The Death Penalty: Does It Deter Killings?

IPR/Law School event considers national report, future research

From left: David Figlio, Daniel Nagin, Charles Manski, and
Max Schanzenbach discuss the National Research Council
committee's findings and ways for researchers to conduct
better studies of the death penalty and its effects.

For decades, scholars have attempted to answer a seemingly simple question: Does the death penalty deter murder? Two National Research Council (NRC) reports, conducted more than three decades apart, have reached basically the same conclusion—we still do not know. 

At a January 9 event, co-sponsored by IPR and Northwestern School of Law’s Searle Center, Carnegie Mellon criminologist Daniel Nagin, chair of the 2012 NRC Committee on Deterrence and the Death Penalty discussed its evaluation of death penalty research.

“Research to date on the effect of capital punishment on homicide is not informative about whether capital punishment decreases, increases, or has no effect on homicide rates,” Nagin said, quoting the 2012 report.

The event also featured commentary by IPR economist Charles F. Manski, who was also a member of the recent NRC committee, and law professor Max Schanzenbach, an IPR associate and director of the Searle Center on Law, Regulation, and Economic Growth.

IPR Director David Figlio and Law School Dean Daniel Rodriguez welcomed participants to the school’s historic Lincoln Hall. Rodriguez also pointed to its particular significance for the event.

“It was in this room that the governor effectively ended Illinois’ death penalty,” said Rodriguez, noting then-Governor George Ryan’s news-making 2003 declaration to commute sentences for 156 inmates on death row.

Examining the Death Penalty

The 2012 NRC committee consisted of criminologists, economists, psychologists, the esteemed and recently deceased political scientist James Q. Wilson, and even a sitting federal appeals court judge. Its mandate was only to review and assess the evidence since the first NRC report in 1978—not to make policy determinations or pass judgment on the death penalty, Nagin began.

Beginning in the 1930s, the number of executions in the United States began a steady decline from a high of almost 200 in 1935 to zero between 1968 and 1976. This included a four-year national moratorium due to the Supreme Court’s 1972 Furman v. Georgia decision. Following the Supreme Court’s 1976 Gregg v. Georgia decision reinstating the death penalty, executions began to rise again to a peak of nearly 100 in the late 1990s. Since that peak, the number of executions has declined to around 50 per year.

The committee, Nagin recounted, was charged with reviewing the evidence in the post-Gregg era and attempting to determine whether this “accumulated evidence” provided any useful information on the magnitude of capital punishment’s deterrent effect. It eventually reached the potentially controversial conclusion that research on the death penalty’s deterrent effect is so flawed that it cannot be used to determine if the death penalty does indeed affect homicide rates.

He singled out the key question from the report: Is capital punishment more or less effective as a deterrent than alternative punishments, such as a life sentence without the possibility of parole?

“None of the studies we reviewed—none—accounted for the noncapital portion of the sanction regime,” Nagin said.

In probing the studies, the NRC committee also found that they did not provide any plausible evidence on how potential murderers perceive, and respond to, capital punishment. Many studies fail to address how perceptions are formed, how they correspond with reality, or how they vary across states or over time. They simply infer that potential murderers respond to the objective risk of execution, Nagin said.

“The calculation of risk is complex,” Nagin continued, citing how difficult it was for even well-informed researchers to grasp it, let alone potential murderers.

In addition to the example of understanding perceptions, he also underlined that only 15 percent of people who have been sentenced to death since 1976 have been executed. Why? He listed many reasons, including that about half of all death sentences are reversed, others are commuted, and many die while on death row, among others.

Despite the bleak picture of post-Gregg research, Nagin said the report offered researchers a difficult but viable way forward. In addition to more research on perceptions, it also called for better data collection on noncapital sanction regimes and better specification of statistical models that incorporate uncertainty.

Explaining the Unexplainable

An expert on criminal sentencing, Schanzenbach generally praised the NRC report and used his comments to provide a critique of existing deterrence literature from a legal institutional perspective.

By showing that murder rates in the United States vary substantially over time, Schanzenbach surmised that there is a very large “unexplained component” determining murder rates. He traced the volatility of U.S. homicide rates from the early 20th century through the early 2000s, noting the few common threads behind the rise and fall of rates.

“Everything went to hell in the 1960s, and I don’t know that it has ever been really credibly explained,” Schanzenbach said.

He pointed to the historically high rates from the 1960s to the 1980s that were also seen in the 1920s and 1930s during Prohibition, but also persisted for a while thereafter. The murder rate rose again in the mid-80s during the nation’s “crack epidemic,” and began its remarkable decline in the mid-1990s, even dropping slightly during the turmoil of the Great Recession.

Schanzenbach cited a number of reasons researchers have given as to why murder rates might have dropped since the 1990s: more police on the streets, more effective policing, more incarceration, a drop in lead exposure, etc. But none provide a fully satisfactory explanation for the declines, and there is at present no consensus.

He also highlighted some vexing holes in the existing death penalty literature caused by inadequate focus on the importance of institutions, some of which were noted in the 2012 NRC report. Following Gregg, for example, the Court required that the death penalty could only be imposed if warranted by aggravating factors. If the death penalty is to have a deterrent effect, it should be concentrated in cases now likely to receive the death penalty, such as those involving multiple homicides or the murder of police officers, government officials, or children, but few studies try to account for this.

In addition, he pointed out that prosecutors are the key actors in the decision to pursue the death penalty.

“[Prosecutors] like to put it on the table and then pull it off again,” Schanzenbach said.

Therefore, a deterrent or incapacitation effect could conceivably operate through the use of the death penalty to induce guilty pleas. He suggested that future work focus on prosecutorial behavior and lamented the lack of good qualitative evidence about how and when prosecutors use the death penalty.

Death Penalty Research: Then and Now

One of the big differences between the 1978 and 2012 reports were in the recommendations each laid out for future research.

The 1978 report gloomily concluded that “research on this topic is not likely to produce findings that will or should have much influence on policymakers.” It acted like a spigot, cutting off death penalty research for the next 15 years.

Then in the 1990s, a new generation of researchers took up the question. But this new generation of researchers committed many of the same mistakes as the previous one.

“It’s like people in the 1990s didn’t even read the 1978 report,” Manski said.

Yet if the 2012 committee concluded that “studies not be used to inform deliberations requiring judgments about the effect of the death penalty on homicide,” their report offers a more hopeful path forward for researchers.

“Its most important contribution is calling for more research—not just on the deterrence of the death penalty, but on deterrence in general,” Manski said.

Future Research: Embracing Uncertainty

In 1975, economist Isaac Ehrlich published a now “infamous” study in the American Economic Review that concluded that eight lives are spared for every execution that takes place.

Though this finding has been widely rebuked since then, the work was up to the journal’s standards for its time, and thus the field, Manski said. So, he quizzed, what standards of proof should we now apply?

Manski turned to his own work for help, drawing on his latest book, Public Policy in an Uncertain World: Analysis and Decisions (Harvard University Press, 2013). When it comes to public policy, what we think we know can hurt us. Policymakers and analysts almost always opt to use models that draw stark black-and-white conclusions, Manski said. But such certainty typically rests on flawed models and may lead to erroneous conclusions. Instead, he suggested reporting a range of estimates for the impact of a particular policy change, derived from a corresponding range of plausible models. This would generate more honesty in policy analysis, he said.

In terms of capital punishment, this means that researchers need data on sentences that do not involve the death penalty and how potential murderers perceive such penalties, in addition to better methods for dealing with inherent uncertainty, he added.

Finally, both Manski and Nagin cautioned against dismissing value judgments in the death penalty debate. The 2012 report sought to avoid the imbroglio of ethical and cost debates. Though it concludes that studies to date are not rigorous enough to prove one way or the other the effects of capital punishment on homicide rates, “the report does point out that judgments about whether capital punishment deters or not are still relevant to policy deliberations,” Nagin stressed. “It’s just that people should just not appeal to this evidence in support of their opinions about whether capital punishment has a deterrent effect or not.”

“The great value of this report is that it clears the air about what we know—and don’t know—about the death penalty,” Manski said.

Charles F. Manski is Board of Trustees Professor in Economics and an IPR fellow. Max Schanzenbach is Director of the Searle Center on Law, Regulation, and Economic Growth, Professor of Law, and an IPR Associate. David Figlio is Director of IPR and Orrington Lunt Professor of Education and Social Policy and of Economics. Daniel Rodriguez is Harold Washington Professor and Dean of the Northwestern School of Law.

National Research Council. Deterrence and the Death Penalty. Washington, D.C.: The National Academies Press, 2012.

View slides from the lecture here.