Estimates of insider trading in the betting on individual races, conditional on the Shin model (Economic Journal, Vol. 103 (1993), pp. 1141–1153), are employed in an analysis of the market anomaly observed by Gabriel and Marsden (Journal of Political Economy, Vol. 98 (1990), pp. 874–885) that Tote payments on winning bets consistently exceed those paid by bookmakers. Use of more appropriate statistical methods suggests that the original anomaly disappears, but that another remains: bookmakers pay more generously than the Tote on winning bets on favourites, but less generously on winning longshot bets. This discrepancy is shown to be associated with the incidence of insider trading in the betting on each race, and it is argued that it cannot be arbitraged away because of the bookmakers’ dominant market position.