EVALUATING THE USEFULNESS OF THE DEA MODEL TO IDENTIFY RATIOS THAT CAN EXPLAIN SOUTH AFRICAN STOCK MARKET RETURNS

Abstract

How to successfully identify ratios that will ensure profitable share selections remains a fundamental question in finance, as the literature has failed to promote a conclusive methodology. This study addressed this issue by being the first to prove that the multi-stage DEA model is a viable ratio selecting tool. The DEA model can capture the interdimensional relationships present and uncover relationships that are unknown to other methodologies (Cooper et al., 2007; Kumar et al., 2014). Data availability limited the study to evaluate 27 financial ratios and variables and 25 risk-adjusted performance ratios’ post-financial crisis explanatory abilities of 176 JSE listed companies. By consulting only, the efficiency scores generated from multiple multi-stage DEA regressions, and after eliminating correlation in portfolio compositions, results indicated that the Calmar ratio should be considered by both passive and active investor. This ratio demonstrated dominance in share selection, leading to market-outperforming portfolios from both a 1-, 3-, and 5-years momentum investment strategy perspective, respectively. In conjunction with the Calmar ratio, preliminary results suggested ratio compositions that differed over time and across sectors and industries. These findings violate the modern portfolio theory assumption of an efficient market and accentuate the importance of consulting time-varying market efficiency during the process of share selection, as market-outperforming decisions are possible.