Do investors care about carbon emissions? Evidence based on stock return co-movement with machine learning-augmented data
image: The models use three categories of variables as features: industry classification, firm fundamentals (including sales, total assets, non-current assets, property, plant, and equipment, and the number of employees), and firms’ carbon awareness. Industry is a significant determinant of a firm’s carbon emission intensity. Additionally, a firm’s fundamental characteristics, such as sales, number of employees, total assets, and other factors, also influence its level of carbon emissions. Furthermore, a company’s carbon awareness increases the likelihood of investing in green production technologies, thereby reducing its carbon emissions.
Credit: Lucas S. Li (Shanghai American School, China) Yan Zhao (City College of the City University of New York, USA)
Background and Motivation
As climate change intensifies, investors and regulators are increasingly focused on the role of carbon emissions in financial markets. However, evidence on whether emissions directly influence investor behaviour has been mixed. China Finance Review International (CFRI) brings you an article titled “Do investors care about carbon emissions? Evidence based on stock return co-movement with machine learning-augmented data”, which examines whether stocks of firms with similar carbon emission levels move together in the market.
Methodology and Scope
The study uses U.S. stock and carbon emissions data from 2004 to 2020. To overcome limited data coverage, the authors apply machine learning models—including Extra Trees, Random Forest, and XGBoost—to predict emissions for non-disclosing firms, nearly quadrupling the sample size. They then test whether pairs of stocks with similar emission intensities exhibit correlated returns, controlling for factors like industry, size, and profitability.
Key Findings and Contributions
- Stocks with similar carbon emission levels show significantly higher return co-movement, especially after 2012.
- This “carbon-based co-movement” is stronger in the machine learning-augmented dataset, confirming the effect holds beyond large emitters.
- Institutional fund flows also align with carbon similarity in the post-2012 period, indicating that investor behaviour has become more emissions-aware.
- The study introduces a novel method to test carbon awareness without relying on expected return models.
Why It Matters
The findings suggest that carbon emissions have become a meaningful factor in investment decisions, particularly over the past decade. This challenges the assumption that investors are indifferent to climate-related risks and supports the integration of environmental factors into financial analysis and policy design.
Practical Applications
- For Investors: Use carbon emission profiles to identify clusters of stocks that may move together, improving risk assessment and portfolio diversification.
- For Companies: Enhance carbon transparency and reduction strategies to attract aligned investor capital and reduce exposure to emission-related market risks.
- For Regulators: Support standardised carbon disclosure to improve market efficiency and help investors accurately price climate risks.
- For Researchers: Adopt machine learning and co-movement analysis to study investor behaviour in other ESG domains with limited data.
Discover high-quality academic insights in finance from this article published in China Finance Review International. Click the DOI below to read the full-text original! Open access for a limited time!