Looking to Live in a Community with Low Murder Rates? Try Committing a Crime

Crime rates have a large influence on the choices people make about where to live. The amazing declines in crime over the last fifteen years have been especially strong in big cities, a factor that helped fuel an urban renaissance. Ironically, however, some of the lowest murder rates are found in places where one might suspect just the opposite to be true: U.S. prisons. The Bureau of Justice Statistics recently released data on the causes of death among inmates in state prisons. In 2005, 56 prisoners were murdered. There are roughly 2 million inmates held in state prisons, meaning that the homicide rate per 100,000 prisoners last year was only 2.8. That number is less than half the rate of New York City (6.6 per 100,000) and an order of magnitude lower than Baltimore (42 per 100,000). Indeed, of the 66 largest cities in the United States, only El Paso, Tex. and Honolulu, Hawaii have lower homicide rates than the state prisons.

Interestingly, suicide rates in prison are about average for the U.S. There were 215 suicides in state prisons in 2005, for a rate of roughly 10 per 100,000. The overall suicide rate for all Americans is 10.6 per 100,000.

These low homicide and suicide rates are both testimony to the fact that prisons are incredibly highly controlled environments. Whenever I have visited prisons, I have been amazed at how safe I felt. In contrast, when doing ride-alongs in police cars, I’ve always had the feeling that something crazy could happen at any moment.

So if you feel there is too much crime in your own neighborhood, there is a simple solution to your problem: just commit a crime yourself. Your new home in prison will likely be a much safer place to live.


Adrian (#12), beyond the label that they slap on the overall facility, there are some known differences. The best known issue is that inmates are safer in a federal prison than in an equivalent state prison, largely due to population differences: most murders (and other violent crimes) are prosecuted under state laws instead of federal laws. Federal prisoners are more likely to be in for, oh, tax evasion or securities fraud.

Inmates are safer when they are away from violent inmates. Inmates are safer when they are away from inmates with uncontrolled mental illness. Inmates are safer when they are in a segregated geriatric or physically disabled population and away from younger and healthier inmates. Inmates are also generally safer when there's basically no change at the facilities. Even positive changes can trigger problems (see "uncontrolled mental illness").

I don't even want to pretend to be able to back this up, but my personal impression is that inmates are also safer when they're predictable, basically nice people with basically nothing of value in their possession, because that kind of person is less likely to set off someone else (see "uncontrolled mental illness" again), or to be robbed (because you can't steal what doesn't exist).

Jacques (#15), I don't know the answer to your question. Someone else might. I know some questions I'd ask, though: Do all the countries have exactly the same rules about what triggers a criminal report after inmate-inmate sexual contact? Are the inmate populations sufficiently similar in all material respects (like age and the prevalence of mental illness and paraphilias) to make such comparisons useful? Given that those countries have essentially universal health care and very different rules about treating mental illness *before* you hurt someone else, I would expect that the prison populations are quite different.



Hello, Steven. I've been a fan of your publications since I read Freakonomics. I live in Costa Rica in Central America and I really love this kind of freaky facts, and this kind of stuff that we never thought. Thanks for publishing interesting articles.


In prison
Killer more likely to be caught, convicted and executed.

Deceased is more likely to belong to a gang who will punish his killers

Most street murders are familial, not a problem in most prisons

Drug and alcohol are a partial explanation for street murder, less so in prison.

So if you live in a high murder area, but get along with family members, don't drink or use drugs, have friends who will avenge your death, and the local police are very good at catching and punishing killers, you might be just as safe outside prison.

Staz S

Guys, he's a columnist here, not an accountant.
Just relaz and have fun.


Silvanus, where in the blog does Levitt suggest that he is analyzing neighborhood quality, or equating this unexamined issue with murder rate? Murder rate is humorously mentioned as a factor that, if employed, indicates little, given that your life may well be lost to the whimsy of the local cul-de-sac "killer".

Further, how might you analyze personal satisfaction indices, unless each contributor happens to factor the same criteria, each of similar relevance, into his/her personal satisfaction index. Are you familiar with elementary statistics or logic, because Levitt is neither offering prospective movers heretofore unknown criteria, relevant to relocation choice, nor is he suggesting murder rates ought be considered with furrowed brow, given their sudden life-threatening presence in cities (other than El Paso or Honolulu). Or, call me crazy, but maybe prisons are safer than expected.

Any hypothesis drawn, by the way, must include prison murder rates during the later Warren Court years, when incarceration rates were probably among the last half century's lowest. I'd guess murder rates were higher then, if only because those few behind bars suffered for their free kindred.



i can't believe what i have had just read. put to the fore facts is fine. but you are really only giving one side of the story. spending time in prison has permanent negative effect on most of your outcomes: health, life expectancy, wages,...


How can we lower crime rates without turning our society into a society controlled like the prisons? I for one would like to live in a low crime area WITHOUT entering prison. haha.


theoldhenk, how many people live on your block? 200? Then at average rates you'd expect to see maybe one murder a century. Obviously such blocks will have a "0 in 100,000" or an "n*500 in 100,000" murder rate in any given year, and trying to backtrack from the final crime statistics to get a risk probability is useless when you have such a small probability and such a small sample size to work with.

The best you can do is to try to group similar neighborhoods together (as is done here by city) to try to get a less stochastic aggregate number. You can surely find many other variables besides "what city am I in" to use for this, and I'd like to eventually see some of the results (starting with average income, say), but I don't think blog entries like this one should be withheld until they can be filled with six dozen interesting facts instead of just six.



/facetious: Why, lordy, how do me? Have I evah had a day of mah life in a numbahs course? Why landsake, no, I guess thatah been the endah me, fo sho!


There are a number of personal satisfaction scales within the literature. All you have to do is follow the named scale with (Smith, 1982- example) along with accompanying studies to make the case why that particular personal satisfaction scale for neighborhood was the right one to use.

I resent your tone btw. However, in case you failed to read, here are Levitt's reasoning sentences..."Crime rates have a large influence on the choices people make about where to live... These low homicide and suicide rates are both testimony to the fact that prisons are incredibly highly controlled environments...So if you feel there is too much crime in your own neighborhood, there is a simple solution to your problem: just commit a crime yourself. Your new home in prison will likely be a much safer place to live."

While Levitt may be facetious (which I typed in my previous post)- the "tongue in cheek" tone is not immediately clear. Making the claim that prisons are safer than regular neighborhoods is a possible conclusion based on the cited references- however, qualitative data is important to form a more holistic picture avoid such a sloppy conclusion. Considering there are two or three commenters who have praised Levitt's conclusion, I'd suggest that perhaps it is in Levitt's interest as an opinion leader to clarify the truth- the worst sin in Christianity is to lead a child away from Christ... so too is the worst sin in philosophy to lead others into error.


Mark in SF

Aw, man, how am I going to suspend my disbelief next time I watch a prison movie? I though inmates got shanked all the time.


Unfortunately the few crimes being committed in El Paso Tx are being done either by County Administration (they are being heavily investigated by the FBI under corruption charges) or in the surrounding area by the Sheriff's department by raiding schools and rounding up kids and entering private houses under false pretenses and requiring citizenship proof...other than that, yes, we are the safest border town in the US with a population over 500,000.


best misleading post ever... Pristine nonsense !

thaddeus buttmunch

He failed to mention that the prison murder rate went down 90% since 1980 when they began segregating violent from non violent offenders.

"The Moneychangers" was heart wrenching in the seventies. Guy wouldnt testify against criminals so he got shower-sodomized. But these days he would get immunity or protctive custody and let's not talk about FUBAR New Mexico prison riot in 1980 which caused all the reforms

Robert S. Porter

But what is the rate of assult and rape...




*Echoes Robert S. Porter's concerns*

In addition- Theft? Intimidation? Guard brutality?

While Levitt appears to be facetious in his analysis, I'd suggest that personal satisfaction indexes are a better indication of neighborhood quality than the criteria Levitt proposes.

Regarding suicide, opportunity is very important to suicide studies. If every inmate were given access to a bottle of Tylenol, I'd suggest the suicide rate would probably increase... and the favored method would be death by Tylenol.


but you wouldn't call that _living life_, would you?


Robert, it's difficult to get good numbers on sexual assaults in prisons. Much of the sexual activity is consensual (meaning that any reasonable person would call it consensual if the same words and acts passed between two university students).

However, since sex is banned in prisons, then an inmate who is caught having sex will often report it as rape. Some inmates see rape allegations as a way of insulating themselves from misconduct charges. Some see it as a way to get special attention or to manipulate the system (to get a different cellmate, for example). Others see it as a way of reinforcing their conformation to a heteronormative roles (as in, "I'm not gay; he raped me!")

Certainly some sex in prisons is non-consensual, but I don't think that is actually a common experience outside of the movies.

(Some of it is purely fictitious, too. Inmates are not known to be the single most truthful population, and trumped-up sexual assualt charges are occasionally used as a lever for extortion.)



Ummm ... is murder the only crime one should worry about in prison? How do the physical and sexual assault rates compare? And how about attempted murder -- not just successful murder? Let's not set aside good social science just for a clever headline.


Why is this surprising? Prisons are extremely thoroughly controlled environments, with guards, bars, solitary confinement for high risk inmates, and most weapons are completely contraband and carefully kept out of the hands of inmates. It's easy to keep murder rates low when you're dealing with a population that has very few individual freedoms and nowhere to hide.