Abstract
Recently, disequilibrium models have been proposed to explain security market disequilibrium conditions caused by noisy information channels, transaction costs, taxes and the costs of acquiring information. While statistical tests have been done to support these models, mechanical trading rule tests have not been done with stock market data. The implications of the disequilibrium models for the construction of trading rules are derived and an apropriate adaptive trading rule is constructed. Empirical tests comparing the adaptive trading rule, the nonadaptive trading rule and the buy-and-hold strategy suggest that the stock market behaves according to a stationary disequilibrium model. However, these disequilibrium conditions cannot be concluded to be inefficient since the trading rule after transaction costs was unable to use the disequilibrium conditions to consistently earn a superior profit.