On Tuesday the Senate voted on an amendment sponsored by Tom Coburn of Oklahoma—political science’s favorite senator—that would have placed a temporary moratorium on the practice known as earmarking. Exemplifying his disdain for the practice, Coburn in the past referred to earmarks as the “gateway drug” to Washington’s addictions. Predictably, the amendment failed 56-39, falling short of the 50 votes needed for passage and well short of the 60 votes needed to end debate (the vote was technically a procedural one). Of the roll-call, David Rogers of Politico noted the “impact of the 2010 elections—as well as 2012 on the horizon—was evident.” Perhaps this is true. But this logic, much like the remainder of Rogers’ article, focuses narrowly on horse race politics. Is knowing the salience or proximity of the next election the only thing needed to explain Congressional action? Fortunately political scientists are not (usually) as myopic in their view. Indeed, there are other plausible (if not better) reasons why a senator would vote for or against the earmark ban; reasons that have nothing to do with “wave elections.” This begs the larger question which I want to address: Why would a senator vote to ban earmarking? Was Tuesday’s roll-call just an electoral signal provided by 2010 coupled with fears of 2012, or are there other dynamics at work?
First, the bad news for 2010-as-electoral-signal-to-2012 meme: by my count only 34.5% of senators who voted for the ban are up for election in 2012 compared to 65.5% who voted against the ban. Further, and perhaps even more damming, there are a number of vulnerable Democrats who voted “no” on Coburn’s amendment (Ben Nelson, Debbie Stabenow and Bob Casey Jr., among others). Of course aggregate statistics can’t tell the whole story, but this fact suggests at the minimum that other factors are at work; at the maximum it suggests the electoral signal is only a minor factor.
So what other factors are there? They are quite simple, actually. Jon Kyl gives us one plausible alternative, stating that his support for the ban is partly “an expression of policy.” Though earmarks account for such a small fraction of the federal budget (a fraction of 1%), deficit reduction is, after all, an ideological issue. Thus, it may be that voting on the earmark ban is explained simply by a senator’s policy preferences. Second, and I think most importantly, we know from empirical work in political science that senators from geographically small states wield disproportionate power in the Senate in their ability to distribute federal dollars (See Lee and Oppenheimer’s Sizing up the Senate). Much like a seat on the Appropriations Committee (also a plausible alternative), it may be that the decision to vote against the earmark ban is a rational one, designed to safeguard a lawmakers' ability to “bring home the bacon.” Thus, electorally vulnerable senators might safeguard their electoral future by doing to opposite of what Rogers suggests (voting against the band). Of course this effect is complicated by another factor—senators can also distribute money to their constituents through block grants. Unlike representatives, who must specify spending for their district if they want to claim credit, senators can distribute money directly to their state, allowing local lawmakers to determine how those funds are spent. Still, it may be that senators from small states are less likely to abdicate a procedure that gives them a disproportion share of power in the upper chamber.
To analyze these competing perspectives I coded data for each senator’s ideology (using Poole’s common space NOMINATE scores), their state’s raw population, whether they are currently serving on Appropriations and whether they are up for reelection in 2012. The dependent variable is their vote on the Coburn Amendment (No. 4697), coded 1 if the senator voted in favor of the earmark ban and 0 if against. Here are those results:
What we see quietly clearly is that a senator’s ideology, their state’s population size and whether they currently serve on Appropriations are all significant determinants of how they voted on the Coburn amendment. The results show that ideological conservatives are more likely to support the ban while senators from small states and who sit on appropriations are less likely to support the ban. However, I find that being up for reelection in 2012 does not have a meaningful effect in this model (in fact the effect is negative). Thus, the ability of a senator to distribute goods to her state trumps (at least in this case) the presumed “wave” or “electoral signal.” As Lisa Murkowski put it in explaining her vote “I’m going to find the best path forward to make sure my state’s interests are represented." Murkowski, of course, is a small-state senator who happens to sit on Appropriations. And even further to the point, she voted against the ban on a similar amendment only months before her recent election fight from conservative Joe Miller. If anyone were to vote for the ban, at least if it’s all about the horse race, it would have been Murkowski.
As an important aside, I tried multiple specifications of the above model with little variation in the overall results. For example, I substituted party and ideology (they are highly collinear) and got the exact same findings. I also added the results for each state from the 2010 midterm and calculated an interaction term between the reelection variable and the two-party vote share. Though we can only model 66 senators in this specification, there is no evidence that senators up for reelection in republican leaning states were more likely as a group to vote for the ban.
Now it’s not that the electoral connection doesn’t matter here; it certainly does in some cases. For example Olympia Snowe, who is facing a challenge from her right in 2012, switched her earlier vote and cast her ballot to ban earmarks. And I don’t mean to be hard on Rogers, Politico or traditional media outlets. Still, this fairly simple analysis took about an hour of data collection time; plus it isn't that hard to consider a range of alternative explanations. But what I do think is that the media’s constant focus on horse race politics overlooks sometimes obvious explanations for congressional behavior. A roll-call vote that is driven by obvious factors such as ideology and/or partisanship and a system of distributing federal dollars that confers disproportionate electoral advantage to some is suddenly framed exclusively against the backdrop of the next election. Fortunately political science has something to offer….
Extra Point: For an excellent discussion of ways Congress can circumvent (potential) curbs on earmarks see the Capitol Confidential blog.