Monday, 31 May 2004

Libertarian futures

More thoughts on libertarianism and the efficacy of the Libertarian Party from Will Baude, who is looking for some sort of cross-party soft-libertarian caucus to take up the mantle of advancing libertarianism. He also writes:

I’d also like to see a group with the microphone and legitimacy of a national party that didn’t run fringe candidates but rather conducted in-depth research into the voting records and announced positions of candidates in every house, senate, and presidential race, and announced how closely they hewed to libertarian orthodoxy (on those things—unlike, say, war—where such an orthodoxy exists).

The Republican Liberty Caucus does this—or at least did this for a while—by compiling two “interest group” ratings for legislators—one each for economic and personal liberty.* Not sure if they still do it… anyway, I used the ratings in at least one iteration of the now-legendary impeachment paper. Unfortunately you can’t find RLC publicity standing on the street outside their headquarters, much less in the halls of power or the public discourse; however, I think that Republicans who want to push their party in a more libertarian direction won’t find a better place to put their money.

* From an econometric point of view, I generally view interest group ratings as near-worthless in comparison to item-response models, but they have clear P.R. value—witness the widespread play the National Journal’s rating of John Kerry received, for example.

2 comments:

Any views expressed in these comments are solely those of their authors; they do not reflect the views of the authors of Signifying Nothing, unless attributed to one of us.

“Near-worthless” because they mis-weight issues (underweighting, for example, issues on which there haven’t been recent votes)? Or for some other reason?

 

Near-worthless, in the sense that actually manually selecting cases doesn’t make the measure any better (and often worse); for the purposes that interest group ratings are used for in day-to-day politics (publicity, fundraising, etc.) they’re OK, but for use as right-hand-side variables in a quantitative model (i.e. what I do for a living) you are better off with something based on a scaling method.

Most interest group ratings, when you boil them down to their essences, are essentially party agreement scores. (Case in point: almost nobody scores in the middle of the ADA rating scale.) For long, boring reasons I won’t go into, it’s better in a model to use ratings that are more continuous.

 
Comments are now closed on this post.