While we are waiting for some response from Glenn Loury to the various points raised in this discussion, I thought that I would respond to the points raised by Bruce Western.
Does the Higher Black Crime Rate Explain the Higher Black Incarceration Rates?
Western cites evidence that 80 percent of the high black incarceration rate during 1979 and only 57.4 percent in 2004 can be explained by the rates that blacks are arrested. These claims are taken as evidence of racism in the criminal justice system. To reach these conclusions, the national prison and arrest rates for whites and blacks are compared. 
Yet, there are many obvious explanations for this so-called “unexplained” gap. It would be more accurate to say that this is a “not tried at all to be explained” gap.
For example, why would blacks who are arrested be more likely to be convicted and spend longer in prison than whites? Two main reasons are the criminal’s criminal history and the severity of the crime. If you are a criminal who has no prior offenses, you are just not going to get as long of a prison term as a career criminal. United States Sentencing Commission data for individuals who were sentenced in the federal courts between October 1, 1991, and September 30, 1994 shows that the average black had a criminal history of 2.4 offenses and the average white 1.8 — a 33 percent higher rate for black criminals.
There is an obvious reason why we punish criminals with a longer or more severe criminal history more heavily. Each additional time the criminal is arrested and convicted the less likely that you are to believe that the next time the criminal is in front of the judge he is there simply by accident. We will punish repeat offenders more heavily simply because we are less worried about punishing innocent individuals. Does anyone want to seriously argue that we don’t want to include criminal history in sentencing?
The severity of the offense will also matter. The overall result only relies on the national arrest and prison shares for each race, but data is reported by type board categories of crime (e.g., “drugs”). Even breaking the national numbers into the single crime category of “drugs” lumps together recreational users and crack cocaine dealers. Depending on the year, the drug category also implies either the largest or second largest black arrest and prison population disparity. Blacks were more likely to have larger amounts of drugs involved in their conviction (Lott, 1992). It is also well known that whites are more likely to be recreational users and blacks more likely to be dealers. Thus, even with the category of “drugs,” you would expect blacks to get longer sentences without any reliance on racial discrimination.
Finally, there are the number of offenses that one is sentenced to prison for. The evidence that Western points to only looks at the worst offense during a particular sentencing — not the number or nature of the other offenses that the criminal was convicted of. Blacks are simply more likely to be charged and convicted of more crimes. That hardly seems surprising since blacks are more likely to be members of criminal organizations such as gangs.
If someone actually did a serious job on all these issues, they would likely find that blacks are underrepresented, not overrepresented, in the general prison population for the crimes they commit.
Beyond these foregoing problems, the work of Tonry and Melewski that Western cites is quite selective in what they were willing to report. For example, while they note the large share of blacks in prison for the death penalty, they neglect to point out that whites have a much higher execution rate. The absence is striking because the rest of their discussion focuses on penalties.
The Cost and Benefits of Deterring Crime
Bruce Western accepts that we should consider both the costs and benefits of law enforcement, but he only provides dollar values for the costs of deterrence in his discussion. He writes: “The modest decline in serious crime over an eight-year period was purchased for $53 billion in additional correctional spending and half a million new prison inmates.” Surely, $53 billion ($67 billion in 2009 dollars) over eight years seems like a large amount of money, though even after adjusting for inflation it is about 5 percent of just two large spending bills that President Obama has just signed into law (the “stimulus” bill and the $410 billion spending bill for the rest of this fiscal year).
The drop in violent crime from its peak in 1991 is huge. Between 1991 and 1999, violent crime fell by 231 violent crimes per hundred thousand people (a 31 percent drop, falling from 748 to 517 per 100,000). The murder rate fell by 42 percent from 9.8 to 5.69 per 100,000 people.
Suppose that the murder rate in 1999 had been the same as it had been in 1991. There would have been about 11,200 more murders that year. In today’s dollars, value of saved life would be $45 billion (Lott, 2009). If increased prison accounts for 10 percent of that benefit, as I argued in my previous post, that is an annual benefit of $4.5 billion from just the reduction in murder ($36 billion = 8 years * $4.5 billion). Of course, there are academic estimates of the victimization costs of rape, robbery, aggravated assault, and various property crimes that would also have to be included, and a rough calculation indicates that this likely doubles the total benefit to over $75 billion over eight years.
The $67 billion expenditure that Western points to over 8 years also has two parts to it — a capital expenditure portion for assets that will last more than eight years and a yearly labor and maintenance cost. I don’t have the time right now to break these numbers down, but this should further increase the size of the benefits relative to the costs.
Other Parts of the Debate
The claim of an “unexplained” gap between blacks’ share of arrests and prison is inaccurate. There are many explanations for this “gap.” Putting dollar values to the costs of punishment and not the costs of crime is also not very useful in conducting any type of cost-benefit calculation.
Yet, the foregoing responses are still playing on Loury and Western’s very narrow field. The points that I have raised about the extremely progressive nature of the total criminal penalty are still being completely ignored. Prison is important but it is just one part of the criminal justice system. Loury raises collateral penalties such as voting rights when it suits him, but he ignores the extremely long list of other collateral penalties when it doesn’t.
In any case, I am still looking forward to Loury’s response to the various points raised.
 The way these numbers for 1979 and 2004 are calculated also exaggerate the size of the gap to readers. It isn’t important to the point that I am making here so I am not going to go through this other than to state the way that I think that these numbers should be reported. Take the results pointed to for 1979. The percentage point difference between the gap Black’s share of the prison population and their share of arrests is 5.69 percentage points. For whites, the same number is 5.70 percentage points. Adding both together indicates a difference of 11.39 percentage points. Since these numbers are already in percent, those differences should be reported that way to show the disparity.
Lott, John R., Jr., “An Attempt at Measuring the Total Monetary Penalty from Drug Convictions: The Importance of an Individual’s Reputation,” Journal of Legal Studies, Vol. 21, no. 1, January 1992: 159-187.
Lott, John R., Jr., More Guns, Less Crime: Analyzing Crime and Gun Control Laws, third edition, University of Chicago Press: Chicago, Illinois (2009 forthcoming).