A pay picture complicated by immigrants, family size

CHAPEL HILL -- In recent weeks we've been blitzed by stories about the stagnation -- even decline -- of family and household incomes in the United States over the past generation. Many leading journalists have written powerful pieces about this subject, the central point being that the boom of the past generation has bypassed most Americans.

These arguments are generally based upon data on median household and median family income, adjusted for inflation. Although the principal point is well taken -- income growth has clearly slowed for middling groups -- greater attention to a few qualifying points would enable us not only to interpret the data more accurately but also to devise more targeted means of addressing the problems revealed.

Groups such as the Economic Policy Institute have done us all a service by disaggregating income data in various ways, most notably, by race and age cohorts. But consider two other considerations: household/family size and immigrant/nonimmigrant status. For whatever one's politics, it is instructive to note that although median income for both households and families has not grown much over the last generation -- and might even have declined a bit between 2000 and 2007, as David Leonhardt recently pointed out in The New York Times -- the median size of both households and families has also fallen over the course of the same period.

This means that, on average, income is being shared by smaller household and family units. In 1980, for example, the median size of U.S. households was 2.75 and the median size of families was 3.27. In 2006, by contrast, median household size had fallen to 2.61 and median family size to 3.20, decreases of about 5 percent and 2 percent respectively. Not great, in other words, but worth noting.

A MORE SERIOUS PROBLEM WITH A BROAD-BRUSH "STAGNATION" THESIS -- or at least with its interpretation -- is related to the sharp rise in immigration (legal and illegal) over the course of the past generation. Because most of these immigrants are situated on the lower half of the income ladder, rapid immigration -- as well as technological change, education gaps, offshore outsourcing, etc. -- is accounting for some part of the observed stagnation in median household and family income.

It is likely, that is to say, that "stagnation" is attributable at least in part to the rapid influx of relatively poorer populations into the United States. Such immigrants may be holding back, or even driving down, median income (and increasing poverty levels) even as they improve upon the income status they held before coming to this country.

Last January the U.S. Department of Health and Human Services set new poverty guidelines, placing the poverty line for a family of four at $22,200. To be sure, such a figure isn't much by U.S. standards, but what about by the standards of Guatemala, Ethiopia or Sierra Leone?

Things get more interesting when framed in this way. And what about nonimmigrant populations in the United States? Is their income also stagnating? It seems to be for nonimmigrant, white, male-led households, at least among certain age cohorts, but what about for other cohorts and groups? Here, too, things quickly get complicated.

Clearly, then, just as disaggregation by age cohort and race is useful, disaggregation by origins and resident status -- and, to a lesser extent, more acknowledgment of changes in household and family size -- would help us to interpret income levels and trends and maybe even to devise better policies to help those in need.

Even if it is relatively poorer immigrant families, for example, that "explain" a good portion of the stagnation in incomes and the sticky or even rising poverty levels in the United States, we would still have problems, of course. But they would be different problems than would be the case if stagnation and poverty affected broader groups in American society.

(Peter A. Coclanis is associate provost for international affairs and Albert R. Newsome professor of history at UNC-Chapel Hill.)