Yesterday Ezra Klein made the claim that "if you had run an algorithm using past voting records to predict last night's roll call [on the budget], you wouldn't have ended with Nelson and Bayh on their lonesome" voting against the bill. Instead, he argues, they did so in order to obtain favorable press for breaking from their party:
The President's budget also passed the Senate last night. Zero Republican votes. Two Democratic defectors. Evan Bayh and Ben Nelson. Interestingly, other senators you might associate with their precise position on the ideological spectrum -- Lincoln and Landrieu and Pryor and Carper -- voted for the budget.
In other words, if you had run an algorithm using past voting records to predict last night's roll call, you wouldn't have ended with Nelson and Bayh on their lonesome. But if you had run an algorithm using the amount of press a given Democratic senator has received for being willing to buck the President, you would have.
However, the Optimal Classification algorithm of UCSD's Keith Poole and UCLA's Jeff Lewis, which Poole uses to order senators in the 110th Congress (2007-2008) based on their roll call votes, ranks Bayh and Nelson as more conservative than any other members of the Senate Democratic caucus, including Lincoln, Landrieu, Pryor, and Carper. If we assume that all new senators from the GOP are to the right of Bayh and Nelson and all new Democratic senators are to their left, then Lewis and Poole's estimated ideal points/rankings can perfectly discriminate yes votes from no votes on the budget (with the cutpoint falling between Bayh and the most moderate Senate Republican, who in the 110th was Olympia Snowe). Given the location of the cutpoint, Klein's statement is false.
Unfortunately, Klein then goes on to engage in mind-reading about Bayh's motivations:
Some wags are noting that none of Bayh's much-heralded Caucus voted with him.
But that's precisely why Bayh voted against -- because no one else was.
This was, in a strange sense, the safe play. Because the budget only requires 50 votes, their opposition didn't seriously imperil the President's budget. If eight more Democrats had signed on, it would have, and there would have been consequences. But the consequences of ineffectual opposition are all positive. Bayh and Nelson have elevated their status as the Democrats willing to imperil the President's priorities. They've assured that the media will say the names "Bayh and Nelson" a lot. They've secured themselves a steady stream of requests to appear on news shows and many calls begging for a quote. They have further cemented their status as power brokers in a closely divided Senate and media stars in a conflict-hungry news environment. It's really a very good day for them.
It's possible that Bayh is grandstanding in order to attract media coverage (or is engaged in some other sort of insincere politically-motivated behavior). It's also possible that he sincerely believed in the explanation he provided for his vote. There's no way to know which is true. Despite his claims to know "precisely why Bayh voted against" the resolution, Klein actually has no idea which of these two explanations is correct.
The speculation about motives continued in a subsequent post suggesting Bayh's support for an estate tax reduction bill was insincere:
WHY DID EVAN BAYH VOTE FOR KYL-LIEBERMAN?
I understand Evan Bayh's decision to vote against the budget. In a Senate with 59 Democrats, the opportunity to emerge as the marquee swing vote is undeniably attractive. It brings with it real power over policy and real celebrity in Washington. And there's even a legitimate argument that Bayh developed in his statement today. He praises the budget for funding "important priorities like affordable health care, energy independence, job creation, and education improvements, rather than tax cuts for the most affluent," but then says that "under this budget, our national debt skyrockets from $11.1 trillion today to an estimated $17 trillion in 2014...I cannot support such results."
Fair enough. But then why vote -- on the very same day -- for the Kyl-Lincoln bill lowering the tax rate on estates over $7 million from 45 percent to 35 percent and reducing charitable giving? That's $250 billion more debt over 10 years. It's in direct conflict with Bayh's statement on the budget. It makes him look insincere.
The obvious answer is that it's important to wealthy contributors in his state. But Bayh doesn't face reelection until 2014. No one will remember -- much less contribute -- based on a vote in 2009. Nor is there a presidential run in his near future. And I simply refuse to believe that Bayh thinks it an important point of principle -- more important than debt reduction or health care -- that extremely wealthy Americans pay 35 percent, rather than 45 percent, on their estates.
Literally the only compelling justification I can come up with is self-interest: Bayh has a net worth estimated between $3 million and $14 million. That's the sort of thing that sensitizes you to the downsides of the estate tax real quick. But even that's not totally satisfying. Calls to Bayh's office went unreturned. And so I wonder.
Correction: Bayh is up in 2010, not 2014 as I mistakenly thought. It's not a particularly contested reelection. He has no serious challengers, a massive advantage in fundraising ($10,000,000 on hand, and that was in 2008, before he'd really started fundraising for reelection), full name recognition, solid approval ratings, and is running in a state that Barack Obama won. So I don't think it's plausible to argue that supporting Kyl-Lincoln is crucial to his reelection. But he might.
Again, we have nothing other than speculation to support Klein's suggestion that Bayh's support for the Kyl-Lieberman bill is insincere. (For what it's worth, I think there's a credible argument that supporting Kyl-Lincoln could helpful to a Democrat running for re-election in Indiana, which is still a conservative state -- remember, the estate tax polls very poorly even at the national level. And while Klein is correct that Bayh's seat seems safe, there's good evidence that members of Congress almost always feel vulnerable to an electoral challenge.)
Update 4/4 7:09 PM: Via email, Klein points me to a later post in which he looks at Bayh's ranking among Senate Democrats when the Lewis-Poole optimal classification algorithm is run separately on the roll call votes from the 107th, 108th, 109th, and 110th Congresses (he says he is using Poole and Rosenthal's rankings but that is not correct):
Ben Nelson genuinely is a conservative Democrat. The Poole-Rosenthal rankings -- which most consider the leading measure of a congressman's relative ideology -- have, for years, ranked Nelson as the most conservative Democrats in the Senate. Call him Democrat #1. Evan Bayh, however, has not traditionally been number two. I went back in the rankings through the 107th Congress -- which began in 2001 -- to compare Bayh and Nelson's ideological consistency. The numbers on the Y axis represent how conservative of a Democrat each senator was. So #1 would be the most conservative Senate Democrat and #5 the fifth most conservative Senate Democrat and so on. The blue line is Bayh. The red line is Nelson.
To say Bayh lacks Nelson's steady hand on the wheel is a bit of an understatement. The two really interesting data points, however, are the 109th Congress, which stretched from 2005 to 2007, and the 110th Congress, which ended in January of this year. In the 109th Congress, Bayh's voting pattern suddenly develops an uncharacteristic liberalism. He becomes the 19th most conservative member, with a record more liberal than, among others, Joe Biden. As context, these were also the years when Bayh was preparing for the presidential run that he eventually aborted.
In the 110th Congress, however, that flash of liberalism gives way to a career-high conservatism: He actually displaces Nelson as the Caucus's most conservative member. He's running for reelection in Indiana this year, but this is also the year that Indiana's tectonic plates shift and the state chooses that Obama guy. So I'm not going to pretend that I fully understand the motivations behind the sharp swings in Bayh's voting record. But they're undeniably present, and seem to be keyed to political campaigns. Bayh is much steadier during the 107th and 108th Congresses, when no elections loom.
I agree with all of this, but it doesn't really justify Klein's original statement. (Note: Poole and Rosenthal's DW-NOMINATE algorithm, which is the most well-known scaling procedure in political science, ranks Bayh as the tenth most conservative Senate Democrat in the 106th and 107th Congresses, ninth most conservative in the 108th, fifth most conservative in the 109th, and fourth in the 110th. However, it only allows for linear change by Congress in estimated ideal points and thus can't handle potential non-linear changes in Bayh's voting record of the sort that Klein describes above.)