Unintended Consequences

John Lott's book, More Guns, Less Crime, is often cited by pro-gun advocates as evidence supporting the right to carry concealed weapons for self-defense. More Guns, Less Crime is based on a study conducted by Lott and David Mustard entitled "Crime, Deterrence, and Right-to-Carry Concealed Handguns," which appeared in the January 1997 issue of The Journal of Legal Studies. The study maintains that arming citizens deters violent crime.

The Claims that Right-to-Carry Laws Reduce Violent Crime are Unsubstantiated, Daniel W. Webster, Johns Hopkins Center for Gun Policy and Research, March 1997. (Available at http://support.jhsph.edu/departments/gunpolicy/research.cfm; INTERNET.)

This paper criticizes the Lott and Mustard study for its methodological and factual flaws. The author notes that Lott and Mustard mistakenly categorize some state laws as shall-issue laws even though a reading of the laws reveals that there is still discretion allowed in granting a concealed carry license. Lott and Mustard's statistical models use arrest ratios (arrests per crime committed in a given year) to predict changes in crime rates�a method deemed inappropriate nearly two decades ago by a National Academy of Sciences panel of experts. As a result of such flawed methodology, Lott and Mustard's findings depart from well-established facts about crime. For example, Lott and Mustard deduce that criminals, in response to shall-issue laws, commit property crime to substitute for crimes which do not involve money or property. The author notes that "no credible criminologic theory can explain why a criminal would steal a car because he felt deterred from assaulting someone."

Concealed Handguns: The Counterfeit Deterrent, Franklin Zimring and Gordon Hawkins, The Responsive Community, Spring 1997, pp. 46-60.

This essay examines two types of flaws in Lott and Mustard's study: their comparisons between states and their assumptions characterizing shall-issue states. First, Zimring and Hawkins write, comparing state crime trends without sufficient controls is misleading because factors such as poverty, drugs, and gang activity vary significantly between right-to-carry states and other states. Second, the authors point out, after confirming that a shall-issue statute was passed, Lott and Mustard make no attempt to measure carrying of handguns by citizens, use of handguns by citizens in self-defense, or offender behavior in relation to street crime. The authors conclude that "what we know from this study about the effects of �shall carry' laws is, therefore, nothing at all."

Two Guns, Four Guns, Six Guns, More Guns: Does Arming the Public Reduce Crime?, Albert W. Alschuler, Valparaiso University Law Review, Spring 1997, pp. 365-373.

This paper summarizes other researchers' critiques of Lott and Mustard's study and raises new questions about their overall conclusions. For example, the author notes that, because gun possession in the home is lawful without concealed carry laws, the deterrent effect of concealed carry should be far greater for stranger homicides than for intra-family homicides. Yet Lott and Mustard report that the proportion of stranger killings increases following the enactment of concealed carry laws, while the proportion of intra-family killings declines. The author concludes, "At this point, there is essentially no reason for an intelligent consumer of social science research to accept the Lott and Mustard findings."

Flawed Gun Policy Research Could Endanger Public Safety, Daniel W. Webster, Jon S. Vernick, Jens Ludwig, and Kathleen J. Lester, American Journal of Public Health, June 1997, pp. 918-921.

This article examines methodological problems in Lott and Mustard's study and finds that several serious shortcomings render Lott and Mustard's conclusions insupportable. For example, relative to other demographic groups, young black males have very high rates of criminal offending and victimization and older black females have much lower rates of offending and victimization. Lott and Mustard's statistical methods are so badly flawed that their results indicate the proportion of young black males in a county's population was only weakly associated with higher rates of crime, while the proportion of black females older than 65 in a county's population had large positive effects on murder and auto theft, while reducing all other violent crimes. The authors conclude that "the flaws in Lott and Mustard's study of shall-issue laws are so substantial, and the findings so at odds with criminological theory and research, that any conclusions about the effects of shall-issue laws based on this study are dubious at best."

Do Right-to-Carry Laws Deter Violent Crime?, Dan A. Black and Daniel S. Nagin, The Journal of Legal Studies, January 1998, pp. 209-219.

This paper demonstrates that Lott and Mustard's results are highly sensitive to small changes in their model and sample. As a result, the data provide no basis for confident conclusions about the impact of right-to-carry laws on violent crime. For example, once Florida is removed from the sample, there is no longer any detectable impact of right-to-carry laws on the rates of murder and rape�the two crimes that Lott and Mustard use to account for 80 percent of the alleged social benefit of right-to-carry laws. The authors conclude that "inference based on the Lott and Mustard model is inappropriate, and their results cannot be used responsibly to formulate public policy."

Concealed Gun Carrying Laws and Violent Crime�Evidence from State Panel Data, Jens Ludwig, International Review of Law and Economics, September 1998, pp. 239-254.

This paper critiques the Lott and Mustard study by focusing on the minimum age requirements for concealed carry licenses. Because only adults can obtain concealed carry licenses, the author hypothesizes that any deterrent benefits of concealed carry laws should be concentrated among adults and should also be reflected in the difference between adult and juvenile victimization rates. However, the author's results find just the opposite�undermining Lott's argument that concealed carry laws have reduced the rate of homicide. In fact, Ludwig writes, his re-analysis of the Lott and Mustard data suggests that "shall-issue [concealed carry] laws have resulted, if anything, in an increase in adult homicide rates."

Book Review of More Guns, Less Crime, David Hemenway, New England Journal of Medicine (NEJM), December 31, 1998, pp. 2029-2030. Follow-up correspondence: NEJM, May 20, 1999, pp. 1599-1600.

This article reviews the Lott book, More Guns, Less Crime, focusing on the core of the book, a large statistical study of state "right-to-carry" laws. The author points out that crime moves in waves, but Lott�s analysis does not include variables�such as gangs, drugs, or community policing�that can explain these cycles. This omission leads to illogical results. For example, according to Lott's results, having fewer older black women in a sample will lead to a more dramatic reduction in homicide rates than increasing arrest rates or enacting shall-issue laws. Additionally, some of the data used in the analysis is simply wrong. The reviewer concludes that while Lott deserves high marks for attempting to study a difficult issue, "he deserves failing marks for pressing policy makers to use his results despite the substantial questions that have been raised about his research."

Myths about Defensive Gun Use and Permissive Gun Carry Laws, Daniel Webster and Jens Ludwig, paper presented at the "Strengthening the Public Health Debate on Handguns, Crime, and Safety" meeting, Chicago, IL, October 1999. (Available at http://support.jhsph.edu/ departments/gunpolicy/research.cfm; INTERNET.)

This paper focuses on problems with the Lott and Mustard study and the authors' interpretations of the findings. Aside from the errors made in the study and analysis, the fundamental problem with Lott's research is that "correlation is not causation." Variables may be related to one another, yet not cause one another. The authors provide the example that, while there is a significant association between a child's shoe size and the child's writing ability, this correlation does not prove that large shoes improve writing ability. Thus the difference in crime rates between Florida and California may mistakenly be attributed to the presence of a permissive concealed carry law in the former, when all or part of the difference may be due to other unmeasured differences, including poverty, drugs, gang activity and police resources, for which Lott does not adequately control. The authors conclude that "the many limitations of Lott's...research indicate that there is no reason to move from [the] view of guns and violence backed by research in previous decades."

Back to Table of Contents