Note from editor R. Sedlock: This article originally appeared in “Analyzing economic impacts and recovery from urban earthquakes: Issues for policy makers,” a volume accompanying a conference sponsored by the Earthquake Engineering Research Insitute and Federal Emergency Management Agency on October 10-11, 1996 in Pasadena, CA. I thank Dr. Russell for letting me add the article to the Geol 112 Web site. As editor of the scanned-in text, I’m responsible for some minor cosmetic changes (citation format, changing figures to tables, typos).

This paper discusses the dramatic changes in earthquake insurance that occurred in California in the mid-1990s. As you read, remember that this was written a few months before then-Governor Wilson signed into law the California Earthquake Authority (CEA). While this paper does not necessarily give you the most current information, it offers outstanding insights into the problems of insuring against disasters.



Introduction

The total financial loss associated with the Northridge earthquake has been estimated to be in the region of $20-25 billion (California Governor Office of Emergency Services). Analysis of the loss data suggests that the potential loss from a future earthquake in the state could exceed $50 billion. How are such losses to be met? In this paper, we use the Northridge experience to review the role of financial institutions, both public and private, in providing the resources necessary to recover from future mega-catastrophes.

There is, of course, nothing new about financial loss due to acts of nature. The fundamental pnnciple of risk sharing, that a major financial loss to one individual will be a much smaller loss if shared among many individuals, has formed the basis for contracts of insurance since the time of ancient Rome. Moreover, this principle remains true even for losses in the $50 billion range, provided these losses are not too frequent.

For example, if every 20 years (on average) a $50 billion earthquake loss occurs somewhere in Califoriiia, and if we assume that 5 million households are at risk each year, it is sufficient for insurance agencies to collect an annual premium of $500 per household per year to break even. (We assume, for simplicity, no allowance for adtistrative costs or the time value of money.) Looked at differently, a $50 billion loss represents 5% of the annual $1 trillion income of the State of California and not much more than 1% of the value of the state’s capital. Again, to assist in framing, between May and August of 1996 the high tech companies of Silicon Valley lost 30% of their equity value in the summer stock market sell-off.

To say that infrequent risks of $50 billion can be shared, however, is not to explain how they will be shared. From the financial point of view, the most important consequence of the Northridge earthquake has been the withdrawal of private agencies from the business of insurance and the resulting shift in the burden of risk sharing towards the public sector. Put simply, after December 1996 when the California Earthquake Authority (CEA) is expected to become operational, private earthquake insurance in the State of California as we know it will cease to exist. How did this state of affairs come about?


Risk Sharing and Mega-Catastrophes - The New Role of the Public Sector

There are at least four ways to apportion the losses causes by an earthquake.

1. The losses can be borne directly by the parties who incur them. This self–insurance can arise either because the parties at risk choose not to purchase insurance, or because the contract of insurance specifies a deductible or co–payment. Large deductibles (15% and higher) are common in earthquake contracts.

2. Losses can be shifted to a private insurance corporation through the purchase of a contract of insurance.

3. Losses on debt-financed real estate can be shifted ex post to the lending institution by the borrower defaulting on the loan. This will only be a rational strategy if the loss has lowered the borrower’s equity to less than zero. In this case, mortgage lenders become de facto insurers even though no explicit contract of insurance is written.

4. Losses can be shifted to public agencies (federal, state, and local). This can be done either through the explicit provision of public insurance ex ante [i.e., beforehand], or through the provision of ex post [i.e., afterward] relief in the form of grants and subsidized loans from agencies such as the Federal Emergency Management Agency (FEMA), the Small Business Administration (SBA), and Housing and Urban Development (HUD).

Using these four categories, it would be very useful to have a breakdown of who bore the $25 billion loss of Northridge. Precise estimates in this form, however, seem to exist only for insurance category 2, which the California Department of Insurance puts at $12 billion. Analysis of federal grants and loans by Lahr and Rose (1995) suggest that the component of category 4 amounted to $8 billion. On the other hand, the only part of a government loan which represents risk sharing is the loan subsidy, since the individual could always borrow themselves at market rates. Estimates by Eguchi et. al. (1996). Table 1 suggests that the state contribution may be around $1 billion. There is no precise estimate of the share of loss taken by mortgage lenders through write-downs and defaults, but anecdotal evidence suggests it was rather small, probably less than $0.5 billion.

For the sake of discussion, the following breakdown of losses from the 1994 Northridge earthquake may not be all that far wrong.

A more refined analysis would reduce the public contribution to the subsidy component of loans, so $9 billion is an upper limit to the public contribution to risk sharing.

In any case, as we can see, approximately half of the risk of financial loss was taken by the private insurance industry. In their view, this exposure was too large. They responded by withdrawing from this insurance line. Since by California law contracts of homeowner’s insurance had to give buyers the option of purchasing earthquake insurance, this required that insurers also withdraw from the homeowner’s line. According to a May 1, 1995 California Department of Insurance survey, insurers representing 93% of the California homeowner’s insurance market stopped writing new or renewal business or severely restricted coverage following Northridge.

The response to this has been to set up a state agency, the California Earthquake Authority (CEA), details of which we will present shortly. By the CEA’s own estimate, if Northridge were to occur again while the CEA is operational, say in early 1997, their total loss, because of the nature of the policy which they offer, would be $4.5 billion. How would the remaining $20. 5 billion be apportioned?

Since federal grant and loan payments are governed by statutory arrangements which use private insurance as an offset, the public agency contribution would certainly be higher. Unfortunately, there seems to be no estimate of the effect of the creation of the CEA on the federal contribution, so we cannot say with any precision by how much it will be higher. Moreover, humanitarian considerations are unlikely to allow individuals to suffer too large a share of the $20.5 billion loss. Again for the sake of discussion, a breakdown of $25 billion in losses as a result of a 1997 [i.e., post-CEA] Northridge-like earthquake may not be far wrong.

Already we see that in the absence of direct private insurance, the public sector provision of risk sharing will rise substantially. This is for a $25 billion loss. What if the state suffered a $50 billion loss? Since the contribution of the CEA is capped at $10.5 billion, some $40 billion dollars in loss will have to be apportioned between public agencies, self insurance, and mortgage write-downs. Again, though no hard numbers exist, we can see that there would be extensive calls on the public purse to provide necessary funds.
?All of which is to say that the Northridge earthquake has drastically altered the insurance landscape in California. The balance of risk sharing arrangements have tipped sharply towards the public sector, both through the creation of a state-run insurance agency, the CEA, and because of the expectation of much higher ex post public relief. Why did this happen, and what are the implications for equity and efficiency?


What Makes Earthquake Risk Uninsurable?

The mass exodus of private insurance companies from the earthquake line in 1994, particularly given that this also forced them to abandon the homeowner’s line, caught many insurance experts by surprise. Certainly there was no argument in previous years that earthquake risk was “uninsurable” (i.e. could not be offered at a price which generates profits), and indeed in 1984 it was the insurance companies which offered to provide the earthquake insurance option thus linking the earthquake and homeowner’s markets. What is it that now makes it impossible for private insurance markets to offer this line?

We have set out in detail elsewhere (Jaffee and Russell, 1996) an explanation for the failure of the catastrophe insurance market. In this section we will review the main features of the argument. We think of an earthquake as an event with a large loss which (blessedly) happens infrequently. In this regard, earthquakes (like hurricanes), are distinguished from almost all other lines of insurance in which losses occur frequently with an average loss which is small.

This difference between earthquakes (and hurricanes) and all other insurable events shows up statistically in the loss ratio of insurance companies, i.e. the ratio of total premiums collected in a year to total losses in that year, times 100. This loss ratio is presented in Table 1.

What is clear from Table 1 is that, in all lines of insurance in which there is no element of “catastrophe,” the loss ratio is less than 100%. This means that in any year losses for that year can be met out of that year’s premiums. For the earthquake line, however, in a bad year, say 1994, there is simply no way to pay annual losses out of that year’s premiums. This is confirrmed in Table 2, which gives earthquake loss ratios by year for the State of California.

To provide earthquake coverage, an insurer must have some way of accumulating premiums in good years against large losses in bad years. Rather surprisingly, U.S. accounting and corporate governance arrangements make it all but impossible for an insurance company to do such necessary “rainy day” planning (Wallace and AIthoff, 1994).

There are two major difficulties.
1. U.S. accounting practice does not permit posting contingency reserves against a loss that is expected to occur but has not yet occurred, even when that loss is virtually certain (FASB 5).

2. Although this does not prevent an insurance company from itself setting aside surplus funds to deal with a future earthquake loss, these funds cannot be earmarked for this purpose, and any firm which carries large reserves of liquid assets draws the attention of corporate takeover specialists. These raiders have no interest in running the company as a growing concern but will strip the company of its cash, closing it down if necessary.
?There is now a large academic literature (e.g., Jensen, 1986; Blanchard et al., 1994) which shows that corporations with large stocks of “free” cash open themselves to takeover. (Think, for example, of the Chrysler Corporation, which accumulated $7 billion in cash against the next business cycle turndown, then had to fend off the unwelcome takeover efforts of Mr. Kirk Kerkorian).

The problem faced by earthquake insurers of how to accumulate and hold a pool of resources to meet a large infrequent loss is not a problem in classical insurance markets at all, but is instead a problem in the organization of capital markets. Recently a number of new capital market instruments have been developed which might, in principle, have allowed private earthquake insurance markets to remain viable. These include

a. Act-of-God bonds. These bonds have the property that if an earthquake occurs, the insurance company is relieved of the obligation to repay principal and interest. (Sometimes the relief is limited to the payment of interest). Such bonds have been issued, though not in large amounts. At the moment, Merrill Lynch is attempting to sell $500 million of these bonds on behalf of USAA, a car and home insurance company in San Antonio. These bonds will be cancelled if USAA suffers hurricane losses of more than $l billion from a single event in a year (W.S.J. Aug.19, 1996). This type of bond revives an old insurance contract, the contract of bottomry, in which money lenders advanced the capital for a voyage to a ship owner, the debt to be forgiven if the ship was lost. This form of debt provides funds to an insurance company if it suffers a loss, but since the bonds must be repaid if no loss occurs, it provides no incentive to raiders to try to take the company over.

b. Catastrophe Options. As a second attempt to find a capital market solution to the problem, the Chicago Board of Trade has begun trading a number of national, regional, and state catastrophe options. For example, since September 29, 1995 a California catastrophe option contract has traded.

During the recent hurricane season, trading in catastrophe options increased sharply. On September 5, 1996 as Hurricane Fran moved towards the east coast, the CBOT traded 3,308 regional cat option contracts. This was equal to the whole of the previous three months combined. Still, in dollar terms, this day’s trading provided only $6 million in insurance capacity, a drop in the potential loss bucket (Chicago Board of Trade, Catastrophe Insurance Options Market Update). Creative as these solutions are, in terms of revitalizing a private insurance market in the State of California they are too little, too late. Providers of private market earthquake insurance have withdrawn from the State of California. What has been left behind?


The Brave New World of Earthquake Insurance: The CEA

With withdrawal of private insurance companies from the market, earthquake insurance will now be provided by a state agency, the California Earthquake Authority (CEA). The CEA will offer a stripped-down “mini” earthquake insurance policy with no coverage for damage to pools, patios, fences, driveways, or detached garages. The deductible will be 15%. The administration of the contract, claims adjustment, claims settlement, etc. will be left to private insurers in the state who will be able to offer the CEA policy if they make a capital commitment to the CEA fund.
?The most interesting feature of the CEA is its method of funding because, although it is a state agency, it receives none of its budget from the state. Capital to meet losses is provided as follows. Initial non-reimbursable capital of $1 billion will be provided by the insurance industry. Re-insurance corporations will provide $2 billion of capital, and, in the event of a loss exceeding $3 billion, insurance companies will be called upon to provide up to another $3 billion. If needed, another $1 billion will be borrowed in the bond market, to be repaid from an assessment on policy holders. If loss exceeds $7 billion, the CEA will borrow $1.5 billion in Act-of-God bonds, and if losses exceed $8.5 billion, the insurance industry will commit another $2 billion, bringing capital to a maximum of $l0.5 billion.

As can be seen, the maximum commitment of capital by the insurance industry is $6 billion, and this commitment will be reduced over 12 years as premiums accumulate. Thus, the insurance industry will never be required to commit more capital towards earthquake losses than 50% of the losses which it incurred following Northridge.

How does the CEA manage to do this? That is to say, how can a public agency which is in receipt of no public funds provide a product (earthquake insurance) which the private sector says it is unable to provide?

In part the answer to this question lies in the reserving capacity of the CEA. Unlike private insurance companies, the CEA can set up a protected reserve fund and accumulate capital without fear of being taken over. Indeed this fund has even been granted tax-exempt status by the IRS. As we discussed earlier, this advantage is not available to the private sector.

In a deeper sense, however, it seems fair to say that the CEA is not really solving the problem of providing credible earthquake insurance at all. This is clear on the face of the contract, which points out that policyholders are subject to a (maximum) 20% premium add-on in the event the CEA has to raise the $1 billion in bonds. It is also clear in the wording of the contract, which states that policyholders will receive pro-rata payment of losses if the CEA’s capital is exhausted. [Editor’s note: be sure you understand the significance of this point.] Pro-rata claims settlement with premium levies if losses are too high is not what one normally understands by the term “insurance.”

This failure to provide credible insurance arises because, assurances notwithstanding, the CEA is desperately undercapitalized to meet bad-luck outcomes. The CEA’s own documents state that “in its first year of operation the CEA will have enough resources to pay all claims from an earthquake more than twice as destructive as the Northridge earthquake.” (CEA FAQ’s Sept. 29, 1996). Because the CEA offers a mini-policy with a 15% deductible, it would have only paid out $4.5 billion of Northridge’s $20-25 billion losses. So, with a capacity of $10.5 billion, this statement is technically correct. However, such an event would exhaust the CEA’s capital resources, and there is no provision for replenishing this fund. What happens, God forbid, if there is a second loss? The insurance companies are free and clear of capital commitments once they pay in $6 billion, and the other commitments from the capital market now lapse.
?No one wishes to see two major earthquakes early on in the CEA’s life, but it is precisely the fear of this that has driven private insurance companies (with combined surplus in excess of $200 billion) out of this market. Finger-crossing is no substitute for a serious discussion of what would happen if the CEA found its resources exhausted by a $50 billion hit, or even two closely spaced $10 billion hits. Looking this unwished-for calamity squarely in the face is surely appropriate.


What Happens if the CEA’s Capital is Exhausted?

If the CEA’s capital is used up, losses will fall on the other three affected parties: the primary party to the loss, mortgage lending institutions, and federal, state, and local relief agencies. Focusing for the moment on lending institutions, it is clear that even though mortgage default losses in Northridge were small, some lenders are beginning to realize that the size of their potential loss is large (e.g., Shah and Rosenbaum, 1996). In the secondary mortgage market, which would bear the brunt of large mortgage defaults, Freddie Mac has responded to the loss data by refusing to buy condominium paper originating in California unless it is accompanied by either earthquake insurance or payment of a special fee. Fannle Mae, the other major buyer of mortgages in the secondary market, has not yet imposed such a restriction but is widely expected to follow the Freddie Mac lead to avoid becoming a dumping ground for uninsured loans.

If the secondary market insists on condo earthquake insurance, the requirement that earthquake insurance be purchased as a condition of a loan may easily spread to primary lenders such as banks and Savings & Loans. Indeed, it is remarkable that mortgage lenders do not insist on earthquake coverage now, the failure to do so probably reflecting the equity cushion provided by rising home values in California over the last 20 years.

If earthquake insurance does become mandatory, even more of a burden on the CEA. Freddie Mac has already indicated that the CEA policy will satisfy their condo insurance requirement. This increased burden makes it all the more important to consider the implications of losses exceeding resources.

However, for an earthquake loss far beyond (say, by $20-30 billion) the CEA’s resources, the real question will not be the role of the mortgage lenders. In this case the question is what will be the role of the federal and state authorities. Could the State of California really enforce the “firewall” between state resources and CEA resources given massive losses on an insurance contract sold by a state agency? What would be the role of the federal government? These questions are very difficult to answer, but in posing them one fact is clear.

By replacing the private contract of insurance with a contact provided by a state agency, the California Legislature has, in effect, condoned the withdrawal of private-enterprise earthquake insurance from the state, and has given birth to a set of expectations that will lead to a greatly expanded role for the public sector at both the state and federal level. This “nationalization” of earthquake insurance is a direct consequence of Northridge. It has received no criticism from the usual source, the private sector, because the CEA allows private insurance to re-enter the lucrative homeowner’s market while capping their total earthquake exposure at 50% of the Northridge loss, $6 billion: It is little wonder that Allstate’s stock price rose 2.5 points on the news that the Governor would sign the CEA bill. Some analysis of the costs and benefits of this shift to the public sector is clearly in order.


Earthquake Risk and the Public Sector

Economists are generally not great lovers of the public sector. For now, however, public sector provision of earthquake insurance is the only game in town. That being the case, it is necessary to give some attention to the design of public sector financial response.

This design must consider the twin criteria of efficiency and equity. From the efficiency viewpoint, it is important that the insurance not be underpriced. If it is, the insured will not face the correct incentives to reduce risk, and society will end up compensating individuals for losses which were ex ante avoidable. For example, underpricing of insurance will induce too many people to live in earthquake-prone areas, and regardless of where they live will cause individuals to under invest in mitigation (Kunreuther, 1996).

The contract price structure offered by the CEA clearly recognizes the need to price insurance according to location, and also sets up incentives for individuals to brace and bolt. The CEA, however, will not be a major source of funds after a large loss. Assuring that this loss is primarily borne by federal and state agencies, as things stand now there is simply blanket ex post funding, so individuals counting on these agencies have no incentives to do anything to mitigate loss.

Moreover, looking at the details of federal relief (see Comerio et al., 1996), we see that the FEMA Minimal Home Repair grants and SBA Disaster Assistance are reduced dollar-for-dollar by insurance payments, so federal programs actually provide a disincentive to purchase more rationally priced insurance. It remains an open question how this federal aid should be structured to provide mitigation incentives, but since this aid will be the major source of funding in a mega-loss, it seems important to give this question some attention.

Similarly, looking at the issue of equity, in many cases federal and state programs reimburse regardless of income (Comerio et al., 1996). Was it really fair for the State of California’s FAIR plan (a predecessor of the CEA set up to provide insurance in inner cities) to reimburse Madonna for the loss of her swimming pool in the Malibu fires? But again, this means testing relief at the time of a disaster is clearly a very difficult task.


Conclusion

The Northridge earthquake has caused major changes in the financial arrangements for the provision of risk sharing in the State of California, as follows:

1. The private insurance market has disappeared.

2. It has been replaced by a limited insurance product offered by a seriously undercapitalized state agency, the CEA.

3. Mortgage lenders, realizing the extent of their potential loss, are beginning to require earthquake insurance as a loan condition, raising more questions about the CEA’s adequacy.

4. State and federal relief agencies, the ex post insurers of last resort, now face far greater loss exposure, though this exposure is very difficult to quantify.
?None of these developments is ideal. In particular, the absence of any type of ex ante pricing on the part of the federal and state relief agencies provides no incentive to mitigate the loss. Indeed, some federal provisions reduce the incentive to buy any insurance at all.

Some of the incentive problems present now would be eliminated by a return to a private insurance market. This does not seem likely to happen. In particular, no efforts are being made to change the tax and accounting practices to allow private insurers to accumulate an earthquake contingency fund tax-free, a privilege now enjoyed by the CEA.

Given this, we can hope that further attention may soon be given to the financial consequences of an earthquake that wipes out the CEA. What would federal and state government do? To whom would mortgage lenders turn if the earthquake policies on which they insisted are worthless?

No one wishes to be an alarmist, and certainly regions have recovered from catastrophes far worse than earthquakes. In WWII, Tokyo suffered property damage 10 times larger than the great Tokyo earthquake, and 1 million lives were lost. But in this new world, there are clearly expectations that someone, most likely federal and state agencies, will be around after a loss. To ask for details in advance of how this will be done does not seem unreasonable. To know in advance what financial arrangements will be made can only help in the process of recovery.


Bibliography

Lahr, M.L. and A.Z. Rose “The Northridge Earthquake: Direct Effects of Business Disruption and Structural Damage” Working Paper Nov.1995.

Eguchi, R.T. et al. “Analyzing Economic Impacts and Recovery from Urban Earthquakes” EQE International 1996.

Jaffee, D.M. and Russell, T. “Catastrophe Insurance, Capital markets, and Uninsurable Risks” Wharton Discussion Paper 1996.

Wallace, W.A. and J.M. Althoff “Retrofitting Away for a Rainy Day" AICPA Case 94.08 1996.

Jensen, M. “Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers” American Economic Review, v. 76, no. 2, p. 323-329, 1986

Blanchard, O.J., F. Lopez-de-Silones, and A. Schilefer “What do firms do with Cash Windfalls?” U Journal of Financial Economics, v. 36, p .337-360 1994.

Comerio, M.C., J.D. Landis, and C.J. Firpo “Residential Earthquake Recovery” Governor, State of California Office of Emergency Services, 1996.

Shah, H.C. and D. Rosenbaum, “Earthquake Risk Shakes Mortgage Industry” Secondary Mortgage Markets 13 v. 2, p.12-19, 1996.

Kunreuther, H. “Policy and Decision Making for Low Probability High Consequence Events,” mimeo 1996.