Saint Louis University Public Law Review

Volume 12, 1993

Symposium: Violence, Crime and Punishment

*285 PERIL OR PROTECTION? THE RISKS AND BENEFITS OF HANDGUN PROHIBITION

David B. Kopel [a]

Copyright © 1993 by the Saint Louis University School of Law; David B. Kopel

[This article is a reply to another article appearing in the same issue. Following this article, Prof. Dixon's reply appears. The Dixon articles are not posted on this website, because they are copyrighted. The hard copy of the issue containing Dixon's articles can be found in law school law libraries. Dixon's opening article makes an empirical case in favor of handgun prohibition. The boldface asterisks (*301) indicate the start of a new page in the printed text. In the printed text, the notes are footnotes, rather than endnotes.]

Dixon's article represents a valuable and intellectually honest approach to the gun control debate. To begin with, he acknowledges that the Second Amendment may pose "strong" legal obstacles to his proposal to confiscate all handguns. [1] Taking many of the familiar *286 arguments for handgun prohibition, Dixon examines them carefully, acknowledging their limitations. Moreover, Dixon has read much of the empirical literature produced by skeptics of handgun prohibition, and he devotes much of the article to meeting the skeptics head-on. Finally, Dixon addresses himself to the issue at hand, and does not *287 engage in invective about handgun owners or the National Rifle Association. Thus, Dixon presents one of the more comprehensive and carefully- reasoned arguments for handgun prohibition that has ever been written.

Despite these virtues, Dixon's argument does not succeed. The blame lies not in his skill in presenting arguments, but in the facts themselves. The more deeply one looks into the issue, the more difficult it becomes to conclude that handgun prohibition will save lives. Prohibition might instead result in a significant increase in the deaths of innocents.

I. THE CASE FOR GUN PROHIBITION

A. Police Exemptions

Before addressing the merits of handgun prohibition, Dixon discusses what some advocates of handgun prohibition consider to be a fatal flaw in his proposal: allowing police officers and security guards to continue to possess handguns. After all, if society acknowledges that handguns have significant defensive value and can help save the lives of police officers and security guards, how can society deny that handguns can also help save the lives of other people?

Dixon replies with two arguments. First of all, "police are entrusted with the protection of society and the prevention and deterrence of crime, it is only to be expected and indeed encouraged that they be given superior force to the rest of society." [2] That police officers protect society, however, does not entitle them to own every type of weapon. No one would suggest that police officers should be allowed to carry hand-held nuclear weapons, if such weapons existed, nor would we want police officers routinely to carry nerve gas, surface to air missiles, grenades, mortars, or a score of other weapons. Thus, even if we assume that police officers are society's legitimate protectors in a way that ordinary persons are not, the police should not necessarily be authorized or entitled to possess handguns.

Perhaps what distinguishes handguns from nerve gas, in Dixon's view, is that the police "need" handguns in a way that they do not need nerve gas. Arguably, police officers, as protectors of society, are uniquely at risk, and thereby have a higher need for *288 defensive weaponry. In regards to officers who perform street patrol in high-crime neighborhoods, police officers face substantially higher risks than most other people. There are, however, many police officers who do work other than patrol of high-crime areas. Officers who process paper work, officers who direct traffic, command-rank officers who do not routinely engage in dangerous missions, and officers who do not patrol unusually dangerous neighborhoods would all be allowed, under the Dixon proposal, to possess handguns. While these officers face certain risks of attack, these risks are no greater than, and sometimes substantially less than, the risks faced by people whom Dixon would disarm, including the owners of businesses such as gas stations or grocery stores in robbery-prone neighborhoods, women who are being threatened by ex-boyfriends, crime witnesses and jurors who are at risk of retaliation from violent criminals, and elderly people who have to walk though dangerous neighborhoods. A police sergeant who sits in a fortified building, in which every entrance is protected by several heavily-armed police officers, is at much less personal risk of attack than is a clerk who works the night-shift at a convenience store on the wrong side of town. Except for police officers assigned to dangerous duty, need cannot justify Dixon's proposal to allow the police to possess handguns if all civilians are forbidden them.

A second argument Dixon offers in support of a comprehensive police exemption is that "[s]ince they are subject to extensive training and strict discipline, police are less likely to abuse handguns than the private citizen ..." [3] The empirical evidence, however, suggests that police misuse of firearms is quite common. [4] Whenever a New York City police officer fires a gun (outside of a target range), police officials review the incident. About 20% of discharges have been determined to be accidental, and another 10% to be intentional discharges in violation of force policy. In other words, only 70% of firearms discharges by police are intentional and in compliance with force policy. [5] According to another study, when police *289 shoot at criminals, they are 5.5 times more likely to hit an innocent person than are civilian shooters. [6]

Many police officers work difficult, stressful jobs for many years. Ordinary citizens, if they find themselves under stress, can simply retreat back to their houses or apartments. Since Dixon argues that too many ordinary, non- criminal citizens lack the emotional stability to be trusted with handguns, how can handgun possession be defended for a group of people who are under significantly higher emotional stress than ordinary people? Not only are police misuses of firearms in the line of duty far from uncommon, police misuse of guns outside the line of duty is all too frequent. When an off-duty New York City policeman fires a gun, one time out of four the firing will be an accident, a suicide, or an act of frustration. [7] The rate of substantiated crimes perpetrated by New York City police officers is approximately 7.5 crimes per year per thousand officers. The number of New York police criminal incidents alleged is 112.7. [8]

Further, Dixon is wrong to claim that police officers receive "extensive training" in use of handguns. More typically, they receive a few dozen hours of training at the police academy, and may be, at most, required every few years to recertify their ability to hit a target. A deplorably large number of handgun-toting officers have not practiced marksmanship since they passed their firearms certification test as a police recruit. The amount of training which police officers have in defensive gun use rarely exceeds what a civilian could learn at a good firearms instruction academy. With the advent of inexpensive *290 indoor laser target systems and high-technology video trainers for "shoot-don't shoot" programs, and the proliferation of civilian firearms schools, citizens willing to invest some time can be schooled in defensive firearms use to at least the same level of competence as the average police officer.

Accordingly, Dixon does not articulate a good reason why police in general should be exempted from his general firearms ban, except for police who patrol high-crime areas, who have unusually high self-defense needs. A fortiori, the proposed exemption for security guards, even rigorously trained ones, lacks a logical basis. If training and licensing make security guards capable of being trusted with guns, then other persons who pass equally rigorous training and licensing systems may also be trusted with guns. As Dixon notes, security guards can be "visible targets for attack," [9] but so are proprietors of "convenience stores, gas stations, and other small businesses," [10] at least in dangerous neighborhoods. If the owners of these stores pass through a licensing and training system equivalent to that of security guards and/or police, there is no basis for denying these persons a permit. To structure the gun permit system so wealthy owners of jewelry stores can hire security guards for protection, but low-income owners of convenience stores, who cannot afford a security guard, are deprived of protection - even though the convenience store owner is as objectively qualified as a security guard to carry a gun - is economic discrimination, and amounts to valuing the property of the jewelry store owner more highly than the life of the convenience store owner. [11]

*291 The overbroad police/security exemption might doom Dixon's gun prohibition scheme from the start. As Justice Brandeis observed, "Government is the great teacher," [12] and the democratic nations which have adopted strict handgun control systems that enjoy a high degree of voluntary civilian compliance, have in part succeeded through the example set by government. While the police have not been entirely disarmed of handguns, they are, in practice, far less oriented towards gun use than their American counterparts. For example, in Japan, which completely prohibits civilian handguns, the police do possess handguns, but in a manner very different from their American counterparts. The police only began carrying guns at the repeated insistence of General MacArthur's government of occupation. [13] Japanese police have merely .38 special revolvers, not the high-power or high-capacity .45 and 9 mm handguns often toted by the American police. [14] No officer would ever carry a second, smaller handgun as a back-up, as many American police do. Policeman may not add individual touches - such as pearl handles or unusual holsters - to dress up their gun. While American police are often required to carry guns while off-duty, and always granted the privilege if they wish (even when retired), Japanese police must always leave their guns at the station. Unlike in the United States, desk-*292 bound police administrators, traffic police, most plainclothes detectives, and even the riot police in Japan do not carry guns. [15]

The official Japanese police culture strongly discourages use or glamorization of guns. One poster on police walls orders: "Don't take it out of the holster, don't put your finger on the trigger, don't point it at people." [16] Shooting at a fleeing felon is unlawful under any circumstance. [17] In an average year, the entire Tokyo police force only fires a half-dozen or so shots. [18]

The police being disarmed, criminals reciprocate. Although guns are available on the black market, there is little use of guns in crime. The riot police leave their guns at the station; and the masses of angry students who confront the riot police also eschew modern weapons. The two sides instead study medieval military tactics, using mass formations of humans as battering rams or as shields. [19]

Comparative criminologist David Bayley, a proponent of stricter American gun controls, suggests that American police attitudes towards guns makes it impossible for gun control to be achieved. As long as the police are armed, writes Bayley, they send the implicit message that armed confrontations with civilians are the norm, and that shootings of police officers, while sad, are nothing extraordinary. [20]

In Britain as well, that the police are mostly disarmed is one of the important reasons why criminal and non-criminal civilians mostly avoid handguns. Even in London, only about 15 percent of the police carry guns. The sole police who are permanently armed are special security forces for "diplomatic, royalty and ministerial *293 protection." [21] The patrolmen who are given guns are more lightly armed than almost all of their American counterparts. Patrol officers do not routinely carry backup guns. The primary guns are Smith and Wesson Model 10 revolvers loaded with .38 special +P ammunition. They must be carried concealed. Bobbies have no powerful Colt .45s on display. They do not carry the same gun from day to day. At the end of every shift, their duty gun goes back into the police safe. [22] High- ranking police administrators almost never carry guns. [23] All officers must follow the policy in the Association of Chief Police Officers' Manual of Guidance. The guidance, although officially secret, is "very largely consistent with Colin Greenwood's Police Tactics in Armed Operations and emphasizes "extreme caution in police use of deadly force." [24] In 1987, only five Britons were shot dead by the police, and even this number was seen as alarmingly high. [25] Police authority has, historically, rested not upon the ability to compel submission, but "upon the benign, non-aggressive image of the unarmed British bobby." [26] The police authority has been achieved "by presenting an image of vulnerability, instead of the invincibility of their Continental or American counterparts." [27] An almost totally disarmed police has long been the British ideal, but starting with the 1967 Shephard's Bush murders (in which criminals with stolen revolvers shot three police officers), a growing minority of police officers have begun to patrol armed. As the police have become more heavily armed, so have criminals; armed robberies and police armament closely correlate. [28]

The British police are mostly unarmed, and the Japanese police hardly ever draw their guns. Few people in either nation own guns. The Canadian police are well-armed, and more likely to use their guns than their British or Japanese counterparts. Canadian police use of guns legitimizes gun use in general, and is one reason why Canadians choose to own so many more guns, including handguns, than do people in Britain or Japan.

While some of Canada's local peace officers are unarmed, the federal R.C.M.P. (some of whom act as provincial or municipal *294 police under contract from the Provincial Attorney General) all carry guns. [29] Most police cruisers carry a shotgun with buckshot loads in the trunk. [30] Although the Canadian police are well-armed, they do not use their firearms as frequently as their southern counterparts. In America, about a person a day is killed by the police. The Canadian per capita homicide by police rate is less than a third of the American rate. [31]

In sum, one of the reasons that severe handgun controls have been partially successful in reducing handgun possession in Japan, Great Britain, and Canada, is that the police in each nation have, to varying degrees, minimized possession and/or use of handguns themselves; the examples set by the police and government may be an important reason why criminal and non-criminal civilians in those nations have, to varying degrees, voluntarily foregone the use of handguns. Accordingly, Dixon's proposal to allow police and security guards a broad exemption from his handgun prohibition may seriously undermine his proposal's chance for success. At the least, his proposal would need to be accompanied by drastic restrictions on the numbers of police authorized to carry guns, as well as major changes in the practices of police who do carry guns.

B. Handgun Density and Handgun Homicide

Would handgun prohibition save lives? Dixon builds his argument primarily by comparing the United States with other nations. He begins by setting forth the estimated per capita handgun ownership rates, and the estimated per capita handgun homicide rates for the United States, Israel, Sweden, Canada, Australia, and Great Britain. With a few minor exceptions (which Dixon carefully notes), the countries with higher handgun density have higher handgun homicide rates. America has about 3.5 handgun homicides per 100,000 population; Israel has about 0.5 per 100,000 population, Sweden about 0.25, and Canada, Australia, and Great Britain all have less than 0.1.

Dixon assumes that any flaws in the data which underestimate or overestimate handgun density are roughly the same for each country. [32] The assumption may not be correct in case of Canada; there may have been a substantial underreporting of gun ownership by Canadians at the time the poll which he cites was taken, since the poll was conducted for the government while the government was *295 considering gun confiscation. [33]

In regards to the United States, there appears to be a distortion of a different nature. Of the countries Dixon discusses, Israel and the United States are the only ones in which use of deadly force in self-defense, and use of a gun for that purpose, is generally approved by the legal system. [34] As Dixon discusses later, Florida State University criminologist Gary Kleck suggests that 1,527 to 2,819 American homicides a year are actually justifiable homicides committed by citizens using a firearm to defend themselves or another person *296 against violent attack. [35] The FBI statistics showing lower self-defense numbers are flawed, Kleck says, because the statistics are based on arrests, rather than final disposition of cases. [36] Other countries, which may also report their homicide data based upon arrests rather than convictions, are less likely to substantially overstate their homicide rate by mistaken counting of justifiable homicides, because such homicides are relatively rare. If we take Kleck's high-end estimate of 2,819 and low-end estimate of 1,427 justifiable or excusable American homicides, and figure that about 70% of those homicides involved a handgun (73% of civilian justifiable homicides recorded by the FBI involve a handgun), [37] then between 999 (1,427 x .7) and 1,973 (2,819 x .7) justifiable or excusable homicides should be subtracted from Dixon's total of 8,634 American homicides involving handguns. As a result, the unlawful American handgun homicide rate per 100,000 population would be somewhere between 3.01 (low-end estimate of justifiable homicides) to 2.66 (high-end estimate of justifiable homicides). [38] Thus, even after self-defense handgun homicides *297 are properly counted, the American criminal handgun homicide rate is still substantially larger than any other country on Dixon's list, although the exclusion of self-defense homicides drops the total American handgun homicide rate significantly.

While the six countries that Dixon lists show a correlation between handgun density and handgun homicide, the correlation appears less firm when other western democracies are considered. As part of a comprehensive international survey of crime and attitudes towards crime, Martin Killias, Pat Mayhew and Jan van Dijk analyzed firearms ownership rates and firearms homicide rates for nine democratic nations. [39] Switzerland, with more guns per capita than all but the United States, had fewer firearms homicides per capita than six of the eight countries. [40] In response, Dixon might note that he is only arguing for the prohibition of handguns, and the fact that long gun density does not correlate with long gun homicide is not *298 inconsistent with his hypothesis.

But even if Dixon's point that higher handgun density relates to higher handgun homicide is true, so what? Unless it is assumed that handguns are some intrinsically evil totem, what difference does it make what kind of weapon is used to kill a person? During one debate on tightening Canadian gun laws, member of parliament Stuart Leggatt praised the benefits of strict gun laws: "New York, which has a fairly respectable Sullivan law, has a 25 percent murder rate by firearms, whereas Dallas, where unrestricted use of firearms is allowed, has a rate of 72 percent of murder by firearms." M.P. Otto Lang replied: "The honourable member has made an interesting case which, if read carefully, shows that murder by knife is a nicer game than murder by gun. I cannot see the point of that." [41] Other gun control advocates make the same point of M.P. Leggatt and Professor Dixon, praising a reduction in handgun homicides per se, without looking to see if overall homicides have gone down. [42]

Obviously it is possible that a greater rate of handgun homicides also correlates with a greater rate of overall homicides, since handguns could be a weapon uniquely suitable to the promotion of homicide. Dixon argues as much later in his article, [43] and I respond below in this article. [44] But in the meantime, it is important to recognize the limitations of what Dixon has demonstrated thus far. At the most, he has shown that there is a relationship between handgun density and handgun homicide. He has not demonstrated a relationship between handgun density and overall homicide rates; he has not even presented figures regarding overall homicide rates. Dixon asserts that the international evidence which he has offered makes, by itself, a prima facie case for handgun restrictions. [45] His point can only be valid if it is believed that a reduction in the handgun homicide rate, unaccompanied by a reduction in the overall homicide rate, would be a good thing in itself. For persons who do not believe that murder by knife or shotgun is a nicer game than murder by handgun, nothing that Dixon has presented, thus far, shows any utility at all from handgun prohibition.

Does low handgun density and/or stricter handgun control *299 lead to a lower total homicide rate? The comparative evidence suggests not necessarily. In Great Britain, handguns may only be obtained after an extremely rigorous licensing process involving police inspections of the applicant's home and months of delay. [46] Carrying a loaded or unloaded handgun is absolutely forbidden without a license, and licenses are virtually never granted. [47] Every handgun transaction must be approved in advance by the police, and every legally-owned handgun is registered. [48] In Switzerland, handguns are readily obtainable after a person obtains a simple police permit which is valid for three months. [49] During the three months, the permit holder may buy as many handguns as he wishes, and purchases are generally not registered. Fifteen of the twenty-six cantons, representing about 57% of the population, have permit procedures for carrying handguns (some of which make permits difficult to obtain); the other cantons, representing 43% of the population, have no rules requiring a person carrying a loaded handgun to obtain any permission at all. [50] In England and Wales, the homicide rate per 100,000 population is 1.1; in Scotland (for which government statistics have always been recorded separately) the rate is 1.7. In Switzerland, where the handgun laws are immensely more lenient than in Great Britain, the rate is 1.1. [51] In the nine-country study detailed above, Switzerland had the third-lowest homicide rate, even though its handgun laws are less restrictive than all countries in the study except the United States. [52] (Indeed, most of the American states with high homicide rates have stricter handgun laws than Switzerland.) [53]

As Dixon points out later, Switzerland has a higher rate of handgun homicide than the other countries he analyzes, such as Australia, Canada and Britain. [54] Yet Switzerland, with a murder rate of 1.1 per 100,000 has a much lower murder rate than Australia (2.7) *300 and Canada (2.5), and a somewhat lower murder rate than Great Britain (1.1 for England and Wales, 1.7 for Scotland). [55] The data suggest that there is not necessarily a relationship between the handgun homicide rate and the overall homicide rate.

American data also fails to provide support for a strict relationship between handgun density and total homicide. Population groups which are highest in handgun ownership rates-namely wealthier people, Protestants, whites, and rural populations-all have lower homicide rates than other groups. [56] In addition, the American homicide rate rose tenfold in the first three decades of the twentieth century [57] but U.S. per capita handgun ownership remained stable. Between 1937 and 1963, handgun ownership rose by 250 percent, but the homicide rate fell by 35.7 percent. Homicide fell again in the early to mid 1980s, even as handgun ownership was surging. [58] Of course there were likely confounding factors in the historical American data. One reason that the American homicide rate rose so sharply in the 1920s was the violence caused by alcohol prohibition, and one reason that the homicide rate fell from 1937 to 1963 was the improved quality of medical care. I am not suggesting that the evidence presented thus far proves that increased handgun density does not cause increased total homicide. I do suggest, however, that the evidence developed so far by Dixon shows no reason to believe that lower handgun density would save lives, although lower handgun density may, arguably, be associated with lower number of handgun homicides.

C. The Burden of Proof

Dixon next proceeds to argue that the international correlation between handgun density and handgun homicide is not merely a co-incidence, but the result of a cause and effect relationship. Even if Dixon is, at the end of the exercise, found to be completely successful in proving his point, he has not yet shown evidence that adopting a handgun ban will save lives, although he will have (if successful), shown that the number of handguns in a society does correlate with the number of homicides by one particular method.

Unlike some proponents of handgun prohibition, Dixon does not claim that handguns are the only cause or the most important *301 cause of the high American handgun homicide rate (or of the high overall murder rate). [59] He offers the more intuitively plausible argument that high handgun density is simply one important cause of the high handgun murder rate in America. In order to prove the cause and effect relationship between handgun density and handgun homicide, Dixon sets up a two part test for himself: 1. Show that no variables other than the cause (high handgun density) better correlate with the effect (high handgun homicide); 2. Provide a "probable theoretical explanation of how the causation occurred." [60]

Immediately after setting up the two part test for what he has to prove, Dixon alters the terms of the debate so that he is certain to win. Dixon had defined item 1 of his burden of proof to be showing that no variables other than handgun density better correlate with high handgun homicide. To prove item 1 of the hypothesis, Dixon would have to look at other variables which might affect handgun homicide (such as poverty, racial problems, or police density), and show how those variables correlate with handgun homicide. Dixon could, by his initial terms, prove item 1 of his test true by showing that handgun density correlates with handgun murder better than do other variables such as poverty. But before beginning the test, Dixon switches the terms. He acknowledges that there are several non-gun causal factor variables which correlate with handgun homicide, but "none of them is nearly strong enough to be considered as the only cause, and hence disprove my hypothesis." [61] Thus, Dixon will consider item 1 of his hypothesis confirmed unless it can be proven that some murder-related variable (such as race or police density) is the only variable in handgun homicide.

Dixon's new standard is illogical. Rather than showing the relationship between alleged cause (handgun density) and effect (handgun homicide) by demonstrating handgun density is the best variable, Dixon claims that cause and effect will be proven simply by showing that no other potential cause can be considered "the only cause." Since almost all human behaviors, including homicide, spring from multiple causes, it is unlikely that any variable which helps cause handgun murder would be "only" cause. Thus, Dixon guarantees in advance that he will satisfy item 1 of his test. [62]

*302 Dixon then attempts to satisfy item 2, which required him to offer a "probable theoretical explanation of how the causation occurred." [63] As to providing a theoretical explanation, Dixon asserts "one need not go beyond common sense." [64] Dixon's defense of "common sense" as a satisfactory offering of a "probable theoretical explanation of how the causation occurred" begins with an assumption which he apparently considers non-controversial: "Assuming human nature to be relatively similar in different developed countries ... one would expect people to be subject to roughly similar amounts to stress, provocation ... and whatever other factors are liable to lead some people to violence." [65] But the assumption that human nature is "relatively similar in different developed countries" is not intuitively obvious. As a simple example of the differences in human nature, consider how people behave on a bus. The Swiss mass transit systems successfully depend on voluntary payment, [66] but any American subway or bus system that depended on the honor system for payment would quickly go bankrupt. Likewise, Swiss pedestrians almost always wait at traffic lights, even when there is no traffic. [67] In most major cities, American pedestrians apparently feel entitled to walk even when traffic is rushing forward. Could it be that the Swiss are, by nature, more cautious, more honest, or more law-abiding than Americans?

Automobile plants, steel mills, and coal mines are generally similar around the world, and if Dixon's assumption about human nature also being generally similar, industrial relations laws which work on one developed nation's coal mines ought to work about as well in another nation's technologically similar coal mines. But attempts to impose one nation's labor laws on another's labor force have failed badly. [68] The British Industrial Relations Act, which was modeled after American statutes, attempted to outlaw the closed shop and to make unions liable for their members' actions, but the law did not work. One commentator concluded that the act's down-fall resulted from attempting to change human nature by statute and to alter existing national patterns of labor relations. [69]

Traffic and transit habits notwithstanding, perhaps the Swiss *303 really are essentially like Americans; and labor law experiences notwithstanding, perhaps working people are the same throughout the developed world; and the perhaps the researchers who believe they have found profound difference in human nature across cultures are wrong. But at the least, the assumption that human nature does not vary significantly among countries is hardly an easily- accepted starting-point for a "common sense" argument.

Having assumed human nature to be similar throughout developed countries, Dixon then assumes that people in those various countries are "subject to roughly similar amounts of stress, provocation, jealousy, anger, desperation, resentment of other people's affluence, and whatever other factors are liable to lead some people to violence." [70] The second assumption is even less supportable than the first. Perhaps people everywhere are subject to roughly the same amount of "resentment of other people's affluence," and perhaps everywhere a given amount of resentment will lead to a given amount of violence. Even so, resentment-based violence would still vary across cultures depending on the types of disparities of affluence. For example, in countries where the income spread between the rich and the poor is very large (as in America and France), there would logically be more resentment (and thus more resentment-based violence) than in countries where income gaps were smaller, such as Japan. [71] Similarly, one might expect different amounts of resentment-based violence in a society where class differences are seen as part of a long-standing social order (such as Great Britain) than in a society whose ideology insists that every person who is not wealthy has only himself to blame (such as the United States).

Similar arguments can be offered regarding rest of Dixon's list of violence- inducing provocations. While all people may react similarly to desperation, there could be considerably less desperation in countries which have effective social welfare systems than in countries which do not. While all people may (arguably) react similarly to jealously, there may be different amounts of jealousy in countries where divorce is often considered socially unacceptable (such as Switzerland) than in countries where divorce is widespread.

Having assumed - perhaps prematurely - that people in all developed countries are basically the same and they are subject to basically the same amounts of factors which induce some people to violence, Dixon comes to the climax of his "common sense" argument: "*304 If one of these nations has a vastly higher rate of private ownership of handguns, one would expect that the similar provocations to violence would spill over into handgun murder far more often than in other nations." [72] Here Dixon has done nothing more than restate the intuition that began his article. Persons who, like Dixon, "expect" that handguns cause handgun homicide may have their intuition reinforced, but nothing Dixon has offered is sufficient to prove anything.

Having announced that he has met his burden of proof, Dixon then suggests that the burden of proof regarding handgun prohibition should be shifted to the opponents of prohibition, and that they should be forced "to produce an alternative causal account which proves that the United States' high handgun murder rate is caused by factors unrelated to its high rate of handgun ownership." [73] He then turns to discussion of several authors to whom he assigns the role of meeting his "burden of proof challenge." [74]

Dixon notes that two of America's most prolific writers on the gun issue, Don Kates and Gary Kleck briefly advert to "factors other than the higher prevalence of handguns in the U.S." which could be responsible for the higher U.S. handgun murder rates. Dixon finds the Kates argument unsatisfactory since Kates simply states that other cultural factors could be at work, but does not detail what they are; [75] Kleck does somewhat better, in Dixon's view, by spending a page and quarter arguing for improved job training and anti-poverty programs. [76] Dixon asks, not unreasonably, for opponents of handgun prohibition to provide considerably more detail to support their argument that factors other than handgun density explain the high American handgun murder rate.

If volume is what Dixon wants, it can be provided. My recent book takes on the challenge of providing many explanations for America's high handgun murder rate. The book runs 442 pages, and contains more than 1,900 endnotes, perhaps enough to meet Dixon's request for detailed analysis. Whether the book is persuasive can only be decided by persons brave and/or foolish enough to venture into such a tome, but a list of reasons for America's high handgun murder rate can be summarized here:

More than any other society in recorded history, the United States is premised on individualism. Mistrust of government, and faith in individual initiative are so deeply ingrained in the American *305 character that Americans may not understand how aberrational they are until they look at other countries. (To give a small example, America's major liberal party, the Democratic Party, supports less government intervention in the economy than do the right-wing parties in Canada and Great Britain.) American mistrust of government means that American police are less powerful legally, and less respected in the community than in most other democratic nations. Further, American police are constrained by Bill of Rights-based protections such as the exclusionary rule and the Miranda warnings that do not exist in other nations. [77] At the same time, the American government has chosen to fight a "war on drugs" with much more intensity than other nations have, and as a result, America's prisons are increasingly filled with small-time drug dealers serving lengthy mandatory minimum sentences, while persons perpetrating serious violent crimes, including crimes with handguns, face a smaller and smaller expected sentence. [78] And of course American rates of abuse for drugs such as cocaine which have a psychoactive effect in lowering inhibitions against violence are higher than in other countries.

While the government is less able to control crime, Americans are more willing than the people of other nations to commit crimes. America's pervasive individualism means that Americans, compared to the peoples of other democratic nations, are less bound to the standards of family, church, employer, community, or state than other peoples. For the most part, American individualism and freedom produces benefits, such as America's great artistic creativity and its high rate of mechanical invention. But the weakening of social control leaves some people without the restraints that might, in other societies, prevent them from becoming criminals. While American criminals are increasingly ruthless and callous toward human life (even their own), even the criminals of other nations retain ties to the social order which would be incomprehensible to Americans. For example, Japanese gangsters, when tipped about an imminent police raid by corrupt informants, will speedily vacate the premises, but will leave a few handguns behind for the police to confiscate, so that the *306 police conducting the raid do not lose face. [79] Few American criminals display much concern for the emotional well-being of the police.

American individualism fits into a capitalist economy where survival of the fittest and blaming of victims for their own plight are emphasized more than in other countries. Income inequality is greater in America than in other nations; the social welfare safety net is considerably worse (and in some respects criminogenic); and the American ideology labels every poor person a failure. All of the above reinforce each other to produce high levels of resentment-based crime. The effect is enhanced by the American ideology of social equality in which persons are taught that there are no legitimate classes, and that one person is as good as another, which leads (in the minds of some poor people) to the conclusion that a poor person has a much right to a rich person's property as does the rich person.

Combined with unintended side effects of a strongly individualist, capitalist economic system are the effects of racism. America is more racially diverse than all developed democracies, and America's racial problem far exceeds those of other nations. Only in America was a major part of the population brought into the nation via kidnapping, enslaved for more than two centuries, viciously oppressed and segregated for another century, and then "liberated" into a destructive welfare system.

In addition, from the first moments of white settlement, America's history has legitimated and encouraged violence. While the whites moving into Australia quickly dispatched resistance from the stone-age Aborigines, and while the Canadian government successfully negotiated Indian treaties to peacefully settle the Canadian frontier, the United States of America was wrested from Indian hands by a savage war of genocide that lasted three centuries. The cruel war between Indians and whites helped inculcate in white Americans the "heroic" willingness to die to protect what one has and the less than heroic willingness to kill to get what one wants. While America had 69 Indian wars, Canada had none. [80]

Unlike Great Britain or Canada or Sweden or Australia (but like Israel and Switzerland), the United States won its independence through a long and difficult war of national liberation, in which ordinary citizens bringing their own weapons to battle with an imperial standing army played a decisive role. Later, the American Civil War led to government-sanctioned killing on a vast scale, as well a *307 sanctification of that killing for the noble ends of abolition or for the South's supposedly glorious "lost cause."

While the Canadian west was peacefully settled under the supervision of a Mounted Police Force that was on hand to provide law and order before settlers began to arrive, the American frontier, whether the frontier was western Pennsylvania in the early 18th century or Nevada in the late 19th, was a Hobbesian, chaotic world, where government was ineffectual, and individuals had to protect themselves with force or die. (The influence of handguns in the American west is discussed below.) [81]

In the urban eastern United States, meanwhile, rapid industrialization and massive immigration proceeded simultaneously (as they did not in most other democratic nations), and immigrants, rather than coming from a relatively small region (such as the British Isles immigrants who settled Australia and New Zealand) came from astonishingly disparate backgrounds, and frequently encountered problems adapting to their new nation. Partly as a result, the violent crime rate in 19th century cities such as Philadelphia was far higher than in America's "wild west." Again and again, Americans displayed an unusual willingness to use violence to achieve their ends. Unlike the more accommodating British capitalists, American captains of industry of the 19th century were quick to use violence to suppress labor militance. In sum, for an immense variety of reasons, America is more violent than other nations, and would be even if handguns had never been invented.

Of course the above paragraphs are only an outline of the arguments detailed in The Samurai, the Mountie, and the Cowboy, and persons who are unpersuaded by the outline might or might not be persuaded by the book itself. Dixon's challenge to provide a causal explanation for something other than handgun density to explain the high American handgun murder rate is, arguably, met by The Samurai, the Mountie, and the Cowboy.

Dixon, for his part, does not deny that there are factors other than handguns at work in the American handgun murder rate, or that we should take steps to address those measures. [82] He does note, correctly, that explication of those other factors does not disprove his *308 hypothesis that handgun density causes handgun murder. [83] True enough, since, as he also notes, it is impossible to disprove a causal hypothesis simply by suggesting alternative causes. [84] But if the historical and cultural analysis of the United States which I offered above is true, then 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. By way of proof, Dixon has thus far done nothing more than to point out that, in six countries for which he compiled statistics, handgun density correlated with handgun homicide, with the United States having the highest levels of both. Given the fact that there are numerous violence-inducing factors at work in the United States which are not at work in the other countries Dixon tallies, Dixon is a very long away from proving that handgun density is a significant cause in America's high handgun murder rate, or even a medium-sized cause.

The weakness of Dixon's assertion that he has proved a relationship between handgun density and handgun homicide is illustrated by the fact that his analytic technique can be used just as easily to "prove" that handgun controls cause handgun homicide. Instead of looking at 6 countries, as Dixon does, let us look at 6 American cities. How about Washington, D.C., New York City, Los Angeles, Memphis, Denver, and Salt Lake City? These cities are listed in decreasing order of homicide rates. [85] The cities are also listed in decreasing severity of handgun laws. [86] Just as Dixon notices that countries with fewer handguns have fewer handgun homicides, I notice that American cities with fewer handgun laws have fewer handgun homicides.

Next, like Dixon, I ask whether the relationship between lenient handgun laws and low handgun homicide is one of cause and effect, or simply a coincidence. Pursuant to Dixon's analytic method, to prove a cause and effect relationship, I must do two things. First, "show that there are no other variables which correlate better with the effect, and would account for the effect better than, or in place of, the posited cause." [87] Regarding the first item, I, like Dixon, "do not rule out the existence of other causes," such as racial, economic, *309 or population density variables as partial explanations of why cities with lenient handguns laws have lower handgun homicide. [88] Like Dixon, I will have satisfied prong one of the burden of proof analysis as long as no one can point to another causal factor that is "strong enough to be considered as the only cause and hence disprove my hypothesis." [89] Since neither race, religion, economics, nor any other likely cause of low handgun homicide rates can be proven to be the only cause, I have (by Dixon's reasoning) met prong one of the burden of proof.

To meet the second prong of the Dixon burden of proof, I must provide a "theoretical explanation" of why lenient handgun laws correlate with low handgun homicide rates. [90] For a theoretical explanation, I, like Dixon, "need not go beyond common sense." [91] He begins by "[a]ssuming human nature to be relatively similar in different developed democratic countries," so I will assume human nature to be relatively similar in different large American cities. Human nature having been assumed to be similar, Dixon then assumes that people in various countries are subject to "roughly similar amounts" of whatever factors "are liable to lead people to violence." I make the same assumption regarding people in various American cities. [92] Finally, I like, Dixon, provide a simple explanation for how my theory could work, simply substituting or adding a few words in Dixon's explanation: "If one of these nations [cities] has a vastly higher rate of private ownership of handguns [thanks to lenient gun laws], one would expect that similar provocations to violence of violence would spill over into handgun murder far more [less] often than in other nations [cities]." Note that Dixon's "theoretical explanation" [and my plagiarization of it] are not a theoretical explanation at all, but simply a statement of a hypothesis. His hypothesis is that handguns make murder more likely. My hypothesis is that handguns in the hands of non-criminal persons make murder less likely, by deterring criminal attack.

If you are not persuaded by my "proof" of my hypothesis - and you should not be - you also should not be persuaded by Dixon's proof of his hypothesis. Dixon's hypothesis could be true, as could mine, but neither of us has done more than present a simplistic and limited data set for which there are so many potential confounding *310 variables that any claim to proof based on the data is hopelessly premature.

D. Snow and Guns: Switzerland and Canada

To support his hypothesis, Dixon turns to two foreign comparisons, one involving Switzerland, and the other involving a study of Seattle and Vancouver. Dixon notes that opponents of gun control often advert to Switzerland, where the government gives every adult male a full automatic assault rifle. [93] Dixon replies that whatever policy the Swiss have towards long guns does not disprove his case for handgun control. [94] In addition, Switzerland, while having a handgun murder rate lower than America, has a handgun murder rate higher than do Sweden, Australia, Canada, and Britain. [95] Dixon is right, but as discussed above, Switzerland has an overall murder rate lower than those other nations. [96] The Swiss evidence suggests, therefore, that while handgun density might correlate with the handgun murder rate, handgun density does not correlate with the total murder rate. Except for persons who think that murder by a handgun is somehow worse than murder with another weapon, handgun prohibition offers no benefits.

Dixon also notes that in Switzerland, gun ownership takes place in the context of "mandatory service in a citizens' militia, with its attendant training and discipline, which bears no comparison with the minimally controlled private handgun ownership in the United States." [97] The observation is not precisely accurate; while militia officers are issued handguns for militia service, every adult in Switzerland can purchase private handguns under conditions less restrictive than those in many American states. [98]

More generally, it might be argued that, although Switzerland has a high rate of handgun ownership, it has a low homicide rate because *311 other factors unique to Switzerland depress the homicide rate. More so than the citizens of most other democratic nations, the Swiss retain close ties to communities where their ancestors have lived for centuries; the patriarchal Swiss family system in strong; violent cinema is censored; and a tight network of social cohesion helps keep violent crime of all types very low. [99] Conversely, the very factors which keep violent crime in Switzerland low are conspicuously absent in the United States. [100] Accordingly, analysis of Switzerland and the United States could suggest that handgun density is no more than a trivial variable in overall homicide rates, being dwarfed by other socio-cultural factors.

After Switzerland, Dixon moves to the most well-known study of the impacts of foreign gun controls, a 1988 comparison of Seattle and Vancouver published in the New England Journal of Medicine. [101] The NEJM article contrasted Seattle, with its higher homicide rate, and Vancouver, with its lower rate. The article observed that Vancouver had stricter handgun laws than Seattle, and a lower handgun homicide rate. Many economic variables were similar in Seattle and Vancouver, and Seattle and Vancouver had roughly comparable robbery and assault rates. Seattle's homicide rate, however, was about 50% higher, and most of the excess homicides in Seattle were the result of handgun homicides.

In reading the Seattle-Vancouver study to demonstrate that handgun density causes handgun homicide, Dixon points to Vancouver's "significantly stricter gun control laws." In particular, "Vancouver does not allow concealed weapons and grants handgun permits for sporting and collecting purposes only. Handguns may be transported by car only if they are stored in the trunk in a locked box." [102] While Vancouver has long had stricter controls on the carrying of handguns, the most important aspect of the Vancouver law - the ban on owning handguns for self-defense (and the consequent prohibition of storing a loaded handgun in the home) was enacted in 1977 and went into effect in 1978. (The laws on carrying handguns outside the home were also tightened in 1977.) [103] If Dixon's hypothesis *312 about handgun density and handgun homicide is correct, then Vancouver's murder rate would be expected to fall after the 1977 law. But in fact, Vancouver's handgun homicide rate (as well as the overall homicide rate) after the law went into effect remained the same as in the years before the law. A study sponsored by the Canadian government (which had an interest in finding the gun laws to be effective) found no evidence that the 1977 law had reduced handgun homicide in Vancouver. [104]

Critics of the Seattle-Vancouver study have also pointed to the importance of race in explaining the differing homicide rates in the two cities. Brandon Centerwall, a professor of Epidemiology at the University of Washington, suggests that if the homicide data "were subjected to a Mantel-Haenszel summary odds ratio, stratifying by race, the differences in homicide rates between Seattle and Vancouver would cease to be statistically significant." [105] J.H. Sloan, lead author of the Seattle/Vancouver study has declined to calculate the ratio.

Although Seattle whites have easier access to handguns than Vancouver whites, Seattle whites are no more prone to commit homicide than Vancouver whites. The fact casts serious doubt on the Dixon hypothesis that handgun density per se causes handgun homicide. In contrast to the Swiss, Seattle whites do not necessarily live under cultural conditions which serve to minimize handgun violence despite a high prevalence of handguns.

That Seattle had a higher overall homicide rate than Vancouver was due entirely to murders involving Blacks and Hispanics, who together comprise about 12% of the Seattle population, but less than 1% of Vancouver's. As Dixon notes, the fact does not itself disprove the need for gun control in America. The Seattle data would be consistent with a weaker version of the Dixon hypothesis: handgun density causes handgun homicide, but only among groups (such as American Blacks and Hispanics) who for other reasons are already vulnerable to violence.

While the Seattle-Vancouver data does not disprove a modified *313 version of the Dixon hypothesis, neither does the data convincingly support the hypothesis. The hypothesis would be supported if Canada, like the United States, had a large racial minority population that had been enslaved and otherwise oppressed and brutalized for more than three centuries. If that comparable Canadian racial group existed in Vancouver and had a substantially lower homicide rate than the Blacks and Hispanics of Seattle, then the modified Dixon hypothesis would be supported. But Blacks and Hispanics are less than 1% of the Vancouver population, and their history in Canada is, to Canada's credit, more benign than the history of Blacks and Hispanics in America. [106] The small numbers of these groups in the Vancouver population make the drawing of statistical inferences regarding their handgun murder rates problematic.

In sum, two non-controversial points can be concluded from the New England Journal of Medicine's Seattle-Vancouver study:

1. Among groups not at risk of violence (whites in Seattle and Vancouver), increased handgun density was not associated with increased risk of handgun homicide;

2. Among high-risk groups (Seattle Blacks and Hispanics), there was a high handgun homicide rate, but it is impossible to conclude that gun control would or would not reduce the rate since a similar population was not present in large numbers in Vancouver.

At this point, it may be useful to note briefly some other data and studies analyzing American and Canadian crime rates to test for the efficacy of gun control laws. The overall death rate for non-hispanic white Americans from all types of shootings (murder, suicide, accident, etc.) is the same as the rate for Canadians, even though American whites own far more handguns per capita than do Canadian whites. [107] In American states which border Canada, the homicide rate is generally no higher (and often lower) than in adjacent Canadian provinces. [108]

A study by Robert J. Mundt study compared twenty-five *314 Canadian cities with twenty-five comparably-sized American cities. When the covariates of "percent Black" and "city size" where considered, the difference between American and Canadian samples diminished to the point of insignificance. [109] Another study by Mundt compared Winnipeg with Minneapolis/St. Paul, and Duluth with Thunder Bay, both city pairs having many similarities. [110] Mundt's four-city study found that racial composition and city size were not sufficient to account for all of the crime rate differences between the American and Canadian cities, and thus Mundt could not exclude the possibility that the Canadian gun laws had crime-reductive effects. [111]

In sum, the Canadian data does not conclusively disprove Dixon's hypothesis, but the data, at best, does no more than weakly support the hypothesis. If, as Dixon seems to acknowledge, handgun ownership presents a problem mainly among at-risk population groups, directly addressing the problems of those groups might be a more efficacious way to reduce handgun homicide than to implement across-the-board handgun prohibition, which, as will be discussed below, would have a devastating effect on the American criminal justice system, possibly leading to increased violent crime, and almost certainly leading to massive erosion of the Bill of Rights. [112]

*315 II. OBJECTIONS TO HANDGUN PROHIBITION

A. Interstate Data

Finished with international comparisons, Dixon turns to interstate analysis. If gun controls reduced crime, then it would be expected that states with stricter gun laws would have lower gun crime rates. But as Dixon acknowledges, states with stricter gun laws have higher crime rates. [113]

There are several possible explanations. First, the states which enacted the strict gun laws had high crime to begin with; that it why the stricter laws were enacted.

A second, not inconsistent explanation, is that gun control itself causes higher crime, making the already high-crime states even worse than they would otherwise be. Dixon deals with the explanation by dismissing it as "perverse." [114] Argument by epithet is not persuasive, and (as will be discussed below), there are plausible reasons to believe that some gun controls may increase crime, and that Dixon's proposed handgun ban would substantially aggravate crime. [115] Dixon admits that at least sometimes handguns prevent crime, [116] so it is hardly "perverse" to suggest that it could be possible that states which weaken the deterrent effect of civilian handgun ownership suffer increased crime.

A third explanation for why states with stern gun laws have more crime than other states is that guns from other states, with looser laws, are smuggled into the high crime states, thus reducing or eliminating the crime-reductive effect of the strict state's law. This explanation is not inconsistent with the first two explanations. Dixon devotes the rest of his interstate discussion to arguing for this third explanation.

Even if Dixon's explanation about leakage is generally true, it remains difficult to account for the dismal performance of many gun controls. For example, in 1976 the Washington, D.C. murder rate stood at 26.9 per 100,000 population, according to FBI statistics. The city council enacted a handgun ban which went into effect in February *316 1977, and since then the Washington rate has always been higher than 26.9 (except in 1985). [117] Today, the rate is three times higher than it was before the ban was enacted. [118] If handgun bans work, why would the homicide rate rise after 1977 (which was years before the "war on drugs" made Washington's homicide problem even worse)? Smuggling guns into Washington, D.C. from other states was no easier in 1980 than it was in 1976. The ban on possession by law-abiding citizens should have reduced the supply of handguns available for Washington, D.C. criminals to steal, and should have prevented law-abiding citizens from shooting each other with handguns in heat-of-passion homicides. The D.C. handgun ban's impact on law-abiding citizens would not be defeated by interstate smuggling, since law- abiding citizens would, be definition, not buy an illegal gun. And yet the Washington homicide rate rose. Similar increases in gun crime in other jurisdictions, such as Chicago after its own handgun ban, [119] and New York City after its severe "Sullivan" handgun licensing law, [120] at least raise doubt about the complete sufficiency of interstate gun smuggling as an explanation for the failure of the gun laws. If interstate smuggling were the whole story, then it would not be expected that crime rates would rise immediately after gun laws were enacted.

These doubts may be enhanced by analysis of gun controls in other nations with uniform national handgun laws, and consequently no problem regarding interstate smuggling. Neither Canada's 1977 gun controls nor the Pistols Act of 1903 in Great Britain nor the Firearms Act of 1920 in Great Britain, were associated with reductions in handgun crime. [121]

To support the theory that interstate gun-running explains why states with severe gun controls have so much more crime than states which do not, Dixon points to a study published by the Bureau of Alcohol, Tobacco and Firearms in 1976. [122] Comparing "crime guns" found in Atlanta, New Orleans, Detroit, and New York City, BATF *317 found that the cities in the states with the stricter gun laws (Detroit and New York City) had smaller percentages of traced guns with origins within their states. The correlation between stricter laws and a higher percentage of out-of-state guns was not, however, exact. New York City has much stricter gun laws than Detroit, but 23% of New York City handguns came from New York State, whereas only 8% of Detroit guns came from Michigan.

Several limitations of the BATF methodology should be noted. First, the "crime" guns included not only guns used in violent crime, but also guns which were simply found on citizens who had not complied with local licensing laws, such as New York City's, which at the time made it nearly impossible to obtain a handgun license. [123] One analysis of handgun seizures (conducted by a handgun prohibitionist), found that 20 to 25% of police handgun seizures were not associated with any crime, not even a licensing violation. Some of the guns may simply have been turned into the police by lawful owners who wanted to get rid of them. [124] BATF made numerous methodological errors, such as counting some guns twice, or counting guns seized outside the time period for the study. [125] More importantly, the study did not detail how the guns moved from one state into another. For example, a gun which was lawfully purchased in Ohio, stolen, and then sold on the black market in Detroit was not distinguished in the BATF data from a gun which was bought illegally in Ohio by a Michigan criminal, as a result of Ohio's "lax" gun laws. [126]

But assume that Dixon's broader point is correct, and that more lenient gun laws in some states undermine more severe laws in other states. Accordingly, suggests Dixon, a comprehensive federal handgun law would be necessary. But would a federal handgun succeed? Dixon acknowledges that even if interstate smuggling of the current American gun stock somehow ended, guns would still be available from illegal manufacture and illegal import. [127] Indeed illegal *318 gun production is already not unheard of; a Bureau of Alcohol, Tobacco and Firearms study found that one-fifth of the guns seized by the police in Washington, D.C., were homemade. [128]

Dixon expects the "fact that such guns are inaccurate and dangerous to the user will also act as a restraint to illegal gun production." [129] How much of a restraint may be open to doubt. While homemade guns will not win target- shooting contests, target shooters will have their own guns (kept at shooting ranges under the Dixon proposal), and homemade guns may suffice for robbery purposes. And most homicides, like most robberies, are perpetrated at very close range where accuracy is not an issue. The risk that a homemade gun could explode in a shooter's hand may deter some otherwise law-abiding citizens who would want to own an illegal handgun for protection. On the other hand, if the person believes that the threats to his or her life and family are serious enough to commit the serious crime of buying an illegal handgun, the additional risk posed by potentially defective handgun may seem small. In addition, newfound popularity for bootleg guns might result in handguns becoming cheaper than they are now, just as in alcohol prohibition days, bootleg gin often cost less than legal alcohol had. If handguns were cheaper, they might become more available to small-time teenage criminals and other low-end miscreants; criminals might end up more widely armed than ever before.

The inevitable black market in homemade and imported illegal handguns would provide a major new revenue source to organized crime. As the black market in alcohol helped create and enrich organized crime in the United States, the new black market in handguns would fund and strengthen organized crime all the more.

Dixon also acknowledges that illegal handguns would also flow in across American borders. [130] Indeed, if small handguns were imported in the same physical volume as marijuana, 20 million would enter the country annually. (Current legal demand for new handguns is about 2.5 million a year). [131]

Responding to the argument that imported guns will render the handgun ban useless, Dixon draws an analogy to the drug war. *319 Illegal imported drugs help to undercut a domestic ban on their possession and sale, but "This possibility would not seriously be entertained as a reason for giving up the fight against dangerous drugs made in the United States. What is called for is an assault on both domestic and imported drugs." [132] The analogy Dixon draws between the drug war and his proposed handgun war is apt, but the drawing of the analogy highlights the danger posed by the proposed handgun assault, as the next section argues.

B. War on Drugs and War on Guns

Some economists argue that drug prohibition should be abandoned because no matter how much a government attempts to prohibit a commodity, the market will always produce enough of the commodity to satisfy consumer demand. [133] Swimming against the basic economic principle that the market will generate supplies of commodities to meet consumer demand, the drug prohibition laws have led to wholesale destruction of civil liberties. The War on Drugs has now become a War on the Constitution, and the American people have become, in the eyes of their government, a society of suspects. [134]

If a Black person buys an airplane ticket with cash, he risks being stopped by police at the airport, and having his money confiscated. Persons who fit "drug courier profiles" may be detained and harassed by the police, although such profiles include getting off the plane early, late, or in the middle as an element of the profile. [135] Infrared sensors spy into people's homes, with no probable cause. [136] Except in the home, the Fourth Amendment's probable cause requirement has been mostly abolished by a "law and order" Supreme Court. [137] Under forfeiture laws, billions of dollars of private property have been seized from persons who have never been charged, let alone convicted of any crime. [138] Pre-trial detention, a gross contradiction *320 of the presumption of innocence, has become routine. Citizens traveling on busses, on trains, or in private cars are liable to be pulled over and searched by police and drug-sniffed by police dogs for no reason at all. [139] Urinalysis has become a routine condition of initial or continued employment, and the medical privacy of many persons taking lawful prescription medication has been compromised as a result. [140] Stalinesque "Drug Abuse Resistance Education" programs in the schools encourage children to turn in their parents for illegal drug possession. [141] Attractive young police officers pretend to be high school students, and pester socially awkward teenagers into selling them drugs. [142] Punishment for crime has become grotesquely disproportionate to the offense, as teenagers in possession of $1,500 worth of LSD are sent to prison for longer terms than kidnappers and arsonists. [143] America has a higher imprisonment rate than any other nation in the world, and yet violent criminals serve less and less time in prison as America's rapidly expanding prison industry takes in more and more young people convicted of drug offenses. [144] The United States Army is conducting domestic law enforcement operations in California and Oregon; the National Guard has been turned into a militarized drug police. [145] Wiretapping has never been more common. [146] Financial privacy has vanished as banks must report currency transactions; car dealers must report *321 customers who buy with cash. [147]

And what has this massive loss of liberty bought? Cocaine and heroin are cheaper, purer, and more widely available than ever. [148] Marijuana use is down, and so now college students who would have quietly gotten high engage in binge drinking. [149] A magnificent legacy of civil liberties, slowly constructed over 200 years, has been squandered to attempt to prevent the American people from choosing which substances they wish to ingest.

As the malignant cancer of the Drug War eats away at the Bill of Rights, Dixon proposes more of the same, by adding handguns to the government's list of prohibited items. The damage to the Bill of Rights could be greater than that resulting from the drug war. Gun controls have always been associated with intrusive searches and seizures in violation of probable cause. Judge David Shields of Chicago's firearms court observed: "Constitutional search and seizure issues are probably more regularly argued in this court than anywhere in America." [150] As early as 1933, one quarter of all weapons arrests in Detroit were dismissed because of illegal searches. [151] According to the American Civil Liberties Union, the St. Louis police have conducted over 25,000 illegal searches under the theory that any Black driving a late-model car must have a handgun. [152] The Chicago Police Department gives an officer a favorable notation in his record for confiscating a gun, even as the result of an illegal search. [153] As a practical matter, one cannot comply with the Fourth Amendment probable cause requirement and also effectively enforce a gun prohibition. Former D.C. Court of Appeals judge Malcolm Wilkey thus bemoaned the fact that the exclusionary rule, "has made unenforceable the gun control laws we now have and will make ineffective any stricter controls which may be devised." [154] Judge Abner Mikva, usually on the opposite side of the conservative Wilkey, joined him in identifying the abolition of the exclusionary rule as the only way to enforce gun control. [155]

*322 Abolishing the exclusionary rule is not the only proposal designed to facilitate searches for illegal guns. Harvard professor James Q. Wilson, the Police Foundation, and other commentators propose widespread street use of hand-held magnetometers and walk-through metal detectors to find illegal guns. [156] The Bush administration began buying magnetometers for city police departments in 1991. The city attorney of Berkeley, California, has advocated setting up "weapons checkpoints" (similar to sobriety checkpoints), where the police would search for weapons all cars passing through selected neighborhoods. [157]

The Police Foundation has also proposed that law enforcement agencies use informers to ferret out illegal gun sales and model their tactics on methods of drug law enforcement. [158] Taking this advice to heart, the Bureau of Alcohol, Tobacco and Firearms already relies heavily on paid informants and on entrapment - techniques originated during alcohol prohibition, and honed in modern drug enforcement. [159]

Gun control laws, already destructive of Bill of Rights liberties, will become considerably more destructive if the Dixon prohibition proposal is enacted. About a quarter of all American families own handguns. [160] A 1979 survey of Illinois gun owners indicated that 73 percent would not comply with a gun prohibition. [161] Thus, the number of new "handgun criminals" will become at least as large as number of drug criminals. Handgun criminals will be much harder to catch than drug criminals, since an illegal handgun owner need only make a one-time buy (or just hold on to what she already has), whereas persons disobeying drug and alcohol prohibitions must buy *323 new supplies as old supplies are consumed. Accordingly, an effective prohibition would likely have to be enforced with house to house searches.

The forfeiture abuses associated with the drug war could also translate easily into the handgun war, as otherwise law-abiding persons lose their homes, automobiles, and businesses because a handgun was found therein, even if the owner of the property was not the owner of the handgun, and had no knowledge about the handgun's presence. Indeed, Chicago Mayor Richard Daley has already initiated forfeiture proceedings against automobiles which contain a handgun in violation of Chicago's prohibition. [162]

Concluded Aryeh Neier, former director of the American Civil Liberties Union:

I want the state to take away people's guns. But I don't want the state to use methods against gun owners that I deplore when used against naughty children, sexual minorities, drug users, and unsightly drinkers. Since such reprehensible police practices are probably needed to make anti-gun laws effective, my proposal to ban all guns should probably be marked a failure before it is even tried. [163]

III. IMPACT OF PROHIBITION ON PUBLIC SAFETY AND SELF-DEFENSE

Having argued for the value and enforceability of handgun confiscation, Dixon devotes the second half of the article to discussion of the impact of handgun prohibition on criminals and other persons. Dixon's detailed analysis of these issues helps advance the gun control debate. While advocates of handgun prohibition often argue for the desirability of their goal, they rarely discuss in depth the results that might flow from their program, nor do they meet head- on the objections raised by opponents of prohibition. Dixon does both.

A. Is Disarming Criminals Sufficient?

Dixon begins by acknowledging that, at least in the first years that his confiscation program is in effect, the program will *324 have little effect on criminals, but will impact mainly non-criminal citizens. [164] Here Dixon takes a more plausible position than some of the more extreme advocates of gun control, who promise that measures far milder than Dixon's (such as a waiting period) will somehow disarm professional criminals such as drug dealers.

Disarming the non-criminal population would, Dixon says, be beneficial in itself, because the non-criminals are the ones who perpetrate most of the handgun murders. He points out that 3/4 of murderers do not have prior felony convictions. Accordingly, a handgun possession ban that applied only to felons would not have a chance of taking handguns away from the 3/4 of murders without previous felony convictions. [165] Dixon notes that many of the murderers without prior felony convictions do have arrest records (two-thirds to four- fifths of homicide offenders have arrest records, frequently for violent felonies [166]), and that in households where a domestic murder takes place, the police almost always have had prior contact (and often repeated contact) with the home because of domestic violence calls. [167] Accordingly, a handgun ban that disarmed only persons who had been arrested or who lived in homes where the police were called in for domestic violence could, theoretically, prevent the vast majority of handgun homicides. But as Dixon points out, outlawing handgun possession merely because a person has been arrested rather than convicted would violate the presumption of innocence, and banning handguns to households involved in domestic violence might raise equal protection problems. [168] If the ban could not (for reasons unrelated to the Second Amendment) be applied only to arrestees or persons living in houses with domestic disturbances, then the ban would have to be applied to the population as a whole.

Dixon's reasoning to this point is plausible, but he fails to take his inquiry one necessary step further. He concedes that criminals would not be affected by the ban in the first years, and therefore *325 we should not expect a quick reduction in homicides perpetrated by convicted felons. If so, it is doubtful that the handgun ban could disarm the 3/4 of the murdering population that does not have a prior felony conviction. Remember, the murderous 3/4 is not composed of persons with clean records and steady jobs who murder their wives in a fit of temporary rage because dinner was overcooked. The murderers without felony records are mostly persons with records of prior arrests, and, in domestic situations, persons with a prior record of (often repeated) domestic violence. In other words, they are criminals who have not yet received a felony conviction. Because the murderers still are criminals (albeit criminals lucky enough not to have been convicted of felony), they likely have access to much of the same social milieu and black market that criminals with felony convictions also have. Since the criminals without felony convictions would still be able to purchase handguns on the criminal black market - just as Dixon acknowledges that criminals with felony convictions would be able to - handgun prohibition would have no impact, in the short term at least, on most of the 3/4 of murderers who did not have felony convictions before their homicide.

The proposition is supported by research regarding victims of near-fatal domestic shootings and stabbings. Homicide victims tend to have sociological resemblances to homicide perpetrators (e.g. urban Black teenagers tend to be murdered by urban Black teenagers), and this is especially true in domestic homicides. In a domestic shooting and knifing victim study, 78 percent of the victims volunteered a history of hard-drug use, and 16 percent admitted using heroin the day of the incident. [169] If the victims have easy access to hard drugs, then it is reasonable to infer that the murderers, whether or not they had prior felony convictions, also have easy access to hand guns. If potential murderers can readily obtain hard drugs such as heroin, which has been strictly prohibited for nearly a century, the potential murderers would also likely be able to obtain handguns, of which there are approximately 66 million currently in private hands.

Accordingly, Dixon does not convincingly demonstrate that there would be any statistically significant reduction in handgun homicide until the day when obtaining illegal handguns became difficult for criminals, if that day ever came. The day when criminals have a hard time obtaining heroin is no closer than it was in 1911 then the Harrison Narcotics Act was enacted. Is it reasonable to expect that criminals will have trouble obtaining handguns 25 years *326 after the Dixon Handgun Confiscation Act is passed - the time period that Dixon proposes enacting the law to test its effectiveness? [170]

In determining whether the Dixon Handgun Confiscation Act would impact more criminals more rapidly than the Harrison Narcotics Act, it should be noted that in the countries with the handgun laws that Dixon admires, police administrators state that handguns remain easily obtainable by criminals. Australian criminologist J. David Fine conducted confidential interviews with high-ranking police administrators throughout Australia. [171] Handguns have been sternly controlled in Australia since the 1930s, and are available only to target shooters and collectors who pass a careful licensing process. [172] According to Fine, senior police officials unanimously asserted that "handguns were still readily available to criminals who wanted them." [173] One officer stated that a criminal need only walk into a certain pub in the State's capital city, "indicate that he wanted a pistol and wasn't fussy about make or calibre, and then wait about an hour." [174] Handguns have been regulated in Canada almost since the turn of the century, and subject to Australian-style laws since 1977; [175] the head of Toronto's detective unit opined that he would not have to walk more than two kilometers "to pick up a hot piece." [176]

B. The Substitution Effect

If handguns were somehow removed from the hands of malfeasants, would the death toll actually increase? Some gun misusers would switch to knives (not much less deadly than small handguns), while others would switch to rifles and shotguns (much more likely to kill than handguns). If enough misusers switched from handguns to long guns, the death toll might therefore increase, or so the "substitution argument" goes. Dixon confronts the substitution argument carefully, and provides one of the most comprehensive critiques of substitution theory ever offered by a handgun prohibitionist.

*327 Dixon is right to take the substitution argument seriously. While handgun wounds are usually survivable, especially if the victim gets medical attention quickly, shotgun blasts at close range are much more likely to be fatal. The shotgun fires a large slug, or from six to more than sixty pellets, with one trigger squeeze. A single shotgun pellet, because it may be of a diameter equal to a small handgun bullet, can inflict nearly as much damage as a small handgun bullet. [177] Wound ballistics and firearms experts concur that at short range, a shotgun is by far the deadliest weapon. [178]

Anti-prohibition writers such as David Hardy, Gary Kleck, and Don Kates have argued that a high level of substitution of long guns for handguns would occur in the case of a hypothetical American handgun ban. Dixon offers a careful rebuttal of their arguments, and concludes that (since he has placed the burden of proof on prohibition opponents) the case for a substitution effect has not been proven convincingly enough to overcome what he considers the strong evidence for handgun prohibition. Overlooked in the discussion of a substitution effect resulting from a hypothetical American handgun ban is non- hypothetical evidence from other countries.

As Dixon showed earlier in his article, countries with more handguns per capita tend to have more handgun homicides per capita. [179] Switzerland, which has, by world standards, relatively lenient handgun laws, has more handgun homicides per capita than countries where handgun laws are tougher. [180] From the handgun density/handgun homicide correlation in Switzerland and other nations (as well as from other evidence detailed supra), Dixon concludes that handgun density strictly correlates with handgun homicide. [181] Let us *328 assume that Dixon is right. In countries such as Australia and Canada, where handgun laws are much stricter than in Switzerland, the handgun homicide rate is lower than in Switzerland, but the total homicide rate is over 100 percent greater. [182] The reason cannot be that Australians and Canadians are more prone to want to kill somebody than the Swiss are - Dixon has explicitly assumed that human nature in developed countries is roughly similar everywhere. [183] So why then do Canada and Australia have more murders, even though they have stricter handgun laws, and fewer handgun murders? One plausible explanation is the substitution effect. A sufficiently large number of Australians and Canadians, unable to obtain handguns, do their shooting with rifles or shotguns; their victims die, whereas if they had been shot with handguns, many would have survived. Although some Australian and Canadian assailants, unable to obtain handguns, switched to less deadly weapons (such as clubs), the number of assailants who switched to rifles and shotguns was sufficiently large to increase the overall death toll. If we have plausible evidence to suggest that a substitution effect may have occurred in Australia and Canada, could a similar effect occur in the United States? [184]

*329 Dixon quotes research developed by Don Kates and Mark Benenson that if 30% of persons attempting homicide switched from handguns to long guns, while the other 70% switched to knives, total homicide would increase substantially. If 50% switched to long guns, the homicide rate could double, even if none of the persons switching to knives killed anyone. [185] A National Institute of Justice study of felons in state prisons found that 72% of the handgun criminals said they would switch to sawed-off shotguns if handguns became unavailable. [186] A 72% substitution rate would lead to an enormous multiplication of the current homicide rate, and Kleck expects that substitution would occur at about 70%.

Dixon retorts that criminals are apt to be braggarts and liars, and might claim that nothing, including a handgun ban, could stop them from committing any crime they chose. Accordingly, the 72% substitution figure might be too high. True enough. But at the same time, at least some criminals may be highly suspicious and mistrustful of authority. Although the National Institute of Justice polling, conducted through written response to written questions, offered the respondents anonymity, some of the prisoners might have believed that their responses would not in fact be anonymous; the polling might be a "setup" to discern their plans after release, and provide a reason for denying parole. Thus, some handgun criminals might have falsely said that they would not substitute sawed-off shotguns for unavailable handguns. Do the number of braggart criminals who falsely said that they would use sawed-off shotguns outnumber the number of mistrustful criminals who falsely said they would not? It is difficult to say with certainty. But since 72% of the criminals said they would substitute, and since only 30% substitution is needed to increase substantially the homicide rate, there is a wide margin for error to assume that bragging criminals outnumber suspicious ones.

Dixon critiques the Benenson and Kates estimate of a homicide rate increase because Benenson and Kates assumed that handgun users who did not switch to long guns would switch "downward" to the next most deadly weapon, knives. Almost certainly, some handgun users would, rather than using knives, turn to even less deadly weapons, such as fists, or would not attempt murder in the first place, absent a handgun. [187] But when calculating expected deaths resulting from substitution, Kates and Benenson assumed that none of the persons who switched to knives would kill anyone; in terms of resulting deaths, therefore, Kates and Benenson underestimated the *330 deaths that would be caused by murderers who switched downward to less lethal weapons. Even assuming that none of the persons who switched down killed anyone, the homicide rate would double if half of the handgun-deprived criminals switched "up" to long guns. [188]

Another tack taken by Dixon is to argue that high rates of substitution are unlikely because long guns are so inferior for most criminal purposes. He notes first of all that less than 10% of murders are currently perpetrated with long guns. [189] This is true, but, as Dixon strenuously argues, handguns are widely preferred as murder weapons, and widely available. Thus, it should not be surprising that more than 6 out of 7 gun murderers chose the "best" tool, a handgun. But what people choose when the "best" option is available does not prove how they would behave if only inferior options were available. Today, virtually all hard liquor drinkers consume the "best" hard liquor available - namely legally-produced hard liquor whose production is regulated by the government to guarantee standards of safety. Probably less than 5% of American hard liquor consumers drink bathtub gin, moonshine, and other home- brewed liquors whose safety cannot be guaranteed. Does the fact prove that very few liquor drinkers would, if legal liquor became unavailable, substitute home-brewed liquor? To the contrary, the experience of alcohol prohibition showed that a large percentage of liquor consumers, if unable to obtain safe, legal liquor, will switch to inferior, dangerous homemade liquor. [190] That murderers only rarely use long guns today does not prove that murders would eschew long guns if handguns were unavailable, any more than drinkers of legal liquor would eschew bathtub gin. [191]

As another argument against substitution, Dixon points out that long guns are less concealable than handguns. Even when sawed off, a shotgun is still about 11 inches long, making it slightly larger than big handguns, and much larger than the small, low-caliber handguns which are frequently used in crime.

Would sawed-off shotguns frequently be substituted in a *331 robberies? Putting an 11 inch shotgun in one's front pocket would not be very effective concealment. On the other hand, sticking the shotgun in the inner pocket of a large coat or jacket would seem reasonably effective. Accordingly, it is plausible to infer that persons who execute planned robberies would substitute concealed shotguns. At the same time, criminals who simply carried handguns with them, and spontaneously perpetrated robberies when the opportunity arose, might not be able to carry concealed shotguns so frequently. Thus, impulsive handgun robberies would suffer less of a substitution effect than would planned robberies. Since casual carrying of firearms in general might decrease, so might the shootings that result from the casual insults and provocations that can occur on the street. Hence, it is reasonable to conclude that an effective handgun ban might prevent some shootings. But again, only a 30% substitution rate would be necessary for total homicides to rise substantially.

What about in the home? It is the home, after all, rather than in robberies of stores, where the larger number of handgun homicides currently occur. Dixon argues that even in the home, the concealability of handguns is important. He asserts that substantial portion of the murders in 1989 involved "friends or acquaintances who may have been unaware that the person they are visiting is carrying a concealed weapon." [192] First of all, there is no evidence as to how many of those murderers actually were carrying a concealed weapon of which the victim was unaware. For the sake of argument, assume that all of the murders would have been prevented had handguns not been available; there is still a long way to go for the substitution ratio to be reduced below 30%, and thus not cause a net increase in homicides.

Next, Dixon writes that "the ease of pulling out the [hand]gun and shooting makes such arguments far more likely to spill over into murder. In contrast, by the time the assaulter has gone into another room to retrieve their [sic] long gun and loaded it, the potential victim has crucial seconds in which to escape." [193] Here, Dixon assumes that the domestic handgun murderers were carrying the handgun on their body, rather than storing the handgun in another room. He likewise assumes that the substituted long gun would be stored in "another room" rather than the room in which the argument was taking place. He further assumes that the handguns used in the domestic shootings were loaded, but the substituted long guns would not be loaded. All of these assumptions may be simultaneously true some of the time, thus making Dixon's escape scenario plausible in *332 some instances. (Although not every potential victim would know that the potential murderer was loading a long gun in the other room, and even then, some might not run away.) [194] But it is highly speculative to assume that Dixon's scenario of the unloaded long gun in the other room replacing the loaded handgun carried on the person would be the predominant scenario. Even if we speculatively assume that the unloaded long gun scenario would transpire more than 50% of the time, all that is needed for an increase in the death rate is a 30% substitution rate.

While Dixon argues convincingly that substitution would not be universal, the evidence easily supports the conclusion that substitution of long guns for handguns would occur in at least 30% of current handgun murder situations, thus leading to a substantial increase in total deaths.

C. Public Safety Benefits of Handguns

As Dixon acknowledges, handguns provide, at least occasionally, protection from criminal attack. Hence, handgun prohibition would deprive some crime victims of a self-defense tool, and a realistic calculus of the benefits of handgun prohibition must therefore consider the harm to public safety that would result from leaving victims with less ability to protect themselves. Rather than simply scoffing at the idea of armed protection, Dixon carefully analyzes the data regarding self-defense handguns, to build his case that self- defense is infrequent enough as to not be a bar to handgun prohibition.

1. Studies of Criminals and Victims

Dixon first of all notes the National Institute of Justice prisoner surveys, in which large percentages of prisoners personally reported being personally deterred from committing a crime because of fear that the victim might have a gun. [195] Dixon critiques the polling by noting that criminals are not particularly honest. But elsewhere Dixon suggests that criminal dishonesty will tend in the direction of "macho" bravado. [196] Accordingly, it does not seem intuitively obvious that a significant percentage of the prisoners who said that they personally had been scared off by a victim who might have a gun were lying, and claiming that they had been scared when in truth they were not.

*333 Dixon also points out that the felony prisoners may be different from criminals as a whole, in that the polled prisoners had all been caught, whereas non-polled criminals who have not been caught may share different characteristics, such as being shrewder. [197] True enough. But the shrewder criminals, one might expect, have not gotten caught because they are more adept than the prisoners were at avoiding risky crimes. Accordingly, it might be that the unapprehended criminals were more risk-averse than the apprehended prisoners, and were therefore more likely than their apprehended fellows to avoid crimes where there was a risk of a victim with a gun.

Moreover, to recognize that there are problems in making precise extrapolations from prisoner surveys does not undermine the fact that 34% of the prisoners personally reported being scared off by an armed victim, and 8% said that the experience had happened frequently. Even if the polled prisoners are representative only of the type of criminals who get caught and sentenced to felony time in state prisons, and even if half the prisoners who said they had been scared were (in a perverse display of anti-machismo) lying, a large percentage of criminals who abandon particular crimes because of armed victims remains. One may nibble at the edges of the percentages without changing the core reality.

A second piece of evidence regarding the frequency of armed defense is Gary Kleck's estimate of 645,000 defensive uses of handguns annually, a figure based on polling of gun owners. Dixon offers several valid critiques of simplistic reliance on gun owners' response to polling data. First of all, some gun owners may have a vested ideological interest in claiming defensive gun use. [198] In addition, some gun owners may have perceived a threat where none existed, or may have reacted with excessive force. [199] Thus, as Dixon points out, not every gun owner who told a pollster that he used a handgun for legitimate self defense may have been truthful or correct. Yet while there were some factors which may have resulted in the Kleck figures being too high, the figures are more likely too low, as a result of a number of factors inherent in the polling process, and the conservative assumptions made by Kleck. [200]

*334 To begin with, it is important to note where Kleck got his data: from a survey conducted on behalf of the pro-control National Alliance Against Violence. The NAAV had hired Peter Hart, a leading Democratic pollster, to survey Americans on guns. The widely-reported "interviewer effect" in polling results in some respondents attempting to match their answers to fit with the interviewer's perceived social values. As a general matter, interviewers for polling firms tend to come disproportionately from segments of the population which favor gun control. [201] And it is possible that in some cases the background biases of the organization paying for the poll, and the organization conducting it, may have been perceived by some respondents. As a result, the "interviewer effect" could tend to depress the numbers of persons claiming to have used guns in self-*335 defense. In addition, the context in which questions are asked can have a significant impact on respondent behavior. If argumentative questions which make the case for gun control precede questions about gun control policies, then respondents are more likely to give pro-control responses. [202] In the case of the National Alliance Against Violence question series, the interviewer's questions included a lengthy battery of questions gauging response to the "fact" that "The Supreme Court has ruled on several occasions that the Constitution does not guarantee individual citizens a right to personally own firearms but that it does guarantee the right of each state to form a state militia or State National Guard, and that either states or the federal government could control or prohibit firearms ownership by individual citizens if they wanted to." [203] (Of course the Court has done no such thing; the Court's few rulings on the gun control issue point towards recognition of an individual right.) [204]

While question wording can further bias a sample, the actual question that Peter Hart used appeared to be written in a neutral manner: "Within the past five years, have you yourself or another member of your household used a handgun, even if it was not fired, for self-protection or protection of property at home, work, or elsewhere, excluding military service or police work?" Six percent answered "yes." Follow-up questions revealed that 3% of the defensive users has used the handgun against a person, 2% against an animal, and 1% against both. [205] (That 4% saying "yes" to defensive gun use against persons meant that about 10% of households where a handgun was owned for protection had actually used the handgun for protection.) [206]

Kleck took Hart's number of 4% "yes" for use of a gun in defense against a person in the last five years. Kleck made the conservative *336 assumption that each "yes" related to only one gun usage in the last five years - that no household used a firearm for self-defense two or more times in the five years. Because 4% of American homicides amounted to 3,224,880 households, Kleck divided by five (since the question had asked about usage in the last five years) to arrive at an estimate for the annual number of uses of a handgun for self-defense: 644,976. (Or roughly once every 48 seconds). The 95% confidence interval for Kleck's estimate is 468,000 to 822,000 uses per year. [207] By assuming that each "yes" answer related to only one defensive gun use per household, Kleck may have substantially underestimated the frequency of total defensive gun use.

In addition, it must be remembered that Kleck's figure of 645,000 annual defensive gun uses was based only on the 4% of respondents whose households had used a gun for protection against a person. Another 2% had used a handgun for protection against an animal; annual defensive uses against animals would amount to about 322,000, by Kleck's methodology. Since a person who is being attacked by a pack of stray dogs has as much legitimate interest in escaping serious bodily injury as does a person who is attacked by a criminal, defensive gun uses against animals should count in the calculus of the social utility of handguns. Accordingly, we have nearly one million total defensive handgun users a year, against persons or animals. Even if we assume that many gun owners lied or misperceived their need for self-defense, Kleck's conservative assumptions have at least partly corrected for respondent misstatement. While there is room to question whether the proper estimate of defensive handgun uses should be 645,000, or 967,000, or perhaps a figure hundreds of thousands less, it still seems likely that there are, at least, several hundred thousand defensive uses of handguns in a given year. Dixon's handgun prohibition plan, if successful, would eliminate them all.

Using data from Detroit and Miami, Kleck has projected that there were about 1,427 to 2,819 self-defense killings in 1980. [208] Dixon points out that even the highest figure of 2,819 defensive killings is still less than the 13,650 murders that occurred that year with firearms, and therefore "Kleck has failed to prove that the self-defensive use of firearms outweighs their abuse in homicide." [209] *337 First of all, as Kleck had explained, the figure of 13,650 murders is based on arrests rather than final disposition. Therefore, most of the 1,400 to 2,800 annual defensive homicides should be subtracted from the total to about 14,000 criminal homicides. Even after the defensive homicides and criminal homicides are properly weighed, there are still more criminal killings than self-defense killings. [210] But before weighing the efficacy of Dixon's handgun prohibition, we must consider whether the prohibition would eliminate all of the criminal and non-criminal homicides (thus producing a net saving of innocent lives), or if the prohibition would eliminate only some of each type of homicide. If the prohibition reduced defensive homicides by 1,000, but reduced criminal homicides by only 200, society would be worse off rather than better.

Turning to Kleck's last piece of evidence - surveys of robbery and assault victims - Dixon acknowledges the evidence that use of a handgun against a criminal attack is the option most likely to result in failure of the criminal attack, and in the victim sustaining no injury. [211] Dixon does not dispute the evidence, but he does note that many of the persons using handguns for self-defense may have been defending themselves against criminals with handguns. "A heavily-armed citizenry might be a rational response if heavily- armed criminals were inevitable; but far more rational would be a society which strives to disarm all private citizens, thus obviating the need to use firearms in self-defense." [212] Here, Dixon makes an unjustified leap in reasoning. Even if no criminals had handguns, handguns would still be useful, and often necessary for self-defense. As Dixon acknowledges in the next section, a handgun allows a person who is smaller than the attacker to protect herself. A handgun likewise allows a person attacked by several criminals at once to keep them all at a distance. And again, Dixon fails to perform the crucial step in his utility calculus of determining whether the number of criminals who are disarmed will exceed the number of crime victims who are disarmed.

*338 2. Women

Dixon's treatment of the issue of armed self-defense by women is, again, remarkable for its willingness to confront serious evidence. He acknowledges the vast body of evidence that firearms are especially useful for defense by women, since they may be less likely to be able to match a male attack with sheer physical force. [213] When an abusive male who has perpetrated numerous felonious assaults against a woman living with him is shot by the woman, Dixon recognizes the shooting to be legitimate self-defense against criminal attack, rather than a "tragic domestic shooting during an argument." (Some judges are, unfortunately, not as progressive as Dixon, and too many women engaged in legitimate self-defense are convicted because standards for self-defense are still based on the context of a fist-fight between men of equal strength, rather than the special circumstances of battered women. [214])

*339 Dixon's response to the self-defense needs of women, particularly battered women, is two-fold. First of all, he points out that there are other ways that women can protect themselves, such as by obtaining restraining orders. [215] But these methods are not even close to sufficient to guarantee that a woman will not be confronted or cornered by an angry ex-mate ready to batter, rape, or kill her. Indeed, most of Dixon's alternative protective steps are premised on the woman being able to leave the home where she is being battered, but it is an attempt to leave that often precipitates a homicidal act by the male. [216] Before the legal system forces women to rely on the protection of the government, the government should begin to offer realistic protection - something far superior to the unenforceable protection of today's temporary restraining orders. Even California, a state noted for its strict gun laws, allows people