St. Louis University Public Law Review
Crime and Punishment Symposium,
vol 12, no. 2, 1993: 361.
Posted for Educational use only.
The printed edition remains canonical. For citational use please visit the local law library or obtain a back issue.
PERILOUS PROTECTION: A REPLY TO KOPEL
Nicholas DixonCopyright © 1993 by the St. Louis University School of Law & Nicholas Dixon
Kopel's response to my paper has several virtues. He avoids the invective that has often pervaded anti-handgun ban writings. He raises some important practical obstacles which need to be faced by proponents of a handgun ban. He furthers the existing anti-prohibition literature by introducing new data and arguments. Most valuable of all, he challenges the causal argument that underlies my paper. In so doing, he forces me to clarify and improve my argument.
Kopel is right that "the facts themselves" are crucial. [1] However, the relevant facts (as opposed to surveys) unwarranted conjectures, and tenuous historical analogies) strongly support my hypothesis that a handgun ban will substantially reduce violent crime in this country. Like the other opponents of handgun prohibition whom I address in my paper, he has failed to meet the burden of proof encumbent on those who advance a hypothesis that flies in the face of the evidence that I have adduced.
I. MY PROPOSAL FOR HANDGUN PROHIBITION
A. Police Exceptions
Kopel criticizes my proposal to exempt police from a handgun ban. [2] One argument I give to justify this exemption is that the police need superior firepower in order to act as protectors of society and law enforcers. [3] Kopel claims that the premise that the police need superior force does not specifically entail that they may carry handguns. The reason why Kopel says the conclusion does not follow is that the police's need for superior force would not justify the use of heavier weaponry such as hand-held nuclear devices. [4] Kopel's [Page 362] reasoning is invalid. The fact that police may not absurdly overarm themselves does not entail that they are not justified in using handguns, which are powerful enough and far better suited for the police's role as protectors of society.
Kopel raises some serious concerns about misuse of firearms by police. [5] He may be right that civilians could be as well trained in defensive firearms use as police. However, as of now, most handgun owners do not go through the "few dozen hours of training" [6] that police currently have at a police academy. Moreover, even if civilians did undergo this training, they are not law enforcers. Police are trained law enforcers, and this is why they have more justification for carrying handguns.
Nonetheless, Kopel makes some helpful recommendations which I accept as amendments to my proposal for a handgun ban. First, some police do not need handguns (e.g. those who deal exclusively with paperwork), and the exception should be confined to officers in the line of fire. [7] Second, he refers to countries such as Britain where police rarely carry guns, and Japan and Canada, where they rarely use them, and proposes that United States police carry and use guns as little as possible. [8] I look forward to the day when gun violence in this country has diminished to such a level that police no longer routinely need to carry guns. Regrettably, Kopel does not display the same zeal in recommending minimal firearms use by civilians.
B. Handgun Homicide and Overall Homicide
Kopel tries to show that, at best, my argument shows a causal relationship between handgun ownership and handgun homicides. Unless I can show what Kopel disputes (that a reduction in handgun homicides would also effect a reduction in overall homicides) the only reason for advancing a handgun ban would be a peculiar preference for homicide by means other than handguns. [9]
Contrary to Kopel's claim, I do present arguments linking a reduction in handgun homicides to a reduction in overall homicide rates in the United States, which is the country for which I propose a handgun ban. In my paper I cite the fact that 49.5% of all homicides in the United States in 1990 were known to be committed with [Page 363] handguns. [10] This percentage has hovered around 50% since at least 1970, and in 1991 rose to 53.1%. [11] My claim that a reduction in handgun homicides in the United States would result in a reduction in the overall homicide rate is based on the high percentage of homicides committed in the United States with handguns, and the fact the handguns are uniquely suited to homicide and other violent crimes due to "their relative cheapness, their small size (and hence greater concealability), and the fact that they are easy to use." [12] I respond at considerable length to objections that a reduction in handgun homicides would only cause an increase in other kinds of homicide. [13]
Kopel is right that I do not cite overall homicide rates for the countries whose handgun homicide rates I compare with that of the United States. But I do not need to do so. The purpose of the international comparison is to establish a causal connection between handgun ownership rates and handgun homicide rates. If my hypothesis is correct, then a reduction in the number of handguns owned in the United States will reduce its handgun homicide rate. In the previous paragraph, I have summarized my argument as to why a reduction in handgun homicides in the United States would reduce its overall homicide rate. [14]
There may be little correlation between the handgun homicide rate and overall homicide rate in some countries. [15] This is doubtless due to the fact that handgun ownership and homicide rates in these countries are too small to make a substantial impact on overall violent crime rates. This is transparently not the case in the United States, in which a huge number and a substantial percentage of murders and other violent crimes are committed with handguns. [16] My proposal is for a handgun ban in the United States, in order to reduce handgun violence and hence overall violence in the United States.
Finally, the overall homicide rate in the United States is almost three times as high as that of Canada, and on average over 5.25 times as high as other countries in my comparison group which [Page 364] includes Switzerland, Australia, Great Britain, and Israel. [17] These statistics undermine Kopel's attempt to deny the importance of handgun homicides for the overall homicide rate in the United States.
C. My Central Causal Argument
Kopel's strategy in responding to my central causal argument--that the United States' high handgun ownership rate is a major determinant of its high handgun homicide rate--is to try to show that my hypothesis is just one of several equally plausible interpretations that could be put on the data. While my interpretation is compatible with the evidence, he argues, it is not established by it. [18]
In my paper I present two adequacy conditions which must be met by those who advance causal arguments. They are minimal requirements for distinguishing causal relationships from accidental correlations.
1. One must show that there are no other variables which [have been shown to] correlate better with the effect, and which would account for the effect better than, or in place of, the posited cause.
2. The second requirement is ... the provision of a probable theoretical explanation of how the causation occurred. [19]
Kopel begins by criticizing my application of the first adequacy test to my handgun hypothesis. [20] Kopel is right to criticize my assertion, which I now retract, that my hypothesis meets this test as long as no other cause has been shown to be the only cause. [21] Virtually every hypothesis in the social sciences, in which actions and events have multiple causes, would pass this version of my test. However, the application of the first adequacy test does not stand or fall with my ill-advised use of the word "only", and is far better expressed a few sentences earlier. To establish my handgun hypothesis I need to show "that there is no other cause that correlates so well with handgun murder as to rule out my own causal hypothesis." [22] While Kopel has correctly criticized one careless phrase of mine, he has not refuted the substance of my first test. In fact, the [Page 365] test is precisely what Kopel himself advocates, i.e., "demonstrating handgun density is the best variable." [23] The task of proving the negative proposition that no better variable than handgun ownership exists would be unreasonably burdensome. Instead, throughout the rest of the paper I show that no other putative cause has been shown to better correlate with handgun homicide in such a way that my hypothesis is ruled out.
Kopel then argues that I have failed to meet my second adequacy test, because of the inadequacy of my theoretical explanation of how high handgun ownership rates cause high handgun homicide rates. He contests my assumption that the countries represented in my international comparison are relatively similar in human nature, and in the psychological causes of murder. He devotes two pages to showing differences between these countries in national character, and to speculations as to differences in "violence-inducing provocations". [24] These differences, and not our rate of handgun ownership, may account for the high rate of handgun homicide in the United States. Kopel is trying to show that his speculations as to the cause of our handgun homicide rates are on a par with my causal hypothesis. He asserts that my theoretical explanation is no better than a restatement of the intuition that underlies my whole article. [25]
Kopel's attack on my theoretical account of the causal connection between handgun ownership and handgun homicide focuses on a minor assumption I make concerning the rough similarity of the United States and the countries to which I compare it. He barely mentions the heart of my theoretical explanation, which, as we have seen, he dismisses as a mere statement of intuition. I confess that I thought that my theory was too obvious to require much explanation. I will now make up for this weakness in my original paper by providing more detail.
My causal theory is very simple. The main raison d'etre of handguns, even when used defensively, is their ability to kill and injure. Their lethality, combined with their cheapness, small size, and ease of use, make them uniquely suited for offensive use. Most important of all, handguns are necessary for handgun homicide. The more handguns people own, the more likely handgun homicides are to occur. By analogy, the more ATV's are driven, the more likely ATV accidents are to occur. The more mosquitoes there are, the more people are likely to be bitten. The burden of proof is on those who would deny these common sense theories. [Page 366]
Underlying Kopel's criticism that I have merely stated an "intuition", instead of providing a theoretical explanation of the causal mechanism, may be the mistaken belief that only a complex explanation counts as a theory. In reality, however, a theory need only be as complex as the data demands. A theory as to why exposure to certain chemicals increases the risk of cancer must be enormously complex. In contrast, a theory as to why the sun rises and sets can remain at the level of common sense: it is because the earth spins. Would Kopel deny this the status of a causal theory, and insist that it is merely an intuition? Of course, a theory alone does not provide a causal argument: data is needed to support the theory. But this is precisely what is provided by both observation of the earth, sun, and other heavenly bodies, and by my international handgun data. I do not claim that the causation of handgun homicide rates is a simple matter. The causes are various and complex. What I do claim is that the handgun ownership rate is one of these causes, and that this particular causal mechanism is very straightforward.
Another problem is that Kopel considers my theoretical explanation of the causal connection between the United States' high handgun ownership and homicide rates in isolation from the rest of my causal argument. He forgets that the starting point of my argument is the empirical data, which indicates a striking correlation between handgun ownership and handgun homicide rates in the countries I have studied. I have then given a theoretical account of why this correlation is a causal connection. Much of my article is devoted to showing that my opponents have failed to produce alternative causal hypotheses that disprove mine. [26]
Let me make my method of causal reasoning even more explicit. I am using Mill's methods of (1) concomitant variation, and (2) difference. [27] First, the method of concomitant variation consists in showing that the amount of the effect (in this case, handgun homicide rates) varies according to the amount of the putative cause (in this case, handgun ownership rates). My international data exhibits exactly such a variation. Second, the method of difference consists in showing that the effect (in this case, exceedingly high handgun homicide rates) is present in the presence of the putative cause (exceedingly [Page 367] high handgun ownership rates), and absent when the putative cause is absent. Again, my data supports use of the method of difference. The handgun homicide rate is exceedingly high (far higher than in any other country in my comparison) only in the United States, which is also the only country in my study with an exceedingly high handgun ownership rate (again, far higher than any other country in my comparison).
The reason why Kopel's alternative account of the international data (in terms of differences between the United States and other countries) is not on a par with my argument is that the data gives prima facie support for my hypothesis, but not for his. His account is essentially a rear-guard action, trying to explain why, appearances notwithstanding, handgun ownership rates are not really causally connected with handgun homicide rates with which they are correlated. Interestingly, the more successful Kopel is at showing the diversity of the countries I study, the less plausible it becomes to dismiss the close correlations between handgun ownership and handgun homicide rates as coincidental. Worse still for his hypothesis, elsewhere in his response, Kopel asserts that handgun ownership helps to reduce homicide. [28] Accordingly, we might expect that the United States would have a vastly lower homicide rate than countries with a lower handgun density. In reality, as we have seen, the overall homicide rate in the United States is substantially higher than that in the comparison countries, and its handgun homicide rate is enormously higher.
Kopel's alternative account of the international data (in terms of alleged differences in national character and in provocations to violence) is a theoretical explanation sorely in need of empirical support, i.e. a speculation. In contrast, the lifeblood of my hypothesis is provided by the empirical evidence I adduce.
It is precisely because the evidence I adduce is so heavily stacked against those who deny my causal hypothesis that the burden of proof is on my opponents. Kopel's "national differences" hypothesis fails to meet the challenge I issued to opponents in my original paper:
[I]t is incumbent on them to produce an alternative causal account proving that the United States' high handgun murder rate is caused by factors unrelated to its high rate of handgun ownership. They must specify what these causes are, quantify their relative presence in the United States as compared to the countries with lower handgun homicide rates, demonstrate that variations in these factors correlate with [Page 368] variations in the handgun murder rate, and provide a plausible theory explaining the causal mechanisms at work. [29]
I now add a further requirement to the burden of proof. My opponents must also provide an explanation of the correlation between handgun ownership and handgun homicide rates, unless, that is, they are willing to dismiss it as a coincidence.
While Kopel has underestimated the force of my causal argument, and wrongly put it on a par with unsubstantiated alternatives, he does at least acknowledge the burden of proof I place on those who advance alternative causal explanations. To his credit, he attempts to meet my demand for greater detail, as he proceeds to give a more substantial version of his alternative causal argument. [30]
Kopel's explanation of the rampant handgun homicide rate in this country appeals to several factors. First, there are various effects of American individualism, including mistrust of police and restrictions on their power, fewer cultural restraints on lawbreaking, and high levels of crime-inducing resentment. Second, the United States has a history of violent racism. Third, more generally, it has a long history of resorting to violence as a means of solving disputes. He concludes that America is more violent than other nations, and would be even if handguns had never been invented. [31] Kopel's book, the argument of which he sketches in his article, appears to be an invaluable piece of scholarship, which enhances our understanding of the causes of the terrible level of violence in this country. [32] However, as summarized in his article, it fails to achieve the purpose for which he advances it here, namely to disprove my argument that the high level of handgun ownership is a significant cause of the high level of handgun homicide in this country.
First, Kopel's explanation is of the high level of violence in general in this country. It does not explain why one half of homicides in this country are committed with handguns. With the exception of Switzerland (where the raw numbers are very low, and guns are more prevalent than in the other countries), the percentage of homicides that are committed with handguns is far higher in the United States than in the comparison countries. [33] Moreover, Kopel's [Page 369] account does not even attempt to explain why the rate of handgun homicide is enormously higher in the United States than in any of the comparison countries. My causal argument, on the other hand, is based on these handgun homicide rates, their correlation with handgun ownership rates, and a theoretical explanation of why the relationship is causal.
Second, Kopel concedes my point that showing additional causal factors does not invalidate my causal argument [34], but he fails to appreciate just how damaging this point is for his case. I can concede that other causes of violence are present, and that we need to address them, while insisting that our enormous rate of handgun ownership is also a significant cause of our alarming handgun homicide rate. Our handgun density exacerbates a situation which, if Kopel is right, is already a fertile ground for violence. Given the historical propensity for violence in this country, allowing the widespread ownership of a particularly lethal and convenient weapon is a recipe for disaster.
Kopel responds that if his analysis is true, "it is certainly possible (although not provable) that handgun density plays a small role, or even no role, in America's violence and murder problem." [35] This indicates the third problem with Kopel's argument, which is his misunderstanding of the nature of empirical hypotheses, and his consequent failure to take seriously the burden of proof incumbent on anyone who challenges my causal hypothesis. I would have granted in advance of reading Kopel's alternative account that my causal hypothesis may be false. It is the very nature of empirical hypotheses that they are not necessarily true. [36] The best we can do is show that they are probably true, and to this end I have adduced statistical evidence, along with a theoretical explanation.
Consequently, it is not enough for Kopel to show that my hypothesis may be false. He must show that it is probably false. To do so, to once again repeat my burden of proof challenge, he must present his own causal argument, backed by statistical evidence of a stronger correlation between his putative cause(s) and handgun homicide, and a plausible theoretical account. His causal argument must be so strong as to rule out mine, since I only claim that handgun density is one of the causes of handgun homicide. Kopel has certainly [Page 370] provided a plausible theoretical account of non handgun-related causes of the higher level of violence in this country. However, it fails to meet the burden of proof. First, as I have already shown, it does not address the specific issue of handgun violence which is my concern. Second, it consists of historical, sociological, and psychological generalizations, and contains no statistical evidence linking these variables with handgun homicide rates. Kopel is right that the key to my argument is my international statistics [37] (though he is wrong to say that this is all I have offered, since in addition I give a theoretical account of the causal mechanism); but he has offered no statistical correlations at all. The international data showing a correlation between handgun ownership and handgun homicide rates remains a thorn in the side of those who deny my causal hypothesis. Only by producing an alternative account of this correlation, and equally impressive statistical correlations between the putative non-handgun causes and handgun homicide rates, can they remove it.
Kopel ends this section with a reductio ad absurdum of my "analytic technique." [38] He argues that my own pattern of reasoning would justify the conclusion that tough handgun restrictions cause higher homicide rates. He lists six United States cities whose ranking in terms of severity of handgun laws coincides with their ranking in terms of homicide rate. [39] To show that this correlation is a causal one, he applies his version of my adequacy test. First, no other variable can plausibly be claimed to be the only cause of handgun homicide rates in these cities. [40] Second, he offers what he considers a common sense theoretical explanation of the causal mechanism at work: assuming people to be roughly similar in these cities, we would expect those cities with higher rates of handgun ownership to have lower rates of handgun murder (because of the deterrent effect of handguns). [41] While Kopel may support the premise that handguns reduce homicides, he certainly does not support this argument. His point is that it is an absurd argument which shows the absurdity of my allegedly similar causal argument.
It does not, however, because my argument pattern would not condone Kopel's argument. I have already disavowed the slip in which I stated that the first step in proving a causal relationship is showing that no other cause is the only cause. The test which I stated before this slip, and which I now reiterate, is showing that no other cause correlates so well as to rule out the putative cause. It is [Page 371] not clear that Kopel's hypothesis would pass this test, i.e., that no other variable(s) correlates so well with the varying handgun homicide rates in these cities that it displaces Kopel's causal explanation in terms of strict gun laws. Socio-economic factors of the kind which Kopel cites, and which I have never disputed, are likely responsible for the different murder rates. However, I will grant for the sake of argument that his hypothesis does pass this test.
Kopel's mock argument fails when the second test, which calls for a "probable theoretical explanation," is applied. [42] He has certainly given a theoretical account, namely that "handguns in the hands of non-criminal persons make murder less likely, by deterring criminal attack." [43] However, it fails to be plausible, much less probable, because there exists a far more probable theoretical explanation of the correlation he examines. Strict gun laws are often introduced in response to historically high levels of violent crime, including handgun homicide, and this is why some cities with high homicide rates have strict gun laws. By reversing the direction of causation, the mock argument thus commits the false cause fallacy.
The biggest weakness of all in Kopel's attempted reductio is the flimsy nature of his data itself. Consider the following table showing homicide rates per 100,000 in these cities since 1960. [44] The homicide rates of the second, third, and fourth cities in his list--New York City, Los Angeles, and Memphis--have been remarkably similar since at least 1960. The "lead" has changed several times among these three cities. Denver's has always been somewhat lower, and Salt Lake City's considerably lower. This data is simply too inconclusive to justify the drawing of conclusions based on the rank ordering of homicide rates in 1991.
Year
District/City
D.C.
N.Y.
L.A.
Memphis
Denver
Salt Lake City
1960
10.7
4.0
4.4
7.8
3.9
1.2
1965
18.6
6.1
6.1
6.2
4.1
1.2
1970
29.5
10.5
9.4
13.0
8.3
4.1
1975
33.0
18.0
14.3
15.5
8.1
2.9
1980
31.3
21.0
23.3
19.5
9.4
4.7
1985
23.5
19.2
24.4
18.6
14.0
8.5
1991
80.6
29.3
28.9
27.3
18.8
8.5
[Page 372]
Moreover, examination of the recent homicide rate in Washington D.C., the city with the highest 1991 homicide rate in Kopel's list, is disastrous for his mock hypothesis that stricter gun laws cause more homicide. In the years 1960-75, before D.C.'s 1977 handgun ban, its homicide rate was substantially higher than all the other cities, usually as least twice as high as that of its closest competitor. This strongly suggests that D.C.'s strict gun laws are a response to high homicide rates, rather than the cause of these rates, thus confirming my claim that Kopel's hypothesis reverses the actual order of causation. Moreover, after the 1977 ban, D.C. was the only city to exhibit a decrease in homicide rate, both between 1975 and 1980, and between 1975 and 1985. Consequently, a comparison between D.C. statistics and those of the other five cities provides a small degree of support for the hypothesis that gun control reduces murder, and zero support for Kopel's mock hypothesis that gun control increases it.
In contrast, the international data on which I base my argument shows an enormous disparity between the United States and the comparison countries in both handgun ownership and handgun homicide rates. Examination of the other Interpol data made available for the decade 1980-90 reveals that this huge disparity has been stable, and was not confined to the 1988 figures which I cite in my first paper. Another advantage of looking at the averages of these five sets of data, rather than just the 1988 figures, is that they exhibit a perfect correlation between the rank orderings in terms of the handgun ownership and handgun homicide rates per 100,000 people. [45]
Country
1991 HANDGUN OWNERSHIP AND HOMICIDE RATES
Ownership
1980
1983
1985
1988
1990
Average
U.S.A.
22,696
4.60
3.60
3.23
3.56
4.22
3.84
Israel
3,716
0.50
N/A
0.39
0.54
N/A
0.48
Sweden
3,700
0.22
0.08
N/A
0.23
0.16
0.17
Canada
2,301
0.03
0.02
0.02
0.03
0.26
0.07
Australia
1,596
0.02
0.06
0.03
0.07
0.06
0.05
U.K.
837
0.01
0.01
0.01
0.01
0.04
0.02
[Page 373]
I recognize that Kopel's "six cities" hypothesis is not intended seriously, but is advanced to show the alleged weakness of my argument. However, I have gone to some trouble to point out the weaknesses of his hypothesis in order to show just how different it is from mine. The fact that a superficially similar argument with flimsy empirical support and highly implausible theoretical explanations can be found in no way invalidates my argument. The details--the quality of evidence and theoretical explanations-- matter; and my argument must be assessed in the light of the details, not by overlooking them.
Kopel is right that both my causal hypothesis and his mock hypothesis are based on a limited data set for which there are many potential confounding variables. [46] However, his conclusion that "any claim to proof based on the data is hopelessly premature" [47] is itself hopelessly premature. The fact that no conclusive proof of a causal hypothesis can be produced from such data does not entail that all interpretations are equally plausible. Only in mathematics and logic can deductively valid proofs be given. In empirical science, the goal is to support hypotheses to the greatest degree of probability that the evidence permits. [48] I have shown that Kopel's mock hypothesis is far less probable than alternative explanations of the data; while my hypothesis that high handgun ownership rates contribute to high handgun homicide rates is immensely more probable than the alternatives suggested by Kopel and others. The burden of proof--as yet [Page 374] unmet-- remains on him to provide a better causal explanation of my data.
D. Seattle/Vancouver Study
In support of my thesis, I discuss a comparison of homicide and other violent crime rates in Seattle, and Vancouver, Canada. [49] Kopel raises two objections to my discussion. [50]
His first criticism is based on evidence that Vancouver's handgun homicide rate did not drop after stricter handgun controls were implemented in 1978. If the lower homicide rate in Vancouver were caused by its tougher gun laws, reasons Kopel, a drop would be expected after 1978. [51] This does little to undercut the study's findings, for two reasons. First, Vancouver's handgun homicide rate was low to begin with, so no noticeable change was to be expected. [52] Second, and most important, my hypothesis is that handgun ownership rates, not handgun laws, are a major determinant of handgun homicide rates. Seattle has a substantially higher ownership rate than Vancouver. [53] In a country in which the demand for handguns is relatively low, [54] the passage of more restrictive laws is unlikely to effect a dramatic drop in handgun ownership, especially in the short term.
Kopel's other main objection to the Seattle-Vancouver study is a reiteration of Wright's argument that racial factors, and not handgun density, explain the difference in homicide rate. [55] Kopel points out that the higher homicide rate in Seattle is confined to nonwhite racial groups. [56] My response to Wright was that the higher homicide rate among nonwhites in Seattle is most plausibly explained by the higher handgun density. [57] Kopel replies that the differences between the makeup and history of the nonwhite populations in the [Page 375] two cities preclude any comparison, and hence invalidate my argument that the differences in homicide rate are due to handguns. [58]
Kopel's response repeats errors which recur among opponents of handgun prohibition. First, he applies an unreasonably strict standard of proof: unless the racial minorities in both cities have an identical history and constitute a similar percentage of the overall population, he allows no comparison. However, the existence of multiple factors does not invalidate causal arguments based on one of those factors, provided, as is the case here, they are supported by clear statistical patterns.
Second, he has once again failed to address the burden of proof which I set for my opponents. The Seattle/Vancouver study gives a documented correlation between gun ownership and homicide rates, especially among racial minorities. It is backed by a theory explaining why Seattle's higher gun ownership rate is a major cause of its higher homicide rate. In contrast, Kopel's response amounts to the assertion that the real cause of Seattle's higher homicide rate may be the differences between the makeup and history of the nonwhite populations in the two cities. He provides no correlations between these differences and homicide rates, and no theoretical explanation of why these alleged non-gun causes affect homicide rates.
Kopel criticizes my citation of Mundt's finding that "the homicide rate among whites alone is almost three times higher in the United States than in Canada." [59] He notes that the American homicide statistics include Hispanics in the category "white", and that there is a much higher percentage of Hispanics in the United States than in Canada. He concludes that it is possible that the homicide rate for Caucasian whites is the same in both countries. [60] However, Kopel fails to substantiate this possibility. He points out the similar rate of overall shooting deaths (including accidents and suicides) among non-Hispanic whites in the United States and Canada. [61] However, this does not show that homicide rates are similar.
Finally, Kopel suggests that "directly addressing the problems" of racial minorities "might be a more efficacious way to reduce handgun homicide than to implement across-the-board handgun [Page 376] prohibition." [62] I advocate both addressing the problems of minorities and banning handguns, which are themselves one of the major problems faced by minorities in this country.
II. OBJECTIONS TO HANDGUN PROHIBITION
Having criticized the arguments I adduce in favor of handgun prohibition, Kopel devotes the rest of his response to presenting data and arguments opposing a handgun ban.
A. Switzerland
Kopel's discussion of the low overall homicide rate in Switzerland, despite its relatively high handgun homicide rate [63], is irrelevant to my thesis. Only in the case of the United States do I assert a connection between overall homicide and handgun homicide rates. [64]
B. Interstate Data
In my original article, I admit that interstate data does not help my hypothesis, since some states with strict gun control have higher violent crime rates than some with looser controls. [65] Kopel accurately explains my two main strategies in dealing with this recalcitrant evidence. [66] (1) Strict gun controls are often a response to crime levels which are already high. (2) The ease of "importing" guns from cities and states with lenient gun laws "masks" any beneficial effect of strict handgun controls in cities and states.
In opposition, he argues that even if interstate smuggling did occur, it cannot account for the increase in Washington, D.C.'s homicide rate after its 1977 handgun ban. [67] However, examination of homicide rates for 1975 and 1980 reveals a very different picture. Whereas the other five cities from Kopel's mock argument earlier in his paper [68] all exhibited an increase in homicide between 1975 and 1980, and the nationwide homicide rate rose from 9.6 to 10.2 per [Page 377] 100,000, [69] Washington D.C.'s rate fell from 33.0 to 31.3. [70] Subsequently, except in 1981, its homicide rate never exceeded the 1975 level until the drug wars caused it to soar beginning in 1987. [71] Consequently, Kopel's assertion that the capital city's handgun ban performed "dismally" is groundless.
Conceding that interstate smuggling may undermine any crime reductive effect of local handgun controls, Kopel tries a different line of attack to criticize the federal handgun ban I propose to overcome the problem of such smuggling. He argues that the very effectiveness of a federal ban would create a demand for illegally made guns. Worse still, the ensuing black market would create an opportunity for organized crime, just as prohibition did decades ago. There is also the problem of imported guns. [72]
Kopel's concerns, while speculative, certainly deserve the serious attention of proponents of a handgun ban. However, if the social benefits of a handgun ban are as great as I have claimed, the possibility of adverse side effects is not a strong objection. Kopel's concerns call for strategies to combat black market guns and other problems, not the abandonment of the effort to prohibit handguns.
C. Analogy with War on Drugs
Kopel is against the Reagan-Bush "war on drugs," and he uses alleged similarities between this war and my proposal for handgun prohibition as an objection to my view. In particular, he is concerned about the violations of civil liberties that have accompanied the war, and he fears that the same will occur if we prohibit handguns. [73]
Kopel's projection that the violations of civil liberties involved in the war on drugs will also accompany a handgun prohibition is of course speculative, and is rendered no less speculative by the fact that he cites other people who make the same projection. Nonetheless, no proponent of handgun prohibition can afford to ignore the dangers he describes.
However, it is not a strong objection to handgun prohibition. Just as one can oppose the methods used in the war on drugs without supporting total decriminalization of drugs, one can support handgun prohibition without condoning violations of individual rights. Kopel's arguments illustrate John Stuart Mill's comment that "there is no difficulty in proving any ethical standard [or in this case, proposed [Page 378] legislation] to work ill if we suppose universal idiocy to be conjoined with it." [74] I propose not an idiotic, but an intelligent implementation of a handgun ban. Instead of the draconian measures used to enforce our overly broad drug controls, I advocate a prohibition targeted specifically at handguns, and enforced in a way that respects civil liberties. Handgun "buyback" schemes have been successfully used throughout Canada and by several cities in the United States, and could be combined with a temporary "amnesty" on ownership of handguns which would be rendered illegal by my prohibition. Any handgun used in crime would be confiscated. Searches for handguns would only be allowed when, consistent with the Fourth Amendment, there is probable cause. By far the greatest benefit of a handgun ban would raise no Fourth Amendment concerns at all: the substantial arsenal of handguns purchased every day at gun stores would be stemmed.
D. Criminals' Access to Handguns
Taking up an objection to handgun prohibition which I discuss in my original article [75], Kopel argues that even a general prohibition would have little effect in keeping handguns out of the hands of those most likely to misuse them. [76] He states that one quarter of murderers have a criminal record [77] and even the non-convicts who murder tend to be a shady bunch, with prior brushes with the law such as arrests and domestic violence calls. [78] They are likely to have continued access to illegal black market guns even after a handgun ban.
Kopel's data is subject to a minor limitation. The figure that one quarter of murderers have prior felony convictions is an estimate made by opponents of gun control, which I granted for the sake of argument. [79] The actual percentage may be lower.
More important is that Kopel's argument shares a crucial weakness with his discussion of the danger of violations of civil liberties. Just as he showed that a ban which respects civil liberties will never achieve the complete elimination of handguns, his discussion in this section shows that some wrongdoers will still get their hands on handguns even after a ban. However, I have never claimed that a complete elimination of the private ownership of handguns is feasible. [Page 379]
My argument all along has been that a ban will reduce the number of handguns available for crime, by ending legal sales, and gradually reducing the pool available for black market resale and theft. A handgun ban does not need to be perfectly effective to lead to a major decrease in handgun violence.
Finally, Kopel's arguments that malfeasants will always be able to obtain handguns undermine the position of other opponents of handgun prohibition, such as Don Kates and Gary Kleck, who support a restriction targeted at those with criminal records. [80] Kopel himself does not explain what handgun restrictions, if any, he would support.
E. Substitution Theory
Despite his energetic attempts to deny my argument that the high handgun ownership rate in the United States is a cause of its high handgun homicide rate, [81] Kopel elsewhere concedes that "lower handgun density may, arguably, be associated with lower numbers of handgun homicides." [82] Perhaps his point is that, while he believes that I have failed to prove this causal connection, it may well exist anyway. However, he insists that lower ownership rates would not necessarily reduce the overall homicide rate. [83] In so doing, he rests his case squarely on the substitution theory that any drop in handgun homicide rate would be outweighed by an increase in homicide by other means. [84]
Since substitution theory is put forward by opponents of handgun prohibition, the burden of proof is on them to support it. They have persistently failed to meet this burden, offering only speculations based on surveys. [85]
Kopel's first defense of substitution theory begins with the fact that Switzerland has far more lenient handgun laws and a far higher handgun homicide rate than Australia and Canada. [86] However, he points out that Switzerland's overall homicide rate is less than half that of the other countries. [87] Kopel argues that the reason for the higher overall homicide rate in Australia and Canada might be that assailants in those countries turned to more easily available, and [Page 380] more lethal, long guns. [88]
Of course it might be! But since Kopel offers no evidence to support this speculation, not even long gun homicide rates in these countries, it can hardly be considered as a vindication of substitution theory.
In my original paper, I attack substitution theory on the grounds that (1) the only evidence offered that the level of substitution necessary to increase the homicide rate would occur is prisoner surveys [89]; and (2) the difficulty of concealing either regular or sawed-off long guns makes widespread substitution unlikely. [90]
Kopel's reply to my first objection is that prisoners might not be as unreliable as I make them out to be. [91] Even if he is right, whether they are representative of handgun owners in general is doubtful. Moreover, even if they are representative, responses to surveys are flimsy evidence indeed for predicting actual behavior, especially when deciding a social policy issue of such great importance.
He then tries to show that the relative inconvenience of long guns for crime does not disprove substitution theory. [92] He grants that, when both handguns and long guns are available, criminals prefer to use handguns. But if they are banned, they are likely to turn to the next best thing: long guns.
First of all, this amounts to a restatement of substitution theory, rather than a new argument for it. One of my original objections was precisely that the relative inconvenience of using long guns will cause many potential violent criminals to simply change their careers if handguns are banned. [93] But let us assume that some people are so hell bent on violent crime that they will upgrade to long guns (perhaps sawed-off) if handguns are prohibited. We can even assume for the sake of argument that enough of these people would cross the 30-40% threshold of substituting long guns for handguns that would supposedly increase the overall homicide rate. [94]
This is where a deeper weakness in substitution theory, which I did not address in my first article, becomes relevant. The substitution threshold percentage cited by Kopel is so often invoked by opponents of handgun prohibition that it has acquired the air of indisputable fact. It is time to question the threshold percentage, which is actually an estimate by Benenson and Kates based on the greater lethality [Page 381] of long gun wounds as compared to those inflicted by handguns. [95] If homicide were like target practice, long guns would clearly be the weapon of choice. In reality, of course, potential homicide victims are not like inanimate targets at shooting ranges. Both potential victims and passers-by are likely to realize the danger and take evasive action as soon as they see the long gun, whereas a handgun can be concealed literally until the second it is fired. In other words, the absence of the factors which cause handguns to be the weapon of choice of violent criminals--especially concealability and ease of "rapid draw"--is likely to minimize the success of long gun attacks. The greater lethality of long gun wounds must be weighed against the greater difficulty in inflicting the wounds in the first place. Consequently, the threshold percentage which is assumed by proponents of substitution theory is groundless.
Kopel concedes that the difficulty of concealing long guns will prevent some shootings. However, he disputes my claim that the greater concealability of handguns facilitates even shootings in the home. [96] He accuses me of making unwarranted assumptions that long guns will be kept unloaded in another room from the one where the shooting occurs, that handguns will be loaded and carried on the body of the shooter, and that victims will be unaware that the murderer is carrying a handgun. [97] Actually, I make no assumption that these contingencies are always the case, and I merely state, in response to Kates and Beneson, that "the 'surprise factor' peculiar to handguns can still apply even in the home." [98]
As I have said before, substitution theory is advanced by opponents of handgun prohibition. The burden remains on Kopel and other proponents of the theory to prove it, by showing that the use of long guns in the home and elsewhere would be sufficient to surpass the substitution threshold percentage. Even if he were to find such evidence--and the surveys he cites do not suffice--the objection I have raised to the threshold percentage itself indicates a fatal weakness in Kopel's defense of substitution theory. Given the centrality of substitution theory to Kopel's case, [99] his failure to substantiate it is very damaging to his argument against handgun prohibition. [Page 382]
F. Defensive Uses of Handguns
1. General
Kopel now advances the objection that prohibition would remove a major benefit of handgun ownership: self-defense. I have already responded at some length to arguments for the defensive benefits of handguns. [100] First, this objection to handgun prohibition depends on evidence that their defensive efficacy outweighs their use in violent crime. The evidence produced by opponents of handgun prohibition is too flimsy, consisting of surveys of prisoners and gun owners, and projections based on a claim that government statistics on justifiable gun homicides are underestimated. [101] Second, unnecessary use of handguns in self-defense needs to be counted with the harmful uses that I want to diminish by prohibition. [102] Third, many of the crimes against which handguns are alleged to offer protection are themselves committed with handguns. [103] Widespread defensive gun ownership, then, is likely to lead to an endless spiral of gun proliferation, with assailants becoming even more heavily armed in order to overpower armed victims. [104]
Kopel first tries to defend the use of surveys as a source of accurate data. As in his discussion of substitution theory, he again suggests that prisoners may be more reliable than I allow. However successful his argument is, his conclusion that "one may nibble at the edges of the percentages without changing the core reality" [105] illicitly transforms responses to surveys into "realities" about the deterrent effect of guns.
Kopel concedes the force of some of my objections to Kleck's use of polls of gun owners, but responds that, for different reasons, Kleck's projections may underestimate the annual number of defensive uses of handguns. [106] The problem with Kopel's arguments about both prisoner and gun owner surveys is that they fail to meet the burden of proof that is encumbent on those who assert the defensive efficacy of handguns. The data consists of nothing more than projections from surveys and the force of Kopel's arguments is that there may be factors that offset my doubts about the accuracy of the [Page 383] data.
In my first paper, I point out that, even if we grant Kleck's estimate of 1,527 to 2,819 self-defensive killings with guns in 1980, it is still far short of the 13,650 recorded firearms homicides in the same year. [107] Contrary to Kopel's claim [108], I do not assume that self-defensive homicide is only justified if it prevents killing. To give Kleck the benefit of the doubt, I allowed that all 2,819 self-defensive killings (the upper limit of his projection) prevented homicide, which is the least controversial occasion for self-defensive killing. Whatever the reason for these alleged 2,819 killings, they are far fewer than the actual firearms homicides the same year.
Kopel correctly asserts that "if the prohibition reduced defensive homicides by 1,000, but reduced criminal homicides by only 200, society would be worse off rather than better." [109] Of course such a scenario is possible. But, ignoring the burden of proof he faces, Kopel has not given a single reason to believe that it would actually happen.
Kopel's final argument for the defensive use of handguns begins with Kleck's evidence that victims of violent crime who carry firearms fare better than those who are unarmed. My responses to Kleck were that (1) gun owners among the survey's respondents have a vested interest in exaggerating the efficacy of their guns, and (2) many of these violent crimes were themselves carried out with handguns. [110] Responding to my second argument, Kopel states that handguns may be necessary for self-defense, even if the attacker has no handgun. [111] However, Kopel gives no data on such uses of handguns. For his response to succeed, he needs to show that the benefits of handguns against attackers with and without handguns outweigh the reduction in attacks with handguns that would follow a prohibition.
In sum, evidence that a reduction in defensive benefits of handguns would outweigh the crime-reductive effects of a handgun ban would be a powerful objection to prohibition. Unfortunately, while Kopel has improved on the existing anti-ban literature, he has failed to overcome the objections I raised in my original paper. [Page 384]
2. Women
The crux of my response to arguments that handguns provide women with a useful form of self-defense is that (1) alternative forms of self-defense for women are sufficiently effective, and (2) the rare cases in which women will suffer injury or death because they are prevented from owning handguns by a prohibition will be outweighed by the reduction of handgun violence against women, and of handgun violence in general. [112]
In response to my first claim, Kopel suggests ways in which my alternative methods of self-defense may fail to protect women. Restraining orders may be ineffective, women may be attacked as they attempt to leave their abusive partners, and mace and martial arts are not infallible. [113]
I have two replies to Kopel. First, since handguns too may fail to protect women, [114] the possibility of failure is not a fatal objection to the alternatives I propose. Second, government protection can be improved. Public spending on policing can be increased, in order to better enforce restraining orders. This would help to alleviate the instances of the police's failure to protect women cited by Kopel. [115] And more funding can be provided for women's shelters, which offer women immediate and secret refuge from violent partners.
Kopel's suggestion that we reform standards of self-defense to remove their sexist bias [116] is well taken, but this can be done without allowing women to use handguns.
As for my second argument, Kopel admits the premise that few women today own handguns; but he still denies my conclusion. First, he asserts that I wrongly assume that a handgun ban will disarm everyone, whereas it is more likely to disarm female victims than male criminals. [117] However, Kopel gives no reason as to why a ban would disarm females more than male assaulters. Moreover, I make no assumption that a ban will disarm everyone. As I have already pointed out, "a handgun ban does not need to be perfectly effective to lead to a major decrease in handgun violence." [118] [Page 385]
Second, Kopel repeats his earlier point that, if a ban did succeed in disarming men and women, these women would lose the benefit of the equalizing effect of handguns. [119] However, this simply begs the question against my argument that handguns are not necessary for women to protect themselves. Moreover, it does not address my argument that any marginal loss in protection for the small number of women who own handguns would be outweighed by the reduction in offensive uses of handguns which a ban would accomplish. Kopel, like the advocates of self-defensive handgun ownership for women whom I discuss in my original paper, has failed to meet the burden of proof encumbent on those who advance this objection to handgun prohibition.
3. Gun Use in the Home
Kopel next addresses the Kellermann and Reay study which I cite as evidence that handguns do more harm than good when used in the home. [120] Their study showed that firearms accounted for 2.94 criminal homicides and accidental deaths for every defensive homicide in the home. [121] Reasonably enough, Kopel states that both the offensive abuse and defensive use of handguns is not confined to a body count of those killed. [122] Nonfatal gunshots also should be counted, both when they are used offensively, and when they are used to wound or frighten off criminals. Consequently, he uses a suggestion from Kellermann and Reay, and estimates that for every justifiable defensive homicide in Seattle during the study's duration, there were 71 murders, suicides, or wrongful woundings. [123] Using Kleck's estimates of defensive uses of handguns, Kopel estimates that in Seattle "there were about 114 fatal and non-fatal defensive uses of handguns for every 71 unlawful handgun fatalities or wounding, including suicides." [124] In other words, handguns were used more often defensively than in homicides, assaults, suicides, or accidents.
Kopel's response to my use of the Kellermann and Reay study is vulnerable to insuperable objections. First is the weakness of the data from Kleck, on which Kopel's argument depends. The number of 645,000 annual defensive handgun uses is cited so often by opponents of handgun prohibition such as Kopel that, by sheer repetition, [Page 386] it may take on the appearance of hard empirical data. But, as explained at length in both my original paper and in this response, it is no more than an estimate based on highly suspect surveys of gun owners. [125]
Second, Kopel's ratio of 114 defensive to 71 "wrongful" uses of handguns underestimates the wrongful uses. Kleck's estimate of defensive uses is based on answers to the question of whether a handgun was used self- defensively, even if it was not fired. [126] In contrast, Kopel's estimate of wrongful uses is an estimate of the actual killings or woundings in the home. Not included are the times a gun was used to perpetrate a crime (e.g. a robbery) without being fired, or at least without wounding anyone.
Third, yet another source of inflation in Kleck's estimate of defensive uses of handguns is revealed by his further estimate that 95% of these uses "do not involve firing the gun." [127] Waving the gun may have been sufficient to deter the would-be assailant, but was it necessary? Could the danger have been warded off by simply walking or running away (remember that Kleck's data is not confined to handgun use in the home), or by calling the police? And was there a real danger in the first place? Several of these sources of inaccuracy are illustrated by Bernhard Goetz's drawing his handgun on a youth who asked him for money. [128] How many of the 95% of alleged gun uses which did not involve firing it were similarly gratuitous?
4. Handguns and MAD
I compare defensive handgun ownership with the Mutual Assured Destruction (MAD) rationale for possessing nuclear weapons, and criticize both on the ground that the very weapons that are hoped to prevent a "first strike" may themselves be used offensively. [129] In the case of handguns, the United States crime statistics which I cite throughout my paper indicate that this is exactly what is happening at an alarming rate. Kopel responds by (1) criticizing the analogy with MAD, on the ground that nuclear weapons are far more destructive than handguns, (2) defending MAD, and (3) suggesting that use of conventional weapons to deter attacks from other nations provides a better analogy with handgun ownership for defensive [Page 387] purposes. [130]
The premise of Kopel's first criticism is correct: there is a vast difference in scale between nuclear weapons and handguns. However, this difference does not invalidate my analogy, which compares the two situations in one respect only--the danger of defensive weapons being used offensively.
Kopel's second objection is his claim that "MAD worked." [131] However, he commits the post hoc, ergo propter hoc variant of the false cause fallacy. The fact that massive nuclear armaments coincided with peace between the superpowers and in Europe does not establish a causal relationship. Many other factors may have been responsible for peace, which may have prevailed in spite of, and not because of, nuclear proliferation. Kopel then compounds his error by committing a straw person fallacy. (i.e. attributing to an opponent a position which she does not hold, and then arguing against this misrepresented position, rather than her actual view.) He asks whether I would have preferred the United States to disarm far more of its nuclear weapons than the former Soviet Union, or even to unilaterally disarm itself. [132] Actually, I think that either of those options would have been better than the nuclear proliferation that actually occurred. However, I support neither of these options nor nuclear proliferation. I support the third alternative of massive bilateral reductions in both countries' nuclear arsenals. Such reductions are compatible with the importance of parity of armaments which Kopel asserts. [133]
Kopel's third response is to compare individual handgun ownership with the deterrent power of the possession of conventional weapons by a country in order to deter or resist external attack. He asks whether I oppose, for instance, the use of weapons by Americans to deter British hostilities at the time of the War of Independence. [134] The weakness of Kopel's argument is his false assumption that, because I oppose both handgun and nuclear proliferation, I oppose all defensive uses of weapons. In reality, I am not a pacifist, and I do support the existence of conventional armed forces to deter external attacks on the United States. I oppose nuclear proliferation because I believe it is not necessary to achieve deterrence, and because it creates such a danger of a nuclear holocaust. Similarly, I oppose private ownership of handguns both because they are not necessary to protect individuals against crime, and because they are used [Page 388] in so many violent crimes.
Kopel insists that handguns are necessary for self-defense, since the government has failed to adequately protect its citizens. [135] I have three responses. First, I have already explained how citizens can effectively defend themselves without using handguns. [136] Second, government protection can be improved by greater expenditures on policing, enforcement of restraining orders, women's shelters, and other methods. Third, one of the biggest reasons for the government's failure to provide better protection is the rampant handgun violence made possible by soaring handgun ownership rates. Proliferation of handguns will only undermine the government's ability to protect society.
G. More International and Historical Comparisons
Up to now, Kopel's paper has wisely focused on challenging the international data and the causal argument on which my case is based, and on furthering standard objections to handgun prohibition. The remainder of his paper, however, is based on data which does little to help his case, and in some instances positively damages it.
1. Burglary
Kopel compares homicide, rape, robbery and burglary rates in the United States and several other democracies. [137] Although the United States has a far higher homicide, rape, and robbery rate than any of the other countries, its burglary rate is one of the lowest. [138] Kopel concludes, reasonably, that a reason for the relatively low burglary rate, despite our overall propensity for violence, may be the deterrent effect on potential burglars of our high rate of gun ownership. [139]
However, the very high rates of violent crime in this country quoted by Kopel are embarrassing for opponents of handgun prohibition, since they are plausibly attributable to the far higher rate of handgun ownership in the United States than in the other countries. In 1990, 50% of its homicides were committed with handguns, and 36.6% of its robberies and 23.1% of its aggravated assaults involved firearms, most of which were handguns. [140] Any increase in burglary [Page 389] would be a bargain price to pay for the decrease in violent crime which would follow a handgun ban.
2. Frontier Communities
To further illustrate the crime deterrent value of widespread gun ownership, Kopel describes the libertarian paradise of nineteenth century American frontier towns. [141] Despite minimal law enforcement and almost universal gun ownership in Aurora, Bodie, and other towns, crimes such as burglary, robbery, and rape were rare. [142] One minor problem was that the homicide rate was "extremely high", and even "astronomical". [143] But not to worry: the people killing each other "at a fearsome rate" were almost all young "drunken males upholding their 'honor'," in " 'fair fights' that they voluntarily engaged in." [144]
I will respond on the assumption that Kopel is seriously advancing these frontier towns as a desirable role model for us. First, his cavalier attitude toward the massive homicide rate, and his belief that it was a worthwhile tradeoff for the low rates of other crime, shows more respect for property than human life.
Second, the fact that the high homicide rate was mostly attributable to young males is not a peculiarity of frontier days. Most of today's homicide victims and perpetrators are also young males. [145] Whether these murders of and by young men are the result of "fair fights" is questionable. That these murders, mostly with handguns, are tragic is certain.
Third, perhaps in the frontier days, macho gunfights were confined to their voluntary participants. In 1993, innocent bystanders in crowded cities are at far greater risk of being caught in the crossfire of voluntary handgun fights. [146] The danger of innocent passers-by being caught in the crossfire is, more generally, a serious limitation of the value of handgun use in self- defense.
Fourth, far more plausible explanations exist for the low rate of non- homicidal crimes than the almost universal gun ownership. The low rape rate is better explained by the small number of women, other than prostitutes, in these frontier towns. [147] The low rate of [Page 390] burglary and robbery is likely explicable by the fact that most people did not own much to steal, since the majority were "young transient males" [148] who were staking claims to land, or trying to find their fortune in gold mining.
3. Jamaica's Handgun Prohibition
The final international comparison made by Kopel is his discussion of Jamaica's prohibition of handguns enacted in 1974. [149] Kopel describes the widespread violations of civil liberties that occurred. Ironically, after falling for a year, violence steadily increased, and reached record highs in 1980. [150] If accurate, Kopel's account is a useful illustration of how handgun bans can be enacted in a draconian fashion, and still fail to reduce violence.
However, nothing in Jamaica's experience offers a serious objection to my proposal for a handgun ban in the United States. First, as I pointed out in my discussion of Kopel's comparison between handgun prohibition and the war on drugs, I propose an intelligent implementation of the prohibition which respects civil liberties.
The second, and most important, reason why Jamaica's experience does not invalidate my proposal for the United States arises from Kopel's admission that "there are immense differences between the United States and Jamaica." [151] By far the most pertinent of these differences is surprisingly neglected by Kopel's account of Jamaica's recent history. The period 1974-80, which immediately followed the handgun ban, was one of the most turbulent in Jamaica's history. Michael Manley's socialist government was trying to introduce egalitarian economic reforms, as well as pursue its status on the international scene as a non-aligned country. Manley's reforms met with strong opposition from conservative forces both abroad and in Jamaica, whose condition was not far removed from a civil war. The terrible violence in 1980 to which Kopel refers is attributable to the bloody election which occurred that year, and which claimed several hundred lives. [152]
This political turmoil is by far the most plausible explanation [Page 391] of the increasing rate of violence in Jamaica after its handgun ban. To draw any general conclusions from these events about the alleged ineffectiveness or dangers of handgun bans is, therefore, utterly unwarranted.
III. CONCLUSION
Neither Kopel's data nor his arguments have seriously undermined the case for handgun prohibition I advanced in my original paper. Like other opponents of a handgun ban, he has failed to meet the burden of proof encumbent on those who challenge my argument. His objections to my hypothesis that a handgun ban would reduce homicide and other violent crime in this country show only that it might be false, which is a possibility I have conceded all along. Like any hypothesis in the social sciences, mine can only be shown to be probably true; and this is exactly what I have done. To respond, Kopel needs to show that it is probably false; and this is exactly what he has failed to do. Too much of his data consists of sociological and psychological generalizations, tenuous historical analogies, unsubstantiated predictions, and projections based on already- suspect surveys.
In contrast, my argument for a handgun ban in the United States is based on actual handgun violence rates. The data itself, in particular the enormous disparity between this country and others in terms of handgun ownership and handgun homicide, is my strongest ally. Until opponents can produce a better explanation of my data, as well as better data of their own, their objections to handgun prohibition lack credibility.
[1]. David B. Kopel, Peril or Protection? The Risks and Benefits of Handgun Prohibition, 12 St. Louis U.Pub.L.Rev. 285 (1993).
[2]. Id. at 287-294.
[3]. Nicholas Dixon, Why We Should Ban Handguns in the United States, 12 St. Louis U.Pub.L.Rev. 243, 246-47 (1993).
[4]. Kopel, supra note 1, at 287.
[5]. Id. at 288-89.
[6]. Id. at 289.
[7]. Id. at 288.
[8]. Id. at 291-294.
[9]. Id. at 294-300.
[10]. Dixon, supra note 3, at 245.
[11]. CRIME IN THE UNITED STATES: UNIFORM CRIME REPORTS (1970-91).
[12]. Dixon, supra note 3, at 245.
[13]. Id. section II.D at 268-272.
[14]. While my international comparison focuses on handgun homicides, throughout my paper I also offer a reduction in other violent crimes, (such as aggravated assault and robbery), as a reason for a handgun ban.
[15]. Compare handgun homicide rates cited by Dixon, supra note 3, at 248- 249, with overall homicide rates cited by DON B. KATES, JR., GUNS, MURDERS, AND THE CONSTITUTION: A REALISTIC ASSESSMENT OF GUN CONTROL 42 (1990).
[16]. See Dixon, supra note 3, at 244-45.
[17]. See KATES, supra note 15, at 42.
[18]. Kopel, supra note 1, at 300-310.
[19]. Dixon, supra note 3, at 252.
[20]. Kopel, supra note 1, at 301.
[21]. Dixon, supra note 3, at 252.
[22]. Id.
[23]. Kopel, supra note 1, at 301.
[24]. Id. at 302-303.
[25]. Id. at 304.
[26]. Failure to recognize the connection between the different elements of my causal argument also underlies Kopel's mistaken claim that my first adequacy test makes a fallacious appeal to ignorance. Id. at note 62. Only if I were to offer the absence of better variables as the sole premise would my causal argument be an appeal to ignorance, but I do not do so. The heart of my argument is the striking correlation between handgun ownership and handgun homicide rates, along with my theoretical explanation of the causal mechanism involved.
[27]. JOHN STUART MILL, A SYSTEM OF LOGIC, Book III., ch. VIII (1843).
[28]. See, e.g., his discussion of the defensive value of handguns, especially for women. Kopel, supra note 1, at 332-349.
[29]. Dixon, supra note 3, at 253.
[30]. Kopel, supra note 1, at 304-307.
[31]. Id. at 307.
[32]. DAVID B. KOPEL, THE SAMURAI, THE MOUNTIE, AND THE COWBOY (1992).
[33]. In the United States, 50% of homicides are with handguns. Based on the sources cited supra note 15, here are the percentages of total homicides that are committed with handguns in the comparison countries.
Australia: 3.6%
Canada: 1.2%
Israel: 27.1%
Switzerland: 72.5%
Britain: 1.8%
[34]. Kopel, supra note 1, at 308.[35]. Id.
[36]. Dixon, supra note 3, at 253-54.
[37]. Kopel, supra note 1, at 308.
[38]. Id. at 308-310.
[39]. Id. at 308.
[40]. Id. at 309.
[41]. Id.
[42]. Dixon, supra note 3, at 252.
[43]. Kopel, supra note 1, at 309.
[44]. CRIME IN THE UNITED STATES: UNIFORM CRIME REPORTS (1960, 1965, 1970, 1975, 1980, 1985, 1991).
[45]. The homicide rates for some countries are likely to be slightly underestimated for the years 1980-88, since I have used the population figures from THE NEW ENCYCLOPEDIA BRITANNICA (1990). These countries probably had slightly lower populations in earlier years.
[46]. Kopel, supra note 1, at 310.
[47]. Id.
[48]. Dixon, supra note 3, at 253-54.
[49]. Id. at 258-62.
[50]. Kopel, supra note 1, at 311-13.
[51]. Id. at 312.
[52]. The firearms homicide rate in Vancouver from 1980-86 was 1 per 100,000. See Sloan et al, Handgun Regulations, Crime, Assaults, and Homicide: A Tale of Two Cities, 319 NEW ENG.J.MED. 1259 (1988). Based on Sloan's data, this translates into 3.59 handgun homicides per year. Since Kopel tells us that the handgun homicide rate did not change after the law was enacted, we may assume that it was similarly low to begin with. The relative risk of being murdered with a firearm was five times higher in Seattle than in Vancouver during the study period. Dixon, supra note 3, at 258.
[53]. Dixon, supra note 3, at 259.
[54]. See id. at 249 for Canada's handgun ownership rate.
[55]. Dixon, supra note 3, at 260.
[56]. Kopel, supra note 1, at 312.
[57]. Dixon, supra note 3, at 261.
[58]. Kopel, supra note 1, at 313.
[59]. Dixon, supra note 3, at 261.
[60]. Kopel, supra note 1, at 314, note 111.
[61]. Id. at 313. The only other evidence he offers is Mundt's finding that the difference in crime rate between some U.S. and Canadian cities disappears when covariates of race and size are considered. Kopel concedes that another of Mundt's studies was unable to account for U.S.-Canadian crime differences by reference to only racial factors. Id. at 314.
[62]. Id. at 314-15.
[63]. Id. at 310-11.
[64]. See supra section I.B.
[65]. Dixon, supra note 3, at 262. Kopel misleadingly says that "states with stricter gun laws have higher crime rates," implying that this applies to all states. Kopel, supra note 1, at 315. In reality, all that is claimed by opponents of handgun prohibition is that interstate data fails to demonstrate that strict controls deter crime. Dixon, supra note 3, at 262, n. 63.
[66]. Kopel, supra note 1, at 315.
[67]. Id. at 316.
[68]. Id. at 308-310.
[69]. UNIFORM CRIME REPORTS (1975, 1980).
[70]. See supra text at 371 for table.
[71]. UNIFORM CRIME REPORTS (1981-89).
[72]. Kopel, supra note 1, at 317-19.
[73]. Id. at 319-323.
[74]. JOHN STUART MILL, UTILITARIANISM, (1861).
[75]. Dixon, supra note 3, at 264-268.
[76]. Kopel, supra note 1, at 324-326.
[77]. Id. at 324.
[78]. Id.
[79]. Dixon, supra note 3, at 266.
[80]. See id. at 265, N. 76.
[81]. For my response to his objections to my central causal argument, see supra text section I.C.
[82]. Kopel, supra note 1, at 300. See also his conclusion, id. at 359.
[83]. Id. at 300.
[84]. For Kopel's defense of substitution theory, see id. at 326-332.
[85]. See Dixon, supra note 3, at section II.D.
[86]. Kopel, supra note 1, at 327-328.
[87]. Id. at 328.
[88]. Id.
[89]. Dixon, supra note 3, at 270.
[90]. Id. at 270-71.
[91]. Kopel, supra note 1, at 329.
[92]. Id. at 330-331.
[93]. Dixon, supra note 3, at 270.
[94]. Kopel, supra note 1, at 329.
[95]. Mark Benenson and Don Kates, Jr., Handgun Prohibition and Homicide: A Plausible Theory Meets the Intractable Facts, in RESTRICTING HANDGUNS: THE LIBERAL SKEPTICS SPEAK OUT 110-11 (Don Kates, Jr. ed., 1979).
[96]. Dixon, supra note 3, at 271.
[97]. Kopel, supra note 1, at 331-32.
[98]. Dixon, supra note 3, at 271.
[99]. See supra text at 379.
[100]. Dixon, supra note 3, section II.E.
[101]. Id. at 273-275.
[102]. Id. at 274.
[103]. Id. at 275.
[104]. Id.
[105]. Kopel, supra note 1, at 333.
[106]. Id. at 333-36.
[107]. Dixon, supra note 3, at 274.
[108]. Kopel, supra note 1, at 337.
[109]. Id. at 337.
[110]. Dixon, supra note 3, at 274-275.
[111]. Kopel, supra note 1, at 337-38.
[112]. Dixon, supra note 3, at 281-82.
[113]. Kopel, supra note 1, at 339-340.
[114]. See Dixon, supra note 3, at 275-77 for a discussion of the dangers of guns kept in the home.
[115]. Kopel, supra note 1, at 339-340 n. 218.
[116]. Id. at 338-339 n. 214.
[117]. Id. at 341.
[118]. See supra text at 378.
[119]. Kopel, supra note 1, at 341.
[120]. Dixon, supra note 3, at 275 n. 112 and accompanying text.
[121]. Id. at 276.
[122]. Kopel, supra note 1, at 342.
[123]. Id. at 343.
[124]. Id. at 343-44.
[125]. Dixon, supra note 3, at 273-74, and supra text at 382.
[126]. Kopel, supra note 1, at 335.
[127]. Kopel, supra note 1, at 344 n. 232.
[128]. Dixon, supra note 3, at 274.
[129]. Id. at 275.
[130]. Kopel, supra note 1, at 350-51.
[131]. Id. at 350.
[132]. Id.
[133]. Id. at 351.
[134]. Id.
[135]. Id.
[136]. Dixon, supra note 3, at 277, 281.
[137]. Kopel, supra note 1, at 345-347.
[138]. Id. at 345.
[139]. Id. at 347.
[140]. Dixon, supra note 3, at 244-45.
[141]. Kopel, supra note 1, at 347-349.
[142]. Id. at 347-348.
[143]. Id.
[144]. Id. at 348.
[145]. In 1991, 49% of homicide victims, and 66% of homicide offenders, were males aged 15-34. UNIFORM CRIME REPORTS 16 (1991).
[146]. In 1880, the United States' population was 50 million. By 1990, it was 249 million. STATISTICAL ABSTRACT OF THE UNITED STATES 8 (1992).
[147]. Kopel states that the population in Aurora and Bodie "was mainly young transient males," supra note 1, at 347.
[148]. Id.
[149]. Id. at 354-359.
[150]. Id. at 355-56.
[151]. Id. at 354.
[152]. For Manley's own account of this period, see MICHAEL MANLEY, STRUGGLE IN THE PERIPHERY (1982). For a conservative account of the same events, see JOHN D. FORBES, JAMAICA: MANAGING POLITICAL AND ECONOMIC CHANGE (1985). Both accounts make clear that, for purely political reasons, Jamaica was undergoing a violent upheaval.