Can ESG news sentiment be utilized by investors to maximize alpha in the short and medium term between the intervals that are covered by ESG ratings agencies? In this analysis, we investigate the relationship between short-term (weekly, monthly) and medium-term (quarterly) ESG news sentiment and stock returns.
Fund managers increasingly are looking towards sustainable investing while simultaneously needing to generate alpha for their investors. As a result, many funds have looked towards ESG rating agencies, such as MSCI, for guidance. MSCI and its peers create thorough ESG scorecards, but released only annually, these scorecards are far from timely. Moreover, the data in the annual ratings is subject to change, and funds with shorter investment horizons may not be able to effectively evaluate a company in August based on ESG ratings that came out months prior.
The Amenity ESG language model employs contextual analysis and language patterns to provide investors with real-time, broadly aggregated, and unbiased environmental-, social- and governance-related evidence and scores. Amenity’s taxonomy consists of 22 events across three dimensions (E, S, and G). These events can be viewed in sum to represent the ESG profile of a given company or analyzed individually. Our approach considers the sentiment, event, and materiality to the company described in the text.
Amenity’s ESG news dataset covers more than 10,000 firms. For this analysis, we removed companies with fewer than one result per day, leaving 1,227 of the largest public US companies to examine. Our sources comprehensively include top international, national, regional, local, and business news along with additional Amenity identified high quality sources. The dataset tracks daily ESG extractions from January 2018 to June 2021. All figures in the analysis make note of the number of companies scored. Table 1 (below) shows the top five mentioned companies. Figure 1 displays the volume of extractions for Alphabet Inc. over the past year.
Alphabet, the most negatively mentioned company in 2020, had its extraction totals inflated by a Justice Department lawsuit against the company for monopolistic business practices in late October of that year. The company saw a similar spike in negative coverage just a month and a half later, when the attorneys general in several states joined the lawsuit.
Other notable events that generated negative coverage include the July 22, 2019, Equifax data breach settlement (2,620 negative extractions), the October 18, 2019, voluntary recall of baby powder by Johnson & Johnson (1,343 negative extractions), and the December 2020 SolarWinds Corp. hack (612 negative extractions).
For this analysis, news sentiment scores were computed on a rolling weekly, monthly, and quarterly basis. The z-score was defined as the percentage of positive extractions for the company, relative to the percentage of positive extractions for the industry. The percentage is calculated as positive mentions divided by total mentions.
In practical terms the score captures the likelihood that the company received positive coverage over the time period assessed, relative to how other companies in its industry were covered. The scores take into account the overall sentiment for a company’s industry as a whole—a highly positive score for one company that is closer to the industry average score will not be seen as favorably as one that is far more positive than the norm. Adding industry context to the scores allows for well-behaving companies in traditionally ESG-negative businesses to be recognized.
In each period, the companies are split into quintiles by ESG score, with the companies in the top 20 percent of ESG scores in the top quintile and the lowest 20 percent in the bottom quintile.
To obtain the complete white paper, which covers our ESG methodology, backtesting results, and more:
This communication does not represent investment advice. Transcript text provided by S&P Global Market Intelligence. Analysis provided in collaboration with Quantitative Management Associates.
Copyright ©2022 Amenity Analytics.