The debate

May 24: The effect is clear, by John Lott

May 23: Data distortion, by Robert Ehrlich

May 22: Less gun control means less crime, by John Lott

May 21: More guns means more guns, by Robert Ehrlich

E-mail us your comments (we may post them)

Reason's gun page

To Prof. Ehrlich, the "basic flaw" in my statistical analysis is that concealed handgun laws are likely to just be accidentally related to changes in the crime rate. He takes a simple example of explaining how the stock market changes over time. Obvious variables to include would be the interest rate and the expected growth in the economy, but many other variables -- many of dubious importance -- could possibly also be included. The problem arises when such variables are correlated to changes in stock prices merely by chance. An extreme case would be including the prices of various grocery store products. A store might sell thousands of items, and one -- say the price of peanut butter -- might happen to be highly correlated with the stock prices over the particular period examined. We know that there is little theoretical reason for peanut butter to explain overall stock prices, but if you go through enough grocery store prices, it just might happen that one of them accidentally moves up and down with the movements in the stock market over a particular period of time. Similar problems can occur with other obviously unrelated variables, such as the incidence of full moons or sunspots.

There are ways to protect against this "dubious variable" problem. One is to expand the original sample period. If no true causal relationship exists between the two variables, this coincidence is unlikely to keep occurring in future years. And this is precisely what I did as more data became available: Originally, I looked at data through 1992, then extended it to 1994, then up until 1996, and, in recent working papers, up through 1998. If Ehrlich understood this, he would realize that this is equivalent to his request that I should try to "predict how the crime rates will change in the future."

Another approach guarding against the "dubious variable" problem is to replicate the same test in many different places. Again, this is exactly what I have done here: I have studied the impact of right-to-carry laws in different states at different times, and I have included new states as more and more states have adopted these laws as the time period has been extended.

As discussed in my previous reply, I have also provided many qualitatively different tests, linking not only the changes in gun laws to changes in crime rates but also the actual issuance of permits; the changes in different types of crimes; rates of murders in public and private places; and comparisons of border counties in states with and without right-to-carry laws. Even if I accidentally found a variable that just happened to be related to crime in one of these dimensions, it seems unlikely that you would get consistent results across all these different tests.

In any case, as far as I know, no one except Ehrlich is arguing that testing whether right-to-carry laws affect crime is the theoretical equivalent of including variables such as full moons. Whatever one's views on the topic, there are legitimate questions over whether these laws increase or decrease crime--and the only way that we can test that is to include them as a variable in the regressions.

However, the bottom line is clear: If Ehrlich believes that there is a particular variable that has been left out and that corresponds with all these changes, I have given him the data set; instead of speculating about what might be, he should actually do the work to see if his concerns are valid. No previous study has accounted for even a fraction of the alternative explanations for changing crime rates as I have and, more important, my regressions explain over 95 percent of the variation in crime rates over time.

His concerns about using before-and-after trends make little sense to me because I report the results in many different ways: linear and nonlinear trends before-and-after, year-to-year changes, and before-and-after averages. Readers of my book can view the graphs with the year-to-year changes and judge for themselves when the change in trends occur.

As I explicitly note in my book (pages 146-7 in the first edition), my graphs showing the nonlinear trends before and after the change in laws are constructed similarly to how other economists have analyzed crime data. No explanation is offered for why I shouldn't have focused on whether there was a decline in crime relative to other states that did not adopt the right-to-carry laws.

Ehrlich might find it amusing that deterrence does work, but the data on guns and crime consistently shows that the greater the likelihood that a person can defend themselves produces more deterrence. William M. Landes and I point to evidence that perpetrators of multiple victim shootings are disproportionately psychotic, deranged, or irrational. Ehrlich and others claim that a law permitting individuals to carry concealed weapons would therefore not deter shooting sprees in public places (though it might reduce the number of people killed or wounded). Yet, a right-to-carry law both will raise the potential perpetrator's cost (he is more likely to be wounded or killed or apprehended if he acts) and lower his expected benefit (he will do less damage if he encounters armed resistance). Even those bent on suicide may refrain from attacking if the harm that they can do is sufficiently limited. Although not all offenders will alter their behavior in response to the law, some individuals might refrain from a shooting spree.

Instead of so casually dismissing our result as "very humorous," Ehrlich and others should rise to the challenge to examine the data and see if they can offer a better explanation for the large drops in multiple victim public shootings when states adopt right-to-carry laws. These crimes have seriously shocked the nation and finding ways to reduce such incidents are very important.

Finally, in both the first and second editions of my book, I have responded to the critics of my work that Ehrlich mentions in his last dispatch (Interested readers should see chapters 7 and 9 of More Guns, Less Crime).

This debate has focused on just my findings dealing with right-to-carry laws, but what is just as important are the overall effects of gun control laws. Despite the best of intentions, law-abiding citizens, not criminals, are most likely to obey the different restrictions that are imposed. Disarming the law-abiding relative to criminals has one consequence: more crime.

John R. Lott, Jr. is a senior research scholar at Yale University School of Law and the author of More Guns, Less Crime: Understanding Crime and Gun Control Laws (University of Chicago Press, 2000)

E-mail us your comments on the gun debate.