Earlier I took issue with blog posts by Ezra Klein and political scientist Michael Miller claiming supposed benefits from so-called “Clean Elections” laws like Arizona’s, which will be considered by the U.S. Supreme Court on Monday.  But perhaps a more nefarious tactic of defenders of such schemes is to deny that they have any effect on speech at all.

 

In a nutshell, Clean Elections’ defenders want us to believe that incentives don’t matter.  Economists would certainly be surprised to learn that.  And the testimony of candidates and independent groups in the case suggest otherwise, as do public funding backers themselves when they argue that “matching funds” are essential to encouraging candidates to sign up for taxpayer funding in the first place.  The claim also defies common sense:  If the government sends a buck to your opponent every time you spend a buck, this will probably have some effect on your spending and, therefore, speaking.

 

 

 

Nonetheless, Miller points to a brief in the Arizona case by Costas Panagopoulos and other political scientists purporting to show that people do not respond to incentives.  First, the brief essentially repeats and extends claims made by others in the case that campaign finance data show no “clustering” around the “trigger” amount set in Arizona law.

 

In plain English, the argument goes that if matching funds discourage speech, you’d expect privately funded candidates facing taxpayer-funded opponents to spend as much as they could without triggering matching funds to the opponents.  This would show up in campaign finance data as a “cluster” of candidates spending just shy of the trigger amount, but Panagopoulos’ data purport to reveal no such effect.

 

This claim is intuitively appealing, but overly simplistic.  There are so many other factors that affect candidate spending decisions, it would actually be somewhat surprising if you saw this kind of clustering.  Matching funds matter, but they are not the only thing that matters.  So do, for example, the competitiveness of a race, the quality of candidates, and the cost of campaigning in a district.

 

Some candidates may not be able to raise up to the trigger amount, some may not need to either because campaigning is cheap in their district or they are shoo-ins.  Some may be holding back in case there are independent expenditures in support of their campaign, which also count toward the trigger.

 

Based on the brief and its appendix alone, it does not appear that Panagopoulos et al. made any attempt to control for such factors.  And their second claim suffers from a similar problem.  The Panagopoulos team looks to see whether private-candidate spending in Arizona went up after the Supreme Court issued an injunction halting matching funds, and finds that it did not.

 

But they look only at total spending and averages across all privately funded candidates, basically lumping all races together and ignoring important distinctions.  You wouldn’t expect, for example, that candidates in non-competitive races would suddenly spend more when matching funds disappear.  If they can win with less, they probably will.  (Or, if they’re sure losers, they probably can’t raise any more regardless of matching funds.) By contrast, candidates in hotly contested races might raise and spend more freed from the threat of matching funds.  The Panagopoulos analysis can’t tell us because its average figures wash away such distinctions.

 

A more complete analysis would look at various subgroups of candidates and control for things like candidate competitiveness.  Thankfully, a more complete and relevant analysis has already been done by political scientist David Primo, an expert in the case.

 

Primo looked to see whether matching funds have an effect on candidate behavior in other ways that just the total amount of money they spend.  He also focused his examination on privately funded candidates most likely to trigger matching funds—these are the folks who may be deterred from spending that additional buck.

 

He found that candidates have figured out a way to avoid matching funds other than just shutting up.  They can delay.  Save up your speech until the very end of the campaign, and any resulting matching funds will be delivered too late to do your opponent any good.  Primo shows that late fundraising and spending of candidates facing the threat of matching funds far outstrips that of other privately funded candidates not facing the threat of matching funds.  Moreover, in a follow-up analysis, he finds that the effect of matching funds is most pronounced in very competitive races.

 

Candidates, political consultants and independent groups say the same thing.  Miller himself wrote an article showing this strategy is now commonplace in Arizona.  A GAO report on Clean Elections programs backs this up.

 

Miller and other Clean Elections backers characterize this effect as mere “gaming” or “strategic choices,” unworthy of First Amendment scrutiny.  But just because candidates have found a way to blunt matching funds’ negative impact, does not mean it is any less of a drag on their speech.  Quite the opposite.  How “free” is your speech if a large part of your campaign strategy is to delay speaking until later than you otherwise would to avoid giving your opponent more cash?

 

It turns out that incentives do matter, and candidates campaign—speak—differently as a result of matching funds.  By delaying, candidates have less time to make an impact and voters have less time to consider the message.  This is not the unfettered exercise of free speech the Framers of the First Amendment had in mind.