The Dismal Science

Going Down Swinging

What if three-strikes laws make criminals less likely to repeat offend—but more violent when they do?

The slammer

On Oct. 1, 1993, a man named Richard Allen Davis kidnapped 12-year-old Polly Klaas during a slumber party at her home in Petaluma, Calif. At the time, Davis was on parole after serving half of a 16-year sentence for a prior kidnapping and had accumulated a 25-year rap sheet with charges ranging from burglary to auto theft to public intoxication. Polly was found raped and murdered a couple of months later, and the public outcry that ensued led to the passage of a California law that mandated stiff prison sentences for convicted felons on their third offense. Davis had more than a dozen convictions when he abducted Polly Klaas.

“Three-strikes” laws have now been enacted in 26 states, often with the stated purpose of keeping society safe from violent criminals like Richard Davis. But a new study released by the National Bureau of Economic Research finds that three-strikes laws like California’s, while discouraging criminals from doing things like smoking pot or shoplifting, may push those who do continue in a life of crime to commit more violent offenses. The study’s author, Radha Iyengar, argues that this is because under such laws, felons with a pair of strikes against them have little to lose (and often much to gain) by committing serious crimes rather than minor offenses.

Why would stiffer penalties increase violent crime? To understand this seeming paradox, you first need to understand the nature of California’s three-strikes law. Not just any offense gets you a first strike. It must be a so-called “record-aggravating” offense, which includes violent crimes like assault and rape as well as serious nonviolent crimes such as burglary or drug sales to minors. But after strike one, strikes two and three can come from any felony, including minor offenses like possession of marijuana or even stealing golf clubs or videotapes. A third strike carries with it a mandatory sentence of at least 25 years in prison.

Now, put yourself in the shoes of a two-strike criminal. The prospect of 25 years behind bars for a third offense is likely to give even a hardened criminal pause before he or she crosses the street against the lights. So we’d expect two-strike felons to commit fewer crimes. But suppose you’ve already decided to break the law—maybe you need to make a quick buck. Are you going to lift a few golf clubs from the local pro shop? Or are you going to hold up a bank? The potential haul from a bank robbery is obviously much greater, and the penalty is the same: Bank robbery will get you decades in the slammer, but if it’s your third offense, so will shoplifting.

Even if you don’t quite have the chutzpah to pull off a bank job, you still might end up committing a more violent crime if you’re in a 0-2 hole. Let’s say you opt for the golf club caper, but as you’re making your getaway, you’re cornered by a store security guard. Do you surrender quietly or pull out a gun? If strike three is looming, it’s all the same to you whether you end up on trial for shoplifting or armed assault, so why not try to shoot your way out of an arrest?

Proponents of three-strikes laws point to declines across the board in crime rates in California during the 1990s, following the passage of the three-strikes law—including rates of violent crime. But crime was dropping around the country during that period, with explanations ranging from new policing tactics to the legalization of abortion. With so much going on, it’s hard to know how much, if any, of the decline comes from fear of a third strike. Instead of analyzing aggregate crime data, Iyengar looks at the lawbreaking choices of individual criminals. She examines how their lawbreaking activities change when the three-strikes law is on the books and also how their lawbreaking activities change depending on how many strikes they have against them.

Using data from all criminal convictions during 1990 through 1999 in California’s three biggest cities—Los Angeles, San Francisco, and San Diego—Iyengar finds that the three-strikes law did indeed have a large effect on the likelihood of recidivating (committing a crime after release from prison) in the two years following a prior offense. For those with one strike, the law reduced recidivism by 14 percent; this doubled to a 28 percent reduction for two-strikers, whose next crime would trigger the minimum 25-year prison term.

But that’s where the good news ends. Three-strike-eligible criminals who actually do get arrested for a third offense commit more serious crimes. Burglars, for example, become robbers—these are both offenses that involve stealing, but robbery has the added element of force. Similarly, while thefts decline overall, assaults during thefts go up under three strikes, suggesting that an increasing number of thieves may, in desperation, be trying to muscle their way out of a third arrest (as in our golf club example). In general, arrests of three-strike-eligible felons are 20 percent more likely to be violent crimes (relative to no-strike criminals).

(A Californian burglar on the verge of a third strike has an even safer option for his next act—take his activities out of state. Just across the border in Arizona, there’s no three-strikes law at all, and in neighboring Nevada, the law is rarely invoked. So rather than breaking and entering in Los Angeles, why not take a road trip to Las Vegas or Phoenix instead? It seems that many criminals do. Iyengar finds that a larger fraction of repeat offenders recidivate out of state after the three-strikes law’s passage.)

Overall, the three-strikes law did have the desired effect of deterring repeat offenders from striking again. But the law’s original intent—motivated as it was by Polly Klaas’ tragic story—was to avert further violent tragedies by putting habitual criminals away for a good, long time. It’s putting away violent criminals, but Iyengar’s study suggests it’s also making criminals more violent. It’s tempting to invoke the law of unintended consequences in thinking about what was perhaps a well-intentioned but flawed piece of legislation. But these consequences could have been entirely anticipated if legislators recognized that criminals, like all of us, often make decisions by rationally weighing the costs and benefits of their actions.