An Inside View of Corporate ESG Practices

By | January 26, 2022

Companies are becoming increasingly vocal about corporate sustainability. Nine out of ten S&P 500 companies voluntarily issued a sustainability report in 2019, detailing their commitments to environmental, social, and governance (ESG) objectives. For example, in Coca Cola’s 2020 ESG report, the company set a goal by 2030 to be “50% led by women” and to match its workforce demographics to the U.S. population demographics. While broadcasting such a commitment can improve a firm’s ESG rating, the commitment may not impact actions within the firm. Consequently, investors aiming to invest in sustainable companies may end up investing in firms that do not follow through on their ESG commitments. 

One way to evaluate whether ESG commitments permeate a firm is to examine how the firm’s employees perceive its ESG practices. In a new study, I analyze 10.4 million anonymous employee reviews on, a career intelligence website, to capture an inside view of ESG practices. Each Glassdoor review contains two open-ended sections in which an employee describes the pros and the cons of working at a company. I capture how a firm’s employees view its ESG practices by how often they mention ESG topics in the pros section relative to the cons section. I do so with the help of machine learning to create comprehensive dictionaries for ESG topics that employees often mention in a review, such as biofuel and recycling on environmental topics, advocacy and social justice on social topics, and malfeasance and embezzlement on governance topics.

The dictionaries indicate that employees pay extensive attention to ESG topics. On average, 43% of reviews mention at least one ESG word in my sample of 1,936 publicly listed companies. Moreover, the employees’ attention to ESG topics, as measured by the percentage of ESG words in a review, has remained stable between 2008 and 2021, suggesting that the employees have always cared about ESG issues. The employees’ attention, nonetheless, spikes around major ESG events. For example, on environmental issues, the employees’ attention spiked in 2015 when world governments signed the Paris Agreement to limit global warming. On social issues, the employees’ attention spiked in 2020, when the COVID Crisis raised health and safety concerns in the US, and again when the death of George Floyd triggered large social unrest throughout the country.  On governance, the attention spiked when companies were put to the test during the 2008 Financial Crisis and the 2020 COVID Crisis.

Ranking firms based on employees’ inside view appears sensible. Ranked first on environmental practices is SunEdison, a solar energy firm, while last is Alpha Natural Resources, a coal producer. On social practices, ranked first is HP Inc., whose employees say the firm “embraces diversity and inclusion,” while ranked last is Colony Capital, whose employees criticize an unfriendly workplace with high turnover. On governance, ranked first is LinkedIn, whose employees say that the firm “practices 100% of its values and cultures,” while ranked last is Sterling Bancorp, whose employees complain about poor transparency and leadership. 

While the inside view appears reliable, it often diverges from outside assessments of ESG practices, i.e., existing ESG ratings. The correlation between the inside view and the MSCI rating is -0.02, 0.11, and 0.07 for the E, S, and G categories, respectively. The correlation between the inside view and the ESG rating from Refinitiv is even lower. This is not surprising because ESG ratings are known to be very imprecise. According to research from the Massachusetts Institute of Technology, even within the most trusted ESG ratings, the correlation among them ranges widely from -0.01 to 0.81 on different ESG categories, even when the ratings are based on essentially the same information from corporate disclosures.

Despite its divergence from existing ESG ratings, the inside view is highly informative about firms’ ESG practices. I find that the inside view not only predicts a firm’s future social misconduct, accounting issues, shareholder activism, and downside risk, but also its valuation, sales growth, and its likelihood of being in Fortune’s 100 Best Companies. In predicting these outcomes, the inside view adds significant power to the MSCI ESG ratings.

Because the inside view is informative about internal ESG practices, I use it to study whether a firm’s stated ESG policies permeate the firm. If a firm’s ESG policies effectively influence its internal ESG practices, then a firm with better ESG policies should have a better inside view. However, this might not be the case if the firm establishes its ESG policies mostly to greenwash its public image. I find evidence consistent with the greenwashing hypothesis. The association between a firm’s stated ESG policies (as indicated by MSCI) and its future inside view is close to zero for all the ESG categories.

A likely reason for the weak link between a firm’s stated ESG policies and its inside view is that the firm often faces little cost in stating its ESG policies or commitments. To investigate this issue, I study two settings in which a firm shows a broad commitment to ESG policies, one without and another with a high cost of not following through with the commitment. The first setting is when firms signed the Business Roundtable (BRT) statement in 2019 committing to serving all stakeholders rather than just shareholders. These firms did not seek the approval of their shareholders or boards of directors before signing the statement, so the BRT commitment was likely cheap talk, according to researchfrom Harvard University. The other setting is when firms participate in the United Nations Global Compact (UNGC), the world largest corporate sustainability initiative. Unlike the BRT commitment, the UNGC commitment carries a likely large compliance and reputational cost because it requires that firms report annual progress or else get publicly expelled. Historically, the UNGC has expelled over 40% of its participants between 2000 and 2020.

As expected, the inside view indicates that the BRT commitment appears less credible than the UNGC commitment. Signing the BRT statement is not significantly associated with a larger improvement in the inside view of ESG practices. Before signing the BRT statement, BRT firms do not have a significantly better inside view either. By contrast, the employees’ inside view on the S and G categories significantly improves three years after a firm joins the UNGC, relative to otherwise similar firms. Overall, the high cost of an ESG commitment appears to make it more likely that the firm will follow through.

Finally, I study cyber-attacks as an unexpected increase in a firm’s incentive to improve its ESG practices. Prior research documents that a firm appears to improve its ESG policies after a cyber-attack to recoup its reputation. If most of the improvement in ESG policies is about shallow ESG activities, such as charity donations, then the firm’s employees will not see its ESG practices improve after the attack. Indeed, I find that the inside view hardly improves several years after a cyber-attack.

Overall, I show that employees discuss ESG topics extensively in their anonymous reviews, thus providing substantial information about a firm’s ESG practices. Nonetheless, the employees’ inside view shows no strong link to a firm’s stated ESG policies or existing ESG ratings, suggesting a prevalence of corporate greenwashing. In various settings where a firm appears to improve its ESG policies, I consistently find that a firm’s ESG policies do not permeate the firm unless the firm sees clear benefits of following through, or clear costs of not following through with the ESG policies.

Hoa Briscoe-Tran is a PhD Candidate at Ohio State University

This post is adapted from his paper, “An Inside View of Corporate ESG Practices” available on SSRN.

The views expressed in this post are those of the author and do not represent the views of the Global Financial Markets Center or Duke Law.

Leave a Reply

Your email address will not be published. Required fields are marked *