Poverty and Crime In response to my post below, Skarl asks:

Higher rates of black crime can be explained very simply: blacks are on average poorer. Is there some air-tight debunk of this elsewhere?

The fact is that poverty does not explain crime. Let me quickly summarize the evidence. I've included all relevant sources so you can check my work. If poverty predicted crime, then we would expect the share of the violent crime rate to be the same as the share of the poor population (approximately). In other words, if 20% of the poor were black, we'd expect about 20% of violent crimes to be committed by blacks. This is not how things shape up. In the US in 2000, there were (in millions):

195 (non hispanic) white americans of which 7.5 % were poor --> 14.6 poor white americans, comprising 48% of poor 34 black americans of which 22 % were poor --> 7.48 poor black americans, 24% of poor 35 hispanic americans of which 21.2 % were poor --> 7.42 poor hispanic americans, 24% of poor 10 asian americans of which 10.7 % were poor --> 1.07 poor asian americans, 3.5% of poor Sources: Population data from US Census Table 24 and poverty data from US Census Poverty Tables.

Also in the US in 2000:

60% of violent crimes were committed by (whites + hispanics), Hispanics 12.5% and Whites 69.1 % of population ---> 60% of violent crimes committed by the (white + hispanic) 82% of population 40% of violent crimes were committed by blacks, 12.1 % blacks in US ---> 40% of violent crimes committed by the black 12% of population Source: FBI UCR, table 43 (Mixed Race, Asians, Amerindians omitted) Note that the FBI reports violent crimes for whites + hispanics together, which dramatically inflates the apparent rate of violent crime committed by whites.

Thus, what we actually find is that 24% of the poor population is black, but 40% of the violent crimes are committed by blacks. This means that poverty is an insufficient explanation for violent crime as we have a substantial mismatch between predicted (24%) and actual (40%) violent crime rates. It's more difficult to see what the violent crime statistics mean for non-blacks due to the FBI's tomfoolery in mixing white and hispanic offender data. It's almost certain that the white violent crime rate is far less than the white poverty rate - the opposite of the black trend. One other important point to note is just how high the black violent crime rate is in absolute terms. The black 12% of the population commits 40% of the violent crimes, while the other 88% of the population commits about 60% of the violent crimes. That means that the black violent crime rate (offenders per capita) is about (.40/.12)/(.60/.88) = ~5 times the non-black crime rate. Again, if the FBI didn't mix white and hispanic offender data, we could compare the black rate to the white/hispanic/asian rates. We would most likely find that the black rate is much higher than simply 5 times the white rate, as the above figure lumps hispanics with the non-black population. We would also find that the white rate is higher than the Asian rate, which is the lowest in the country. So what is the explanation? What we would like is an explanatory variable X. We'd like to say, "whites and blacks of equal X do just as well", meaning that they have (say) the same crime rates and the same incomes. Poverty didn't work as variable X, so what does? It's my opinion that the explanatory variable will be a combination of variables that predict criminality, such as low IQ and MAO promoter polymorphisms. Followup comments from Godless : Going to be busy for the foreseeable future, so posting intensity will be down. Most of the common objections have been addressed in the comments section. Here are some quick observations: 1) Are blacks falsely accused of crime? If so, then we would expect the race distribution of offenders as reported by victims to be different from that reported by law enforcement. This is not the case, as the two distributions track very closely. It seems reasonable to assert that even if the justice system leaves a lot to be desired, it isn't accusing hundreds of thousands of black men in place of white criminals. 2) Are there are non-genetic explanations for African-American criminality? Yes, there are, and some of them are plausible. But first, let us dispense with the ones that aren't plausible. I find poverty unconvincing both because of the above analysis and because blacks aren't that poor. As Glenn Reynolds has pointed out, the average African-American has a higher per capita income than Swedes! Urbans blacks aren't victims of grinding hunger-over half of black women are overweight. In terms of environmental quality, the ghettos of the early 20th century were surely safer than those of the late 20th century, yet the latter are richer materially (with cars, TVs, etc.). The black ghettos of the mid-20th century were safer than those of the late 20th century. And yet blacks have become wealthier and better educated. Thus "deprivation" is not a contributing factor in any obvious sense. Perhaps the most important environmental factor is the issue David raised: the preponderance of single-parent households. 60-70% of black children are born out of wedlock, and in a large portion of these the father is not in the picture. I have argued before that this maybe a response to forced imposition of Western norms (monogamy) on a group of people who have conventionally practice polygamy. But culture and genetics aren't always easy to separate in my opinion, as blacks might employ different mating strategies than whites. Polygamy might very well be a response to the conditions of Africa. (see James. Q. Wilson's article on this issue) 3) What then is your view on genetics and criminality? Only science can provide us with a clear answer. There are likely to be factors independent of IQ that contribute to criminality, such as genes that cause high testosterone levels. IQ is not the be-all and end-all of criminality research, though it is the be-all and end-all of income disparity research (as referenced in the Bell Curve and excerpted in the comments section.) The body of evidence to date is conclusive: criminality is partially genetic. Several independent lines of evidence point to this: -Low IQ strongly correlates with criminality because it indicates a reduced ability to make long-term decisions, and IQ is highly heritable. -Studies indicate that a biological parent's criminal record more closely predicts an adopted child's behavior than the adoptive parents environment. -Large scale statistical evidence (e.g. the crime rates of populations) -Physiological and genetic evidence from individuals and populations (e.g. salivary testosterone levels, MAO promoter polymorphisms) The case for a genetic contribution to criminality is conclusive. What remains to be figured out are exactly which genes contribute to criminality and how they exert their effects . These questions will be addressed with the tools of molecular biology, and it will be far more difficult to assail the results as "biased".