A Case for Theory-Driven Experimental Enquiry

We argue that finance theorists and practitioners need to examine the reasons behind a seeming anomaly. The behavioral anomalies in the finance literature can be classified as price and return effects, volume and volatility effects, time-series patterns, and miscellaneous effects. For each category, the empirical literature offers a multitude of explanations. Our main thesis is that theory-driven experimental analysis will allow clarification among competing explanations and should complement existing empirical paradigms. We present a theoretical information-processing framework for examining the psychology of financial decision making. The framework comprises both cognitive and motivational antecedents of bias in financial decision making and provides a grounding for many behavioral anomalies noted in the literature. To illustrate use of the model, we examine financial decision-making biases in the choice between underpriced (value) and overpriced (glamour) stocks. The main thesis of this article is that theory-driven experimental enquiry is an appropriate methodological tool at this stage of the life cycle of behavioral finance. Many empirical irregularities have been found and shown to be robust, and numerous compelling explanations have been proffered for their existence. But either multiple explanations are argued for the same effect, or alternative explanations are posited to be driving the effect. Theory-based empirical enquiry can eliminate alternative explanations and isolate key causes. It can also highlight potentially important gaps in the literature and direct future empirical investigation. The biases have been demonstrated and replicated; what remains is to isolate why they occur.We develop a theoretical information-processing framework to examine the psychology of financial decision making. Many behavioral anomalies noted in the finance literature can be derived from this framework. A brief survey of the literature results in a classification of common behavioral anomalies into three primary categories—price and return effects, volume and volatility effects, and time-series patterns—and a fourth miscellaneous category for other effects. For each category, we found that the empirical literature offers a multitude of explanations.The theoretical framework we suggest comprises both cognitive and motivational antecedents of bias in financial decision making. The model posits five stages at which cognitive biases may arise: perception, memory retrieval, information integration, judgment making, and behavior. Motivational effects are theorized either to directly affect the manner in which information is processed or to indirectly moderate the likelihood that cognitive biases will affect decisions.We used the proposed model to examine biases in the choice between underpriced (value) stocks and overpriced (glamour) stocks. A systematic series of five mini-experiments demonstrates the following results: First, there is experimental support for the hypothesis that when prior information about a stock is negative, people overreact to good news about the stock. Second, those results do not depend on whether the participant is trading for his or her own account or for a client. Third, the systematic preference for one stock versus another vanishes when the trade decision is a buy instead of a sell. Fourth, in a buy situation, when one stock's prior condition is positive, one stock's prior condition is negative, but all the news is positive, the stocks with previous negative results are significantly preferred to those with previous positive results. In summary, the primary antecedent driving choices between value and glamour stocks appears to be cognitive; thus, results differ according to information conditions and decision tasks. But our experimental exercise implies that psychological antecedents moderate the cognitive effects in the glamour versus value decision.Practitioners as well as finance theorists and researchers need to examine the why of a seeming anomaly. Practitioners need to understand underlying causes if they are to design appropriate, successful training, risk-management, and product communications. Theorists and researchers need to incorporate a psychological theory-driven experimental paradigm at the micro level to complement their existing work in behavioral finance. The model presented here should be a good first step toward a comprehensive theory of how financial decisions are made.