CPRC in the Associated Press on background checks

From Matt Stroud at the Associated Press:

. . . John R. Lott, Jr., author of More Guns Less Crime, doesn’t think Hunt’s gun purchase indicates that gun regulations should change.

He said studies consistently show that background checks don’t reduce violent crime, but they can produce “false positive” results that block law-abiding citizens from purchasing a gun.

NICS data provided to the AP shows that about 1.25 percent of those whose names are submitted to NICS for background checks are initially denied the right to purchase a firearm. About half of those denials are later overturned after appeal.

The additional time it takes to identify law-abiding gun purchasers “can disarm law-abiding citizens and make them more vulnerable to violent crime,” Lott said.

But saying that half the denials are later overturned after appeal gives a misleading impression of the number of mistakes that were made by the NICS system.  Not everyone is willing to go through the process to make an appeal, which for most people would require hiring a lawyer.

Take the numbers for 2009. There were 71,010 initial denials. Of those, only 4,681, or 6.6 percent, were referred to the BATF field offices for further investigation. As a report on these denials by the U.S. Department of Justice indicates, “The remaining denials (66,329 – 93%) did not meet referral guidelines or were overturned after review by Brady Operations or after the FBI received additional information.”  They are making sure that the reason that the person has been flagged is for a reason that could prohibit them from getting a gun, that they have the right person, and that the reason that they have been flagged is actually correct.  To put it differently, the initial review didn’t find that these individuals had a record that prevented them from buying a gun.  (Numbers for 2010 are very similar and are available here.  The Obama administration has stopped releasing this data after 2010.)

Still that isn’t the end of the story. Of these 4,681 referrals, over 51 percent, or 2,390 cases, involve “delayed denials,” cases where a check hasn’t even been completed. Of the rest, 2,291 covered cases where initial reviews indicated that the person should have been denied buying a gun. But the government admits that upon further review another 572 of these referrals were found “not [to be] a prohibited person,” leaving about 4,154 cases. That implies an initial false positive rate of roughly 94.2%. And it still doesn’t mean that the government hasn’t made a mistake on the remaining cases. In some cases for example, a person’s criminal record was supposed to be expunged, and it had not been.

Of the cases referred to the BATF field offices there were still a number of false positives.  A 2004 sample found out that about 21 percent of these cases were found to be false positives (the percentage is slightly higher if a weighted sample is used).  A discussion of this second round of checks in the Washington Post breaks down the 4,732 referrals to the field office in 2010 this way: 10.1% “not a prohibited person,” 8.4% “no potential or unfounded,” 10% “closed by supervisor,” 35% “no prosecutorial merit,” and 23% “guidelines not met.”  Adding together the 10.1% “not a prohibited person” and 8.4% “no potential or unfounded” gives a total of 18.5% (which is pretty close to the 21 percent found in 2004) and seems to imply cases where mistakes were clearly made (either because they had the wrong person or they had misidentified something that the person has supposedly done).

Up until this point, no discretion about the merits of the case has entered the picture. If a review of the records indicates that someone is a prohibited individual, they are included. But of these 4,154 cases, only 140 cases involving banned individuals trying to purchase guns being referred to federal prosecutors, just 60 of which involved providing false information when buying a firearm. Of those 140 cases, federal prosecutors thought the evidence was strong enough to bring a case only 77 times.  As the Washington Post evaluation of the 2010 data notes that 35% of the cases show “no prosecutorial merit” these cases aren’t being dropped because of judgments over whether the case will be difficult to win.  They are being dropped because there is some mistake in the case.

The 2010 annual report (p. 5) makes clear that cases that merit prosecution in the second stage of review include, but are not limited to, “Cases involving restraining orders, domestic violence misdemeanors, non-immigrant aliens, violent felonies, warrants, and indictments.”  The only category not included in this list that can bar someone from owning a gun is non-violent felonies.

State prosecutions take up some of these 4,154, but the state level numbers on prosecutions and convictions are spotty and some of their denials come from some states’ own background check systems.  In Oregon, background check related arrests amount to just 0.04% of background checks.

Data for Colorado confirms the very low rate of prosecutions at that level.  The Coloradoan March 29, 2015:

Twenty months into implementation, . . . Three people have been convicted for failing to obey the new background check laws since a bill to expand the requirements went into effect in July 2013. . . .

From July 2013 through February 2015, the state conducted more than 512,028 background checks for gun transfers made at licensed gun dealers’ stores, at gun shows or between private parties. About 24,000 background checks were conducted on private transfers during that time, about 5 percent of the state’s total. . . .

Supporters of the background check law point to Colorado Bureau of Investigation data showing that in the first year of the law, there were 68 denials for private sales after background checks were performed at gun shows and another 98 denials for private purchases outside gun shows.

The denials were for people with previous charges that included sex offenses, restraining orders and assaults. From July 2013 through February 2015 in Colorado, there were 10,412 total denials from all background checks, which totaled 512,028, resulting in 2 percent of all background checks being denied. . . . .

UPDATED:
As if all the mistakes weren’t bad enough, from October 2015 to at least January 2016, the Obama administration has suspended processing thousands of appeals from buyers who were falsely denied being able to purchase guns.  From USA Today:

The surge of criminal background checks required of new gun purchasers has been so unrelenting in recent months that the FBI had been forced to temporarily halt the processing of thousands of appeals from prospective buyers whose firearm purchase attempts have been denied.

Since October, the bureau’s entire cadre of appeal examiners— about 70 analysts — was redeployed here to help keep pace with waves of incoming background investigations that continued through December when a record 3.3 million firearm sales were processed.

The transfer of examiners, which had left a backlog of 7,100 appeals, is only part of a makeshift reorganization that FBI Assistant Director Stephen Morris said has become necessary to handle a burgeoning workload that expands in the wake of every mass shooting and call for increased gun control that invariably prompt firearms sales binges across the country. . . .

UPDATED 2: Discussion of Department of Justice IG report titled “Audit of the Handling of Firearms Purchase Denials Through the National Instant Criminal Background Check System.”  By contrast to what was in the note above, the claim here is that there is an “accuracy rate that ranges from 99.3 percent to 99.8 percent.”

“We judgmentally selected 447 denied transactions and found that only 1 transaction was incorrectly denied, resulting in a 99.8 percent accuracy rate.” 

“Using the judgmentally selected 447 of the 373,900 FBI denied firearm applications (standard and delayed denials) for FY 2008 through FY 2012″