June 9, 2001

BEING POOR ISN'T A RACE THING

SCRIPPS HOWARD NEWS SERVICE

Everybody knows that black children are far more likely to be poor than white children. But like so much that "everybody knows," it isn't true, at least not in the way that everybody understands it. Many things affect whether a child is poor, and after all the others are accounted for, being black isn't one of them.

Surprising, yes? But that's what the data show, according to a May 23 report by Robert Rector, Kirk Johnson and Patrick Fagin of the Heritage Foundation (on the Web at heritage.org/library/cda).

There's no question about the absolute numbers: In 1999, 33.1 percent of black children and 13.5 percent of white children were poor, not counting supplements to income such as the earned income tax credit or housing subsidies.

But some of the disparity certainly results from factors distinct from race.

Single-parent families are more likely to be poor than married-couple families, and black children are more likely to live with a single parent; welfare families are almost all poor, and black children are more likely than white to live in families that are or were on welfare; people with limited educational achievement are more likely to be low-income, and black children are more likely to live with mothers who have limited educational achievement.

The researchers used data from the National Longitudinal Survey of Youth, a federal research project begun in 1979 with a sample of 12,686 people then between the ages of 14 and 22. Follow-up information has been gathered about the same people every year or two since, and also about the 6,120 children born to them between 1979 and 1996.

Which of those factors matter most? Far and away the most important independent influence is what percentage of the child's life has been spent on welfare (under the pre-reform system, since the data go up to 1996).

Second most important is percentage of time in a single-parent family. Those numbers are stark. On average, the child of a never-married mother had spent 44.5 percent of his or her lifetime on Aid to Families with Dependent Children, as welfare was called then. If the mother married after the birth of her child, it was 20.4 percent; if she was married at the time of the birth but later divorced, 10.7 percent; and if in an intact marriage, 2.5percent.

Third most important is the mother's math and verbal skill level. The next three are number of children, living in the South, and the mother's age at first birth. Being black is not statistically significant.

You might wonder whether they've got the causality backward. Isn't it true that people go on welfare because they're poor?

Individually, no doubt some do. But statistically it's the other way around.

Being on welfare influences subsequent poverty more strongly than being poor influences being on welfare subsequently.

That's not entirely counterintuitive. First, welfare payments were rarely high enough to keep people out of poverty, just high enough to pull them out of the labor market. And second, being out of the labor market for a time meant fewer years of experience, depressing income even after the family left welfare, if they ever did.

But people could argue endlessly about which explanation was correct, perhaps subtly influenced by which explanation they believed ought to be correct. What Heritage is able to do is run the numbers both ways, and get an answer that is not dependent on the researchers' presuppositions.

At a journalism conference I went to recently, a colleague bragged that he'd never use any material from Heritage because they were conservative.

So they are, and proudly so; but they are equally proud, with good reason, of the research capabilities of their Center for Data Analysis, which published this report. The center's mission is to offer "methodologically transparent and peer-reviewed analyses" to members of Congress (from both parties) and - despite my colleague's skepticism - they work with journalists too, aiming to shift reporting from away from anecdotes toward research.

"Improved public understanding is a desirable result, regardless of ideological perspective," they say.

It's certainly a much better guide to sound public policy.

(674 words)