Earnings Calls Reveal How Firms Really View Climate Risk

When the CEO Says “Climate,” the Market Listens

In 2019, the CEO of a major European energy company stood in front of a room of analysts and said the word “climate” 14 times in a single earnings call. Two years earlier, he had said it twice. The analysts didn’t just hear a change in vocabulary. They heard a signal. Within weeks, the company’s stock options started pricing in a new kind of risk, one that had nothing to do with oil prices or geopolitical instability.

That moment, captured in a dataset of 10,000 firms across 34 countries, is exactly what Zacharias Sautner, Laurence van Lent, Grigory Vilkov, and Ruishen Zhang set out to measure in their 2023 paper for The Journal of Finance. They didn’t ask CEOs to fill out surveys. They didn’t comb through sustainability reports. Instead, they listened to what companies say when they think no one is keeping score. Or rather, when they know everyone is.

Earnings calls are strange rituals. Executives and analysts speak in code, hedging every claim with caveats, burying bad news in jargon. But Sautner and his coauthors found something unexpected: when companies talk about climate change on these calls, they aren’t just performing. They are revealing what they actually plan to do.

What the Algorithm Found That Humans Missed

The authors developed a machine learning keyword discovery algorithm that doesn’t rely on a prewritten dictionary of climate terms. Instead, it scans transcripts of earnings calls and identifies words that appear more frequently in calls where climate is already being discussed. The algorithm then learns to spot climate talk even when the specific vocabulary shifts. This matters because companies rarely say “climate change” directly. They talk about “regulatory uncertainty,” “carbon pricing,” “weather volatility,” and “green technology investments.”

The method captured three distinct types of climate exposure:

• ▸Opportunity exposure: How much a firm stands to gain from the transition to a low-carbon economy.
• ▸Physical exposure: How vulnerable a firm is to floods, fires, storms, and heatwaves.
• ▸Regulatory exposure: How much a firm’s business model depends on or is threatened by climate regulations.

Between 2002 and 2020, the authors tracked these exposures for over 10,000 firms in 34 countries (Sautner et al., 2023). The data reveal a clear pattern: firms that talk more about climate opportunities are the ones that actually invest in green technologies. Firms that talk only about regulatory risk tend to do nothing.

Why Earnings Calls Are Better Than Surveys

Most research on corporate climate action relies on self-reported data. Companies fill out questionnaires for CDP (formerly the Carbon Disclosure Project) or release glossy sustainability reports. The problem is obvious: companies have every incentive to look good. They can claim to care about climate without changing a thing.

Earnings calls are different. They are legally binding. Executives can be sued for misleading statements. Analysts ask pointed questions. The conversation is recorded and transcribed. It’s the closest thing we have to a corporate brain scan.

Sautner and his team validated their measure against real outcomes. They found that firms with higher climate opportunity exposure subsequently created more jobs in disruptive green technologies and filed more green patents (Sautner et al., 2023). In other words, when executives talk about climate as a business opportunity, they actually follow through. When they only mention it as a risk, they don’t.

The Market Already Knows

Here is where the paper gets genuinely spooky. The authors found that their climate exposure measures contain information that is already priced into options and equity markets (Sautner et al., 2023). This means traders are listening to these calls too. They are betting on which companies will adapt and which will get crushed.

Consider a utility company that spends 10 minutes of its earnings call discussing investments in solar and battery storage. The algorithm flags that as high opportunity exposure. The options market, in turn, starts pricing in lower volatility for that stock. Investors are effectively saying: “This company has a plan. We trust it.”

Now consider a mining company that spends 10 minutes complaining about carbon taxes and regulatory uncertainty. The algorithm flags that as high regulatory exposure but low opportunity exposure. The options market prices in higher volatility. Investors are saying: “This company is stuck. We don’t trust it.”

The authors don’t claim that markets are perfectly efficient. But the data suggest that climate talk is not cheap. It moves money.

What the Algorithm Missed

The method is clever, but it has limits. The algorithm can’t distinguish between genuine commitment and sophisticated greenwashing. A CEO might talk extensively about climate opportunities because the company is actually pivoting, or because they know analysts want to hear it. The algorithm picks up the signal but not the intention.

The authors acknowledge this. They write that their measure captures “attention paid” to climate exposure, not the quality of that attention (Sautner et al., 2023). A firm that talks a lot about climate could still be doing nothing. The data show that, on average, firms that talk more about opportunities do more. But averages hide individual cases.

There is also a geographic bias. Firms in Europe and North America appear more frequently in the dataset. Companies in developing countries, where physical climate risks are highest, are underrepresented. The algorithm works best where earnings calls are common and transcripts are available. That leaves out a lot of the world.

The Three Flavors of Climate Talk

The authors’ breakdown of climate exposure into opportunity, physical, and regulatory categories is more than a taxonomy. It reveals a strategic logic.

Opportunity Exposure: The First Mover Signal

Firms that score high on opportunity exposure are not just talking about climate. They are talking about specific technologies, markets, and business models. They mention “green hydrogen,” “electric vehicle charging networks,” “carbon capture,” and “sustainable aviation fuel.” These are not buzzwords. They are investment signals.

The authors found that a one standard deviation increase in opportunity exposure predicts a significant increase in green patenting in the following year (Sautner et al., 2023). This is not correlation. It is causal direction: the talk comes before the patents.

Physical Exposure: The Silent Sufferer

Physical exposure is harder to talk about. Firms don’t like admitting they are vulnerable. A real estate company might mention “extreme weather events” without specifying that half its portfolio is in flood zones. An agricultural firm might discuss “crop yield variability” without saying that droughts are already cutting profits.

The algorithm picks up these euphemisms. It finds that physical exposure is most common in industries like agriculture, insurance, and real estate. But it also finds something surprising: firms with high physical exposure do not necessarily invest in adaptation. They talk about the problem but often fail to act.

Regulatory Exposure: The Complainer’s Trap

Regulatory exposure is the most common form of climate talk, but it is also the least predictive of action. Firms that complain about carbon taxes, emissions standards, and reporting requirements tend to have lower green patenting and fewer green jobs (Sautner et al., 2023). They are stuck in a defensive posture.

This makes intuitive sense. If a company views climate regulation as an external threat rather than an internal opportunity, it will lobby against change rather than invest in it. The paper provides empirical evidence for what activists have long suspected: the loudest complainers about climate regulation are often the ones doing the least to adapt.

What This Means for Investors

The paper has direct implications for anyone managing money. If you are a portfolio manager, you can now track climate exposure in real time. You don’t have to wait for annual sustainability reports. You can listen to earnings calls and measure whether a company is positioning itself for the transition or digging in its heels.

The authors show that their measures predict stock returns and option prices (Sautner et al., 2023). This is not a niche finding. It suggests that climate risk is already being priced, but only for firms that talk about it clearly. Companies that stay silent may be hiding bigger risks.

What This Means for Policymakers

Regulators are trying to force companies to disclose climate risk. The SEC, the EU, and the UK have all proposed mandatory disclosure rules. This paper suggests that voluntary disclosure, at least in earnings calls, already contains useful information. The problem is that it is unstructured and hard to compare across firms.

A regulator could use the authors’ method to score every public company on climate exposure. That would create a standardized metric that investors could trust. It would also reveal which companies are talking the talk without walking the walk.

What This Does Not Prove

The paper is careful not to overclaim. It does not prove that climate talk causes green investment. It could be that companies planning to invest in green technology simply talk about it more. The authors use statistical methods to address reverse causality, but they cannot fully rule it out.

The paper also does not prove that markets are correctly pricing climate risk. It only shows that markets react to the information in earnings calls. If the calls themselves are misleading, the market reaction could be wrong.

Finally, the paper does not address the biggest question: will firms actually reduce emissions? Green patents and job creation are proxies, not guarantees. A company could patent a green technology and never commercialize it. It could create green jobs while expanding fossil fuel operations elsewhere.

What This Actually Means

• ▸Listen to what executives say about opportunity, not risk. When a CEO talks about climate as a business opportunity, they are more likely to invest. When they only mention regulation, they are likely stalling. This is a usable signal for investors and analysts.

• ▸Earnings calls are underutilized data sources. The authors show that unstructured text from calls predicts real outcomes better than structured surveys. Any serious analysis of corporate climate action should start with transcripts, not questionnaires.

• ▸Greenwashing is detectable, but not perfectly. The algorithm catches talk but not intent. A firm that talks about opportunity could still be greenwashing. But on average, talk predicts action. The burden of proof should shift: if a company talks about climate opportunity, it should be expected to show results.

• ▸Regulators should standardize climate exposure metrics. The authors’ method could be scaled to score every public company. This would give investors a consistent, comparable measure of climate risk. It would also expose firms that are silent on climate, which may be hiding the biggest risks.

• ▸The market is already pricing some climate risk. Options and equity markets react to the information in earnings calls. This means climate risk is not a future problem. It is already moving capital. Firms that ignore it are not just harming the planet. They are hurting their own stock price.