1993. Wilson is James A. Collins professor of management and public policy at UCLA.

Copyright 1993 by The Brookings Institution. All rights reserved. Reprinted with permission.

Policy elites, whether liberal or conservative, usually explain the problems of the urban underclass in terms of wrong incentives. Liberals rail at the lack of benefits and opportunities afforded the underclass; conservatives, at the excess of benefits offered without corresponding obligations. Liberals blame crime on poverty and joblessness; conservatives, on insufficiently severe criminal sanctions. Invariably, the argument boils down to incentives -- rewards and penalties.

Ordinary citizens see the matter differently. Though acknowledging that the incentives may be poor -- education inadequate, jobs scarce, and the criminal justice system ineffective -- the public tends to stress the attitudes of the permanently poor and the habitually criminal, attitudes formed in the family (or increasingly, the nonfamily) and reinforced by the culture. Those attitudes -- which, I would emphasize, are not unique to the underclass, but are particularly destructive in their effect there -- are characterized this way: a belief in rights but not in responsibilities, an emphasis on "me" and a neglect of "we," a preference for immediate gratification over investments for the future, and an expectation that if one is lucky or clever enough one can get something for nothing.

Some of these attitudes may be the result of scarce legitimate jobs and abundant criminal opportunities. But if those factors were the whole story, the results of carefully evaluated efforts to change behavior by supplying jobs, providing training, or altering penalties would not be as discouraging as they are. And if the objective conditions of inner-city life were all that mattered, then we would not see such great differences in behavior among individuals and groups confronting similar conditions. Without denying the importance of incentives, I want to side with the view of the average citizen who believes that poor self-control and indifference to the rights of others arise in large part from poor family training.

The chief means by which every society induces its members to exercise a modicum of self-control and to assign a reasonable value to the preferences of others is the family. Developmental psychologists are in broad agreement that the parental practices most likely to achieve these goals involve a combination of affection and discipline such that the child's attachment to the parents is strong and the rules of everyday behavior are clearly understood andconsistently enforced. Fortunately, most babies are biologically eager for attachment and predisposed to socialization, and most parents love their children and invest without compensation in their rearing.

But individuals differ in the extent to which they have (or reveal) prosocial impulses, and so some children are difficult and some parents incompetent. Unfortunately, since temperament is to a significant degree under genetic control, there is an elevated probability that difficult children will be born to incompetent parents. Socialization failures, if uncorrected, can breed -- literally -- more failures. Crime runs in families, alcoholism runs in families, impulsive and sensation-seeking behavior runs in families.

Matters become worse if families cease to exist or are transformed into pseudo-families. Poverty, as Daniel Patrick Moynihan steadily and rightly reminds us, has now become a children's problem owing chiefly to the fact that an increased proportion of children live for long periods, sometimes their entire childhood, in mother-only families, a large fraction of which are also poor. A 1988 Department of Health and Human Services study found that at every income level save the very highest (over $50,000 a year), children living with never-married mothers were more likely than their counterparts in two-parent families to have been expelled or suspended from school, to display emotional problems, and to engage in antisocial behavior.

If the family is headed by a teen-age mother, the risks are even greater. Children of teen-age black mothers are less able to control their impulses, have a lower tolerance for frustration, are more likely to be hyperactive, have more difficulty adapting to school, and, if boys, are more likely to be hostile, assertive, and willful than children of older mothers.

Is there any way to come to the aid of a weakened family, specifically to help it instill in its children the habits of mind and character that will enable them to take better advantage of whatever opportunities they have?

In principle, the answer is yes. Certainly, that is the governing assumption of every religious and of many secular efforts to help social outcasts -- alcoholics, drug abusers, and school dropouts -- and those efforts have produced many successes. The problem arises in trying to imagine a program that by plan and in the hands of ordinary managers will achieve the necessary personal redemption for large numbers of people.

Efforts to do this on a large scale and by bureaucratic processes have not, on the whole, proved very successful. For example, efforts to rehabilitate large numbers of delinquents or criminals have met with more failures than successes. Though there are some success stories, no one should suppose that we know how to convert large numbers of 18-year-old delinquents into law-abiding citizens.

Nor can one take much hope from two recent occasions in which the military had an opportunity to test whether its intensive training and strict discipline could improve the prospects of difficult boys. The first experiment began in 1966, when Secretary of Defense Robert McNamara undertook Project 100,000 as a way of contributing to the War on Poverty, and ended in 1971. The second took place by accident during 1976-80, when the military enlistment test was misnormed and low-aptitude individuals were inadvertently recruited.

Investigators studied veterans of both these quasi-experiments, comparing their economic, educational, and family status with that of people of similarly low aptitude who had never served in the military. Although most of the low-aptitude veterans said that their military experience had been good for them, primarily because it taught them discipline and made them more mature, the belief did not correspond to reality.

Project 100,000 veterans were worse off than nonveterans in employment status, educational achievement, and income; misnormed veterans were no better off than nonveterans. All were more likely to be divorced that the nonveterans. As the authors of the study concluded, "The military doesn't appear to be a panacea for struggling youth."

Clearly, addressing the problems of temperamentally difficult or low-aptitude youth is not easily done if one waits until they have reached their teens or young adulthood years. What if the intervention could begin earlier? Evidence from early childhood programs is more encouraging but still fragmentary and in some cases inconsistent.

Scarcely any governmental program is more popular today than Head Start, but even its long-term effects are in doubt (no lasting effects on IQ have been found, and only a little evidence exists on Head Start's effects on such matters as pregnancy, welfare, and crime). The strongest evidence of long-term effects of preschool education comes from one (non-Head Start) program, the Perry Preschool in Ypsilanti, Mich., whose "graduates" were followed for many years. Compared with a control group, the Perry students were found to be less likely to drop out before finishing high school, less likely to go on welfare, more likely to be working after leaving school, and less likely to have been arrested.

Why did the Perry program (and possibly a few other model programs) do so well? One reason is that they were model programs conducted by capable people who had received intensive training and ample budgets. Another is that the Perry project was not limited to providing children with preschool experiences for 12.5 hours a week. It also involved an extensive program of home visits.

Short-term results have been reported by the Infant Health and Development Program, which provided intensive services to nearly 1,000 premature infants in eight cities. The infants, at risk for retardation, behavioral problems, and learning difficulties, were randomly assigned to treatment and control groups. Infants and parents in the experimental group received weekly (later biweekly) home visits by trained counselors.

After their first birthday the babies attended child development centers five days a week. And parents attended biweekly meetings at which they were given information and could share experiences. At age three, children in the experimental group had higher IQs and fewer behavioral problems than those in the control group. Gains were greatest for children who had the most disadvantaged mothers.

Suppose that the long-term results from the premature baby project parallel those of the Perry Project. What lessons can we infer? The most obvious and (to some) perhaps the most troubling is that intervention programs produce more benefits the more deeply they intervene. For at-risk children, the more the programs either assume parental functions or alter the behavior of parents, the greater the benefit to the child.

Another possibility, albeit one that has as yet only fragmentary evidence, is that long-lasting interventions are likely to make more of a difference than short-term ones. Children cannot be inoculated against behavioral problems as they are against smallpox. Yet beginning at age five or six the only intervention program aimed at children and generally under government control is the school.

The central role of the school has led Americans to focus their hopes for character formation on it, hopes that receive some support from studies suggesting that the schools doing the best job of educating children are also those that do the best job of controlling their behavior. But even the best schools rarely occupy a child more than six hours a day, half the days of the year. And even the best school cannot offset the threats of disorderly streets, the neglect of absent parents, or the discord of unhappy homes.

Families with the necessary financial resources have always had an important way of coping with hard-to-socialize children or of escaping theirresponsibility for socializing them -- the boarding school. Families of lesser means have heretofore had no such option for their at-risk children. Could public resources be used to enable families in underclass neighborhoods voluntarily to enroll their children, beginning at an early age, in boarding schools?

I have in mind no single model of what such a boarding school would do, only a set of guidelines. In the elementary years, a boarding school would simply extend the number of hours the child was under school rather than parental supervision. School might become an all-day affair, with the child given breakfast and dinner and supervised after-class play opportunities in addition to regular instruction. In the extreme case of a child with no competent parent at all, sleeping quarters would be provided. As the child got older, say in the junior high years, it would become a full-time boarding school with home visits arranged by mutual agreement.

The schools could be operated by private as well as by public agencies. Enrollment would be voluntary but encouraged for at-risk children. The primary object would be to provide a safe, consistent, and enjoyable mechanism for the habituation of the child -- that is, for the inculcation of the ordinary virtues of politeness, self-control, and social skills. Another goal for these schools would be either to place their students into a college or to qualify them for entry into an occupation by means of an apprenticeship program.

Instead of leaving matters entirely to chance or voluntary participation in a boarding school program, perhaps better homes could be supplied in other ways. Suppose that unmarried mothers seeking welfare were given a choice: as a condition of receiving financial aid, they must either live with their parents or in group homes where they would be instructed in child care, receive a regular education, and conform to rules governing personal conduct and group responsibilities. The key elements in this idea are threefold. Do not allow welfare to be used for subsidizing independent but dysfunctional households; do not require the mothers of small children to work outside the home; and provide the best and most structured start in life for the next generation of children.

Boarding schools may be especially important for boys growing up in a fatherless family. Neighborhood standards and values may be set by mothers, but they are enforced by fathers, or at least by adult males. Neighborhoods without fathers are neighborhoods without men able and willing to confront errant youth, chase threatening gangs, and reproach delinquent fathers.

I do not know any way of requiring this generation of errant fathers to take up their responsibilities. The reach of the law has been lengthened, but we should not be optimistic that this will result in more than a modest increase in the size of the family support payments received by some mothers. Our chief goal ought to be reducing the number of errant fathers produced by the next generation -- that is, increasing the number of young urban males who marry and remain married.

Of all the institutions through which people may pass -- schools, employers, the military -- marriage has the largest effect. For every race and at every age, married men live longer than unmarried ones and have lower rates of homicide, suicide, accidents, and mental illness. Crime rates are lower for married than unmarried men, and incomes are higher.

Marriage not only involves screening people for their capacity for self-control, it also provides inducements -- the need to support mate, care for a child, and maintain a home -- that increase the capacity.

In the past, the institutions that have produced effective male socialization have been private. Today we expect government programs to accomplish what families, villages, and churches once did. I think we will be disappointed. Government programs, whether aimed at farmers, professors, or welfare mothers, tend to produce dependence, not self-reliance. Our policy ought to be to identify, evaluate, and encourage those local, private efforts that seem to do the best job at reducing drug abuse, inducing lovers to marry, persuading fathers to take responsibility for their children, and exercising informal social control over neighborhood streets.

The federal government is a powerful but clumsy giant, not very adept at identifying, evaluating, and encouraging. What it is good at is passing laws.

There may be a better way. Public funds might be sent to private organizations that in turn do the identifying, evaluating, and encouraging, all on the basis of carefully negotiated charters that free these intermediaries from most governmental constraints. I know of no example, but people who wish to think seriously about changing the culture of poverty had better start inventing one.