Lawrence Mead addresses the problem of nonwork among low-income men, particularly low-income black men, and its implications for families and children. The poor work effort, he says, appears to be caused partly by falling wages and other opportunity constraints but principally by an oppositional culture and a breakdown of work discipline. Mead argues that if government policies are to increase work among poor men, they must not merely improve wages and skills but enforce work in available jobs. Using the same "help with hassle" approach that welfare reform has used successfully to increase work among poor mothers, policymakers should adapt the child support enforcement and criminal justice systems so that both actively help their clients find employment and then back up that help with a requirement that they work. Men with unpaid child support judgments and parolees leaving prison would be told to get a job or pay up, as they are now. But if they did not, they would be remanded to a required work program where their efforts to work would be closely supervised. They would have to participate and get a private job and have their subsequent employment verified. Failing that, they would be assigned to work crews, where again compliance would be verified. Men who failed to participate and work steadily would--unless there were good cause--be sent back to the child support or parole authorities to be imprisoned. But men who complied would be freed from the work program after a year or two. They would then revert to the looser supervision practiced by the regular child support and parole systems. If their employment record deteriorated, they could again be remanded to the work program. Mead estimates that such a program would involve as many as 1.5 million men who are already in the child support and criminal justice systems and would cost $2.4 billion to $4.8 billion a year. It is premature, says Mead, for such a program to be mandated nationwide. Rather, the best role for national policy at this point is to establish and evaluate promising model programs to see which work best.