Public Sentiment and the Price of Corporate Sustainability

George Serafeim is Professor of Business Administration at Harvard Business School. This post is based on a recent paper authored by Professor Serafeim. Related research from the Program on Corporate Governance includes Socially Responsible Firms by Alan Ferrell, Hao Liang, and Luc Renneboog (discussed on the Forum here) and Social Responsibility Resolutions by Scott Hirst (discussed on the Forum here).

News about firms’ impact on society is an everyday phenomenon. In this paper, I analyze how public sentiment influences the market pricing of firms’ sustainability activities and thereby the future stock returns of portfolios that integrate ESG data. This is the first paper that combines analyst driven ESG ratings from MSCI with big ESG data from machine learning and artificial intelligence from TruValue Labs.

Thousands of companies are investing resources to reduce energy consumption, waste and carbon emissions and to provide products that improve environmental and social outcomes. For example, developments in healthy nutrition, access to wellbeing services, low carbon transportation, and green buildings have provided billions in revenues for companies that developed products for these markets (Generation Investment Management 2017). Similarly, companies spend significant resources to improve employee safety and well-being and to conduct business with integrity avoiding corruption. These activities, typically referred by companies as corporate sustainability activities, are under the supervision of a Chief Sustainability Officer and are disclosed in sustainability reports (Miller and Serafeim 2015). The data from sustainability reports and other sources that might also reflect controversies around human rights, pollution, discrimination and corruption, are collected by data providers and form the basis of measures of company’s performance on environmental, social and governance (ESG) issues.

I combine ESG performance scores from MSCI, the largest provider of ESG data to investors, with big data from TruValue Labs that measure public sentiment momentum around ESG issues between 2009 and 2018. The latter is a measure of whether sentiment has turned negative or positive for a company, from a set of vetted, credible, and reputable resources (e.g. NGOs, industry analysts, think tanks, media) in the past twelve months. I expect a lower valuation of corporate ESG performance in the presence of negative sentiment momentum for multiple reasons. First, firms with strong ESG performance and occasional or temporary societal controversies might be judged as weak ESG performers. Similarly, firms with weak ESG performance that have strong marketing campaigns to advertise their ESG activities might be judged as strong ESG performers. Moreover, negative news on a specific topic (e.g. supply chain controversies) might affect investor views about other ESG issues (e.g. climate change strategy or human capital development) leading them to undervalue strong ESG performance in those other issues. Second, investors might assign a higher discount rate to a firm’s ESG performance in the presence of negative sentiment momentum because they expect future reputational, legal, or operating costs. Third, even if investor views about a firm’s ESG performance are unaffected by sentiment, their incentives might lead them to ignore firms with strong ESG performance and negative sentiment momentum (or to hold firms with weak ESG performance and positive sentiment momentum). If institutional asset owners and retail investors value holding (avoiding) companies with positive (negative) sentiment momentum, asset managers will act to satisfy their clients’ preferences (Amel-Zadeh and Serafeim 2018). The key insights are below:

  • I find that the valuation of corporate ESG performance increases as a function of public sentiment. Figure 1 shows that the price of corporate sustainability performance has increased over time. This is the estimated premium (if positive) or discount (if negative) that firms with better sustainability performance trade relative to peers after accounting for several factors such as current profitability, size, leverage, past returns and other firm characteristics. This is good news for companies that perform better on material sustainability dimensions (as defined by MSCI) as the market rewards them with a higher multiple.
  • The positive association between ESG performance and market valuation is stronger for firms with more positive public sentiment momentum. An increase in a firm’s ESG performance has nearly two to three times the effect on a firm’s market valuation for a firm with positive relative to a firm with negative public sentiment momentum. In the presence of negative sentiment a firm’s sustainability performance is largely discounted. This has fundamental implications for Chief Sustainability Officers and business leaders in how they work with the broader ecosystem.
  • The higher price of corporate sustainability poses a challenge for ESG investors. Do you get a good value for money from your investments? It is not only a matter of the value of corporate sustainability anymore, but it is also a function of the price you are paying for it. Moreover, the differential pricing of sustainability activities based on public sentiment momentum raises the question if the price paid for these activities is efficient, or the market undervalues strong ESG performance in the presence of negative sentiment or overvalues strong ESG performance in the presence of positive sentiment. The paper shows that combining big data and analyst driven ESG data allows one to identify “value” opportunities in the ESG space and as a result construct an investment strategy that delivers alpha while investing in companies with greatly superior ESG performance scores.
  • I construct ESG factors following the standard approach in the literature (Fama and French 2018), based both on the levels and/or the change in ESG performance scores. Importantly, I separate the ESG factor to a high and low sentiment. The high sentiment ESG factor goes long on firms with strong ESG performance and positive sentiment momentum and short on firms with weak ESG performance and negative sentiment momentum. Therefore, the factor has a positive spread both on ESG performance and sentiment momentum. The low sentiment ESG factor goes long on firms with strong ESG performance and negative sentiment momentum and short on firms with weak ESG performance and positive sentiment momentum. Therefore, the factor has a positive spread on ESG performance and a negative spread on sentiment momentum.
  • The low sentiment ESG factor produces significant positive alpha of about 4-5% annually. It exhibits a higher Sharpe ratio than other factors during the period of study and does not exhibit significant correlation with any of the six factors introduced in the literature (Novy-Marx 2013; Fama-French 2016). Moreover, the long portfolio of the ESG factor has dramatically better ESG profile than the short portfolio. The average ESG score assigned by MSCI is close to 100% higher in the long portfolio and the average change in ESG score for the long (short) portfolio is an increase (a decrease) of close to the sample standard deviation of ESG score. This suggests that the ESG factor goes long on firms with significantly greater positive social impact than the firms in the short portfolio, if MSCI ESG ratings are correlated with social impact.
  • In contrast, the high sentiment ESG factor exhibits insignificant alpha. Moreover, it exhibits very strong correlations with many other factors. Importantly, it has a strong negative correlation with the value factor suggesting high sentiment firm portfolios with better ESG characteristics have returns that resemble those of growth stocks. I find some evidence of overvaluation in more recent years, with the high sentiment ESG factor yielding negative alphas in the years between 2015 and 2018, although the estimates are insignificant. The low sentiment ESG factor yields a significant positive alpha in those years.
  • I complement the results using US data with data across 37 more countries to understand if the role of sentiment generalizes to other markets. The low sentiment ESG factor delivers even higher alpha, between 2010 and 2018, in a sample of international firms traded in European and Asian-Pacific stock exchanges. The four-factor alpha in the international sample ranges between 44 and 57 basis points monthly (6-8% annually). As in the US, in the international sample the high sentiment ESG factor yields an insignificant alpha. Overall, the results support the interpretation that the market undervalues sustainability activities in the presence of negative sentiment.

The complete paper is available for download here.

Both comments and trackbacks are currently closed.