In many ways, the debate over curbing carbon emissions is more interesting than the debate over health care. The latter is more or less a straightforward discussion about economics — how much to subsidize health care for lower-income taxpayers at the direct or indirect expense of an increased tax burden on higher-income taxpayers, and to what extent private-sector insurers, warts and all, can be expected to deliver more efficient solutions than public-sector ones.
The debate over cap-and-trade, on the other hand, is a genuine moral dilemma, pitting the interests of present-day Americans against those of future generations both here and abroad. I have absolutely no sympathy for those who voted against the climate bill because they don’t believe in global warming; I do have some sympathy for those who weren’t willing to sacrifice jobs in carbon-intensive industries in their districts now to possibly save a village of Bangladeshi children who will be born 40 years hence. That is not meant to sound sarcastic; it is naive to pretend that there wouldn’t be losers from a bill that sought to increase the cost of carbon and it is naive to assume that a member of a legislative body who is subject to re-election every two years might not err on the side of his present-day constituents. Although the bill barely got through the House, in some ways it is amazing that it did in the midst of the worst recession in 70 years.
In any event, it would be useful to see how the 434 sitting members of the House (one seat in California remains vacant) tried to navigate the waters. As I did for health care, I built a logistic regression model that attempted to predict the likelihood of a particular congressman voting for the cap-and-trade bill as the result of a variety of factors. After much trial and error, the factors that look to be most significant are as follows — factors are listed roughly in declining order of significance:
Ideology. The overall liberal-conservative bent of a Representative, as determined by DW-NOMINATE scores, which run from -1 for very liberal to +1 for very conservative. In this instance, I use the “common space” version of DW-NOMINATE scores, which are slightly less robust overall but place Representatives and Senators on a level playing field, which will come in handy later when we try and predict (as we will in a subsequent post) how the Senate will vote on the bill. Scores are as of the 110th Congress; for freshman Congressmen, they are extrapolated from Progressive Punch scores.
District Partisan Lean. The PVI (Partisan Voting Index) in a district was a highly significant variable; Congressman in Democratic-leaning districts were more likely to vote for Waxman-Market and those in Republican ones more likely to vote against it, all else being equal.
Lobbying Money. As in the case of health care, funds raised from certain types of PACs are a significant predictor of a representative’s vote, although the money in this case cuts both ways. Whereas receiving contributions from coal industry PACs decreased the likelihood of a vote for Waxman-Markey, contributions from nuclear and alternative energy providers significantly increased it. I also looked at contributions from oil and gas industry PACs, public utility PACs, and agribusiness PACs, but these had no statistically significant effects. All data is taken from the Center for Responsive Politics and covers the 2004 cycle forward; contributions are divided by the number of cycles a Representative has participated in as a Congressman or as a candidate.
Carbon Emissions. I use county-by-county data on the amount of carbon emissions per capita in a particular area, as determined by Project Vulcan. This requires us to map the county data onto congressional districts by dividing the population of a county evenly among all congressional districts that occupy a part of its geography. Estimates are in metric tons of carbon consumed annually per capita. The carbony-ist district is the At-Large one in Wyoming, which produces 36.3 metric tons of carbon per capita; the least carbon-intensive are the 10th and 11th Congressional Districts of New York, which are both located in Brooklyn and are responsible for 1.1 metric tons of carbon per capita.
Poverty Rate. Although the Waxman-Markey bill contains provisions to refund a portion of increased energy costs to lower-income consumers, it was nevertheless more likely to receive support in districts where the poverty rate is low. Alternate measures of economic welfare like per-capita income work almost as well in the model and could serve as reasonable substitutes for the poverty rate.
Employment in Carbon-Intensive Industries. Lastly, the fraction of a district’s jobs that are in manufacturing, mining or agriculture was a good predictor of voting on Waxman-Markey (although this variable was significant only at the 90 percent level and not at the 95 percent level).
*-*
Overall, this set of variables is pretty useful and explains about three-quarters (R-squared = .74) of a particular Congressman’s vote on the climate bill. The model predicted 401 of 431 votes correctly.
The congressmen deemed most likely to vote in favor of cap-and-trade are as follows:
These are liberal Democrats in liberal areas with relatively low carbon output. All on this list did indeed vote for Waxman-Markey. Waxman and Markey themselves, incidentally, ranked as the 11th and 15th most likely Congressman to vote for the bill, respectively. All on this list did indeed vote for the bill.
The congressmen deemed least likely to vote for Waxman-Markey are these:
There’s Ron Paul! Few surprises here either; these are some very economically conservative Republicans in districts that tend to consume a lot of carbon. Cynthia Lummis was given only about a 1-in-4.5 million chance of voting for the bill — she didn’t. Neither did any of the other Congressmen on this list, although Jeff Flake of Arizona missed the vote.
Where were the surprises then? These are the Congressmen the model thinks were most likely to vote for Waxman-Markey but in fact didn’t:
The first three names on this list — Pete Stark, Dennis Kucinch, and Peter DeFazio, apparently all cast nay votes on the bill because they they thought it was too conservative. One imagines that they might have voted for the bill nevertheless if their votes were necessary to secure passage — but as it actually went down, they didn’t. Not listed here is Alcee Hastings of Florida, who was given a 99.8%+ likelihood of voting for the bill but did not cast a vote either way.
Next, here are the least likely yes votes.
Note of these yea votes were truly all that unlikely, the closest thing to an exception being John McHugh from upstate New York, who is generally fairly conservative and represents a somewhat poor district.
So what are the general takeaways here?
— People on the whole were pretty rational in trying to balance “selfish” traits (their own ideology; lobbying influences) against “unselfish” ones (the economic and political characteristics of their districts).
— Nevertheless, the playing field is fairly broad, as there are quite a few representatives for whom these traits balance out in ambiguous ways. Some 95 representatives — about 20 percent of the House — were deemed to have between a 10 percent and a 90 percent chance of voting for the bill and can reasonably be described as swing votes.
— Cap-and-trade differs from health care in that there are particular private sector groups that would appear to benefit from its passage: nuclear power and renewable energy providers. Although the nuclear energy lobby is small, and the alternative energy industry lobby is very small, they nevertheless appear to have had some influence; nuclear is a big, untold part of this story. On the other hand, the effects of the agricultural lobby appear to have been mostly neutralized, perhaps because of concessions made in the bill to farm-state Democrats.
— This bill faces long, but not impossible, odds in the Senate — we will cover that in more detail tomorrow.