Bryan Caplan  

Immigration and Crime: Tell Me What You See

PRINT
The Value of the Right of Lega... Leon Panetta: Wars Will Go On ...
Compare these GAO statistics on "Federal Prison Illegal Alien Inmates" (p.23) to the Bureau of Justice Statistics table on the "Number of Sentenced Inmates in Federal Prisons" (p.10).  Both sets of figures come from c.2003.  Tell me what you see.



COMMENTS (18 to date)
Liberal Roman writes:

Wow. Scary, scary depressing and maddening stuff. 90% of people in federal jails shouldn't even be there. Insane.

Liberal Roman writes:

When Tyler Cowen writes of the Great Stagnation and that we have eaten all the low hanging fruits, why does he not consider this?

How much lower can this fruit hang? Stop jailing people for things that shouldn't be crimes.

Kevin L writes:

Well, for one, incarceration rates among illegals is quite a bit lower than the general population: 550 per 100,000 (assuming 10 million illegal aliens in the country) vs. 712 per 100,000 for the 2003 population overall. I didn't see recidivism rates in the DoJ report. Interesting how the GAO report doesn't directly mention any estimates of incarceration rates among illegals.

Miguel Madeira writes:

1 - The most popular "crime" within illegal aliens is... immigration itself.

2 - violent crimes are less common in illegal aliens than within general population.

Kevin L writes:

Here's my comment on Google Reader after looking at the numbers for a few more minutes:

According to the numbers in the GAO report, illegals have about a 30% lower incarceration rate than the general population. Also, they accounted for only 2.7% of the total prison population.

If you exclude "immigration" offenses - since the rest of the population doesn't get convicted for those - their rate drops to 438 per 100,000 (40% less than the general population) and 2.1% of the prison population.

By these numbers, illegals are making our country a safer place.

John Jenkins writes:

What I see is someone who has nothing to do with the criminal justice system cherry-picking data and suffering from availability bias. Immigrants of all types are no more likely to be criminals than anyone else, arguments to the contrary notwithstanding, but, these data fail to account for a few things.

First, we most often commit crimes against people most like ourselves, and undocumented immigrants are the people least likely to report a crime because they are more worried about forcible removal than crime. This is a serious concern for police in a lot of cities (who generally just don't care about immigration "offenses").

Second, non-citizens who are not LPRs can simply be removed on criminal conviction (and LPRs can be expelled on certain convictions), so they would not show up in these statistics.

Finally, the federal system is mostly well-stocked with drug offenders, who are more likely to be citizens (drug addicts and dealers don't tend to immigrate; they buy local or hire local people because it's cheaper that way), so I suspect the discrepancy would close substantially by including state-level data (which is not readily available), and if you removed drug "offenses" would disappear entirely.

Jason Collins writes:

A possible explanation for the observation that "violent crimes are less common in illegal aliens than within general population" is that the potential for deportation acts as a deterrent against crime.

Garrett writes:

I see that illegal aliens are more sneaky at their crimes and are better at not getting caught.

Mr. Econotarian writes:

It makes sense that illegal aliens are less likely to be criminal, as they have far more to lose than a citizen from even a minor conviction (deportation), or an arrest for that matter. Besides, they are people self-selected for seeking a better place to work.

Their (citizen) kids may be a different story though...

Tom writes:

The GAO report only lists the crime that has the longest sentence (in the footnote). The previous two tables show that 98% of illegals were arrested for 2 or more offenses, and the most common offense was drug related. (We should end the war on drugs but that is a separate topic.)

That makes the comparison with the BJS problematic since it is unclear what criteria they are using to sort prisoners. The footnote says that the length of sentence was not considered, but then how do you decide between immigration and the other crimes they've committed? I sincerely hope the surge in immigration incarcerations is not simply due to changing definitions (incredibly misleading) but I am not exactly enamored with government competence.

In the end, I'm not sure what conclusion we are expected to draw from this.

Eric Hammer writes:

Definitely seems like illegal immigrants are much less likely to commit crimes. I agree with those above who suggest that they probably self select for being inclined to work and avoid trouble that would get them kicked out. I might expect minor "quality of life" crimes, but not likely anything too serious.

In addition though, I wonder how often the government locks up immigrants for being here illegally when they suspect but can't prove they have committed another crime. Sort of like Al Capone going to jail for tax evasion.

I also wonder why we would lock up illegals instead of deporting them, that's weird. I bet there is some information lost between why the authorities were looking at certain people and what they were eventually convicted of. Still, it supports my views that immigrants in general are a good sort.

Clay writes:

The children and grandchildren of immigrant populations is generally where you see the crime and other social dysfunction. Adult immigrants themselves, due to common life experiences and the self-selection involved in immigration, tend to be low crime.

Aren't genetics a strong casual factor of individual criminal behavior and social dysfunction? If so, why wouldn't countries want to screen for that kind of thing?

Evan writes:
A possible explanation for the observation that "violent crimes are less common in illegal aliens than within general population" is that the potential for deportation acts as a deterrent against crime.
A simple way to test this theory would be to look at a part of America where it was possible to deport citizens for committing crimes. Such a place existed at one point: the Panama Canal Zone. If you misbehaved there they'd send you back to the States (I wonder if other nonstate territories, like military bases, have similar policies). Now, I don't have any crime statistics for that area on hand, but someone who used to live there told me that the crime rate was very low. So that theory seems to hold water.

It would be interesting to see why deportation serves as more effective deterrent than normal punishment. My best guess would be that it's faster and more certain and there's less opportunity for appeal. Therefore, native violent crime might be more easily deterred if punishment is faster and more certain (this might be achieved by making punishments less severe, since the more severe a sentence is, the easier it is to appeal). Of course, you also have to consider the possibility of getting the wrong guy. I have no idea how many illegals who are deported for violent crime are not guilty of it.

Aren't genetics a strong casual factor of individual criminal behavior and social dysfunction? If so, why wouldn't countries want to screen for that kind of thing?
Because genetically based screening policies remind people of Jim Crow. That episode in our history was so harmful no one wants to implement any policy that even vaguely resembles it.
MikeP writes:

It would be interesting to see why deportation serves as more effective deterrent than normal punishment. My best guess would be that it's faster and more certain and there's less opportunity for appeal.

My best guess is that the net present value of living and working in the US versus a developing country for a 20-year-old is on the order of half a million dollars while the NPV of staying out of jail for a year is a tenth of that.

Showing up on ICE's radar and being deported are simply far greater punishments than spending a few months behind bars. Thus they are greater deterrents.

Nathan Smith writes:

It's nice to see that immigration substantially reduces crime rates. However, even if immigrants committed more crimes than natives, that would be a small price to pay for the opportunity to do business and have cultural interactions with millions of hard-working and enterprising people. In any case, it would be somewhat unjust to exclude non-criminal immigrants for the crimes of immigrants. So it is worth insisting that even in an alternative world where immigration increased crime, open borders would still be the right policy. It's nice that crime reduction is an added bonus from immigration, though.

George writes:

Here's a simple model: 1% of natives are violent criminals, 1% of illegal aliens are violent criminals, and there are equal numbers of (total) natives and illegal aliens.

Whenever they are free, either type is certain to commit a violent crime, certain to be caught and convicted, and certain to spend 5 years in prison.

At the end of a native's sentence, he is released, commits another crime, and is caught, convicted, and sent back to prison.

At the end of an illegal alien's sentence, he is deported, and thus unable to commit another violent crime in the U.S.

What will the prison population look like? If a violent criminal survives for 9 sentences, then the population will be 90% native violent criminals, even though natives offend at exactly the same rate as illegal aliens, and the proportion of total natives exactly equals the proportion of total illegal aliens.

Bryan, I think you're suffering from confirmation bias here. If you were on the other side of the issue, you would have poked about a dozen holes of this size in the "look at the numbers" argument by now. (Though, to be fair, you only said, "Tell me what you see.")

(P.S. to Evan: I'm not sure the Canal Zone is all that comparable: if Mexico punished violent criminals by deporting them to the U.S, I think their violent crime rate would go up.)

David C writes:

Nobody has spotted it yet?

Page 12 of the GAO report
Number of convicted criminal aliens as of Dec 27, 2003 that ICE determined to have entered the country illegally: 55,322

Page 10 of the BJS report
ICE reported 23,514 detainees at yearend 2003

That's a pretty big difference.

AirmanSpryShark writes:

@David C: You're closer than the others, but you're on the wrong page in the first (check the original post).

p. 23 GAO: 12,694
p. 10 BJS: 16,903

That's a 33% discrepancy in the count of those incarcerated with Immigration as their 'most serious offense'.

Comments for this entry have been closed
Return to top