Imagine the moments leading up to a murder. Are we to believe a person about to commit a homicide would consider whether or not his state has a death penalty as opposed to just life imprisonment without the possibility of release? Common sense suggests that lucid, rational calculations about capital punishment statutes are not among the things that would cross the mind of a would-be murderer. Yet as illogical as it sounds, new reports are asking us to accept this very line of reasoning.
The reports, including a Wall Street Journal op-ed by two professors from Pepperdine University, have revived the long-discredited argument alleging that capital punishment has a deterrent effect on homicide rates. They are certainly not precise or even consistent with each other, claiming that each execution deters anywhere between three and 74 murders.
Nevertheless, the new reports have captured widespread media attention, which in turn has accorded them a degree of unjustified legitimacy in the public arena. Detailed follow-up studies by prominent social scientists have demonstrated that the available data don't -- and because of inherent limitations probably can't -- support the conclusion that capital punishment deters murder.
The first glaring problem is the studies' failure to consider alternate explanations for declining murder rates in recent years. The possible alternatives include, among other things: the increase of life sentences without release, the improved police "clearance" rates for felonies, and the waning of acute drug epidemics such as the crack problem of the early 1990s. Any study that doesn't take into account these obvious factors can't be taken seriously.
Another problem -- inherent to any statistical study of deterrence -- is that the number of executions is small and insufficient to determine with any certainty the effect of executions on homicide rates.
Economist and law professor John Donohue of Yale University, together with economist Justin Wolfers of the University of Pennsylvania, conducted an exhaustive analysis of the datasets used by several deterrence proponents. They reported that, while the evidence pointed more strongly to an increase in homicides following executions, "there exists profound uncertainty" about capital punishment's deterrent or anti-deterrent effect because of the enormous disparity between the number of homicides and the number of executions.
Numerous other studies by distinguished social scientists have reached essentially the same conclusion -- that, like an election poll that samples too few voters, the limited available data render any conclusions meaningless. For example, Jeffrey Fagan of Columbia University, along with other experts, examined the data and found that each execution may save 10 lives -- plus or minus 14 lives. In other words, executions may result in deterrence, no deterrence or more homicides -- not exactly a confidence-inspiring result.
Moreover, the small number of executions makes any conclusions about their influence on homicide rates highly vulnerable to slight changes in the statistical models used. In one of many examples in their analysis, Donohue and Wolfers looked at a deterrence study that had attempted to account for a number of factors that might have influenced execution and homicide rates, including the effect of "partisan influence." This study had concluded that each execution decreases homicides by 18. When Donohue and Wolfers used a slightly different method to account for partisan influence, they found that each execution increases homicides by 18; when they dropped the factor altogether, they found that each execution adds over 50 homicides.
Like the old deterrence studies, the new deterrence studies have been thoroughly debunked -- there is no legitimate or reliable statistical evidence of deterrence. Capital punishment -- the deliberate killing of defenseless human beings -- cannot be statistically, morally or legally defended on the basis that it deters homicides.