Catastrophes are extreme events. They are extreme in that they impose huge losses. (This is what affords them such media attention.) However, catastrophes are also extreme in that, by our fortune, they are infrequent. Elementary arithmetic shows that an insurance industry, which was able to take the long view, could spread a large infrequent risk (say a 20-year, $50 billion, event) over time with annual premiums no less affordable than auto insurance.
Principal justification or not, it is clear that private insurance markets are currently having a difficult time providing coverage for catastrophe risk. Public officials now take it for granted that if catastrophe insurance is to be available at all, it must be provided by a public agency either State or Federal. Catastrophes have thus become what insurance textbook writers call an ‘uninsurable risk’.
Is Catastrophe Risk Uninsurable?
Let us begin by considering the factors that cause market participants to consider catastrophe risks uninsurable. There seems to be no agreed upon definition of an uninsurable risk. The insurance literature, however, often identifies three factors which are viewed as impediments to the successful operation of a private insurance market:
Problems of selection and moral hazard;
The insured risk is ‘too large’ in some sense;
The probability of loss is not susceptible to precise actuarial calculation.
The practical problem for the reinsurance industry is that catastrophic risks have reached record values. Currently, reinsurers face two difficulties in underwriting such coverage: volatile climate conditions and the inapplicability of the law of large numbers in predicting catastrophe losses.
Climatic Conditions : In recent years, the increase in weather volatility has heightened the difficulty in predicting catastrophic property losses, rendering standard actuarial tables unreliable. Such climate volatility is often associated with global climate change. This is blamed by some on the emission of greenhouse gases into the planet’s atmosphere by organization for Economic Cooperation and Development (OECD) countries, the group of the world’s most industrialized nations.
If so, volatility is only going to increase. The problem is not likely to go away. As climate becomes more volatile, actuarial tables have become unreliable, creating the risk of using the wrong table for predictive purposes For example, one table could predict five hurricanes over a 5-year period, with average strength and associated loss of $3 billion each. Another, equally reliable, table could predict 10 hurricanes during the same period, having losses of about $2 billion each. Although one can take an average of the two source’s opinions in creating a new actuarial table, this does not work in actual practice. If both scenarios are equally plausible, for example, then taking the average guarantees that, most of the time, the exposure to risks will either be over insured or underinsured. Both lead to costly risks. The former leads to financial losses since insurance is expensive. Underinsurance is even more costly; underinsurance leads to financial risks of default.
Inapplicability of the Law of Large Numbers : The second problem associated with predicting the incidence of catastrophic property loss is that insurance does not work very well under these circumstances. The law of large numbers requires that risks be “independent,” behaving, for example, as car accidents or fire hazards. These conditions produce reliable actuarial tables, which form the scientific foundation for pricing in the insurance industry. However, when large-scale catastrophic property losses occur, risks are no longer independent because a hurricane affecting one insurer will also affect every other insurer writing coverage in the same geographical area. In effect, catastrophic property losses are highly correlated risks-as opposed to being independent risks. And since large-scale property catastrophes impact a significant part of the insurer population both in physical and in financial terms, the law of large numbers does not operate under these circumstances, making it impossible for reinsurers to diversify risks.
Catastrophe Insurance And Loss Ratios
For most lines of insurance, the value of loss per dollar of insurance varies little from year to year. For that reason, a dynamic premium strategy, in which premiums are set so that the loss ratio (based on say the average of the last three years of loss) attains some target level, will produce a time path of premiums which is reasonably smooth. The actual current year loss ratio will also be reasonably smooth and will be close to the target.
In this case, it is both possible and appropriate for the company to plan to pay today’ losses out of today’s premiums. Of course, some capital surplus is necessary to cover unexpected losses, but the quantity of surplus is small if the loss pattern is smooth, and in this case no reserve fund is needed to cover future losses.
In contrast, in the case of catastrophe insurance, the annual pattern of losses is highly non smooth and dynamic premium strategies based on a few years of experience will lead to loss ratios in some years which are far from the target. In particular, in bad years these loss ratios could require a large stock of surplus capital.
The fundamental problem of catastrophe insurance thus seems clear. Unlike every other line of insurance, the contract of catastrophe insurance, as presently structured, requires that the seller have access to a large pool of liquid capital every year in which the contract stands. Since such large pools of capital do not exist, firms have withdrawn from this market rather than bear the risk of insolvency.