I am sad now

Dear Politicians,

It has come to my attention that you are more or less incapable of doing anything in America at the moment.  Please quit mucking about and do something useful.  A couple basic guidelines:

1. Evidence - please use it, and please present it in context.  Large numbers are scary, but large numbers when compared to other historical or projected large numbers are much less so.
2. Consistency - I know it is bad politics to hold logically consistent beliefs, but I think that maybe if you don't someone should call you on it, and in a public easy to understand way, not in the clever but not terribly accessible way my favorite bloggers do.

Point 1 is for the pragmatists.  If you care about the impact of your legislation, and not terribly much on the abstract values behind the way it works, then you should be a slave to Evidence.  For those of you who are primarily interested in a legislative enactment of some value-system, you should be checking to make sure your enactment doesn't undermine your values by its effects.

Point 2 is for values-politicians.  If you're holding a core set of beliefs and want to enshrine them in law, then you should at the very least make sure that they don't contradict each other.  For pragmatists, this might not matter as much in your implementation, but you should be watching the aggregate effects to make sure they still do what you hope they will.

I know this is a hopeless and empty plea, but I'm burnt out on real world politics and needed to get that off my chest.  Expect more bounds on taxation and other abstract policy discussions in the near future.

Love,
Rory

Prognostications

This article was interesting, in that it gave strong evidence for trusting Krugman when he makes predictions about the economy.  I also found it interesting that lawyers were worse prognosticators than other professionals.  However, the liberal-means-right thing is very clearly biased by the time-frame and nature of the predictions studied.  All of the predictions came out of the period of time immediately before the 2008 elections, and many of them directly related to election outcomes.  Liberals were therefore far more likely to be good predictors merely because their side happened to win that time.  I'd love to see a similar study over a longer timeframe and including multiple changes in party power.

Internet Digest: the things that are interesting in my news feed

Rather than pretending to come up with original content today, I'll just link you to all the things I read and vaguely cared about this morning.

This post on unemployment and education level was interesting, but I wish it had a graduate degree line.  Might help inform some of our earlier discussions about the goodness/badness of grad school.

The Japanese regulate the darndest things.

Apparently Paul Krugman's is the right blog to read.

Here, we use Paul Krugman, a New York Times columnist, as an example to show how our PVS system works.

From a random sample of his columns and television appearances, we tallied 17 testable predictions. Krugman was awarded one point if he either predicted an event would happen and that event took place, or if he predicted an event not happening and that event ultimately did not take place. When Krugman made a prediction and the opposite outcome occurred, we subtracted one point from his total. Hedged predictions did not receive any points.

Of the 17 predictions, there were 15 instances where Krugman made a correct prediction and only one instance where he made an incorrect prediction. With the single hedged prediction included, Krugman’s final tally was +14. We took the final tally (14) and divided that by the total number of testable predictions made (17), which we then multiplied by ten. Krugman’s final score was 8.235#.

All in all the internet has given me a very validating morning.