His point is based on the rates of crime commission, not the rates of incarceration. The source cited by the article you link to says that "Yes, African Americans and Latinos disproportionately commit certain crimes", and so actually provides some support for ESR's argument. They do say that for drug use crimes, the crime rates are about the same for blacks and whites, but ESR was specifically talking about violent crime.
Also, they say that the differences in crime rates by race can be explained by socioeconomic disadvantages. But among the effects of socioeconomic disadvantages are poor early childhood nutrition and more likely exposure to toxic environments (hello, old buildings with lead paint!). These things tend to hold back cognitive development, which tends to result in lower IQ scores. So to deny a link between lower IQ on average and higher crime rate on average is in effect denying a link between socioeconomic disadvantages and higher crime rates.
Whether or not someone is being racist when writing about the link between IQ and crime depends on what they think is the causality behind the link between IQ and socioeconomic disadvantages, and the factors behind whichever one of those they think is the cause of the other. ESR may indeed be a racist, but you've not shown it, and the source behind your link could be reasonably cited to support the ESR statement you quoted.
We need to find some way to get some statistical sense into people at an early age, so as they grow up it becomes routine and automatic to recognize when you should care about group averages and when you should not.
In almost every situation where you are dealing with individuals, the fact that one individual comes from a group whose bell curve for some characteristic of interest is slightly displaced from the bell curve of the group that some other individual comes from is meaningless trivia when it comes to deciding which individual to choose for something that depends on that characteristic.
