The Science of Everything Podcast - Episode 40: Gun Control

Episode Date: December 18, 2012

A science-based discussion of the issue of gun control in the United States. After reviewing some basic statistics about crime and gun ownership, I examine the empirical evidence concerning the relationship between gun ownership and violent crime, highlighting the many difficulties associated with such studies and the limitations of this literature. I then discuss some more specific aspects of the gun control debate, including the question of who commits most violent crimes, how deadly guns are compared to other weapons, and how criminals acquire guns. I conclude with a discussion of the evidence for and against various types of gun control measures, and some thoughts on contemporary US gun culture and its impact on gun policy. If you enjoyed the podcast please consider supporting the show by making a paypal donation or becoming a patreon supporter. https://www.patreon.com/jamesfodor https://www.paypal.me/ScienceofEverything

You're listening to The Science of Everything podcast episode 41. Gun Control. And I'm your host, James Fodor. So this is somewhat of a special episode in that it's the first episode ever to be inspired by actual current events in the world. In this case, the recent tragedy at Connecticut, in Connecticut, in the United States. So what we're going to be looking at in today's episode is the science and research that's been done on the issue of gun control and also the relationship between guns and violent crime. Now, there's certainly a lot of ideology that's put out there in the news and other sources about this issue. but in my view, far too little actual science.

And so that's what we're going to be looking at in today's episode. Specifically, we're going to be looking at, first of all, I'll just outline the two basic sides of the argument in terms of guns, increasing or reducing violent crime. Then we're going to talk about some of the problems with measuring the relationship between crime and gun prevalence. We'll also then go into look at some of the results of measurements and other studies that have been done and what they can and cannot tell us.

After that, we're going to look at a few more specific aspects of the issue, including various more specific hypotheses that have been made about guns, including what I call the average, Joe thesis that guns are the hypothesis that guns are more deadly than alternate weapons that criminals might use. Whether or not guns are actually useful for home defence and the issue of whether the US has a gun culture that's significantly different from other countries. We'll also look at this some of the evidence on where criminals acquire their guns and what

can be potentially done to stop that. And finally, we'll end the episode by looking at some potential principles of regulations of firearms that may be evidence-based in terms of reducing violent crime. and so we'll look at both what he is evidence-based and what he's not. Okay, so let's get started. The two basic arguments about gun control are either that guns reduce crime or guns increase crime. The anti-gun argument is that a criminal with a gun is a more efficient criminal. They're more likely to be able to get the victim to do what they want by just simply threatening with the weapon.

And if they do choose the weapon, it will be able to do more damage, and therefore more likely to kill or seriously injure the person. A related issue is that having a gun could emboldened individuals to take more risk or engage in more risky or criminal. activities that they otherwise would not do. Yet another part of this argument is that individuals in, say, domestic disputes or other arguments in the presence of guns are more likely to boil over and shoot someone possibly killing or injuring them than they would be if they did not have such deadly weapons within such easy access. So these are some of the sort of arguments for why guns might increase violence and violent crime. On the other side is the argument that people also

acquire guns for self-defense and that the larger number of people there are in general society who have guns, the more easily people are able to protect themselves from criminals. And so the argument is that by being able to thought criminals in this way, it may be actually possible that more guns can reduce the amount of violent crime, because people can defend themselves better and warn off criminals. So these are the two sides of the argument, basically. I think I'll just say right now that there is an element of truth to both of these sides, as you might have expected. However, that doesn't mean that the truth is somewhere exactly in the middle. So we'll be unpacking that idea about exactly what elements of these two stories are true and not true throughout podcast.

Having presented the two basic sides, I now want to just present some basic statistical information about accidents, homicide, and gun prevalence in the US. By the way, most of the studies and facts I'm going to be talking about in this episode relate to the United States, given that its number of firearm homicides are something like similar to the entire rest of the developed world combined. So, firearm violence is much more prevalent in the US than elsewhere. And it's also a much greater political issue there than elsewhere. And also, finally, the available evidence mostly relates to the US.

And so if I was focusing on a different country, they just wouldn't be enough evidence to really say very much. There's a little enough as there is, as we'll see in a moment. Regarding the US, the number of guns in the US is, according to my statistics, around 270 million guns, which is around 89 guns for every 100 people. This is by far the highest ratio of guns to people in the entire world. For comparison, according to my statistics, the second highest developed countries, developed country that in terms of guns per capita is Switzerland, which has about 46, so only about half as many as the US does. And many developed countries like Sweden, Norway, France and Germany have only around 30 guns per capita.

Others like UK and Australia have less than 10 guns per capita. Sorry, this is guns per 100 people, so much of the capital. So we can see from these statistics that the US clearly has dramatically more weapons per person than any other country in the world, even underdeveloped countries. And especially compared to developed countries, the ratio is just enormously higher. So that's gun data. In 2005, there were about 10,000 gun homicides in the US, about 17,000 gun suicides and about 800 accidental gun deaths. The numbers fluctuate plus or minus a few hundred or even a thousand or so

for those, what, for gun homicides and suicides every year, a few hundred maybe for accidental gun deaths, but those statistics are fairly representative. So what we see is something like twice as many gun suicides as gun homicides, not quite twice as many, but on the order, and many, many fewer accidental gun deaths. So based on the fact that there are so few accidental gun deaths compared to homicides, mind you, 800 is not trivial, but it's quite small compared to 10,000. I'm going to focus in this episode about the impact of firearms on homicides and not talk so much about accidental deaths. And as I just said, the number of gun suicides is dramatically higher than the number of gun homicides. So you might think that it would be more important to talk about suicides.

The reason I wanted to focus on homicides is one, because this is an area that's been researched somewhat more than suicides. And the second issue is that suicides that has a great deal of mental health and other public health aspects to it, which would fit better into the context of a different episode. In 2007, the overall homicide rate in the US was about six per 100,000 This is the standard way that homicide is measured per 100,000 people because it just gives us easy numbers to work with. Now, this is, as you might expect, much higher than most developed countries who have a level of around 1 to 2, UK, France, Germany, most of them have around about 1, Japan's less than that, fewer somewhat more. But however, you pick at the US is several times higher, probably three to four times higher than the average for developed countries.

Interestingly, however, the homicide rate has been falling for the last 15 to 20 years or so. it peaked in the early 90s, so that's one good thing. But a very interesting phenomenon is that if you break up the homicide rate by race in the US, you see a dramatic, dramatically different picture. For white non-Hispanics, the homicide rate, this is in 2007, was only 2.7. Now I say only 2.7. That's still a lot higher than most developed countries, but it's a long way down from six. However, for non-Hispanic blacks, the homicide rate was about 23,

which is the same level as many African countries and other, you know, South American and other undeveloped countries. So we see a very clear bifurcation here in terms of homicide biracial groups. And some of the reasons for that we'll look at a little bit later, but I think it's almost certainly socio-economic. And I think this gives us a clue as to some of the reason for the gun problems in the US, because actually, although, as we've just established, the rate of homicides among blacks is something like eight times the rate of homicides amongst non-Hispanic whites, about 44% of whites are born owning a gun, that is, that there's a gun in their home,

compared to only 27% of blacks. In other words, fewer blacks than whites own firearms, but many, many, more blacks than whites are killed every year. And that's not because blacks are killing whites, or whites are killing blacks, actually, the overwhelming majority of homicide, something like 90% stay within the same race, if we want to use those terms. In other words, whites mostly kill whites, blacks mostly killed blacks. It's rather somewhat odd that that's the way it is, but that's what the statistics show. In the black population, if we want to talk about it in those terms, there are fewer guns, but many more homicides per person than in the white population. This is a point that I

want to touch on later again. One final statistic before we dive into the problems of measurement, around 60 to 65%, I mean, give or take, depends on the year. Of all homicides in the US are gun homicides. So maybe you'd say around two thirds as a rough figure, somewhere between half to two thirds. And about half all suicides are committed by gun as well. So those figures I gave before, about 10,000 gun homicides and 17,000 gun suicides, those are just for guns, not the total level. But the majority of homicides and about half suicides in the US are by gun. And this is much higher percentages than in most other developed countries.

Okay, so that's some basic statistics, and I'll keep bringing those up every now and then just to, to set the stage and help to frame the subject. Moving on to the problems of measurement. So you might think it would be relatively easy just to count up the number of guns in a given US state or country or something, or county or whatever, and count up the number of homicides, compare the two and see if there's a relationship.

And that's broadly what people try and do, but there are many, many complicating factors. First and foremost, most states in the US, and certainly the federal government, do not have any reporting requirements for firearms sales or ownership. So in other words, we don't know how many guns there are in US states or in most US states

and certainly in counties. We don't have very good data on this. We have a fairly good idea of how many guns there are in the US as a whole, in large part because we know how many are manufactured and not too many are imported, and we can also work that out. But in terms of who owns them and where they are in the US, we really don't know. We don't have good data on this because there's no national registry,

or most states don't have registries either. Most countries do, so it's much easy to get data on this. So this is a first and foremost problem. We don't know exactly where the guns are in the US and who owns them and what states have more and so on. There are other problems as well. Generally, the most common measure that is used is survey data where researchers call up or sometimes talk to people in person and ask,

is there a gun in your household? That's the general way of doing it. And by the way, the statistics there vary between sources, but it's somewhere around like 40%, although I've seen as low as 30 and as high as 50. So we don't know how many households in the OS have guns, but it's somewhere between 30 and 50%. I think 40% is probably roughly it. But the problem with this measurement is that it doesn't say anything about how many firearms there are in a particular home or what type of firearms there are. The highest risk type of firearms are handguns, that is basically pistols or similar small

weapons, which are easier to conceal and are preferred by criminals generally, and these are the most common guns used in homicides and also suicides. So they're the ones we're most concerned about, but when people say they have a firearm at home, we don't know if that's a shotgun, or if it's a hunting rifle, or if it's an assault weapon, or a handgun, or what? And that actually makes a big difference because some weapons are much more dangerous than others, and are much more likely to be used for criminal activities. It also doesn't tell us who has the guns in a given area. For example, if we find out that 50% of people in given county have guns and only 30% of the people in an adjacent county have guns, or percent of

households, it could well be that the guns are differently distributed within the two different counties, such that criminals or mentally unstable persons are more likely to have them in the 30 gun county than in the 50% gun county. It's perfectly possible, and we don't really know it because we don't have good data on this. Yet another problem when we're trying to measure these things is that the number of guns that people have may, well, in fact, we would expect it to be endogenous with respect to the levels of crime. Now, this is a technical term in statistics, but the basic way of thinking about it is it's quite possible that causation runs in multiple directions. That is, it could be that areas that have more guns have more crime

because guns lead to crime, or also suicides, but we'll think mostly about crime from now on. So it could be that guns cause crime. However, it could also be that areas that have more crime, people buy more guns because they're more interested and more concerned about defending themselves. Or it could also be, another possibility, is that areas that have more crime enact more stringent gun control laws, which then make it more difficult for people to get access to weapons, and therefore reduce gun ownership. So in this situation, it could be the case that you see a correlation between low numbers of guns and high crime, not because guns stop crime, just because the high amount of crime led to street of gun controls.

So there's many potential ways that the causation could run. And just by running a correlation, that is just by measuring firearm prevalence and level of crime and comparing the two in different counties or different states in the US or even different countries. And seeing if there's a relationship between those variables, doesn't tell you which way that the causation is going. And this is a massive problem which really is not very well addressed

in the literature at all, as we'll see as we look at some of these other studies. Another problem is that there are many other factors that affect crime aside from similar gun levels. So you can have one county or one state that has many more guns than an adjacent state, but it might have a lower crime. Suppose guns do cause violence, even if that was the case, you still might not see a very strong relationship because there's so many other things that cause crime as well, including levels of poverty, inequality, socioeconomic factors, education, and job, job possibilities,

quality of the police force, health services, and on and on. If you don't hold these other factors constant, then you're going to see false correlations in the data potentially, or not see correlations when there actually should be. Another possibility is that instead of looking at two different states or two different counties or two different countries at the same time and comparing them. That's called cross-sectional analysis, where you look at multiple entities at the same time. We could look at the same entity over a period of times.

So look at how much crime there was and how many guns there were at a given state in 1980, and they look again in 1981 and again in 1982 and so on. This is called time series analysis, or cross-section, you go over time. There's problems with this as well, though, of course, because suppose the gun control laws are enacted, and that reduces the number of guns in a given state, and suppose crime goes down? Well, did crime go down because of the gun control laws,

or did it go down for some other reason? because there's many things that are going to be happening over the course of time. You know, police funding and resources are likely to be changing. Socioeconomic situations are changing. The state of the economy is changing. The way the judicial system works is likely to be changing as well. The situation in the drugs market, which is a heavily influence on guns and vice versa, could well be changing as well.

There's so many other factors that could be affecting the crime rate that you can't say just because gun controls were enacted and crime went down or even if crime went up, that one caused the other. So with these multitude of problems, you can see why it's so difficult to know what's actually happening, what the actual relationship between violent crime and gun ownership is. One way that researchers have adopted to deal with the problem of lack of statistics about how many guns different areas have, like a prevalence of gun ownership, is to use proxies.

A proxy is a measurement that you use instead of the thing you actually want to measure. So what we actually want to measure is the prevalence of gun ownership, but we can't measure that directly. So instead, we take something else that is related to the thing we want to measure. It's not the same, but it's related to the thing we want to measure, and we use that data instead. And so there's a number of proxy variables that are used instead of gun ownership. And of course, the question you have to ask when you're dealing with your proxy measurements is how valid is the proxy. If we used hair color, for example, or eye color as a proxy for gun ownership, that's a really bad proxy because it has nothing to do with gun ownership. So

if you measure eye color, would it be measuring nothing? So what kind of things might you pick? Well, one thing that's been chosen is that's been used is the percentage of suicides that are committed via guns. The idea is the more prevalent guns are, the high percentage of suicides will be committed with guns. And there's some, there's some justification for this. Another proxy that's used is membership of the NRA, the National Rifle Association, which again you would expect to be related to the number of guns people own, yet another one is subscription to gun magazines like guns and ammo. You would again expect that to be related to the number of guns that people own.

And so this is one way that researchers have tried to get around the problem of inadequate measurement of guns by using these proxy variables. But of course, the trouble is that it's hard to know exactly how reliable a given proxy is and how reliable it is might differ between different states. The only way you can tell how reliable a proxy is is to take data when you know both gun ownership and the proxy data, let's say magazine subscription and gun ownership, you know it directly in this state, and you compare the measurements and you see, okay, they're pretty close to each other. There's a pretty good

correlation. You look at that at a few states where you've got the data, and then you just sort of assume that the same relationship applies in the states where you don't have the data. So it's like, in states one through ten, guns and ammo subscriptions and firearm ownership correlate very well together, so they are strongly related to each other. So we're going to assume that they are also strongly related to each other in states 11 through 20, where we don't have data on gun ownership. And that's an assumption that you can't really prove. And it might not be true because there could be systematic differences between the two groups of states. And so if you don't like someone's research, you can just say, well, the proxies that they

used was rubbish, and there's many other things you can say as well. So I'm trying to emphasize how difficult it is to make accurate measurements of this sort of thing and how much skepticism we need to have and reservation we need to have in any of the results we get from many studies. And one other thing, you remember before how I said that you have to, that there are many other factors like education and income and socioeconomic factors and all other things like that that also affect crime? Well, in order to control for these extraneous variables, which we want to get rid of,

because we're just interested in the effect of guns on crime, not all these others. We need to introduce control variables into our equations. So, for example, what we do is we measure income, let's say, or income inequality, and we say controlling for that, what is the effect of guns on crime, or vice versa, crimes on guns, in a given state or any given area. The trouble with that is that which control variables do you include? In order for your results to be valid, you need to include all the relevant control variables. But if you include some that aren't relevant, that can lead problems.

But if you don't include some that you should include, that can also lead the problems. And also if the control variables aren't measured very well. So, for example, you want to control for how racially divided given city is or how much of a job opportunities there are or levels of trust in that city. There aren't very good measurements for any of those, especially levels of trust, for example, so you have to use proxy variables for those. But if you've mismeasured these control variables, then that could also stuff up your results. So there's so many factors that could affect a validity of the results for these type of things. Having made all of those many qualifications, we're finally ready to start looking at the actual

evidence. To put it bluntly, the evidence is pretty pretty crumbed. A number of studies have been done, but not nearly as many as they should, given the importance of this issue, and many of the studies are not very good quality. Bottom line seems to be that using regression results that is trying to compare the levels of crime and gun ownership and then controlling for all these other variables. This is called a regression analysis. The majority of these seem to indicate that at least within the US, comparing different states in the US and comparing different cities in the US, higher levels of gun prevalence are associated with higher homicide rates, primarily due to higher firearm homicide rates. that is, it seems that more guns are correlated with more firearm homicides, but not with other types of homicides. However, this is by no means a extremely robust result, which means that

there are many studies that don't show any such relationation. Not too many studies show that a negative correlation, that is, that more guns are associated with less homicides. Almost always the result is either positive relationship or no relationship. Now, the level of the number of positive relationships that you've seen these regressions are generally higher than the number of the number of zero relationships. However, you have to consider publication bias, especially because a lot of these studies are published in health journals and other things like that, where you suspect there may be a sort of a bias on the part of the researcher against guns in favour of gun control. That doesn't invalidate the study, it just raises the bar of evidence a little bit higher.

I still think that there is a clear trend in the studies towards a positive correlation, and there's very little evidence that guns reduce levels of homicides. This, though, only really applies to homicides. There's very, very little evidence that gun prevalence is associated with higher rates of robbery or aggravated assault. so other types of violent crime. The main effect seems to be on homicide. Again, that doesn't mean there is no relationship with, say, robbery or assault. It's just the data for these is much, much weaker, and indeed, they're studied much less. So I'm going to primarily talk about homicide where the evidence is strongest. There are many studies that I've read and got referenced, and there's too many to go

through individually, and many of them are a similar story as to each other, so it gets a bit repetitive. They try and find a good measurement of gun prevalence. They try and find a good measurement of homicide. That's not as difficult, and they compare different cities in the US or different states in the US, and a few of them do international comparisons as well, although the international ones aren't good because it's really hard to get data for these. Some of the best ones actually compared very similar cities in Canada compared to in the US, which were, that is the cities were similar to each other in terms of level of economic development, in terms of geographic location, and in many other aspects as well. The main difference is that the ones in the US have

much high levels of gun ownership, and there's also generally racial differences between the cities. I can't specifically remember which cities were compared. The reason this is a good study, because it's a fairly good control. We've got two different cities, the main difference between them being that one has much higher levels of guns than the other. So it's a relatively good control. Of course, it's not perfect because there are other differences between Canadian and US cities, even fairly similar ones, but it's one of the better studies that I've seen, and this did demonstrate positive relationship between guns and crime. Another really good study that I read was much better than many of the other ones, because they didn't just run

one or two regressions. That is, they didn't just do the equation a couple of times. They actually did it something like a thousand times. The reason this is good is because, remember, before, I was talking about all the different control variables you have to include, and it's not always easy to tell how to measure the control variables or which ones to include, and a source of remember I talk about the different proxy variables you can use instead of, you know, in order to try and measure how many guns people have, and there are also other factors that you can vary as well

that will affect your results. Well, instead of just doing one or two of these and presenting and publishing that as a paper, which many of the studies do, this one did about a thousand of them, changing the way they measure things and the variables they include and the structure of the model and so on each time. And this is good because it allows you to see how the way we use is robust, so that is how

robust the results are to change in the model specification. So if you only get, if you do lots of regressions, you run lots of equations, you only get some of the results being significant, that is some positive results, then it's not very persuasive because you expect some positive results that come out just by chance. And you don't really know if the research around 10 equations or 10 regressions and picked the one that happened to come out as significant. This is not that uncommon in the literature, unfortunately, and so it's a problem that you have to be wary of. A study like this that does a thousand

regressions and tells you about the results for all of them is a bit more credible in that respect. So in this particular paper, they were mainly comparing measures, various measures of gun ownership, as I said, they used multiple proxies compared to various indexes of barrenes. crime. Specifically, what we're interested in is the one about homicide. So they did 1,024 regressions and they got 190 significant results at the 10% level. So don't worry too much about what that means exactly. But basically, the strength of the criteria that they used, you would expect just by chance to get about 100 significant results, and they got 190. So this is an indication that there's some real relationship there, because it's more than

just by chance you expect. And another good thing about this study, by the way, is that they didn't just test one direction. They tested whether guns led to an increasing crime, and also where the crimes led to an increase in guns. Now, before I, earlier in the podcast, I said this is very difficult to do, and it is. The way they tried to do is by using time series data. What they do is they take the measurement. As an example, they might take the measurement of the number of guns in an area in 2005, and then compare that to crime in 2006.

So it's pretty hard to argue that the amount of crime in 2006 affects the amount of guns in the year before, because, you know, can't go back in time. It's called a lagged variable. You compare a variable earlier in time to one later in time, and that should help to separate out the effects. So if you compare guns in 2005 to crime in 2006, if there's a relationship between crime and guns, that should appear, but any relationship between that goes in the other direction

shouldn't appear because of the lag you've introduced. And they did this testing for causation in both directions, guns to crime and crimes to guns. And so these 190 significant results include significant results in both directions. However, the number of significant results in terms of guns leading to crime and death was 67 positive and 43 negative. That is, of about 100 significant results they got for different regressions. 67 of them said that guns caused more.

more death and crimes. Forty-three of them said that guns cause less death in crimes, and of course the rest that were not significant than say anything. So this is not a massive difference, but it's significant enough that indicates that probably guns do cause more deaths and crime. But it's fairly weak. It's not huge. The second direction had 75 positive results saying that increased crime causes more guns, and only three negative results saying that increased crime causes less guns. So that's much clearer. It seems from this evidence, and this is consistent with whatever in other studies as well, that it's very clear that area.

areas that have higher rates of crime do tend to, as a result of that, have more guns. So the causation does definitely run in that direction. And the evidence for that, I think, is actually stronger than the evidence of the causation in the other direction, that is that guns cause more crime. However, this study, the 70 of the 67 compared to the 43, and also other studies, do seem to indicate that there is moderately good evidence that, I'd say, weak to moderate good evidence that more guns do cause more violent crime, particularly homicides. As we said, the evidence for assaults and so on is not as good as for homicides.

Another reason that study that did the 1,000 regressions was really good is because it used the lagged variable approach to try and work out the difference direction of causation. Most of the other studies didn't do that. They just look at the correlations. It's still a bit if you're there, but it's relatively stronger that there is definitely a positive relationship between guns and high levels of violent crime. When you look at the try and break their damage to causal elements is the guns causing the crime or the crime causing the guns, it's much, much harder. It seems that the evidence for crimes causing guns is strong, or, you know, comparatively strong at least. The evidence, to try and break their end to cause the end. It's the evidence. for guns causing crime is weaker but still there. It seems, in other words, that both of these things are happening. And this seems to make sense, because if both of these things were happening, then they would both be contributing to the relationship, the positive correlation between guns and crime and crime and causing guns, then you put those two things together and you're getting an even stronger overall relationship if they're both causing each other.

One thing that's fairly clear from the literature, however, there's virtually no evidence that guns reduce crime. So the hypothesis, the self-defense hypothesis, is not borne out. That is, it does not, although guns may be used to deter criminal, It does not, the evidence seems to be that the absolute level, the aggregate level of crime is not reduced by the presence of more guns. Another type of studies have been done called case control studies, where instead of looking at, you know, big groups like big aggregated areas like states or counties or countries, they actually look at specific households. And what they do is they pick a bunch of households where homicides occurred in a given period, say, five-year period one study. And then they pick another bunch of households that are very similar to the households where the homicides occurred in every respect, except no homicides occurred in these other houses.

So this second lot of houses are called a control group. So you try and match the control group to the study group, that is, you try and pick households that have similar levels that have similar incomes, similar racial, composition, similar size, and all those other factors, you try and, and this is again controlling for all those other variables that can affect rates of crime. And while we do this, well, what we're interested in seeing is homicide houses versus control households, is there a difference in the number of guns. And according to one study, well, a number of studies have done this, but the particular one I'm looking at the moment, 36% of case households, that is where homicides occurred, had guns, handguns in particular, this was looking at, compared to only 23% of the control households. So that's not a massive difference, but it is a significant difference. And so, in other words, that difference is not likely to be due to chance. It's likely to be a real effect. Other studies have come up with similar results, that it does seem like there is a substantially,

not massively, but substantially higher percentage of guns in households where homicides occur compared to households where homicides don't occur. However, there are some limitations to this line of study. One is that none of the studies are able to determine whether the gun in the household was actually used for the homicide or not, so we don't know if that was actually what caused the homicide. Also, it could well be that people who are more likely to commit homicide or just households where homicide is more likely to occur are simply more likely to get guns, the various reasons, perhaps the people are intrinsically more violent, or their situations are worse for other reasons

that are hard to measure. And that's not something that can be ruled out by these studies. So, in other words, these studies show that there's a correlation between where guns are kept and households that have guns and households where there's high levels of violent crime. However, it doesn't say, it doesn't show us what direction of course that the causation runs in. Other studies have also been done which show that people who buy guns are more likely to be murdered and two murder than people who don't buy guns. But again, this doesn't show us the relationship in terms of causation. However, I think probably the best study I've seen that does demonstrate the causation

would probably be the comparing the Canadian to the US cities and the lagged variables' thousand regression study that I talked about earlier. Most of the other studies can't show causation. Those two sort of can, not very well, but about the best there is. A quick word on the international studies where instead of comparing US states or counties or even households in the US, they compare countries across the world. One big problem with these studies is that it's hard to introduce enough control variables because there is not enough data. The more control variables you have, the more data

you need for reasons that are to do with degrees of freedom, which is, if you don't know what that is, don't worry about it. There aren't very many developed countries in the world, a few dozen dozen countries in the few dozen countries, and gun ownership, so we don't have with developed countries. But there, we don't have very many of them. And there's, again, generally, there is a relationship between firearms and, and homicides, that is countries that have more firearms tend to have higher rates of homicide, especially gun homicide, not so much non-gun a homicide. However, some of these regressions are a bit sensitive to which countries you include. So if you include some countries, the relationship gets stronger or it can even go away,

the US is a big outlier, but it does seem that even excluding the US, the relationship still holds, although it does get weaker when you take out the US. That is, even in countries apart from the US, more guns do seem to be related to more homicides. And if you're interested in more details about these studies, because I've really only scratched the surface here, it's impossible to go through all the details that I've gathered. I'll be posting up some of my references and other notes on the website and probably also the Facebook page. So have a look there if you're interested. So to sum up, It seems the evidence, the empirical evidence, shows a reasonably strong correlation between firearm prevalence and homicide, internationally between the US states and also within households and counties within the US.

So that one's reasonably robust, although it's not, the evidence is not super fantastic. In terms of causation, there seems to be weaker evidence, but still, I think, persuasive evidence that the causation runs in both directions, both from more firearms to more firearms to more firearms, with the more crime to more firearms having somewhat greater evidence in favour of it than the reverse. I'm really only going from two or three studies there in terms of the direction of causation because almost no studies have been able to demonstrate this. Most of the studies on the issue just look at a simple correlation and that can't tell you anything about causation. And even there, a lot of the studies are very crappy.

They don't include proper control variables or they don't validate their proxy variables and many other problems that they have. Overall, I would say that the causation runs in both directions, but the question that most people are interested in is, do more firearms cause, more homicides? And the answer seems to be yes. However, it's a very qualified yes in that this effect is not really strong and the evidence for it is not great. It's, I think, persuasive enough, it's by no means very strong. Also, we really have no idea about the magnitude of this effect. Let me put very vague figures on it

from my reading of the research, which you should by no means, you know, think of as gospel or anything like that, but it's just from what I read. So the US homicide rate per 100,000 people is around six, and most other developed countries, some between one or two. So suppose the number of firearms in the US was brought down from 88 per 100 people to, like 30 per 100 people, comparable to other European countries. So that would be roughly reducing the number of guns in the US by 66% or something like that to one-third of its current levels. If that happened, I highly doubt that crime, and suppose also nothing else changed, which is a bit unrealistic, but just hypothetically.

If that happened, I highly doubt that crime, that homicide levels in the US would drop to European levels. And one of the big reasons to remember is because of the difference between gun ownership between African-Americans and non-Hispanic whites and the difference in homicide rates there, that and many other pieces of evidence, and I don't have time to go into all of those in this podcast, but we'll touch on some of them, seems to indicate that there's a huge socioeconomic, socioeconomic factors that are related to the US's much higher homicide rate, also related to drug and gang warfare and education system, another factor too.

So, I don't think the US homicide rate would go down to two or even probably three per 100,000 if the gun ownership levels were brought down to European levels. I think it probably would go down more than, it wouldn't just go down from 6 to 5.9 or something like that. It would be, I think, more significant than that. The evidence seems to be that the effect is big enough to detect. So, I don't know. Maybe the homicide rate would go down from 6 to 5.5 or 5 or 5 or 5 or something like that. I don't think it would go down like 2 or 3, and I think it's going to go down by more than, like, 0.1 or 0.

So that gives some magnitude of the effects, but that's purely subjective. It is based on what I read, but we really don't have any good data on this. And even the coefficients that we get in the regressions that have been done are just so hard to interpret because of the way the variables are measured and so on that it's very difficult to, and they're so inconsistent anyway that it's very difficult to be more precise about that. So I would call that a moderate effect, but not a massive effect. So far, however, we haven't said anything about some more specific hypotheses that are made about guns and gun control, and we haven't said anything about gun control yet.

We've just been talking about a number of guns, even if it's the case that guns, lead to homicides and more crime, that doesn't necessarily mean that gun control would actually help the issue in any way. So that is a separate question, which we'll look at momentarily. But first, I want to look at a few more specific hypotheses or statements that are made in the gun control debate. One is what I'm going to call the average Joe thesis. I stole this name from another article. I don't know if it's in wide use, but I'm going to use it anyway. Now, the thesis goes that most or many, many or even most homicides are the result of impulsive actions taken by ordinary law-abiding citizens who have little or no criminal background. And so the

idea is that if many homicides are impulsive, then having a loss of guns around will dramatically increase the number of homicide because it's easier to get access to deadly weapons in the moment of fury or rage or whatever. And similarly, that this thesis also would seem to indicate that focusing on controlling the access to criminals to guns is not the most important or not the most useful use of resources because it's not criminals who are doing most of the homicides, most of the killing, it's actually ordinary people. And so this would argue for a broader gun control rather than specifically focusing on

criminals. So this is a statement that most homicides are committed by ordinary people taking impulsive actions. How accurate is this? Well, it's sort of true and sort of false. So part of the truth, part of the element of truth here is that it is true that most homicides are committed by people who know the victim, either family members or friends or acquaintances. So being killed by a stranger is relatively rare. I think the stats are something like 20% or something like that.

It does vary, of course, and it's hard to get great measurements. But most homicides are committed by people you know. So that aspect of it is, and that was part of the thesis as well, that it's sort of like family swabbles or domestic disputes. And that is a large category of homicides. So that part is sort of true. It's also true that most people who commit homicide do not have a past felony record. According to my statistics, something like 75% of those convicted of homicide have no past felony conviction against them. And so that is consistent with the fact with the idea that most homicides are not committed by hardened criminals. However, there's also many things that are misleading about

this average Joe thesis statement. Although it's true that most convicted murderers don't have a felony conviction, most convicted murderers have, do have a criminal record and an arrest record. So in other words, many have committed misdemeanors or been arrested for other reasons, even if they haven't been convicted of a felony. The data I have varies, but it seems something around 75% of US murderers have adult criminal records, compared to about 15% of the population as a whole. So the rate of, in other words, murderers are much, much more likely to have a criminal record before they are convicted of murder, obviously, because obviously once you're convicted of murder, you have a criminal record, but I'm talking about

before that, they're much, much more likely 75% compared to 15% to have a criminal record. And so this is dramatically inconsistent with the hypothesis that normal individuals, like ordinary law-biting citizens, are mostly the ones committing murders. In fact, most of the people who commit murders have prior arrest or criminal records. A related fact is that, although the average Joe thesis relates to domestic violence, the idea is that domestic disputes flare up and one person grabs a gun and shoots the other in an impulsive action when, you know, thinking about it rationally they wouldn't have done so. That is true that a large fraction of homicides in the US are resolved. that, not just in the US, by the way. However, something like two-thirds to, well, the statistics that I have very bit, a rough figure would be like 70, 75%. So an overwhelming majority of homicide between intimate partners, that is, people who are married or partnered together. So when one

kills the other in an intimate partner relationship, there was physical abuse of the female in the male before the murder, no matter which partner was killed, whether the husband kills the wife or the wife kills the husband, basically. In roughly 75% of cases where that happens, there was past previous physical abuse of the female by the male. And studies that I've seen show something like 80 or 90% of cases where you have a domestic homicide, in something like 80 or 90% of those cases, police had already previously been called in to deal with domestic violence disputes or had cases reported to police and so on. So although in most cases, you know, no felony convictions have been recorded, it seems that the type of people who are likely to

commit a murder are also much more likely prior to the murder to come up on police radar through being called to a domestic dispute or committing a misdemeanor or being arrested or something like that. So it's not the case that people who commit homicides are just sort of ordinary people who happened to explode in a moment of impulse. It is actually possible to sort of not identifying exactly who was going to commit homicide, but to identify sub-populations who are at much greater risk of committing homicide than the ordinary population. Another study that I read from California found that the sale of guns to people with at least one prior misdemeanor conviction. You can't sell guns to people with a felony conviction, by the way.

So sale of guns to people with at least a prior misdemeanor conviction was six times more likely to be followed by a violent offense from that person than the sale of guns to people with no criminal record history. Those who had two or more prior misdemeanors instead of just the one were 15 times more likely to be charged with homicide, rape or robbery or aggravated assault. That is, people who have been convicted of a misdemeanor are much more likely six times more likely, in fact, to go on to commit a violent crime after purchase of a weapon. And the more misdemeanors you've committed, the more likely it is that you commit a violent crime after having purchased a weapon. So this is further evidence that we can statistically at least tell what kind of people are more likely to commit homicide than others. In fact, it also seems to be the case that the overwhelming majority, or at least a very large proportion, of violent crimes, this includes assaults, robberies, and so on, are committed by a very small number of people. One study that I found, in a one year period, an average offender in this small group of recidivist offenders, that is repeat offenders, the average one of these committed eight assaults, 63 robberies, 172 burglaries, 1,252 drug deals, and 214 other thefts.

That's in a single year period by one person. Those include self-reports and also police reports about their activities. And there's other evidence to support this idea as well, that a lot of the crime is committed by a small number of people, and this includes a violent crime as well. So, to sum up, the average Joe thesis is partly true in that most of the people who commit murder don't have prior felony convictions. However, it's not true in the sense that most of the people who commit homicides

do have appeared on the police radar in the past, either through domestic violence issues or through misdemeanors, arrest records, or other things like that. And also, it does seem that most violent crimes are committed by a small number of offenders, not just why your average Joe are losing their cool Monday and being up a gun. Of course, that does happen, but it does not seem to be the majority of cases or even a substantial minority. So, that's the average Joe thesis.

Mostly not true. Another argument that people put forward in this debate is the guns more deadly thesis, as I've called it. This is the idea that guns are more deadly or more likely to cause death or serious injury compared to weapons that criminals would use if they didn't have access to weapons, because this is an important argument that a gun proponents sometimes make it well. If criminals couldn't get access to guns, they'll just use something else, A knife, for example.

The counter to that is, well, guns are more deadly than other weapons. But is this true? This is an empirical question. Indeed, it does seem to be true. Guns do seem to be more deadly than alternate weapons. The main alternate weapon that people would use for, say, a robbery or homicide or something like that, it seems to be a knife. And it's, again, really hard to get good data on this. I've got a few different studies reference,

but the basic figure seems to be somewhere between, like, three or four times as many people die from stab wounds compared to gunshot wounds, up to maybe 10 times as many. I'd say maybe it's like, you know, four or five times, something like that. Different studies come up with different results, it depends when and where how it was measured and all sorts of things. But it seems to be several times greater likelihood that you'll die from wounds of stab wounds, sorry, more likely that you'll die of gunshot wounds compared to stab wounds.

However, this doesn't necessarily prove that the gun's more deadly hypothesis is correct, because it could well be that criminals who are inherently more violent or just more predisposed to using violence or think that it's more likely that they will actually have to use their weapon as opposed to just threatening people with it. It could be that such people preferentially choose guns. That is, it's not that guns make the criminal more violent, that more violent criminals choose more deadly weapons. That might sound silly, but if you think about it, criminals are not going to choose their weapons

just by random chance. They're going to have a reason for why they choose a particular weapon, a knife over gun, or vice versa. Availability will be one factor, but it's certainly not the only factor. And so it's not at all clear that the fact that people are more likely to die of gunshot wounds is wholly caused by the fact that guns are more deadly than knives. It could be that people who use guns are more aggressive and more violent than people who are likely to use knives. I think that is a valid argument. However, I don't think it's enough to completely overturn the substantial size

of the effect, which is several times. It's not like 50%, it's several times more likely to die from firearms than stab wounds. So I think there is evidence to support the guns more deadly thesis, although qualified evidence. Next, next feature we're going to look at is the thesis that guns are actually useless for defense. So pro-gun advocates will say that guns are useful for self-defense, defending the home, scaring off criminals, that sort of thing. Opponents of guns will say that people don't use them for that or very rarely use them for that, and they're actually not helpful. So it's really hard to get data on this, and I said that about 50s. 50 times, but I'm going to keep saying it because it's very important to understand. It's hard to know

how often people use guns and self-defense. You can do surveys, and this has been done, but there are many problems with asking people about things like this, in particular because of memory reliability. Look at the previous episode I did on the reliability of memory. There's also the issue that people might like to think that their gun is more useful as scaring-all criminals than it actually is, self-serving bias and confirmation bias and other things like that could be playing a role here. However, there is still some evidence. So one study interviewed, rather than gun owners, and interviewed prison inmates. 36% of imprisoned felons reported, having decided at least, quote, a few times, end quote,

not to commit a crime because they believed the potential victim was armed. So about one-third of imprisoned felons said that gun's gun ownership on the part of the victim was a deterrent to them for committing a crime. Now, there's some problems with this. First of all, at least a few times is a bit vague, especially if we consider that many imprisoned felons are going to commit hundreds of crimes, and if it's only a few times that gun ownership deters them, that's not necessarily very often.

But it might be much more than that. We don't really know because they didn't, you know, for a number or anything. It's also possible that in-prisoned felons are not giving very accurate information, although I don't see any particular reason why the answers would be biased one way or the other. It does seem, though, that looking at police reports and other factors, and other pieces of evidence, that criminals are concerned about the fact that their victims might have guns. When victims pull guns, criminals generally flee, they get the heck out of there because they don't want to be shot, and, as I've said, this and other surveys of prisoners have tended to show

that criminals are worried about guns and are deterred by them. Another piece of evidence is that 43% of British residential burglaries, so burglaries is when someone comes into a, breaks into a building and tries to steal things, 43% of those in Britain are committed while victims are at home, whereas only 9% are residential burglars in the US are committed when the victims are at home, when people are at home. And we don't exactly know why this is, but one potential reason is that there are many, many more, like, nine times more, I think, maybe more guns per person in the US compared to Britain. In other words, many, many more households in the US have guns compared to Britain. And so the argument is, well, criminals know that, and so they're much less likely to try to burglarize a house when the residents a home in the US and thereby to routinely shoot them compared to Britain where they're much less concerned about that

because much fewer people have guns. This statistic that I quoted the 43 to 9% doesn't prove that but it's consistent with our hypothesis and it seems like a reasonable interpretation. Now onto the specific issue of asking people do you use your gun for defence, there's a lot of controversy about these figures here. Estimates of how often this defensive gun use occurs vary from

several million annually to only about 100,000 annually. Probably the real figure is a few hundred thousand, maybe 500,000 or something like that. That is incident. of people using guns to defend themselves or their homes. So it does happen. I mean, say there's a million just to pick an easy figure. That's one in, I don't know how many households there are in the US,

probably something like 100 million. That's so maybe one in 100 households, one in 200 households, uses a gun every year to defend itself, according to the best estimates we have, which aren't that great. But, of course, it's not just actually using the gun to deter criminals, it's also the simple deterrence factor if the criminal doesn't go there in the first place. Both of those, based on the prison surveys

and the individual surveys seem to be happening. In other words, people do use their gun to actually physically deter criminals, and the presence of guns also does deter criminals in a more general sense. Injury of the person who is doing the defending is also very rare as well. This idea that the person who pulls a gun to defend themselves is more likely to get injured is not really supported by the evidence. It's very rare for someone to be injured while using a gun in a defensive way. According to one study, I have something like 8% only about 8% of defensive uses

were injured after using the weapon. And even that overstates the case a little bit because they also found that victims who used a gun in defense were in more dangerous and difficult circumstances. For example, they were alone or the offender was more heavily armed than those who did not use a gun. So it seems that guns are useful for defence, or at least they are used for defence to some degree, they do really deter criminals. And the risk to the person you're using the gun is, at least in terms of being injured in the act of using the gun as defence, is not especially high.

Remember, I'm not talking about suicide or accidents or other risks. I'm just talking about the risk in an actual incident of using the gun in defence. It does not seem to be particularly high. Another interesting little factor or situation here that I like to mention is what's called the weapon effect. The idea here is that guns have a association of violence and violent action with people, and so the actual presence of gun itself is likely to elicit a violent response. The sort of catchphrase for this is the trigger pulls the finger, instead of the finger pulls the trigger.

That is the fact that someone has a gun makes them more likely to use it because it generates aggression and other connotations and feelings within the person, so it makes them more likely to use it. There's not too much evidence for this. Some laboratory studies have been done where they do get associations between, like showing pictures of guns and firearms and then eliciting aggressive responses and things like this. But having an effect like that shown in the laboratory, and having it shown to be relevant in real-life circumstances are two very different things,

and evidence to try and show the effect of this weapons effect in realistic circumstances have not generally demonstrated very much of an effect. So it's possible that this effect does exist, that is, that the trigger pulls the finger to some extent, but I personally am not very compelled by the evidence available, so I won't spend too much, I won't spend any more time on that. Next thing I want to talk about is the hypothesis that often often brought up by pro-gun advocates that if guns are made illegal or mostly illegal, then only criminals will have guns, because, of course, criminals don't care about what's illegal or legal.

So all you'll do is be disarming the victims of crime whilst keeping the perpetrators armed. Now, this was one of the hypotheses that I could get least evidence on. I don't think I found really any empirical evidence for this one way or the other, although there's some inferential things we can say about it. I did find one very interesting game theoretic. So it's a theoretical study, but I think it's relevant, because what it basically did is tried to model the motivation of criminals compared to the motivation of non-criminals to own firearms.

And the results that it came out with was basically that the more likely criminals were to have firearms. then the greater the benefit there is of a non-criminal to have a firearm because it's essentially more likely you'll need it to defend yourself. Conversely, though, the larger proportion of the non-criminal class, that's ordinary citizens, who have firearms, the greater the benefit is to criminals to own firearms as well, because it's more likely that they'll need to use the firearm to threaten individuals and so on, because the law-abiding citizens are better on themselves. So it's sort of like as an arms race between the criminals and the ordinary citizens.

So based on this model, if we reduce the number of guns that law-abiding citizens have, that could potentially reduce the number of guns that criminals have. purely voluntarily. In other words, criminals don't find it as necessary to carry guns around all the time or to carry as high-caliable weapons and so on. This might sound odd because wouldn't criminals just do it anyway. Well, no, not necessarily, because criminals know that if they're found with a gun, they can be, the sentence they can get is much higher and they're more likely to be prosecuted and things like that, more likely to be convicted. People who are engaging in robbery, in burglary and drug deals, don't want to die and generally don't want to kill people either,

because that's just going to increase the chances they'll get caught. They want to just take the stuff and get out of there. So if they can do it without having to have a gun, which increases, you know, chances of being involved in something like a homicide or, and also, as I mentioned, it will increase their prison sentence or potentially could increase their prison sentence, and is itself expensive and can potentially represent other problems. I mean, guns aren't always expensive, but they can be. You know, they'll do without one if they can. At least some will.

On the margin, there will be some effect. So it's not clear that if guns were illegal, only criminals would have guns. It seems to be that the number of guns that criminals and non-criminals have will actually go together. And this seems to be consistent with the empirical work that we see. whereby the essentially there is a pretty strong correlation between crime levels and gun holdings. And remember, we had the bi-direction clause alley where more guns lay to more crime and more crime led to more guns, according to our best interpretation of the regression results. So if guns were illegal-only criminals will have guns, probably true to some extent,

in other words, that criminals will be more likely to have guns than ordinary people, if guns are harder to get, but not completely true because it's not clear that criminals will just always get guns. They actually respond to incentives as well. And this leads me into the next topic that I want to talk about, which is how criminals actually get guns. This is very relevant to the issue of gun control, because how criminals get access to firearms is very relevant to what control measures we want to put in place, because we want to try and prevent criminals from getting access to firearms, but if they get access to them through some method, then another method than some other method of gun control that doesn't actually relate to how the criminals get guns

will probably not be very effective. So we need to know how criminals actually get access to weapons. Remember, felons are not allowed to buy guns themselves, at least from a licensed vendor in the US. You have to do a background check, although the degree to which this is enforced and probably done is not nearly as good as it should be in a lot of places in the US. A lot of records aren't online, for example, and there's not necessarily a long enough period for the background check to be properly carried out, and there are some corrupt dealers as well.

Definitely in the minority, but they do exist. So it's not easy for felons to get by weapons legally, at least. However, various studies have found that criminals, and not just felons, but also people who have committed misdemeanors and others, but especially felons, have a very wide range of potential sources of guns. So according to one survey, for example, of state prison inmates, among the inmates who possessed a handgun, 9% got it through theft,

28% acquired it through an illegal market, such as a drug dealer, or a fence, and I'll talk about what a fence is in a moment, and of all inmates, 10% had stolen at least one gun, and 11% had sold or traded stolen guns. So a lot of these criminals are getting guns through theft or through illegal purchases. In another sample, 50% had stolen at least one gun in their lives,

and 24% had stolen the gun that they are most recently acquired, so very high rates of theft of guns. Other studies that I've seen in terms of sources of guns, One study found that 52% of guns were borrowed or bought from friends, 32% obtained from theft, 16% bought from store. Another study, 30% came from friends, 22% came from the street, 21% came from drug dealers or drug addicts, 12% came from theft, 6% came from family members, 7% came from a store, and many others, also, especially if they're juveniles, they just ask adults to buy the guns for them. So repeatedly in the studies we see a very wide range of sources of guns. Stealing them is very common, buying or borrowing them from friends or families is very common, obtaining them from drug dealers or other people on the street, very common, and buying through fences is also relatively common as well. Now, a fence is someone

who buys a gun legally, because, you know, the fence will not be a felon themselves, and then illegally sells the gun on, or sometimes legally, actually. I'm not sure if it's always illegal to sell a gun to a known felon. It might be, I don't know, I'm not 100% sure about that, but anyway, regardless of whether it's legal or not, the fence buys the gun and then sells it on to another person who would themselves not be able to buy the gun legally. So the conclusion from this body of evidence is that there's no one big source of guns that criminals have. It seems they have a wide range of sources. Indeed, many criminals, especially juveniles, don't seem to be very, they don't need to be very proactive in terms of looking for guns because it's relatively easy to come across

them if they're involved in the drug trade. A lot of drug dealers or, um, drugies have access to guns. A lot of people associated with gang culture or other, um, another sort of crime-prone environments will have someone, a family member or an acquaintance or friend whom they know they can get guns from. Many people who engage in theft or, uh, burglary steal guns. For example, that study that I said that 50% of juvenile inmates had stolen at least one gun in their lives. Many of them aren't even actually looking for guns to steal, but because so many US homes have guns, they break in, they look for valuables, and they just find a gun, and so they take it because it's something relatively comparatively valuable compared to its weight and easy to resell. So, it doesn't seem like there's any big one source of guns to criminals that we can focus the regulation on, because there are many sources. And this, this to me, is a depressing result because it seems to indicate that even if we significantly focused regulation on, say, stopping fence selling, that is people buying guns on behalf of other people, or on cracking down on guns in the drug trade, or in cracking down,

on theft of guns or something like that. Any one particular source, it seems very likely that criminals will simply be able to increase their reliance on other sources. Now, the particular evidence for this is not very strong. In other words, I didn't really find any evidence that specifically looked at the effectiveness of policies focusing on different sources of gun control. So it's possible that some sources are more important than others, and it may be easier to stop some sources of guns fly than others.

But the data I saw didn't seem to indicate that, and it seems that there are many sources of guns, and if you stop any one of them, criminals will just be out relatively easily in the current US climate, relatively easily be able to get guns somewhere. else. By the way, an interesting fact is that most gun theft is the byproduct of residential burglary. Less than 2% of stolen guns are stolen from gun dealers. So that's one thing that, likely, we don't need to worry too much about, is theft from gun dealers. Mostly it's theft from residential burglaries. Okay, so we've looked at all the specific hypotheses about, well, many of the specific hypotheses that are mentioned in the gun debate. The average Joe thesis,

the guns are more deadly thesis, guns, guns, use-olds for defense thesis, the weapon effect, and we've also looked at how criminals obtain guns. Oh, and we also looked at the issue of if guns were illegal, only criminals will have guns. We see that the way of phrase most of these, most of these were false, or at least significantly misleading, with the exception that guns more deadly does seem to be true. Now, where does this leave us? Because the title of this podcast is actually gun control. It's not guns and violence, although that's mostly what I talked about so far. What can we actually do to try and reduce, or what can the US do, specifically to US government

to reduce the number of gun homicides and other violent crimes, particularly homicides, though, because that's the most pressing issue, arguably, because it's the most common violent crime in which guns are used, that is guns are used in homicides much more frequently than in other type of violent crimes in the US. Well, you remember I said that the evidence for the relations, you know, between violence and gun ownership is pretty bad. Well, the evidence for gun control, the effectiveness or otherwise of gun control regulations, that is the actual laws, is even worse. It's really, really, really bad. There's just very few studies of this sort of thing, and most of the

studies that do exist, are poorly controlled, are smaller. There's so many problems with them, so you can't necessarily tell very much. Basically, there's a great review article which I found. Again, I'll post a link to this, so with the show notes, that went through all the different, many, many different types of regulations that have been tried and studies that have been done on them, again, not very many studies, so we can't be very confident about the results. But they concluded that there's basically not, almost all types of gun control legislation that are proposed in the US, at least, have insufficient evidence to recommend them as being particularly effective in reducing the levels of violent crime, especially homicide.

So some of these specific regulations that they looked at are bans on specific types of firearms or ammunition, didn't seem to do much, restrictions on who can access or acquire firearms, again, didn't seem to help. Increasing waiting periods of firearms acquisition, requiring registration and or licensing of firearms, concealed handgun, concealed weapons, Laws, child access prevention programs make it more difficult for children to accidentally get guns, zero tolerance of firearms in schools, and various index combinations that try and put different firearms laws together and measure the overall effect, overall strength of firearms regulations in

different counties. In all of these cases, there's insufficient evidence that these policies do very much. Now, remember, this is not a negative result. In other words, it doesn't mean that these laws make it the situation worse, and it doesn't mean that we definitely know they don't work. There's just not enough evidence either way, so they might work, but there's very, the evidence we have is very poor, and it's insufficient to say that these are really good programs. And there's good reasons to suppose that at least many of these things would not be especially effective.

For example, child access prevention laws, only 800, again, I say only, 800 is too many, but it's a lot less than 10,000. So only 800 gun accidents occur in the US, gun accident deaths occur in the US every year. And very few of those are for children, only a few dozen, I think. Go back to 10,000 homicides. So focusing on child access prevention is probably not going to produce the biggest result in terms of life saves for money spent, which is what we're looking at here. We're looking for effective and cost-effective results. because simply the number of people who die from accidental discharge of guns in the presence of young children is very small.

Similarly with the zero tolerance of firearms in schools, the recent incidents in the US notwithstanding, the number of people who are actually killed by firearms in US schools every year is very low. It's only a few dozen. Now, of course, that's a few dozen too many. But when we're looking at interventions that are going to be cost effective in terms of producing, combating the areas where gun violence is the worst, schools is actually not one of the places you're going to go to. And also in regards to bans on specific types of firearms or acquisition restrictions and so on, it seems that there are many, many ways that people can get firearms, and so restricting particular ways of doing it are just probably not going to be very effective.

Another policy that's been trying to gun buy-back programs, where basically the government buys back guns from the public, this is probably one of the worst policies that I've seen, that the only a few small studies have been done on it, and it doesn't seem to have had much of an effect. But one of the reasons that these programs that I strongly recommend against these programs, apart from the fact that the NRA is not going to like it at all, is because they're very poorly targeted. In other words, they're just buying up guns from all over the place. In fact, generally, it's more likely you're going to buy guns from law or buying citizens

compared to criminals. And so you're probably just spending a lot of money, because these products can be very expensive. You know, $100 a gun or something like that. If you tried to buy most of the US firearms, it would cost billions of dollars. And the ones that firearms are going to be getting are probably not from the people who really want them, like criminals or other people who are more likely to be using them, in other words, probably. So this is very poorly targeted, and so very unlikely to have a good, even if it does anything, it's very unlikely to have a strong cost effectiveness.

So, where does that lead? If most of the, if it seems to be that the significant number of guns in the US is causing more violence, but the gun control regulations that have been proposed don't have very good evidence in favour of them, and there are theoretical reasons based on other evidence that we've seen that they probably wouldn't work, or even if they did, they wouldn't, they would be very expensive and hard to enforce and so on, not to mention the fact that they'd be politically very infeasible to implement because of the very strong, powerful gun lobby in the US. Where does that leave us?

What we need to look at is for interventions that are going to be politically feasible, so something that you could potentially get the NRA to sign onto, something that they don't feel is going to significantly curb their, you know, their perceived right to it to bear arms. And I'll talk a little bit more about that later on. Also, we're going to need something that is enforceable. So something like outright bans on all handguns. It's just not enforceable in the US. There's just too many handguns and too many people who want to own them.

The cost would be horrendous. And that leads to the third issue, which is the proposed investment needs to be cost-effective. It needs to achieve a relatively good return on the amount of money invested in it. If you spend $100 million and only save, like, one life. Of course, that's good for the person whose life was saved. The thing is you could have taken that money and spent it on, say, maternal health or vaccinations or something like that, and saved thousands of lives instead, potentially. There's an opportunity cost for everything, and if you're spending huge amounts of money on security or gun control or whatever and getting very little return for it in terms of reduced homicide, then that's not a good use of resources. So we want to find some gun control that's going to be cost-effective and enforceable and politically possible to implement.

Does any such thing exist? Well, of course, it's not obvious that something does. Of course, if there was really obvious thing, it would have been done already. But there are a few policies that I think are potentially promising. The evidence for them is not nearly as good as it needs to be. but there's some evidence in favour of them, and they're things that we might expect that could work. So let me talk about a couple of them. One is gun patrols.

So a couple of, a couple of experiments of these have been done where basically uniformed officers patrol known areas of significant levels of violent crime and where they, you know, know a lot of people are illegally carrying weapons, and they just basically go patrolling around the streets, seizing illegally carried weapons. One of the most famous examples, this was the Kansas City Gun Experiment in 1992. They found that gun crimes dropped by 50% as a result of at the same time as this experiment was carried out. You'll notice I was careful not to say as a result of this. experiment because we can't know that for certain than it was a result of this

experiment, but 50% is a very large drop in a very short period of time. So it seems to be that this intervention of the gun patrol did help a lot. Something else that's more promising, but that's also related, is the idea of problem-oriented policing. Now, the idea of problem oriented policing is that instead of being reactive, that is the police just responding to cause they get or responding to crimes after they've happened or even while they're happening, the police actually are proactive in that they try and deal with the underlying causes of the problems and trying to prevent crimes before they happen as much as possible. in a couple of these cases where this was done,

so Boston Operation Cies FIRE, New York's Comstatt and Richard Virginia's Project Exile that are different cases where this problem-oriented policing strategy has been implemented. They try to focus specifically on, you know, known crime hotspots, known elements of gun control problems, known illicit gun trafficking and gang violence areas, tried to actually get on the ground, talk to gang members, deal with these problems and try and prevent crime before it happened.

The specific ways in which they did is quite interesting, and we don't really have time to get into them in this podcast. It seems kind of obvious, but the police have so many things to do and they're under-resourced that they don't necessarily have time to do this sort of proactive stuff a lot of the time. But in these particular programs, they made a real push to do that sort of thing. And they did find significant drops in levels of violent crime, a particular total homicide. So one study that I found showed that the drops in firearm homicide in Boston and Richmond, and Richmond were significantly different from Sierra, whereas in New York, the results weren't so clear.

It might have been a positive result, might not have been. But certainly in Operation Seasfire in Boston and in Project Exile in Richmond, Virginia, there seems to be very quite good evidence, I should say, because it's only one study that I've got the specific data here, but reasonably good evidence is that the program did have a significant reduction in firearm homicide, this problem-oriented policing approach. So how can we apply these problem-oriented policing to gun control? Well, apart from using it in a very direct way that the Operation Sleas Farron and so on did, in terms of trying to specifically identify problem areas, you know, gangs and drug deals and so on,

I think we could potentially think of extending the problem-oriented policing approach to other aspects of gun control, that is focus on the people we know. are more likely to misuse guns. So before we said that people with a, there's evidence that people with a prior misdemeanor, at least one, are six times more likely to commit a homicide or a lot of environment, crime in general, forget which, after purchasing a weapon than people without one.

And people who had two or more misdemeanor were 15 times more likely. So it would seem that police should try and focus on regulating or at least keeping a closer eye on gun transactions that involve people who with prior misdemeanors, and especially multiple prior misdemeanors, especially people have been convicted of a previous gun offense. So rather than trying to ban outright classes of guns,

maybe a better way to go would be to try and focus on these types of people to try and get really good background checks and to try and really stamp down on not just ex-felons, but also ex-safeelons, but also, say, people who can be convicted of misdemeanors or other similar offenses are from buying guns as well. And also more resources into cracking down on the illegal gun trade, although that's, of course, difficult, but potentially can be done. Rather than, say, putting resources into a general gun buyback or an overall handgun ban, I think those resources would much more effectively go into trying to try and reducing

the illegal gun trade. Again, it's not easy to do, but what the principle of effective regulation that we're interested in here is focusing resources that are limited on areas where it's going to have maximum benefit. And overall bands don't seem to have to promise that benefit, apart from the fact that they also just very difficult politically to get implemented. One thing that I wanted to bring up is this issue of the gun show loophole. That is that guns purchased at gun shows, you don't have to have background checks for those. This is potentially an issue, and that could be something that regulation could look into

in terms of getting background checks at these as well. However, they are not as big a problem as some anti-gun activists present the issue, basically because not many gun owners seem to actually purchase their guns from gun shows. And remember, criminals have many sources of alternative sources of guns as well. Very few criminals seem to get their guns from gun shows either. Only 4% of gun owners surveyed in 1994 had purchased their guns at gun shows, and very, very small percentages of convicted criminals and so on and juvenile offenders had got their guns at gun shows either.

So it just doesn't seem that gun shows are a big source of guns that are used on for homicides or on crime. Now, of course, you know, any guns that are used for violent crime, you want them to be cracked down on, and you should, you'd probably want all gun purchases to have some element of background checks. But it just seems that making a particularly big issue about the gun show loophole is probably not, again, the best use of resource, because it just doesn't seem to be empirically that important.

So what type of gun regulation would I like to see issues? What I would like to see is targeted, enforceable restrictions and regulations that focus on the people who are most likely to misuse guns. An example would be greater degrees of follow-ups on domestic violence cases and potentially trying to create more greatly restrict or regulate in some way the guns that are owned by people who have a history of domestic violence. Might be one example. Problem-oriented policing approaches in terms of trying to reduce problems of

gang violence and guns relate to the drug trade and other issues like that. Potentially a use of uniform gun patrols in cities to try and confiscate illegally carry firearms. A better record keeping in terms of exactly how many guns different states have. I mean, a national registry would be really nice. The thing is I think that that would be probably deemed unconstitutional. I think it actually has been deemed unconstitutional, yes. But also it would be prohibitively expensive. The main reason, though, that I think it will be important to at least get better data on how

many guns are in different parts of the US is so we can better track the statistics. because remember, I said one of the big reasons that it's hard to measure the effect is because we don't know how prevalent guns are in different regions. So better data on that, even if it's not a natural registration, if it's some other means, better data would be very helpful. And I don't think it would actually be that expensive to develop some mechanism for having better measurements. But overall, I think the most important thing to say is that what the US needs is solid evidence-based

gun control. That is, well, they should introduce different policies in different places and try and do it in such way that we can have controlled trials, but we can introduce different gun control measures in, say, similar counties or similar cities or something like that, compare the results and do this on multiple occasions so we can find out what works and what doesn't. It's quite possible that things that work in some cases don't work in other cases. It might depend on the city, or whether it's north or south of US, or something like that. But we need much better information than we have at the moment.

I think that this approach of sort of trying to gradually introduce new potential reforms and gun control measures and trying to work out what works and then focusing on the areas that offer the highest return in terms of invested resources offers a much better approach and in the long run will save more lives than trying to introduce widespread controls, which will be very hard to do anyway and very expensive to enforce and so on. And we'd probably never know if they work, because we wouldn't have the controls we can compare them to. This leads me to the final thing that I want to talk about, which is a US gun culture. There is an argument, and I think it's true to a good substantial extent, as I'll explain in a moment, that the attitude that, not all, of course, Americans,

but a lot of Americans have towards guns is very different to that that exists in other countries, especially countries like Switzerland and Finland, which are two countries that have a very high gun ownership rates as well. In fact, in Switzerland, most males, something like two-thirds of males, between the ages of 20 and 30 or so, are conscripted into the militia. And as part of it, they have to own a gun and keep it at home. They don't actually have ammo for that gun, although it's, if they really wanted to, I imagine that it's possible to get such ammo. And the gun that they have is not to any gun, but it's actually an assault rifle. So in other words, something like two-thirds-plus of all the young men in Switzerland

must own an assault rifle, and yet gun homicide rates in Switzerland are very low. We see a similar thing in Finland, where there are many, which has one of the high rates of gun ownership in the world. Many of them are hunting rifles because they do a lot of hunting in Finland. But again, the gun homicide rates in Finland are quite low, certainly much, much lower than the US. And even within the US, we see that the homicide rates differ dramatically between states and differ dramatically between non-Hispanic whites and non-Hispanic blacks, even though the gun ownership rates are actually greater for whites than for blacks. So these pieces of evidence seem to point to the fact that it's not guns themselves that

are doing most of the work. Yes, there is some evidence, as we pointed out earlier in the podcast episode that guns do directly lead to an increase in violence, a violent crime. However, there seem to be many other more important factors, and what I think is one of them is gun culture. So, in the US, there is a conception of a constitutional right to bear arms, and that

everyone has a, at least many states have a, what's called a castle dropper, whereby every, a person has a right to shoot potentially, legally, someone who is invading their property or trying to, you know, trying to burglarize their home or something like that, the idea that the home is a castle and you have a right to defend it through deadly force if necessary. This idea of the right to own weapons as part of a self-defense and as part of an independent culture and things like that, you know, independent-minded culture.

That sort of thing, potentially being able to resist the government if they turn tyrannical. That's another sort of idea that's embedded in American culture, at least, especially in the South. This kind of idea does not seem to exist in other countries that have large gun industry. So in Switzerland, for example, even though they have very high gun ownership, especially for young men, the idea that someone has a right to own a gun as part of a self-defense or as a part of a constitution right or something like that just doesn't exist. same thing in Finland. In Finland, people view firearms as something that's used for sport and for hunting and so on, but the idea of having a constitutional right to it or defending yourself with it in the home is, again, just not there. The attitude towards guns is very different.

It's a little bit hard for me to express exactly the way in which this is the case, and of course the evidence for this in a verbal podcast like this, you have to sort of read some of the accounts from these countries about the way their regulations are shaped what people say about their attitudes to guns and so on to get a feel for this. And I think this is part of the problem, because it's not so much that this attitude directly leads to the increased violence. It's all part of the package. That is, in the US, it seems to be quite rational for someone to want to have a gun to defend their home, because there's actually evidence that it does help.

However, part of the reason that it's necessary is because so many people have guns that it's so easy for criminals to get access to guns, that they tend to be used a lot more in crime, especially violent crimes, than they are in other countries. So it's sort of a vicious circle there, and that more guns around, so more criminals have guns, so it's more necessary to have guns for self-defense, so that more people need guns, and so more criminals have guns and so on. And this fact that guns are necessary and actually potentially useful for defense and other factors, throwing in the historical elements also with the First Amendment and so on.

the Second Amendment and so on, or go into propagating this culture of an attitude to guns. And that's why the NRA is just so powerful in the US, and it's just not a force in other countries, to anything like the same extent. Tied to this is this issue of the gun culture are the socioeconomic factors in the US, which I think are predominantly responsible for the higher rates of crime there. And again, I point to some eight-fold higher homicide rates among African-Americans compared to non-Hispanic whites. And some of those go away if you control for differences in income and education, but not all of them. So there do seem to be significant problems, especially in a lot of urban areas,

especially in the South, especially in African-American, urban areas of difficulty in accessing education, poor employment prospects, poor what we call social capital or a cultural capital of the families. I won't go into that, but also a heavy influence of drugs and gang subculture in these areas. And of course, this happens, you see this sort of thing in all countries around the world, but it seems to be substantially worse in the US, and especially in some urban areas in the US than it is in other countries.

And this seems to be reflected in the much higher homicide rate, again, especially amongst African-Americans. So my overall diagnosis is if we're going to use such language is. that primarily the reason for the US's much higher homicide rate than comparable countries is their gun culture and their socioeconomic problems. Now, fixing these problems is by no means easy because they're connected to so many different things. The political system, the cultural heritage, social ideals, the education system, the health system, the employment prospects, the economy. It's all tied together. So simply introducing a few pieces of gun legislation is not going to solve these problems.

The gun legislation might help ameliorate the situation, but it's not fundamentally, I think, going to address the socio-economic or gun culture issues. And I think until the US finds a way of addressing these deeper problems, they're going to experience substantially higher rates of violent crime and homicide compared to other developed areas of the world, even if they'd introduce much harsher anti-gun legislation, which of course they probably won't do, given all the other factors of how politically sensitive this issue is. So that's somewhat of a depressing note, because there's no easy way to deal with those problems of the socioeconomic differences in the gun culture and so on. But I think, as I mentioned before, there are some principles of good regulation that can

potentially be followed, including, to recap the gun patrols, the problem-oriented policing, the focusing on people who are more likely to commit offenses, such as people who committed to crime misdemeanors or households where there's known domestic feuds, and generally just to try and experiment and collect better evidence and better data about what works and what does it. So this has been a very long episode, but I think I won't split it up because it's a bit of a special episode in the sense that I specifically wanted to address this area of current issue after the recent incident in the US. So I hope you enjoy the episode. I imagine probably this topic is, well, I mean, this topic's obviously controversial, so if you disagree with my conclusions or you think

that there's a particular piece of evidence that I missed or something like that, please feel free to email me. My address is Fods12 at gmail.com. That's F-O-D-S-12 at gmail.com. Even better than that, you can go to the Facebook page. Just type in the Science of Everything podcast on Facebook and give us a like. But you can also post comments or rebuttles or ask questions there about the episode or past episodes as well. If you enjoy the episode, please recommend the podcast to friends or family. Spread the word about the podcast. Jump onto iTunes and give me a

favorable review and rating. It helps to increase listenership. Thanks for listening and I'll talk to you next time.