There's been some debate among the climate hawks about last night's election returns. Politico posted a story suggesting that the toll was especially hard on Democrats who supported the landmark climate change legislation passed by the House last summer. Kate Sheppard observed that quite a few of the Democrats who opposed the bill also lost their seats, and Chris Mims (formerly of Scienceblogs, now at Grist) argues that the election wasn't a referendum on climate change, pointing to the significant numbers of bill supporters who survived last night. NRDC finds the same thing.
The tricky thing about just looking at the raw numbers is that some of the people who lost last night would have lost no matter how they voted, while others might have held their seat if they had voted differently. Most of the Democrats who supported the bill would never have lost re-election. Rick Boucher is the foremost example of a Democrat who lost because of his vote on the bill, with his former chief of staff arguing compellingly that Boucher's aggressive work to pass the bill was the main thing that caused his coal-economy constituents to drop him. But he didn't just vote for the bill, he actively negotiated to pass it, and that's different.
To really get at this issue we have to do some statistical analysis. The hypothesis we're testing is straightforward: Representatives who voted for the climate change bill performed worse in this election than they should have.
Using Charlie Cook's PVI as a basic proxy for the underlying politics of the district (a statistic measuring the partisan trend in a district based on previous presidential races), I ran a series of regressions against the fraction of the vote going to the Democrat in each House district, as reported at the New York Times (for statistical reasons, I left out races with no Democratic candidate, no Republican challenger, or where the Democrat got less than 10% or more than 90% of the vote). As you'd expect, the regression is highly predictive, with the regression explaining 85% of the variability. Adding in information on the party of the Representative raises that predictive power to 90% (by accounting for some of the effects of incumbency).
I then added two variables: the vote on final passage of Affordable Care, and vote on the climate change bill. Only 36 members crossed party lines on Affordable Care, and only 52 crossed party lines on climate change. I include Affordable Care in the model because that vote got a lot more attention, and there's good reason to think that responses to the debate over healthcare reform would have a significant impact on voting behavior. It also helps us separate the effects of party-line voting from the effects of actually voting for the climate bill.
Adding the Affordable Care vote only boosted the model's predictive power to 92%, but it is strongly statistically significant. Adding the climate change vote on top of that has no further effect (slightly negative, actually) on predictive power, and the effect of voting for the bill is not statistically different from zero.
The figure here shows the residuals for the model without the climate change vote. Residuals show the difference between observed data and the model, and help us see if there is any systematic trend in that unmodeled variation. The higher up you go on the graph, the better a candidate did relative to the model's predictions, the lower they are on the graph, the worse they did relative to expectations. The green dots are Democrats who voted against the bill, and Democrats who lost their seats are marked with an x. If voting against the bill were necessary to save Democrats, we should see those green dots clustered toward the top of the graph, but they are distributed fairly evenly.
There's no evidence that voting for or against the climate bill did Democrats any harm at the polls.
There's no hiding the House Democrats' bloodbath, with more than two dozen members who voted for the Pelosi-led climate bill losing their seats, and more likely to fall as the final tallies come in. The outcome sends a strong signal to moderate lawmakers as they consider any risky votes in future Congress' on energy and environmental issues.
"It's going to be cap and tax forever more, and I don't think any of these guys are ever going to touch it again," said Linda Stuntz, an industry attorney who held a top Energy Department spot during the George W. Bush administration. "I think anyone who thinks there's vitality left is kidding themselves."
Boucher's defeat is perhaps the most stinging given the central role he played in brokering key pieces of the legislation to make it more friendly to his home state's coal industry. Over the last 18 months, Boucher has defended his work on the climate bill, saying it's much better than the alternative of Environmental Protection Agency emission control regulations.
But his Republican opponent, state House Majority Leader Morgan Griffith, still found political leverage and ultimately won, 51 percent to 47 percent.
"I don't think there's any question about it, cap and trade was the issue in the campaign," Andy Wright, a former Boucher chief of staff, told POLITICO. "If Rick had voted no, he wouldn't have had a serious contest."
"...members who voted for the Pelosi-led climate bill losing their seats..."
That's exactly the talking point that this analysis convincingly refutes. Repeating it doesn't make it any more true.
I would be very interested in seeing more of your analysis, including the data. Using what I think is the same data and a very straight-forward (i.e. simple) regression, I get slightly different results, though the practical result is the same. For my regressions, all three factors show as being statistically significant, though only CPVI has any practical significance, with R-squared hovering around 80% as long as PVI is included.
I am opposed to both legal and illegal immigration. This country is overpopulated, and 21 million Americans are out of work.
Dear Joshua. I'm an Oberlin College student writing a research paper on climate change in the 2010 election. Your analysis has been very useful, I'm wondering if you'd be open to having a conversation about your findings. What i'm curious is Rick Boucher seeing as his was a case where his position on climate change was actively pushed upon voters. If you factor in the style of campaigning does this make any difference in your findings? Also, these democrats who voted against climate change legislation, are they blue dog democrats? If so, does this also effect your findings?