42 Comments

Compelling analysis. This is a phenomenon that is fairly taboo, but screams for attention.

https://carolinacurmudgeon.substack.com/p/that-which-it-would-be-impolite-to

Expand full comment

Great write up, thank you for sharing this. I had seen arguments that the US murder rate looks a lot like other Western countries once one controls for race, but this does a very good job of highlighting just what is going on there.

Expand full comment

I actually have a different post tackling exactly that question, though it is one of my posts for paid subscribers. The short answer is that, while demographics accounts for most of the homicide rate difference between the United States and other comparably developed Western nations, a nontrivial difference remains after such controls.

https://inquisitivebird.substack.com/p/demographics-and-american-homicides

Expand full comment

Excellent, thank you!

Expand full comment

Can you comment on how your analysis differs from this?

https://randomcriticalanalysis.com/2015/11/16/racial-differences-in-homicide-rates-are-poorly-explained-by-economics/

It seems your methods are pretty similar, and the results are too. It is nice though to have an updated analysis.

This discussion reminds me of the debate between JonTron and Destiny.

https://youtu.be/1qQNYukh-n0?t=409

"Wealthy Blacks commit more crime than poor Whites"

Depending on how you define wealthy and poor, this certainly doesn't seem like an incorrect statement.

Expand full comment

Hi. Yes, the difference is that, as I understand it, RCA does not separate economic variables by race within counties. He uses the economic indicators for the counties at large. Because I separate them by race within county, I can make sure that my analysis is not biased by within-county economic disparities between black and white people. That is, I can compare people who are economically similar.

Expand full comment

Hmm... I'm not sure if this is true. Or maybe I'm just missing something. Take for example this graph:

https://i0.wp.com/randomcriticalanalysis.com/wp-content/uploads/2015/11/sep_hom_by_mfi.png?ssl=1

If he does not differentiate counties by race, shouldn't the three dots of one county simply be vertically shifted? However, there are a bunch of blue dots on the right. Where are the corresponding red and green dots? His graph only makes sense to me if he already separated incomes by race within county.

Expand full comment

Oh, right. He also has a section where he does disaggregate economic indicators by race it seems. Yes, then that section should be very similar to my analysis.

Expand full comment

In many black communities, the image of a strong, rebellious black man is perpetuated by cultural influencers in music, videos, and movies. Add that to the majority of back children are born into homes with no father and a mother trying to work and raise her children alone, and it becomes a recipe for young boys to run the streets.

Expand full comment

People need to stop making excuses and have a conversation on a national level. Something needs to change.

Expand full comment

Brilliant. Let's just ask people very nicely to stop shooting people

Expand full comment

the Obama presidency will go down in history as the greatest missed opportunity EVER.

Expand full comment

It's nice that you've done the analysis here. This along with births to single mothers are the statistical anomalies amongst poor black communities that need to be addressed and discussed prior to any woke solution to close white/black disparities.

Many on left refuse to even acknowledge these and seem surprised/offended when I bring them up. I genuinely don't understand how we can ever reduce despair in poor black communities until these things change.

Expand full comment

How much of the differences in homicide rates can be accounted for by age/sex. I know it can account for a decent amount of the difference in incarceration rates in the ACS(Hispanic-white ratio is reduced to roughly 1.5x). If you don't double control for age ( https://www.aei.org/society-and-culture/one-last-note-on-the-hispanic-crime-debate/).

Expand full comment

Sex: accounting for sex would have virtually no practical impact, but the difference it does make would be in the direction of making the disparities larger. This is because the female percent is slightly higher for black people than white people, and obviously males have higher homicide rate.

Age: Accounting for age would make the disparities smaller, but the effect is much smaller than often suggested. Whites have homicide victimization rate of 2.88, but age-standardized rate of 2.97. Blacks have homicide victimization rate of 28.29, but age-standardized rate of 27.72. So black people have 9.8 [28.29/2.88] times as high homicide victimization rate, and 9.3 [27.72/2.97] times as high age-standardized homicide victimization rate. So the disparity is reduced by about 5% [1-9.3/9.8] when adjusting for age.

Expand full comment

Fair point the difference in age between black and white Americans is not that large, 34.5 and 43.8 in 2021. Hispanic are even lower at 30.5 though(the lowest for any race/ethnic group besides mixed race Americans), and the gap between Asians and whites would increase because of the former groups lower age. I'll double check the result for age and sex adjusted incarceration rates because I thought it made a bigger difference(but I only looked at 18-65 because above that it was the elderly in institutions besides prison).

Expand full comment

Using a Logistic regression for Census/ACS data, the unadjusted odds ratio(1870-2022 and 1990-2022, aged 18-64 because of elderly homes), is 5.085 and 4.811 for blacks, 2.110 and 1.857 for hispanics(compared to NHW). Adjusting for age and sex in the same years(with no filter for age) reduces this to 3.677 and 3.485 for black americans and 1.626 and 1.440 for hispanics. So adjusting for age reduces this by 27.6-27.7% for black americans and 22.5-22.9% for hispanics, larger than for homicide victimization, and similar across more recent years(the difference could be explained by the age-crime relationship weakening over time).

Expand full comment

Are the data available to analyze a correlation with family composition (two parents, single father, single mother, etc.)?

Expand full comment

i wonder, is there some manner in which to try to establish the directionality of causality on poverty and murder rate?

my prior would be that it is not poverty that causes murderousness but rather the sorts of personality traits (poor impulse control, poor forward planning, over aggression, etc) that lead to murder also make one difficult to employ but i lack any real empirical basis to support it.

i wonder if there is some possible vertical study but they look problematic to me as if you follow the same people across different wealth you get selection bias for success and a likely age skew.

has anyone done any work there? maybe tried to identify such traits in kids and then followed them to measure life outcomes vs peers?

Expand full comment

I shortly discuss this exact point in the article. See the section "Poverty as a cause of homicide". As I note in that section, I plan on writing a piece in the future that is dedicated to exactly that question.

Expand full comment

It is worth noting as well that increases in homicide in an area would seem likely to cause increased poverty a people and businesses leave. It would be interesting to see if homicide rates increase ahead of poverty rates in time, although it is quite possible that the data isn't fine grained enough if people start moving within a year or so of perceived increases in homicide (which are presumably also preceded by increases in general crime rates.)

Expand full comment

Excellent analysis. Young black men (15-34) are just 2% of the population and commit about half of the nation's homicides. A rate an astounding 49 times that of the average American. Most of their victims are other young black men. A major reason no one cares. They are the country's gun violence problem. Saying that truth makes me a racist in today's world.

The roots of the problem are: the lack of respect for education (read up on the disruption in any inner city classroom and the refusal of black administrators to address it by imposing needed discipline), the casual acceptance of criminal behavior in the black community (remember the racist OJ jury?), and the failure of many (most?) black fathers to love and care for their children and especially their boy children. Those who object to this analysis deny black people any agency over their own lives. They are the true racists. Fix those issues and you have a shot at reducing gun violence in America.

Expand full comment

Would you say that approximately 5% of the population is causing 56% of all murders. Accounting for mostly male and of a certain age..

Expand full comment

"For example, the last two bars illustrate that, in the counties where black people have per-capita incomes exceeding $40,000, black people in those counties have a homicide victimization risk of 13.1 per 100,000. Similarly, in the counties where white people have per-capita incomes exceeding $40,000, white people in those counties have a homicide victimization risk of 2.5 per 100,000..."

https://substackcdn.com/image/fetch/w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F304b1a64-2a5b-4869-a740-af3f721c298d_1124x833.webp

But the current median per capita income in the US in those years was around $70,000/yr. You're eliding all US counties with incomes greater than $40,000 into one cohort at the end, and the other five entries on your appended chart consist of the poorest counties in the US. The number of US counties with a per capita income of less than $35,000 amount to a small percentage of the total. https://vividmaps.com/wp-content/uploads/2021/12/Income-us-county.jpg

Just from eyeballing the map, counties with less than $35,000 per capita income comprise a single-digit percentage.

You don't give a demographic breakdown of the populations of those counties to show the percentage of residents by race. As a well-traveled, knowledgeable native American, I can confidently assert that most if not all of the counties shown on that map with <$35,000 per capita income have a white population that's far less than the national median. Some of those counties have a low African-American population, too, leading me to wonder about the criteria used in the chart to define the label "black."

Your chart aggregates the years 2018-2021. In the US, the year 2020 saw the largest single year increase in homicides in decades, and the rate continued to increase in 2021 before dropping slightly in 2022. http://cdn.statcdn.com/Infographic/images/normal/31062.jpeg

Expand full comment

You're wrong. The median personal income in these years was lower than $40,000, and substantially lower for black people. You're probably thinking of household income, which includes the incomes of more than one person.

In the figure you reference, I show how many counties apply in each bin. There are few counties where black people had per-capita incomes exceeding $40,000. The analysis includes all >3,000 counties, except for very few where there is missing data.

I also provide the image for all counties without binning incomes. I also conducted a regression analysis that covers the full range without binning incomes.

Expand full comment

The map that I linked is labeled as showing “per capita personal income”—the same language as the chart you linked.

https://vividmaps.com/wp-content/uploads/2021/12/Income-us-county.jpg

https://substackcdn.com/image/fetch/w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F304b1a64-2a5b-4869-a740-af3f721c298d_1124x833.webp

I’m not sure how to account for that disparity. Not at present, anyway.

The number I gave for median income--around $70,000-- was indeed drawn from a link to household income that appeared at the top of search results (despite the fact that I specified [median per capita income] in my keyword search, which is a sad commentary on what's happened to search engines over the last ten years.) But your statement that "median personal income in those years (2018-2021) was lower than $40,000" is incorrect; it was actually closer to $42,000/yr.

https://fred.stlouisfed.org/series/MEPAINUSA672N

So even using the lower figure, you've still elided more than 50% of personal incomes in order to arrive at the figures shown on the rightmost bar on your chart.

Your linked chart also includes a notable demographic disparity between the total number of counties provided for the white residential population and the total number provided for the black residential population: White, 3124; Black, 2023. That's more than a discrepancy involving only a "very few [counties] where there is missing data."

The number of US counties without black residents does not explain that disparity: according to the link below, there are only 6 of those. Even when the measure is changed to “counties with 25 or fewer black residents”, the number only rises by 200, to a total of 206.* Around 93% of all US counties have 26 or more black residents.

https://vividmaps.com/wp-content/uploads/2021/10/Black-Americans-by-US-county.jpg

[*there’s also a slight difference given between the total number of counties enumerated in your chart—3124—and the total of the counties listed in my map link: 3145.]

Somehow, more than 1/3 of the US counties with 26 or more black residents are not counted in the graph you used for your reference. That's more than a "very few" missing counties, for the black population.

You did the calculations for that graph yourself, yes?

You've listed the sources for your raw data. Do you have page links to those data sets?

Expand full comment

There is no disparity with respect to the income, other than the initial misunderstanding that the median should be $70,000 (which is only true for household income). As my own analysis shows, a huge number of counties have per-capita incomes exceeding $40,000... for white people. It's the biggest bin for white people. To be precise, about 100 million white people live in counties where they have per-capita incomes $40,000 or higher. But it's a different story for black people, who have substantially lower incomes.

The number of US counties with next to no black residents *does* explain that disparity. Since economic data is based on surveys, there has to be a significant number of people to get a county estimate, and that is not your arbitrarily set threshold of just 25. The survey sample has to contain an adequate amount of observations for a given race.

Because missingness comes in the counties where there are the fewest people, it has negligible effect on the total population estimates (regressions are population-weighted regardless). I can tell you that my regression analysis covers the counties where 99.06% of black people live.

Homicides by county and race come from CDC Wonder (cause of death code, X85-Y09 Assault). Economic data can be retrieved easily from, e.g., IPUMS. It's the American Community Survey 5-year 2018-2022 data.

Expand full comment

I think you're missing the point that you can only compare the range of incomes for which there is overlap between blacks and whites. Sure, a large portion of the country's counties have per-capita incomes exceeding $40,000. But that is not true for black people. You cannot compare numbers much higher than $40,000 because there are simply too few counties where black people have per-capita incomes that high. To see this better, check out the scatter-plot I made that does not bin incomes.

Expand full comment

I think you're enthralled with the chance to draw sweeping conclusions from a handful of data point correlations. You don't understand the requirement to critique your own findings based on estimating the part of the iceberg that you're unable to see, because you don't even seem to understand that it exists.

Expand full comment

No, no, no. You're drawing arbitrary lines--and relying on arbitrary lines. You discount 1000 counties where some black residents live, on the basis that the numbers are too small to be significant. But then you do no weighting at all in terms of population density when it's at a maximum. You treat county lines as if they had probative significance, when physical geography demonstrates otherwise. The poorest ward in Washington DC abuts what's probably still the wealthiest black majority county in the US, but there's no way to consider those impacts on the basis of your threadbare attempt at drawing conclusions based on three all too easily reduced factors--race, income, and "counties", those wildly disparate geographic regions both physically and culturally,, with their legacy histories. Exactly what makes it important that only a few affluent enclaves of black Americans exist, as compared with the large number of affluent white enclaves? Intergenerational wealth factors count for nothing, in maintaining the status quo ante from coastal Maine to Palm Beach, Aspen to Los Gatos? ? Do you even know how people live? What it's like to be an educated middle class striver with relatives on the edge of precarity, or in jail? It's all just metrics and dessicated statistical analysis and regression analyses that can be massaged to prove any damn thing if you just ask the questions designed to yield the result you prefer? You think all data "science" is created equally, because Numbers? Do you ever think to offer your own reflective appraisal of possible confounding factors? Do you know what questions to ask? Do you even think there are any?

Expand full comment

In the chart showing income and crime by race, you stopped at > $40K for the final bucket. Do you have data for income buckets at 50K, 75K, 100K+? I read somewhere that blacks from families making $100K+ a year still commit more crime than whites in poverty.

Expand full comment

No, I don't have data going that far. You'd need individual-level data for that. I am basing this on county-level data, and though there are many individuals earning that much, there are no counties where the black *average* income is that high.

Expand full comment

Interesting what is weird if you look at poverty and homicide in Latin American countries the relationship is non existent. Colombia is richer and more dangerous than Bolivia which and poorer and safer than Mexico. It is hard to think of any reasons that explain this. Although I would love someone to do so.

Expand full comment

a valid question would related to cumulative, aka generational poverty, correlating to murder rates.

If a person who was raised in non-poverty family background, through misfortune became poor, we could posit this statistical group will be less likely to commit murder. Conversely, a group who are the 3rd straight generation to endure poverty, may show a higher rate of murders committed.

.

it seems clear that the poor commit more murders than the non-poor; which black poor committing a higher rate than white poor. This question, why stats show black poor have a higher murder rate than white poor, needs to be explored. And it is possible that there is a time or generational influence, where multi-generational poverty (white or black) may be likelier to murder; with blacks being more prone to multi-generational poverty.

.

Expand full comment

If I really wanted to argue that SE causes racial crime gaps, here is what I would say: Maybe it's not just absolute poverty that causes crime but relative poverty. Let's say in County A Whites earn 50k and Blacks 30k and in County B Whites earn 30k and Blacks 20k. Then we find that Whites who earn 30k are less violent than Blacks who earn 30k. But maybe this is just the result of relative poverty. In both cases, Blacks are poor relative to the county mean. I wonder what happens if we plot crime against some sort of racial equality index. Are Blacks also more violent than Whites when both earn the same in a county? Does relative poverty "explain" anything?

Expand full comment

I did a quick test where I instead looking at how within-county economic differences were associated with within-county homicide rate differences. Even as the within-county economic differences approach zero, the homicide rate is still substantially higher.

Expand full comment

So what is the X factor that causes high homicidality? Murder rates fluctuate quite a bit over time so it can't be all genetic. There must be a unique cultural cause, no?

Expand full comment

The group difference is mostly genetic, but fluctuations can be caused by historical events and changes in policing. For example, the police pullout from black areas during the Floyd protests/riots.

Expand full comment

There's no genetic argument for it cause if we seen the homicide data of other majority Black countries before 1960 they had one of the lowest murder rates than America. South Africa even before Apartiad the Bantu had a lower homicide rate compare to both White and Black american. Caribbean countries before independence had a lower homicide rate with in comparison to Americans and certain European countries.

Poverty is clearly has to be one factor here and this man is trying to do the same shtick of saying we're naturally violent and dumb despite plenty of criminologists across the board and across other countries clearly stating poverty does contribute to high crime. So clearly he has anti Black agenda as usual.

Expand full comment

Are you unable to read the article? The article shows, using very clear data, that poverty cannot explain the difference between the homicide rate of Whites and Blacks. I'm just repeating the points from the article, but basically, Blacks are much more overrepresented in homicides than they are in poverty. For each given income level, Blacks commit many more homicides than Whites.

The other thing the article notes is that even when there is an association between poverty and crime, it does not necessarily mean that poverty is the cause of crime. The causation could go the other way, with crime causing poverty, and/or it could be that crime and poverty are caused by other factors that influence both such as low IQ and poor impulse control. It is almost certainly true that crime has bad economic effects, and it is almost certainly true that selection is a huge part of any association between poverty and crime.

Regarding the idea that African/Black countries had low homicide rates, the first and most obvious problem with this idea is that African/Black countries tend to be horrible at keeping good statistics. The second problem with your idea is that it contradicts your theory that poverty causes crime, because obviously those countries had much higher poverty than the Black population in the US today.

Expand full comment

I did, data he made not based or verify otherwise other criminologist would state this. And said poverty was one factor. Does he account for density, the fact that man poor Whites are in Rural areas with less data reporting, is he comparing poor urban Blacks to poor rural Whites? Does he not take into account that rich Black neighborhoods are next to poor Black ones and that why it proceeds to assume that higher rates of violence accore in those areas. White rich areas do not live next to poor ones. Black commit more homicides than Whites because again they always systemically discriminated compare to Whites and other groups, so of course they would be targeted more. Black americans are the only group aside from Natives that aren't immigrants, they were always meant and design to be in the bottom by White society.

Where did the get the stats showing that rich black with higher income commit more homicides? Low IQ doesn't necessarly mean violence, Black people and Whites people back then had iq that you would consider to be borderline retarded, yet they had less crime compare to now, in fact some areas and Black towns were more safer compare to White cities and towns. Low IQ shouldn't be factor in violence otherwise many SEA countries should have higher rates of violence. As for poor impulse control that has to do with multiple factors outside of genetics, where was this poor impulse control when Black americans had towns ran by them and barely any crime or during reconstruction when many Blacks had schools and in politics?

When I was talking about African/Carib countries I meant when it was under European control where they had better track at keeping good statistics.

https://www.researchgate.net/figure/Homicide-rate-in-South-Africa-1938-2003_fig1_301275939

https://www.jamaicaobserver.com/2023/03/11/a-sad-tale-of-two-islands-jamaica-and-trinidad/

USA also had higher poverty back than compare today as well.

So I ask better question, why Black americans had a high crime rate compare to South African bantus and Caribs despite Jim Crow harsh polices against them?

Expand full comment

Notice every time when Black americans kept to themselves they had a lower crime in their towns compare to living in White cities and towns in South or North. And those towns were actively destroyed or sabotage in someway by White supremacists. Liberia from 1960s again despite it's faults were relatively safe and had growing economy and was destroyed by CIA led Natives who turn it into a shithole. Or again Haiti during 1950s and early 1990s had coup back by Whites due to trying to take control of their economy and society. Same Grenada, same other countries in africa, LA, Asia.

Expand full comment