Please turn on JavaScript

Brooks Wilson's Economics Blog: models
Showing posts with label models. Show all posts
Showing posts with label models. Show all posts

Wednesday, December 10, 2008

$650 Million Did Not Buy Obama the Election!

A spate of news stories and blogs are declaring that President elect Barak Obama bought the presidency for $650 million, a point I dispute (see Big Money Bought Obama the White House or £500m election fund 'bought Obama victory').

The stories are of two types, by raising so much money Mr. Obama was able to bamboozle voters with tons of TV, radio and other advertising, or that the money was somehow illegally raised. As to the second point, illegally raising money, I will only comment briefly. Mr. Obama should comply with the law no matter how difficult the compliance and stupid the law. Without any evidence, I believe that when all is said and done, Mr. Obama will be shown to have complied with the law about as well as anyone else.

Money and votes seem to be correlated, but the question is one of causality. Does money buy votes or do good candidates raise more money. Steven Levitt and Stephen Dubner answer that question based on Levitt's peer reviewed work in their best seller, Freakonomics, certainly a valuable and entertaining read. Levitt looked at elections in which the same congressional candidates ran against each other in consecutive elections. This happened a lot, over 1,000 times in his sample.

Levitt then compared the votes and money raised in the two elections. He found that a candidate can cut his or her spending in half and only lose about one percent of the votes. Candidate Obama won by a 7.2 percent spread. It wasn't even close. Mr. Obama won because, given the economic and social conditions, he was the better candidate.


Read more!

Friday, December 5, 2008

Football Ranking Model

The ugly turn of events in the Big 12, the loss of OU to Texas, Texas to Tech, and Tech to OU has yet again placed the BCS in a predicament. Two teams play for the "National Championship" and only one has a perfect record. The remaining top ranked teams have a loss. Texas, Florida, OU, Tech, USC and Penn State can all make a claim that they should have a shot at the title game. Even a couple of undefeated teams with weak schedules will tell you that they should be in the championship.

Lacking a playoff system, the BCS relies on a combination of human rankings and computer rankings to select teams that play in the championship bowl.

But my question is not whether you prefer human rankings or computer rankings, but if you had to construct a computer model to rank teams, which variables would you use?


Read more!