Building a Socially Responsible Equity Portfolio Using Data Envelopment Analysis

Abstract

This paper uses two techniques to build a socially responsible portfolio of U.S. equities and examines prospective perf ormance us ing publicly avai lable data. The first technique eliminates stocks from considerat ion using categorical exclusions with a restrictive Environment, Social and Governance (ESG) screen. The paper shows that stocks surviving the screen have a significantly higher average projected Value Line alpha and are more likely to have a Morningstar 5-star rating. Using cate gorical exclusions, however, introduces a sector bias in that the ESG screen is more likely to restrict stocks from the manufacturing sector than the service sector. The second technique does not introdu ce a sector bias because it uses a best-in-class optimization approach in place of screening. The paper introduces a linear programming model called Data Envelopment Analysis (DEA) to the application of SRI portfolio development to find the best financially and socially performing companies within each industry sector. When compared to a categorical exclusions portfolio, a DEA portfolio is rated significantly higher by Morningstar and Value Line. Depending on the specific needs of a socially responsible in vestor, t he DEA technique could be a better tool in developing a financially and socially balanced equity portfolio.