Back to my web page at http://www.claytoncramer.com.
[This appeared in a slightly different form in "California's Handgun Waiting Period Law, 1952-1990: Did It Work?", American Rifleman, April, 1993] All of these figures appeared in the original American Rifleman; only one of them appeared in the San Jose Mercury-News version.
How effective are handgun waiting periods in preventing murder? Do they decrease the murder rate? Since a national waiting period for handgun purchases has been touted as a possible solution to America's murder problem, it is appropriate to examine how such waiting period laws have actually worked.
Unfortunately, we don't have the option of comparing equivalent populations, one with waiting periods, and one without, for the simple reason that the factors that determine murder rates are not perfectly understood. The best we can hope for is to examine the year to year changes in murder rates before and after a waiting period is created or extended.
In much the same way that the populations of adjoining states are not exactly comparable, the demographics of a state change from year to year. But while the population characteristics of two adjoining states may be dramatically different in subtle and unobvious ways, the demographics within a state change very slowly from year to year, as people age, change their behavior, or move in or out of the state. As a consequence, while looking at a state's crime rates through time is still a technique with limitations, it is nowhere near as prone to problems of non-comparability as comparing different states.
What result should we expect a waiting period to have? There are two purposes usually given for firearms waiting periods: time for a background check on the purchaser; and a "cooling off" period, while passions abate. As instant background checks become feasible, and the waiting period is no longer needed for that purpose, the argument for a handgun waiting period relies on the claim that waiting periods save lives because of the "cooling off" of passions. What evidence does California's experience provide to judge the accuracy of such a statement?
California is very nearly a perfect example of a laboratory for waiting period laws. California is sufficiently large, and its urban populations are so remote from its borders with adjoining states, that a waiting period is enforced not only by state law, but also by the limits of geography. In addition to the Gun Control Act of 1968, which prohibited interstate purchases of handguns, most of California's murders happen in a small number of urban counties that are at least six to eight hours driving time from other states. Even assuming that a seller could be found immediately over the border in Nevada, Oregon, or Arizona, willing to break federal law, the travel time alone imposes a waiting period.
California has also had a waiting period for handgun sales since at least 1923. The California Legislature increased the handgun waiting period from one to three days in 1955, to five days in 1965, and to the current 15 days in 1975. Figure 1, "California Handgun Waiting Periods & Murder Rates" plots the murder rate per 100,000 Californians during the period 1952 through 1990. (The use of a murder rate, which counts murders relative to the size of the population, eliminates changes in the number of murders caused by changes in the number of people living in California.)
The increase from one to three days in 1955, and from three to five days in 1965, had no apparent effect on rising murder rates. Indeed, the California murder rate went from a bit above 2/100,000 people in 1952, to over 10/100,000 by 1975. While it is certainly true that murder rates rose throughout the United States during this period, as Figure 1 shows, California's murder rate rose even faster than the murder rate for the rest of the United States.
The first full year of the fifteen day waiting period, 1976, showed a 1% decline in murder rates - followed by continually rising murder rates, peaking in 1980. In fact, murder rates didn't start to decline until 1981, five full years after the new waiting period took effect. Can the advocates of waiting periods take heart from the fact that California's murder rates eventually fell?
No. Murder rates in the rest of the United States roughly rise and fall in parallel with the California murder rates. Indeed, there is a 0.93 correlation between the California murder rate, and the murder rate for the rest of the United States -- astonishingly high. See Figure 3. From 1974 onward, the strong similarity in these curves suggests that factors beyond the reach of California's laws were by far the most important determiner of California's murder rate. The California murder rate fell 36% from 1980 to 1985, but the murder rate for the rest of the U.S. (the U.S. murder rate with California's population and murders removed) fell 26% during the same period.
In much the same way, the murder rate for the rest of the U.S. rose 13% from 1985 to 1990; during the same period, the California murder rate also rose 13%. In both cases, a small "hump" of murders appeared in 1986, there were declines in 1987 and 1988, then the murder rate increased in 1989 and 1990. Indeed, where California murder rates were always lower than for the rest of the United States before the 15 day waiting period took effect in 1976, California's murder rate from 1976 onward was always higher than for the rest of the United States. See Figure 2
What conclusions can we reach from the data? Positive correlation (that two events happen together with high frequency) does not prove causality (that one event causes the other event). As an example, higher ice cream cone sales are positively correlated with increased rape rates - but this doesn't mean that ice cream cone sales cause rape. Rather, one underlying cause of both ice cream cone sales and rape is hot weather. (It is probably because there are more unlocked windows in the hot summer months, allowing easier access to victims at home). It is also true that zero correlation between two events does not disprove causality -- but if no correlation (positive or negative) between two events can be shown, it should certainly make us skeptical that a causal relationship exists.
There are a number of possibilities that might explain the astonishingly high 0.91 correlation between handgun waiting periods and murder rates in California:
1. Waiting periods may be completely useless for reducing murder rates.
This doesn't mean that a waiting period doesn't occasionally prevent
a murder from happening; there are probably cases where delay may cause
reflection, and the passions of the moment may subside. So what are the
circumstances that would cancel out the murders prevented by a waiting
period? o That so-called "crimes of passion" are not readily deterred by waiting
periods because of substitution of other weapons.
o That murderers are already armed, and thus waiting periods do not
o That murderers don't bother buying handguns legally.
o For every murder thus prevented, a victim was similarly
prevented from buying a handgun for self-defense.
o That so-called "crimes of passion" are not readily deterred by waiting periods because of substitution of other weapons.
o That murderers are already armed, and thus waiting periods do not delay them.
o That murderers don't bother buying handguns legally.
2. Waiting periods may be of some limited utility for reducing murder rates, but were overwhelmed by other factors.
3. Waiting periods may increase murder rates, but the effect that we see may have been caused by other factors, and the waiting period made only a minor contribution.
4. Waiting periods may increase murder rates, and the effect was dramatic enough to explain what we see.
5. Waiting periods may be ineffective, and the political energy devoted to them (both in support and in opposition), meant that more effective strategies for reducing murder simply fell by the wayside, and the murder rate was unaffected.
Unfortunately, it is impossible, based on the available data, to determine which of these five possibilities is correct. In a sense, though, it doesn't really much matter which of these propositions is the answer to the question. If possibility #2 above (the most positive interpretation of the data for advocates of handgun waiting periods) is true, the effectiveness of handgun waiting periods as crime control is so slight that it doesn't jump out at us from the data. The political energy devoted to waiting periods thus makes no sense -- especially since opposition to the Brady Bill centers on the waiting period aspect of the law. (Contrary to popular perceptions, the National Rifle Association has supported, and continues to reluctantly support an instant background check law as an alternative to a national waiting period. Their support for such laws is reluctant, partly because of skepticism that any background check law is going to solve the fundamental causes of violence in our society.)
The most plausible explanation for the continued popularity of waiting periods for gun purchase, in spite of their ineffectiveness, is that they represent a cheap consensus among politicians. A major division exists in our society about the best method for solving the problems of murder and other violent crimes. Conservatives generally favor longer sentences for violent felonies. This requires more prisons; prisons cost money. Liberals generally favor more spending on welfare programs and education as a method of reducing violent crime; these aren't cheap, either. Politicians are reluctant to anger taxpayers by use of either strategy. But creating and lengthening waiting periods for handgun purchase, because they require no expenditure of public funds, are an especially tempting "solution" to the violent crime problem - even when the evidence suggests that they are at best, of only minor value, and at worst, completely ineffective.
Research assistance for this article was provided by Charles R. Hardy.
2 Cal. Penal Code Ann. 12071 (1991).
3 California murder rates for 1952-1989 in this article come from California Dept. of Justice, Crimes Reported, 1952-1989, Number and Rate per 100,000 Population. California murder rates for 1990 are from California Dept. of Justice, California Criminal Justice Profile 1990, 5. U.S. murder rates were derived from population and raw murder figures in FBI Uniform Crime Reports for the United States, 1952-1990.
4 This is an important distinction, because California's population is large enough, especially in the latter part of the period studied, that its murders significantly influence the U.S. murder rate. Hence, throughout this article, we are comparing the California murder rate, with the murder rate of the U.S. minus California.
Year,Waiting Period (in days),California Murder Rate,US - Cal. Murder Rate '52,1,2.40,4.75 '53,1,2.28,4.62 '54,1,3.33,4.28 '55,1,3.20,4.21 '56,3,3.49,4.18 '57,3,3.50,4.77 '58,3,3.71,4.77 '59,3,3.37,4.96 '60,3,3.91,5.17 '61,3,3.70,4.78 '62,3,3.94,4.56 '63,3,3.71,4.57 '64,3,4.16,4.89 '65,3,4.76,5.10 '66,5,4.69,5.65 '67,5,5.40,6.16 '68,5,5.99,6.89 '69,5,6.93,7.23 '70,5,6.76,7.81 '71,5,8.03,8.54 '72,5,8.69,8.84 '73,5,8.92,9.24 '74,5,9.30,9.67 '75,5,10.20,9.42 '76,15,10.09,8.45 '77,15,11.10,8.41 '78,15,11.39,8.49 '79,15,12.65,9.17 '80,15,14.33,9.63 '81,15,12.98,9.42 '82,15,11.24,8.79 '83,15,10.49,7.97 '84,15,10.63,7.58 '85,15,10.55,7.64 '86,15,11.23,8.23 '87,15,10.59,7.98 '88,15,10.41,8.18 '89,15,10.87,8.40 '90,15,11.88,9.03