure to work in the preceding year, only 6 percent of women and 12 percent of men said they were unable to find work. The remainder cited reasons such as their obligations to school, job training, and family for their lack of employment. Clearly, many of the jobs available to those with limited education are low-paying and disagreeable. For these reasons, and because they often have access to other sources of income such as welfare, illegal earnings, or help from relatives, many of the unemployed are not interested in taking these jobs. Still others are depressed, addicted, or have prison records that make regular employment difficult if not impossible. Providing these hard-to-employ individuals with a welfare or disability check may be better than leaving them homeless, especially if they have children. But such cash assistance does not address their more basic behavioral problems. One unintended consequence of the tough welfare law enacted in 1996 is that it has forced welfare offices to grapple with these hard-to-help cases-cases that in the past could be conveniently forgotten once the application had been processed and the welfare check mailed.
In his book When Work Disappears, Harvard sociologist William Julius Wilson argued that joblessness was the source of most of the behavioral problems encountered in inner-city areas. The decline in well-paying manufacturing jobs for unskilled men combined with the seeming inability of inner-city schools to prepare children for the better-paying jobs in the new economy have, to be sure, had devastating effects. But, as he noted, these economic trends have interacted with cultural influences such as the paucity of positive role models in inner-city neighborhoods, and the corrosive effects of living in places where the discipline and routines of work have disappeared. Ghetto residents face tough obstacles, but they aren't helpless victims. Without greater efforts on their part, no government program-especially one of cash aid alone-is going to move them out of poverty and into the middle class. Liberals, in their eagerness to help, have been much too willing to ignore or excuse the cultural underpinnings of poverty.
What is needed is what some have called "tough love"-an insistence that people work combined with generous supports for people who make the effort. This means subsidized health care and child care for those in low-wage jobs. It means supplementing low wages when necessary and ensuring that the poor have opportunities to obtain the training and education needed to advance. For those who can't find private-sector jobs, for whatever reason, it means providing work in the community through which they can contribute something to society in return for financial support. Our experience with programs providing such "jobs of last resort" suggests that they are rarely needed, but that having them available makes work requirements more tenable and enforceable.
When the state of Wisconsin required that its welfare recipients work as a condition of aid, and offered to provide jobs to those unable to find them, most people were able to land jobs on their own. The state backed up its work requirement with subsidized child care and health care. Among families receiving welfare, caseloads in the state declined by 76 percent between 1995 and 2001. Although Wisconsin's success was especially notable, a similar pattern emerged in other states across the nation. A push from the welfare system along with a strong economy in the late 1990s produced a dramatic increase in employment and earnings among single mothers, the group most likely to wind up on welfare, and a drop in child poverty. It turns out that the number of involuntarily unemployed is much smaller than commonly believed, although that number clearly waxes and wanes with the condition of the economy.
The Aid to Families with Dependent Children (AFDC) system was created in the 1930s with the explicit purpose of enabling single parents to care for children in their own homes. The assumption was not only that single mothers shouldn't be expected to work, but also that no one becomes a single parent by choice. In those days, most of the women on AFDC were widows, and almost all children were born inside marriage. By the 1990s, this was no longer the case. Most modern-day single mothers on welfare have never been married. They typically have their first baby as a teenager and go onto welfare shortly thereafter. In the meantime, the vast majority of middle-class mothers, including those with young children, now work and pay the taxes that help to support these stay-at-home moms. No wonder welfare is unpopular with the public. President Clinton understood this when he pledged to "end welfare as we know it" in 1992. Egged on by Republicans in Congress, he signed the tough welfare-reform bill of 1996. Congress and the Bush administration are now attempting to reauthorize the law, and once again the debate is about whether, or how much, recipients should be required to work and whether resources should be devoted to encouraging marriage.
Many scholars continue to argue that marriage is an unrealistic goal for many of the poor. They point to the lack of employed men in low-income neighborhoods and cite the significant number that are in prison. But they also fail to stress the importance of finishing school and getting a steady job before taking on the responsibility of raising a child. Although a good job is no substitute for a second parent, it at least ensures that the child will not grow up in poverty. With the availability of birth control and legal abortions having made unplanned parenthood unnecessary, it is hard to understand why so many women are having babies that they cannot support. Many suspect the welfare system has been an enabling factor in these women's bad choices-choices that are further reinforced by the decline in social stigma.
As evidence of the benefits to children of growing up in a two-parent family has strengthened, liberals have become less likely to question the value of marriage. But many liberals still argue that marriage is no cure for poverty. About this, they are wrong. Adam Thomas, now a doctoral student at Harvard, and I recently conducted a statistical exercise in which we identified all of the single mothers interviewed by the Census Bureau and then matched them up with unmarried men of the same race, education, and age. If enough marriages had taken place to return the incidence of single parenting to 1970 levels, and the incomes of the men and women involved were combined, the poverty rate among children in 1998 would have fallen by about a third. (For certain subgroups of African-American women we did find a shortage of eligible men, some of which may be due to the difficulty the Census Bureau has in locating these men and some of which is due to the large number who have died at a young age or been incarcerated.)
The reasons for the decline in marriage remain unclear. Economic factors, such as the poor earnings prospects of less-educated men, cannot explain more than a modest portion of the rise in single parenting. Cultural changes have likely played a larger role. Acceptance of unwed childbearing has increased dramatically among the youngest generation. According to Dafoe and Whitehead, fully 50 percent of teenagers in 2000 accepted out-of-wedlock childbearing as a "worthwhile life style." My own view is that it is impossible to ever completely separate economic and cultural explanations. Behaviors that had their roots in conditions fostered by slavery and were later exacerbated by racial discrimination, the welfare system, and the poor inner-city job market made single parenting the most viable option in many communities. Now, through a process of cultural conditioning, these behaviors have become endemic and self-sustaining.
Though marriage has a number of important benefits, especially for children, it is not presently clear what can be done to encourage it. There is little evidence that we can convince people to marry by enrolling them in marriage-education or counseling programs. In some low-income minority communities, marriage is now so rare that one wonders what, if anything, could ever reverse the tide. Making tax and benefit programs more marriage-friendly may help, but the cost will be large and the benefits will almost certainly be small. Nonetheless, government programs should at least send the right message. Nongovernmental organizations, including those that are faith-based, also have a role to play.
A much shrewder, if less obvious, solution would be to redouble efforts to prevent early childbearing. As I have argued before in these pages, the recent emphasis on marriage in public debates about welfare and taxes has missed a crucial point-that the problem is not purely a lack of marriage. More than 90 percent of women have been married by the age of 45 (although the racial differences are large). The real problem is that women are having babies at a very young age, usually before they are even old enough to marry. Half of all childbearing outside marriage begins in the teenage years. Since teenage marriages are notoriously unstable, teenagers shouldn't be encouraged to marry. They should instead be encouraged to finish school, get a job, and defer having a child until they have found a lifetime partner with whom to share the responsibilities of parenting.
The good news about marriage is that it appears to be making a small comeback. After years in which the proportion of children being raised in a two-parent family declined, the trend leveled off and reversed slightly in the late 1990s. Although it is not clear what is driving this more favorable trend, I suspect that the new messages embedded in welfare reform are having an effect. To a young woman, the new system says, "Full-time motherhood at public expense is no longer an option," and to young men it says, "If you father a child, you will be expected to contribute to that child's support."
Even if these public messages begin to turn the tide, it will be a long time before out-of-wedlock childbearing drops substantially. One-third of all children in America are born outside of marriage, and the figure for African-American children now stands at a staggering two-thirds. Single parents will be with us for at least the short term. We should insist that they work, but we should also provide child care or early childhood education for their children. This will relieve single parents of an expense that their low-wage jobs make unaffordable. And if done right, it will help to break the cycle of poverty as well.
A behavioral divide
Policies that support work and marriage are not just an antidote to short-term poverty. They also have a role to play in preventing our country's economic and social divisions from growing even wider. The differences in attitudes toward work and marriage between rich and poor have been steadily increasing, with disturbing implications for the children of the poor.
Consider trends in employment. Historically, the poor have worked long hours to compensate for their lack of skills and low pay, and leisure has been the province of the rich. But, as Gary Burtless at the Brookings Institution has documented, in recent decades people at the bottom of the income distribution have worked fewer hours than those at the top. Since 1968, total hours of work have increased significantly among the entire working-age population, mainly because more women have entered the labor force. But the increases have been largest for those at the top, while the number of hours has actually dropped for those at the bottom. Hours of work increased slightly among men at the top but fell by one-third among adult men in the bottom income quintile. Women's hours increased for both income groups, but the gains were much larger at the top. These changes in hours worked are exacerbating preexisting income gaps between rich and poor. Changes in hourly rates of pay over this period have also favored the more advantaged, but this growing salary gap has been greatly amplified by a growing hours gap.
The story surrounding marriage is similar. Not only are fewer people marrying than in the past, but the trends have affected the haves and have-nots quite differently. Several years ago, I studied the family environments of American children under the age of six. I found that over the past two decades an increasing proportion were being born into a high-risk environment, defined as one in which the child's mother is a high school dropout, is not married, had her first child as a teen, and has a poverty-level income. Most of the children growing up in these circumstances will likely end up repeating the cycle of poverty. Many will do poorly in school, become unwed parents, and will have difficulty supporting themselves as adults. While some children will overcome the odds, their prospects are seriously compromised.
At the same time as more children were being born into these high-risk families, another large and growing segment was being raised in very privileged circumstances-that is, in families where the mother is a college graduate, is married, and had her first child in her mid-twenties or later. These families have on average four times as much income as a poverty-level family. In other words, as a result of changes in work and family patterns, today's children are getting a much less equal start in life than the children who were born a few decades ago.
Similarly, when David Ellwood and Christopher Jencks at Harvard University began looking at marriage and childbearing patterns among women with varying levels of education, they were surprised to find how much the trends had diverged. For women aged 25 to 34 in the bottom third of the education distribution, the percentage of single parents had risen from 7 percent in the mid 1960s to nearly 20 percent today. For comparably aged women in the top third of the education distribution, the percentages had barely changed from the 5 percent of the 1960s. Both well-educated and less well-educated women are delaying marriage (and in the case of many African-American women not marrying at all). The difference is that the less well-educated are not delaying childbearing, while the better-educated are. The result is very high rates of childbearing outside of marriage among less-educated women and low rates among more advantaged women.
This bifurcation in children's early environments implies much larger economic and social divisions in American society a few decades from now. These future divisions are not inevitable, but they are likely unless something happens to change the disparities in children's family environments.
Thinking about the future
Where does this evidence lead? It suggests we will become even more a divided nation of haves and havenots unless strong measures are undertaken to change these poverty-inducing behaviors at the bottom and ward off the damage they inflict on the next generation. These measures should not deny assistance to the poor, but should link assistance to a change in behavior.
Policies that are inconsistent with this premise must be discarded. Expanding the Earned Income Tax Credit (EITC), on the other hand, makes enormous sense. The EITC is an income supplement provided to low-income families with working parents. A parent earning $10,000 a year can qualify for as much as $4,000 a year in tax refunds. This is the equivalent of turning a $6-an-hour job into an $8.40-an-hour job. Unlike an increase in the minimum wage, it doesn't affect employers' labor costs and hence their willingness to hire low-skilled workers. And it has a powerful record of encouraging work. Single mothers, the group most affected by recent expansions in the EITC, have moved into the work force in droves. Their employment rate increased from 57 percent to 74 percent between 1992 and 2000. Unfortunately, the EITC's effects on marriage are not so benign. If a single mother marries another low-wage worker, she will lose most of this benefit. Congress took some steps toward reducing the EITC's marriage penalty in 2001, but more needs to be done.
Another perverse incentive is health insurance. As long as a mother is on welfare, she and her children are automatically eligible for Medicaid. But if she goes to work, she often loses coverage for herself and may also lose coverage for her children, depending on her income and the state in which she resides. Considerable progress has been made in recent years in extending health insurance to low-income children through Medicaid and the new State Children's Health Insurance Program enacted in 1997. Finishing the job should be a high priority. Finding affordable child care is another stumbling block for low-income parents. Some subsidies are available, and some mothers have friends and relatives who can care for their children. But not everyone is so fortunate. For many mothers, it makes little sense to take a job that provides, say, $12,000 per year in take-home earnings if they have to pay $5,000 per year for child care.
Whatever assistance is provided must be both generous enough and sufficiently tied to desirable behavior to be effective. One successful example is a Canadian experiment called the Self Sufficiency Program (SSP) that has operated in two provinces. Adults who agreed to leave welfare and work full-time were provided with wage supplements that effectively boosted their income by as much as $400 a month. The program has been carefully evaluated by the Manpower Demonstration Research Corporation in New York City, which found that when work-incentive programs are conditioned on full-time work, employment, earnings, and total income increase. In addition, the taxes paid on the increased earnings, coupled with decreased welfare payments, completely offset the cost of SSP's generous earnings supplements. This is a model "tough love" program. Some observers claim that a program like SSP could never be enacted in the United States, given our more conservative politics and less generous social welfare system. But in the early 1990s the state of Minnesota experimented with a similar, though less generous, program. This experiment was so successful that many facets of it have been adopted into law in that state.
My purpose is not to recommend any particular program. Rather, it is to stress the importance of aligning policy with what we know about the importance of certain behaviors in reducing poverty and inequality. There is both reason to applaud the steps that have been taken to require and reward work in the 1990s and room for experimentation with bolder measures. Not only will these "tough love" policies be more effective than the cash welfare policies of the past, but they will be more popular with the public. We cannot afford to allow American society to fragment into two cultures, one rich and one poor.