By Richard Jong-A-Pin

On March 17th, we have parliamentary elections in the Netherlands. And therefore, I’d like to share with you the joke of that economics professor who went out to vote...(S)he didn’t go...

Well, that is, if you believe the rational choice voting theory of Anthony Downs (Downs, 1957). According to Downs, the professor realizes that the benefit of voting does not outweigh the costs of travelling to the voting station, since her/his vote is most likely not going to be decisive for the election outcome.

There is a problem though. According to this theory nobody would vote. Yet, in reality we do observe massive turnout in large democratic elections. Since the theory is at odds with what is observed in reality, it is better known as the “Downsian Paradox” or “the paradox that ate rational choice theory.”

Throughout the years, a number of theories have been proposed to overcome the Downsian Paradox. These theories have widened the range of determinants in the cost-benefit analysis by including phenomena like warm-glow, altruism, peer-pressure, and expressive behavior. In particular, expressive behavior is a determinant that is widely studied and together with Rasmus Wiese, I also added my 2 cents (Wiese and Jong-A-Pin, 2017).

Even though it makes a lot of sense to include these non-monetary factors into a cost-benefit analysis, a dissatisfying feature of these theories is that they impose the solution to the paradox in an ad-hoc way. Naturally, there also have been studies that tried to overcome this ad hocery. These studies use game theory and are very successful in explaining voter turnout for small elections. However, they still perform poorly in explaining voter turnout for large elections.

In a recent research article published in the Journal of Economic Behavior and Organization, Oliver Herrmann, Bert Schoonbeek and I combine the aforementioned strands of research and present a prospect-theory model of voter turnout (Herrmann et al., 2019). On the one hand, we rely on a game-theoretic framework and, on the other hand, we depart from the standard rational choice approach by including prospect-theory preferences that replace standard expected- utility preferences. Central to prospect theory is that individuals evaluate perceived losses in a different way than perceived gains, and whether outcomes are perceived as gains or losses depends on a subjective reference point. Furthermore, individuals assign subjective (biased) weights to objective probabilities using non-standard weighting functions.

Our behavioral model provides new insight in the voting decision individuals make. We find that the reference point of an individual has a large impact on the propensity to vote. That is, individuals with a reference point corresponding to the act of voting have a much higher likelihood to vote than those whose reference point is based on the act of abstaining. This implies that if you believe that voting is your civic duty, then you are more likely to vote in comparison to people that lack civil duty. Even if you are (most likely) not going to influence the election outcome!

Another interesting finding is that non-standard probability weighting (as opposed to the assumption of voters using objective probabilities) leads turnout to decrease at a slower pace and leads to higher turnout predictions in larger electorates. This is because individuals tend to overvalue small probabilities. Consequently, the (small) chance of being pivotal in large elections is overweighted, turnout is increased, and the Downsian result is way less strong.

Now you may wonder what the conclusion is of this blog in which I briefly presented our prospect-theory model of voter turnout.

Here it is: Go Vote! It may not be rational in the traditional sense, but it is very human to do so!