The investment universe consists of stocks with required data (environmental indicators related to emissions, resource use, and innovation – such as Total Water Withdrawal, Electricity purchased, Total Energy used, Waste Recycling Ration, CO2 E.E. Indirect – S3; all indicators that were left after data cleaning in the research are listed in Table 1 of the paper) available on Refinitiv. Microcap stocks and stocks with a price of less than $5 are excluded. Furthermore, all firm-year observations ought to have non-missing total CO2 emissions. Then, only variables that have at least 60% coverage of firm-year observations are included. If for a firm-year observation, over 20% of the remaining variables are still lacking value, the firm-year observation is omitted.
ASSET CLASS: stock | REGION: United States | FREQUENCY:
Monthly | MARKET: equities | KEYWORD: Machine Learning, Strategy, Equities
I. STRATEGY IN A NUTSHELL
The strategy uses ESG-related data from Refinitiv, filters out incomplete records, and applies Random Forest machine learning to predict monthly stock returns. Portfolios are built by longing high-return predictions and shorting low-return ones.
II. ECONOMIC RATIONALE
Machine learning captures the hidden impact of detailed environmental indicators on stock performance better than aggregate ESG scores. This enhances predictive power and strengthens the link between sustainability and future returns.
III. SOURCE PAPER
Environmental Variables and Stock Returns [Click to Open PDF]
William O. Brown, University of North Carolina (UNC) at Greensboro; Xiaoli Gao, University of North Carolina (UNC) at Charlotte – Finance; Yufeng Han, University of North Carolina (UNC) at Greensboro – Bryan School of Business & Economics; Dayong Huang, Central Washington University – College of Business; Fang Wang
<Abstract>
Individual environmental variables may contain information that is obscured in aggregate environmental scores. We apply machine learning methods to granular environmental variables and study whether they are associated with future stock returns in the cross-section. A long-short portfolio that longs stocks with high forecasted returns and sells stocks with low forecasted returns earns more than one percent per month. Stocks with high forecasted returns are associated with high environmental operational standards. One interesting finding is that Scope 3 emissions are more important than Scope 1 and Scope 2 emissions in predicting future stock returns. Consistent with Pastor, Stambaugh, and Taylor (2022), the long-short portfolio performs better when climate concerns in the media are more intense.
IV. BACKTEST PERFORMANCE
| Annualised Return | 16.08% |
| Volatility | 12.86% |
| Beta | N/A |
| Sharpe Ratio | 1.25 |
| Sortino Ratio | N/A |
| Maximum Drawdown | N/A |
| Win Rate | N/A |
