Full article available on Network for Business Sustainability
The author, Tal Yifat, describes the evolution and approaches to tracking sustainability data, or ESG data, detailing three different approaches:
- Traditional ESG Data
- ESG Ratings Automation through measuring sentiment
- Company Culture data
TruValue Labs is covered in the second group, as shown below.
Case 2: Automating ESG Rating by Measuring Sentiments
Like Sustainalytics, TruValue Labs develops corporate ESG ratings and links them with financial performance. But where Sustainalytics uses human analysts to develop ratings, TruValue Labs draws on big data analytics. The company’s Insight360 platform rates companies’ ESG performance by analyzing text about the company from multiple Internet sources, such as news and social media. This automated approach allows ratings to be updated daily instead of annually.
Insight360’s automated approach uses sentiment analysis, a technology that identifies an author’s attitude to the topic he or she is writing about. When the system flags a reference to an ESG issue in relation to a company, it also assesses how positive or negative the discussion is.
Ratings reflect the overall attitude toward the company across a wide range of analyzed sources. The resulting score can be called “an objective sentiment,” suggests Dr. Jim Hawley, TruValue’s Head of Applied Research and professor emeritus at Saint Mary’s College of California.
TruValue Labs claims that the ESG intelligence uncovered by its platform can yield substantial financial gains. In a backtest on historical data from 2013-2017, a portfolio guided by Insight360 beat a benchmark of S&P 500 stocks by 5% per year. The study found that a measure of how much ESG performance trended up or down over the last 12 months was especially predictive of future financial performance.
Hawley says that these insights fit with academic findings. “Momentum has been found in the ESG literature to actually be a very good indicator for increasing long-term value. But the data that has been used to track momentum has been typically annual data, so what you get is momentum in the rearview mirror.” By scanning tens of thousands of data sources daily, TruValue Labs “can track that much close to real time, in a much more granular way, because we have much more data.”