Just when everybody is starting to understand the "feminization of poverty" -- the crucial link between family break-up and poverty -- along comes Mary Jo Banes to tell us it ain't necessarily so.

It used to be. Between 1960 and the late 1970s, among both blacks and whites, family-structure changes served to keep the poverty rate higher than it otherwise would have been.

It's not true now. Family-composition changes "contributed almost nothing" to the sharp increase in poverty since 1979.

Bane, executive deputy commissioner of the New York State Department of Social Services, told a recent congressional hearing that she was surprised by her own research findings -- so surprised, in fact, that she did further analysis. What she found was an unexpected distinction between "event-caused" poverty among whites and "reshuffled" poverty among blacks.

That is, while three-fourths of poor, female-headed, white families became poor as a result of the family-break-up, two-thirds of the poor, female-headed, black families had been poor before their families broke up.

At least among blacks, the frequently remarked increase in the number of poor families does not necessarily reflect an increase in the number of poor individuals. The jobless man who leaves his wife and children increases the number poor families by one, but leaves the number of poor individuals unchanged.

This is no debate over angels and pinheads. A crucial part of what is coming to be conventional wisdom is the Charles Murray message that the availability of public welfare induces marriage break-ups and out-of-wedlock births and, thereby, adds to poverty. What Bane found is that welfare has no discernible effect on birthrates to unmarried women, but does influence the decision of whether the young mother will live separately from her own parents, thereby increasing the number of poor families, though not the amount of poverty.

The public policy implications are obvious. If welfare somehow "causes" poverty, then it may make sense to change or even reduce availability of welfare. But if Bane is correct in her belief that the increase in the number of poor families is, to a substantial degree, the result of "reshuffling," then we have to look elsewhere for poverty's basic causes.

Bane says she has found one. "The decline in real benefits levels since the mid-1970s has been an important contributor to poverty among female- headed families, and to their increasing poverty rates since 1979. I have found nothing in my research that would suggest that raising welfare-benefit levels would do anything to make things worse, and much to suggest that raising them would make things better."

She is not suggesting that adolescent pregnancy is other than what it is: a growing disaster that is "ruining innumerable lives among both white and black young women and severely limiting the opportunities available to their children." Nor is she saying that the preservation of the family is of no societal concern, or that welfare ought to be more than an economic Band- Aid, with increased educational and employment opportunity the only long-term solution. She is only urging that we not let unsound analysis lead us to faulty policy choices.

The family-disintegration idea, she believes, is an appealing but false explanation that may serve to lead us away from the deeply rooted-and worsening-poverty that is besetting a growing segment of our society.