Method, ESG Score or Data? What Matters Most in Capturing ESG Risk Factors

The paper discusses how to better construct ESG risk factors. It constructs and compares the environmental and social factor using different methods, ESG categories and sets of data. In terms of different methods, the paper first constructs E/S risk factors under the two most commonly used risk factors construction methods: the Fama/French (FF) and the Fama/MacBeth (FM) method. Then, it tests these factors in asset pricing models. In addition, the paper compares four kinds of ESG categories to construct ESG risk factors. Finally, it discusses if a change of sets of data helps identify ESG risk factors. The paper finds that the FF method performs better than the FM method in capturing ESG risk premiums. The ESG category and sets of data are more important than methods. A change of ESG category will significantly improve asset pricing results. When using a more focused data set, in our case the energy and utility sector, the ESG risk factor can better explain asset returns. Therefore, in order to have a better estimation of ESG risk factors, policy-makers, academics and investors should be well aware of which aspect of ESG they want to measure (carefully choose the ESG category) and which group of stocks they want to model (sets of data).