Using stochastic frontier analysis instead of data envelopment analysis in modelling investment performance

Abstract We introduce methods to apply stochastic frontier analysis (SFA) financial assets as an alternative data envelopment analysis, because SFA allows us fit a with noisy data. In contrast conventional SFA, we wish deal estimation risk, heteroscedasticity in noise and inefficiency terms. investigate measurement error the risk return measures using simulation–extrapolation method develop residual plots test model fit. find that shrinkage estimators for makes striking difference fit, dealing only improves confidence model, are vital establishing The important they allow under assumption risks returns not known exactly.