The predictive ability of relative efficiency for future earnings: an application using data envelopment analysis to Spanish SMEs

Abstract

Research on earnings prediction has documented that the transitory component of current earnings undermines its predictive ability about future earnings. The implication of this finding is that a measure that better captures the underlying persistent component of earnings may prove very useful in predicting future earnings when used along with current earnings. Our study, based on a large sample of 1939 Spanish Small and medium Enterprises (SMEs), investigates whether an alternative measure of performance ignored in previous research on earnings forecasting–i.e. relative efficiency–has predictive ability over and above current earnings and book value of equity. Relative efficiency captures the inherent ability of a firm–as compared to other similar firms–to make the most productive use of available resources and is measured using Data Envelopment Analysis (DEA) nonparametric technique. Our findings highlight that our DEA-based efficiency measure has an incremental predictive ability over and above current earnings and book value of equity for predicting future earnings. Moreover, we have further validated the models in a holdout sample and our results evidence the highest forecast accuracy of the model that includes our efficiency measure as an additional predictor to current earnings and book value of equity.