Do horses like vodka and sponging? On market manipulation and the favourite-longshot bias

Abstract

One of the most striking empirical patterns of horse race betting markets is the favorite-longshot bias: Bets on favorites have dramatically higher expected returns than bets on longshots. The literature offers a couple of different, though not mutually exclusive, explanations based on risk preferences and probability perceptions. This article adds a new possible explanation: The favorite-longshot bias may be the rational answer of an honest audience to a simple, but highly lucrative cheating opportunity of insiders. We provide anecdotal evidence that the type of cheating we model here really takes place. What is more, by employing a large scale German data set we are able to demonstrate that the pattern of the favourite-longshot bias changes as the opportunity of cheating vanishes. The changes we observe are in accord with the cheating model we suggest.