For centuries the death penalty, often accompanied by barbarous refinements, has been trying to hold crime in check; yet crime persists. Why? Because the instincts that are warring in man are not, as the law claims, constant forces in a state of equilibrium.
--Albert Camus (1)
Ever since a young Edwin Sutherland first published research on the issue in 1925, (2) criminologists have been interested in the question of whether the death penalty is a more effective deterrent to criminal homicide than long-term imprisonment. At least until a decade ago, there was widespread consensus among criminologists that the death penalty could not be justified on deterrence grounds. In November 1989, in part because "social science research ha[d] found no consistent evidence of crime deterrence through execution," the American Society of Criminology passed a resolution condemning the death penalty, one of only two public policy positions the organization has ever taken. (3) In 1996, Radelet and Akers surveyed sixty-seven leading American criminologists regarding their opinion about the empirical research on deterrence and found that the overwhelming majority of the experts agreed that the death penalty never has been, is not, and never could be superior to long prison sentences as a deterrent to criminal violence. (4)
The research reported in this Article was designed to update the 1996 study and assess if any recent deterrence studies have modified the beliefs of the world's leading criminologists. The results indicate that only a small minority of top criminologists--10% or less, depending on how the question is phrased--believes that the weight of empirical research studies supports the deterrence justification for the death penalty.
These results come despite the publication of several widely-cited studies conducted in the last half dozen years (primarily by economists) that claim to show the death penalty has deterrent effects that criminologists have not spotted. (5) In 2002, the Washington Post published an article under the catchy title Murderous Pardons? about research by econometrician Naci Mocan purporting to find that each execution led to 5-6 fewer homicides, and for every three additional "pardons" of a death row inmate, there were 1-1.5 additional homicides. (6) A few months later, Emory University economist Paul H. Rubin and his colleagues began to publicize their work which found that each execution deterred approximately eighteen homicides. (7) Later that year, Dale Cloninger and Roberto Marchesini, economists in the School of Business Administration at the University of Houston, published a letter in the Wall Street Journal claiming that their research showed that each execution in Texas prevented between eleven and eighteen homicides. (8) In 2007, Professors Roy Adler and Michael Summers (a professor of Marketing and a professor of Quantitative Methods at Pepperdine University, respectively) (9) published an op-ed in the Wall Street Journal claiming that their data showed each execution in the United States, from 1979-2004, prevented some seventy-four murders in the following year. (10) By late 2007, an article on the front page of the New York Times entitled Does the Death Penalty Save Lives? A New Debate announced that the recent articles on deterrence were "setting off an intense new debate about one of the central justifications for capital punishment." (11)
Are these new studies really "setting off an intense new debate"? What should the general public conclude about this morass of conflicting results and opinions? To be sure, most of the recent research that purports to find a deterrent effect has been critiqued (as we will discuss below), but that still leaves the layperson trying to decide between "he said, she said" exchanges and complex statistical debates that few can understand. Therefore, we decided to find some sort of answer by replicating the study conducted a dozen years ago by Michael Radelet and Ronald Akers in which they surveyed sixty-seven leading criminologists to see if there was consensus on whether the death penalty was superior as a deterrent to long term imprisonment. In this study, we use a different sample of expert criminologists to see if the opinions of the country's top criminologists have changed.
The importance of the deterrence justification for capital punishment has declined precipitously in recent years among the general public. In the mid-twentieth century (12) and up through the 1970s, it was unquestionably the top argument in favor of executions. (13) In a 1985 Gallup Poll, 62% of the respondents answered yes to the question, "Do you feel that the death penalty acts as a deterrent to the commitment of murder, that it lowers the murder rate, or not?" (14) This fell to 34% in 2006, (15) when the question was last asked. Conversely, the proportion of respondents who stated that the death penalty was not a deterrent doubled by 2004, from 31% to 62%. (16) Similarly, a 1995 national survey of nearly 400 police chiefs and county sheriffs found that two-thirds did not believe the death penalty significantly lowered the number of murders. (17)
No doubt part of this declining support for the deterrence hypothesis is a consequence of empirical research by criminologists. Led by the pioneering work of Thorsten Sellin, (18) scores of researchers have examined the possibility that the death penalty has a greater deterrent effect on homicide rates than does long-term imprisonment. (19) While some econometric studies in the 1970s claimed to find deterrent effects, (20) these studies were exhaustively criticized and largely discredited. (21) A panel set up by the National Academy of Sciences and chaired by Nobel Laureate Lawrence R. Klein to examine the studies--primarily those published by economist Isaac Ehrlich--concluded that "the available studies provide no useful evidence on the deterrent effect of capital punishment" and "research on the deterrent effects of capital sanctions is not likely to provide results that will or should have much influence on policy makers." (22) In retrospect, that finding seemed to settle the scholarly debate, at least for the next twenty-five years.
THE MOCAN-GITTINGS STUDY
Against this background, the article entitled Murderous Pardons? that was published by the Washington Post in 2002 raised the eyebrows of many criminologists. (23) The study discussed in the article was authored by University of Colorado-Denver economist Naci Mocan (24) and one of his (then) graduate students, Kaj Gittings. (25) They examined 6,143 death sentences imposed in the United States between 1977 and 1997, and built a data set with 1,050 observations (one observation per state for twenty-one years). (26) Their results indicated that each execution resulted in five fewer homicides, and each commutation (27) of a death sentence to a long or life prison term resulted in five additional homicides. (28) Further, each additional removal from death row--primarily occurring when appellate courts vacate death sentences that were imposed with various improprieties by trial courts--resulted in one additional homicide. (29)
At least two prominent criminologists have found serious flaws in the Mocan-Gittings work. Richard Berk noted that the execution figures by state by year for the 1977 to 1997 period were highly skewed. (30) Berk specifically noted that most states--accounting for 859 of the 1,000 observations (31)--had zero executions in a given year, and only a few states had more than a handful in a few years (n=11), with most of these being from Texas. (32) He used a straightforward procedure to assess the implications of this skewed measure: using Mocan and Gittings's original data set, he removed the Texas data and ran the model exactly as the original authors did, albeit only for the other forty-nine states. (33) The deterrent effect of executions disappeared. (34) Berk concluded that "it would be bad statistics and bad social policy to generalize from the 11 observations to the remaining 989." (35)
A second reexamination of the Mocan-Gittings study was conducted by Jeffrey Fagan. (36) Fagan's work is the most comprehensive review of the theoretical and methodological shortcomings of deterrence studies published after 2000. He first improved Mocan's measure of deterrence, which is the number of executions in a given state divided by the number of death sentences imposed six years earlier. (37) Because of the impossibility of computing this measure if the denominator is zero, Mocan and Gittings coded years with no death sentences as .99. (38) Fagan reanalyzed the data using .01 (which is closer to zero) in the denominator rather than .99. That simple improvement made all the deterrent effects found by Mocan and Gittings disappear. (39)
Furthermore, Fagan noted that potential offenders are unlikely to remember the number of death sentences imposed in their states six or seven years prior to their crime. (40) Instead, he computed a variable measuring deterrence by calculating the number of executions in the previous year divided by the number of death sentences handed down two years earlier (rather than six). Again, this minor adjustment makes the deterrent effect observed by Mocan and Gittings disappear.
Fagan also showed that alternative statistical models that consider the strong correlation of homicide rates from year to year within a given state also produce results that eliminate any deterrent effects. (41) In addition, because the data set used by Mocan and Gittings to count homicides has wide gaps with missing data, Fagan used Morbidity and Mortality data from the National Center for Health Statistics to improve the measure of homicides. (42) Again, these minor adjustments and corrections eliminated the relationship between executions and homicide rates.
Rather than prove that Mocan and Gittings erred in their...