In this article, we discuss the the effect of capital punishment, or death penalty, on crime and murder using U.S. data.
Capital punishment, once a widely-used form of punishment, is now abolished in many countries.
According to the Amnesty International, 102 counties formally abolished the death penalty and 32 counties abolished the punishment in practice (i.e., they have not executed for 10 years or more). Only 58 countries retain it. But even in these countries, capital punishment is usually reserved for a small number of offenders who commit unusually atrocious crimes. In the United States, for example, the death penalty is used almost exclusively for homicide with an aggravating factor (for example, homicide committed in the course of other violent crimes, serial killer, murder of a child or a police officer).
While the debate on whether to abolish or reinstate the death penalty involves complex moral and ethical arguments, one of the common arguments for/against it concerns the deterrent effect of capital punishment on murder. On the one hand, the possibility of capital punishment should make potential murderers think twice before committing the crime and thus deter some murders. Critics, on the other hand, would argue that many murders are committed in the heat of passion and would not be deterred by the remote possibility of capital punishment.
Economics of crime, with its strong emphasis on causal analysis, seems particularly well-suited to study the deterrent effect of death penalty on murder (or lack thereof). Many economists have investigated this important policy issue, but there is no clear consensus yet. Some find that the death penalty significantly reduces murders. (Economists Hashem Dezhbakhsh, Paul H. Rubin, and Joanna M. Shepherd report that each execution leads to a reduction of eighteen murders.) Others argue that the estimated deterrent effect of the death penalty is much weaker when alternative regression specifications are used.
Going into the technical details of the debate is out of the scope of this article. But we can still take a glimpse of a few important challenges that economists face when studying the causal effect of the death penalty on crime.
Studies based on the U.S Data
Most of research evidence on the deterrent effect of capital punishment come from the U.S. This is partly due to the fact that each U.S. state has its own law regarding the use of the death penalty, and the large variation in the death penalty eligibility and execution patterns across different states provides researchers with a natural “laboratory” setting to study the effect of capital punishment. (Of course, it may have more to do with the fact that many countries have completely abolished the death penalty already. Imagine running an empirical analysis on capital punishment and homicide using data from the EU countries.)
Taking advantage of the variation in the use of capital punishment across different states, researchers investigate whether states that practice capital punishment have fewer homicides than states that do not (after accounting for important socioeconomic differences such as per capita income, unemployment, and expenditures on police protection). However, there are a few fundamental problems that make it difficult to interpret these regression results.
First, dividing the 50 U.S. states into two groups (abolitionists and non-abolitionists) is inadequate to describe the large variation in how capital punishment is actually used among non-abolitionist states. Consider the number of death sentences and the rate of actual executions between Texas, Virginia and California. Between 1976 and 2009, there were 927 death sentences and 13 executions in California (1.4%), 1040 death sentences and 447 executions in Texas (43.0%), and 150 death sentences and 105 executions in Virginia (70.0%). Although a few executions take place in California once in a while, it is unclear how seriously potential murderers in California would take this remote chance of an execution. (This problem cannot be solved by using an alternative measure in the empirical analysis, such as the number of death sentences, the number of executions, or the share of death sentences leading to executions. None of these measures gives us a satisfactory picture of the chance of an execution upon committing a homicide by itself.)
Relatedly, one would expect that a higher number of executions deter homicides because it would increase the cost of committing murder. But very few prisoners on the death row are actually executed. Out of 40 states that ever used a death sentence since 1973, the share of actual executions over the number of death sentences is less than 10 percent in 25 state, and less than 5 percent for 15 states. In these states, how much do you think potential murders will be deterred by the possibility of capital punishment? Even for a small number of prisoners who are actually executed, it usually takes a long time between the time of sentence and the time of an actual execution. (The average delay, as of 2010, was 10 years.) For many murderers on death row, there is little difference between the death penalty and a life sentence.
More importantly, comparing homicide rates between abolitionist and non-abolitionist states can at best pick up additional deterrent effects of capital punishment relative to long prison sentences.
High risk murderers receive long prison sentences in both abolitionist and non-abolitionist states: The only difference is that a murderer in non-abolitionist states may face an execution, on top of a long prison sentence. There is an important problem, however. The severity of non-capital-punishment for murderers may be systematically different between abolitionist and non-abolitionist states. (For example, non-abolitionist states may have tougher sentencing systems against all criminals in general.) In this case, how can we find out whether potential murderers in non-abolitionist states are deterred by the death penalty or tougher non-capital-punishment?
Lastly, think carefully about why
more executions should deter murder. Suppose you live in Texas. Now, if you find out that there were unusually many executions last month, would you take this as a sign that the probability of executions for murders is now higher in Texas?
Some of you may see this as a sign that the probability will be lower
in a short run because the number of murderers on the death row decreased and there will be fewer executions in the near future. Others may expect that there would be no substantial change in the probability of executions for murderers. We all know that Texas has been and will continue to actively use the death penalty for high-risk murderers. A small increase in the number of executions last month or last year will not be enough for us to adjust our expectations.
In sum, at least three issues make it difficult to assess the causal effect of capital punishment on murder using the U.S. data: (Note: The three limitations below are from the 2012 National Research Council report
. Interested readers should see the report, which offers much more in-depth discussions about the deterrent effect of the death penalty.)
1) the large differences in how capital punishment has been used among non-abolitionist states.2) a potential difference in the non-capital component of the punishment between abolitionist and non-abolitionist states.3) insufficient understanding of how potential murderers form their expectations on the probability of an execution upon committing a murder.
Can you think of other empirical settings (from different time or different countries) where such concerns would be less relevant?
- Dezhbakhsh, Hashem, Paul H. Rubin, and Joanna M. Shepherd. “Does Capital Punishment Have a Deterrent Effect? New Evidence from Post-moratorium Panel Data.” American Law and Economics Review 5.2 (2003): 344-376.
- Donohue, John, and Justin Wolfers. “Estimating the Impact of the Death Penalty on Murder.” American Law and Economics Review 11.2 (2009): 249-309.
- Donohue, John, and Justin J. Wolfers. “The Death Penalty: No Evidence for Deterrence.” The Economists’ Voice 3.5 (2006).
- Mocan, N., and Gittings, K. (2010). The Impact of Incentives on Human Behavior: Can We Make It Disappear? The Case of the Death Penalty. In R.E.S. Di Tella and E. Schargrodsky (Eds.), The Economics of Crime: Lessons for and from Latin America(pp. 379-420). Chicago: University of Chicago Press.
- Nagin, Daniel S., and John V. Pepper, eds. Deterrence and the Death Penalty. National Academies Press, 2012.
© Songman Kang, Hanyang University