Republic, Lost_ How Money Corrupts Congress--And a Plan to Stop It - Lawrence Lessig [60]
0. It Matters Not at All
Some believe that this dependence upon money does nothing. That it is harmless. Or at least, they insist, we have no good evidence that this dependence does anything, and since we’ve got no evidence, we’ve got no good reason to change it.
By “evidence,” these conservatives (with a small c—they could well be politically liberal; my point is that they’re scientifically conservative) mean numbers. Statistics. Regressions that show an input (campaign contributions) and an output (a change in votes). There is no good evidence, these scholars insist, that campaign contributions are changing political results. There may be many such contributions. Securing them may well occupy a huge chunk of a congressman’s life. But we don’t have the data to support the claim that this money is buying results that otherwise would not have been obtained.28 As Frank Baumgartner and his colleagues summarize the research, there is “no smoking gun, no systematic relationship between campaign contributions and policy success.”29
The most prominent work making this claim is by political scientists Stephen Ansolabehere, John M. de Figueiredo, and James M. Snyder. In an important paper published in 2003, “Why Is There So Little Money in U.S. Politics?,”30 these authors question just about every strand of the commonsense view that money is buying results in Congress.
The most important bit of their argument for our purposes questions whether campaign contributions actually affect legislative decisions. Ansolabehere and his colleagues first collect about forty articles that tried to measure the effect of PAC contributions on congressional voting behavior. Looking across this range of studies, they conclude, “PAC contributions show relatively few effects.” “In three out of four instances, campaign contributions had no statistically significant effects on legislation or had the ‘wrong’ sign….”31
Ansolabehere and his colleagues then identified a number of statistical problems in some of the studies they collected. This led them to perform their own statistical analysis. That analysis used the voting score produced by the U.S. Chamber of Commerce as the dependent variable. They then estimated six models that mirrored the range of their original forty studies and that included campaign contributions among the independent variables.
Their conclusions are not good for the commonsense view (even if they sound promising for the republic). While they did find some evidence that contributions had an effect on voting patterns, that effect was small relative to other factors. Much of that effect, moreover, was eliminated once they controlled for voter preference. And once they controlled for legislator-fixed effects (such as the party of the legislator), they were able to “eliminate the effects of contributions entirely.”32 As they conclude: “Indicators of party, ideology and district preference account for most of the systematic variation in legislators’ roll call voting behavior. Interest group contributions account for at most a small amount of the variation. In fact, after controlling adequately for legislator ideology, these contributions have no detectable effects on legislator behavior.”33
In understanding the significance of this claim, we should first be very careful about what exactly is being argued here. Ansolabehere and his colleagues are themselves careful to insist that they are not saying that contributions have no effect. Indeed, as one version of their paper asserts, “It is still possible that campaign contributions have significant effects on economic policies.”34 How would that happen, given the data they’ve studied?
To raise sufficient funds, candidates might skew policies in ways preferred by donors. Campaign contributions might therefore