Analyzing the Bankruptcy Vote

I and others have thrown a lot of rocks at the Bankruptcy Bill, and there have been a lot of loose accusations that congressmen who voted for it were "bought off" by the credit card and car companies. Now I think charges of corruption are pretty serious, and worthy of futher investigation.

To sort out the dynamics of who voted for the bill and why, I decided to dust off my political science skills. For the last week or so I have been collecting data on who voted for the bill, the median income of the congressional districts (both from techpolitics), campaign contributions from those interest groups supporting the bankruptcy bill (from Open Secrets), and who is on the finance committee (from the House website). I then ran the numbers through a statistical program to figure out if there was a relationship.

There are several hypotheses of why a member of the House would vote for the bill: they are taking large campaign contributions (the corruption thesis), they are from wealthy districts that don't care about the law (the representation thesis), and they are on the finance committee and are "getting along" (the committee thesis).

There is some evidence that members from rich districts and who received lots of money were more likely to vote for the bill, but the relationship is weak. Being on the finance committee doesn't seem to have any effect. The importance of these variables is dwarfed by political party, which should be no surprise: Republicans overwhelmingly supported the bill and only half the Democrats voted for it.

But our real interest is why Democrats would support bankruptcy "reform." If we look just at them, there are some interesting results. There is a slight (and I mean slight) suggestion that members from poorer districts opposed the bill. And there is a correlation between campaign contributions and the vote. These results support the worst suspicions of the left: Democrats were bought off by wealthy interests. However, the statistics are less informative than they might first appear. The entire model only has an adjusted r-squared of 6%. In other words, all three factors (money, district income, committee status) only explained 6% of the vote. The other 94% percent remains unexplained by these variables.

This is both good news and bad news. The bad news is that House members didn't really represent the economic interests of their districts. If they had, then the median income of the district would be a powerful predictor of the vote. But there is good news as well: House Democrats weren't bought off by big corporations. They may have voted for reasons we don't like - they wanted to appear pro-business, or just agreed with the substance of the issue - but they aren't shills for credit card moguls.

So we can (and should) be mad at those Democrats for being wrong. But the evidence simply does not support the idea that they are corrupt.