The NBER site should have Duggan's full working paper:
www.nber.org/papers/w7967 .
Duggan claims that circulation data for Guns & Ammo magazine can serve as a
good proxy for gun ownership rates. Using this proxy, he goes on to show
that higher gun ownership causes higher crime rates.
My worry in this is the proxy. There is no reason to believe that the
proportion of gun owners who purchase Guns & Ammo magazine remains constant
over varying county rates of gun ownership. Plausible arguments can be made
that the proportion will vary with ownership rates. It may be the case
that, in high ownership areas, fewer owners will purchase the magazine as
they can more readily acquire information from other owners; while in low
ownership areas, more owners might purchase the magazine to learn about
upcoming gun shows, technical info on new ammunition and such. Or, it could
be the case that there are network effects influencing readership -- owners
may be more likely to purchase the magazine if they know others will be
reading the magazine. In this case, ownership levels will be relatively
underestimated in low ownership areas and overestimated in high ownership
areas. Duggan finds that a 10% increase in ownership as proxied by
readership increases the homicide rate by 2%. If the first case above is
correct, a 10% increase in readership corresponds to a much higher increase
in ownership, making for a much smaller marginal effect. If the second case
is correct, the increase in ownership necessary for a 10% increase in
readership is smaller and the marginal effect cited is underestimated (which
would strengthen his results).
Duggan also argues that Lott & Mustard's study of the effect of concealed
carry legislation is flawed because their standard errors assume
county-level variation in concealed carry, while changes in such legislation
is enacted at the state level. Revised standard errors reduce their results
to insignificance.
I'm hoping that someone out there with more technical expertise will find
the holes in Duggan's work...I'd sorely miss being able to cite Lott &
Mustard in debate....
Eric Crampton
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of
[EMAIL PROTECTED]
Sent: Thursday, January 18, 2001 2:02 PM
To: [EMAIL PROTECTED]
Subject: More Guns, Less Crime?
I hope I'm not starting a flame war here, but I just read an article in _The
Economist_ that mentions the now-famous Lott and Mustard study of concealed
carry laws and two subsequent studies whose conclusions disagree with it.
After
searching on the web, I was able to read the Lott and Mustard article and
the
Derzhbakhsh (?) and Rubin article. (The third one is a working paper that
hasn't been published yet, so I could only find its abstract.)
I'm having trouble deciphering the competing claims, and I was wondering if
anyone on this list who doesn't have too strong an ideological or emotional
investment in the debate (a naive hope?) might be able to enlighten me
further
on this issue.
Many thanks,
James
