Misuse of Texas Data Understates Illegal Immigrant Criminality

The distribution is not static. Over time, authorities could identify additional illegal immigrants who were previously categorized as “other/unknown” and reclassify them as illegal, shrinking the other/unknown share and increasing the illegal share.¹ Due to resource constraints and offenders’ limited durations in custody, not all illegal immigrants are identified as such by either DHS or DCJ. Therefore, the number of illegal immigrants arrested or incarcerated in Texas is in a perpetual state of undercount.

Illustration
Figure 1 demonstrates how failing to understand this undercount can lead to erroneous conclusions about illegal immigrant crime rates. If we counted only the illegal immigrants who were initially identified by DHS when arrested for homicide, the 2012 homicide rate (convictions divided by population) would be 2.7 per 100,000 illegal immigrants in the state, as depicted in the light-blue bar. This rate, while hardly trivial, is lower than the state’s overall homicide conviction rate of 3.0 (orange bar) in 2012.

However, once the illegal immigrants identified during their prison stays are added to the total count of illegals, their homicide rate rises to 3.9 per 100,000 (dark-blue bar) in 2012. This rate is higher than the state average, and it could rise even further as more illegal immigrants continue to be identified in prison.

Figure 1 also illustrates the tendency for the gap between the blue bars to increase over time. Notice, for example, that the rise from the light-blue bar to the dark-blue bar in 2012 is 3.9 – 2.7 = 1.2, but the rise in 2019 is just 0.5. Because Figure 1 is based on data requested in 2021, the state has had less time to identify illegal immigrants convicted in 2019 compared to those convicted in 2012. In other words, the more recent the conviction year, the more likely illegal immigrants are to be left undiscovered in the “other/unknown” category at the time of the data request.

Full Data
Table 1 contains the complete DPS data that we received from our 2021 request, along with the crime rates that we calculated using Census population data. It is sorted first by year, then by the offenses for which illegal immigrants are most likely to be convicted relative to the state population as a whole.

As the table indicates, illegal immigrants in Texas appear to be convicted of crimes such as homicide, sexual assault (shown in Figure 2), and kidnapping at higher rates than the state average. By contrast, they appear to be convicted at lower rates for crimes such as robbery and drugs.

There are two crucial caveats associated with Table 1. First, the “Illegal Conviction Rate” is still underestimated. It is simply the number of illegal immigrant convictions identified by DHS plus the number of illegal immigrant convictions identified subsequently in prison, all divided by the total illegal immigrant population in Texas.² More illegal immigrants could move out of the other/unknown category over time — especially in the most recent years, as discussed above.

A second caveat is that comparisons of the types of crimes that illegal immigrants commit are inherently skewed. The longer people with unknown status are in custody, the more likely it is that Texas will correctly ascertain their immigration status. DHS and Texas DCJ have extensive time and incentive to investigate an individual’s immigration status when the crime is murder or sexual assault. Lesser offenses (e.g., larceny) carry shorter sentences and are a lower priority for deportation purposes, resulting in fewer unknown statuses moving to the “illegal immigrant identified in prison” category over time.

Therefore, the most serious crimes tend to generate the most accurate illegal immigrant conviction rates. Since serious crimes are only a small portion of all crimes, however, any estimate of illegal immigrants’ conviction rate for “all crimes” is practically meaningless, given the limitations of the data.

Mistaken Studies
Recent studies using DPS data understate the extent of illegal immigrant crime by ignoring or downplaying the fact that many arrestees and convicts move from “other/unknown” to illegal over time.

Cato Institute. In February 2018, the Cato Institute’s Alex Nowrasteh released his analysis of 2015 DPS data.³ He concluded that “the conviction and arrest rates for illegal immigrants were lower than those for native-born Americans.” This conclusion was highlighted by the New York Times, Washington Post, NPR, and many other outlets.

Nowrasteh’s error was to treat as native-born anyone who had not yet been categorized as a legal or illegal immigrant. He failed to understand the DPS “other/unknown” category and the movement of illegal immigrants out of that category over time. As discussed above, “native-born” is not a category verified by DPS. Native-born Americans are grouped with yet-to-be-identified immigrants in a catch-all category called “other/unknown”. The number of unknowns shrinks during incarceration as Texas updates the figures upon identification of an inmate’s immigration status.

It is easy to observe how the identification of additional illegal immigrants changes the relative conviction rates over time. For example, in Cato’s analysis of 2015 DPS data, the illegal immigrant homicide conviction rate appeared to be 8 percent lower than the state’s homicide rate. By the time we requested updated data in 2021, after more illegal immigrants had been identified, the 2015 illegal immigrant homicide conviction rate was 20 percent higher than the state rate.

Nowrasteh later published new studies in 2019 (with DPS data for 2017) and in 2021 (with DPS data for 2019).⁴ By now it should be clear that such recent data will not provide an accurate picture because the state has had too little time to identify illegals who are still categorized as “other/unknown”.⁵

PNAS. A 2020 study in the Proceedings of the National Academy of Sciences (PNAS) also claimed low rates of illegal immigrant crime based on Texas DPS data.⁶ Lead author Michael Light, a sociologist at UW-Madison, found that felony arrest rates for illegal immigrants were half that of the “native-born”. His study has been favorably cited by the federal government in response to lawsuits filed by the state of Texas over illegal immigration.⁷

But Light makes the same mistake as Nowrasteh in treating illegals as fully identified by DHS at intake, even though DCJ will go on to identify more illegals who are initially placed in the DPS “other/unknown” category.

Unlike Nowrasteh, Light then relies on unverified claims made by arrestees about their citizenship and place of birth to both supplement the “legal” arrest category and create a “native-born” category. Not appreciating that arrestees of any claimed status could turn out to be DCJ-identified illegal immigrants, Light inadvertently places some illegals in his “legal” or even “native-born” categories. This is perhaps why he reaches the doubtful conclusion (opposite to Cato’s) that legal immigrants commit crimes at higher rates than illegal immigrants.⁸

Conclusion
Studies purporting to show low illegal immigrant crime rates in Texas fail to account for the fact that illegal immigrants are not always identified immediately upon arrest. In many cases, illegal immigrants are identified only after they are imprisoned. Given sufficient time for data collection, it appears that illegal immigrants have above average conviction rates for homicide and sexual assault, while they have lower rates for robbery and drugs. Significant uncertainty persists, however, as to how many illegals may remain unidentified, especially those who committed lesser offenses requiring little or no prison time. While strong claims about the overall criminality of illegal immigrants are not possible with the current data, prior research has understated it substantially.

¹ We learned about this process through discussions with personnel at the Crime Records Division within DPS.

² For the total illegal immigrant population in Texas each year through 2019, we rely on estimates from the Center for Migration Studies (CMS). DHS and Pew have not yet published estimates for 2019, but DHS counts are usually slightly higher than CMS and Pew, resulting in a slightly lower illegal immigrant crime rate. For example, in 2012 the illegal immigrant homicide rate per 100,000 was 3.9 with both CMS and Pew data, but 3.6 with DHS data. Occasionally, one hears dubious claims that the illegal immigrant population is far greater than the estimates made by any of these organizations. If true, the illegal immigrant crime rate would be far lower.

³ Alex Nowrasteh, “Criminal Immigrants in Texas: Illegal Immigrant Conviction and Arrest Rates for Homicide, Sex Crimes, Larceny, and Other Crimes”, Cato Institute, February 26, 2018.

⁴ Alex Nowrasteh, “Criminal Immigrants in Texas in 2017: Illegal Immigrant Conviction and Arrest Rates for Homicide, Sex Crimes, Larceny, and Other Crimes”, Cato Institute, August 27, 2019; and Alex Nowrasteh, “Criminal Immigrants in Texas in 2019: Illegal Immigrant Conviction and Arrest Rates for Homicide, Sex Crimes, Larceny, and Other Crimes”, Cato Institute, May 11, 2021.

⁵ In both of these newer studies, Nowrasteh acknowledges that some illegal immigrants are not identified until they are in prison, but he suggests they could have been legal at the point of arrest before becoming illegal while in prison. DPS personnel with whom we spoke cast doubt on this suggestion. When DHS identifies arrestees as legal immigrants at intake, they will remain in the “legal” category in the DPS data even if DCJ were to later determine that they have become illegal.

⁶ Michael T. Light, Jingying He, and Jason P. Robey, “Comparing crime rates between undocumented immigrants, legal immigrants, and native-born US citizens in Texas”, Proceedings of the National Academy of Sciences, Vol. 117, No. 51 (December 2020).

⁷ Department of Homeland Security, “Significant Considerations in Developing Updated Guidelines for the Enforcement of Civil Immigration Law”, September 30, 2021, FN 46.

⁸ The conclusion is doubtful because the average legal immigrant is older, has higher socioeconomic status, and has undergone more vetting than the average illegal immigrant.

Sean Kennedy is a visiting fellow at the Maryland Public Policy Institute. Jason Richwine is a resident scholar and Steven A. Camarota is the director of research at the Center.This article is published courtesy of the Center for Immigration Studies (CIS).