Incorporating ESG into the Black-Litterman Model: A Penalization Approach Using Swedish Equity Data

Abstract

This thesis investigated how the ESG factors – environmental, social and governance – can be integrated into the Black-Litterman portfolio optimization model. By utilizing scaled ESG scores to modify the model by a penalization scheme on the covariance matrices, an adjusted Black-Litterman model was obtained. Thereby, the perceived risk of low-ESG scoring assets was increased, without excluding them from an optimization process. Monthly rebalanced portfolios were constructed over the period 2015–2023 using a 5-year rolling estimation window on a subset of OMXS30 stocks. Two portfolios were analyzed, a baseline portfolio using the Black-Litterman model and an ESG-integrated portfolio using the modified Black-Litterman model. The thesis emphasized the careful formulation and implementation of the optimization process itself, ensuring methodological transparency and comparability between the two approaches. By this analysis, the study aimed to explore one way to integrate sustainability factors into portfolio construction within an established optimization framework without a trade-off of risk-adjusted return. The results showed that the ESG penalization affects asset allocation, ESG exposure and risk-adjusted return. However, to what degree was dependent on the input of the model, the assumptions made, and the dataset used. Further, the two models’ sensitivity to parameter changes was analyzed to assess the robustness.