2001 study: white mortality in U.S. metropolitan areas is a function of the size of the local black population

Max P. writes:

In regards to your recent posts concerning the attacks on an Atlanta commuter train and in a Maryland McDonald’s, I bring to your attention this study from Princeton University from October 2001, “Mortality, Income Inequality, and Race in America,” by Angus Deaton and Darren Lubotsky.

The study was set to explore the correlation between income inequality and mortality, but the authors found little evidence of this correlation. Instead they found that racial composition plays a key role in mortality rates. “Specifically, they show that white mortality rates are higher in states and MSA’s [“metropolitan statistical area.”] where the black population is a larger fraction of the total.”

The full 38 page report is available online. Here is a two page executive summary. From the executive summary, some key passages:

“Their results indicate that the positive relationship between inequality and white mortality can be accounted for by regional racial composition. Specifically, they show that white mortality rates are higher in states and MSAs where the black population is a larger fraction of the total.”

“Mortality among both whites and blacks and fraction of the population that is black are higher in southern cities and states than in other parts of the United States. However, the authors find that the strong positive correlation between white mortality and the black fraction holds within all four broad Census regions of the country, and within states. Thus, the relationship is not driven by any particular region of the country, or by a comparison of geographically dissimilar places.”

“Deaton and Lubotsky’s results indicate that for white men, the risk of death is 14% higher in Mississippi, which is one-third black, than in Vermont, which has virtually no blacks. The corresponding risk is 9% higher among white women.”

And finally from the conclusion:

“In American cities and states, the fraction of the population that is black significantly affects white mortality rates.”

LA replies:

Very interesting, thanks for sending. I wonder if there was a response to their findings and if their findings have held up. On the face of it, it doesn’t seem likely that black-on-white homicides by themselves could account for such a substantial difference in white mortality rates in different metropolitan areas. But it would seem probably that there are other factors related to a large black population that would come into play, such as greater general disorder; inefficiency of institutions including hospitals and police; car accidents; fires in black apartments and homes in which white firemen are killed (most of the deaths in the overwhelmingly white New York City Fire Department occur from fighting fires in black and Hispanic residences); and things of that nature.

We know from the National Longitudinal Study of Youth, discussed in absorbing detail by Murray and Herrnstein in part two of The Bell Curve, that there is a large correlation between low IQ and accident proneness. A black population, with its much lower mean IQ, will be much more accident prone than a white population. A metropolitan area with a large black population will have many more accidents, including car accidents and accidentally set fires, than an area without blacks, and the greater number of car accidents and fires would affect non-blacks as well.

- end of initial entry -

April 26 Forta Leza writes (April 23):

I am skeptical of this study. Although it’s true that blacks cause a lot of problems, they also tend to show up in places where the whites are pretty wealthy.

Anyway, I looked at the report and it seems that the data it used was from the “compressed mortality file” of the CDC from 1980 and 1990.

A description of that data can be found here.

Significantly, in 1980 and 1990, the data did not keep track of Hispanic origin. It just categorized people as “white,” “black,” or “other.” So I am guessing that as with a lot of crime data, Hispanics are lumped in with whites. I would guess that if you are categorized as “white,” and you live in a city with a lot of blacks, there is a much better chance that you are actually Hispanic than if you are white in an area with few blacks. And that Hispanics have less favorable mortality than non-Hispanic whites, which would explain the disparity in the study.

Posted by Lawrence Auster at April 23, 2011 08:08 AM | Send