This paper presents empirical tests of market rationality using data from the point spread betting market on National Football League games. Data from this market avoid many common pitfalls of tests of rationality in conventional financial markets. The authors test for rationality with two types of tests, statistical and economic. Results of the tests reveal that the statistical tests cannot reject market rationality while the economic tests do reject market rationality.

This paper examines betting line changes from the opening to the closing of the point spread betting market on National Basketball Association games for evidence of informed trader betting. We show that within-betting period line changes significantly improve the accuracy of betting lines as forecasts of game outcomes. We examine individual line change magnitudes and show that these are directly and proportionately related to biases in opening lines. Further, line changes are of sufficient magnitude to remove these biases by the close of betting. We interpret these results as evidence that informed traders are influential in this market.DO DAY-TO-DAY AND WITHIN-DAY changes in asset prices reflect the incorporation of information into price? A number of recent market microstructure papers have cast doubt on the incorporation of information explanation. For example, studies of day-to-day price changes by French and Roll (1986), Roll (1988), and Cutler, Poterba, and Summers (1989 show that market volatility is much larger over periods when the market is open than over comparable periods when the market is closed, and that external news is not responsible for the majority of large daily stock price changes. The implication of this research is either that the arrival of private information causes price volatility or that trading itself introduces noise into prices. The withinday price change studies by Amihud and Mendelson (1987) and Stoll and Whaley (1990) demonstrate that prices formed at the opening of trading are more volatile than prices determined at other times of the day. These authors suggest that trading mechanisms used at the start of trading may be responsible for noisier opening prices.Efforts to validate the incorporation of information explanation have proceeded apace. At the theoretical level, for example, Romer (1993) develops a model allowing for rational changes in the market's assessment of funda-385 mentals, and hence, rational price movements, in the absence of the arrival of outside news. There are also numerous empirical studies. For example, Ederington and Lee (1993), building on the work of Harvey and Huang (1991), show that scheduled macroeconomic news announcements are responsible for most of the observed time-of-day and day-of-the-week volatility patterns in interest rate and foreign exchange futures markets. Gerety and Mulherin (1994), using forty years of hourly price index data from the New York Stock Exchange, find that transitory price volatility declines steadily over the trading day. This price volatility pattern is consistent with the hypothesis that trading aids price formation rather than the alternative arguments that trading mechanisms add to price volatility at the opening or that price changes are simply noise.An impediment for all empirical research into asset price changes is the lack of an endpoint at which an objective fundamental value can be observed and compared with actual prices. As a result, neither the price formation, trading mechanism, nor noise trading hypothese...

No abstract

No abstract

A recent study of the fixed-odds betting market on baseball games, while finding that the betting market is generally efficient, also found evidence of an underbetting on underdog teams. This article examines the evidence for this new anomaly. It corrects Woodland and Woodland's estimates of the commission, subjective win probabilities and test statistics. The efficiency null hypothesis cannot be rejected for all of their tests when revised test statistics are calculated for their sample period (however, like them, it was found that slight underdogs are underbet). It is also shown that their bias is not simply a bias involving favourites and underdogs. Whether underdogs are playing at home or away also seems to matter in their sample period. As well a positive relationship between returns and subjective probabilities was found for underdogs and favourites, a relationship suggestive of a favourite-longshot bias rather than its reverse. It is concluded that there is insufficient evidence to claim that this bias is a ‘true market inefficiency’.