Epic Battle 3: Who are the swing Senators?

Part 1 looked at an analysis of the House vote by stat master Nate Silver of Silver “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.”

Now I’ll look at the same model applied to the Senate vote by Silver, “How Can the Climate Bill Get to 60 Votes?” Silver’s is not the only analysis out there of swing Senators — I will be looking at others and I definitely have some issues with Silver’s picks. But he does have the only quantitative analysis, and an impressive track record of picking winners and losers. Even where you disagree, he makes you say, “Hmmm.”

While Silver’s analysis is first-rate, he often buries the lede. In this case you have to wade through a staggering amount of data to learn the bottom line:

Overall, this is a slightly better assessment than I expected. Although the model considers only 52 Senators to be more likely than not to vote for the bill, there are somewhere between 62–66 votes that are perhaps potentially in play…. Further compromises would almost certainly be needed, some of them designed to placate as few as one senator. The question is how many ornaments the Democrats could place on the Christmas Tree before it starts to collapse under its own weight.

Still, Silver assumes, correctly I think, “that there won’t be nay votes from the left, of which there were almost certainly three (DeFazio, Kucinich, Stark) in the House.” I can’t imagine a progressive would be the one to bring down the first serious climate bill in US history.


A few key caveats that Silver makes are “the Senate will be voting on a different bill than the House’s version” and “The overall political tides may have shifted by the time the Senate considers the bill because of changes to the economy, Obama’s approval rating, gas prices, and perhaps even the weather, etc.” Indeed, the economy would not appear to be moving in a favorable direction, although if the economy were robust then oil prices would be over $3 a gallon, so you the opponents will always have their do-nothing talking points.

The key vote in the Senate is not really the vote on final passage, which this model would seem to get at, but rather the vote on cloture, or breaking the filibuster, which would require 60 votes. It’s not all that uncommon for a senator to vote for cloture and then against the underlying bill, or vice versa, although it seems to happen less often for major issues like climate change legislation.

That was the point of Senate battle, Part 2: Sherrod Brown (D-OH) says he won’t filibuster climate bill.


Recall from Part 1 that Silver found a “pretty useful” set of variables that “explains about three-quarters (R-squared = .74) of a particular Congressman’s vote on the climate bill.”

Here are the factors affecting votes “listed roughly in declining order of significance”:

Ideology. The overall liberal-conservative bent of a Representative….District Partisan Lean. The PVI (Partisan Voting Index) in a district was a highly significant variable….

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….

Carbon Emissions. I use county-by-county data on the amount of carbon emissions per capita in a particular area….

Poverty Rate.

Employment in Carbon-Intensive Industries.

I’ll skip his three tables of the “nearly certain” and “extremely likely” and “highly likely” yes votes, those with a Probability of a Yes Vote “” Pr(Y) greater than 90%. Those 44 senators do include Lieberman, Carper, Webb, Stabenow, Levin, and Brown. “At the end of the day, a lot has already been done to assist the auto industry, and Brown, Levin and Stabenow are mainline liberal Democrats,” so they are very likely to vote for the bill (or at least cloture). Still, “their votes on the climate bill aren’t as certain as the analysis indicates.” Also, I would not put Webb at near 95%, but again I think he’ll vote for the bill. Heck, Rick Boucher (D-VA) voted for it.Then we get to “Likely Yes” votes:


These numbers seem high — I’d drop them all by at least 20 percentage points. Baucus seems the toughest of this lot. Yet Silver notes:

Montana consumes a fair amount of carbon, and Max Baucus is pretty conservative, which seems like a bad combination — until you see how much money he’s getting from alt/nuclear PACs — the most of any senator on a per-cycle basis. Arlen Specter was generally thought to be sympathetic to cap-and-trade legislation — and that was before he turned into a Democrat. North Carolina’s economy is fairly low-carbon, which should help to prevent Kay Hagan from defecting. The other two senators on this list, however, could be more problematic for the Democrats, as Claire McCaskill has already tweeted her concerns about cap-and-trade, and as Tim Johnson voted against cloture last year — although South Dakota’s economy, for whatever reason, is much less carbon-intensive than North Dakota’s.

Then we get to the “Possible Maybes”:

Silver notes:

These are three senators for whom the percentages tended to move quite a bit based on relatively small tweaks to the model. Snowe and Collins are almost certainly going to be necessary parts of any path to 60 votes and are almost certainly going to be easier gets than at least half a dozen Democrats. And I tend to think the model has erred a bit pessimistic on them here. But that doesn’t mean their votes are assured.

No, the Maine Senators’ votes are not assured, but I would add at least 20 percentage points to their probability of a yes vote.


Begich’s number looks about right, since his energy plan states, “Mark Begich will support national legislation to reduce greenhouse gas emissions by 80 percent by 2050 through a cap and trade system.” He may ask for (and receive), “revenues put into an Alaska Adaptation Fund to provide immediate assistance to Alaska communities already suffering the impacts of climate change and to help protect Alaska infrastructure from future damage caused by climate change.”

Sadly, while all of these Senators would provide the 53 votes needed to move climate legislation absent a filibuster, we will still need seven more to stop opponents from killing the bill.

I’m gonna say that Bayh will vote for the bill, but the deal with China that the administration is currently working on may be crucial (see “Does a serious bill need action from China?” and “Exclusive: Have China and the U.S. been holding secret talks aimed at a climate deal this fall?”)I also think Rockefeller will vote yes. The biggest omission from Silver’s Senate analysis is that he doesn’t factor in date of reelection. “Rockefeller was re-elected in 1990, 1996, 2002 and 2008 by substantial margins.” If he doesn’t plan to run again in 2014 at the age of 77, then he has no reason whatsoever to vote against this bill. But even if he does, we’re still talking 5 years from now. As Silver notes, “Rockefeller, though, voted aye on cloture on last year’s bill and is probably attainable.”Will Landrieu vote to destroy her state, which is certainly one of the five most vulnerable to climate change? Also, the refineries have been whining that they haven’t gotten their fair share of allowances. So I can certainly see a reasonable deal, if she wants a face-saving way to do the right thing. Still, a tough get.

I think Martinez and McCain are way too low. As I said in Part 2, joining the filibuster of the climate bill would certainly make McCain the greatest hypocrite in the history of the Senate. That doesn’t mean he will vote for the bill or cloture — he probably won’t — just that a “no vote” has got to be much tougher than 5% for him. Silver notes:

Martinez voted ‘yea’ on cloture last year and Florida is a low-carbon state which might suffer significantly from a sea-level rise or an increase in Atlantic Hurricanes, but the fact that he’s retiring may actually harm the Democrats, since I’d gather that cap-and-trade is reasonably popular down in Florida (which passed a statewide permutation on the policy last year).

Finally, in Silver’s list of the GOP Hail Mary’s and No-shots is one Lisa Murkowski of Alaska, listed as only a 2.37% probability of yes. That is at least 25 percentage points to low, I think. She understands the global warming is destroying her state. She is likely to vote against the bill, but I don’t think it is a 97% sure thing.

Many other analysts have put forward slightly different lists. Since we have at least three months before the final Senate vote, future posts will look at some of those other lists — and at individual swing senators and states.