Problems with GAO Illegal Alien Crime Stats

 

On August 8, 2015 Breitbart published an article by Tom Tancredo describing illegal alien homicide statistics. This article borrowed heavily from a PowerPoint presentation I gave in July. The YouTube video of my presentation was embedded in the article. Politifact challenged both Tancredo and my statistics and rated our assertions as “false.” While Politifact’s analysis was fraught with errors of its own, its conclusions were correct. Both Tancredo and I made errors. This article is presented to explain and correct those errors.

Tancredo misquoted my statistics as covering the years 2008 – 2014, when I had said they cover the years 2005 – 2008, implying that the numbers were much more current. But I made a number of mistakes in my analysis that exaggerated the illegal alien homicide rate, noting for example that 38% of all murders in five states between 2005 an 2008 were committed by illegals. Tancredo reiterated those statistics, and Sean Hannity repeated the erroneous 38 percent number on his radio show. Media Matters called Hannity on it, citing the Politifact article.

My analysis was based on a 2011 Government Accountability Office report, Criminal Alien Statistics. In particular, GAO had provided illegal alien homicide and other crime convictions as a percentage of total illegal alien convictions in five states: New York, -Florida, Texas, Arizona and California. I filed a FOIA for the raw data used to calculate those percentages. (It was not included in the published report.) Following are the numbers GAO provided. These data measure all illegal aliens convicted of homicide in the five states that were incarcerated in those states in 2008. They are not estimates but actual counts:

 

Arizona

326
California 2,895
Florida 1,762
New York 1,168
Texas 934
Total 7,085

 

Politifact erroneously tallied this number at 5,300-5,400, applying percentages provided in charts, to rough incarcerated population estimates, not the raw data I obtained from GAO.

There were a total of only 835 murders committed in New York in 2008. Checking other states, the results were similar. In California for example, there were a total of 2,143 murders. Prior years had similar numbers. How could there be fewer murders committed in a given year than illegals convicted of murder? A closer reading of the report found that GAO reviewed conviction data for those five states, “from fiscal years 2005 through 2008.”

I assumed from GAO’s explanation that the above conviction figures must be totals for those four years. Comparing that with the total number of murders committed in those states over those four years brought the following results:

 

  Total   Alien %
  Murders   Convictions Total
Arizona          1,832     326 17.8%
California 9,392   2,895 30.8%
Florida 4,382   1,762 40.2%
New York 3,432   1,168 34.0%
Texas 5,581         934 16.7%
Total 24,614   7,085 28.8%

 

Pretty stunning results, no? Definitely worthy of a headline. However, I made three mistakes:

  1. In my original presentation, an erroneous total murder number of 18,643 made it into the publicized table instead of the correct 24,614 number. Don’t know how that happened, but as a result, the average percentage of total murders committed by illegals in those five states was miscalculated at 38.0% instead of the 28.8% shown above.
  2. The data provided by GAO was for illegal aliens incarcerated in 2008 for homicide, not 2005 – 2008 homicide convictions. In my defense I will say that GAO did not make this at all clear. The Politifact rebuttal to my article made the same misinterpretation.
  3. Most importantly, homicide and murder is not the same thing. Did you know that? Homicide is a much broader category, including manslaughter and vehicular homicide.

The correct comparison then, is between illegal aliens incarcerated for homicide in 2008, and all prisoners incarcerated that year for homicide.[*] The following table presents these corrected results.

 

Homicide Incarcerations – 2008
         
    Illegal % Illegal %
  Total Aliens Total Total Pop
New York1   10,435     1,168 11.2% 2.9%
California2   27,967     2,895 10.4% 7.3%
Texas3   16,012       934 5.8% 6.8%
Arizona4     3,494       326 9.3% 5.2%
Florida   12,296     1,762 14.3% 3.7%
Total   70,204     7,085 10.1% 5.6%

 

Sources: State incarceration databases  
GAO Unpublished statistics SCAAP Illegal Aliens

1 As of January 1. 2009

2 As of December 31, 2008

3 As of August 31, 2008

4 As of 9/1/2009.    

 

Separately, the GAO report also analyzed 249,000 arrest records of criminal aliens (their legal status was not specified), incarcerated in state and local prisons throughout the country spanning the years 2004 – 2008 and in federal prisons at the end of 2008. It chose a random sample of 1,000 from this pool of records and applied the results to the entire population, estimating that 25,064 criminal aliens had been arrested for homicide.

I made the same mistake with these data, comparing the 25,064 illegals imprisoned for homicide with total murders committed between 2004 and 2008 (83,289), concluding that 30% of murders during that period were committed by illegals. Once again, the more accurate comparison is the total number of people imprisoned for homicide in the relevant period (about 185,000). Using this figure, aliens represent about 14% of the total. However, this only holds if the aliens are imprisoned on the homicide charge. The population studied had been arrested an average of 7 times, and GAO did not specify for which crime they were incarcerated.

In criticizing my analysis, Politifact committed some of the same mistakes I did, misreading the GAO report and calculating incorrect numbers as a result. Politifact compounded its mistake by conflating the two separate analyses of criminal alien statistics described above, claiming “Simpson’s figures must be based on some statistical assumptions, not raw data, because his total is larger than the number of individual cases examined by the GAO.”  They were erroneously referring to the 1,000 samples taken from the nationwide analysis, rather than the separate analysis of state conviction data. To try to fully explain their mistakes would take another article. But their point has been taken. I made mistakes in the initial analysis and accept full responsibility for them. The original claims in my analysis were indeed “false.”

Politifact also challenged Texas homicide statistics included in my analysis and Tancredo’s article, referring to public documents suggesting the rates were much lower than claimed in the article. The source for those statistics was an unpublished Texas Public Safety Department study. I did not see the study but the reporter that did is a well-regarded former Justice Department attorney. He could be mistaken also, but without seeing the report it is impossible to draw conclusions. I will withhold judgment until that report becomes available.

In conclusion I will say that while the corrected numbers described in this article are not quite as earth shattering as those originally presented, they still show that illegal aliens have committed a ghastly number of homicides at rates out of proportion to their populations. The GAO study claimed that the five states studied held approximately 70% of the criminal alien population in 2008. If their homicide rates can be extrapolated to the rest of the country, then approximately 10,000 illegals were incarcerated for homicide nationwide in 2008. The conclusion I have drawn still holds: not one of those 10,000 people needed to die. Furthermore, that reflects only those who were caught and convicted. The actual number of homicides committed by illegals is likely much higher. Those who promote open borders, amnesty and sanctuary cities, as well as their press supporters, are accessories to mass murder.

[*] The illegal alien statistics cover inmates incarcerated for homicide between July 1, 2007 and June 30, 2008, while statistics for total homicide incarcerations measured slightly different dates, as indicated in the table footnotes. These were the closest comparable data available.