Clean Elections vs. Political Speech

One of the goals of The Monkey Cage is to publicize political science research. We wish to do so for many reasons, but obviously one of the most important is to try to catch the eye of policy makers who can (hopefully!) improve public policy as a result. However, there is another way in which political science already directly contributes to public policy, which is when the research of political scientists is used to help resolve court cases. With that in mind, we welcome the following guest post from Professor Michael Miller of the University of Illinois, Springfield. Due to the length of the post, it can also be downloaded as a .pdf file here.


Last year one of the most-analyzed Supreme Court cases was a First Amendment challenge to campaign finance law in Citizens United v. F.E.C. Next week, in McComish v. Bennett, the Court will hear arguments in another campaign finance case, and this one has similar potential to alter the role of money in American politics. The case will essentially decide the Constitutionality of state public election funding programs that according to the vast majority of academic studies have made elections more competitive and voter-centric. While a decision on a relatively obscure campaign finance law may not appear to be particularly important, this case is going to be a big deal for any state or municipality seeking ways to reform its elections. Everyone with an interest in how money affects elections should be watching closely.

Here’s the background in a nutshell: Since 2000, Arizona and Maine began new programs (dubbed “Clean Elections” by supporters) that offer state candidates public election subsidies covering the entire cost of a campaign. Connecticut started a very similar program in 2008, and at least three other states offer full funding in some elections. In exchange for these grants, which they qualify for by raising a not-insignificant number of $5 contributions from citizens, candidates agree to raise no additional money from private sources and to limit their spending to the amount of the grant. So if John Smith from Tucson decides that he wants to run for the Arizona House, so long as he can convince 220 people to contribute exactly five dollars, the state will give him more than $30,000, or enough to run a well-funded general election campaign for Arizona state office with no donor money. In short, public funding is supposed to make money—and donors—less important.

The issue in McComish turns on matching funds provisions that guarantee financial parity (up to a point) for candidates who take public money. Participation in Clean Elections is optional, meaning candidates may opt out to raise and spend their own private money. Yet, when these “traditional” candidates spend large sums against publicly funded opponents, the latter receive extra grants matching traditional candidate spending dollar-for-dollar. So if John Smith’s opponent raises $20,000 more than Smith’s original grant, the state will cut Smith an additional check in that amount. This provision is designed to guarantee that candidates who participate in public funding will compete at financial parity, even with extremely rich opponents.

Enter strategy. When Clean Elections was enacted in Arizona, it wasn’t long before traditional candidates figured out that every dollar they raised above the spending limit, in the words of one such candidate I interviewed for a 2008 article (gated), “feeds the alligator trying to eat (them).” Thus, while traditional candidates may want to raise money for ads and events, they recognize that doing so will essentially amount to a campaign contribution for their opponent. Since they have an incentive to restrict their spending during the campaign, traditional candidates argue that Clean Elections chills their political speech.

Yet efforts of the petitioners (who cite my work) in McComish to depict Clean Elections as a nefarious brake on speech present only half of the story. There is no existing evidence that non-participating candidates raise less money in Clean Elections systems. In an excellent amicus brief, Costas Panagopoulos, Ryan Enos, Conor Dowling, and Anthony Fowler report no evidence of reduced spending by non-participating candidates. Furthermore, in the very piece that the petitioners cite to justify their case, I show that rather than halting their spending, traditional candidates have merely adjusted their financial activity to game the system; they still raise and spend money, but at the very last minute—too late for their opponents to receive matching funds. In 2006, for instance, 40% of all matching funds allocations in Arizona came on the last day of the election. Traditional candidates face no legal restriction on their fundraising, instead strategically choosing to delay spending. If they bombard their districts with campaign materials in the election’s closing days, traditional candidates should argue more accurately that matching funds affect not what they say, but when they say it.

Moreover, political scientists have consistently shown that Clean Elections alters the way in which campaigns are waged. The most studied effect is in the area of electoral competition. Neil Malhotra (gated) found evidence that when challengers accept Clean Elections subsidies, elections are closer, which is consistent with earlier evaluations. And while evidence in state senate elections is less clear, the figures below show downward trends in instances of unopposed house incumbents in both states and in the Maine Senate through 2006, which is the last year for which I presently have data.


But Clean Elections also has broader, less observable effects. In my doctoral dissertation, I collected surveys from state house candidates in 18 states during the 2008 election, and did field interviews with Arizona politicians following the 2006 campaign. I asked them about how they spent their time, why they chose the funding plan that they did, and how they felt about their campaigns. The figure below shows the percentage of time that candidates in each state devoted to various activities.


“Field” work—knocking on doors, placing signs, and related activities—comprises by far the greatest proportion of candidates’ campaign efforts. But when I compared the behavior of candidates who accepted Clean Elections subsidies to candidates who raise money from private sources only (after doing a bit of matching), I found that because they spend no time raising money, Clean Elections candidates devote significantly more effort—about 10 percentage points of their overall time—to direct engagement with voters and groups compared to traditional candidates. Moreover, as shown in the chart below, the effect of public money on campaign time is present only for candidates who take the full funding (with matching funds) challenged in McComish, and not smaller partial subsidies like those available in Minnesota, Wisconsin, and Hawaii:


When this effect is projected over dozens of campaign weeks, it is safe to conclude that Clean Elections candidates are making hundreds of voter contacts that would not have otherwise occurred. In the same paper, I find reasonably strong evidence that when a publicly funded candidate is present in a given legislative district, increased interaction with citizens leads to fewer people “rolling off” (abstaining from voting) in state house elections. Finally, while we cannot say for certain whether public funding has diminished the influence of large organized interests in government, if participating candidates focus on voters because they can ignore donors, it seems safe to conclude that Clean Elections probably has done no harm.

No campaign finance system is perfect. However, while the petitioner’s argument in McComish rests on the narrative of a handful of candidates, almost all of the available empirical evidence refutes the story of “chilled speech” in Clean Elections systems. To the contrary, political scientists have consistently found that Clean Elections leads to a different, voter-focused sort of campaign, and it has historically enjoyed the support of a substantial majority of citizens where it exists. Clean Elections cannot eliminate all “problems” in state elections, and there is still much that we do not know for certain about its effects. But it is hard to argue that the effects we have observed—greater interaction between candidates and voters, more competition, and reduced risk of corruption—are anything but improvements over the state of elections in much of America.

The McComish case presents an excellent opportunity for political science to inform the judiciary. I hope that the justices will reach their decision based upon the evidence that we have provided them, rather than the anecdotal claims of a few candidates.

10 Responses to Clean Elections vs. Political Speech

  1. anonymous coward March 23, 2011 at 12:56 pm #

    But it is hard to argue that the effects we have observed–greater interaction between candidates and voters, more competition, and reduced risk of corruption–are anything but improvements over the state of elections in much of America.

    Yes, but.

    Another (partial) outcome of the election law is the current crop of Arizona legislators. And, frankly, I do not think you would find a more wretched hive of scum and villainy this side of Mos Eisley.

    I’ve seen at least informal arguments that public financing has played a role in the increasing dysfunction of Arizona government as it comes under the control of far-right nutjobs more intent on stunts than on governing. The argument is that public financing allows people whose positions are too extreme to fund-raise effectively, but who are charismatic, to gain the funding they need to get elected.

    Of course, that doesn’t mean that the “chilled speech” argument holds any water. But I don’t think anyone should hold up Arizona as anything but a cautionary example.

  2. Jack Lechelt March 23, 2011 at 11:23 pm #

    This is a most interesting post. And yet the more important point is whether or not clean elections lead to “better” government, however one defines better.

  3. Michael G. Miller March 24, 2011 at 12:06 am #

    To anonymous coward:
    I’m really glad you raised that issue, because I think it is worth talking about. The best I can do is say that with the help of a coauthor who is much smarter than me, I have recently begun sketching out an empirical exploration of whether there is anything to that argument. So stay tuned, is what I guess I’m saying.

  4. beowulf March 24, 2011 at 1:17 am #

    “Another (partial) outcome of the election law is the current crop of Arizona legislators. And, frankly, I do not think you would find a more wretched hive of scum and villainy this side of Mos Eisley.”

    The trouble with an electoral reform that produces public officials who accurately reflect the values and beliefs of the electorate is that, unfortunately, it produces public officials who accurately reflect the values and beliefs of the electorate.

  5. Eric Friedman March 24, 2011 at 10:08 am #

    You note a comparison with “partial subsidy” programs like those in Wisconsin, Minnesota, or Hawaii. Have you looked at programs designed to match small contributions, like New York City’s program? To a certain extent, I believe that candidates engaging in low-dollar fundraising are accomplishing two objectives at once: collecting resources and engaging voters. I think your results are interesting, but to the extent there is overlap among the various activities, and that they seem to be self-reported by candidates after the fact (please correct me if I have the wrong impression) could there be a bit of flex there?

    All in all, it is an interesting way to get at a really vexing issue — how and what do you measure to evaluate the impacts of public campaign financing? — and I commend you for it.

    As a final note: though our program works differently, I would also say that here in NYC our results are similar in one important way: we have seen no evidence whatsoever that bonuses for candidates facing high-spending opposition have deterred non-participants’ spending.


  6. paul March 24, 2011 at 11:34 am #

    The speech-chilling effect on small donors was certainly not considered in the CU case. Does that mean it will be controlling here?

  7. dlw March 24, 2011 at 3:23 pm #

    I wonder if folks might consider an approach that combined both progressive taxation of $peech with public funding such that it was revenue neutral?


  8. Michael G. Miller March 24, 2011 at 7:24 pm #

    I have not personally done anything with NYC or any other municipalities that offer any form of public funding (Portland and Albuquerque spring to mind). Jeff Kraus had a 2006 paper that was pretty pessimistic on NCY, though; it’s worth a read if you’re interested.

    Yes the activities are self-reported, but the survey was (mostly) conducted during the campaign. I think the approach is certainly open to critique, but I’d say that if we’re going to study campaign activities in a large N quant. study, this is really the only practical way I can think of. Plus, the candidates in the publicly funded states were not clued in to the purpose of the study, and if the numbers are flawed by self-reporting, it’s not clear that the disturbance would exist in a systematic way (but it’s also certainly not impossible). So I guess what I’m saying is I got at the issue–altered campaign strategy/activity–the best way I could. As to categories, the graph on this page depicts an additive index of publicly oriented campaign activities, which would mitigate that problem if you believe that campaign time overall is accurately reported. Of course, we can still debate the categories that go into that index.

  9. Todd Lang March 25, 2011 at 12:41 pm #

    I look forward to your study on whether clean elections produces a different “class” of elected officials.
    I think not for two main reasons: 1. Arizona is no different than the rest of our country in regard to “Tea-Party” takeovers or the success of extreme conservatives. If Clean Elections was the cause, then Arizona would be an isolated case. 2. Arizona does not have competitive districts. I know academics love to debate this, but right now there is no incentive for primary voters to “moderate” their choice –as the primary winner is the de facto general election winner.

    Regardless, those who complain that clean elections allows “bad candidates” to win are missing the point about what the law does. It fulfills our democratic ideals by allowing more folks to run and giving the voters more choices. Critics are simply unhappy with the choices made by voters. Instead of educating voters to their point of view, these critics would simply like to take the choice away from voters and return to the “smoke filled rooms” and party insiders of yesteryear. To me, letting the voters make a bad decision is better than giving them no choice at all.

  10. Eric Friedman March 28, 2011 at 12:40 pm #

    Thanks, Michael. I would just caution that “partial” programs are not all the same. Wisconsin and Hawaii, for instance, provide maybe 5 percent or less of the money candidates have to spend. With a matching rate of 6:1 for small contributions (up to $175), our program is generous enough to provide most participating candidates with half of their funding or even more in many cases. Prof. Kraus’ paper notwithstanding, NYC’s program has brought a many of the benefits you’ve found in Clean Elections states. NYC has done a lot of the same things Todd’s program in AZ has — prospective candidates don’t need big money or county bosses on their side to put their message before the voters and compete for office.