Joseph Olson writes:

Ask any gun dealer about women buying handguns for self protection.
A dealer friend tells me that over =BD of his handgun purschasers are
women but also tells me that most of them are very concerned that NO ONE
will know that they have a gun. It’s his opinion, after talking with
these REAL buyers, that none of them would admit gun ownership to a
surveyor. None of them subscribe to “gun magazines.” And, if they
commit suicide, I’ll bet $10 they don’t use the gun. So there is a HUGH
hole in the information base. None of these women kill baby animals
either.

And none of this matters. Even if we (falsely) assume that women
never use guns for suicide (so their gun ownership is not counted by
Cook and Ludwig’s measure) how on earth is that going to create a
spurious correlation between gun ownership and burglary?

I find this result more credible than Lott’s (More guns, less crime)
and Duggan’s (More guns, more crime).

In the case of Lott for two reasons:

(1) In Lott’s case ony 1 or 2 per cent of people get permits (and those
people were from low crime risk groups). This small difference in the
number of people with guns is not sufficient to plausibly cause the
changes in crime rates that Lott attributed to the carry laws.

In the Cook-Ludwig study there is a much larger variation in gun
ownership, almost 50 percentage points if I recall correctly. It does
seem plausible that these large differences couls cause relatively
small differences in burglary rates.

(2) In Lott’s case when you break crimes down in smaller categories you
don’t see the differences that the deterrence theory predicts. For
example, juvenile homicides decline as much as adult homicides, even
though they are not allowed to get permits.

In the Cook-Ludwig study, the analogous breakdown is hot
burglaries/non-hot burglaries. Gun ownership did not deter hot
burglaries, once again contradicting the deterrence theory. This
actually a stronger argument against deterrence than Lott’s results
since you could plausibly argue that because of reason 1 above, Lott
was not a good test of deterrence.

In the case of Duggan the reason is that the suicide fraction appears
to be a better proxy than gun magazine subscriptions. Changes to the
prices of the magazine would surely affect the number of subscriptions
without their being any change in the gun ownership rate.