Watching the courage of ordinary low-income people as they deal with the aftermath of Katrina and Rita, it is hard to decide which politicians are more contemptible--Democrats who are rediscovering poverty and blaming it on George W. Bush, or Republicans who are rediscovering poverty and claiming that the government can fix it. Both sides are unwilling to face reality: We haven't rediscovered poverty, we have rediscovered the underclass; the underclass has been growing during all the years that people were ignoring it, including the Clinton years; and the programs politicians tout as solutions are a mismatch for the people who constitute the problem.

• We have rediscovered the underclass. Newspapers and television understandably prefer to feature low-income people who are trying hard--the middle-aged man working two jobs, the mother worrying about how to get her children into school in a strange city. These people are rightly the objects of an outpouring of help from around the country, but their troubles are relatively easy to resolve. Tell the man where a job is, and he will take it. Tell the mother where a school is, and she will get her children into it. Other images show us the face of the hard problem: those of the looters and thugs, and those of inert women doing nothing to help themselves or their children. They are the underclass.

We in the better parts of town haven't had to deal with the underclass for many years, having successfully erected screens that keep them from troubling us. We no longer have to send our children to school with their children. Except in the most progressive cities, the homeless have been taken off the streets. And most importantly, we have dealt with crime. This has led to a curious paradox: falling crime and a growing underclass.

• The underclass has been growing. The crime rate has been dropping for 13 years. But the proportion of young men who grow up unsocialized and who, given the opportunity, commit crimes, has not.

A rough operational measure of criminality is the percentage of the population under correctional supervision. This is less sensitive to changes in correctional fashion than imprisonment rates, since people convicted of a crime get some sort of correctional supervision regardless of the political climate. When Ronald Reagan took office, 0.9% of the population was under correctional supervision. That figure has continued to rise. When crime began to fall in 1992, it stood at 1.9%. In 2003 it was 2.4%. Crime has dropped, but criminality has continued to rise.

This doesn't matter to the middle and upper classes, because we figured out how to deal with it. Partly we created enclaves where criminals have a harder time getting at us, and instead must be content with preying on their own neighbors. But mainly we locked 'em up, a radical change from the 1960s and 1970s. Consider this statistic: The ratio of prisoners to crimes that prevailed when Ronald Reagan took office, applied to the number of crimes reported in 2003, corresponds to a prison population of 490,000. The actual prison population in 2003 was 2,086,000, a difference of 1.6 million. If you doubt that criminality has increased, imagine the crime rate tomorrow if today we released 1.6 million people from our jails and prisons.

Criminality is the most extreme manifestation of the unsocialized young male. Another is the proportion of young males who choose not to work. Among black males ages 20-24, for example, the percentage who were not working or looking for work when the first numbers were gathered in 1954 was 9%. That figure grew during the 1960s and 1970s, stabilizing at around 20% during the 1980s. The proportion rose again, reaching 30% in 1999, a year when employers were frantically seeking workers for every level of job. The dropout rate among young white males is lower, but has been increasing faster than among blacks.

These increases are not explained by changes in college enrollment or any other benign cause. Large numbers of healthy young men, at ages when labor force participation used to be close to universal, have dropped out. Remember that these numbers ignore young males already in prison. Include them in the calculation, and the evidence of the deteriorating socialization of young males, concentrated in low income groups, is overwhelming.

Why has the proportion of unsocialized young males risen so relentlessly? In large part, I would argue, because the proportion of young males who have grown up without fathers has also risen relentlessly. The indicator here is the illegitimacy ratio--the percentage of live births that occur to single women. It was a minuscule 4% in the early 1950s, and it has risen substantially in every subsequent decade. The ratio reached the 25% milestone in 1988 and the 33% milestone in 1999. As of 2003, the figure was 35%--of all births, including whites. The black illegitimacy ratio in 2003 was 68%. By way of comparison: The illegitimacy ratio that caused Daniel Patrick Moynihan to proclaim the breakdown of the black family in the early 1960s was 24%.

But illegitimacy is now common throughout the population, right? No, it is heavily concentrated in low-income groups. Perhaps illegitimacy isn't as bad as we used to think it was? No, during the past decade the evidence about the problems caused by illegitimacy has grown stronger. What about all the good news about falling teenage births? About plunging welfare rolls? Both trends are welcome, but neither has anything to do with the proportion of children being born and raised without fathers, and that proportion is the indicator that predicts the size of the underclass in the next generation.

• The government hasn't a clue. Versions of every program being proposed in the aftermath of Katrina have been tried before and evaluated. We already know that the programs are mismatched with the characteristics of the underclass. Job training? Unemployment in the underclass is not caused by lack of jobs or of job skills, but by the inability to get up every morning and go to work. A homesteading act? The lack of home ownership is not caused by the inability to save money from meager earnings, but because the concept of thrift is alien. You name it, we've tried it. It doesn't work with the underclass.

Perhaps the programs now being proposed by the administration will help ordinary poor people whose socialization is just fine and need nothing more than a chance. It is comforting to think so, but past experience with similar programs does not give reason for optimism--it is hard to exaggerate how ineffectually they have been administered. In any case, poor people who are not part of the underclass seldom need help to get out of poverty. Despite the exceptions that get the newspaper ink, the statistical reality is that people who get into the American job market and stay there seldom remain poor unless they do something self-destructive. And behaving self-destructively is the hallmark of the underclass.

Hurricane Katrina temporarily blew away the screens that we have erected to keep the underclass out of sight and out of mind. We are now to be treated to a flurry of government efforts from politicians who are shocked, shocked, by what they saw. What comes next is depressingly predictable. Five years from now, the official evaluations will report that there were no statistically significant differences between the subsequent lives of people who got the government help and the lives of people in a control group. Newspapers will not carry that story, because no one will be interested any longer. No one will be interested because we will have long since replaced the screens, and long since forgotten.