DEA Portfolio Modeling - The Case of Socially Responsible Investing

Abstract

In this paper, we expand the literature on multi-criteria portfolio modeling using data envelopment analysis (DEA). We do not solely use DEA as a positive screening mechanism, but also exploit the information contained in DEA efficiency scores directly in order to build public equity investment portfolios. With the main focus of our paper on multi-criteria portfolio modeling for investors interested in socially responsible investments (SRIs), we conduct a broad empirical analysis of this approach with market data from the USA, Europe, Asia and Oceania going back to 2005. Evaluating the out-of-sample performance of our models and challenging the predominant view of the literature on DEA stock selection, we find that also less efficient firms have a positive impact on portfolio performance but that the approach is not suited when considering non-financial criteria. In addition, we show that it is indeed beneficial to directly apply DEA efficiency scores to portfolio modeling. All results indicate that these portfolios are compliant both with respect to financial and non-financial criteria.