I saw David Myatt present this paper in Oxford this Fall. It made quite a splash with some surprising results about voter turnout rates consistent with “rational choice” theory. Voting theories based on the assumption that voters calculate the costs and benefits have always been thought to imply very low turnout in order to generate sufficiently high probabilities of close elections.
Consider a region with a population of 100,000 where 75% of the inhabitants are eligible to vote. Suppose that a 95% confidence interval for the popularity of the leading candidate stretches from 56% to 61%. If each voter is willing to participate in exchange for a 1-in-2,500 chance of influencing the outcome of the election, then turnout will exceed 50%. Greater turnout for the underdog offsets her disadvantage.