Contact FutureLearn for Support
Skip main navigation
We use cookies to give you a better experience, if that’s ok you can close this message and carry on browsing. For more info read our cookies policy.
We use cookies to give you a better experience. Carry on browsing if you're happy with this, or read our cookies policy for more information.

Factors that explain and do not explain a crime drop

Levitt argues that there were four factors mainly responsible for the crime drop in the 1990s: an increase in the number of police, a rising prison population, receding crack epidemic, and legalization of abortion.

His claim that legalization of abortion during the 1970s caused crime to fall during the 1990s remains controversial (more on this in Week 5), but the first three are widely accepted by researchers as key reasons of the crime drop.

Based on existing research evidence, Levitt argues that the observed increase in police force size was associated with a 5.5 percent drop in violent crime in the U.S. between 1991 and 2001, the increase in the prison population with a 12 percent drop, receding crack epidemics with a 3 percent drop, and legalization of abortion with a 10 percent drop.

Adding up these numbers gives us a predicted crime drop of 30.5 percent in violent crime, which is close to the observed drop of 34 percent in violent crime rates.

Of the four factors, the first two, namely, increased police force size and increased prison population, are directly related to the rational choice model we saw last week. As the number of police increases, potential offenders would figure that the probability of arrest and punishment is now higher, and respond to this change by committing fewer crimes. Similarly, when prison population rises because of tougher sentencing laws, would-be offenders would realize that the severity of punishment is now higher, and thus commit fewer crimes.

Deterrent Effect and Incapacitation Effect

The rational choice model of crime predicts that increased police and prison population will deter crime by increasing the certainty and severity of punishment (“deterrent effect”). However, increased police presence and larger prison population can also lower crime by incapacitating criminals and preventing them from committing new crimes while incarcerated (“incapacitation effect”).

Which effect do you think would be greater? Measuring the magnitudes of deterrent and incapacitation effects is of great interest for at least two reasons. First, the deterrent effect is directly related to the empirical relevance of the rational choice model of crime. Recall that the model describes a rational individual’s decision to commit crime or not, and predicts that increased certainty and severity of punishment would deter potential criminals. (On the other hand, we do not need a theoretical model to see why criminals already incarcerated cannot commit new crimes while in prison.) Thus, if empirical findings suggest that the extent of deterrent effects is large, we can take this as evidence that the rational choice model can be useful in explaining and predicting individuals’ criminal decisions.

Secondly, while both deterrent and incapacitation effects should lower crime, lowering crime through incapacitation would be a lot more costly than through deterrence. Suppose the incapacitation effect is very strong, while the deterrent effect is minimal. Then, if we want to reduce a lot of crimes, we would have to catch a lot of criminals and punish them with long prison sentences (very costly). On the other hand, if the deterrent effect is large, the presence of large police forces and tough sentencing laws should deter many would-be criminals from actually committing the crime. Since only few would actually commit crime, the cost of punishment would be low(not as costly). The magnitudes of the two effects are thus closely related to the aggregate cost of crime-control and the socially optimal level of crime.

In many empirical studies, however, the distinction between deterrent and incapacitation effects can become unclear. For example, when estimating the effect of additional police officers on local crime rate, the estimated effect will necessarily include both deterrent effects (increased police presence deters would-be criminals) and incapacitation effects (many criminals would be caught by police and sent to prison). Nevertheless, we will encounter several empirical studies that exclusively focus on either deterrent or incapacitation effect.

Effects of Police and Prison on Crime

How large are the effects of police and prison on crime?
Existing research evidence suggests that, when police size increases by 10 percent, crime can fall somewhere between 3 and 15 percent. The number of U.S. police officers per capita increased by 14 percent during the 1990s. Based on the research evidence, the 14 percent increase in police would have resulted in a 4-to-20 percent crime reduction.

Levitt then presents a simple cost-benefit analysis. Estimates on the cost of police hiring suggest that it cost approximately $8.4 billion to increase the police force size by 14 percent. Estimates on the cost of crime victimization suggest that a 4 percent crime reduction is associated with the social gain of $17 billion. Even with a conservative estimate of the police effect on crime (a 4 percent crime reduction), the increase in police force size likely have generated a lot more social gains than its cost.

It is straightforward to extend this cost-benefit analysis on the effect of prison on crime. The adult incarceration rate in the U.S. (the number of individuals held in state and federal prisons and local jails divided by the number of adult population) increased by about 40% between 1990 and 2000.

Again, available research evidence suggests that, when prison population increases by 10 percent, crime falls by somewhere between 1 and 4 percent. Then, a 40 percent increase in prison population should have reduced crime by 4-to-16 percent, which roughly corresponds to social gains between $17 billion and $47 billion. On the other hand, the annual operating cost per inmate in 2001 was $22,650, and imprisoning 800,000 more criminals should have cost approximately $18 billion. Unless we take a very conservative view of the crime-reducing effect of prison, increasing prison population also seems to pass the cost-benefit test. However, it seems that fighting crime through increasing the police force size probably had a higher benefit-cost ratio than the prison population increase.

Levitt also describes five factors that many believe have contributed to the crime drop but seemed to have little impact:

1) strong economy of the 1990s
2) changing demographics</br>
3) improved police strategies</br>
4) gun control laws</br>
5) increased use of capital punishment</br>

Although his assessment may be true, many economists spent considerable efforts on understanding the effects on crime of strong economy and labor market, changing demographics, different police strategies, gun control laws, and capital punishment. We will examine some of these research findings next.


  • Levitt, Steven D. “Understanding Why Crime Fell in the 1990s: Four Factors that Explain the Decline and Six that Do Not.” Journal of Economic Perspectives 18.1 (2004): 163-190.

Share this article:

This article is from the free online course:

Economics of Crime

Hanyang University

Course highlights Get a taste of this course before you join:

  • Welcome!

    Welcome to the course and introduce yourself

  • Welcome to Week 3
    Welcome to Week 3

    an increased probability of punishment seems to deter potential criminals?