Following dramatic declines in the welfare rolls, there's more good news on the anti-poverty front: The urban underclass seems to be breaking up.

The underclass, you may recall, was the concept popularized in the 1980s to describe inner-city neighborhoods plagued by high rates of unemployment, school dropouts, crime, out-of-wedlock births, and welfare dependence. Veteran poverty warriors hated the term, saying it smacked of "blaming the victims." Yet the idea that dysfunctional behavior and mores were perpetuating poverty and making it worse jibed with the moral intuition of most Americans. Social pathologies isolated poor communities from the mainstream.

By the early 1990s, the underclass phenomenon and the growing concentration of poverty in urban neighborhoods had provoked rising public angst about social welfare policies that failed to promote responsible behavior. President Clinton brilliantly crystallized the new consensus in his pledge to "end welfare as we know it."

Since Clinton signed the historic 1996 welfare reform bill, welfare rolls have been cut in half, and they stayed down even during the brief Bush recession. Poverty rates also have come down, while employment and earnings for single mothers are way up. Not every welfare recipient has found a job, and not every child is better off. But taken as a whole, the new regime of work and personal responsibility is producing far better results than the old system of cash assistance and dependence.

Now comes a University of Texas researcher with a striking finding: Concentrated poverty shrank dramatically in the 1990s. According to Paul Jargowsky, the number of people living in neighborhoods where the poverty rate is 40 percent or higher plunged by 24 percent, or 2.5 million people, in the 1990s. This marked an abrupt turnaround from the previous two decades, during which the population in these high-poverty neighborhoods doubled.

The underclass and high-poverty neighborhoods aren't entirely overlapping categories; the former is defined by behavior, the latter by income. But Jargowsky, in papers written for The Brookings Institution, also reports that the decline in the number of underclass neighborhoods was even greater -- about 33 percent, or about 1.2 million people.

The deconcentration of poverty affected all groups, but none more than blacks. The share of blacks living in high-poverty neighborhoods fell from 30 percent in 1990 to less than 19 percent in 2000.

Yet the study also flashes a caution: While the number of high-poverty neighborhoods fell in rural areas and central cities, there was little change in the suburbs. In fact, poverty grew in some older suburbs. "If poverty in these areas rose during the strongest economy we can reasonably expect to enjoy, then they may well have a bleak future and develop many of the same fiscal and social concerns that plagued central cities in earlier periods," Jargowsky warns.

Nonetheless, the break-up of concentrated poverty and the shrinking underclass offer fresh proof that the nation's social problems are not intractable. It means that fewer poor Americans are living in neighborhoods with very high unemployment, school dropouts, families headed by single mothers, and households dependent on government handouts.

What's behind these welcome trends? The surging growth of the late 1990s, which, for once, was broadly shared, was critically important. It brought unskilled workers their first sustained rise in real wages since the 1960s.

Equally important were the social policy changes of the Clinton years: a much more generous Earned Income Tax Credit, which encouraged and rewarded work; replacement of the welfare entitlement with a work obligation; new child care supports to enable more single parents to take jobs; and the demolition of some of America's worst high-rise public housing projects, such as Chicago's infamous Robert Taylor Homes, which packed the poorest of the poor in like sardines.

All embody the central insight of progressive, post-welfare policy: Changing behavior, not just doling out cash, is the best way to reduce poverty.

"There is no state where a welfare check will raise a four-person family above the government's official poverty line," note Isabel Sawhill and Ron Haskins of Brookings. Their research instead stresses behavioral changes: If we could get more poor Americans to finish school, defer having children until they are married, and work full time, poverty rates would fall sharply. In fact, full-time work reduces poverty by almost 50 percent among families with children. To get equally deep cuts in poverty rates the old way, you'd have to triple welfare benefits. Leaving aside the vexing problem that more generous cash assistance would undermine incentives to work, there's clearly no public support for such an approach.

For progressives, the lesson is equally clear: Using public benefits to reward work and responsible behavior is the right way to fight poverty in America.

This solution was originally published in the Progressive Policy Institute's Blueprint Magazine on July 23, 2005.