What's Safer At Reducing Murders: Having Richer Neighbors Or Fewer Black Neighbors?
Print Friendly and PDF

From the Journal of the American Medical Association:

Neighborhood Racial Composition and Gun Homicides

November 30, 2020
Chaeyoung Cheon; Yuzhou Lin; David J. Harding, PhD; Wei Wang, PhD; Dylan S. Small, PhD

Substantial racial disparities exist for gun homicide deaths in the US: the 2003 to 2017 age-adjusted rate was 7.4 times higher for Black individuals than for White individuals.

That is that blacks were 7.4 times as likely to be murder victims as whites, from CDC data. (Note that CDC victimization data does a pretty good job of breaking out whites from Hispanics, unlike FBI offending data.) The ratio of black murder offending rates goes unsaid.

Walker et al found an even greater disparity in Chicago and suggested that because race may be a surrogate for income, public health interventions should aim to alleviate poverty in order to reduce gun violence. However, in Philadelphia, Black individuals of the same income level as White individuals were at higher risk of being shot, suggesting that public policies to reduce gun violence and racial disparities might need to go beyond alleviating poverty. In this cross-sectional study, we examine disparities in gun homicide rates among neighborhoods of different racial composition for fixed levels of socioeconomic status….

We obtained from the Gun Violence Archive information on gun homicide deaths by US Census tract from 2014 to 2018. To measure US Census tract racial composition and socioeconomic status, we used data from the American Community Survey’s 2014 to 2018 US Census tract profiles and computed the deprivation index, which is a 0 to 1 scale based on poverty rate; median household income; residents older than 24 years without a high school degree; residents without health insurance; residents receiving public assistance income, food stamps, or the Supplemental Nutrition Assistance Program; and vacant housing units.

The higher the percentile ranking of a neighborhood on the deprivation index, the worse it is.


For a fixed socioeconomic status of a US Census tract—high, medium or low—US Census tracts with a higher proportion of Black residents have higher gun homicide rates. The US remains highly residentially segregated by race despite improvements since the 1960s. Besides residential segregation reducing Black individuals’ socioeconomic status by such mechanisms as inhibiting wealth accumulation through housing value and limiting access to high-quality schools, our findings suggest that even among neighborhoods of the same socioeconomic status, residential segregation may put Black individuals at higher risk of gun homicide. Potential explanations include the following being more prevalent in higher proportion Black neighborhoods: lack of institutional resources and opportunities caused by racial wealth gaps and underinvestment, the legacy of punitive law enforcement leading to difficulties controlling crime, lower collective efficacy due to lack of political power or city responsiveness, geographic proximity to poor neighborhoods, and gang networks or interconnections.

Also, one other possible reason why blacker neighborhoods tend to have higher murder rates than nonblacker neighborhoods of the same average deprivation level is that black neighborhoods have more blacks:

Further studies should be conducted to investigate these explanations and design policies to reduce gun homicides. Limitations of this study include the inability to pinpoint the demographic characteristics of the people affected by gun homicides.

I’ve adapted their table slightly to make it easier to work with, changing homicides from per 1000 to per 100,000 as used in most other statistics. (E.g., Baltimore, which is 5/8ths black, had a homicide rate per 100,000 in 2020 of 54.5 and Chicago, which is 30% black, had a murder rate of 28.4.)

In this table, the rows represent neighborhoods of worsening deprivation percentiles as you go down. E.g., the top column represent neighborhoods at only the 10th percentile on the deprivation index (i.e., 90th percentile on the niceness index). Meanwhile, the columns moving from left to right indicate increase percentages of black residents.

The lowest murder rate is seen, not surprisingly, in the top left corner, which represents neighborhoods at only the 10th percentile of deprivation (90th percentile of privilege) and only 1% black residents. According to the researchers big multiple regression model, these neighborhoods can be expected to see 1.0 homicide per year.

That 1.0 figure is a lucky coincidence, since you can therefore read both the homicide rate and the ratio to the least homicidal neighborhood from the same numbers. E.g., in the worst neighborhoods in the lower right corner that are at the 99th percentile of deprivation and are 90% black, the homicide rate is 51.1 per 100k, and also the homicide rate there is 51.1 times that of the best neighborhoods in the upper left corner:

Table: Mean Incidence Rate of Gun Homicide Deaths per 100,000 People per Year over 5-Year Period.
From 2014 to 2018 by a US Census Tracts Deprivation Index Level and Percentage Black Residents

Point estimate

Percentile, deprivation index level (higher is worse) 1% black residents 13% black residents 50% black residents 90% black residents Ratio: 90% Black / 1% Black
10th, 0.20 1.0 1.6 2.9 4.8 4.8
33rd, 0.28 1.7 2.5 4.7 7.7 4.5
50th, 0.34 2.2 3.3 6.3 10.9 5.0
67th, 0.40 2.9 4.1 8.7 16.8 5.8
99th 0.72 5.9 11.8 25.5 51.1 8.7
Ratio: 99th deprivation/10th deprivation 5.9 7.4 8.8 10.6  

The italicized figures are the ratios of the worst neighborhood in each row or column to the best neighborhood. For example, in neighborhoods at the 50th percentile (or the median) of deprivation, murders are 5.0 times as common in the ones that are 90% black as in ones that 1% black. Similarly, in neighborhoods that are 50% black, murders are 8.8 times as common as in ones that are at the 99th percentile of deprivation as in ones that are at only the 10th percentile.

What it looks like is that at most deprivation levels, 90% black neighborhoods are about 5 times as homicidal as 99% nonblack neighborhoods, except at the very bottom of the social scale, where they rise to 8.7 times as lethal.

Of course, a big limitation in this data is that the Gun Violence Archive doesn’t break out victims, much less shooters, by race. So what we’d like to know is what percentage of the murderers in each type of neighborhood are black.

And also we’d like to know what the average black deprivation level is within each neighborhood. And what the deprivation level of each black shooter was.

Raj Chetty’s theory appears to be that if you can get ghetto moms to move to a nice suburb before their sons are headed for the street gang thug life track, the sons will do a lot better. But it’s hard to evaluate this using this data.

[Comment at Unz.com]

Print Friendly and PDF