A Comment on “ESG Risk Factors and Tail-Risk Mitigation”
E. Mark Curtis, PhD
Assistant Professor, Department of Economics, Wake Forest University
In recent years there has been a push from investors, activists and consumers to obtain more information on the sustainability practices of the companies in which they invest and from whom they purchase goods and services. As such, a number of ratings systems have been developed that attempt to quantify various aspects of a company’s corporate social responsibility (CSR). A critical question for investors is whether they can use companies’ CSR measures to predict firm performance and mitigate portfolio risk. Whether and how institutional investors should incorporate CSR metrics into their investment decisions remains an open question.
There are two differing views on the CSR–firm outcome relationship. First, classical models taught in economics suggest that firms will not voluntarily contribute to public goods or internalize any externalities they create. Under these models, CSR indicators such as voluntary pollution reductions, investing in community welfare, and increasing worker wages above market rates are all likely to come at the expense of a firm’s profit. As such, firms partaking in these activities would be expected to have lower returns and make for a less attractive investment opportunity. Indeed, CSR activity could represent what economists call “agency problems,” which occur when managers in the firm partake in CSR activity to curry favor for themselves in their local community rather than to increase the firm’s profits (Tirole 2001). For example a bank may make large donations to a local charity because the CEO values the increased clout she will gain in the community. If this, rather than the Bank’s profit, is the motivation for the giving, then higher CSR activity will be associated with lower company returns.
Alternatively, there are a variety of mechanisms through which CSR activity may increase profits and limit firms’ exposure to risk (Edmans 2011; Servaes and Tamayo 2013). CSR may increase brand-value and appeal to consumers’ preference for products and companies that do good for their communities (Kotchen 2006). Similarly, CSR measures may reflect firms’ compliance with government environmental and safety regulations and minimize risk of large government fines such as those recently experienced by BP following the Gulf of Mexico oil spill and Volkswagen’s emissions scandal. When there are government regulations such as a carbon tax, reductions in pollution can increase profits and lead to higher CSR scores.
Given the noisy signal that CSR activity provides for firm performance, how can institutional investors best make use of CSR measures when balancing portfolios? Research by Simon and Legnazzi provides a model to shed light on this question by taking 34 CSR measures from MSCI’s Environmental, Social, and Governance Index and examining whether they are correlated with reductions in portfolio tail risk for U.S. and European companies from 2013 to 2018. Ideally, the measures they use could disentangle CSR activity that reflects agency problems with activity that reflects profit-maximizing behavior. Doing so is not easy, but increasingly disaggregated CSR measures allow for investors to choose CSR measures that they feel best predict firm performance. These factors may differ based on the country and industry of the firm. For example, it is not surprising that the CSR factors that predict better firm outcomes in Europe are not the same factors as those that predict firm outcomes in the United States. On average, European consumers are likely to have a greater preference for environmentally friendly companies, and the European regulatory system is more likely to punish firms that are not already engaging in CSR activities.
A few features of the model and data are particularly critical to understanding the results. First, the model is univariate, meaning that the results are largely correlational. This is useful to the extent that investors make decisions based solely on CSR factors. However, this is unlikely to be the case. Investors make decisions based on a variety of variables and CSR measures are likely to be highly correlated with, for example, the size, age, and market power of the firm. Portfolio allocations based on CSR will likely lead to selecting firms that are disproportionately old and large. This may bias investors away from younger firms who, ironically, are more likely to be developing environmentally friendly technologies. Models that predict firm outcomes should incorporate many features of a firm. Understanding how CSR predicts firm outcomes conditional on other firm characteristics would be of considerable use. More convincing models could also use causal inference techniques now common in economics and finance to identify the causal effect of higher CSR scores on firm outcomes.
A separate important feature of the model is that the CSR scores are normalized by industry. As such, when taking these results and using them to compare firms, the results are only applicable when comparing firms that are in the same industry. These results should not be used when investors decide their exposure to particular industries. Once investors have decided their desired industry allocation, these results can be used to compare firms within a chosen industry. Other caveats apply as well. The results are specific to the 2013–2018 time period for U.S. and European companies. Investors should always ask whether changing market and regulatory conditions might shift the CSR–firm outcome relationship in the future.
Edmans, A. 2011. “Does the Stock Market Fully Value Intangibles? Employee Satisfaction and Equity Prices.” Journal of Financial Economics 10(3): 621–640.
Kotchen, M. 2006. “Green Markets and Private Provision of Public Goods.” Journal of Political Economy 114: 816–34.
Servaes, H., and A. Tamayo. 2013. “The Impact of Corporate Social Responsibility on Firm Value: The Role of Customer Awareness.” Management Science 59(5): 1045–1061. Tirole, J. 2001. “Corporate Governance.” Econometrica 69(1): 1–35.