Evidence Based Or Ideology Based? Policy Choices

The 1995 legislation also moved about half of all handguns into a “prohibited” category based on technical details such as calibre or barrel length. Although there was no evidence of any kind that guns with these technical details were in any way more prevalent in criminal gun use it was felt that they “might be” and therefore must be banned. Somewhat irrationally, these “too dangerous”, now prohibited guns were then left with their owners who could buy and sell them freely with other “prohibited” gun owners. It became a criminal offence for anyone else to possess one. (This s one of the biggest “registration to confiscation” events that has ever happened in North America, about half a million guns were confiscated, even though there current owners can keep them. The guns are to be confiscated when the current owner dies.) Calculating Cost/Benefits The number of guns to be registered is an important variable since the more guns there are, the more cost and effort is involved in the registration process. In Canada the number of guns in the country is the subject of some debate. In 1974, a Statistics Canada survey established the number at 11,186,000. Historical import/ export estimates would place the current number at anywhere from 18 to 21 million. In 1991, the Justice Department estimated the total number to be 6 million. For the purpose of this analysis, I have chosen 7 million as the number of guns to be registered. This estimate is on the low side and may underestimate the registration effort required, as such it slants the analysis in favour of gun registration. One measure of cost/benefit is the ratio of total effort to effective effort, in this case the ratio of all guns registered (total effort) to registered - and therefore traceable – crime guns (the “effective” against crime bit) We can calculate this ratio in a number of ways. Homicides In 2003 there were 548 homicides in Canada, 161 resulted from shooting. Assuming that each shooting involved a separate gun these figures yield the following cost-benefit ratio. Assuming there are only 7 million guns we get the following calculation. 161crime guns/7,000,000 not-crime guns = 0.000023 or .0023%, a cost-benefit ratio of 23 ten thousandths of 1% Therefore, in terms of the effort expended, 99.9977% of our effort was wasted registering non-crime guns and only .0023% of our effort landed on guns used in homicide (our crime prevention pay-off). Crimes of Violence We can expand this analysis to include the more common “crimes of violence.” There were 302,000 “crimes of violence” in Canada in 2003. Statistics Canada indicates that 5% of these crimes “involved” a firearm.* Assuming that each incidence involved a different firearm, this works out to (302,000 X .05) 15,100 firearms “involved” in a “crime of violence” (includes homicides). This yields the following cost-benefit ratio. Low estimate 7million guns 15,100/7,000,000 = 0.0022 or .22%, a cost-benefit ratio of 22 hundredths of 1% Accordingly, 99.78% of our effort was wasted on non-crime guns and only .22% of our effort affected guns used in crime. This analysis highlights the basic cost/benefit problem of gun registration as crime control. Even if we make the clearly incorrect assumption that each crime involves a different gun the best “return on investment” we can hope for is that less than ¼ of 1% of our efforts will involve crime guns. Arguably, this is a inefficient way to spend scarce law enforcement resources. The Registration Problem The above analysis contains other assumptions besides the one gun per crime assumption. It assumes that all crime guns are registered. It should be noted that this is a crucial assumption at the very centre of the debate. The whole gun registration equals crime control argument rests heavily on the assumption that gun registration will apply to the guns of criminals as well as the guns of law-abiding citizens. The percentage of crime guns that are registered is subject to considerable variation. Only 31% of guns used in homicides are recovered and of these only 28% were registered. Citing another example, we can note that the Department of Justice reported that 70% of guns “recovered” from criminals in Metropolitan Toronto were registered. This gives us a wide variation in crime gun registration rates, from 28% (recovered homicide guns that were registered) to 70% (guns seized from criminals). We can identify the best-possible return for our efforts if we pick the higher registration rate (a gun registry has zero connection to guns that are not registered). If we apply this best-case solution and use a crime gun registration rate of 70% we find that our most favourable “return on investment” becomes (.0022 X .7 = 0.00154) .154% or about 15 hundredths of 1%. Therefore 99.846% of our effort will be wasted on non-crime guns. A best case “return on investment” ratio of 15 hundredths of 1% indicates that gun registration is a rather inefficient crime control measure. The Connecting the Guns Problem The above analysis is based on another assumption. It assumes that, in every case, we will be able to “link” each registered crime gun to the criminal involved. This is a critical assumption. The entire registration equals crime control argument rests heavily on the assumption that the police can use the registry to trace registered crime guns back to criminals. (A registry cannot be used to trace unregistered guns, nor can it be used to trace guns that have not been recovered for identification.) This raises real problems. Not all crime guns are recovered for identification. Accordingly, we must conclude that the 15 hundredths of 1% best case “return on investment” is too optimistic and must be reduced even further to factor in unrecovered crime guns. However, there are still the registered guns that are recovered so we can argue that there is still some return on investment. Unfortunately, taking the next step, moving from gun recognition to evidence against a criminal is not always easy. In order to use a registry to “link” a crime gun to a particular criminal all of the following things must be true: the gun must be left behind at the scene of the crime, or otherwise recovered; the gun must be linked to the crime by being found at the scene or by ballistic evidence; the criminal must have registered the gun, using his true name and identity; the gun was not stolen, which would break the link between gun and owner leaving the criminal unidentified; and the criminal who was the registered crime gun’s owner did not claim that the gun had been stolen or lost prior to the crime. Obviously, not all of these conditions are true in every gun crime. Our best case “return on investment” already below 15 hundredths of 1% must be reduced again. Conclusion The above is a statistical analysis of the cost/benefits of gun registration as a crime control tool. Since it is good analytical practice to cross check statistical analysis against other known empirical data we should confirm our findings with an empirical test. Fortunately, here in Canada we have a readily available existing gun registry that can be used as a cross check. In Canada, handguns have been strictly registered since 1934, sixty-one years later in 1995 the Department of Justice admitted that they could not identify a single instance where the handgun registry had ever “helped” solve a crime. Accordingly, our analysis would indicate that gun registration is not a cost/effective way of reducing crime or spending our limited crime control budget. *(The Statistics Canada number (5%) is misleading: guns “involved” in crime are guns picked up by the police from a crime scene - say ones that were found in a closet. Guns are actually used in violent crime about 1.9% of the time. The higher number is used because it was the one used in Parliament to justify the long gun registry.) [Proofreader's note: this article was edited for spelling and typos on February 19, 2006]