An article in the Economist from last year, arguing for open borders says:

If lots of people migrated from war-torn Syria, gangster-plagued Guatemala or chaotic Congo, would they bring mayhem with them? It is an understandable fear (and one that anti-immigrant politicians play on), but there is little besides conjecture and anecdotal evidence to support it. Granted, some immigrants commit crimes, or even headline-grabbing acts of terrorism. But in America the foreign-born are only a fifth as likely to be incarcerated as the native-born. In some European countries, such as Sweden, migrants are more likely to get into trouble than locals, but this is mostly because they are more likely to be young and male. A study of migration flows among 145 countries between 1970 and 2000 by researchers at the University of Warwick found that migration was more likely to reduce terrorism than increase it, largely because migration fosters economic growth.

Is the highlighted claim about the US (America = USA, no doubt) found[ed] in credible research? It obviously doesn't mesh too well with US presidential claims... and the difference in odds claimed is pretty large (and not fitting the pattern from some EU countries, mentioned in the next sentence.)

  • If a source doesn't include per-capita numbers (or enough info to calculate them), i'd suspect equivocation between "1/5 as likely to be incarcerated" and "1/5 of incarcerated people are immigrants".
    – cHao
    Jul 10, 2018 at 13:00
  • There's a difference between "being incarcerated" and "comminging a crime". The former implies successful apprehention and prosecution. There's multiple factors that bias the former number against being as likely among foreign born.
    – user5341
    Jul 10, 2018 at 14:06
  • @user5341: it occurred to me that some illegal immigrants might just be deported (instead). But without looking at some study (hopefully the number is not from thin air) it's hard to tell what exactly is included or not in the stats. Jul 10, 2018 at 14:22
  • 2
    "doesn't mesh too well with US presidential claims". These days that's not a reason to doubt it. Jul 10, 2018 at 15:10
  • 2
    Something to consider here: Most of those people in the US had to keep their noses clean to be able to immigrate in the first place. The EU currently has a large number of people whose backgrounds haven't been checked. Of course you would expect more criminality amongst the EU immigrants. Jul 11, 2018 at 4:55

3 Answers 3


It's not one fifth as likely, it's about one third as likely. Depending on exactly what you mean.

The Cato Institute Study of Criminal Immigrant Demographics compared figures on incarceration rates of native born Americans, legal immigrants and illegal immigrants.

The incarceration rate was 1.53 percent for natives, 0.85 percent for illegal immigrants, and 0.47 percent for legal immigrants

In other words 1.53% of all native born Americans were imprisoned, compared with 0.85% of those in the US illegally and 0.47% of those who came to the US legally.

Statistically a native born American is about 3 times as likely to be imprisoned as a legal immigrant, and about twice as likely as an illegal immigrant. It's about three times more compared with an immigrant of any kind.

  • The report linked in the answer says there were 122,939 incarcerated illegal immigrants ages 18-54. How is that 0.85%? There are only 12 million illegal immigrants, and only 10 million in that age range. dhs.gov/sites/default/files/publications/… It seems that the report uses a big number for illegal immigrant population, to get a low incarceration rate. Figure 2 even claims that illegal immigrants are 9.06% of the population.
    – DavePhD
    Jul 10, 2018 at 19:25
  • Definite figures for illegal immigration are hard to find (surprise!). Since the question is about immigration in general, it doesn't actually make a lot of difference. Jul 10, 2018 at 19:36
  • 3
    @DavePhD The article doesn't say "non-citizen foreign-born" so why would you assume naturalized citizens should be omitted from the calculation?
    – 1006a
    Jul 10, 2018 at 22:49
  • 1
    According to Wikipedia about 39 million of the US are immigrants (including citizens), and about a third are probably in the 18-65 age range, so the figures seem right to me. Jul 11, 2018 at 0:28
  • 2
    I am worried about a confounding factor: Are undocumented (and documented) immigrants simply deported rather than incarcerated, giving misleading figures? [This is from ignorance; I haven't read deep enough to know either way.]
    – Oddthinking
    Jul 11, 2018 at 8:55

The fraction of people in the United States who were born elsewhere was 12.9% according to the most recent census. (Many of these are citizens, only 7.2% of the population is foreign born non-citizens).

According to the Federal Bureau of Prisons 20.1% of inmates were not citizens of the United States.

So, for federal incarceration, foreign born are clearly incarcerated at a higher rate.

According to the Congressional Research Service document Interior Immigration Enforcement: Criminal Alien Programs, when state and local inmates are included, non-citizens were 7.4% of the incarcerated population in 2012 (the last year that California cooperated with federal requests for citizenship information).

So as far as is known, non-citizens are incarcerated at similar rates to citizens, not one fifth the rate.

  • 2
    Hmm, your stats give a much worse ratio than even the Cato study... and I suspect they're not pro-immigration much. I wonder how may of those incarcerated in Federal prisons have committed their crimes outside the US, e.g. drug dealers extradited or flow-in more or less legally. Jul 10, 2018 at 14:33
  • 2
    @Fizz mine are only for federal inmates. If a good source is found for state and local inmates, the answer could be changed. However, the Department of Justice says that some states, such as California, refuse to provide citizenship data of prisoners, and other states under count non-citizen prisoners. See page 13 here: bjs.gov/content/pub/pdf/p16.pdf
    – DavePhD
    Jul 10, 2018 at 14:38
  • 4
    @Fizz - Cato is likely very pro-immigration. They are a libertarian think tank, not a conservative one, even if left wingers don't make fine distinctions
    – user5341
    Jul 10, 2018 at 14:38
  • 5
    But I see there's a critique of this method: bipartisanpolicy.org/blog/… It does say (as I suspected): "It does appear, however, that the prevalence of undocumented and other immigrants is largely the result of immigration enforcement priorities, not necessarily increased rates of overall criminality among immigrants. Experts have noted that the large number of individuals imprisoned for federal immigration offenses contributes significantly to the size of the federal prison foreign population. " Jul 10, 2018 at 14:45
  • 6
    @Fizz According to this: politifact.com/punditfact/statements/2015/jul/10/larry-oconnor/… 19,000 of 48,470 federal non-citizen inmates are incarcerated for immigration law violations, so if the OP claim was changed to exclude immigration crimes, the 20.1% number could be reduced to about 12%.
    – DavePhD
    Jul 10, 2018 at 15:12

Whenever immigration is discussed, you always have to be wary about conflating "immigration" with "illegal immigration". Or, more relevantly, people purposely misleading based on it.

In this case, the argument is, "People in the US that weren't born in the US commit less crime, so we should have more legal immigration." But the problem is, they're including the stat for legal immigrants - people that are already part of the population being admitted for citizenry (which will be part of the country whether we have 'Open Borders' or not.) Instead, the question is, are illegal immigrants causing more/less crime than native citizens?

That's a lot tighter of a question. From some quick googling, it looks like it's true - illegal immigrants cause less crime (https://www.cato.org/publications/immigration-reform-bulletin/criminal-immigrants-their-numbers-demographics-countries) - but it's definitely not the "a fifth as likely".

Short Story: Basically, the person making the claim is using slight-of-hand to get a more compelling number... but the actual number still goes in their favor (just not to an extent that it should probably drive immigration policy.)

  • 1
    It would help if you added some numbers to your answer. If we go with the Cato study, the percentage of incarcerated legal immigrants is 0.47%, while for illegals is 0.85%, compared to 1.53% for natives, so that's a approx 1:3 ratio when only legals vs natives are considered and slightly.less than 1:2 (1:1.8) on illegal vs native. Jul 10, 2018 at 14:30
  • 5
    No the question does not mention 'illegal immigration' at all. Nor does the article it is sourced from. You are not answering the question asked. Jul 10, 2018 at 15:10
  • 2
    The question didn't explicitly say "illegal immigration", but it did say "Open Borders". (As a general rule of thumb, people arguing for Open Borders tend not to ever use the phrase 'Illegal Immigration', since it kinda emphasizes that the people it's talking about are doing an illegal act in circumventing regular immigration.) Anyways, my answer is basically, "It doesn't matter whether or not the one-fifth stat is correct, because it's not relevant to the discourse anyways."
    – Kevin
    Jul 10, 2018 at 15:19
  • 5
    Open Borders is not synonymous with 'illegal immigration'. Even if it were the question is explicitly asking about immigrants, not illegal immigrants. Feel free to read whatever subtext you like into the article, but the claim is clear and it's about immigration in general. Jul 10, 2018 at 18:24
  • 3
    -1: The question is about foreign born. You answer is about "illegal" immigrants. If you want to add a comment warning people to not conflate foreign-born versus native born, citizen versus non-citizen, and documented versus undocumented, great. If you want to also give other stats for context, to show the original question is misleading, great. But you should answer the right question. I would delete this answer outright as "not an answer", but it just so happens that your reference ALSO addresses the actual question, so maybe this is hope for a fix here.
    – Oddthinking
    Jul 10, 2018 at 22:35

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .