11. See Sam Peltzman, "The Effects of Automobile Safety Regulation," Journal of Political Economy 83 (Aug. 1975): 677-725.
12. To be more precise, a one-standard-deviation change in the probability of arrest accounts for 3 to 11 percent of a one-standard-deviation change in the various crime rates.
13. Translating this into statistical terms, a one-standard-deviation change in the percentage of the population that is black, male, and between 10 and 19 years of age explains 22 percent of the ups and downs in the crime rate.
14. This is particularly observed when there are more black females between the ages of 20 and 39, more white females between the ages of 10 and 39 and over 65, and females of other races between 20 and 29.
15. In other words, the second number shows how a one-standard-deviation change in an explanatory variable explains a certain percent of a one-standard-deviation change in the various crime rates.
16. While I believe that such variables as the arrest rate should be included in any regressions on crime, one concern with the results reported in the various tables is over whether the relationship between the "shall-issue" variable and the crime rates occurs even when all the other variables are not controlled for. Using weighted least squares and reporting only the "shall-issue" coefficients, I estimated the following regression coeffi-
How do average crime rates differ among states with and without nondiscretionary laws?
NOTES TO PAGES 58-62/281
Note: The only factors included are the presence of the law and/or year-specific effects. All these differences are statistically significant at least at the 1 percent level for a two-tailed t-test. To calculate these percentages, I used the approximation 100 [exp(coefficient) — 1).
17. The time-trend variable ranges from 1 to 16: for the first year in the sample, it equals 1; for the last year, it is 16.
18. Other differences arise in the other control variables, such as those relating to the portion of the population of a certain race, sex, and age. For example, the percent of black males in the population between 10 and 19 is no longer statistically significant.
19. If the task instead had been to determine the difference in crime rates when either all states or no states adopt the right-to-carry handgun laws, the case of all states adopting concealed-handgun laws would have produced 2,048 fewer murders, 6,618 fewer rapes, 129,114 fewer aggravated assaults, and 86,459 fewer robberies. Non-arson property crimes also would have fallen by 511,940.
20. Generally, aggregation is frowned on in statistics anyway, as it reduces the amount of information yielded by the data set. Lumping data together into a group cannot yield any new information that did not exist before; it only reduces the richness of the data.
21. Eric Rasmusen, "Stigma and Self-Fulfilling Expectations of Criminality," Journal of Law and Economics 39 (Oct. 1996): 519-44.
22. In January 1996, women held 118,728 permits in Washington and 17,930 permits in Oregon. The time-series data available for Oregon during the sample period even indicate that 17.6 percent of all permit holders were women in 1991. The Washington state data were obtained from Joe Vincent of the Department of Licensing Firearms Unit in Olym-pia, Washington. The Oregon state data were obtained from Mike Woodward of the Law Enforcement Data System, Department of State Police, Salem, Oregon. Recent evidence from Texas indicates that about 28 percent of applicants were women ("NRA poll: Salespeople No. 1 for Permit Applications" Dallas Morning News, Apr. 19, 1996, p. 32A).
23. For an interesting discussion of the benefits to women of owning guns, see Paxton Quigley, Armed and Female (New York: E. P. Dutton, 1989).
24. Unpublished information obtained by Kleck and Gertz in their 1995 National Self-Defense Survey implies that women were as likely as men to use handguns in self-defense in or near their homes (defined as in the yard, carport, apartment hall, street adjacent to home, detached garage, etc.), but that women were less than half as likely to use a gun in self-defense away from home. See Gary Kleck and Marc Gertz, "Armed Resistance to Crime: The Prevalence and Nature of Self-Defense with a Gun," Journal of Criminal Law and Criminology 86 (Fall 1995): 249-87.
282 / NOTES TO PAGES 63-73
25. Counties with real personal income of about $15,000 in real 1983 dollars experienced 8 percent drops in murder, while mean-income counties experienced a 5.5 percent drop.
26. Lori Montgomery, "More Blacks Say Guns Are Answer to Urban Violence," Houston Chronicle, July 9, 1995, p. Al. This article argues that while the opposition to guns in the black community is strong, more people are coming to understand the benefits of self-protection.
27. For an excellent overview of the role of race in gun control, see Robert J. Cottrol and Raymond T. Diamond, "The Second Amendment: Toward an Afro-Americanist Reconsideration," Georgetown Law Review 80 (Dec. 1991): 309.
28. See William Van Alstyne, "The Second Amendment Right to Arms," Duke Law Review 43 (Apr. 1994): 1236-55. In slave states prior to the Civil War, the freedoms guaranteed under the Bill of Rights were regularly restricted by states because of the fear that free reign might lead to an insurrection. As Akhil Reed Amar writes, "In a society that saw itself under siege after Nat Turner's rebellion, access to firearms had to be strictly restricted, especially to free blacks." See Akhil Reed Amar, "The Bill of Rights and the Fourteenth Amendment," Yale Law Journal 101 (Apr. 1992): 1193.
29. Associated Press Newswire, May 9, 1997, 4:37 p.m. EDT. As the Washington Times recently noted, this story "comes at an awkward time for the administration, since President Clinton has spent the last week or two berating Republicans for failing to include in anti-crime legislation a provision requiring that child safety locks be sold with guns to keep children from hurting themselves" (Editorial, "The Story of a Gun and a Kid," Washington Times, May 22, 1997, p. A18).
30. The conversation took place on March 18, 1997, though regrettably I have misplaced the note containing the representative's name.
31. John Carpenter, "Six Other States Have Same Law," Chicago Sun-Times, Mar. 11, 1997, p. 8.
32. John J. Dilulio, Jr., "The Question of Black Crime," The Public Interest 117 (Fall 1994): 3—24. Similar concerns about the inability of minorities to rely on the police was also expressed to me by Assemblyman Rod Wright (D—Los Angeles) during testimony before the California Assembly's Public Safety Committee on November 18, 1997.
33. One additional minor change is made in two of the earlier specifications. In order to avoid any artificial collinearity either between violent crime and robbery or between property crimes and burglary, violent crimes net of robbery and property crimes net of burglary are used as the endogenous variables when robbery or burglary are controlled for.
34. The Pearson correlation coefficient between robbery and the other crime categories ranges between .49 and .80, and all are so statistically significant that a negative correlation would only appear randomly once out of every ten thousand times. For burglary, the correlations range from 0.45 to 0.68, and they are also equally statistically significant.
35. All the results in tables 4.1 and 4.4 as well as the regressions related to both parts of figure 4.1 were reestimated to deal with the concerns raised in chapter 3 over the "noise" in arrest rates arising from the timing of offenses and arrests and the possibility of multiple offenders. I reran all the regressions in this section by limiting the sample to those counties with populations over 10,000, over 100,000, and then over 200,000 people. The more the sample was restricted to larger-population counties, the stronger and more statistically significant was the relationship between concealed-handgun laws and the previously reported effects on crime. This is consistent with the evidence reported in figure 4. 1. The arrest-rate results also tended to be stronger and more significant. I further reestimated all the regressions by redefining the arrest rate as the number of arrests over the last three years divided by the total number of offenses over the last three years. Despite the reduced sample size, the results remained similar to those already repo
rted.
NOTES TO PAGES 81-97/283
the coefficients for the period before the law produces results that yield the same pattern after the passage of the law. Using both the time-trend and the time-trend-squared relationships, the F-tests reject the hypothesis that the before and after relationships are the same, at least at the 10 percent level, for all the crime categories except aggravated assault and larceny, for which the F-tests are only significant at the 20 percent level. Using only the time-trend relationship, the F-tests reject the hypothesis in all the cases.
37. The main exception was West Virginia, which showed large drops in murder but not in other crime categories.
38. See Thomas B. Marvell and Carlisle E. Moody, "The Impact of Enhanced Prison Terms for Felonies Committed with Guns," Criminology 33 (May 1995): 259—60.
39. I should note, however, that the "nondiscretionary" coefficients for robbery in the county-level regressions and for property crimes using the state levels are no longer statistically significant.
40. Toni Heinzl, "Police Groups Oppose Concealed-Weapons Bill," Omaha World-Herald, Mar. 18, 1997, p. 9SF.
41. A simple dummy variable is used for whether the limit was 18 or 21 years of age.
42. Here is one example: "Mrs. Elmasri, a Wisconsin woman whose estranged husband had threatened her and her children, called a firearms instructor for advice on how to buy a gun for self-defense. She was advised that, under Wisconsin's progressive handgun law, she would have to wait 48 hours so that the police could perform the required background check.
"Twenty-four hours later,... Mrs. Elmasri's husband murdered the defenseless woman and her two children" (William P. Cheshire, "Gun Laws No Answer for Crime," Arizona Republic, Jan. 10, 1993, p. CI.) Other examples can be found in David B. Kopel, "Background Checks and Waiting Periods," in Guns: Who Should Have Them, ed. David B. Kopel (Amherst, NY: Prometheus Books, 1995.) Other examples tell of women who successfully evaded these restrictions to obtain guns.
In September 1990, mail carrier Catherine Latta of Charlotte, N. C, went to the police to obtain permission to buy a handgun. Her ex-boyfriend had previously robbed her, assaulted her several times, and raped her. The clerk at the sheriffs office informed her that processing a gun permit would take two to four weeks. "I told her I'd be dead by then," Latta recalled.
That afternoon, Latta bought an illegal $20 semiautomatic pistol on the street. Five hours later, her ex-boyfriend attacked her outside her house. She shot him dead. The county prosecutor decided not to prosecute Latta for either the self-defense homicide or the illegal gun. (Quoted from David B. Kopel, "Guns and Crime: Does Restricting Firearms Really Reduce Violence?" San Diego Union-Tribune, May 9, 1993, p. G4.)
For another example where a woman's ability to defend herself would have been impaired by a waiting period, see "Waiting Period Law Might Have Cost Mother's Life," USA Today, May 27, 1994, p. 10A.
43. Quoted in David Armstrong, "Cities' Crime Moves to Suburbs," Boston Globe, May 19, 1997, pp. 1 and B6.
CHAPTER FIVE
1. While county-level data were provided in the Supplementary Homicide Reports, matching these county observations with those used in the Uniform Crime Reports proved unusually difficult. A unique county identifier was used in the Supplementary Homicide Reports that was not consistent across years. In addition, some caution is necessary in using both the Mortality Detail Records and the Supplementary Homicide Reports, since the murder rates reported
284 / NOTES TO PAGES 98-106
in both sources have relatively low correlations of less than .7 with the murder rates reported in the Uniform Crime Reports. This is especially surprising for the supplementary reports, which are derived from the Uniform Crime Reports. See U.S. Department of Justice, FBI staff, Uniform Crime Reports (Washington, DC: US. Govt. Printing Office) for the years 1977 to 1992.
2. Indeed, the average age of permit holders is frequently in the mid- to late forties (see, for example, "NRA poll: Salespeople No. 1 for Permit Applications," Dallas Morning News, Apr. 19, 1996, p. 32A.) In Kentucky the average age of permit holders is about fifty (see Terry Flynn, "Gun-Toting Kentuckians Hold Their Fire," Cincinnati Enquirer, June 16, 1997, p. Al).
3. This is the significance for a two-tailed t-test.
4. Similar breakdowns for deaths and injuries are explored in much more depth in a paper that I have written with William Landes; see William Landes and John R. Lott, Jr., "Mass Public Shootings, Bombings, and Right-to-Carry Concealed-Handgun Laws," University of Chicago working paper, 1997.
5. A second change was also made. Because of the large number of observations noting no deaths or injuries from mass public shootings in a given year, I used a statistical technique known as Tobit that is particularly well suited to this situation.
6. The results shown below provide the estimates for the simple linear time trends before and after the adoption of the law. They demonstrate that for each year leading up to the passage of the law, total deaths or injuries from mass public shootings rose by 1.5 more per 10 million people and that after the passage of the law, total deaths or injuries fell by 4 more per 10 million people. The difference in these two trends is statistically significant at the 1 percent level for a two-tailed t-test. It is interesting to note that higher murder arrest rates, although they deter murderers, do not seem to deter perpetrators of mass public shootings.
Linear time trends for deaths and injuries from mass public shootings before and after adoption of concealed-handgun law
Total deaths and injuries per 100,000 population
Average annual change for years after adoption
of the law -0.04*
Average annual change for years before adoption
of the law 0.015*
Arrest rate for Murder —0.0003
*Statistically significant at least at the 10 percent level for a two-tailed t-test Note: numbers are negative; years furthest beyond adoption are the largest
7. See appendix 4 for the means and standard deviations of the variables used in these regressions.
8. Again, this is stating that a one-standard-deviation change in arrest rates explains more than 15 percent of a one-standard-deviation change in crime rates.
9. Running the regressions for all Pennsylvania counties (not just those with more than 200,000 people) produced similar signs for the coefficient for the change in concealed-handgun permits, though the coefficients were no longer statistically significant for violent crimes, rape, and aggravated assault. Alan Krug, who provided us with the Pennsylvania handgun-permit data, told us that one reason for the large increase in concealed-handgun permits in some rural counties was that people used the guns for
NOTES TO PAGES 106-117/285
hunting. He told us that the number of permits issued in these low-population, rural counties tended to increase most sharply in the fall around hunting season. If people were in fact getting large numbers of permits in low-population counties (which already have extremely low crime rates) for some reason other than crime, it would be more difficult to pick up the deterrent effect of concealed handguns on crime that was occurring in the larger counties.
10. A one-standard-deviation change in conviction rates explains 4 to 20 percent of a one-standard-deviation change in the corresponding crime rates.
11. I reran these regressions using the natural logs of the arrest and conviction rates, and I consistently found statistically larger and even economically more important effects for the arrest rates than for the conviction rates.
12. For example, see Dan M. Kahan, "What Do Alternative Sanctions Mean?" University of Chicago Law Review 63 (1996): 591-653.
13. See John R. Lott, Jr., "The Effect of Conviction on the Legitimate Income of Criminals," Economics Letters 34 (Dec. 1990): 381—85; John R. Lott, Jr., "An Attempt at Measuring the Total Monetary Penalty from Drug Convictions: The Importance of an Individual's Reputation," Journal of Legal Studies 21 (Jan. 1992): 159-87; John R. Lott, Jr., "Do We Punish High-
Income Criminals Too Heavily?" Economic Inquiry 30 (Oct. 1992): 583-608.
14. Put differently, six of the specifications imply that a one-standard-deviation change in the number of concealed-handgun permits explains at least 8 percent of a one-standard-deviation change in the corresponding crime rates.
15. Philip Heymann, a former deputy attorney general in the Clinton administration and currently a law professor at Harvard University, wrote, "None of this [the drop in crime rates] is the result of... the Brady Act (for most guns were never bought by youth from licensed gun dealers)." See "The Limits of Federal Crime-Fighting," Washington Post, Jan. 5, 1997, p. C7.
16. For a discussion of externalities (both benefits and costs) from crime, see Kermit Daniel and John R. Lott, Jr., "Should Criminal Penalties Include Third-Party Avoidance Costs?" Journal of Legal Studies 24 (June 1995): 523-34.
17. Alix M. Freedman, "Tinier, Deadlier Pocket Pistols Are in Vogue," Wall Street Journal, Sept. 12, 1996, pp. Bl, B16.
18. One hundred and eighty-two million people lived in states without these laws in 1991, so the regressions would have also implied nine more accidental deaths from handguns in that year.
19. Given the very small number of accidental deaths from handguns in the United States, the rate of such deaths in the vast majority of counties is zero, and the last two columns of table 5.6 again use Tobit regressions to deal with this problem. Limitations in statistical packages, however, prevented me from being able to control for all the county dummies, and I opted to rerun these regressions with only state dummy variables.
20. For example, see Nicholas D. Kristof, "Guns: One Nation Bars, the Other Requires," New York Times, Mar. 10, 1996, sec. 4, p. 3. For some evidence on international gun ownership rates see Munday and Stevenson, Guns and Violence (1996): 30.
21. See Ian Ayres and Steven Levitt, "Measuring Positive Externalities from Unobserv-able Victim Precaution: An Empirical Analysis of Lojack," NBER working paper 5928 (1997); and John Donohue and Peter Siegelman, "Is the United States at the Optimal Rate of Crime?" Journal of Legal Studies 27 (Jan. 1998).
More Guns Less Crime Page 34