Climate Finance | NBER #NewsUnitedStates

Climate change poses a critical global challenge, impacting not only the social and geopolitical spheres but also economic activity, asset values, and financial stability. Addressing this complex issue requires a multisector approach, with financial markets playing a crucial role in advancing a sustainable transition. Financial markets can mitigate climate risks by channeling resources into sustainable activities and green innovation while also facilitating adaptation through the sharing of climate risks. The emerging field of “climate finance” studies these and other contributions of the financial sector to climate solutions. This article reviews key findings from our recent work in this area.

Climate Change and Financial Markets

For financial markets to help address the climate challenge, market participants must recognize and respond to climate risks. In turn, this responsiveness by market participants should lead to the integration of climate risk information into asset prices. Our research in this space examines how asset prices currently reflect, and should ideally reflect, climate risks.¹

Our work’s first contribution is to provide a benchmark theoretical model that illustrates the channels and mechanisms through which climate change can affect asset prices. The model features an integrated description of the dynamics of both climate change and the economy, thus incorporating the economy-climate two-way feedback introduced by integrated assessment models. It captures how economic activity generates CO₂, leading to temperature increases and a higher probability of a “climate disaster” scenario (similar to a tipping point that is central in the climate literature). The model also features an asset (calibrated to represent the housing market), which reveals the effects of climate change on asset pricing.

The theoretical analysis emphasizes the importance of understanding the drivers of climate risks and uncertainty for evaluating their pricing implications. One possible scenario considers climate uncertainty as predominantly driven by economic uncertainty. In this view, strong global economic growth leads to increased carbon emissions and potentially severe climate damages, while economic slowdowns naturally reduce emissions and the likelihood of climate damages. This scenario, common in many climate-economics models, suggests that climate damages would occur in favorable economic conditions where it would be easier for society to bear them. Consequently, climate risk would only be moderately bad for agents, which reduces the equilibrium climate risk premium in financial markets.

However, if the fundamental uncertainty is directly about the dynamics of the climate—such as uncertainty about sea level rise, the timing of tipping points, or the economic damage of weather shocks — then climate change can itself become a key driver of economic growth. In that case, the realization of a bad climate scenario could trigger a substantial economic decline, and climate damages would instead coincide with “bad” economic states. As a result, investors would perceive climate risks as particularly threatening and therefore demand high risk premia for exposure to those risks.

This model delineates the channels through which asset prices can move with climate risks and provides a framework for interpreting climate risk premia estimated in the literature in terms of climate exposures and investor preferences. It also emphasizes the importance of this modeling for choosing discount rates to apply to investments that mitigate the effects of climate change. Mitigation investments reduce the probability and damages from climate change and, therefore, act like insurance against climate damages. As finance theory tells us, the appropriate discount rate for these investments will contain a risk premium that will be the opposite of the risk premium associated with the risk to be mitigated: calibrations in which climate risk coincides with bad economic outcomes and where it commands a high risk premium will also imply that very low discount rates apply to climate mitigation and adaptation investments (because they will earn a negative risk premium).²

Pricing Climate Risks

A direct implication of different climate risk theories is that risk exposures should be reflected in asset prices: everything else equal, assets with higher exposure to climate risk should command a lower price. We test this implication by focusing on the pricing of real estate in Florida, New Jersey, North Carolina, and South Carolina.³ We analyze property vulnerability to sea level rise by combining transaction-level house price data with flood risk projections from the National Oceanic and Atmospheric Administration for a six-foot sea level increase.

Identifying climate risk pricing in housing markets is challenging as coastal properties — often at risk from sea level rise — have other characteristics affecting their value, such as nicer ocean views or beach access. To address potential confounding factors in valuing vulnerable properties, we developed a “climate attention index” based on the proportion of climate risk mentions in real estate listings by zip code and year. Using a difference-in-differences approach to control for other property characteristics, we compute changes in price differences between exposed and unexposed properties when attention to climate change changes. We find that properties exposed to sea level rise face more significant (relative) price discounts in markets with high climate risk awareness, indicating a direct climate risk effect on house prices. Additionally, we observe that, conditional on our controls, there is no difference in annual rents between exposed and unexposed properties, suggesting that the observed price impacts likely reflect concerns about future climate risk realizations rather than current property damages, which would also be reflected in current rents. Overall, these findings highlight that climate risks already substantially affect real estate valuations.

We have also conducted a similar analysis studying the pricing of climate risk in equity markets.⁴ Firms face two main types of climate risk exposure: physical risks, where extreme weather events disrupt operations and supply chains, and transition risks, mostly related to regulation aimed at pushing towards a net-zero economy. Our theoretical models predict that both types of risk should influence stock prices.

In our work, we find evidence that, indeed, stocks of firms that are more exposed to these risks have relatively lower returns in periods when bad news about future risk realizations becomes available. This finding suggests that equity investors are already considering the possible effects of climate risk realizations when valuing and pricing stocks.

Beyond real estate and equities, there is growing evidence that climate risk exposures are priced across a wide range of asset classes, from municipal and corporate bonds to mortgage-backed securities to options. We have reviewed some of this evidence in our recent work.⁵

While researchers have been able to test and reject the null hypothesis of “no pricing of climate risks,” a much more difficult question is whether climate risks are priced adequately: does the current pricing of climate risk correctly reflect the climate risk exposures of different assets? This question requires researchers to adopt a stance on the likelihood of various types of physical and transition risk realizations, their impacts on cash flows, and the appropriate discount rates for different states of the world. Ideally, future work building on integrated asset pricing/climate models like the one we have discussed above will lead to quantitative answers to such questions.

While research to provide this information is ongoing, our recent research has addressed this question from a different angle. We survey professional economists, investors, and policymakers and find that a large majority of respondents believe that climate risks are not yet fully priced in real estate and equity markets.⁶

The Management of Climate Risks: Hedging and Stress Testing

Given that climate risk is priced in asset markets, a natural question is whether investors can exploit this fact to reduce their exposure to these risks.

Ideally, investors could hedge climate risk by purchasing long-term insurance-like securities whose payoffs are directly linked to climate outcomes. However, such instruments do not currently exist. Our results on the pricing of climate risk, however, suggest an alternative approach: investors can create synthetic insurance by constructing portfolios long on low-climate-exposure equities and short on high-exposure ones. The key idea is that this strategy would earn returns if climate risks materialize, as highly exposed companies would lose value relative to less exposed ones.

For such a portfolio to succeed, frequent rebalancing to hedge against evolving climate risk information (e.g., news about long-term risks) over time is required. Building on this idea, we construct equity portfolios that best hedge the high-frequency arrival of climate news.

We construct an empirical proxy for the arrival of climate news. We do so by creating a new index of climate news using textual analysis in the Wall Street Journal.⁷ Figure 1 shows the news series from the Wall Street Journal, which peaks concurrently with important climate events.

Our research explores different methods to construct portfolios that would hedge against innovations in this news series. In our first paper, we propose forming long-short portfolios based on firms’ environmental, social, and governance (ESG) scores as proxies of their climate risk exposures. While portfolios tilting towards stocks with high ESG scores (and against stocks with low ESG scores) showed positive correlations with negative climate news, these correlations were small and unstable across time horizons, possibly due to well-known data quality issues with available ESG scores. Statistical approaches inferring climate risk exposures from past correlations of asset prices with news realizations also proved unreliable, hampered by short time series and relatively infrequent aggregate climate news.

Our more recent work proposes a new approach to determining optimal hedge portfolios against climate news. This method combines information on individual traders’ idiosyncratically varying climate change concerns — measured, for example, by the tone of their climate risk discussions in investor disclosures — and their trading responses.⁸ The premise is that assets bought by investors as their climate risk concerns increase should do well when aggregate climate news materializes. As more investors become concerned and demand those assets, this increased demand will push the price up, increasing returns for the climate risk hedging portfolio that had held those assets. Our work shows that this approach tends to outperform traditional methods of hedging aggregate climate news.

While financial investors can manage exposures to climate risk realizations using the approaches described above, banks and financial regulators focus on understanding the risks to financial stability through banks’ loan books. Globally, regulatory stress tests and “scenario analyses” have become widespread, requiring banks to assess the impacts of particular climate risk scenarios (e.g., a hurricane hitting the Northeastern US) on their loan books. Given the importance of these types of stress tests in managing the effects of climate risk on financial stability, we advocate for more academic research on optimizing the methodologies of scenario design and stress tests.⁹

Retail Investor Views of Climate Risks

The work in the previous section highlights that various assets — including traditional “green” assets such as renewable energy firms — have returns that covary with news about climate risks. According to standard asset pricing theory, these stocks should have lower expected returns than the aggregate market, with the lower returns approximating an “insurance premium” that investors are willing to pay due to the assets’ covariance with aggregate climate risks.

In recent work, we consider whether retail investors understand these trade-offs and whether they are still willing to hold these assets even if they understand their return properties.¹⁰ We collaborated with Vanguard to survey retail investors about their motives and expectations regarding ESG investments and linked these to their portfolio holdings. We document large heterogeneity both in the motivations retail investors associate with ESG investing (e.g., ethical motives, climate-hedging motives, or pecuniary motives) and in the returns they expect from these investments. For example, Figure 2 reports the histogram of the annualized 10-year returns expected by our respondents for ESG investments, in excess of the market return, across all survey waves and all respondents. On average, investors in our sample expect significant (2 percent per year) underperformance of ESG investments relative to the market, consistent with standard theories about equilibrium returns. What is also striking is the considerable heterogeneity in expected excess returns visible in Figure 2.

In addition to studying investors’ beliefs about ESG investments, we also explore how ESG motivations and expectations are linked to actual ESG investment behavior. Most interestingly, we find that investors’ portfolio choices strongly reflect a trade-off between nonpecuniary motivations and pecuniary ones (expected returns from ESG investments). While it is expected that investors would invest more in ESG assets when they anticipate higher returns, this preference remains strong even for investors who argue for ESG investments for nonpecuniary reasons. Figure 3 shows that holdings of ESG investments drop very significantly when investors are pessimistic about ESG returns. Furthermore, Figure 4 highlights that this trend holds even for investors who report ethical reasons to hold ESG investments. These results underscore the importance of understanding the full range of motivations for ESG investments (both pecuniary and nonpecuniary) in order to better comprehend investor behavior and inform climate-related policies.

Biodiversity and Nature Risks — The Next Frontier

Much of the research on the interaction between nature and economic activity has focused on the role played by climate change. However, a similarly large but less well understood risk to economic activity comes from the significant and ongoing loss of nature and biodiversity over the past decades. While biodiversity loss and climate change are heavily correlated, biodiversity loss has distinct effects on the economy. In survey data, we find that these risks are an increasing source of concern for investors and regulators around the world.

Unlike climate risks, where the channels through which physical risk affects economic activity are relatively clear, the mechanisms through which biodiversity loss, such as species extinctions, affect economic output are less well understood. Part of the challenge is that economic models typically consider nature as a single, monolithic “stock of natural capital” within the aggregate production function rather than account for its diverse and complex configuration.

In recent work, we thus develop a new ecologically founded model of the economic effects of biodiversity loss that explicitly considers the interaction of different species in the production of the aggregate “ecosystem services” that enter the production function.¹¹

In our model, aggregate ecosystem services are produced by combining several non-substitutable ecosystem functions such as pollination and water filtration, each provided by many substitutable species playing similar roles. As a result, economic output is an increasing and concave function of species richness.

The marginal economic value of a species depends on three factors: (i) the number of similar species within its ecosystem function, (ii) the marginal importance of the species’ affected function for overall ecosystem productivity, and (iii) the extent to which ecosystem services constrain economic output in the economy. Using our framework, we derive expressions for the fragility of ecosystem service provision and its evolution over time, influenced by the distribution of biodiversity losses across ecosystem functions.

We discuss how these fragility measures can help policymakers assess the risks induced by biodiversity loss and prioritize conservation efforts. We also integrate our model of ecosystem service production with a standard economic model to study optimal land use when land use raises output at the cost of reducing biodiversity. We find that even in settings where species loss does not reduce output substantially today, it lowers growth opportunities and reduces resilience to future species loss, especially when past species loss has been asymmetric across functions.

There is increasing evidence that investors are recognizing and pricing assets’ exposure to biodiversity risk.¹² Along with our work on climate risk, we have developed a measure of negative aggregate news about biodiversity loss. We make this time series, along with other data, available at www.biodiversityrisk.org.

At the country level, we observe that credit default swap (CDS) spreads move with negative realizations of biodiversity news. Consistent with our model described above, this effect is more pronounced in countries with more depleted ecosystems. These findings highlight that investors in CDS markets appear to appreciate that biodiversity loss affects economic tail risk probabilities for countries.

We also find that biodiversity risk affects US stock prices. In our analysis, we measure the biodiversity risk exposures across different firms and industries using information from firms’ 10-K statements. We further demonstrate that portfolios that underweight firms that are negatively exposed to biodiversity risks increase in value upon the realization of negative biodiversity news, thus providing investors with a new approach to constructing biodiversity-hedge portfolios akin to the ones aimed at hedging climate risks that we explored in previous research.