Strange Empirical Claims About Mass Shootings and Stock Prices

I have a few "alerts" set up on Google Scholar, with one of them set to notify me whenever something containing the phrase "Second Amendment" is published.  The system has a few kinks to work out (my "Second Amendment" alert emailed me the link to this robotics article this evening...) but the system is an interesting way to stumble across things I might otherwise miss.

This evening's alert clued me in on an article by Benjamin W. Cross and Stephen W. Pruitt entitled Dark Knights Rising: The Aurora Theater and Newtown School Massacres and Shareholder Wealth.  The Article appears in Volume 41 of the Journal of Criminal Justice.  Here is the abstract:

Purpose 

This study analyzes the stock price impact of the Aurora theater and Newtown (Sandy Hook) school massacres on both domestic (US) and foreign theater operators and US gun manufacturers in an effort to document the economic effects of these tragedies. 

Methods 

The well-established “event study” methodology from the fields of economics and finance is employed to assess the impact of the shootings on the affected companies after controlling for risk and overall market movements. 

Results 

The Aurora theater shooting resulted in striking declines for Cinemark (the targeted theater) as well as major US competitors, but had no impact on overseas theater chains. Smith & Wesson (maker of the gun used in Aurora) showed no response, whereas Ruger (a competitor) exhibited large gains. Both Smith & Wesson and Ruger plunged after the Newtown shooting, although neither made the weapons used in the shooting. 

Conclusions 

Contrary to prior research on workplace homicides, the results show that random mass shootings have profound effects on targeted companies. In addition, the results suggest the presence of a very strong “contagion effect” (where negative events affecting one company impact others in the same industry). The negative responses of both publicly-traded US firearms manufacturers to the Newtown shooting suggests a “sea-change” in the debate over gun ownership in the US.

Normally I would write off an article like this as involving subject matter that is typically outside of my area of expertise -- but the abstract's last claim about a "sea-change" in the gun ownership debate caught my eye.  I decided to read through the article to see what the authors had found.

As it turns out, the article is blatantly flawed and highly misleading.


The abstract's "Conclusions" sections makes a number of claims to draw in readers.  The authors note that the study indicates that random mass shootings have "profound effects" on companies, and that this is "[c]ontrary to prior research on workplace homicides."

The prior research that the authors primarily cite as an example of contrary findings is this article by Michael J. McNamara and Stephen W. Pruitt, entitled, "There's a Guy in the Center Aisle With a Gun!"--Workplace Homicides and Shareholder Wealth.  Note that Pruitt is a coauthor in both this article and the Dark Knights article.

Cross and Pruitt evaluate the case of the Aurora Theater shooting and note that it had a significant, negative impact on the stock prices for Cinemark, the theater company that owned the Aurora Theater, as well as for several other movie theater companies.  They then note:

Cinemark’s strikingly negative response to the Aurora shooting might be viscerally expected, but is actually somewhat puzzling given that it is completely counter to the null response of stock prices to “random” workplace homicides previously documented by McNamara and Pruitt (2006). To reiterate, McNamara and Pruitt found no stock price responses whatsoever to workplace homicides perpetrated by individuals with no prior employment relationship with the targeted company (as was the case with Cinemark and James Eagan Holmes). Conversely, Cinemark’s losses were almost identical in magnitude to the authors’ findings of the adjusted losses of up to 7% in share values for homicides committed by employees or former employees of the targeted firms—events, it must be emphasized, with a clear line of liability running straight to the affected firms.

There are several big problems here.

First, and most obviously, the McNamara and Pruitt study involved a sample of 40 workplace shootings, all but one of which resulted in less than ten deaths.  The Cross and Pruitt study, on the other hand, had a sample size of one shooting, which resulted in 12 deaths, 58 injuries, and which coincided with the massively publicized premier of the film, The Dark Knight Rises.  It is already sketchy that Cross and Pruitt imply that their results are operating on the same level as another study with a sample size that is 40 times larger -- and the sheer difference in the character of the shootings addressed in both studies amplifies the flaw.  The Aurora Shooting was far more likely to gain attention of the wider public due to its relationship to a popular movie release and high death toll, and comparing it to the McNamara and Pruitt study's sample is downright misleading.

Second, at the risk of sounding slightly repetitive, it's odd that Cross and Pruitt are making claims about general industry dynamics based on the stock fluctuations in the theater industry resulting from a single shooting incident.

Third, the Aurora shooting's occurrence at a movie theater is likely to perpetuate fear among prospective moviegoers and thereby harm the revenues in the movie industry.  Compare this to several of the companies involved in the McNamara and Pruitt study -- companies that include Xerox, Lockheed Martin, Edgewater Technology, and Ford, to name a few.  While widely publicized news of a shooting at a theater may make people more fearful of going to a theater, similarly widely publicized news of a shooting at Xerox or Ford is far less likely to make people afraid of purchasing copying machines or vehicles.  The nature of the businesses and the consumer fear that is unique to the Aurora scenario makes the Aurora shooting and its impact on the theater business fundamentally different from many of the McNamara and Pruitt sample shootings.

There are other problematic areas of the article, but I only have the space here to focus on one other issue -- the Second Amendment "sea change" that initially caught my eye.

Cross and Pruitt note that after the Aurora shooting, the stock prices of Ruger and Smith & Wesson increased -- likely due to a surge in purchases due to fear that stricter gun control laws would be enacted.  After the Newtown shootings, however, the stock prices of these companies fell, despite the fact that the Newtown shooter did not use guns from either company. Cross and Pruitt note:

Clearly, the extraordinarily negative price responses of Smith & Wesson and Ruger following the Newtown massacre was the polar opposite of the zero-to-positive responses registered by the two companies following the Aurora theater shooting and suggests strong support for the hypothesis that investors did, indeed, believe—at least initially—that the national horror following the Newtown shooting could represent a significant “sea change” in the national debate over Second Amendment rights.

This is actually a pretty decent conclusion.  Granted, the sample size is only one, so drawing broad conclusions may not be the best strategy, but clarifying that the "sea change" is one that was subjectively perceived by the gun industry at the time?  That could be sensible.

But then there's the article's conclusion...

However, the prices of both Smith & Wesson and Ruger declined dramatically following the Newtown school massacre, despite the fact that neither company’s products were used by the killer. . . . Indeed, it is suggestive of a considerable “sea change” in the debate concerning gun ownership and weapons sales in the United States.

The subjective perception of a "sea change" has suddenly become an objective reality.  The authors take their potentially reasonable conclusion and make it into a hyperbolic, empirically denied assertion.  Newtown got everybody's attention, and it still serves as a rallying cry for gun control advocates.  But it is downright naive to claim that there has been a "sea change" in the debate over gun control based on shootings' impacts on the stock of gun manufacturers.

My first impression of Cross and Pruitt's article was that it said something unique and made some intriguing claims.  It is a shame that these claims are based on such sloppy analysis.