What If There Were More than 30 Million Illegal Immigrants?

Most nativist narratives would collapse

Vice President JD Vance and a collection of online conservative commentators have argued that Biden let in 20 or 30 million illegal immigrants, swelling their population to 30-50 million. These numbers are an exaggeration, but there was a large increase in the illegal immigrant population during the Biden administration. Based on pre-2020 population trends, net immigration exceeded estimates by about 8.7 million, with illegal immigrants accounting for about two-thirds of the increase. This means there was a net increase in the illegal immigrant population of 5.5 to 6 million during Biden’s administration. What if the alarmists are correct and there are more than 33 million illegal immigrants in the United States? If true, then illegal immigrants would have much less of an impact on American society than nativists claim.

Background

Other research organizations and the Department of Homeland Security (DHS) still claim there are about 10 to 12 million illegal immigrants, which is implausible. My research used a residual statistical method to estimate almost 14 million nationwide in 2022, and I expect there to be a few million more that I can't statistically identify because they're even less likely to answer surveys than demographers estimate. Older research has controversially claimed that the number of illegal immigrants currently residing in the United States is at least 50 percent greater than previously thought, and some of their arguments are aging well.

One of the best pieces of evidence against the belief that there are so many more illegal immigrants than is commonly believed is the data regarding their children born in the US. If there were a higher number of illegal immigrants, we would expect to see more children born here, reflected in birth data and school records. However, this is not the case; there are significantly fewer children born here than demographers would expect if there were more illegal immigrants. That's not as convincing as it first sounds because illegal immigrants are rational economic actors, and it would make sense for them to reduce their fertility while living and working in the black market. People will tend to have fewer children if they are in a precarious and unpredictable legal situation that could result in arrest, detention, and deportation.

However, there are many reasons to be skeptical of the estimated population size of 10-12 million illegal immigrants because the residual method used to estimate it relies on some unusual assumptions. The residual method is just a top-down subtraction approach: it starts with the entire foreign-born population and subtracts legal immigrants (or illegal, depending on the exact methods used) by applying various filters. Those remaining individuals are therefore illegal immigrants. Often, demographers will also compare the known flows of legal and illegal immigrants and consult surveys of migrants after they’ve returned to their home countries. So far, so good. If one wants to know the number of men in a population, subtracting the number of women from the total population is a perfectly sound way to do so.

The obvious problem with the above is that counting the entire population is difficult. For example, demographers must make sampling choices, foreign-born residents are more difficult to count, and illegal immigrants are the most difficult to count because they don’t want to be identified by the government. As a result, the American Community Survey periodically updates its methods, while other surveys adjust for undercounts in ways that increase the illegal immigrant population by different rates over time.

George Borjas’ comments on DHS’s undercount assumptions are worth considering. He wrote that the undercount assumption for the 2000 Census

[C]omes from data reported in an unpublished study that looks at the undercount rate of Mexicans in Los Angeles County in 2000. It is doubtful that this statistic, even if it were correct in that very narrow context, provides much information about the undercount rate for the much larger population of undocumented immigrants today.

Given the shaky foundation for this key assumption in the DHS calculation, it is understandable that observers seem perpetually surprised whenever a government program tries to predict the number of undocumented immigrants that will apply for some benefit and ‘unexpectedly’ many more show up. As the Sacramento Bee reported in April 2015: ‘A surge of undocumented immigrants seeking driver’s licenses has surprised the California Department of Motor Vehicles, pouring in at twice the rate officials expected.’

The residual method is unusual and perhaps insensible because of its assumptions. Regardless of the true number of illegal immigrants, it's almost certainly higher than the 11-12 million number commonly cited, but not by as much as 20 million. Additionally, many illegal immigrants did not enter or stay during the Biden administration. But let’s think through the implications: what if there were 30 million illegal immigrants in the United States, about double my current estimate?

What if There Were 20 Million More Illegal Immigrants?

Illegal immigrant crime and incarceration rates would be among the lowest of any subpopulation in the United States. These rates are calculated by dividing the incarcerated population by the total population, multiplied by 100,000. If the total population increases, the crime rate mechanically falls (the denominator rises relative to the numerator). Since we know the number of immigrants in prison and have a good sense of the number of incarcerated illegal immigrants, more than doubling the illegal immigrant population would mechanically halve the incarceration rate.

If there were 20 million more illegal immigrants in 2023, the nationwide incarceration rate for illegal immigrants would be 272 per 100,000, about 15 percent below that of legal immigrants, and almost 80 percent below the native-born rate. This would make illegal immigrants the most law-abiding subpopulation in the United States by nativity. If applied uniformly across illegal immigrant racial and ethnic groups, it would mean that the Asian illegal immigrant incarceration rate would be over 13 times lower than that of native-born white individuals.

Assuming a uniform geographic distribution, an illegal immigrant population of more than 30 million would also lower the Texas illegal immigrant criminal conviction rate for homicide from 3.1 per 100,000 in 2022 to 1.3, about a third of the native-born American rate and 28 percent below that of legal immigrants. But if more of the recent claimed massive increase in illegal immigrants was concentrated in Texas, which seems likely because it’s a border state, then the rate would be even lower. Illegal immigrants look like they have a homicide conviction rate in Texas lower than that of Europeans in Europe.

Terrorists who entered as illegal immigrants have murdered zero people in attacks on US soil since 1975. Even if the population of illegal immigrants who crossed the border or otherwise entered without inspection was somehow many times greater than most demographers estimate, it would only indicate that the terrorist threat is even smaller than I estimate. The US government already spends much more on counterterrorism than it should, but the overspending would be even more egregious with such a large illegal immigrant population because that would mean the threat is vastly smaller than estimated.

Illegal immigrants voting in American elections is a statistical nonissue of such unimportance that it's an affront to the gods of efficiency that I even have to write about it, but such is the democratic process. A statistically insignificant number likely registered to vote, and few actually cast ballots. However, if there were more than twice as many illegal immigrants, we'd expect more mistakes where they registered to vote. Or even where some of them criminally did so. If there really are that many more illegal immigrants here, their criminal (or accidental) election interference is a tiny fraction of an already insignificant number.

Fewer economists discuss the wage effects of immigrants anymore because they are so small and clustered around zero. However, more than twice as many illegal immigrants means that the negative wage elasticities reported by economists like George Borjas are even smaller than he reports, and the substitutability of new illegal immigrants with other immigrants who arrived earlier is greatly exaggerated. Of course, most of them would also be working in the black market without reporting their taxable income.

Commentators instead focus on the fiscal impact of immigrants because the wage effect is so muted and would be even less significant with tens of millions of illegal immigrants living here in the shadows. If there were more than 30 million illegal immigrants, per capita welfare use among noncitizens would be about $2000, rather than the $4,158 my colleague Jerome Famularo and I discovered. Relatedly, fewer illegal immigrants would be children, fewer of them still would be having children (as mentioned above), and almost none of the additional population would be retiring here. Consequently, illegal immigrants would be receiving fewer benefits and paying more in taxes than we estimated. The following result would be a far more positive fiscal effect from immigration.

Suppose there were more than 30 million illegal immigrants, which would more than double the number of noncitizens living in the United States. In that case, it means the effect of immigrants on domestic housing prices is currently overstated by around 25 percent. Housing would still be the market where immigrants affect prices the most, but much less than previously thought.

Immigrants and their children are assimilating well into American culture and the economy. A larger illegal immigrant population would mean that they are even more successful than most of us think, because they would be more embedded in their own communities and have less contact with the outside world. That they would still be assimilating so well with even less contact would really do much to alleviate the worriers who wrongly think American culture isn't attractive enough to encourage assimilation.

The Trump administration claimed that it would deport most or all illegal immigrants, but it won’t be able to do that over four years. Even if the administration somehow convinces half of the illegal immigrants to leave voluntarily, with a population of over 30 million, they’d have to deport an average of 11,396 individuals per day every day for four years without allowing any more illegal immigrants to join the population by entering unlawfully or overstaying their visas. If you thought deporting 13 million illegal immigrants would be difficult, deporting more than 30 million would be impossible.

There are rhetorical advantages to exaggerating the number of illegal immigrants, and most people who do so probably believe it, but there are also many problems. Exaggerating the numbers makes the issue seem more urgent and dire, perhaps increasing support for mass deportation. On the other hand, it would also mean that mass deportations are impossible and may seem daunting to voters. If there are 20 million more illegal immigrants than estimated, they would number about 33 million (according to my residual estimates), and they would be about 9.4 percent of the country’s total population.

To put this in perspective, no country has ever deported so high a percentage of its population during peacetime, and those that came close were not democratic and usually genocidal. For instance, during and after World War II, Stalin deported 10-20 percent of the population of the Baltic states to other internal destinations. Deporting 13 million illegal immigrants was already impossible, and deporting a number almost three times higher, even more so.

Conclusion

Many commentators, journalists, and Vice President JD Vance, among others, have embraced the claim that there are more than 30 million illegal immigrants in the United States. That is extremely unlikely, but if there were, then the overall effects of immigration on the United States would be much smaller than commonly perceived. They would be one of the least crime prone and dangerous subpopulations who have zero effect on the labor market, pay more taxes than they consume in benefits, and are assimilating just fine into a culture that is somewhat hostile. In all, the presence of more than 30 million illegal immigrants would be confirmation of the tremendous contribution of immigrants to the US even when the law is a tragic mess.