Social Choice in the Real World

Abstract

This chapter begins by assessing the empirical evidence about the social choice paradoxes, especially the evidence about the Condorcet Paradox. I discuss briefly the possible reasons of the paucity of the empirical evidence. Then, I compare two real-life political institutions: the traditional Finnish presidential election system, which was in use from 1925 to 1988, and its US-American counterpart. For several reasons, the two Electoral Colleges provide an interesting subject for applied social choice. By using historical sources I try to show that the history of the Finnish Electoral College exhibits two, perhaps three instances of the Condorcet paradox, and provides dramatic examples of the effects of strategic voting. In contrast, in the US Electoral College there are no instances of the Condorcet paradox, for the unit-rule version of plurality tends to produce unambiguous majorities. This is acquired at a cost: the US system of electing the College violates the Arrovian independence and path-independence conditions far more often than the proportional Finnish system. For this reason, the connection between the distribution of the popular vote and the composition of the College is very weak. Both the US and the Finnish system may exhibit quasi-chaotic behaviour—“chaotic” in the sense that very small, unpredictable perturbations may sometimes change the outcome. This is one consequence of Arrow’s theorem.