Colbert Busch vs. Sanford: Who Will They Represent?

Elizabeth Colbert Busch and Mark Sanford met last night for the first—and presumably only—televised debate for South Carolina’s vacant 1^st congressional seat.  There were a few notable exchanges.  For example, Colbert Busch hit Sanford on his extramarital affair, and after the moderate noted “she just went there,” Sanford replied wryly he “didn’t hear her.”  But Sanford had perhaps the best line of the night, turning a question about his vote to impeach Bill Clinton on its head, asking Democrats in the room whether they think Clinton should have been “condemned” for his affair.
But if we strip away the prepared comebacks and terse exchanges, one issue took center stage: representation.  Sanford repeatedly linked Colbert Bush to Nancy Pelosi, national Democrats, and worst, unions.  Attacking Colbert Bush for taking labor donations, Sanford asked: “Whose voice you will carry?”  Sanford has hit Colbert Busch in recent ads for accepting nearly $30,000 from “big labor.”

Elizabeth Colbert Busch, the aggressor for most of the evening, fired back against Sanford’s questions, stating numerous times that “nobody tells me what to do.”  I thought this was a successful response on her part.  Consistent with her position as a “centrist,” Colbert Bush positioned herself throughout the debate as a moderate voice.  So who would she represent in Washington?  The answer she gave was simple: Voters in the 1^st district.

Thus, in last night’s debate we had two competing narratives about how representation functions in Congress.  Sanford claims that moneyed interests—in this case, unions—hold sway over lawmakers.  Colbert Bush argued the rater simple point that lawmakers are beholden to their constituents.  Who’s right?  Here are two charts.

The first chart (top) is a lawmaker’s estimated ideology (derived from their roll-call record) on the y-axis.  Liberals toward the bottom, conservatives on top.  The x-axis is the (logged) amount of contributions the lawmaker accepted for labor groups in the 2010 cycle.  The former data come from Keith Poole at VoteView while latter is available at OpenSecrets.  All data is for members of the 112^th Congress.  The second figure (bottom) is a lawmaker’s estimated ideology, again on the y-axis, and the percentage vote for Obama in their district in 2008.  If lawmakers represent their constituents—not outside interests—we would expect liberals to represent districts with a higher percentage of Obama voters and conservatives to represent districts with a lower percentage of Obama voters.

In the two figures the relationship appears to be as both Sanford and Colbert Bush claimed.  Liberals take more money for labor groups and reside in districts with more Democrats and vice versa for conservatives.  But there’s a problem with inferring causality from the first figure.  An old adage in the social sciences is that “correlation does not equal causation.”  In this context, we can’t tell from the first figure whether labor unions donate to candidates who become liberal as a consequence, or whether unions donate to candidates who are already liberal.  It’s a point we addressed earlier with respect to the NRA.  Simply put, there’s not a lot of evidence that interest groups like the NRA “buy votes” in Congress (as satisfying a conclusion as that might be).  Interest groups have other effects to be sure—for example, lending “policy expertise” or by directing lawmakers’ attention to an issue—but there’s just not a lot of evidence that they change minds in Congress (for a thorough review of the academic literature on how interest groups affect lawmaking, see this post).  In short, a number of researchers think causality goes in the opposite direction that Sanford is claiming.

But let’s assume for a moment that Sanford’s claim is indeed true: interest groups, like unions, can actively “buy votes” in Congress.  Under this assumption, we can run a multiple regression comparing the effect of both factors—labor contribution and the district’s partisanship.  The estimated model is to the above (click for bigger image).  From the results we can see that both factors are statistically significant and correctly signed.  Moreover, the model does well, explaining 73% of the variation in a lawmaker’s roll-call record.

Ok.  Both factors correlate with a lawmaker’s ideology.  But what does this mean for the candidates running in 1^st district?

First, let’s take into consideration Sanford’s figure that Colbert Bush has accepted $30,000 from “big labor” (see the first image, a screenshot from a recent Sanford ad).  Second, we also know that Obama only received 40% of the vote in the 1^st district.  We can use two figures to simulate Colbert Bush’s ideology if she wins election.

According to the regression model and the simulation, Colbert Bush’s ideology in the 113^th Congress is estimated to be 0.58.  That’s comfortably on the conservative end of the spectrum.  We can see in the first scatterplot that Colbert Bush would be more closely aligned with moderate Republicans (the top cluster).  In fact, she would rank as more conservative than about 1/3^rd of all Republicans.  One important caveat: I doubt very seriously this represents her actual ideology if she in fact wins.  The point, rather, is to compare the relative effects of her district versus union donations.

How does this make sense?

First, political scientists who study this issue know that lawmakers do a fairly decent job representing their constituents.  That’s an unexciting conclusion for sure, but lawmakers care deeply about winning reelection.  This effect is also stronger in the House, as one would expect.  The simple—but nonetheless important—point is that the 1^st district is a Republican leaning district.  Mitt Romney carried the 1^st by a whopping 18 points.  Last night we saw Colbert Bush stake her position as a moderate, claiming, for example, that “Obamacare” is “deeply problematic.”  So if the past is any predictor, and I think it is, the district’s partisan leanings will keep Colbert Bush away from Nancy Pelosi and “big labor.”

But second, even if we assume for the sake of argument that interests groups can “buy” votes in Congress, the effect is not large in magnitude.  For example, based on the regression model, the simulated difference between taking $0 in campaign contributions from unions (Mark Sanford) and $30,000 (Elizabeth Colbert Bush) is estimated to move a lawmaker in the liberal direction by 0.12.  In the 112^th Congress, that’s the difference between the voting record of South Carolinians Mick Mulvaney and former 1st district representative Tim Scott.  I would challenge any Sanford supporter to articulate the policy differences between these two individuals (both members of the Tea Party Caucus).