Choosing for others
New paper on decision-making by Sandro Ambuehl and B. Douglas Bernheim, published in the American Economic Review.
There are many ways to make decisions for groups of people (up to entire nations). For example, you can give each person one vote and choose the majority. Or you can use multi‑stage systems in which the least popular alternatives are filtered out first, and then the remaining options are voted on. Mathematically, it can be proven that there is no optimal system for making such decisions — this is known as Arrow’s impossibility theorem. Every imaginable system has substantial disadvantages.
But which systems produce outcomes that people are most likely to perceive as fair? Most methods used in practice (such as those in political elections in Switzerland) consider only which candidates people prefer most, but not which candidates they least want — you can vote for candidates you like, but only very limitedly against candidates you dislike. This new study by Sandro Ambuehl and B. Douglas Bernheim, published in the American Economic Review shows that this is exactly the opposite of what people consider fair.
Across different countries, people consistently find it far more important to ensure that as few people as possible must live with a very unpopular result than to ensure that as many people as possible get their most preferred result. In this sense, they have a strong preference for compromise, in contrast to the majority rule embedded in many common voting mechanisms.
Sandro Ambühl’s website