Discussion of: “Contextual Fundamental Analysis Through the Prediction of Extreme Returns”

Abstract

Beneish, Lee and Tarpley (2000) represents one of a small, but growing number of studies that de- velop and test contextual fundamental analysis techniques. Such studies offer great promise for increasing our understanding of the role of accounting information in evaluating firm performance. However, these studies also introduce their own unique research design issues. In this paper, I discuss the opportunities and research design issues facing this new line of research, using Beneish, Lee and Tarpley to illustrate my points. Research on fundamental analysis has seen a resurgence in the last decade. Following the lead of Ou and Penman (1989), a large number of papers have demonstrated the useful- ness of financial ratios based on accounting statement data to predict future performance. However, with just a few exceptions, most of this work has been conducted at a very gen- eral level and empirical results have been documented using broad cross-sections of firms. Such broad sample research is useful for demonstrating the promise of fundamental re- search and for providing some very general guidelines. However, it seems clear that most practical applications of fundamental analysis should be carefully tailored to the charac- teristics of the firms being analyzed. For example, a cursory analysis of sell-side research reports reveals that analysis techniques vary systematically as a function of characteristics such as industry membership and growth opportunities. Despite the widespread practical use of contextual fundamental analysis, relatively little academic research has attempted to develop and test techniques of fundamental analysis that are tailored to specific con- texts. Thus, there are great potential opportunities for research on contextual fundamental analysis.