The Profit vs. Planet Paradox: Unraveling the Corporate Tradeoffs with AI

As crazy smog envelops New York City, it seems very appropriate to revisit the human impact on climate. As a financial researcher, I am always particularly interested to see how investors can influence the markets, in this case applied to the latest climate developments, perhaps by encouraging certain technologies over others. As an Artificial Intelligence investigator, I am always fascinated by what truly autonomous AI can tell us about the directions we are heading in and the outcomes these initiatives can bring.

As detailed in a recent research paper I had a privilege to co-author with the super talented Payton Martin, there are multiple initiatives that are regularly deployed and reported upon by public companies. The data these initiatives collectively accumulate is a fertile ground for analysis of application and impact of some techniques over others.

In our research, we find that many public U.S.-based companies experience a divergence between the management actions and the management rhetoric. This is really not such a surprising finding. Specifically, we find that the corporate management loves to talk about three broad dimensions: 1) diversity, 2) hazardous materials and 3) greenhouse gasses. The statements along these three dimensions tend to powerfully impress investors and give a temporary boost to the companies’ stock price.

When we apply AI to analyze the management actions, however, we find that the management tends to lump together the issues of greenhouse gasses, data security and inclusion and treat them as the same necessary corporate prophylactic. On the issue of emissions, our AI models suggest that the corporations treat corporate emissions as a tradeoff to product quality, with product quality often winning this battle. Finally, the management spends a lot of time finagling the product labeling to get the messaging exactly right to reflect their preferred party line. What is most interesting is that once these AI inferences are taken into account, the temporary price impact of the management environment-friendly rhetoric disappears. The price changes in the longer term are then really driven by the management actions. In other words, while investors may be encouraged by the corporate CEO’s latest round of promises, it is the corporate actions that matter in the end.

What are the key implications of this analysis? First, we can clearly see the corporate tradeoff for profits vs. environmental externalities, especially in the emissions vs. product quality dilemma. Second, regulations do matter, and corporations do follow them, if reluctantly. This explains the corporations tendency to bag together greenhouse gasses, data security and inclusion as a lump regulatory issue to deal with. Most importantly, we see that AI can really help investors to navigate through corporate marketing and language embellishments to clearly identify the corporate actions, separating them from the rhetoric.

Of course, the results we have obtained hold on average across a vast sample of companies. What we can do, however, is to accurately rate every public company on their 1) rhetoric, and 2) actions scales as deemed by AI. Since, per our analysis, the actions are much more predictive of the longer-term returns, the investors may indeed be very interested in following these. After all, who wants to look at the orange sky all day?

Irene Aldridge will be discussing the subject of AI in ESG at the upcoming Responsible Asset Owners' Symposium taking place on June 28, 2023, in New York City.