So I’ve decided to
move to Blogger permanently. And to celebrate, let’s post on something of
broader interest: gun laws. When there’s a shooting in Australia, the normal
reaction is that we have to have tougher gun laws to stop them. In America
(and, sadly, sometimes in Australia),
the reaction is sometimes that there needs to be weaker gun laws so more
civilians can stop them. The reasoning here is fairly simple. If everyone has a
gun, and everyone knows everyone has a gun, and everyone knows that everyone is
ready, willing and able to use their gun on the smallest provocation, and no
one wants to be shot, no one will shoot anyone because they know they will be
shot. There looks like there are a few possible problems with this theory, but
who knows, maybe it works.

Of course, there is one rather dramatic
problem with this theory. The US has the most liberal gun laws in the western
world and has the highest homicide rates in the western world. How, oh how,
then are we meant to believe that more guns leads to less crime? The answer is
that we’re apparently comparing apples and oranges. The things we’re allowed to
compare are regions that are adjacent in space or time. And if we compare
region r at t with region r at an immediate subsequent
time just after gun laws are tightened, the crime rate rises. (Or maybe it
doesn’t rise, but then they say it
would have fallen anyway, and would have fallen more without tight gun laws
Spot them that.) Or if we compare region r with tight gun laws with
adjacent region s with looser gun laws, region s has less crime.
In America, the regions are counties or states, which can change gun laws
between areas. And that’s meant to be the conclusive evidence.

But it seems to me that this is less than
conclusive, especially given the rather large elephant in the room that is the
horrifically large American murder rate. It seems to me that there are three
possible effects of tightening gun laws.

(a)      Decreasing
the effect of having lots and lots of gun around, hence
increasing crime.

(b)      Decreasing the access to guns of
those wanting to commit crimes, hence decreasing crime.

(c)      Decreasing the cultural penetration of guns, and hence
decreasing gun-related crime, and probably crime overall.

Which of these effects is the strongest?
Well, that’s an empirical question, and we shouldn’t try and answer it from the
armchair. But we can tell some things from what we know about those three
effects. First, if© is an effect at all, it will be very delayed and very
diffuse. So you wouldn’t expect to see it in the side-by-side analysis that
gun-nuts run with, and would expect to see it in the country by country
analysis that we gun-banners run with. That doesn’t prove anything, because if
there’s no cultural penetration effect you wouldn’t expect to see it either.

The more interesting thing is that if (b) is
an effect at all, and frankly it beggars belief to think that it isn’t, you
also would expect it would be much weaker in side-by-side regions as
compared to regions from quite different areas. The fact that guns can be
carried across county borders and state borders with no hassle at all will mean
that the access of guns in a region is a function of the gun laws in that region
and adjoining regions. And plausibly something similar happens
immediately after a tightening of gun laws, because non-criminals move into
compliance with the laws before criminals do. So while there will be an
immediate and local drop in the deterrence effect, there will be a slow and
diffuse drop in the access and cultural penetration effects.

What does that imply? If (b) and© are
large effects, then we’d expect that the side-by-side region data would not
show this up, but would show that there’s a short sharp cost to tightening gun
laws, because the only spatially and temporally local effect is the deterrence
effect. In effect, various regions (or, more precisely, time-slices of regions)
are playing a prisoners dilemma. Each has a motivation to slacken, or at least
not tighten, gun laws. But the community of regions is best off if everyone
tightens gun laws. Of course, trying to convince libertarians that a particular
situation is a prisoners’ dilemma is a little harder than trying to convince an
Australian of something so silly as that more guns leads to less gun crime, so
I don’t expect this argument will get very far, but one can only try.