Trend: The bots are coming (to ratings and reporting)
Automation and artificial intelligence are being leveraged to both generate and evaluate ESG data. Is that a good thing?
The following is adapted from State of Green Business 2020, published by GreenBiz in partnership with Trucost, part of financial information and analytics giant S&P Global.
Corporate reporting on sustainability — including environmental, social and governance (ESG) performance and achievements — has grown more than fivefold in the past 10 years. Roughly 20 percent of S&P 500 companies published a sustainability report in 2011. In 2018, that number rose to 86 percent. During that time, sustainability professionals have fretted about whether anybody reads their reports.
What we’re beginning to see is that it may not be "who" but "what." Automation and artificial intelligence (AI) are being leveraged to both generate and evaluate ESG data.
The bots and AI are largely in response to the confusing world of ESG reporting. There are more than 600 ESG ratings agencies globally, according to the Global Initiative for Sustainability Ratings, as ESG data becomes a greater factor in a company’s valuation and access to capital. The challenge is that current corporate ESG disclosures lack consistency and standardization.
What’s a corporate reporter to do?
But in the past year or so, there has been increased interest in understanding the differences among the various ratings and rankings organizations. This has become more pronounced since Institutional Shareholder Services, and its main competitor, Glass Lewis, started focusing more on E&S and not just G. These two prominent proxy advisory services provide institutional investors with assistance in voting their shares at corporate annual meetings.
The frustration for many corporate sustainability reporters is the general lack of transparency as to how their company is scored. Firms such as CSRHub seek to synthesize data from myriad sources, ranging from "best of" lists to ratings agencies, but the scores don’t contain enough information and context for most investors.
Subscribing to a service such as CSRHub or Sustainalytics is often more about the data than the rankings. Firms such as these provide data services where software known as APIs can pluck data and populate a firm’s database, where its internally developed algorithms can test and validate various investment hypotheses.
Taking this a step further are firms such as Sensefolio and Arabesque, which complement traditional ESG data with feeds from news reports, social media posts, job postings and review websites such as Glassdoor. This data is then leveraged with self-learning quantitative models to assess the performance and sustainability of globally listed companies.
These are strategies and technologies that mainstream investors have been deploying for some time, although it is still early days. According to MarketWatch, financial markets don’t produce enough data to get the most out of AI and machine learning. AI functions best on billions of data points rather than millions, but three decades of daily share-price data for the benchmark S&P 500 Index would yield only about 4 million data points, a mere drop in the big-data bucket.
The takeaway is that AI works best when humans develop an investment thesis and machines test that theory.
For many investors, the technology doesn’t have to be that exotic. For example, bot searches of companies’ 10-Q and 10-K filings with the U.S. Securities and Exchange Commission can track and redline what has changed when it comes to sustainability and ESG topics. Investors take notice when a phrase in what normally may be seen as boilerplate shifts from "probable" to "likely" from one report to the next. A machine is more likely to spot such subtleties.
In response, investor relations and sustainability teams are striving to discover the best keywords to use to highlight strategic information. Regular checkups on a Bloomberg terminal of a company’s publicly available information can help make sure the bots are getting the right data — and getting the data right. Organizations are also subscribing to software-as-a-service providers such as Datamaran to identify and monitor nonfinancial risks. The service tracks 100 nonfinancial topics for thousands of companies by sifting and analyzing millions of data points from publicly available sources.
It isn’t only external data that’s automatically collected, sifted and analyzed. For years there has been a niche market of vendors that have offered carbon accounting software. Adoption has been slow to scale as the cost of the software and even more so the services to implement and support it outweighs managing by old-fashioned spreadsheet.
One barrier to wider adoption has been CIO skepticism in buying from small vendors. That’s where the Salesforce Sustainability Cloud may gain greater acceptance: The company’s customer relations management, or CRM, software is already installed at more than 150,000 customers and has 3.75 million subscribers. The sustainability application focuses primarily on measuring and tracking energy consumption, climate emissions, waste generation and environmental data. Early users claim the ability to produce environmental data as fast as, or even father than, financial data. (Typically, environmental data lagged financial data by one or more quarters.) This will free up time for sustainability managers to focus on more strategic efforts.
The increase in automation is changing reporting in a significant way. In the past, sustainability executives felt pressured to keep their reports short and sweet. Now companies are expanding the amount of data they offer. Some are supplementing the annual sustainability report and creating a separate ESG information site on the investor web page.
Think of it as a welcome mat for the bots.