Special issue on climate change and natural disasters
本特刊探讨气候变化加剧自然灾害风险对保险业的影响，包括保险定价、风险分担、政府补贴及监管问题，适合研究气候经济与保险的学者。
Climate change, with the associated rise in sea levels and extreme weather including more severe droughts and floods, stronger hurricanes, and higher average temperatures, has a profound impact on all parts of the economy. One direct consequence of rising natural disaster risks is an increased need for greater risk diversification and insurance for businesses and households to combat the rising costs of extreme weather events. Figure 1 shows that over the last 40 years the cost of natural disasters has quintupled. Besides the increased frequency and severity of climate events, the increase in construction in disaster prone areas has also played a fundamental role in magnifying natural disaster costs (Brakenridge et al., 2021; Brody et al., 2007). Despite an increasing awareness about climate change and its consequences, current projections suggest that efforts to reduce greenhouse gas emissions (holding the increase in global average temperatures below 1.5°C) are insufficient, and that the consequences of climate change are likely to intensify in the short- and medium-run. Human-related greenhouse gas emissions, which are the main source of higher average temperatures, have increased steadily in the last century (Figure 2). Commitments to reduce these emissions are partially binding only by 2030 and fully binding by 2050, as called for in the Paris Agreement. Similarly, recent technologies are still in their infancy, and it will take time before they can have a material impact on the climate change trajectory. The increase in climate change-related costs has already affected many dimensions of insurance (Issler et al., 2020; Keys & Mulder, 2020), and this trend is likely to continue. Some implications are easy to gauge, such as increases in catastrophe losses leading to higher property and casualty (P&C) insurance costs. Other, more indirect consequences could be harder to evaluate. Consider the extreme case in which some coastal structures might become uninsurable by P&C insurers because they are in soon-to-become permanently flooded areas. Or consider changes in temperature patterns towards more episodes of extreme heat that could increase mortality risk across all ages, potentially posing a burden to life insurers. Insurance markets play a natural role in helping businesses and households mitigate increased exposure to climate-related risks. You and Kousky (2024) demonstrate that greater insurance take-up rates influence postdisaster visitations to local commercial establishments. They find that “insured households are less likely to experience high financial burdens in both the short and longer run and are less likely to have unmet funding needs than those without insurance.” Nonetheless, recent research suggests that insurance markets may not function efficiently and hence are not optimally mitigating financial consequences of climate change that businesses and households are experiencing (see Jeziorski & Ramnath, 2021; Oh et al., 2022; Sastry et al., 2023; You & Kousky, 2024). Philippi and Schiller (2024) find, however, that government subsidies can support the development of private insurance, especially where government disaster relief has been available. Because disaster relief is often relied upon in instances where the market breaks down, we would anticipate the value of subsidies to be greatest in those instances, including when climate change brings new risk. As Philippi and Schiller state in their conclusions: “When people are not aware of risks and do not buy insurance, states are often pressured into making disaster relief payments when losses are large. The resulting anticipation of future disaster relief payments depresses insurance demand. Premium subsidies … are one potential policy change to eliminate charity hazard and to promote private insurance.” In addition to financial support provided by insurance markets, insurers can also play a crucial role in mitigation. Insurance design can help align incentives so that individuals internalize climate change consequences. For instance, homeowners' insurance makes individuals bear the cost of their geographic location decision, and pricing can provide incentives for further risk mitigation (e.g., hurricane-proofing of houses). The paper by Hinck (2024) published in this special issue considers the optimal insurance contract with probabilistic government disaster relief. That paper shows the propensity to get private insurance decreases when governemnt payments are expected, but the reduction is smaller when the governemnt aid depends on the magnitude of losses. Specifically, when payments are expected regardless of the loss size, a simple deductible contract is optimal, (and this result is robust to introducing ambiguity about the probability of government payments). But, when the probability of payment depends on the magnitude of losses, the optimal contract includes coinsurance. The introduction of coinsurance, reduces the incentives to relocate into risky areas, relative to a model without coinsurance. The results in Hinck (2024) might explain why subsidies, in the form of ex-post disaster payments, like those given by FEMA, might have reduced the propensity to pay for homeowners insurance, and might have caused a disproportionate historical settlements in flood prone areas (Rentschler et al., 2023). One concern with current insurance regulation is that the rate-setting process may result in price distortions and rigidities that preclude insurance companies from correctly pricing climate change-related risk into premiums. Therefore, insurers may not be able to cover the underwriting costs of policies (Born & Klimaszewski-Blettner, 2013). For example, in California—a state that recently experienced several severe wildfires—current insurance regulations forbid insurers from including reinsurance costs in their policies, and bar the use of forward-looking modeling (Insurance Information Institute, 2023), both of which would help insurers correctly price policies affected by climate change. In response to these restrictions, we have already seen the withdrawal of several major insurers (McDaniel, 2023). Similarly, the financial burden of climate change might be cross-subsidized from states with high price frictions to states with low price frictions, affecting the long-term allocation of insurance (Oh et al., 2022). Florida is often in the news regarding climate change. Fifteen Floridian insurance companies stopped writing new policies and seven have been declared insolvent between 2022 and the first half of 2023 (McDaniel, 2023). Some of these insurers have cited climate change risk as the main reason for their decision; others have stated the disproportionate lawyers' fees charged in the claims as the reason for their withdrawal (McDaniel, 2023). For example, out of the $51 million paid in insurance claims over a period of 10 years, 71% went to lawyer fees (Friedlander, 2024). Climate change will likely increase litigation, and absent a regulatory change, may trigger the exit of more insurers in a state with significant exposures to flooding and wildfires. The exit of long-standing insurers from high climate change risk areas has been in part filled by smaller and less solvent insurers as shown in Sastry et al. (2023), who show that these insurers meet Government Sponsored Enterprises (GSE) rating eligibility thresholds using emerging credit rating agencies that are lenient relative to traditional agencies. Lenders have responded to the decrease in insurance quality by selling a large portion of loans to GSEs. Loading (federal-level) GSEs with low insurance-quality mortgages can result in the cross-subsidizing of climate change costs from high- to low-risk areas. Even without price distortions, insurance markets might break down when insurance costs are above individual purchasing power. Consider the case of a household purchasing a home. Households are likely to be less able to predict the consequences of climate change in the medium run, and hence, future insurance premiums. While house buyers arguably consider the homeowners insurance cost at the time of purchase, after the purchase, an unexpected, rapid increase in insurance premiums stemming from the consequences of climate change may lead the household to no longer be able to afford insurance. In addition, households' beliefs over climate change can also lead to myopic behavior that can exacerbate the problem increasingrelocation into cheaper (of insurance costs) disaster-prone areas.1 One article in this special issue considers the climate change implications of balance sheet policies and risk management by insurers. The paper by Berry-Stoelzle et al. (2024) uses a textual analysis of insurers' climate disclosures from a mandatory survey to construct a novel climate risk management quality index. The authors document that natural disasters in an insurer's geographic market do not appear to lead to timely changes in risk management practices with one exception: the number of natural disasters in the home state leads to an increase in climate risk management. The findings of this study are therefore consistent with salience theory, which raises concerns about risk mitigation and risk management practices of insurers that are not frequently exposed to disasters in their home states. A similar paper—Chiang (2024)—studies how P&C insurers change their balance sheet policies after catastrophic events. Rather than the smaller events in Berry-Stoelzle et al. (2024), this study considers exposure to concentrated losses that can deplete insurers' capital, increasing insolvency events. The study documents that insurers mitigate capital losses due to a first catastrophic event by ex-post increasing capital more than unaffected insurers and increasing cash holdings to self-insure against large payouts in the future. At the same time, the author finds no evidence for increased risk-taking on the assets side, though the author also finds no evidence of increased risk mitigation through increased used of re-insurance. Taken together, these findings suggest that new experiences with large losses lead to a revision in balance sheet policies before a first catastrophic event. Surprisingly, the effect diminishes for the second and third catastrophic event. Affected insurers do not significantly increase their capital ratios or change their asset composition which, on net, suggests a larger risk of insolvency following subsequent catastrophic events. In addition to the balance sheet and risk management policies examined by these two papers, climate change could also affect such policies in other ways. For instance, larger catastrophes and changes in weather patterns could increase geographic correlation of insurance claims, making it more difficult for insurers to diversify risk.2 The reluctance of insurers to offer contracts exposed to correlated risk has been in part addressed by the creation of the Federal Emergency Management Agency (FEMA), but whether FEMA programs provide efficient risk-sharing has been an open question. One paper in this special issue, Boonen et al. (2024), studies pareto-efficient risk sharing in centralized insurance markets and characterizes ex-post efficiency of risk transfers in terms of minimizing the sum of the insured's risk positions. The authors apply their theoretical results to flood risk insurance, the National Flood Insurance Program (NFIP) run by FEMA that acts as a centralized provider of flood insurance. The authors assess the welfare gains of pooling specific geographically distant US states and find that centralized insurance exhibits considerable gains due to diversification. A different, market-based solution for increased losses and correlation of losses includes larger reliance on reinsurance. One caveat of more reliance on reinsurance is that a large enough catastrophe might be financially stressful even for a large reinsurer. The 1906 San Francisco earthquake and fire that resulted in losses of about 1% of the gross national product serves a cautionary tale. Losses of this catastrophe did not only wipe out decades of US insurers' profits, but also had to be partly covered by foreign insurers through Lloyd's of London, resulting in large capital outflows of the United Kingdom (Odell & Weidenmier, 2004). An alternative to reinsurance that has emerged is catastrophe (CAT) bonds. These bonds allow issuers to diversify catastrophe insurance costs among a larger set of financial market participants (Polacek, 2018). So far, the contribution of CAT bonds to risk mitigation has received somewhat limited attention. Notable exceptions include Cummins (2008), who surveys the development of the CAT bond market, and Lakdawalla and Zanjani (2012), who show that the full collateralization of the underlying risk transfer in CAT bonds limits their market penetration but may still be an improvement over reinsurance if contracts do not specific all state-contingent payouts. One paper in this special issue, Li and Su (2024), contributes to this strand of literature by studying the effect of CAT bonds for insurers providing wildfire coverage. They show that CAT bonds linked to the insurer's losses can substantially reduce volatility and tail risk. Notably, they show that CAT bonds whose payments are linked to aggregate wildfire losses rather than the specific losses of the issuer are also effective in reducing risk. These types of bonds have the advantage of better liquidity and access to a broader investor base. Insurers are also potentially exposed to climate risk through their asset side. In addition to a considerable exposure to commercial real estate mortgages and the associated climate risks, insurers are the largest investors in corporate bonds. Hence the insurance industry is also affected by climate change transition risk. Different from physical risk (the destruction of assets), transition risk includes when assets become less productive or, in the extreme case, turn into stranded assets. Concretely, proposals to achieve a carbon neutral (net zero) economy often involve carbon taxes that reduce firm profits and increase the risk of bankruptcy, potentially impairing returns. Ozdagli and Wei (2024) show that after the introduction of the mandated Climate Risk Disclosure Survey, affected insurers reduced their exposure to high-climate risk corporate bonds but, on average, still have a significant exposure to high-climate risk sectors. There are myriad unanswered questions, both from a policy and a research perspective. What is the optimal government involvement in climate-related insurance provision? In the United States, there is a strong trade-off among humanitarian considerations, economic incentives, and crossed-individuals and crossed-states subsidies. What is the optimal regulatory framework to incentivize risk sharing and achieve optimal pricing? What can we learn about households' beliefs and the level of desired insurance, and how does this level compare to optimal insurance? How should CAT bonds be structured to maximize risk sharing? Are there other risk-sharing agreements that may mitigate the financial impact of large catastrophes? Are reinsurers sufficiently diversified, or have changes in correlations reduced the benefits of reinsurance? The effects of climate change will impact us for generations. In the meantime, the insurance industry will contribute to aligning incentives, sharing risk, and mitigating the social consequences. Doing so efficiently will require a deep understanding of the questions raised above. While the papers in this special issue are first steps in providing some answers, we hope this issue fosters more high-quality research that improves our understanding of the consequences and potential solutions of climate change for the insurance sector.