In response to The Truth-Teller
I’ve been thinking a lot about this topic over the past week or so, between a new study of Arizona by John Lott and what Peter Kirsanow wrote here over the weekend. I find it to be an especially thorny issue, even setting aside the sketchiness of our data on both illegal-immigrant crime and the size of the illegal-immigrant population — the numerator and denominator, respectively, in any calculation of crime rates.
Here’s a long and winding discussion of the assorted complications, so buckle up.
Probably the biggest stumbling block is that illegal immigrants and natives have completely different demographic profiles. According to estimates from the Migration Policy Institute, illegal immigrants are highly concentrated in the 16–34 age range where criminal behavior is most common: That range covers 43 percent of illegal immigrants but only 26 percent of the native-born. (In part this is because illegal immigrants’ U.S.-born children are automatically made citizens and thus not counted in the illegal-immigrant population.) They are also significantly sex-skewed, at just 46 percent female, to 51 percent for natives.
This demographic gap has profound implications for crime rates. This weekend, using numbers from MPI, the Census, and the Justice Department, I took a stab at estimating what the murder-rate gap between illegal immigrants and natural-born citizens “should” be, based on arrest data for the general population broken out by age and sex and the two groups’ demographic profiles. I’m a journalism major, so my math is always suspect — and there was some outright guesswork involved (owing to a lack of arrest data for those over 65) — but my conclusion was that they should have a murder rate something like 60 percent higher than natural-born citizens do, based on demographics alone.
Yikes! But to put that a different way, if in fact their murder rate is anything less than 60 percent higher, they’re actually less likely to commit murder than natives of the same age and sex.
That generally seems to be the case if Kirsanow’s homicide-incarceration data from five states are indicative. (Indeed, illegal immigrants’ raw rate is lower in two.) But there are caveats. New York is a huge exception, with a 3.5:1 disparity. The tallies exclude a great number of aliens whose legal status is unknown. And as Kirsanow notes, Lott’s estimates for Arizona (which use better data) differ significantly from Kirsanow’s source, with illegal immigrants far more overrepresented among criminals, including murderers, and including in the 15–35 age range. So, however, does a recent Cato study that tries to identify illegal immigrants age 18–54 in Census data (which Kirsanow also mentions), which concludes they’re significantly less likely to be incarcerated than natives.
In general, especially if adjusting for demographics is appropriate, I’d say the picture is pretty murky. But is adjusting for demographics appropriate? I think it depends what kind of immigrants we’re talking about.
If young men are coming here, working for a few years, and going home, they’re pumping up the U.S. crime rate simply by virtue of their being young men, and it makes no sense to wipe that fact out with statistical controls. (A sudden influx of young men during the fracking boom had a similar effect in North Dakota. The costs are very real.) But if they’re staying for the long haul, they’ll grow old eventually, and will have a roughly even mix of male and female children, just like the rest of us — so their current crime rates don’t tell us much about the long-run implications of their presence unless they’re adjusted for age and sex. MPI reports that more than four-fifths of illegal immigrants have been here five years or more, and more than half have been here ten or more.
But the complications don’t end there. For one thing, the overall U.S. murder rate is a blend of the rates of various subgroups — and it’s pushed up especially by the distressingly high rate among blacks, who constitute 13 percent of the population but commit about half of the murders for which the offenders’ races are known. That’s a national emergency to address, not an excuse to let in immigrants with higher crime rates than we would otherwise tolerate. Using overall native crime rates as a benchmark for immigrant crime treats it as the latter. (This also throws a wrench into state-by-state comparisons; some states are much more violent than others, in part owing to different racial demographics, which will make a given group of illegal immigrants look better or worse relative to natives depending on where they live.)
For another, it’s well-known that immigrants in general have low crime rates, possibly because being in a new place induces caution. This effect doesn’t persist in future generations, however. If there’s a similar effect among illegal immigrants — and they have a crime rate similar to natives’ despite it — that could spell trouble in future generations if assimilation falters. Indeed there are already some signs that, on numerous metrics, assimilation stalls after the second generation for Mexican Americans.
It’s getting old now, but a 2006 report from MPI draws together the above two points nicely. (I’m quite sure this wasn’t their intention.) Generally speaking, immigrants have pretty low incarceration rates no matter where they come from. But among the native-born, many nationalities dominated by second- and third-generation immigrants, particularly from Latin America, have considerably higher rates than U.S.-born whites do.
Finally, here’s a point I once made in The American Conservative: “If one immigrant group has the same crime rate as natives, but another has a lower rate, should we not prefer the latter? What we want here is not assimilation, but lower crime.” Illegal immigrants unquestionably have higher crime rates than legal immigrants do, to say nothing of the college graduates we’d let in under a merit-based system.
This is one reason why, going forward, I’d like to get illegal immigration under control through a DACA deal and reorient the legal-immigration system around skills. Illegal immigrants might have unusually high crime rates or they might not — the data are really not clear — but either way we’ll be better off if we actively select our immigrants, rather than failing to police our borders and getting stuck with whoever decides to cross them.