Obama administration people still pushing out junk science on gun issues

Nov 16, 2017 | Featured

A new very short report from a few people at the Harvard Business School claims that waiting periods reduce homicides, though they find that there is no consistent effect on suicides.  The report is titled “Handgun waiting periods reduce gun deaths,” and it was released by the National Academy of Sciences.  While neither the report nor the media coverage notes it, the report indicates that background checks are more likely to increase homicides and particularly suicides than to decrease them.

From the abstract: “We find that waiting periods reduce gun homicides by roughly 17%.  We provide further support for the causal impact of waiting periods on homicides by exploiting a natural experiment resulting from a federal laws in 1994 that imposed a temporary waiting period on a subset of states.”

Here are some comments on the paper.

— There are lots of papers that have looked at waiting periods and not found any benefit.  Yet, this paper somehow seems to be unable to mention those other papers or to bother explaining why it does things differently than those other papers.  The study even cites work by Dr. Lott, but while it mentions right-to-carry laws, it ignores that the very same research also deals extensively with waiting periods.

There are basically four reasons that their results are different from what others have found.

— They have coded the adoption and end dates of waiting periods and background checks quite differently than others.

— One of the most profound differences is that this paper assumes that all waiting periods are the same.  But waiting periods can vary from 2-days to 6 months, though during the period of this study a lot of states had the 5-day Brady Act waiting period.  Does it make sense to assume the impact of waiting periods is the same across states no matter how long the waiting period is?  Dr. Lott’s work has always tried to account for these differences.

— No explanation is offered for why the paper just looks at homicides and suicides for those over 21.  Implicitly they are assuming that gun control regulations are impacting homicide and suicide rates for those over 21 relative to those younger than that.  There is also the issue of why they look at homicides and not murders.  Homicides includes murders plus justifiable homicides.  Waiting periods could reduce the number of justifable homicides, but it isn’t obvious why anyone would consider that to be a good outcome.

— The control variables are quite different than other research.  For example, economists and criminologists try to explain crime rates (homicide) by accounting for policing and prison.  For suicide, suicide rates for other age groups or non-firearm suicide rates are included as control variables to account for factors that impact suicide generally.  In this case, since they seem to believe that the gun control regulations primarily effect those over 21, they could use suicide rates for those under 21 to help explain suicide for those over 21.

Bottom line.

— If one believes these results, they usually show that benefits from waiting periods (though not always statistically significant), but they also consistently show no benefits and frequently significant costs from background checks.

— In Table S6 where they account for all the gun control laws that they have in their regressions, while waiting periods reduce homicides, background checks tend to increase total suicides.  Over the 1970 to 2014 period, the number of lives reportedly saved from waiting periods reducing homicides, is almost exactly offset by the increase in total suicides from background checks.



  1. D. Martin

    I can see why background checks might increase homicides by keeping potential victims (e.g. those fleeing abusive relationships) disarmed. But why might background checks increase suicides?

    • johnrlott

      That is simply a result of their poorly constructed research.

      • mobius

        Shhh… You’re not supposed to mention the rigging.