Did the Bipartisan Campaign Reform Act “Bend the Cost Curve” on Campaign Spending?

Apparently, it did!

On Thursday, I produced a graph and some older papers in economics that made the case that there is a pretty clear trend in campaign spending that was completely unaffected by the 2002 BCRA. However, I’m a philosopher, not an econometrician, so I left off the most important part: comparing growth in campaign spending to growth in inflation and GDP. The numbers I used were absolute totals, and there didn’t seem to have been any effect from the BCRA or Citizens United. Today I sat down to expand on the earlier point, and produced the following chart:

This graph tells a different story than the last one: we can see a clear “bend” after 2002, and another after 2008. Thus, it’s plausible to suppose that BCRA did bend the cost curve: we spent less of our GDP on elections while its important provisions were in effect.

My point yesterday was to offer some predictions and beliefs about the effects of BCRA that militated in favor of overturning it. I maintain that norms are engaged in a reflective equilibrium with our beliefs about the facts of the matter, and that sometimes inaccurate predictions can masquerade as principles. I still think this is true, and I’d only expand that claim: visual representations of facts can and do make arguments. Thursday I made a bad graph, and thus a bad argument. Today I retract it.

These new facts must still be interpreted in light of principles: perhaps the effect size is too small to justify the criminalization of partisan political speech. Perhaps the Rent Seeking Model still applies, and politicians were able to extract fewer rents from businesses during the reign of BCRA. (Perhaps, too, we owe it to shareholders to protect them from the unwanted expenditures of the companies they own.) For now I just want to point out that this does give us evidence against the “no effect” hypothesis.

Democratic Facts and Norms: Testable Hypotheses about Citizens United

So I’ve just completed grading 55 papers on Citizens United v FEC, and though I’d kind of like to reflect on it a bit, I’m also finding that grading has totally exhausted my interest in the legal questions. (But seriously: the personhood question is a red herring!) Maybe later this week I’ll post the best arguments I culled from the lot. Instead, I’ve been thinking about some of the contested facts that ground our judgments about Citizens United. After reading so much about it, it’s become clear that there are frequently empirically-testable claims underwriting judgments about that case, and that these claims can be converted into predictions that are going to be verified or falsified over the next few years. These aren’t values questions, themselves, but frequently I’ve found that my judgments on the legal questions are in part generated by my differing beliefs about the particular likelihoods behind these predictions. Consider the following hypotheses as predictions that can be tested.

The Supreme Court’s decision in Citizens United will…

  • …massively increase the amount of spending on elections (Massive Increase Hypothesis)
  • …have no effect on the levels of campaign spending  on elections (No Effect Hypothesis)

What reasons do we have to believe that “massive increase” predictions are more likely than “no effect” predictions? 2010 was a more expensive election than 2006, but that doesn’t necessarily prove that the decision caused the increase in spending.  2006 was more expensive than 2002, too. Here’s how the total spending trends look (in billions):

The law overturned by Citizens United, the Bipartisan Campaign Reform Act, was passed in 2002 but the limits didn’t effect the 2002 elections. A small piece of it was overturned in 2003 in Wisconsin Right to Life v. the FECCitizens United was decided by in January 2010, in time to effect the midterms. Here’s the data… 1998: $1.61 billion, 2000: $3.1 billion, 2002: At least $2.376 billion (source), 2004: $4.14 billion, 2006: $2.85 billion, 2008: $5.3 billion , 2010: $3.7 billion (source), 2012 prediction: $7 billion (source) These numbers are not inflation-adjusted.

Back in 1973, Gordon Tullock argued that the value of campaign contributions did not reflect the potential benefits of regulatory capture. In a recent paper, “Why is there so little money in US politics?” Ansolabehere, Figueiredo, and Snyder argue that campaign spending has tracked GDP for more than a century. So increases in campaign spending over the last decade, both before and after Citizens United, may simply track the economy as a whole.

If that trend continues, then we might want to accept an alternative model:

  • Campaign contributions are rents extracted from businesses by politicians, especially incumbents and front-runners. (Rent Seeking Model)

Why should we believe this? Well, the traditional liberal model of political capture assumes that politicians are primarily motivated by the money they can receive from businesses for beneficial regulation. The problem is that on many regulatory questions, there are competing economic interests, and politicians can only preserve the benefits of “selling out” so long as they keep their seats. Thus, it is much more likely that a politician will choose positions amenable to her constituents, especially if there are campaign contributions available for proponents on both sides. (More complicated “lean towards the green” models allow politicians to take up political causes on which their constituents are indifferent, however: for instance, interest groups spent millions of dollars trying to sway votes on debit card swipe fees at the beginning of 2011, while most voters worried about other matters.) What do businesses get for their campaign contributions? Not much it seems:

Overall, PAC contributions show relatively few effects on voting behavior. In three out of four instances, campaign contributions had no statistically significant effects on legislation or had the “wrong” sign—suggesting that more contributions lead to less support. (Ansolabehere 2003, pg. 114)

Even worse, money spent on elections has been shown to be tremendously inefficient: several different analyses of House races have shown that there is less than 1% effect on the vote per $100,000 spent. (Here’s one from Stephen Levitt, which also argues against public financing.) These findings are consistent with normative judgments like the concern that “money buys access” or that these trends undermine “democratic integrity” by their sheer size.

  • …reduce the transparency of campaign spending.
  • …have no effect on the transparency of campaign spending.
  • …increase the transparency of campaign spending.

While it seems tremendously unlikely that anything that the Supreme Court has done will lead to a decrease in domestic investment, legalizing political spending seems like it might make spending more transparent if a lot of previous spending was illegal and thus secretive. The new information on Super PACs is going to be an interesting test of this question.

  • shift campaign spending from direct candidate donations to indirect issue and advocacy advertisements.

Whether or not campaign contributions continue to hug the ratio-to-GDP trend-line, the most likely effect of Citizens United will be to shift funding away from the direct control of candidates advocating for themselves. I think we already saw this in 2010, and the alternative seems less likely.

I’d also like to propose some more general hypotheses for testing:

  • Does increased campaign spending increase interest in political campaigns?
  • Does increased campaign spending overemphasize a citizens’ role as voter, and in so doing crowd out other kinds of citizenship?
  • Does increased campaign spending exacerbate partisanship or increase negative advertising?