Last Thursday, Tucker Carlson invited Peter Kirsanow onto his top-rated Fox News show Tucker Carlson Tonight to discuss illegal immigration and crime. They began the segment by playing a recent clip of me and Carlson arguing about data on illegal immigrant criminality in Texas. In that earlier segment, Carlson said we don’t have good data on illegal immigrant criminality and I said we do, specifically from the state of Texas. The data show that illegal immigrants have a lower murder conviction rate than native-born Americans. 

Kirsanow responded to my clip in a multi-minute near-monologue. Unfortunately, Kirsanow made many errors and misstatements. His comments on television parroted a piece that he wrote earlier this year in National Review. That piece made so many mathematical, definitional, and logical errors that I rebutted it in detail in Reason this February.

Since I was not invited on Thursday’s segment to debate Kirsanow while he criticized my points and presented his own, I've decided to respond here.  Below are Kirsanow’s quotes from his recent appearance on Tucker Carlson Tonight, followed by my rebuttal.

There’s something called the State Criminal Alien Assistance Program and you can extrapolate from that and get pretty reliable data.

No, you cannot extrapolate from the State Criminal Alien Assistance Program (SCAAP) data to get reliable national estimates of illegal immigrant criminality. The subsequent statistics that Kirsanow uses in his segment are nearly all from a 2011 Government Accountability Office (GAO) report that specifically says, “[w]hile our analysis provides insight into the costs associated with incarcerating criminal aliens in these states and localities, the results of this analysis are not generalizable to other states and localities.” A follow-up GAO report on SCAAP in 2018 repeated the same warning that “[o]verall, our findings are not generalizable to criminal aliens not included in our federal and state and local study populations.” Data from the report that Kirsanow relies upon cannot be used for Kirsanow’s purposes.

SCAAP is a federal program that is supposed to compensate states and localities for incarcerating some illegal immigrants, but it is not a reliable program. As Kirsanow himself admitted, SCAAP only “partially reimburses states and localities for the cost of incarcerating certain criminal aliens [emphasis added].”  States also must choose whom to report to the federal government for SCAAP refunds, which are often small compared to the cost of incarceration, so requests are inconsistent, partial, and the criteria for reporting vary considerably by state. 

He [Alex Nowrasteh] conveniently mentioned Texas to claim that the homicide rates among illegal aliens is 44 percent lower than that of lawful residents. He chose the one state where it is true that the homicide rate is lower for illegal aliens, by 15 percent, not 44 percent.

Kirsanow is mixing and matching his sources here. First, I said that the homicide conviction rate for illegal immigrants in Texas was 44 percent below that of natives in 2016. Unique among all American states, Texas records criminal convictions by crime and the immigration status of the person convicted or arrested. I requested and received data on this from the Texas Department of Public Safety and then made public information requests to every state to see if they kept similar data, but none had. 

Second, Kirsanow said that Texas is the one state where illegal immigrant homicide rates are below those of natives. Even if we analyze the SCAAP data in the GAO reports in the incorrect way that Kirsanow does, there is no evidence for his claim or that the homicide rate for illegal immigrates in Texas in 15 percent below that of native-born Americans.  Kirsanow was likely citing a Cato Immigration Research and Policy Brief that looked at the relative rates of homicide convictions in Texas in 2015, but he got the percentage wrong. In 2015, illegal immigrants had a homicide conviction rate that was 16 percent below that of native-born Americans according to our Brief. 

There are over 300,000 illegal aliens incarcerated.

Kirsanow got this number from the 2011 GAO report mentioned above. That GAO report does state that there were 295,959 incarcerations of criminal aliens in state and local prisons over the course of 2009. Kirsanow incorrectly interpreted what that number meant and made many other errors. 

First, the GAO’s definition of criminal aliens is “[n]oncitizens who are residing in the United States legally or illegally and are convicted of a crime.” Thus, the data on criminal aliens also include legal immigrants who have not yet become citizens. On television, Kirsanow erroneously assumed that the term criminal aliens is synonymous with illegal immigrants, even though he previously acknowledged the distinction in a National Review article, in which he wrote “[a]ccording to GAO, in FY 2009 295,959 SCAAP criminal aliens, of whom approximately 227,600 are illegal aliens, were incarcerated in state jails and prisons.”  

Second, the 295,959 number is the total number of incarcerations of criminal aliens in 2011, not the number of individual criminal aliens incarcerated. The 2011 GAO report states this bluntly: “SCAAP data do not represent the number of unique individuals since these individuals could be incarcerated in multiple SCAAP jurisdictions during the reporting period.”

In other words, the 295,959 number includes many of the same people who have been incarcerated multiple times. If an individual criminal alien was incarcerated for 10 short sentences, released after each one, and then re-incarcerated, then that single alien would account for 10 incarcerations. But Kirsnaow counted him as 10 separate individuals. In Kirsanow’s piece for National Review on this subject, he then compared the number of native-born individuals incarcerated with their total population to estimate relative incarceration rates. In other words, Kirsanow compares the flow of criminal aliens into prison to the stock of all aliens with the stock of natives in prison compared to the stock of all natives in the entire population. Kirsanow confused stocks and flows and his nonsensical apples-to-oranges comparison produced a relatively higher, but incorrect, illegal immigrant incarceration rate.

Third, a quick look at the American Community Survey shows just how wrong Kirsanow is. In 2009, the ACS reported that there were 162,579 non-citizens incarcerated in all federal, state, and local adult correctional facilities (S2601B, 1-year). This is slightly-more than half of the 295,959 incarcerations that SCAAP reports in just state and local prisons. That makes it logically impossible for the 295,959 number to refer to the total number of criminal aliens incarcerated. The ACS counts stocks at a specific time, the GAO counted some flows. Kirsanow is incorrect for talking about the SCAAP figures as if they are stocks of illegal immigrants incarcerated. 

They don’t count the millions of offenses and crimes committed by illegal aliens.

I wish we could count the millions of crimes committed by people that are unsolved or unreported and then study the demographics of the people who committed them, but that’s impossible. Furthermore, we would have to also have that information for all native-born Americans to make a comparison between the illegal immigrant and native-born crime rates. To go even further, I wish we could count everything that didn’t happen as it would immensely improve our world and social science. In the real world, Kirsanow’s statement does not have much relevance. 

John Lott did probably the most methodologically rigorous and comprehensive examination of this type using Arizona Department of Corrections Data.

Kirsanow approvingly cited this working paper by economist John R. Lott Jr. of the Crime Prevention Research Center, in which he purported to find that illegal immigrants in Arizona from 1985 through 2017 have a far higher prison admission rate than U.S. citizens. However, Lott made a small but fatal error that undermined his entire finding: He misidentified a variable in the dataset. Lott wrote his paper based on a dataset he obtained from the Arizona Department of Corrections (ADC) that lists all admitted prisoners in Arizona. According to Lott, the data allowed him to identify “whether they [the prisoners] are illegal or legal residents.” Yet the dataset does not allow him or anybody else to identify illegal immigrants.

The variable that Lott focused on is “CITIZEN.” That variable is broken down into seven categories. Lott erroneously assumed that the third category, called “non-US citizen and deportable,” only counted illegal immigrants. That is not true because non-US citizen and deportable immigrants are not all illegal immigrants, as confirmed by the ADC – the source of Lott’s data. A significant proportion of non-U.S. citizens who are deported every year are legal immigrants who violate the terms of their visas in one way or the other, frequently by committing crimes. According to the American Immigration Council, about 10 percent of people deported annually are Lawful Permanent Residents or green card holders—and that doesn’t include the non-immigrants on other visas who were lawfully present in the United States and then deported. 

Lott mistakenly chose a variable that combines an unknown number of legal immigrants with an unknown number of illegal immigrants and assumed that it only counted illegal immigrants. Lott correctly observed that “[l]umping together documented and undocumented immigrants (and often naturalized citizens) may mean combining very different groups of people.” Unfortunately, the variable he chose also lumped together legal immigrants and illegal immigrants. I wrote about the fatal flaw in Lott’s paper here in February. Lott and I had an exchange here. Kirsanow should have known that Lott’s paper was not methodologically sound because he misidentified the only variable that mattered for his analysis. Lott’s working paper is not the slam dunk that Kirsanow claimed it was.   

Alex is very knowledgeable and that’s why it’s puzzling that he won’t acknowledge the overwhelming amount of data that shows that illegal aliens not only commit more crimes, at a higher rate that is, than lawful residents but more serious crimes at a far higher rate than legal residents.

As I’ve shown above, Kirsanow misread, misinterpreted, and incorrectly defined numerous terms in the GAO report that was his near-exclusive source of information to make an intellectually indefensible case that illegal immigrants are more likely to be criminals than native-born Americans. What’s even more puzzling is that Kirsanow is aware of his errors after a previous exchange that he and I had on this very issue but he chose to repeat them on television regardless. 

Cato scholars have produced much original research on illegal immigrant criminality.  Based on data from the state of Texas in 2015, we found that illegal immigrants have a lower criminal conviction rate than native-born Americans for most crimes in that state (number of convictions), the rate of homicide convictions for illegal immigrants is below that of native-born Americans in 2016 (the number of people in each subpopulation convicted), and that the incarceration rates for illegal immigrants are below those of native-born Americans (but above those of legal immigrants).  Peer-reviewed research also points in roughly the same direction.   

Policy analysts, commentators, politicians, and members of the media have a duty to honestly parse the facts and debate these complex issues in good faith.