ABSTRACT
In a recent contribution to Applied Economics Letters, Vaughan Williams and Paton (1997) (hereafter VWP) develop the debate relating to market efficiency in the context of betting markets by examining the proposition that information efficiency is dependent upon a perceived element of information inefficiency in the relevant market. Information efficiency is defined as ‘an absence of unexploited opportunities for arbitrage’. The specific test of efficiency employed relates to the existence or otherwise of a differential between returns to winning horses in a sample of parallel UK pari-mutuel and bookmaker-based horse-race betting markets. Perceived information inefficiency is identified as being associated with ‘a marked movement in bookmakers’ odds’. The conclusion suggests that the presence of such information inefficiency signals induces bettors to act so as to eliminate the cross-market returns differential, whereas in the absence of these inefficiency signals, information inefficiency endures in the form of a returns differential. The paper raises interesting questions, but invites critical comment relating to the terminology employed, the conclusion drawn and aspects of the methodology and analysis. A general observation which is pertinent to the majority of the points developed below is that the investigation of data relating to betting markets, and the subsequent interpretation of results, demands an understanding of and willingness to accommodate the institutional peculiarities of this type of market. The failure of the paper under discussion to acknowledge the idiosyncrasies of betting market structure and process results in inappropriate methodologies and misleading interpretation. In making this general observation, it should equally be noted that shortcomings in context awareness are not uncommon in studies of betting markets, where the susceptibility of data to analysis of general market efficiency issues may deflect the researcher from influences associated with the specific context. A recent example of this is the influential work of Gabriel and Marsden (1990, 1991) (and see, for a critique, Bruce and Johnson, 1996).
