Hannu J. Nurmi

This book deals with 20 voting procedures used or proposed for use in elections resulting in the choice of a single winner. These procedures are evaluated in terms of their ability to avoid five important paradoxes in a restricted domain, viz., when a Condorcet winner exists and is elected in the initial profile. Together with the two companion volumes by the same authors, published by Springer in 2017 and 2018, this book aims at giving a comprehensive overview of the most important advantages and disadvantages of voting procedures thereby assisting decision makers in the choice of a voting procedure that would best suit their purposes.

This book deals with 18 voting procedures used or proposed for use in elections resulting in the choice of a single winner. These procedures are evaluated in terms of their ability to avoid paradoxical outcomes. Together with a companion volume by the same authors, Monotonicity Failures Afflicting Procedures for Electing a Single Candidate, published by Springer in 2017, this book aims at giving a comprehensive overview of the most important advantages and disadvantages of procedures thereby assisting decision makers in the choice of a voting procedure that would best suit their purposes.

It is well-known that different social choice procedures often result in different choice sets. The article focuses on how often this is likely to happen in impartial cultures. The focus is on Borda count, plurality method, max-min method and Copeland's procedure. The probabilities of Condorcet violations of the Borda count and plurality method are also reported. Although blatantly false as a descriptive hypothesis, the impartial culture assumption can be given an interpretation which makes the results obtained in impartial cultures particularly significant, viz. the probabilities of deviations in choice sets indicate how far apart are the intuitions underlying various choice procedures.

The difficulty of making social choices seems to take on two forms: one that is related to both preferences and the method used in aggregating them and one which is related to the preferences only. In the former type the difficulty has to do with the discrepancies of outcomes resulting from various preference aggregation methods and the computation of winners in elections. Some approaches and results which take their motivation from the computability theory are discussed. The latter ‘institution-free’ type of difficulty pertains to solution theory of the voting games. We discuss the relationships between various solution concepts, e.g. uncovered set, Banks set, Copeland winners. Finally rough sets are utilized in an effort to measure the difficulty of making social choices.