Capital Punishment as a Deterrent: What the Research Shows

Introduction: Inconsistency in the Death Penalty Debate

When it comes to the death penalty, the United States is anything but consistent. Although the country has leaned toward capital punishment for much of its history, there are periods when the trend moves in the opposite direction. Overall, the annual number of death sentences dropped by approximately 60% since the 1990s, when it was close to 300 (Death Penalty Information Center, 2008). Even though polls are not empirical research, Americans continue to view death as the necessary punishment for certain crimes. The last Gallup Poll conducted in October 2008 found support for capital punishment at 64%, a decline from the 69% support recorded in 2007 and considerably down from 80% in 1994.

In the years surrounding this debate, U.S. Supreme Court Justice Stevens called the death penalty "the pointless and needless extinction of life" and declared that he no longer believed it to be constitutional. The question of whether the death penalty functions as a deterrent is equally unsettled. Researchers who believe it deters crime and those who do not can each point to studies supporting their position, resulting in what amounts to a "he said, she said" situation. As Delegate Donald B. Elliott questioned: "Why are we studying this again?" (Olson & Dechter, 2008). Researchers continue to study the issue, and the results remain persistently inconsistent, leaving it to each individual to determine what to believe.

Early and Foundational Research on Deterrence

One of the first decisive studies in favor of the death penalty as a deterrent was conducted in 1975 by Ehrlich, who concluded that executions do prevent murders. Hood (2002) reported that data collected from 1973 to 1984 showed murder rates in states without the death penalty were consistently lower, averaging 63% of the corresponding rates in states that retained it. By contrast, Thorsten Sellin (1959) was among the earliest researchers to compare abolitionist and retentionist states. He found that homicide numbers in abolitionist states were not significantly different from those in retentionist states and concluded that executions do not have a meaningful impact on homicide death rates.

Researchers have also examined the death penalty's impact by analyzing a single state and comparing outcomes during periods when capital punishment was and was not in effect. For example, in 1999, Sorensen et al. tested the deterrence hypothesis in Texas, which carries out the greatest number of executions in the United States. Looking at results from 1984 to 1997, the authors found that "within a context so ideally suited for finding any potential deterrent effects, this study confirmed the results of previous ones that failed to find any evidence of deterrence resulting from capital punishment" (Sorensen et al., 1999, p. 483).

State-Level Studies and Mixed Findings

Mocan and Gittings (2003) arrived at contrary results. They merged a state-level panel data set that included crime and deterrence measures and state characteristics with information on all death sentences handed down between 1977 and 1997. Because the specific month and year of each execution and removal from death row could be determined, these events were matched with state-level criminal activity during the relevant time periods. The authors examined homicide rates based on execution rates, commutation and removal rates, homicide arrest rates, sentencing, imprisonment, and prison deaths. Their results showed that each additional execution reduced homicides by approximately five, each additional commutation raised homicides by a similar amount, and each additional removal from death row led to one additional murder.

In a similar vein, Dezhbakhsh and Shepherd (2003) examined data for all 50 states during the period from 1960 to 2000, comparing murder rates immediately before and after changes in state death penalty laws. Their findings pointed in the opposite direction from the most recent studies at the time: "Results suggest that capital punishment has a deterrent effect, and that executions have a distinct effect which compounds the deterrent effect of merely (re)instating the death penalty" (p. 319).

Berk (2005) challenged studies such as those by Dezhbakhsh and Shepherd, arguing that while numerous papers claimed to show that executions deter serious crime, many suffered from serious statistical problems. He highlighted the problem of "influence," which occurs when a very small and atypical fraction of the data dominates the statistical results. In this case, the number of executions by state and year is the primary explanatory variable, yet most states in most years execute no one. Very few states in any given year execute more than five prisoners, representing only about 1% of the available observations.

Shepherd (2005) used a large data set covering all U.S. counties from 1977 to 1996 to examine whether capital punishment's effect on murder rates differed across states. She found that "of the 27 states in which at least one execution occurred during the sample period, capital punishment deters murder in only 6 states" (p. 203). In 13 states, capital punishment actually increased murder, while in eight states it had no effect at all.

In only 22% of states did executions produce a deterrent effect. At the same time, executions had induced additional murders in 48% of states and created no deterrence in 78% of states. In other words, the findings were completely mixed.

Shepherd's (2005) paper also examined a threshold effect to better understand why a handful of states showed a deterrence effect while many did not. She stated that "on average, the states where capital punishment deters murder execute many more people than do the states where capital punishment does not deter murder" (p. 208). In states that carried out more executions than the threshold, each additional execution deterred murder on average. However, in states that performed fewer executions than the threshold, executions were associated with increased murder rates. This suggests that when a state executes many people, there may be greater public credibility that the state is serious about capital punishment, thereby producing a deterrent effect.

Given these inconsistencies, Keckler (2006) approached the issue from another angle. He noted that most researchers on both sides of the debate rely on simple assumptions about criminal behavior. Instead, he "examined whether it is plausible to suppose there is a marginal increase in deterrence created by increasing the penalty from life imprisonment without parole to capital punishment" (p. 101). He argued that a subset of offenders who are relatively unafraid of prison may be deterred effectively only by the death penalty. Empirically, criminals appear to fear a capital sentence and often attempt to avoid it through plea bargaining. Thus, at least for some offenders, the death penalty may encourage greater caution in the use of lethal violence, and any statistically observed deterrent effect may stem from behavioral changes among this group.

Zimmerman (2006) used state-level data spanning 1978 to 2000 to determine whether the method of execution influenced the per-capita incidence of murder. He analyzed various measures related to the perceived likelihood of execution and considered the specific timing of executions. He concluded that "the deterrent effect of capital punishment is driven primarily by executions conducted by electrocution. However, not one of the other four methods of execution — lethal injection, gas chamber asphyxiation, hanging, and/or firing squad — has been found to have a statistically significant impact on the per capita incidence of murder" (p. 920).