Murderous Impact of Three Strikes

Since the early 1990s, the states have passed many get-tough-on-crime laws, boosted spending to double prison and juvenile detention space, and improved policing. Contrary to many government programs, these measures mostly worked and deserve much of the credit for our one-third drop in crime over the last decade.

But not everything has worked. Between 1993 and 1995, 24 states passed "three strikes and you're out" laws that prolong prison terms for comparatively few criminals with a particular pattern of prior felony convictions. All but one of the 24 states already had provisions that allowed the courts to "throw away the key" on habitual criminals, so nothing was radically new about the idea.

Here's the logical problem with the new statutes: While these laws may encourage some career criminals to cease and desist, those who continue to commit crimes face little or no additional sanction for committing a more grievous crime.

That behavior may cost the lives of more victims and crime scene witnesses. Under some circumstances, the criminal can reduce his chances of apprehension and conviction by killing a witness, without any practical disincentive in terms of a more severe punishment. If apprehended and convicted of a third serious violent crime, for instance, the penalty would amount to that for homicide anyway (an average of 89 months served for those released), whether the criminal takes a life or not. So, too often the criminal will kill in order to reduce the chances of being caught.

The laws, in other words, violate the principle that marginal punishment must be held in reserve to stop marginal harm.

"No Justification"

Since 1991 the national murder rate has declined 40 percent. But it has not fallen as much in three-strikes states, a difference of roughly 10 percentage points between three-strikes and other states. On first examination of the data, that amounts to at least 1,400 extra murders per year in the three-strikes states.

Because homicides are quite rare events, even if eliminating witnesses occurs in only a tiny portion of violent crimes, homicides can still increase significantly on a percentage basis.

Thomas Marvell and Carlisle Moody, two veteran economists at the College of William and Mary in Virginia, have carefully explored this connection between homicide and three-strikes laws in their January 2001 article in the University of Chicago's Journal of Legal Studies. They conclude, "Given their unintended consequences in terms of human lives, we see no justification for three-strikes laws."

The authors base this strong conclusion on their analysis of the data from all 50 states during the years 1970-1998. The basic finding is that three-strikes laws produce a 10-12 percent increase in homicides within the first two years after passage of the laws, while over the longer haul the increase rises to as much as 29 percent more homicides. That makes the long-run impact 3,300 more homicides each year. With some economists valuing a human life at $3.2 million, that would be a social cost of nearly $11 billion each year.

But isn't this staggering loss offset by favorable effects of the law in reducing crime? Apparently not. Marvell and Moody search diligently using impressive statistical procedures, but find relatively little evidence that the laws deter much crime. A few states, like California and Georgia, with the most severe laws seem to suffer the smallest increases in homicide, suggesting the increased murders to evade capture in these states may be offset by some deterrence and incapacitation effects. Also, their data end in 1998 and "it is possible that there is an incapacitation impact delayed many years." This seems unlikely, though, because the laws have yet to register a detectable impact on the prison population, suggesting no major incapacitation effect will be found.

Can a single study settle the issue? Of course not. For example, the study does not account for state differences in arrest rates, and the results might change if they were incorporated. We'll have to see if other researchers confirm the Marvell-Moody conclusion using different models and data. In the meantime, it looks like three-strikes laws have boomeranged badly.

Morgan Reynolds is chief economist for the U.S. Department of Labor and on leave as director of the Criminal Justice Center at the National Center for Public Policy.

For more information ...

Three Strikes, You're Out: A Review. Washington State's three strikes policy was approved by 76 percent of the voters in 1993. Between 1993 and 1997, violent crime fell, criminals seemed to be changing their behavior, and the program's costs were less than opponents predicted. (Washington Institute, 1997)

request PolicyBot documents #1377403 (part 1, 14pp.) and #1377404 (part 2, 14pp.)