As several of Donald Trump’s Republican opponents have pointed out, his fiery statements on any given subject strike a chord among a set of voters who feel the political system has failed them. Tom Tancredo, a former GOP politician turned pundit, credits Trump for "forcing all 17 Republican candidates to talk about the social costs of illegal immigration."
In an Aug. 8, 2015, article for the conservative website Breitbart, Tancredo said the "mainstream media – including, sadly, major segments of the presumably conservative media, like the Wall Street Journal — are working overtime to keep the American public and the American voters in the dark on the scope of illegal alien crime."
Tancredo provided these bullet points:
"Between 2008 and 2014, 40 percent of all murder convictions in Florida were criminal aliens. In New York it was 34 percent and Arizona 17.8 percent."
"During those years, criminal aliens accounted for 38 percent of all murder convictions in the five states of California, Texas, Arizona, Florida and New York, while illegal aliens constitute only 5.6 percent of the total population in those states."
Tancredo said his source was a presentation from a conservative think tank, the Center for Security Policy.
Several readers asked us to look into Tancredo’s core statement that between 2008 and 2014, "criminal aliens accounted for 38 percent of all murder convictions in the five states of California, Texas, Arizona, Florida and New York."
There are serious glitches in his claim. He botches the takeaway of the presentation, which has plenty of issues on its own.
A botched summary
We tried a variety of ways to reach Tancredo and did not hear back.
In his column, he embedded a video of the presentation by the Center for Security Policy and invited readers to watch the whole thing.
We did. The first thing that jumped out was Tancredo mangled the dates.
He said "criminal aliens" accounted for 38 percent of murder convictions in five states between 2008 and 2014.
In fact, the presentation offered numbers for 2005 to 2008.
That's not the only issue. The presentation’s author, James Simpson, told us he had emailed Breitbart about Tancredo’s use of his presentation. "(Tancredo) quoted the whole thing incorrectly," Simpson told PunditFact.
Our research found that even if Tancredo had quoted the presentation as it was given, there would still be plenty of concern about its accuracy.
State data not prevalent
Simpson’s report includes findings from two separate sources: an article that describes 2008-14 data from the Texas Department of Public Safety, and his digest of a 2011 Government Accountability Office report, which uses 2005-08 numbers.
We got into the weeds of each source, and the fact is neither provides an accurate look at the percentage of murders committed by criminal aliens in those five states.
The real figure may be impossible to know; Texas appears to be the only one of the five states that actually keeps track of convictions of criminal aliens. The "criminal aliens" label applies to noncitizens who have either legal or illegal immigration status. (It is incorrect to consider all of them as illegal immigrants.)
With various estimates floating around in Tancredo’s article, it’s useful to start with the solid numbers out of Texas.
The Texas Department of Public Safety continuously updates its tally of criminal aliens booked into local jails, tracking the charges filed and whether the person was convicted.
The agency’s latest report covers June 1, 2011, to July 31, 2015. In that time, 344 noncitizens were convicted of homicide. In about the same period, Texas had 4,571 murders. (There’s a difference between calendar and fiscal years, but as of this writing, the differences balance out.) So based on counts of actual cases, criminal aliens account for 7.5 percent of all homicides in Texas.
That figure is striking because it is one-fifth as large as the number Simpson gave in his presentation. Simpson said, "Illegal aliens have committed 35 percent of all murders in Texas since 2008."
Simpson told us he had not seen the official Texas report. He had relied solely on an article for PJ Media.
In our view, official hard data beats something in an article.
Murky federal data
The numbers from the U.S. Government Accountability Office, a widely respected source, are also imprecise — which is obvious from all of the report’s cautionary notes.
In 2011, the GAO aimed to learn the cost of keeping criminal aliens behind bars. The government analysts wrote that due to the large number of arrest files, they randomly selected 1,000 criminal aliens in five states — Arizona, California, Florida, New York and Texas — and took a close look at their arrest records.
The researchers detailed a host of caveats: At the state level, it’s possible that some people are counted twice if, for example, they are first kept at a county jail and are then transferred to a state prison. And the margin of error for the overall tally of homicides, as well as other crimes, was +/- 20 percent.
The GAO was clear about the lack of precision in its results. A close look shows why anyone should use them with great caution.
Because this was a study of costs, the only state prisoners examined fell under a federal reimbursement program called the State Criminal Alien Assistance Program (SCAAP). Working with broad estimates, the GAO gave the percentage of convicted murders in each state’s population under that program. We used those percentages to come up with the number of criminal alien murderers in each state (see our chart).
Without wading too deep into the weeds, we found that our analysis of the publicly available data wasn’t matching up with Simpson’s. Simpson’s total of murders is about 7,085, and our estimate was much lower at 5,300-5,400. Further, in Florida, he reported three times as many killers as our estimate, and his percentage of all homicides due to criminal immigrants in Texas was about double the official number (albeit for a different time period).
We should note that Simpson’s numbers are the basis for Tancredo’s claim that criminal aliens accounted for 38 percent of all murder convictions in the five states of California, Texas, Arizona, Florida and New York.
When we asked Simpson about his figures, he said after several months of writing to the GAO, he got the actual raw numbers.
But 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. And Simpson told us "the whole thing is difficult to understand frankly, and they couldn’t explain it very well over the phone either. I may have to go back to them for clarification."
Also, Simpson compared the number of immigrants convicted of murder to the total number of murders. Tancredo said the percentages referred to all murder convictions. Since many murders go unsolved, that statement is clearly incorrect. This error underscores another way that Tancredo misquoted Simpson's work.
We do have the number of people arrested for homicide in Texas 2011-2014 (a shorter period than the one for the number of criminal aliens convicted of homicide). Convicted aliens represent 12 percent of all homicide arrests, but it's important to note that due to the mismatched time periods, that figure is too high.
The bottom line is this: Even the man who generated the numbers, which Tancredo then misquoted, expresses uncertainty about their precision.
Tancredo said that between 2008 and 2014, over one-third of the murder convictions in Arizona, California, Florida, New York and Texas were committed by illegal immigrants. The man who presented the data Tancredo cites said Tancredo "quoted the whole thing incorrectly."
Tancredo used the wrong time period. He thought the baseline number was homicide convictions when it was actually all homicides. Most important of all, he took the presentation he relied on at face value and ignored the hard numbers available from the Texas Department of Public Safety.
And while Tancredo might not have known it, the researcher whose work he used himself has questions about the underlying data he used. We have hard data from Texas that refutes the big and estimated numbers Tancredo used. Undocumented immigrants do commit murder, but perhaps only one-fifth as often as Tancredo said.
We rate the claim False.
Update: We added the Texas homicide arrest data after we first published.