Economic and Statistical Properties of Implementable Trading Strategies

Abstract

Capital market studies typically report one set of results that tests for both economic and statistical significance. Statistical significance is necessary, but not sufficient, for economic significance, which requires that a trading scheme earns abnormal returns, after information costs, transactions costs, the opportunity cost of capital, and adjustment for risk preferences. Such a strategy must be implementable in real time, not CRSP/Compustat time. Typical research designs are evaluated in the context of economic and statistical properties of implementable trading schemes. Applying the analysis to well known accounting papers suggests that abnormal returns cannot be earned based on publicly available information.

Keywords: Trading profits, Abnormal returns, Transactions costs, Implementable

JEL Classification: G11, G14, M41