What Causes Crime?

The field of criminology is now flooded with some back and forth over the nature-nurture debate. Let’s try and settle some of it.

This post will not focus heavily on race and sex differences in criminality, but rather what does and doesn’t cause crime. It will overview many environmental theories as well as, more controversially, some theories related to biology and genetics. That said, we must come to the conclusion race and sex must be included. They are too important of factors, as will be displayed, that we can not mention them at all.

Environmental Effects

Single Motherhood:

A prominent argument made by pundits, like Thomas Sowell and Ben Shapiro, is that crime arises due to single motherhood – a position also supported by the Men’s Rights Activist Warren Farrell. In order to test this hypothesis, we can see what the correlation is between broken homes, growing up with a single parent and being a criminal later in the future.

According to Petrosino et al. (2009) which reviewed 5 meta-analysis on the relationship between delinquency and family structure, the correlations produced between someones family breaking up and them becoming delinquent or becoming a criminal later in life was .07, .09, .09, .10, and .10. Thus, the effects of single-motherhood and later-in-life delinquency/criminality was statistically weak. Table 2 gives a list of factors and their correlations:

lmadlmmadmadm.png
dslmlmmdsal.png

From this data, we can see that the correlation between growing up with a single parent and in a broken home have a weak correlation with becoming a criminal/becoming delinquent later in life.

Poverty:

As a starter, we should acknowledge current studies on poverty and crime do not control for a number of important variables, such as what might create poverty? Some of these poverty-creating factors are criminality itself, race, genetics, and IQ. Intelligence, as a factor, is highly heritable (Panizzon et al. 2012) and correlates with poverty to a high degree (Kodila-Tedika and Bolito-Losembe, 2014). Mani et al. (2013) claims poverty and economic pressure decrease IQ to a large degree, but suffers a number of methodological flaws (see this post).

One theory is that living in poverty has a larger environmental impact on intelligence than being wealthy does. Essentially, the heritability of IQ is higher for wealthy people than for those in poverty. The largest study to date (21,640 twins) puts this hypothesis to the test and they find there is no evidence for the theory (Figlio et al. 2017).

In addition, the heritability of living in deprived neighborhoods is found to be 65% in a twin study conducted by Sariaslan et al. (2016) and various factors related to poverty are heritable as well. This makes the relationship of poverty to crime possibly true, but not noting the more important, original factor. Hence, we advise taking caution in interpreting the results of current literature on poverty and crime.

lmdlmda.png

Raphael and Winter-Ebmer (2001) find that decreases in unemployment played a significant, causal, role in the drop of crime in the 1990’s, specifically on property crime rates but not on violent crime rates, while poverty did not have a large effect across the board.

One of the funniest studies I found was a measure of the effect of poverty on witch-killing. In this study they used the amount of excess rainfall or lack of rainfall to estimate when poverty would increase largely. They conclude that large droughts and floods resulted in the increase in killing of “witches” – aka elderly women – but not anyone else (Miguel, 2003).

Similarly, a study looks at the effect of rainfall on rye prices in 19th century Germany – an increase of which leading to more poverty – and how it related to property and violent crime. In terms of property crime, there was a small but significant effect, but there was typically a lack of effect on violent crime when the decrease of alcohol was controlled for (Mehlum, Miguel, and Torvik 2005).

Flango and Sherbenou (1976) claims to prove poverty is a large indicator of crime. Remarkably, they simultaneously state it explains about 25% of the variance on average. It may be the best predictor out of those which they tested, but it is no end-all be-all of the matter. Instead, it is still a fairly low correlation.

Pratt and Cullen (2005) conducted a very large meta-analysis on the effect of poverty on crime and report the following:

Sample:Effect on Crime:% which were significant
153 studies0.25359%

Perhaps one of the better meta-analyses of the data comes from Beaver, Ellis, and Wright (2009). In this they report, the following findings:

Sample:Economic Conditions:Effect on Crime:
17 studiesImprovesRises
10 studiesWeakensRises
5 studiesNo relationship

Similarly, Vieratis (2000) finds these results:

Type of Crime:Positive-SignificantPositive-InsignificantNegative-SignificantNegative-Insignificant
Homicide0.530.160.120.09
Rape0.270.370.090.27
Robbery0.230.230.030.51
Assault.10.70.000.2

The effect of unemployment on crime shouldn’t be ignored as it was found to be fairly significant in Raphael and Winter-Ebmer (2001). In regards to that, we may refer to Vieratis (2000) again as they meta-analyze a large number of studies and find the effect of unemployment is very small as well.

The Color of Crime (2005) analyzed 250 cities and ran a regression analysis between poverty rates and violent criminality. It found a low correlation of 0.36. It also analyzed the effect of unemployment on violent crime and finds a low correlation of 0.35.

Sariaslan et al. (2014) looked at over half a million Swedish people and found that people in poor SES were more likely to grow up to become criminals later in life when compared to those who grew up in higher SES. However, the same held true for the siblings of poor kids who’s families were becoming wealthier as they aged. This suggest that the likelihood of someone committing crime doesn’t magically go away as their SES increases. If better SES conditions did lower crime, then the studies outcomes should’ve been different.

From these results, it would appear poverty does not have a large effect on crime. Even a study which asserts it to be actually finds a low effect size of 25%. The results overall, once again, fail to include various genetic factors such as intelligence into the equation. Therefore, even these studies should be taken with a grain of salt as to how high the effect truly is.

Abortion:

This argument was most popularly made by Steven Levitt and John J. Donohue III in 2000. In this paper, the two argued the legalization of abortion that came with Roe v. Wade in 1973 was responsible for the drop in crime in the 1990s. The logic sounds weird at first but give it a listen and maybe it will follow. Less kids are born into bad conditions because they’re aborted -> less kids exist that would commit crimes -> less crimes exist. Now, it makes more sense. This theory is best contested by Steve Sailer (see Sailer [2005a] and Sailer [2005b])

But, this theory is based on a false premise that all children aborted would’ve lived in bad conditions. So far we have one major study (as far as I know) regarding this. What this study finds is only 28% of women have abortions because they can’t afford a baby right now (Finer et al. 2007).

lmfdlsmfdsm.png

Regardless of the amount of children that get brought out of poverty, this whole relay Levitt and Donohue rely on to explain how abortion leads to crime relies on the idea poverty must create crime.A meta analysis overviews different life factors and how they relate to delinquency later in life. Family socio-economic status is very poorly correlated (Derzon (2005).

Donohue and Levitt also fail to account for the amount of abortions happening before 1973. All Roe v. Wade did was make the first three months of abortion legal at the federal level, essentially taking away the states right to choose when abortions may happen. Plenty of states had abortion legal beforehand and even in general, abortions happened for the mothers’ health.

Lott and Whitley take this into account in their own analysis of the data and state,

“We find evidence that legalizing abortion increased murder rates by around about 0.5 to 7 percent.”

Technically Lott and Whitley only looked at murder rates according to that quotation. In addition, it should be noted this is implied to be causation, rather than a simple correlation. (Lott and Whitley, 2005)

Furthermore, we all know population has been massively increasing, especially since the 1990s. Despite abortion being legalized and the citizens depopulating themselves, immigrants still come in and plenty of children are born. Between 1990 and 1999, nearly 10 million legal immigrants came into the United States, generally known to be disproportionately young men. This on it’s own would make up for a large amount of the children who disappeared. Not to mention the 20 million other new people who were accounted for in the population that year.

Levitt and Donohue’s analysis doesn’t take into account a number of important variables and mishandles the data according to Sailer, so it should likely be thrown out.

Lead:

Lead is one of the environmental hypotheses with more empirical merit.

Feigenbaum and Muller (2016) proposes lead exposure at earlier ages plays a large role in homicide rates later in life. They looked at areas with more lead piping compared to those without lead piping and compared the crime rates of the populations 20 years later. They found a strong relation of lead early in life to homicide rates later in life.

While this critique holds true for most lead-crime studies, it is important to mention it here. Lead is a cheaper material compared to steel hence its usage for piping may more be related to poverty (or those primitive causes which influence poverty). Therefore, the relation of lead to crime may be a middle variable, rather than the full causal story.

To possibly counter-act this, another study looks at lead exposure in people’s blood and finds a strong relation of it to school suspensions and juvenile detention (Aizer and Curry, 2017). That said, higher levels of lead in someone’s blood may be because of something else such as poverty. In fact, it likely is. But, even still, this study fights against that. First, they show that even across different economic classes, lead exposure has an effect. They also draw causal evidence by measuring the effect of de-leading gasoline in local communities.

In relation to race, the effect seems smaller on black people, those who would be most affected by this issue, than it does on whites. As they say in the study,

“For example, a unit of lead increases the probability of incarceration for black males by only 6.6 percent compared to 16.6 percent for white males.“

A meta-analysis by Marcus et al. (2010) looks at 19 different studies that review the relation of lead exposure to later-in-life delinquent conduct. In regards to this, they find a significant correlation of 0.19, indicating lead exposure explains 19% of the variance in delinquent conduct among adults.

Mielke and Zahran (2012) looks within cities and finds a relationship between variation of lead exposure and crime, stating in the Abstract,

“Other things held equal, a 1% increase in tonnages of air Pb released 22 years prior raises the present period aggravated assault rate by 0.46% (95% CI, 0.28 to 0.64). Overall our model explains 90% of the variation in aggravated assault across the cities examined. In the case of New Orleans, 85% of temporal variation in the aggravated assault rate is explained by the annual rise and fall of air Pb (total=10,179 metric tons) released on the population of New Orleans 22 years earlier.”

Another longitudinal analysis done by Wright et al. (2008) overviews a large, supermajority black sample and shows that 5 micrograms of lead exposure plays a large impact in the violent crime rates among the sample. Overall, I would argue, lead plays a big impact on crime.

Sean Last has previously argued that while lead exposure plays a significant role in crime, it doesn’t play a large role into the racial differences in crime,

Given this, even though lead impacts crime, the fact that the races barely differ in terms of lead exposure as children and not at all as adults suggests that lead probably plays little to no role in the American Black-White IQ gap. This conclusion is consistent with Feigenbaum and Muller (2016) and Nevin (2007) who found that the proportion of an area which was Black continued to predict its crime rate even after its degree of lead exposure was controlled for. On the other hand, lead exposure is more common in the developing world and more play a role in the high crime rates of nations in Africa.”

Sean Last cites Tsoi et al. (2016) and Pirkle et al. (1994) to make a graph of the rate by which blacks and whites are exposed to lead:

dsmfsmlfdmld.png

But, this looked at the total population rather than just at children. In Pirkle et al. (1994) they say that in 1975, the mean blood-lead levels for white children was 13.7ug and was 20.2ug for black children, leaving over 6ug in between the two. Hence, the exposure is still significantly higher for black people and could translate to a good amount of the differences in crime rates then. As for more modern estimates, as the gaps of lead exposure have closed between blacks and whites, only time will tell, but it will be no shocker as we continue to see black people having crime rates significantly above that of white people. The Color of Crime (2016) shows some evidence that things really aren’t changing.

The blogger NotPoliticallyCorrect responds to Sean Last with a couple of points, largely a semantic dispute as well as a point on how Last need not mention that differences in lead exposure disappear by adult-age as the development during childhood is most important, hence Last’s argument faces some flaws,

“Sean Last argues that, while he believes that lead does contribute to crime, that the racial gaps have closed in the recent decades, therefore blood-lead levels cannot be a source of some of the variance in crime between blacks and whites, and even cites the CDC ‘lowering its “safe” values’ for lead, even though there is no such thing as a safe level of lead exposure (references cited above). White, Bonilha, and Ellis Jr., (2015) also show that minorities—blacks in particular—have higher rates of lead in their blood. Either way, Last seems to downplay large differences in lead exposure between whites and blacks at young ages, even though that’s when critical development of the mind/brain and other important functioning occurs. There is no safe level of lead exposure—pre- or post-natal—nor are there safe levels at adulthood. Even a small difference in blood lead levels would have some pretty large effects on criminal behavior.”

As I stated before, this is largely a semantic dispute. Sean Last never makes the claim there is a safe level of lead exposure and I doubt anyone would seriously argue that. Last even uses quotation marks around the word “safe” to be sure it is not confused that way. NotPoliticallyCorrect continues,

“Sean Last also writes that “Black children had a mean BLL which was 1 ug/dl higher than White children and that this BLL gap shrank to 0.9 ug/dl in samples taken between 2003 and 2006, and to 0.5 ug/dl in samples taken between 2007 and 2010.” Though, still, there are problems here too: “After adjustment, a 1 microgram per deciliter increase in average childhood blood lead level significantly predicts 0.06 (95% confidence interval [CI] = 0.01, 0.12) and 0.09 (95% CI = 0.03, 0.16) SD increases and a 0.37 (95% CI = 0.11, 0.64) point increase in adolescent impulsivity, anxiety or depression, and body mass index, respectively, following ordinary least squares regression. Results following matching and instrumental variable strategies are very similar” (Winter and Sampson, 2017).”

What he is saying here is even this 1ug difference in blood exposure to lead can create large effects on the black-white criminality gap. Of course, we don’t know yet how much the lead will actually have an effect on the black-white crime multiples as we haven’t hit the point where these children are at peak-crime committing age. But, I would assume things aren’t really going to change much.

All in all, lead does seem to be a pretty big determinant of crime. It would be fair to rule this as something that of course needs to be fixed. The effect it has on the black-white crime gap is much more shaky, but it does seem to play some role there. Though as some evidence has shown (as Last points out) the gap remains even after controlling for lead exposure.

Church Going and Religiosity:

To be honest, I expected there to be not a huge effect. And I’m not wrong – the effect isn’t huge. But, it’s there.

Johnson (2002) measures the effect of religious involvement in lowering African American crime rates. They find a negative, but fairly small correlation between religious involvement and crime rates. The higher the level of neighborhood disorder, the higher the effect of religious involvement – but as I stated, the relationship stays pretty small. Most trials were statistically insignificant and those which were barely significant, depending on your scale (p=0.06). This study provides some framework for an effect, but not a very large one.

Demond et al. (2010) finds religious youth more likely to have higher self control and this in turn causes them to be less likely to do drugs. Similar results are reviewed in Chitwood et al. (2008). Wallace et al. (2007) finds when schools are more religious, the students are less likely to engage in marijuana use, drink alcohol, and smoke cigarettes and similar outcomes are found among religious people in general.

Ulmer and Harris (2013) include race into the equation of how much religiosity affects criminality. They find that religious homogeneity results in statistically significant reductions in black and Latino criminality. They also find the higher the country’s population that is religious, the lower the crime rate is for whites and blacks. I’d note the effect size is fairly low, ranging in the r^2 range of 0.1-0.3.

Olson (1990) finds similar results to that which we have seen – church attendance is negatively correlated with criminality. They note as well this is more prominent with Protestantism than with Catholicism.

Overall, I’d note religiosity does seem to have some effect on criminality, but it’s worth being wary of the effect size of the variation and the significance. Ulmer and Harris (2013) and Johnson (2002) both found some statistically insignificant results (both of which they reported fairly.

Though when it comes to racial differences in crime rates, there seems to be an inverse correlation between religiosity and crime rates, meaning this would probably not be an explanatory factor for racial differences in crime rates. Pew Research Center (2014) finds black people are the most likely to believe in God, with mixed race people, white people, and Asian people following successively. Chatters et al. (2009) finds black people and Latinos have higher levels of religiosity (ergo more religious devotion) than white people as well.

Self Esteem:

A meta analysis by Mier and Ladny (2017) overviews 42 different studies and argues a self esteem boost would result in a drop in criminality.

In addition to this, Oser (2006) conducted a study on 134 individuals, controlling for a number of variables, and finds criminals have a lower self esteem. They find that criminals who do better in prison (more participation, emotionally stable, etc.) have higher self esteem.

Trzesniewski et al. (2006) finds low self esteem during childhood predicts worse outcomes later in life – criminality, lower earnings, etc.

Baumeister et al. (1996) rejects the popular notion that low self esteem causes crime. This study finds that egotism and feelings of self-superiority are better predictors of criminality.

Ostrowsky (2010) argues the current evidence is much more mixed. There seems to be a lot in favor of the idea lower self esteem causes violent crime. Simultaneously, there is evidence that narcissism has a strong link to crime (see Blinkhorn et al. 2017 [also Abbey Sereno links narcissism to antisocial behavior, hence why it would lead to criminality).

A very recent study, Kalemi et al. (2019), looks at female prisoners and finds higher levels of aggression of prisoners compared to non-prisoners could be predicted through higher levels of narcissism, sociability, and lower self esteem (among other traits).

Stanton Samenow from Psychology Today argues studies on the self esteem of criminals being lower is a reversal of cause and effect:

“Such thinking inverts cause and effect. In most instances, the criminal has rejected his family, teachers, and the world of work long before they ever rejected him. By refusing to cope with adversity constructively and by exerting little to no effort in responsible endeavors, he has accomplished little that is substantive. If a person throws away opportunities and resorts to deception, intimidation, or force to make his way in the world, is it not realistic for him to have low self-esteem, at least by the standards of the responsible world?”

I’d add self esteem is relatively heritable, estimates ranging from 0.3-0.8 (Roy et al. (1995);Raevuori et al. (2007);Jonassaint et al. (2010)) so considering this an entirely changeable, environmental effect would be incorrect.

Racial Demographics:       

It is generally recognized that blacks commit a disproportionate amount of crime. This is not a trend only noticed today, but all throughout American history.

In the 1850s, Frederick Law Olmstead traveled the South and wrote three books describing what he saw on his tour. He noted that slaves were generally unproductive and avoided actually working. Olmstead says, “The overseer rode about among them, on a horse, carrying in his hand a rawhide whip, constantly directing and encouraging them; but, as my companion and I, both, several times noticed, as often as he visited one end of the line of operations, the hands at the other end would discontinue their labor, until he turned to ride towards them again.”

Slaves often skipped work, according to Olmstead, as he says, “[t]he slave, if he is indisposed to work, and especially if he is not treated well, or does not like the master who has hired him, will sham sickness – even make himself sick or lame – that he need not work.”. Russell also reports of actual empirical data that might suggest slaves very commonly faked sick to leave work:

“On the Wheeles plantation, one out of seven working days was lost to slaves claiming they were too sick to work. On the Bowles plantation, of the 159 days missed due to illness in one year, only five were Sundays, when there was the least work to do. The Leigh plantation, where only thirty slaves worked, reported 398 sick days in one year. At these plantations, the rates of sickness peaked on Saturdays and during the planting and harvest seasons, when there was the most work to be done.”

Later on, Russell describes how the slaves were not exposed to a culture of self control and many thought this had a lot to do with how they were often misbehaved – often being hedonistic and licentious. Slaves were not held under the same laws as whites and hence were not forced to respond to the same culture. This caused many slaves to sleep around (even before marriage), get divorced more, and dress promiscuously. Russell continues as well,

“Angry slaves were also dangerous slaves, and in addition to the documented cases of slaves’ lethal vengeance, there were many stories of poison or ground glass mixed in with the master’s food and white children under the care of slave women who died unexpectedly.”

This itself is pretty disturbing.

Anyways, it does seem slaves were not the best behaved and generally not conforming to white values, which were much stricter and constrained. This is possibly evidence that blacks were more likely to commit violent crimes and acts at the time.

I also looked to the book, Violence and Culture in the Antebellum South, for some evidence. This provided some of the most telling evidence the slaves had higher rates of violence than whites, besides the rebellions themselves. The author, Bruce Dickson Jr., describes that slave masters would sometimes pit slaves against each other for fun (think to that scene in Django: Unchained) but that a lot of the time, the masters didn’t need to, because slaves would fight each other regardless and the master trying to get them to fight, just made them not want to. One paragraph stuck out to me in particular:

“Fights occurred in the slave community for a variety of reasons, and only a few ex-slaves agreed with the assessment, approved by Southern whites, that “slaves warn’t civilized folks den – all dey knowed was to fuss and fight and kill one ‘nother.” In general, Eugene D. Genovese’s comment that violence grew out of a “flash of passion” seems to account for most of the fighting that took place. Jealousy, a spur of the moment quarrel, or even fighting for fun were all causes of violence – and the problem of motives that so troubled elite white Southerners was not so much importance to those who talked about violence in the slave community.”

Dickson goes on to say that masters attempts to limit violence among slaves was largely appreciated by ex-slaves, and that many who “looked back at on the institution with some fondness, this was its main selling point.” After the Civil War, Dickson describes that ex-slaves resorted to the same “flashes of passion” as before for the violence in their communities. He says,

“Disputes could start over the merest trifles, and fighting out differences when they arose was a normal, even an encouraged mode of action for children and adults alike.”

Slaves and ex-slaves alike got into petty fights and fights for personal purposes very commonly. They were certainly violent against each other and this seems to show here. For a full description, I’d recommend reading the chapter “Slavery and Violence: The Slaves’ View” from the book for a fuller description.

The violence displayed by slaves is incredibly prevalent to this day. Even since before segregation, blacks have committed a disproportionate amount of crime relative to their population, but to the degree of its disproportionate has fluctuated over time.

African Americans as a Percent of State/Federal Prisoners and the General Population
DecadePercent of State and Federal PrisonersPercent of General Population
1920’s21%9.9%
1930’s24%9.7%
1940’s30%9.8%
1950’s30%10%
1960’s33%10.6%
1970’s38%11%
1980’s44%11.7%
1990’s50%12.1%
2000’s43%12.3%
2010’s37%12.6%
Chart borrowed from The Alternative Hypothesis

The Color of Crime (2016) might provide one of the most extensive analyses of racial differences in crime rate. Edwin Rubenstein uses NCVS data to compile data for New York City, Chicago and California and the multiple of the white rate each race/ethnicity produces. This is particularly more important than the current FBI data available as most police stations lump Hispanics in with whites as perpetrators, hence skewing the data (an effect called the Hispanic effect). At the same time, many are completely okay with separating the Hispanics for victims.

Rubenstein hence comes up with the following data:

It also reports the following main findings:

The evidence suggests that if there is police racial bias in arrests it is negligible. Victim and witness surveys show that police arrest violent criminals in close proportion to the rates at which criminals of different races commit violent crimes.

  • There are dramatic race differences in crime rates. Asians have the lowest rates, followed by whites, and then Hispanics. Blacks have notably high crime rates. This pattern holds true for virtually all crime categories and for virtually all age groups.
  • In 2014 in New York City, a black was 31 times more likely than a white to be arrested for murder, and a Hispanic was 12.4 times more likely. For the crime of “shooting”—defined as firing a bullet that hits someone—a black was 98.4 times more likely than a white to be arrested, and a Hispanic was 23.6 times more likely.
  • If New York City were all white, the murder rate would drop by 91 percent, the robbery rate by 81 percent, and the shootings rate by 97 percent.
  • In an all-white Chicago, murder would decline 90 percent, rape by 81 percent, and robbery by 90 percent.

Some commentators may be quick to point at Rubinstein’s data and how it comes from AMREN, a “white nationalist” source – making the data unreliable. If this is the case, then it seems weird that Beaver, Ellis, and Wright (2009) also found that blacks commit more crimes than whites around the world.

4.jpg

I know of a few meta analyses of % black of an area compared to crime or violent crime.

Unz (2013) meta-analyzes % black in a city to violent crime and finds a correlation ranging from 0.75-0.85.

The Color of Crime (2005) finds a correlation of 0.81 for % black/Hispanic and violent crime (n=50).

Bartels et al. (2010) finds a correlation of % black to murder of 0.57 and to robbery of 0.61. The other two correlations were statistically insignificant.

The correlation between murder and percent black for the 50 states is .77 (Whitney 1995). It rises to .82 when Washington D.C. is added (Whitney 1990b). For the 170 most populous cities the correlation is .69 (Whitney 1995). In the same cities, the correlation between percent black and robbery and rape is is .75 and .52, respectively (Kleck and Patterson 1993).

I tried to run a state by state regression analysis on this myself but 1) state by state regression analyses are already flawed because there is so much within-state difference, essentially making the results kind of weird and 2) the results were statistically insignificant regardless. This is what I know of so far in relation of % black to crime.

In terms of Hispanic crime…

In 2010, conservative writer Ron Unz wrote an essay for The American Conservative called “The Myth of Hispanic Crime” where he argued Hispanics did not have crime rates that far above whites. He said the right wing hysteria about this was usually based on anecdotes and bad data. The main criticism Unz gives on the data comparing whites to Hispanics is that they don’t account for age. The average Hispanic age at the time was around 29, at the peak crime committing age, while the average white age was around 45, after their crime committing age.

So, Unz attempts to find the best data he can. The conclusion he comes to: Hispanics don’t commit much more crime than whites. The rate is somewhat higher in certain areas and also lower in certain areas. As a whole, not much higher. This is represented in this graph for example:

hispanic crime rates compared to whites.gif

This evidence outraged conservatives at the time. It didn’t take long for many to write responses back, most of which Unz responded to, but was not happy about for their shrillness. One, in particular actually came through as a good response – a VDare article by Edwin Rubenstein.

Rubenstein’s article utilizes data on the incarceration rates of different racial/ethnic groups brought about by the Bureau of Justice Statistics which actually separates whites and Hispanics as perpetrators unlike the FBI’s Uniform Crime Report. He states,

While the Hispanic/white differential shrinks somewhat in high-crime age brackets, it remains in the vicinity of two to one:

Sentenced male prisoners under state orfederal jurisdiction, December 31, 2008
White non-HispanicBlack non-Hispanic Hispanic
Number
Total—all ages477,500562,800295,000
Ages 25-2966,000102,80060,000
Ages 30-3470,70096,80054,400
Ages 35-3975,20090,50045,900
Rate per 100,000 U.S. residents
Total—all ages4873,1611,200
Ages 25-291,0177,1302,612
Ages 30-341,2178,0322,411
Ages 35-391,1717,3922,263
Data Source: BJS, “Prisoners in 2008,” December 2009. Appendix Tables 13 and 14Note: Imprisonment rates are per residents in each population group.

Rubenstein throws this away right away though, because he admits Unz’s point against this data: it includes illegal immigrants as well as people in prison for immigrating; it does not limit itself to the crimes where there is an immediate victim or to people who got here by illegal means. Now, Rubenstein also acts like this doesn’t matter by sarcastically talking about it as if it does matter. But it does, because it doesn’t actually tell us about the real violent crime rates and property crime rates for example that we’re actually wary of.

Rubenstein then presents data to argue Hispanics commit violent crimes at a higher rate,

Sentenced Prisoners in State Prisons by race, Hispanic origin and offense, 2006 and 2006
20002006Change,2000-06% change
All inmates1,206,4001,331,100124,70010.3%
    Violent589,100667,90078,80013.4%
    non-violent(a)617,300663,20045,9007.4%
White, non Hispanic436,700474,20037,5008.6%
    Violent212,400227,50015,1007.1%
    non-violent(a)224,300246,70022,40010.0%
Black, non Hispanic562,000508,700-53,300-9.5%
    Violent273,400267,900-5,500-2.0%
    non-violent(a)288,600240,800-47,800-16.6%
Hispanic178,500248,90070,40039.4%
    Violent87,100141,60054,50062.6%
    non-violent(a)91,400107,30015,90017.4%
a. Property, drugs, and other non-violent crimes.Data Source: BJS, “Prisoners in 2008,” December 2009. Table 7. 

The number of Hispanics serving time for violent crimes increased by 63% in just six years, 2000 to 2006. Over the same period the number of whites incarcerated for such offenses rose by only 7.1%, while the corresponding number of Blacks actually fell by 2%.

Put differently, Hispanics account for more than half (56%) of the recent increase in violent crime. That is four times larger than their share of the adult population.

But, first of all this doesn’t address Unz’s main concern which is that the age groups are not being separated. This is not going to effectively and fairly allow us to compare crime rates by race. Unz also brings up in his response to this article how the BJS data does not take into account the people incarcerated in local jails which is what Unz analyzed.

Unz addresses the argument at the end that Hispanics are largely responsible for the increase in violent crime,

“Much more importantly, a central argument of my article had been that ethnic criminality is best inferred not from raw incarceration statistics but from incarceration rates relative to the size of the high-crime age ethnic population; and even just during the few years from 2000-2006, there was actually a substantial change in the relative numbers of Hispanic and white males aged 15-44 (a demographic trend which should come as no great surprise to regular VDare readers!).

Based on the Census/ACS data, the number of white males in that age range dropped by almost 10% during those years, while the corresponding number of Hispanics increased by around 15%. Thus, the age-adjusted increase in violent white inmates was actually between 15 and 20% and for Hispanics was about 40%. Although we still find a significant divergence in white/Hispanic violent criminals, simply adjusting for age cuts the nominal gap in half.”

To be fair, Unz doesn’t adequately address the rest of the 20%, but we will ignore that for now.

For now, it seems Rubenstein’s 2010 article does not argue against Unz’s essay very well. The data he presents either includes the people we shouldn’t be concerned with for America’s safety or conflates age groups with each other whilst ignoring the adequate points Unz makes against this.

Later, Edwin Rubenstein utilized NCVS data to write The Color of Crime 2016. This study analyzed crime data by race/ethnicity in Chicago, New York City, and in California. In New York City, the Hispanic multiple of white arrest rate ranged from 1.9x for grand larceny to 23.6x for shooting. In Chicago, the Hispanic multiple of the white arrest rate ranged from 1.2 for larceny to 6.7 for murder.

Unz took this data as pretty critical and to be very good. But, for every area except the state of California, the rates were not age-adjusted. For California, Unz notes, the multiples for Hispanics were not too far off from Unz’s estimates. They were also not very much higher than the white rate and nowhere near as high as the black rate. In violent crimes, Hispanics on average only had a multiple of 1.4 times the white rate.

The state prison estimates, while not age adjusted did show varying disparity in white vs. Hispanic crime. But, of course, a good portion of this would disappear if it had been age adjusted.

Unz also notes the varying heterogeneity in Hispanics of the main areas. In Unz’s original essay he writes areas that are more Carribean will have higher crime rates while Meso-Americans don’t seem to have that problem. This hypothesis might hold. Chicago’s Hispanic community is nearly 75% Mexican, followed by Puerto Ricans, Ecuadorians, and Guatemalans, the only group from the Carribean from this being Puerto Ricans which was 15% of all Latinos there. New York Hispanics are over 58% Caribbean and California is only 11% non-Mexican. The crime multiples seem to match up somewhat to these Caribbean percentages.

It seems odd this would be causation. Mexico’s crime rates are massively higher than Puerto Rico’s. But for now, we can accept it as one theory as to why the crime rates differ between these areas. Of course this also shows you can’t say all Hispanics have low rates of crime – Caribbean Hispanics are still an issue but you can say Meso-Americans may not commit so much crime.

Overall, the Unz-Rubenstein debate does not go to Rubenstein’s favor. Unz’s arguments about the age groupings still have to be disputed and of course haven’t been. Worth noting as well, using state by state data isn’t the best for finding correlations because of massive differences in lifestyles among states – California and Texas being the biggest examples of this. But using California, we still find Hispanic crime rates are not much higher than white rates, especially in comparison to blacks.

Overall, demographics play a major role in the crime in an area.

Bio-Criminology

Heritability of Criminality:

For this section, we have made it easiest by creating a graph of the heritability of traits which either may lead to criminality, are likely to lead to criminality, or are purely criminal acts. Heritability in the broad sense is simply the phenotypic variation in a trait which is due to genetic influence.

Trait/Crime:HeritabilityStudy:
Living in deprived neighborhoods0.65Sariaslan et al. (2016)
General intelligence0.86Panizzon et al. (2014)
IQ0.85Bouchard and McGue (2003)
Schizophrenia, bipolar disorder, substance misuse, and violent crime0.53-0.71Sariaslan et al. (2016)
Violent crime0.46Frisell et al. (2012)
Incarceration0.6Boutwell and Connolly (2017)
Aggression0.5
0.5-0.8
0.41
0.27
0.69 (boys) – 0.72 (girls)
Rushton et al. (1986)  
Porsch et al. (2016)
Niv et al. (2014)
Olson et al. (1998)
Bartels et al. (2003)
Impulsiveness0.23-0.45  
~0.5 0.26
Plomin et al. (1998)  
Keller et al. (2005)
Zietsch et al. (2010)
Olson et al. (1998)
Extroversion0.53  
0.58-0.69 (men)
0.6-0.71 (women)
0.5 (men) – 0.58 (women)
Jang et al. (1996)  
Viken et al. (1994)
Pedersen et al. (1980)
Self Esteem0.52  
0.32 (girls) – 0.82 (boys)
0.77
Roy et al. (1995)  
Raevuori et al. (2007) Jonassaint et al. (2010)
Rule Breaking0.41  
0.56 (girls) – 0.79 (boys)
Niv et al. (2014)  
Bartels et al. (2003)
Sexual offense0.4Langstrom et al. (2015)

Neurology of Crime:

Criminals, on average, are typically anti-social – which would make sense considering that criminals have impaired brains. These brain differences will deviate from the average non-criminal brain, and this will cause crime differences.

Lencz et al. (2000) found that people with persistent anti-social behavior had an 11% reduction in the volume of gray matter in the pre-frontal cortex. Yang and Raine (2009) found that in their meta-analysis, the pre-frontal cortex of the brain is structurally impaired in offenders.

The amyglade, which is a set of neurons located deep in the brain’s medial temporal lobe that helps process emotions, is impaired in psychopaths. On average, psychopaths had an 18% reduction in the volume on the right side of the amyglade (Yang et al. 2009; Paradini et al. 2014). People with cavum septi pellucidi maldeveloppment (the CSP is the space-like gap in the middle of the brain) were prone to psychopathy, antisocial personality disorder, and had more charges and convictions for criminal offenses (Raine et al. 2014). This maldevelopment has been linked to lifelong anti-social behavior (i.e. reckless disregarded for lack of others, lack of remorse and aggression).

Aggression-Related Genes:

It should be no surprise that aggression is in the genes. Recently, 40 genes related to aggression have been found in both humans and mice (James et al. 2018). In a review of the MAOA gene, it was found that it does influence aggression, although the environment may mediate such effects (Sohrabi 2015). The DAT1 gene is also a gene that influences aggression, as found in Guo, Roetegger, and Shih (2007).

While we have found some genes that influence aggression, we don’t need to find each gene to infer the heritability of a trait such as aggression. Using twin and adoption studies, we have found that the heritability of aggression is 50%; this means that 50% of aggression in humans can be attributed to genetics (Tuvblad and Baker 2013).

Great caution should be taken when citing one single gene for aggression. Because a gene may influence aggression, this doesn’t mean that the specific gene will explain a large variance in aggression. Regardless, we can hypothesize that these genes might differ by sex.

Men are more aggressive than women, which explains why more men are in jail than women. It might be that these genes also differ by race, which explains why certain groups are in jail more.

Conclusion

Overall, there is clear evidence of a few different causal explanations for criminality. The main ones seem to be lead, demographics, and genetics. Many popular environmental explanations explain little to none of the variance in criminality, and are relatively weak.

This is a collaborative post between Unorthodox Theory and DissAcad.

You can find the former at: https://radicalstatistica.wordpress.com/

and the latter at: https://dissentingacademia.org/

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s