Performance Criteria for Politicians

In this age of quantification, how do we judge a politician's performance? By doing before/after comparisons of public opinion dynamics? That sounds silly. By, calculating the change in the "misery index or murder rate"? In that case, you're assuming that the politician is very powerful in shaping events. Ronald Reagan asked "are you better off than you were 4 years ago?" So, random macro economic events determine a politicians' fate? Economists have written many empirical papers on whether CEO's are rewarded in terms of high pay for luck or high performance.

Now, how do we judge Big City Mayors. Is it all about personality (Koch, Rudi)? Unfortunately, the LA Times letters to the editor appear on a 1 day lag but Matthew Duggan has a very funny letter that will appear tomorrow on line. I will update.

Given that politicians make thousands of decisions and given the diversity of the voter base and our individual priorities over these voting decisions on a wide range of issues ranging from abortion, to climate change, to welfare to China, how do we rank them?

A structural IO economist would map these decisions into attribute space (this is sort of what Keith Poole does with his factor analysis of voting in Voteview). His approach essentially returns to Hotelling's ice cream selling on a long narrow beach as all politicians are mapped into a liberal to conservative continuum.

Now in the case of IO, I know the attributes of a Ford versus a GM car before I make my purchase decision. In the case of politicians; the attributes are staggered. I see the choices that the incumbent made but I don't observe what choices his challenger would have made in the same situation.

The politicians who run for re-election know this and can make choices over how to market themselves to the public during the election season. So, my point is that the political market is pretty funky.