Tao Li
--------------
www.people.fas.harvard.edu/~li7
---------- Forwarded message ----------
Date: Thu, 2 Jan 2003 13:22:32 -0500 (EST)
From: Gary King <king(a)harvard.edu>
To: Phillip Y. Lipscy <lipscy(a)fas.harvard.edu>
Cc: Tao Li <li7(a)fas.harvard.edu>du>, dkane(a)latte.harvard.edu
Subject: Re: GK Question
this is a fine question and well put. the answer, however, is
theoretical. I.e., we chose one definition of incumbency advantage. You
certainly could choose another. Whether one chooses one vs the other is
not a matter of inference but one of theory or preference or normative
choice.
Gary
On Wed, 1 Jan 2003, Phillip Y. Lipscy wrote:
Basic question about the GK paper page 1143, the
definition of incumbency:
The definition of incumbency incorporates candidate quality by not making the
counterfactual that candidates are of similar quality. I'm not sure what to
make of this... Does it really make sense to include this as a part of
incumbency advantage?
ie, assume that we conduct elections and winning the election confers no perqs
whatsoever. According to this setup, wouldn't we get a positive psi because
incumbents have a tendency to be higher quality candidates (by virtue of having
won an election before, a selection effect)? I'm kind of uncomfortable calling
this "incumbency advantage" because there's correlation but no causation
here:
i.e.
not:
incumbency -> increase in votes
but rather:
1. high quality candidate -> incumbency
2. high quality candidate -> increase in votes
Doesn't this mean psi is biased upwards as an estimator of incumbency advantage?
thanks,
Phillip.
-------------------------------------------------
Phillip Y. Lipscy
Perkins Hall Room #129
35 Oxford Street
Cambridge, MA 02138
(617)493-4893
lipscy(a)fas.harvard.edu
First Year Student, Ph.D. Program
Harvard University, FAS, Department of Government
-------------------------------------------------