What Voting Signals

From Arthur B. at The Distributed Republic:

This paper should be familiar. It was featured by Levitt in the New York Times and has been discussed at length in these circles. It describes what happened when a Swiss Canton allowed mailing ballots in an election. Participation fell dramatically. The article concludes that most people vote because showing up at the poll booth signals you are participating in the election.

I was recently reminded of this paper by an article on Slashdot telling a similar story:

Voting fell 83% in an all digital election

Why should we care? … A low turn out indicates that people do not really care about the election. … Better, it attacks the myth that democracy is representative. … A democracy with a lot of eager participation is a recipe for collectivist arguments about the “will of the people”. Low turnout allows the state to be viewed as a separate parasite.

That NYT article elaborated:

The motivation could be hope for social esteem, benefits from being perceived as a cooperator or just the avoidance of informal sanctions. … The Swiss study suggests that we may be driven to vote less by a financial incentive than a social one. It may be that the most valuable payoff of voting is simply being seen at the polling place by your friends or co-workers.

One is reminded of the Soviet Union bragging about its 100% voter participation rate with only one candidate for each office.   So why are folks so eager to increase voter participation, when it lowers the quality of decisions and doesn’t obviously mean people accept the government?

I’d guess self-deception is key.  People enticed to the polls by signaling incentives would rather not admit such motives; they’d rather believe in their own power and altruism, and in the virtues of their polity.  So encouraging more signaling via votes encourages people to believe that the little guy has influence, and that voters really care about each other.  So induced voters end up believing more that their democracy works.  Even in the Soviet Union.

Added: OK, uncle, the NYT did severely misrepresent the paper, as several comments have mentioned.  This does not offer much support for the signaling theory, though it doesn’t go much against it either.  Let me vow to never again blog an academic paper I haven’t perused.

GD Star Rating
Tagged as: ,
Trackback URL:
  • Cyan

    People enticed to the polls by signaling incentives would rather not admit such motives; they’d rather believe in their own power and altruism, and in the virtues of their polity.

    Is it that they would rather not admit such motives, or that they are consciously unaware of such motives as an adaptation to keep the signals looking genuine?

  • Dan

    “Voting fell 83% in an all digital election”

    OK that IS scary. The problem is that the nuts ALWAYS vote. So for example imagine the world being run by paultards or slashtards, if the election is all digital…
    Reminds me of that VOTE! youtube video with the exhortation “Vote because “they” are sure as hell is voting”.

  • Stephen

    How about an experiment where two votes are taken from the participant pool: the first voluntary, the second mandatory. I wonder whether/how precisely the outcome of the voluntary voting would predict the outcome of the mandatory voting.

    In other words, maybe low voter turnout is a more efficient and adequately accurate mouthpiece for the whole population.

    It is possible that the signaling-induced voters are accidentally making democracy work.

  • Marlon

    “Participation fell dramatically”? How does that jibe with “the effect on aggregate turnout was small” (from the abstract of Funk & Fabra’s study)?

    Also, looking at voter turnout trends in a US state which switched to entirely vote by mail in 1998, Oregon, there doesn’t seem to be a dramatic effect. In fact, the last three presidential elections (which were vote by mail) have seen higher turnouts than the 1996 presidential election, where ballots were cast at polling locations. A quick perusal of other elections doesn’t seem to show a dramatic drop in voter participation post-1997. [http://www.sos.state.or.us/elections/other.info/stelec.htm]

    “Voting as signaling”, if true, seems to be strongly dependent on both community size and regional culture. I doubt its effect is significant.

  • Grrrr

    So induced voters end up believing more that their democracy works. Even in the Soviet Union.

    I don’t think the Soviet Union voting counts here. Given the total deficit of many kinds of rather basic foodstuffs, the voters were enticed to the voting stations simply by a network of kiosks beside the booths selling low-priced salami and the like on the voting day. Worked like a charm.

  • JH

    Do people actually run into many of their friends and family at the poll? I can see it in small towns, maybe, but not larger metropolitan areas. Is there are negative correlation between population and participation? People in smaller, rural areas are the ones who most likely have to answer to mom if they don’t vote.

  • Psychohistorian

    Your data are flawed.

    The digital example does not seem generalizable; they completely eliminated balloting stations, so the issue may have been more people not knowing how to vote / that there was an election. An election official explicitly mentioned that ignorance may have been a cause of low turnout. It’s also possible that there are a lot of older / non-tech-savvy voters in the area. It may be that the election was not advertised as well because there were no real places to vote.

    The Swiss example is just blatantly misrepresented:
    “In a difference-in-difference framework, I study the impact
    of optional postal voting on Cantonal turnout. The effect is small (an increase in 2 percentage
    points), and not statistically different from zero,” from the actual paper. That’s a positive effect, if an insignificant one, not “Participation fell dramatically.” There’s more detail, and there is some support for social pressure, as small communities with inconvenient voting times saw no meaningful increase in turnout. I didn’t see any negative changes exceeding a few percentage points, though I admittedly skimmed the paper. But, at the most, this paper suggests that the demand to vote is relatively inelastic with respect to the immediate inconvenience costs of voting. Also, social pressure appears to play a role, but it is unclear if this is properly described as signaling.

    Methinks you need a lot more (and better) data before you conclude too much from this.

  • Hal Finney

    I just googled “vote by mail participation rate” and this was the first hit:
    Voting by Mail – Political Participation and Institutional Reform.

    Oregon is the first state to exclusively vote by mail. In this new system, all ballots are mailed to the voter’s home, eliminating the polling-booth. To determine voting by mail’s impact on turnout, I analyze changes in Oregon’s voting rates and socioeconomic factors, national and regional voting trends, and other states’ turnout rates. I find that Oregon’s turnout is up by 10.1 percent of registered voters when compared to previous elections. I show that other suggested reasons for this turnout increase are not the cause of Oregon’s change in voting rates.

  • http://hanson.gmu.edu Robin Hanson

    I added to the post.

  • http://entitledtoanopinion.wordpress.com TGGP

    I recall in John Lott’s “Freedomnomics” an analysis of what policies reduced turnout. One of his conclusions was that before the secret ballot large numbers of people were basically being bribed to vote in certain ways but afterward those agreements were impossible to monitor.

  • Constant

    the NYT did severely misrepresent the paper

    Falling standards, or same old NYT?

  • http://www.geoffreyfalk.com/blog/blog.asp Geoffrey Falk

    That’s a positive effect, if an insignificant one, not “Participation fell dramatically”….

    Falling standards, or same old NYT?

    “Standards fell dramatically.” 🙂

  • Pingback: Elections et point focal « Rationalité Limitée

  • Pingback: Voting Signals and GTD « RAD Thoughts