Voters are Irrationally Irrational. Or Why We Should not Worry about Gibbard-Satterthwaite
Note: In spite of what this post may suggest, I didn’t vote for Mélenchon, so my critical appraisal of voters’ behavior is not triggered by an ideological disappointment!
The results of the first round of the 2022 French Presidential elections are now known:
While there were twelve candidates, the first three gather almost 73% of the votes. Interestingly, these results were largely predicted by the latest polls that have been published Friday: Macron and Le Pen were predicted between 23 and 27%, and Mélenchon between 17 and 19%. No other candidate, including Zemmour and Pécresse, was predicted above the 10%. Even if polls are naturally subject to a margin of error, the gap was such that the probability that Macron/Le Pen/Mélenchon would not be the three first places was very low. However, it was indeed far from being clear who the first two would be, a very important fact as only those two qualify for the second round. Today, leftist voters and politicians are highly disappointed, as they realize that they have missed the opportunity to have a candidate in the second round for less than 500000 votes. Except for Mélenchon, candidates at the left of the political spectrum (Hidalgo, Jadot, Roussel, Arthaud, Poutou) combine for more than 4 million votes. If one out of seven of those voters had settled for Mélenchon instead, the second round of the Presidential election would have a fairly different shape. The same could have been true on the other side of the political spectrum: Le Pen was not far of being behind Mélenchon, and this is case the more than 2 million votes given to Zemmour would have been wasted.
Considering that polls are public information, one may wonder what crossed the mind of the 27% of voters who voted for candidates for which it was virtually certain that they would be eliminated. Can we make sense of the fact that these voters decided to not vote strategically (i.e., for their most preferred candidate among those that have a realistic chance to go to the second round) in this context? Here are a few hypotheses.
H1: These people do trust polls or underestimate their accuracy.
H2: These people consider that the marginal effect of their vote is virtually null.
H3: These people do not care about the outcomes of the elections, and/or have other and stronger motivations.
H4: The three top candidates did not sufficiently match their political preferences.
I’m skeptical regarding hypothesis H4. Of course, there are differences between the programs of Le Pen and Zemmour, as well as between the programs of Mélenchon and the other leftist candidates. But these differences are arguably anecdotical compared to the differences between the three favorite candidates. If H4 is the correct explanation, that would mean that voters (one-quarter of them at least) have very fine-grained political preferences. This seems implausible.
Hypothesis H1 is plausible. There is a relative distrust in polls, which is due, in the French case to the painful memory of the 2002 Presidential elections when Le Pen (the father) qualified for the second round, an event that nobody among the pollsters saw coming. This skepticism is also fed by the politicians themselves who tend, as part of a populist strategy, to present polls as attempts from the elite to manipulate voters. Anyway, the track-record of polls is actually fairly good, especially for the cases of the last three French Presidential elections.
Hypothesis H2 sounds reasonable. It is well-known that the probability that one’s vote will be decisive in an election like this one is almost null. But then, consistency would demand that these people don’t vote in the first place! So, either people are inconsistent or this is not the correct explanation.
Hypothesis H3 is the most plausible explanation. But then, let me put my Weberian commitment to refrain from making value judgments aside for a minute. This behavior doesn’t make sense! There are many ways and many opportunities to reveal “expressive political preferences” both in politics (e.g., by protesting) and in the economy (e.g., by buying local products). It could be argued that it is possible that because a candidate makes a decent score, the ideas she/he represents will be taken seriously over the next mandate. But recent experience suggests that this is wishful thinking, at least in the French case. The point of elections is to allocate political power where people want. And to have power, you have to win the elections!
A significant proportion of voters are thus very bad at strategic voting, which makes them doubly irrational: not only they are voting, but they are doing it very badly, not in terms of the preferences they express but how they are expressing them!
On a more analytical note, I would suggest that this form of irrationality is made more obvious by the weakening of traditional political parties. As Anthony Downs explained in his positive economic theory of democracy, political parties play an important role in helping voters to cope with uncertainty and the imperfection of the information. Parties help to channel the expression of political preferences by defining clear political alternatives. The rise of Macron’s center-right political movement five years ago contributed to the disintegration of the French political landscape, already made fragile by the two preceding chaotic presidencies under the traditional right and left parties. What yesterday’s results indicate is that the landscape is starting to reconfigure under a stable equilibrium with three major political forces. But obviously, some voters were lost during the transition!