(and Guatemala's famed "Pollo Campero," which is their knockoff KFC,
although the locals all swear that Campero is the original).
But I'm getting ready to roll on, probably up to Uspantan, up in the
central highlands. I've spent about two weeks hanging around Santa
Catarina, and the last couple interviews haven't turned up a lot of new
information, though they have helped me to crystallize my thoughts a bit
about the place.
So I've started to write up an account of Santa Catarina, which as I see
it, does confirm the effects of institutional mechanisms posited in
Krister's earlier work (as it should, since it's right on the regression
line.)
But I've been thinking a lot about a potential problem in this
work--this is an idea which I wrote about earlier, but it's been preying
on my mind quite a bit the last couple of times I've sat down to type
this case study up...
One of the institutional mechanisms which scholars of Krister's ilk
believe is important in promoting desirable forestry outcomes is
electoral accountability. If local voters punish local politicians who
fail to promote the kinds of forestry outcomes the locals want to see,
then politicians will tend to pursue those desired outcomes.
There's a problem, though, that has to do with the way we're measuring
"citizen demands" for forestry.
On one hand, where people demand better forestry outcomes, you're likely
to see better outcomes over the long term. Over the short term, though,
the places with a large number of citizen demands for forestry are
likely the ones that have really screwed up forestry outcomes. People
complain--make "citizen demands" for forestry--when things aren't going
the way they want them to go, and they won't complain so much when
things are going all right.
So there's an endogeneity problem.
There's an issue of strategic interaction here that creates the
problem. Mayors that know locals are going to demand improved forestry
outcomes--like in Sta. Catarina, where the mayor knows that if a lot of
deforestation takes place, the deforestation will mess up the watershed,
and the sources of water will start to dry up. The people know this,
too, but it hasn't happened yet. And it probably won't happen, either.
The mayor knows that he'll be out of a job if he lets the water sources
run dry (because of deforestation), so he will work pretty hard to make
sure that said deforestation doesn't take place.
And as long as he does his job, there will be nothing to complain about.
So what the mayor is concerned about is not the citizen
complaints/demands. Because they haven't happened yet. Because he's
done his job pretty well.
What he's worried about are the potential for future complaints/demands,
and future electoral punishment, as a result of a failure to prevent a
big problem with the water.
But if we measure citizen concern for water protection or forestry
management by using a variable that is the "frequency of citizen
demands," we'll be missing out on this strategic interaction effect. So
we need to figure out a way to instrument for "citizen demands"
Of course, instrumental variables are hard to come by. But I wonder if
a geographic instrumental variable might do the trick. Edwin
Castellanos tells me that this water-protection dynamic is primarily a
phenomenon in the highlands. So I might be able to construct an
instrument using ArcGIS, or just by using a geographic variable like
"altitude," which I have already.
But then that leaves out the possibility that people might be demanding
forest conservation for other reasons--the most notable one maybe being
the desire to protect resources in a sustainable fashion for economic
reasons--maybe for the purposes of logging.
mmm...
No comments:
Post a Comment