Skip to main content


Overview: Death Penalty

The death penalty remains one of the most controversial aspects of the American criminal justice system. America stands alone among wealthy democratic countries in its imposition of the death penalty, and per capita stands above all other countries in the frequency with which death is imposed. Courts, policy-makers, politicians, and citizens face complex questions: should the death penalty be imposed at all, constitutionally, morally, and economically? If so, for which crimes? Upon which criminals? Legal answers to these questions are further circumscribed by the 8th Amendment’s ban on cruel and unusual punishment, generally interpreted to relate to the values of retribution and deterrence, respectively.


One of the most straightforward questions social science can address about the death penalty is simply this: does the death penalty deter? Without a deterrent effect, justifications for the death penalty must rely solely on retributive and not utilitarian grounds. Arguments for and against deterrence typically fall under the cruelty prong of the 8th amendment’s ban. As the most extreme form of punishment allowed by the Constitution, the death penalty carries substantial costs: constitutional appellate rights ensure that imposing the death penalty is frequently more costly to governments than imposing life imprisonment; the government must ensure the punishment is being imposed fairly and accurately; and the United States’ insistence on the legality of the death penalty can strain international legal norms and international relations. The effect of an execution, beyond the permanent incapacitation of the convict in question, is vital to defend the practice on both moral and economic grounds.

Assuming that an execution is worse for the convict than life in prison, by imposing the death penalty the state raises the cost of committing death-penalty eligible crimes—principally homicide—in the United States. Yet despite a large amount of data, the answer to the deterrence question remains frustratingly inconclusive.  One of the persistent problems in proving that executions cause less crime is the small number of executions compared to the large fluctuations in crime rates from year to year. Numerous other factors seem to influence homicide rates. For example, the chart below compares homicide rates between the U.S. and Canada. Despite drastically different approaches to death penalty policy, homicide rates have largely mirrored each other, implying the relative insignificance of the death penalty among the many factors that contribute to homicide rates.

Source: John J. Donohue III, Justin Wolfers, Uses and Abuses of Empirical Evidence in the Death Penalty Debate, Discussion Paper No. 1949 (January 2006) available at

The earliest influential study on the value of the death penalty as a deterrent was Isaac Elhrich’s national time-series analysis published in 1976 in the American Economic Review, comparing executions and homicides from 1933 to 1969. Elhrich estimated that each execution saved eight lives—a study lauded and sharply criticized when the Supreme Court reinstated the death penalty in 1976. Two years later, a reanalysis by the Panel on Research on Deterrent and Incapacitative Effects at the National Academy of Sciences concluded that Elhrich’s time-series analysis could not justify the conclusion that the death penalty deters crime. The years 1933–1961 showed no relationship between executions and homicides; only the years 1962–1969 showed a negative correlation. But this was a time when all crime rates rose precipitously, potentially implicating a confounding variable unconnected to the frequency of executions. Time-series analysis can make isolating the variable difficult, if not impossible. Certainly no clear trend emerges, even from later time-series analyses: sometimes executions and homicide rates move in opposition to one another, implying a deterrence effect; other times they move in tandem.

Cross-state analysis also leads to unclear conclusions. The graph below shows homicide rates for each state, death penalty and non-death penalty in 1998. Comparing, say Michigan to South Dakota supports the idea that the death penalty deters while comparing Iowa to Louisiana supports the exact opposite conclusion: the death penalty raises homicide rates. Looking at the data as a whole, no clear picture emerges.


A recent analysis by Hashem Dezbakhsh et al. in the American Law and Economic Review (2003) using more sophisticated forms of regression analysis reignited the debate over deterrence and morality. Dezbakhsh and his colleagues estimated that every execution prevented eighteen murders—a number substantially higher than Elrich’s. Following their analysis, Cass Sunstein and Adrian Vermeule released a working paper in 2005 arguing that if Dezbakhsh’s analysis was accurate, the state has a moral duty to impose the death penalty (as opposed to merely no moral duty not to impose the death penalty). Both Dezbakhsh’s analysis and conclusions drawn from it by Sunstein, Vermeule, and others have been praised, criticized, and debated. Donohue and Wolfers in a 2006 paper argued that the data was simply too vulnerable to small model changes to be of use one way or the other.

Another factor complicating calculations of deterrence is the execution process. Most death row inmates will never be executed. The appellate process, political dynamics, execution innovations, constitutional challenges, and resource limits make predicting when an inmate will be executed after being sentenced to death difficult. Several states give death sentences but rarely execute inmates; others have indefinite or de facto moratoria on the death penalty. Does the mere fact that a state permits capital punishment deter potential criminals? Or must execution commence? How would a potential criminal know the likelihood of getting the death penalty for a potential crime beyond state reputation? Most studies find that extreme punishments can have a profound effect on compliance, but the lengthier and more unpredictable the interval between crime and punishment the less power extreme punishments have.

This brings us to the issue of criminal decision-making. Justice Stewart observed in Gregg v. Georgia that the death penalty might deter some criminals and not others: murderers for hire may be deterred while those who act in passion may not be. Psychologists have since classified criminals into two categories: reactive (a criminal who commits out of passion) and instrumental (a criminal who plans the crime). It’s easy to imagine the planned murderer as the rational actor, but neurological explanations for the behavior of each are equally compelling, complicating our traditional ideas of responsibility. Decision-making researchers have also frequently explored the limits and nuances of human rationality—cognitive biases may prevent potential criminals from fully calculating the possibility of a death sentence into their criminal decisions. If a potential criminal does not calculate the cost of the risk of the death penalty then the mechanism of deterrence may be a fiction.

The deterrent effect of the death penalty remains, and perhaps will always remain, unclear. The death penalty may play a small role, if it plays any role at all, in deterring homicide, but other factors seem to be at play. The absence of clear evidence for deterrence means that the legal burden takes on increasing importance in the legal outcomes. If, as some justices have suggested, the burden is on advocates of capital punishment to prove a deterrent effect, then capital punishment cannot be justified until such an effect is proved. If, on the other hand, the burden is on opponents of capital punishment to prove no deterrent effect, then capital punishment can be justified until such proof exists.


Racial discrimination continues to plague the American justice system, and the death penalty context is no exception. Arguments for and against the existence of discrimination typically fall under the unusual prong of the 8th Amendment’s ban. Like research on deterrence, the existence and extent of racial discriminatory death penalty outcomes remains a point of debate. Initial research by Wolfgang and Riedel in 1973 explored the disparity in death sentencing between black and white defendants in rape cases in southern states, finding that blacks were sentenced at nearly six times the rate of white defendants, and more recent research has found that the single largest predictor of a death sentence is the race of the victim, not the race of the defendant. Prosecutors are more likely to pursue the death penalty when the victim is white and juries are more likely to sentence defendants who have killed white victims to death—even though white and black victims are killed in roughly equal numbers, almost 80% of defendants sentenced to death have killed white victims.

David Baldus and his colleagues performed the earliest research on victim race with data from Georgia in Criminal Law & Criminology in 1983, finding that black defendants who were convicted of killing white defendants were sentenced to death 22% of the time; those convicted of killing black defendants were sentenced to death 1% of the time. Further research has substantiated this conclusion for other states as well: black defendants who kill white victims are consistently sentenced to death more frequently than other categories of defendants and victims. Researchers have also found that in black-defendant/white-victim cases, blacks who have stereotypically black features are sentenced to death more frequently than those that do not. Defendants designated as stereotypically black (those in the top half of the distribution of stereotypicality) were more than twice as likely to face a death sentence than non-stereotypically black defendants.

Despite the early Baldus study, in McCleskey v. Kemp the Supreme Court rejected that Georgia’s imposition of the death penalty on McCleskey violated either the Equal Protection Clause of the Fourteenth Amendment or the Cruel and Unusual Clause of the Eighth. The Court noted that the statistical analysis could not show that the state discriminated against McKleskey individually simply because he was black and killed a white victim, it mentioned that cosmetic disparities in sentencing were inevitable, and it ended by deferring to the legislature to weigh and evaluate such statistical studies. Though the same patterns have been found in states across the country, no new legal challenge has been mounted in the Supreme Court.

Several state actors play roles in the capital punishment process, and the racial makeup of these actors may play a role in explaining the racial disparities in sentencing.  The state legislature and to a lesser extent the state courts determine whether the state can seek the death penalty, but most states give prosecutors wide discretion to pursue the death penalty as they see fit. A national survey of every death penalty jurisdiction by Jeffrey Polanak in 1998 found that 97.5% of all prosecutors were white, 2.1% were black and 1.2% were Hispanic. Polanak suggests that this could explain both defendant-race and, more importantly, victim-race effects. A number of factors also affect the diversity of jury pools. The structure of jury lists has traditionally diminished minority representation in jury pools: using voter registration lists under-represents minorities, and undeliverable and nonreturnable summons contribute to this effect. Death qualification, where jurors who would never sentence a defendant to death are disqualified from serving on the sentencing jury, further diminishes minority representation.

Commuting a death sentence can offer a final, and largely discretionary, decision in the death penalty process. Surveys of department of justice commutation data yield quite different trends from sentencing data. Of the small number of death row inmates who ask for their death sentence to be commuted, governors commute nonwhites at about twice the rate of whites, women significantly more often than men, and those incarcerated in northern states at about three times the rate of those incarcerated in southern states. Unfortunately, it is very difficult to know the cause of these observations. Commutation may be a corrective measure that implicitly seeks to remedy sentencing biases. The racial difference may also simply be a reflection of the victim-race bias we noted earlier—thus far no study has compared commutation rates with the race of the victim, which would allow researchers to isolate this potential confound.

In 2006, Stephen P. Klein and his colleagues at the Rand Corporation released a study contesting findings of racial discrimination using data from the Department of Justice from 1995 to 2000 and statistical regression methods testing a variety of explanatory variables. Klein found that prosecutors were more likely to seek the death penalty when white defendants killed white victims than when nonwhite defendants killed nonwhite victims, but that this discrepancy disappeared when controlling for the heinousness of the crime, though other scholars debate these conclusions.

After the Supreme Court banned the death penalty as applied in 1972 in Furman v. Georgia in part because of racially discriminatory results, states reorganized their sentencing processes. Though significant evidence of racially discriminatory results persists, its unclear what changes to the existing capital punishment regime would have to change to resolve these troubling findings.


The Supreme Court has issued two opinions in recent years significantly limiting the application of the death penalty: Atkins v. Virginia established that executing mentally retarded individuals violates the 8th Amendment; Roper v. Simmons found that executing individuals for crimes they committed before the age of eighteen violates the 8th amendment.

The Court in Atkins relied upon two compelling sources of evidence to disqualify the executions of mentally retarded individuals under the 8th Amendment. First, the majority found a growing consensus among the states. Among death penalty states, 18 had recently exempted mentally retarded individuals from being executed, in addition to the federal exemption. Second, the Court found that the death penalty failed to serve penological purposes for mentally retarded individuals. Without the same capacity to understand ones actions and the actions of others, the death penalty would serve a lesser deterrent effect. Jurors may also misinterpret the behavior of such individuals as remorseless and mentally retarded individuals are typically more susceptible to suggestion making it more likely that a jury would impose a death sentence despite potential mitigating evidence

Like in Atkins, the majority in Roper relied heavily on a host of neuroscientific studies finding distinctions between the brains of adolescents and adults. The well-substantiated developmental gap, combined with evidence of evolving standards of decency, meant that individuals under the age of eighteen could not be sentenced to death, the Court concluded. It justified the age eighteen as one generally used to demarcate the line between childhood and adulthood in our society.

Both cases were also battlegrounds over how changes in national opinion should be treated by the courts. Besides rejecting that a national consensus had been reached on either issue, dissents in both cases also rejected the validity of an argument from national consensus, regardless of how careful or compelling the polling may be.


Perhaps the most disturbing element of capital punishment is the prospect of executing individuals who are innocent of the crime they have been convicted of. Estimates of the number of innocent people on death row are notoriously difficult to pin down. While the Supreme Court has stopped short of declaring claims of actual innocence constitutional, since 1973, one hundred and thirty-nine inmates have been exonerated from a death sentence because of evidence of their innocence. Proponents of death penalty abolition also name several inmates who have been executed even with compelling evidence of their innocence.

For Further Study


Isaac Ehrlich, The Deterrent Effect of Capital Punishment: A Matter of Life and Death, 65 Amer. Econ. Rev. 97 (1975).

John Monahan & Laurens Walker, Social Science in Law, Sixth Edition, Foundation Press (2006).

Gregg v. Georgia, 428 U.S. 153 (1976).

Eisenberg, Death Sentence Rates and County Demographics: an empirical study, 90 Cornell L. Rev. 347 (2005).

Brendan O'flaherty, Why Repeated Criminal Opportunities Matter: A Dynamic Stochastic Analysis of Criminal Decision-Making, 14 Journal of Law, Econ., & Org. 232 (1998).

John J. Donohue III & Justin Wolfers, Uses and Abuses of Empirical Evidence in the Death Penalty Debate, IZA Discussion Paper No. 1949 (January 2006) available at

Cass R. Sunstein & Adrian Vermeule, Is Capital Punishment Morally Required? The Relevance of Life-Life Tradeoffs, Public Law and Legal Theory Research Paper Series No. 85, John M. Olin Law & Economics Working Paper Series No. 239, AEI – Brookings Joint Center For Regulatory Studies Research Paper No. 05-06 (2005), available at

Hashem Dezbakhsh et al., Does Capital Punishment Have a Deterrent Effect? New Evidence from Postmoratorium Panel Data, 5 AM. L. & ECON. REV 344 (2003).

Reid Griffith Fontaine, Disentangling the Psychology and Law of Instrumental and Reactive Subtypes of Aggression, 13 Psychology, Public Policy, and Law 143 (2007).


Marvin E. Wolfgang & Marc Riedel, Race, Judicial Discretion, and the Death Penalty, 407 Annals of the American Academy of Political and Social Science 119 (May 1973).

David C. Baldus, Charles Pulaski, & George Woodworth, Comparative Review of Death Sentences: An Empirical Study of the Georgia Experience 74 J. Crim. L. & Criminology 661 (1983).

David Baldus, Charles Pulaski, & George Woodworth, Equal Justice and the Death Penalty (Boston: Northeastern University Press 1990).

McCleskey v. Kemp, 481 U.S. 279 (1987).

Jennifer Eberhardt et al., Looking Deathworthy: Perceived Stereotypicality of Black Defendants Predicts Capital-Sentencing Outcomes, 17 Psychological Science 383 (2006).

John Kraemer, An Empirical Examination of the Factors Associated with the Commutation of State Death Row Prisoners’ sentences between 1986 and 2005, 45 Am. Crim. L. Rev. 1389 (2008).

Maxine Goodman, A Death Penalty Wake-up Call: Reducing the Risk of Racial Discrimination in Capital Punishment, 12 Berkeley J. Crim. L. 29 (2007).

Stephen P. Klein, Richard A. Berk, & Laura J. Hickman, Race and the Decision to Seek the Death Penalty in Federal Cases, Rand Corporation (2006), available at

United States of America: Death by discrimination—the continuing role of race in capital cases, Amnesty International (2003) at


John Monahan & Laurens Walker, Social Science in Law, Sixth Edition, Foundation Press (2006).

Atkins v. Virginia, 536 U.S. 304 (2002).

Roper v. Simmons, 543 U.S. 551 (2005).


David Gram, Trial by Fire: Did Texas execute an innocent man? New Yorker, September 7, 2009, available at (about the Cameron Todd Willingham investigation, and Willingham’s potential innocence).

The Innocence Project.

Errol Morris, The Thin Blue Line (1988) (the film responsible for Randall Adam’s exoneration after being convicted of murder and sentenced to death in Texas).

Other Articles and Links of Interest

Furman v. Georgia, 408 U.S. 238 (1972).

James S. Leibmen, The Overproduction of Death, 100 Colum. L. Rev. 2030 (2000) (examining the imbalance between trial-level death sentences production and appellate-level death sentence correction).

Vanderbilt Law School Death Penalty Debate, featuring Joshua Marquis, District Attorney of Clatsop County, Oregon, arguing for the death penalty, and Dr. Ken Haas of University of Delaware advocating against the death penalty.

Related Articles: Death Qualification, Future Dangerousness