Irrational Economist_ Making Decisions in a Dangerous World - Erwann Michel-Kerjan [111]
ALTERNATIVE DISPOSITIONS OF STAGE 1 RISK
Stage 1 risk is uninsured, and stage 2 risk is insured through short-term insurance contracts. This is “business as usual”; insurers continue to offer year-by-year contracts with premiums based on updated estimates of the risk level. But even if contracts continue to look as they do at present, climate change presents enormous challenges to insurers. Increasing risk will require more capital, and climate instability will lead to problems in estimating contemporaneous loss distributions and therefore difficulties in setting premiums. Re-estimating loss distributions based on rapidly evolving hazard levels could lead to violent premium changes, and this premium volatility is a risk that policyholders face. Compounded difficulties would arise if policyholders and insurers differed in their assessment of the likelihood of the risk, which, in turn, could lead to excess supply or demand of insurance. As mentioned earlier, additional challenges will arise if climate change leads to changes in the correlation structure of the world insurance portfolio. (For example, El Nino/La Nina currently present correlated weather patterns across widely separate parts of the world, but we know little about how these would change with global warming.)
Stage 1 risk is insured in constant premium long-term insurance contracts. Can insurance cover both the stage 1 and stage 2 risks? A paper by Howard Kunreuther, Erwann Michel-Kerjan (at Wharton), and Dwight Jaffee (at UC Berkeley) advocates long-term insurance in which much of the premium risk is absorbed by the insurer.6 The appeal of this proposal lies with the demand side of the market—risk-averse homeowners who would clearly benefit from offloading the joint risks of actual losses, volatile premiums, and the future availability of insurance.
The problems lie on the supply side. This long-term insurance proposal cleverly links insurance to preexisting long-term contracts that are rooted in value (i.e., mortgages). In doing so, they acknowledge that lenders already accept a long-term risk position insofar as the revealed risk (either decline in house prices or uninsured losses) will affect their collateral. Could one go further and formally establish a market for long-term insurance contracts that offer some protection against premium volatility (whereby, for example, premiums and coverage remain constant over the whole period of the contract, or are guaranteed for sub-intervals such as three years, five years, etc.)?
There is a precedent. Life insurance is routinely sold on terms of twenty, thirty (etc.) years with uniform premiums. But comparing life insurance with long-term catastrophe insurance can be alarming. The three conditions that facilitate long-term life insurance—predictable risk, uncorrelated trends in risk, and a beneficial trend in risk (declining mortality)—are unlikely to be present in the context of climate hazards. If, as one fears, an upward trend in catastrophe risk is associated with (serially and contemporaneously) correlated ambiguity, then the stage 1 risk would not be amenable to diversification. In this event, the amount of insurer capital required to secure that risk at an acceptable credit rating could be enormous. On the other hand, if the pool was not adequately prefunded with capital, then credit quality would deteriorate perhaps to the level of unacceptable counterparty risk.
I will mention another issue that is addressed by the long-term insurance proposal: Unless such contracts are binding on policyholders, then competing insurers offering short-term contracts can pick off those policyholders for whom risk might decrease. Such slippage will, of course, diminish the degree of risk