4. is particularly persuasive.  The old adage in politics is that if your
goal is to find a candidate that you agree with on every issue, run.
Otherwise voters have some beliefs held more deeply than others and accept
that the politician who supports the view on taxation they prefer does other
things they don't like, but do not value as much. They are "buying" a
package.

Given the likelihood of being the deciding vote and the costs of getting
good answers from politicians to tough questions it is a wonder that anyone
votes at all...

Regards,

Brian Moore
ESI Corporation




-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of
fabio guillermo rojas
Sent: Monday, August 26, 2002 9:03 PM
To: [EMAIL PROTECTED]
Subject: Re: Median Voter and Sampling


> So what are you getting at?  Since there is a series of elections, each
> with a different median voter, the MVT doesn't actually predict that the
> median general voter gets his way?  Or what?
>                         Prof. Bryan Caplan

I think that applications of MVT are very, very sloppy. Four
criticisms:

1. You seem to assume that policy responds quite well to public
opinion. You assume that if opinion shifts, policy will quickly follow.
I believe that policy is very "sticky" with respect to public
opinion. To make it econo-talk, I think policy is not very
elastic with respect to changes in the median voter.

2. Institutions are designed to prevent policy from being overly
sensitive to public opinion. Ie, we don't have elections every
day. We create rules that allow policy makers to resist
every whim of opinion. Examples: rules for changing the constitution,
judicial dependence on precedents, etc. In a sense, institutions
play the role that contracts do in the labor market - set
practices over some time period (ie, you've bought labor
at price X and the employee can't leave just because the price
is now more than X).

3. When people (ahem, Mr. B.C.) say "look - puzzle - people want
X but we get Y" - the poll that measures opinion is probably
a random sample of adults, or maybe voters. But as I've argued
before, this might not be the relevant group. Maybe it's
party activists, or party-rank and file. Policies may have
select audiences and there is no puzzle until you show that
the relevant audience does in fact strongly oppose a policy.

4. Cognitive limitations: I'm no expert, but my hunch is that
many people are only willing to get worked up over a small
# of issues - taxes, abortion, immigration, defense... and
the dedicated might add their favorites like gun control
or affirmative action. Therefore, it's no risk to screw
the voter on an issue as long as you don't do it on certain
big issues. Therefore it's easy to get a list of dozens
of issues and find a descrepancy - what's so puzzling about
that?

So my beef isn't the MVT per se, but the knee jerk use of it.

Fabio








Reply via email to