Household Income in Census Tracts Where Someone Was Killed by Police

police killings
Author

Matt Carson

Published

July 22, 2023

I’m working on a project analyzing various class indicators (or levels of deprivation/privilege, if you prefer) in census tracts1 where someone was killed by law enforcement.2

There is a longer history of the debate around whether or not police killings are a function of capitalist political economy and the managing of the “surplus population” or whether the phenomenon largely derives from discrimination against black Americans3. I may return to this in more detail in a future post, but for now I wanted to share the plots I have made for my honors thesis project. Hopefully someone will find them useful in helping us to better understand the phenomenon. Any feedback is appreciated.4

The following plot shows the per 10 million rate of police killings in census tracts based on their median household income as a proportion of the county median household income in which they are situated. I used this measure rather than raw incomes to adjust for differences in what is considered low income and high income.5

This plot shows that the incidence of police killings varies quite a lot depending on tract income level. Low income tracts experience a rate nearly twice that of average income tracts, and high income tracts experience a rate less than one-third of low income tracts.

This plot is constructed the same as the one above, except that it is subset to only include tracts with a white population > 50%. The rates were similar to the plot above except for the low income tracts.

Another plot subset to only include tracts with a black population > 50%. The rates for low income and moderate-low income were quite a bit higher than average or greater income. Average to high income tracts had roughly the same rate.

Latino/Hispanic-majority tracts. Pretty clear trend here as well. The higher the income, the lower the rate.

In majority-Asian tracts the lowest income tracts also experienced the highest rate with higher income tracts experiencing lower rates except high income tracts had a rate greater than moderate-high tracts.

Next I looked at tracts in which there was no racial or ethnic majority, but more than one race/ethnicity comprised > 30% of the population.

White-black tracts.

White-Hispanic/Latino

Black-Hispanic/Latino. (No police killings occurred in high income tracts.)

Update (7/26): I just want to add a plot that shows all the majority one race rates together:

Here is the table with the rates.

Letter Definition
W White
B Black
A Asian
H Hispanic/Latino
Race Income Level Total Population # of Police Killings Average per Year Per 10 Million Person-years
A Average 756235 9 1.5000000 19.835104
A High 1142337 11 1.8333333 16.048971
A Low 568765 9 1.5000000 26.372931
A Moderate-High 600245 5 0.8333333 13.883220
A Moderate-Low 574605 8 1.3333333 23.204346
A Unknown 15139 NA NA NA
AH Average 287183 10 1.6666667 58.035004
AH High 175125 2 0.3333333 19.034023
AH Low 218377 6 1.0000000 45.792368
AH Moderate-High 167022 3 0.5000000 29.936176
AH Moderate-Low 328824 5 0.8333333 25.342838
AH Unknown 26 NA NA NA
B Average 3277934 58 9.6666667 29.490120
B High 1306465 20 3.3333333 25.514142
B Low 10061133 494 82.3333333 81.833063
B Moderate-High 1736993 30 5.0000000 28.785378
B Moderate-Low 5144285 152 25.3333333 49.245587
B Unknown 238475 7 1.1666667 48.921970
BA Average 21324 NA NA NA
BA High 19647 NA NA NA
BA Low 52918 3 0.5000000 94.485808
BA Moderate-High 28913 NA NA NA
BA Moderate-Low 8478 NA NA NA
BH Average 316018 8 1.3333333 42.191690
BH High 137442 NA NA NA
BH Low 1385288 56 9.3333333 67.374678
BH Moderate-High 187009 2 0.3333333 17.824454
BH Moderate-Low 706614 25 4.1666667 58.966659
BH Unknown 42508 1 0.1666667 39.208306
H Average 6617160 181 30.1666667 45.588541
H High 2966472 56 9.3333333 31.462739
H Low 12482813 498 83.0000000 66.491423
H Moderate-High 3166311 66 11.0000000 34.740744
H Moderate-Low 10308869 312 52.0000000 50.442003
H Unknown 159463 6 1.0000000 62.710472
NA Average 4345624 99 16.5000000 37.969231
NA High 3307656 54 9.0000000 27.209601
NA Low 4803133 177 29.5000000 61.418245
NA Moderate-High 2983665 56 9.3333333 31.281439
NA Moderate-Low 5212294 110 18.3333333 35.173253
NA Unknown 209091 16 2.6666667 127.536176
W Average 56688110 1096 182.6666667 32.223101
W High 60786693 657 109.5000000 18.013811
W Low 16293469 461 76.8333333 47.155909
W Moderate-High 45173294 724 120.6666667 26.711948
W Moderate-Low 37231765 903 150.5000000 40.422473
W Unknown 598252 14 2.3333333 39.002516
WA Average 725798 6 1.0000000 13.777938
WA High 1207621 4 0.6666667 5.520496
WA Low 252639 3 0.5000000 19.791085
WA Moderate-High 638750 7 1.1666667 18.264840
WA Moderate-Low 514456 8 1.3333333 25.917344
WA Unknown 24595 NA NA NA
WAH Average 2356 NA NA NA
WAH High 4040 NA NA NA
WAH Moderate-High 6370 NA NA NA
WAH Moderate-Low 3725 NA NA NA
WB Average 1673154 29 4.8333333 28.887558
WB High 773203 9 1.5000000 19.399821
WB Low 2065931 78 13.0000000 62.925625
WB Moderate-High 872883 17 2.8333333 32.459486
WB Moderate-Low 2030464 42 7.0000000 34.474879
WB Unknown 95110 NA NA NA
WBH Average 9868 1 0.1666667 168.896095
WBH High 2043 NA NA NA
WBH Low 59452 2 0.3333333 56.067640
WBH Moderate-Low 37333 NA NA NA
WH Average 2970689 60 10.0000000 33.662225
WH High 2070409 43 7.1666667 34.614739
WH Low 2463263 144 24.0000000 97.431740
WH Moderate-High 1881275 48 8.0000000 42.524352
WH Moderate-Low 3302617 83 13.8333333 41.885975
WH Unknown 41803 NA NA NA

Footnotes

  1. These are “small, relatively permanent statistical subdivisions of a county or statistically equivalent entity” that “generally have a population size between 1,200 and 8,000 people, with an optimum size of 4,000 people” (US Census Bureau).
    ↩︎

  2. I will refer to this phenomenon as “police killings” throughout the rest of the blog. This is not to be confused with law enforcement officers being killed.↩︎

  3. Or both, but still the question hinges on to what extent racist discrimination can explain the phenomenon, and to what extent that explanation is inadequate.↩︎

  4. carsonmatt [at] ucla [dot] edu↩︎

  5. For example, the median income in San Francisco County is much greater than other regions and thus what is considered “low income” and “high income” for that county is much greater than in other regions.↩︎