One of the most pressing questions facing today’s climate research is whether climate change risks are reflected in stock prices. In a peer-reviewed study* recently accepted for publication in Journal of Banking and Finance, we found that investors only care about climate change risks when policymakers intervene, not about physical climate risks.
Our results also suggest that investors hedge against the realisation of imminent transition risks, and that their hedging portfolios do not necessarily contain green firms.
To the best of our knowledge, our study is the first to shed light on the question of which climate change risks, if any, are priced in securities markets from a market-wide perspective. Our study has been presented to policymakers, supervisors and regulators, including the European Parliament’s Future of Science and Technology Panel, the European Banking Authority, EIOPA, and ESMA, the EU’s financial markets watchdog. It has been cited in the annual report of the Task Force on Climate-Related Financial Disclosures and in the OECD report on Climate-resilient Finance and Investment (released as OECD Policy Perspectives, which form policy recommendations). The US Securities and Exchange Commission has posted our study on its website as a comment to its March 2022 proposal on climate-related disclosures for investors.
We based our analysis on news articles about climate change published by Thomson Reuters, with our initial sample consisting of more than 13 million articles published from 1 January 2000 to 31 December 2018. We disentangled our textual data to form individual proxies of physical and transition risk, doing so by employing state-of-the-art methods from natural language processing. These methods are unsupervised, meaning that it is the method itself, rather than the user, that performs the disentanglement. This is in stark contrast to dictionary-based methods such as sentiment analysis, where the user essentially defines the proxies in advance by manually specifying which words are most likely to characterise them.
Applying this method to our textual data revealed that climate change risks can be decomposed in the following four factors or ‘topics’: natural disasters, global warming, international summits, and the US climate policy debate. The first three are inherently longer-running physical and transition risks, elicited by regulatory risks stemming from international summits and news on the scientific evidence surrounding global warming and the occurrence of natural disasters. The fourth is a comparatively shorter-term transition risk as it encompasses events such as presidential addresses, the enactment of laws, and appointments to key regulators.
As a sanity check, we computed the average daily occurrence of climate-related news articles per topic and, as expected, spikes in these time series coincide with relevant and significant events. This strongly corroborated that our factors serve as measures for the underlying risks.
Armed with our new factors, we were then able to assess whether investors are aware of climate change risks. We did so by a standard portfolio-sorts approach, where we started by sorting stocks into portfolios based on their sensitivity to each factor. We then formed a long-short value-weighted spread portfolio consisting of going long in the portfolio that includes stocks with the highest climate beta and going short in the portfolio that includes stocks with the smallest climate beta. We examined whether this spread-portfolio yields a statistically significant abnormal performance after controlling for common equity factors. If it does, this would suggest that the climate risk proxied by the specific factor is priced.
Intertemporal hedging explains risk premium
A clear picture emerged after running a multitude of tests: Only securities that are more exposed to risks deriving from the US climate policy debate tend to pay a significantly higher return over those less exposed. Specifically, the average return after controlling for all risk factors is up to approximately six percentage points larger per annum. This is the climate change alpha that institutional investors can reap by going long (short) in stocks with high (low) climate betas estimated with respect to our risk measures. But our findings also mean that substantial sources of longer-running climate risks stemming from natural disasters and global warming may not be adequately priced, nor decisions taken at international summits on climate, as these decisions are not binding.
Interestingly, in contrast to what one would expect, we find that investors do not necessarily invest in ‘green’ firms to hedge climate policy risk, but they may also be using ‘brown’ firms, as the latter can be substantial innovators of green technology.
It is natural to ask why physical risks are not reflected in stock prices in the first place. There are several plausible explanations.
One possibility is limited attention from financial investors. Under this view, the US political arena serves as a ‘wake up’ call. A second possibility is that investors lack information on the exposure of businesses to physical climate change risks. This view seems to be shared by several institutions and financial regulators and it is inspiring new regulation on the disclosure of climate risks. A third possibility is that financial investors are myopic, in that they are only focused on risks that have immediate financial effects. All three explanations imply mispricing of physical climate risks, and they are in line with the view expressed by a number of policymakers that policy intervention is required to address the market failure underlying this mispricing.
Another very different potential explanation for our results is that climate change per se does not pose a material financial risk, and hence is not expected to be priced. In contrast, any government intervention is expected to be priced, yet it can threaten financial stability by stranding assets and hurting firms’ profitability.
Disentangling these different views was beyond the scope of our study, but research must pursue this question. Incorporating financial and non-financial risks in asset prices is of paramount importance to institutional investors and policymakers alike. This is a lesson we should have learnt by now, given that it was mispricing of mortgage-backed securities in the early 2000s that played a key role in fostering the conditions that eventually led to the Great Recession.
Renato Faccini is deputy head of research at Danmarks Nationalbank; Rastin Matin is senior quantitative analyst at PFA Asset Management; George Skiadopoulos is professor of finance in the School of Economics and Finance at Queen Mary University of London and in the department of banking and financial management, University of Piraeus