Sentencing for Homicides in the U.S.

08/01/2000
Summary of working paper 7676
Featured in print Digest

Victims' race, age, and criminal record all determine sentence length, even in vehicular homicides. Drivers who kill black victims get substantially shorter sentences, but drivers who kill women receive significantly longer sentences.

Why do murderers receive different sentence lengths for the same crime? In The Determinants of Punishment: Deterrence, Incapacitation and Vengeance (NBER Working Paper No. 7676), Edward Glaeser and Bruce Sacerdote use Bureau of Justice Statistics data on murders to test the economic model of optimal punishment against alternative theories. The authors find much support for the optimal punishment view. For example, sentences are longest for crimes in which the apprehension rate is lowest, and criminals with higher rates of recidivism are incarcerated for longer periods of time. Yet certain aspects of punishment are hard to reconcile with the optimal punishment view. Characteristics of the victims including gender, race, and occupation seem to matter in a way that suggests that vengeance plays a role in determining sentences. The most striking fact is that even in vehicular homicides where the choice of victim is fairly random, offenders killing blacks receive sentences lengths that are 53 percent lower than the sentences received by offenders who kill whites. Offenders killing men receive sentences that are 56 percent lower than sentences received by offenders who kill women.

No matter what type of murder is committed, as the probability of apprehending the criminal rises, the length of the sentence falls. Glaeser and Sacerdote conclude that lovers' quarrels, arguments, and brawls involving alcohol have the highest chance of apprehension, and consequently the lowest sentence lengths. The crime of arson, which has an exceptionally low probability of apprehension, has the highest average sentence length. But "long prison sentences for arsonists may have more to do with the great potential for social damage created by arson," the researchers concede.

The authors point out that if victims are chosen randomly, victim characteristics would not affect the criminal's sentence length under an optimal system of punishment. To test this theory, they examine vehicular homicide data from the state of Alabama and the Bureau of Justice Statistics. Vehicular homicides typically involve substance abuse and reckless driving, and the victims are fairly random. Still, Glaeser and Sacerdote find that the drivers' sentences depend heavily on victim characteristics. According to their data, victims' race, age, and criminal record all determine sentence length, even in vehicular homicides. Drivers who kill black victims get substantially shorter sentences, but drivers who kill women receive significantly longer sentences.

How do the authors explain these findings? They propose that sentence lengths are driven, in part, by a taste for vengeance. They argue, "there is likely to be a much stronger visceral response to a drunk driver who accidentally kills an eight-year-old girl than to a drunk driver who accidentally kills a 22-year-old gang member."

-- Marie A. Bussing-Burks