Taking Exception

Misreading the Poverty Data

By Robert Greenstein
Tuesday, September 18, 2007

In his Sept. 5 op-ed, " Importing Poverty," Robert J. Samuelson assailed the Census Bureau, the American Enterprise Institute, the Center on Budget and Policy Priorities, and the media for missing what he views as the core of the poverty story. When discussing the figures that the Census Bureau released Aug. 28, we all failed, he said, to explain that poverty "is increasingly a problem associated with immigration," driven by the large numbers of poor Hispanics entering the country.

But a careful look at the data does not support Samuelson's narrow view of how immigrants in general, and Hispanic immigrants in particular, affect poverty trends.

The poverty rate in 2006 was 12.3 percent. If immigration had not increased, and immigrants and their family members comprised the same share of the population in 2006 as in 1993 (the first year for which these Census Bureau data are available), the poverty rate would be nearly the same, about 12 percent.

There is debate on whether immigration lowers the wages of natives -- and the research on that subject is mixed -- but even if it does, the added effect on the poverty rate would be small. Immigrants do experience more poverty than native-born citizens, but they are not driving the nation's poverty rate.

In addition, the overall drop in the poverty rate and the rise in national median income in 2006, compared with 2005, were driven by improvement among Hispanics. Hispanic poverty fell, and the median income of Hispanic households rose. Non-Hispanic whites and African Americans, by contrast, experienced no such improvement.

Indeed, since 2001, Hispanics have made considerably more progress against poverty than the other groups. Their poverty rate is lower than it was in 2000, before the last recession -- it stands at its lowest level on record -- while poverty rates for non-Hispanic whites and blacks remain well above their pre-recession levels.

Samuelson focused on longer-term trends and, in particular, on changes in the number of poor Hispanics since 1990, using Hispanics as a proxy for immigrants. But in doing so, he told only one side of the story.

In the 1990s, the number of poor Hispanics did increase substantially even as the number of non-Hispanic poor declined. So Hispanics accounted for the entire increase in the poverty population in that decade. But that's not true since 2000. The Pew Hispanic Center has found that newly arrived Hispanic immigrant workers were better educated and much less likely to be low-wage earners in 2005 than in 1995.

Moreover, even while the number of poor Hispanics rose markedly in the 1990s, the Hispanic poverty rate -- that is, the percentage of Hispanics in poverty -- fell, and it fell more than the poverty rate for the rest of the country. Since then, the Hispanic poverty rate has continued falling (except for temporary increases related to recessions), even as large numbers of new immigrants continue to enter the United States.

How could this be? How could the number of poor Hispanics rise but the percentage of Hispanics who are poor fall sharply? Because the Hispanic population is growing so quickly. Samuelson complained that there were 3.2 million more poor Hispanics in 2006 than in 1990. He did not mention that the number of non-poor Hispanics -- including doctors, teachers, small-business owners, waiters and members of the armed forces -- grew by 20 million over the same period. (And contrary to Samuelson's implication that the Center on Budget and Policy Priorities ignores poverty increases among Hispanics to avoid upsetting its supporters, we reported frequently in the 1990s on the rising numbers of poor Hispanics and the role of immigration in contributing to that increase.)

Nor was poverty the only issue on which Samuelson's focus was too narrow. He noted, correctly, that Hispanics accounted for 41 percent of the increase, since 2000, in the number of Americans who lack health insurance. That sounds alarming, until you realize that Hispanic population growth accounted for 51 percent of total U.S. population growth over this period.

In fact, Hispanics also accounted for 60 percent of the increase in the number of people with insurance. And the percentage of Hispanics who are uninsured grew more slowly than the percentage of non-Hispanics who lack insurance.

Poverty, race, ethnicity and immigration are complicated and controversial issues, and they arouse strong passions. That's all the more reason that we should be careful how we use data to tell a story. We should not oversimplify a complicated story, as the normally careful Samuelson has done here.

The writer is executive director of theCenter on Budget and Policy Priorities.


© 2007 The Washington Post Company