Tag: illegal immigrants

FAIR’s Confused Criticism of Our Immigration Crime Research

Spencer Raley at the Federation for American Immigration Reform (FAIR) recently wrote a criticism of a recent Cato brief that estimates illegal immigrant incarceration rates in the United States.  Much of Raley’s critique is perplexing as following his methodology advice would not only lead to an erroneous result but it would reduce illegal immigrant incarceration rates – which is the opposite result that he and his organization desire.  Raley’s points are quoted below, my responses follow.

The authors rely on faulty, voluntary data from the Census Bureau’s American Community Survey (ACS).  Even mainstream organizations like Pew Research acknowledge that many illegal aliens are slow to volunteer information about themselves to the federal government.  That’s why reputable research organizations assume a certain undercount when relying on ACS data.  Hesitation to self-report personal information is only increased when surveys include questions about criminal history. So, from the start, the primary source used in this study will yield an undercount of incarcerated illegal aliens because it relies on self-reported data.

The responses of prisoners are recorded by Census officials who interview a sample under the supervision of prison officials who also supply information like immigration status and country of birth.  Although it’s easy for people outside of prison to avoid a Census official, it’s quite difficult for a prisoner to do so if he or she has been selected for an interview by the ACS.  Since the ACS doesn’t ask about the respondent’s criminal histories, Raley’s criticism here is perplexing.  If anything, using the ACS would yield an undercount of the illegal immigrant population – which would increase the illegal immigrant incarceration rate in our brief. 

They also misstate illegal alien crime data from Texas. The authors sliced and diced data from Texas’ Department of Public Safety, claiming that the original data offered by the state was far too high, and that illegal aliens in Texas are half as likely to be incarcerated as U.S. citizens. The real numbers, however, tell a different story. Based on data compiled between June 2011 and February 2019, 25,000 illegal aliens are booked into Texas state and local jails annually, on average.

Raley makes several errors in summarizing my Texas crime research.  First, I didn’t “slice and dice” any data from Texas.  I took the numbers released by the Texas Department of Public Safety, divided them by the relevant subpopulation of Texas in 2015, and then multiplied the result by 100,000 to get a criminal conviction rate.  Second, I didn’t compare illegal immigrants to U.S. citizens.  I compared illegal immigrants to native-born Americans and legal immigrants separately.  Third, my Texas study did not analyze incarceration rates.  My Texas study looked at criminal conviction rates.  Incarceration rates and criminal conviction rates are different. 

Raley’s other criticisms are answered by reading the methodology section of our brief.  This is the most relevant section here which explains why we looked at the 18-54 population:

Another limitation of the ACS data is that not all inmates in group quarters are in correctional facilities. Although most inmates in the public-use microdata version of the ACS are in correctional facilities, the data also include those in mental health and elderly care institutions, as well as those in institutions for people with disabilities. These inclusions add ambiguity to our findings about the illegal immigrant population but not about the immigrant population as a whole, because the ACS releases macrodemographic snapshots of inmates in correctional facilities, which allows us to check our work.

The ambiguity in illegal immigrant incarceration rates mentioned above prompted us to narrow the age range to those who are ages 18-54. This age range excludes most inmates in mental health and retirement facilities. Few prisoners are under age 18, many in mental health facilities are juveniles, and many of those over age 54 are in elderly care institutions. Additionally, few illegal immigrants are elderly, whereas those in elderly care institutions are typically over age 54. As a result, narrowing the age range does not exclude many individuals from our analysis. We are more confident that our methods do not cut out many prisoners because winnowing the 18-54 age range reduces their numbers to about 4.5 percent above that of the ACS snapshot. Natives in our results include both those born in the United States and those born abroad to American parents.

 

Illegal Immigrants and Crime – Assessing the Evidence

Whether illegal immigrants bring a significant amount of crime to the United States is one of the most important questions to answer in the debate over immigration policy.  President Trump also seems to think so as he launched his campaign in 2015 with the now infamous quote: “[Mexican illegal immigrants] are bringing drugs.  They’re bringing crime.  They’re rapists.  And some, I assume, are good people.”  From executive orders to major talking points to the President’s speeches, which Vox reporter Dara Lind has aptly described as “immigrants are coming over the border to kill you,” Trump is interested in this important topic.  

It is difficult to know whether illegal immigrants are more likely to commit crimes than native-born Americans are.  All immigrants have a lower criminal incarceration rate and there are lower crime rates in the neighborhoods where they live, according to the near-unanimous findings of the peer-reviewed evidence.  Since 1911, large nationwide federal immigration commissions have asked whether immigrants are more crime-prone than native-born Americans and each one of them answered no, even when the rest of their reports unjustifiably blamed immigrants for virtually every problem in the United States.  From the 1911 Immigration Commission, also known as the Dillingham Commission, to the 1931 Wickersham Commission, and 1994’s Barbara Jordan Commission, each has reported that immigrants are less likely to commit crimes than native-born Americans. 

That research combines legal and illegal immigrants to calculate a crime rate for all immigrants, but the modern debate is over the crime rates of illegal immigrants.  Most people seem to accept that legal immigrants have lower crime rates than natives.  Measuring illegal immigrant crime rates is challenging for several reasons.  First, the American Community Survey does not ask which inmates in adult correctional facilities are illegal immigrants.  Second, federal data on the number of illegal immigrants incarcerated on the state and local level is recorded through the State Criminal Alien Assistance Program (SCAAP), which is a combination of stocks and flows that is incomparable to any other measure of inmates.  Third, 49 states do not record the immigration statuses of those in prison or convicted.  Until recently, these data limitations allowed pundits to say anything about illegal immigrant crime without fear of being fact-checked. 

Cato scholars have since published numerous Immigration Research and Policy Briefs to shed light on this topic.  Michelangelo Landgrave, a doctoral student in political science at the University of California, Riverside, and I released a paper today that estimates that illegal immigrant incarceration rates are about half those of native-born Americans in 2017.  In the same year, legal immigrant incarceration rates are then again half those of illegal immigrants.  Those results are similar to what Landgrave and I published for the years 2014 and 2016.  We estimated illegal immigrant incarceration rates by using the same residual method that demographers use to estimate the number of illegal immigrants in the United States, only we also applied that method to the prison population.  We used the same method to also find that the incarceration rate for young illegal immigrants brought here as children and theoretically eligible for deferred action is slightly below those of native-born Americans.

The second strand of research from Cato looks at criminal conviction rates by immigration status in the state of Texas.  Unlike every other state, Texas keeps track of the immigration statuses of convicted criminals and the crimes that they committed.  Texas is a wonderful state to study because it borders Mexico, has a large illegal immigrant population, is a politically conservative state governed by Republicans, had no jurisdictions that limited its cooperation with federal immigration enforcement in 2015, and it has a law and order reputation for strictly enforcing criminal laws.  If anything, Texas is more serious about enforcing laws against illegal immigrant criminals than other states.  But even here, illegal immigrant conviction rates are about half those of native-born Americans – without any controls for age, education, ethnicity, or any other characteristic.  The illegal immigrant conviction rates for homicide, larceny, and sex crimes are also below those of native-born Americans.  The criminal conviction rates for legal immigrants are the lowest of all.

FAIR SCAAP Crime Report Has Many Serious Problems

The Federation for American Immigration Reform (FAIR) recently released a report on illegal immigrant incarceration rates that is poorly contrived and terribly executed.  FAIR uses the number of illegal immigrants counted under the State Criminal Alien Assistance Program (SCAAP), which compensates local and state governments for incarcerating some illegal immigrants, to calculate the illegal immigrant incarceration rate.  FAIR finds that illegal immigrants, as measured by the number of SCAAP aliens, are incarcerated at much higher rates than natives and legal immigrants according to their faulty methods. 

There are many problems with FAIR’s report. 

The first problem is that it doesn’t state what year or period of years that it is analyzing.  I can’t even work backward because their citations only indicate when they accessed the relevant data, not which annual data they accessed.  This means that it would take many dozens of hours of guesswork to recreate their work.

The second problem with using SCAAP data is that they are the oddest crime data.  The SCAAP aliens are a measure of the stock and flow of illegal immigrants into state and local correctional facilities over a year.  SCAAP is used to compensate local communities and states for the costs of incarcerating some illegal immigrants, it is not designed to be used to estimate illegal immigrant incarceration rates.  As a result, the normal operation of dividing the number of prisoners who are in SCAAP by the stock of all prisoners to calculate a rate is not useful here because the proper denominator is the stock of the prison population at the beginning of a year plus the number of people admitted throughout the rest of that year.  However, FAIR uses just the stock of the population in state prisons at the end of the year plus estimates from a private group for the number of people incarcerated in local facilities.  FAIR does not add the flow to the stock to estimate a proper denominator. 

As a result, there is no correct way to use SCAAP data to estimate incarceration rates because there is not a similar number published by the government that can serve as a denominator.  Using the Bureau of Justice Statistics (BJS), one can estimate a flow plus stock denominator for some years without eliminating the significant problem of overcounting.  Although I caution against drawing a lesson from this back of the envelope calculation, Stephen Dinan of the Washington Times piqued my curiosity and I could not resist.  This is the only way to use SCAAP data to estimate illegal immigrant incarceration rates in state prisons.  

My back of the envelope calculation uses only the state-level incarceration numbers and excludes the number of local SCAAP prisoners and people jailed on the local level.  This is because the BJS does not collect the local numbers and they are not available for many years.  The number of illegal immigrants nationwide comes from the Department of Homeland Security (DHS) while the number of non-illegal immigrants (legal immigrants and native-born Americans) comes from the American Community Survey.  I merely divided the number of SCAAP aliens nationwide who were incarcerated and who entered state correctional facilities by the number of illegal immigrants nationwide in each year.  I then compared that rate to the same rate for non-illegal immigrants who were in correctional institutions that year or that entered them.  Lastly, I presented the figure per 100,000 for each subpopulation because that is common in the crime literature.

Figure 1 shows the nationwide results: Illegal immigrants in the SCAAP program are incarcerated at a lower rate than non-illegal immigrants.  Just to reiterate, this is back-of-the-envelope calculation with significant problems.  It excludes all prisoners incarcerated on the federal and local levels.  There is a lot of double counting.  SCAAP does not count all illegal immigrants.  The non-illegal immigrant number is a combination of legal immigrants and native-born Americans, which makes the latter look more peaceful and the former more dangerous than they really are.  These figures underestimate the incarceration rates for illegal immigrants and non-illegal immigrants

 

Figure 1
Illegal Immigrants and Non-Illegal Immigrant SCAAP-Adjusted Incarceration Rates, Per 100,000

 

Sources: Bureau of Justice Statistics, American Community Survey, and the Department of Homeland Security.

Note: The per 100,000 is for each subpopulation. 

 

The SCAAP figures are not refined enough to use for estimating an illegal immigrant incarceration rate because there is not a good enough denominator available.  However, using the best data and methods available still shows that the nation-wide incarceration rate for SCAAP aliens is below that of non-illegal immigrants – which includes native-born Americans and legal immigrants.  FAIR’s report is a prime example of poor scholarship, but the authors are only partly to blame.  The lack of adequate crime and incarceration data make such desperate attempts tempting.  

Illegal Immigrant Conviction Rates Are Low, Even When Factoring in Recidivism

 Over the last two years, Cato has published three Immigration Research and Policy Briefs on illegal immigrant criminality.  In each one, we found that illegal immigrants have lower criminal conviction rates in the state of Texas and lower nationwide incarceration rates relative to native-born Americans.  Although nobody has criticized our methods or the data, we answer other criticisms that arise.

The best recent criticism is that illegal immigrant conviction rates are low because they are deported after they serve their sentences, which reduces their recidivism rates relative to native-born Americans who cannot be deported after being released from prison.  Thus, the illegal immigrant incarceration or conviction rates are lower than those of native-born Americans because it is more difficult for them to recidivate as they would have to enter the country illegally again to do so.  This has been a difficult criticism to address as data limitations are severe, but we attempted to do so after making some assumptions.  We focused on comparing first-time criminal conviction rates.

We estimate that native-born Texans had a first-time criminal conviction rate of 683 per 100,000 natives in 2016.  In the same year, we estimate that illegal immigrants had a first-time criminal conviction rate of 462 per 100,000 illegal immigrants – 32 percent below that of native-born Americans.  Thus, about 36 percent of the gap that we observed in criminal conviction rates between illegal immigrants and native-born Americans can be explained by lower illegal immigrant recidivism that is likely due to their deportation. 

This question could have been easily resolved by comparing the immigration statuses of first-time offenders.  Of course, such data do not exist.  Regardless, this is still an important question even if our estimate results from a back of the envelope estimate.  You can judge for yourself how we came to this estimate.  This is how we did it. 

First, we used the Arizona state prison data from 2016 for those admitted to state prison that year.  Of U.S. citizens sent to prison that year, 58 percent had previously been to prison at some point since 1984.  The subpopulation of deportable non-citizens, which includes illegal immigrants but is not limited to them, had a recidivism rate of 47 percent – below those of U.S. citizens, but not that much below. 

Police Clearance Rates Are Not Lower in States with Many Illegal Immigrants

Immigrant criminality and its impact on the United States is one of the most important issues in the public debate over immigration. In order to provide new insight into this topic, my coauthor Michelangelo Landgrave and I have attempted to estimate the illegal immigrant incarceration rate. I have also written a short paper on Texas criminal conviction rates by immigration status and crime based on data provided by the state of Texas. All three papers found that illegal immigrants were less likely to be convicted or incarcerated for crimes than native-born Americans.

My paper on illegal immigrant crime rates in Texas is based on data from the Texas Department of Public Safety (DPS) that I obtained through a Public Information Act request. The Texas DPS data separately show the number of convictions and arrests of illegal immigrants, legal immigrants, and native-born Americans for 44 and 46 different crimes, respectively, in the state of Texas by year from January 1, 2011, to November 15, 2017.

One of the persistent criticisms of my paper on Texas criminal conviction rates is that the DPS data do not record the number of illegal immigrants who commit crimes but are not convicted. Given data limitations, that is probably an impossible question to answer in a satisfactory way for immigrants and for natives.  However, I try to address this criticism in my Texas paper by showing that the gap between the arrest rates and conviction rates for illegal immigrants and the gap between the arrest rates and conviction rates for native-born Americans are similar, indicating that there are few illegal immigrants who are arrested for offenses who then disappear or are deported before their convictions relative to natives who are arrested and then not convicted.

A related criticism is that illegal immigrants flee Texas and then go back to their home countries after they commit crimes, which means that the Texas state conviction data would not count them. Thus, the criminal conviction rate for illegal immigrants is so low because they commit their crimes and flee – an option that few native-born American criminals possess. This argument makes a certain amount of sense in Texas as it shares a long border with Mexico, the source of a majority of illegal immigrants in Texas.

To answer that second criticism, we decided to investigate whether police clearance rates are correlated with the number of illegal immigrants on the state level. According to the FBI’s Uniform Crime Reporting (UCR) Program, law enforcement agencies can clear offenses by one of two means. The first is called “cleared by arrest” whereby a person must be arrested, charged with an offense, and turned over to a court for prosecution.  The second is called “cleared by exceptional means,” whereby the police must identify the offender, gather enough evidence for an arrest and to charge them with a crime, identify the offender’s exact location, and have encountered a circumstance out of law enforcement’s control that prevents an arrest. The death of the offender or the lack of an extradition treaty with the country harboring a suspected criminal are common causes of clearances by “exceptional means.” Mexico and the United States have an extradition treaty. An offense is cleared when the police have taken certain actions to solve the underlying crime short of a criminal conviction.

Landgrave ran many regressions between clearance rates (logged) and the proportion of the population of each state who were illegal immigrants (logged) with state-year and region-year fixed effects. The regressions control for demographic characteristics, the number of police officers for every 100,000 residents, education, and population density. He ran regressions for clearance rates by major crime and the entire crime index. All he found is that motor vehicle theft and burglary clearance rates are positively correlated with the proportion of the population who are illegal immigrants, but only at the 10 percent level for the state-year fixed effects (Table 1, click for larger version). There were no other statistically significant results.

Table 1: Correlation between State Police Clearance Rates and Illegal Immigrant Population

As a quick exercise to test this persistent criticism, these results reveal that there is no nationwide link between clearance rates and the proportion of the population who are illegal immigrants. The only exception is that police clear more motor vehicle and burglary offenses in states with more illegal immigrants as a proportion of their population, but only in one permutation and only at the 10 percent level. Although the theory that illegal immigrants commit crimes and then flee states seems plausible, we see no evidence of that in the aggregate clearance rates.

Peter Kirsanow’s Numerous Errors

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.

The White House’s Misleading & Error Ridden Narrative on Immigrants and Crime

President Trump recently held an event with some of the relatives of people killed by illegal immigrants in the United States. Afterward, the White House sent out a press release with some statistics to back up the President’s claims about the scale of illegal immigrant criminality.  The President’s claims are in quotes and my responses follow.

According to a 2011 government report, the arrests attached to the criminal alien population included an estimated 25,000 people for homicide.

Criminal aliens is defined as non-U.S. citizen foreigners, which includes legal immigrants who have not naturalized and illegal immigrants. The 25,064 homicide arrests he referred to occurred from August 1955 through April 2010 – a 55-year period.  During that time, there were about 934,000 homicides in the United States. As a side note, I had to estimate the number of homicides for 1955-1959 by working backward.  Assuming that those 25,064 arrested aliens actually were convicted of 25,064 homicides, then criminal aliens would have been responsible for 2.7 percent of all murders during that time period. During the same time, the average non-citizen resident population of the United States was about 4.6 percent per year. According to that simple back of the envelope calculation, non-citizen residents were underrepresented among murderers.

In Texas alone, within the last seven years, more than a quarter million criminal aliens have been arrested and charged with over 600,000 criminal offenses.  

We recently published a research brief examining the Texas data on criminal convictions and arrests by immigration status and crime. In 2015, Texas police made 815,689 arrests of native-born Americans, 37,776 arrests of illegal immigrants, and 20,323 arrests of legal immigrants. For every 100,000 people in each subgroup, there were 3,578 arrests of natives, 2,149 arrests of illegal immigrants, and 698 arrests of legal immigrants. The arrest rate for illegal immigrants was 40 percent below that of native-born Americans. The arrest rate for all immigrants and legal immigrants was 65 percent and 81 percent below that of native-born Americans, respectively. The homicide arrest rate for native-born Americans was about 5.4 per 100,000 natives, about 46 percent higher than the illegal immigrant homicide arrest rate of 3.7 per 100,000.  Related to this, the United States Citizenship and Immigration Services recently released data that showed the arrest rate for DACA recipients about 46 percent below that of the resident non-DACA population.

More important than arrests are convictions. Native-born Americans were convicted of 409,708 crimes, illegal immigrants were convicted of 15,803 crimes, and legal immigrants were convicted of 17,643 crimes in Texas in 2015. Thus, there were 1,797 criminal convictions of natives for every 100,000 natives, 899 criminal convictions of illegal immigrants for every 100,000 illegal immigrants, and 611 criminal convictions of legal immigrants for every 100,000 legal immigrants. As a percentage of their respective populations, there were 50 percent fewer criminal convictions of illegal immigrants than of native-born Americans in Texas in 2015. The criminal conviction rate for legal immigrants was about 85 percent below the native-born rate.

Figure 1: Criminal conviction rates by immigration status in Texas

Murder understandably garners the most attention. There were 785 total homicide convictions in Texas in 2015. Of those, native-born Americans were convicted of 709 homicides, illegal immigrants were convicted of 46 homicides, and legal immigrants were convicted of 30 homicides. The homicide conviction rate for native-born Americans was 3.1 per 100,000, 2.6 per 100,000 for illegal immigrants, and 1 per 100,000 for legal immigrants. In 2015, homicide conviction rates for illegal and legal immigrants were 16 percent and 67 percent below those of natives, respectively.

Figure 2: Homicide conviction rates by immigration status in Texas

Murderers should be punished severely no matter where they are from or what their immigration status is. There are murderers and criminals in any large population, including illegal immigrants. But we should not tolerate the peddling of misleading statistics without context. What matters is how dangerous these subpopulations are relative to each other so the government can allocate resources to prevent the greatest number of murders possible. Thus, enforcing immigration law more harshly is an ineffective way to punish a population that is less likely to murder or commit crimes than native-born Americans. Illegal immigrants, non-citizens, and legal immigrants are less likely to be incarcerated, convicted, or arrested for crimes than native-born Americans are. 

Pages