The Potential Role of Government in Financing Catastrophic Risk
April 23, 1998

Testimony before the House Committee on Banking and Financial Services
U.S. House of Representatives

Christopher M. Lewis
Policy Economics, Risk Management and Regulatory Practice
Ernst & Young LLP Economics

Chairman Leach and Members of the Committee:

Thank you for the opportunity to appear before you today to discuss the important topic of whether and how the Federal Government could improve our nation’s ability to finance the losses created by large natural disasters. As many of the other panelist have asserted, a strong economic and financial case for federal involvement in improving the allocation of disaster risk in the United States can be made, especially in the areas of pre-disaster mitigation and the financing of large insured losses from catastrophic events. However, the need for federal involvement is also limited in scope and requires a very carefully designed program that does not interfere or distort the existing private insurance market or regulatory structure.

To assist in this effort, I appear before you today on my own accord – not representing Ernst & Young LLP,any client or interest group—to give you the benefit of my experience in studying this issue over the past several years both as a member of the White House Working Group on Natural Disasters and as a private risk management consultant in the insurance industry. Furthermore, while cost-effective mitigation would be an important component of any federal disaster policy agenda, my expertise resides more in the financing of catastrophic risk. Hence, the majority of my comments will focus on the financial aspects of the current catastrophic risk problem facing the nation today.

What is the Nature of the Problem?

Over the past 10 years, the United States has witnessed a significant rise in the magnitude of insurance industry losses from natural disasters. In nominal dollars, nine out of the 10 largest U.S. catastrophes in history have occurred since 1989. In fact, after adjusting for housing price inflation, insured losses over the 1989-1995 period totaled almost $75 billion, more than five times the average real insured losses during the prior four decades. More importantly, however, is the fact that one of the major factors driving this increase in disaster losses is the continued development in high-risk areas of the country.

To illustrate this fact, Table 1 examines the potential magnitude of prior disaster events if they were to occur today given today’s build-up in property values in high-risk areas of the country. Specifically, Table 1 adjusts the losses associated with all disaster events experienced over the 1949-1997 period for the migration of U.S. population into higher risk regions of the country and the resulting increase in the value of housing stock in these areas. After making these adjustments, Table 1 ranked the 10 worst disasters in terms of $1994 losses.

Not surprisingly, three of the top 5 disaster losses were Hurricane Andrew ($18 billion), the Northridge Earthquake ($12 billion) and Hurricane Hugo ($6 billion). More surprisingly, however, is that six out of the 10 largest disasters in history actually occurred during the twenty-year period from 1950-1970. That is, if the same disaster events that occurred in the period 1950-1970 were to reoccur today with the current concentrations in population and housing values in these regions, industry losses from these events would total almost $34 billion.

More troubling is that this trend in growing exposures is continuing. Insured coastal property values in the United States grew 69 percent from 1988 to 1993 to $3.15 trillion. Similarly, the average annual growth rates in the population per square foot in California (2 percent) and Florida (3.2 percent) over the past fifteen years have been double and triple the average national growth rate for the whole U.S. (U.S. Bureau of the Census, 1994)¹. Furthermore, research on frequency and magnitude of hurricanes and earthquakes measured by the Saffir-Simpson and Richter scales, respectively, indicates a strong potential for disaster activity to remain at or above cyclically high levels over the next 20 years – strong indications that disaster losses will continue to be a real problem facing this country.

This increased recognition of disaster exposure has sent reverberations throughout the private sector financial markets. Reinsurance companies responded to this increased exposure quickly by raising rates. Discussions with insurance company executives indicate that reinsurance rates increased by as much as 150 percent from 1993 to 1995. With a rise in reinsurance rates, an increased catastrophe exposure retention, and a realization of their overexposure to disaster risk, primary insurers sought comparable rate increases. When these rate increase proposals were pared down in the process of state insurance rate approval, insurers started to withdraw from the market, causing a drop in the availability of insurance coverage for individuals living in high risk areas of the country. In response, states instituted new regulations restricting insurer exits, established new state insurance facilities and approved modest increases in primary insurance rates. The net result, however, was a continued overexposure of insurance companies to natural disaster risk. Similar disruptions occurred after the Northridge earthquake.

The concern over natural disaster expenditures after Hurricane Andrew was not limited to the insurance industry. In a 1994 Senate Task Force Report, the U.S. Congress raised concerns over the growth in long-term disaster recovery expenses incurred by the federal government. Thus, at the same time the insurance industry started seeking federal assistance in reducing their catastrophe exposure, the federal government was concerned with the budgetary implications of disaster recovery expenses that already were being incurred.

How Catastrophic Losses are Financed Today

Ultimately, the losses from natural disasters always are paid by only one constituency – the American people. However, the method through which the costs of natural disasters accrue to the American people (e.g., taxes vs. policy premiums vs. direct losses) is critical in optimizing the overall welfare of society. For example, if all losses were financed through federal post-disaster assistance, which in turn is financed by general tax revenues, the general tax code would determine who bore the costs associated with disaster losses. In this case, the burden of financing disaster risk would not be linked to the individuals creating the disaster exposures and individuals would continue to increase society’s overall disaster exposure by building in high-risk areas – clearly an undesirable solution.

Alternatively, if all disaster losses were completely financed up front through the purchase of insurance contracts at an actuarially-fair premium, individuals would have to factor the costs of insurance into their decisions on where to locate and how much disaster mitigation to incorporate into their building decisions. As a result, individuals would have an incentive to build in lower risk areas of the country, reducing the growth in the country’s overall disaster risk and optimizing the allocation of risk in the economy. Self-insurance accomplishes the same goal of risk internalization, but without the benefits of diversification offered by insurance.

Ideally, a comprehensive system for financing catastrophic losses would (1) create strong incentives for individuals to mitigate or reduce their exposure to loss from disaster events (and thereby, the overall exposure of society); (2) improve the efficiency and effectiveness of how losses that do occur are financed; and (3) provide for quick response in assisting individuals that do need assistance in the wake of a disaster. The question before the Committee is how can Federal policy help foster a market-based solution that meets these objectives? To address this question, however, requires a closer examination of the problems in the insurance markets that are preventing a market-based solution today.

The Role of Insurance

In financing property/casualty disaster risk, insurance companies are the primary intermediary in the U.S. economy. Individuals living in disaster-prone areas are exposed to considerable losses from disaster events. In providing disaster insurance, an insurance company offers to assume a portion (e.g., over a deductible) of the policyholder’s disaster risk exposure in exchange for a premium. After accumulating these policyholder positions, the insurance company can diversify its disaster risks through risk pooling (aggregation), risk identification and segregation, risk monitoring, and risk mitigation. That is, insurance companies serve a valuable role in identifying, monitoring, pricing, and controlling the individual risks associated with property coverage: risks that are too asymmetrically under the control of the insured to be effectively traded directly in the capital markets where more public information on risk is required.

An insurance company’s ability to diversify insurance risks through pooling is extremely effective when risks are independent and identically-distributed like fire insurance. However, in the case of large idiosyncratic risks like disaster risk, where multiple policies are effected by the same event, the benefits of pooling are significantly less. For these events, diversification must occur across time with insurers financing disaster exposure through the purchase of reinsurance, the issuance of capital market securities or through stockholder equity. For most disaster risks, the problem does not reside in whether the events are "insurable," but whether the insurance company can establish an effective mechanism for accomplishing this intertemporal smoothing of claims.

Historically, insurance companies have relied largely on reinsurance or self-insurance to smooth claims over time. Reinsurance provides a useful source of capital for regional and local insurance firms and plays an integral role in expanding capacity in the primary insurance market. The primary advantage of reinsurance is to achieve a greater degree of spatial diversification through pooling and better price disclosure, although reinsurance also can assist in the intertemporal diversification of disaster losses through the maintenance of long-term (implicit contract) relationships. However, given their own limited liability and exposure constraints, even large reinsurers lack adequate capacity for intertemporally diversifying large disaster losses. Limited liability, regulatory pressure and a competitive market place similar constraints on an insurance company’s ability to accumulate large capital reserves against disaster risk. In addition, even if an insurer could accumulate adequate reserves over time, if a disaster occurs before the reserve is adequately funded, the insurer risks bankruptcy.

In response, considerable attention has been focused on creating alternative market mechanisms for insurance companies to hedge their exposure to natural disaster risk. As a result, in the early 1990s, insurers started looking for alternative forms of inexpensive capital. The natural place to look was the $19 trillion capital markets.

Capital Market for Catastrophe Securities

The prospect of finding another source of capital through disaster derivatives or securitization was alluring to insurance and reinsurance firms because of the sheer size of the capital market. If financed through the capital markets, natural disaster losses of the magnitudes of the Northridge earthquake and Hurricane Andrew would be swamped in the normal trading volatility in the securities markets. Furthermore, in the market of rising reinsurance rates following Hurricane Andrew, insurers saw the capital markets as a possible source of cheaper funding.

Catastrophe securities also were seen as offering advantages to institutional investors. On the investor side, the attraction of securities in disaster risk is the ability to better diversify their investment portfolio by adding a new security with a return that is largely uncorrelated with the returns associated with stock and bond portfolios. Furthermore, an examination of the reinsurance market suggests that the potential investor return from catastrophe securities could be significant. Thus, catastrophe securities can be structured to yield an attractive risk-adjusted rate of return for investors.

Recognizing these advantages, insurers and investment banks have made considerable investments in developing new capital market products that are attractive to both insurers and investors. These products include such instruments as:

• Cat Bonds: Cat Bonds are debt instruments that carry above market interest rates, but where the interest or principal payments on the bonds vary based on the loss experience of an underlying index. The underlying index has been based on either the insurer’s own loss experience (indemnity bond), an industry index of loss experience (index bond), or on some combination of both (e.g., double trigger). In a typical Cat Bond transaction, a Special Purpose Reinsurance company (SPR) is established off-shore to provide direct reinsurance to an insurer. The SPR is then financed through the issuance of the Cat Bond securities.
• CatEPuts: Catastrophe Equity Put options (CatEPuts) are contingent equity arrangements that give the holder (i.e., the insurance company) the option to call on additional capital investment from the option writer (i.e., reinsurer or investor) following a catastrophic loss. Essentially, a CatEPut allows an insurer to shore-up its balance sheet after a catastrophic loss.
• Contingent Surplus Notes (CSN): Under a CSN arrangement, investors make a principal investment in an insurance company trust collateralized with Treasury securities that pays an above market interest rate. In return, the insurer retains the option to replace the Treasury collateral with company surplus notes in the event of a (catastrophic) event.
• Chicago Board of Trade PCS Options: The CBOT created the first catastrophe derivatives market back in 1992 with the advent of catastrophe futures and options. Currently, the CBOT offers options benchmarked to the PCS (Property Claims Service) index of insured catastrophe losses in nine US regions. The PCS cat insurance options offer a risk transferring option for insurers and are typically sold as call spread options, which are similar to traditional excess-of-loss reinsurance contracts. Buyers simultaneously buy and sell call options with different strike values to create a desired coverage layer.
• CATEX: Starting October 1996, the Catastrophe Risk Exchange (CATEX) was established in New York and, more recently Bermuda (CATEX Offshore), as a swap market for insurance risk trying to further diversify geographically.
• BCE Options: The Bermuda Commodities Exchange (BCE) which opened in November, 1997 offers an option similar to the CBOT PCS option and is based on the Guy Carpenter Index of industry losses. The BCE options are for windstorm damage to homeowners policies only, but are available at the zip code level.

Following several years of lackluster performance, the catastrophe securities market registered significant progress as a new asset class for securitization in 1997. During the year, the market registered over $1 billion in new issuances, including transactions by USAA, Swiss Re, Winterthur and Tokio Marine. In fact, several of these transactions were oversubscribed – demonstrating a latent demand by investors for these new instruments. All indications suggest that 1998 will exceed the $1 billion mark set in 1997; and if the catastrophe market continues to follow the evolutionary path of other asset-backed securities markets, new issuance could grow significantly over the next five years.

Notwithstanding this potential, the securities market faces many obstacles that need to be overcome before a fully developed market can be established. For example, the market faces a lack of standardization in risk measurement and in structuring transactions, a lack of a generally-accepted index on which to base payouts, and high transactions costs vis-a-vis traditional sources of reinsurance. Thus, even if the private capital market is able to grow over the next several years, catastrophe securities will be more likely to provide additional capacity for insurers in the low to medium range of disaster losses. That is, capital market solutions are unlikely to provide the answer for financing truly catastrophic events (e.g., disaster events resulting in over $30 billion in industry losses) for quite some time, if at all.

The Need for a Federal Government Role

Given the shortcomings within the existing markets for financing disaster claims over time, insurance companies, reinsurers, states, and other constituents have raised the issue of a Federal role in the financing of disaster insurance. Historically, the insurance industry, which is regulated by state governments, has opposed federal intervention in the insurance marketplace. However, the large scale disruptions created by Hurricane Andrew and the Northridge earthquake in recent years have prompted many members of the insurance industry to call for some type of federal assistance for catastrophe claims given the magnitude of insurer disaster exposures and the lack of a mechanism to smooth large disaster claims overtime.

To their credit, some states have worked to develop short-term solutions. For example, in 1996, California started the California Earthquake Authority(CEA). The CEA began operations with participating companies contributing the initial $700 million in capitalization. Similarly, the state of Florida has created the Florida Hurricane Catastrophe Fund – a mandatory catastrophe reinsurance pool for property insurers writing business in Florida. The CEA and the Florida Catastrophe Fund clearly have the capability to finance a small or moderate disaster. However, states even the size of California and Florida do not have the financial capacity for financing major disasters, even given their substantial tax advantage over private insurance companies. Furthermore, existing state programs have raised concerns within the insurance industry over residual insurer liabilities to state funds and the creation of "thinly capitalized" insurance affiliates within the state. As a result, there is a growing recognition that these states funds are only a partial solution.

Market-enhancing Federal Policy

Evaluating the potential role of the Federal Government in the financing of catastrophic risk is an important issue. In examining the current problems facing the insurance industry and in the financing of catastrophic risk more broadly, two shortcomings in the current allocation of catastrophic risk become apparent – the lack of internalization of risk in individual decision-making and the absence of a mechanism that would allow the insurance industry to smooth large disaster claims over time. In addition to cost-effective mitigation, the encouragement of actuarially-priced insurance is one mechanism for improving the internalization of risk at the individual level. However, encouraging the use of insurance will be ineffectual unless insurance companies can provide the insurance policies without jeopardizing their solvency. Thus, insurers also would need a solution that enhances their ability to finance disaster claims over time, thereby expanding their capacity to provide insurance protection at a given insurance premium.

To develop a federal program that can expand the capacity of the insurance industry without interfering in the regulatory framework of the insurance industry and while continuing to support innovation and development in the insurance and capital markets presents a considerable challenge. Equally as difficult will be developing an approach that is narrowly-defined and targeted at providing a mechanism for intertemporal risk diversification without encouraging additional risk-shifting to the Treasury or to other insurers. In this respect, the Federal Government’s unique position in the debt markets is a great advantage. However, the difficulty is designing a program sufficiently targeted to only utilize this unique borrowing authority. Other issues that such a program would face include the following:

• Avoiding any replacement of displacement in the insurance industry, including favoring one segment of the industry (e.g., state funds) over other segments of the industry (e.g., direct insurance writers).

• Setting appropriate attachment points and caps on the federal government’s overall exposure so as to allow the private sector to continue to provide coverage layers of disaster losses consistent with industry losses of up to $25 or $30 billion.

• Maintaining the current state regulatory framework where the critical link between state regulation of insurance company solvency, insurance rate approval, and the state guaranty fund structure is maintained.

• Fostering the development and innovation of the capital markets, effectively allowing private sector innovation to crowd out the public sector program.

• Pricing the program at "normal" market rates that include such factors as the average expected cost, a charge for measured variance in expected costs, an administrative load, and a cost-of-capital charge consistent with normal reinsurance rates (5-7 percent over Treasury securities) and obtaining full information on changing market perceptions with respect to risk.

• Avoiding the creation of another federal bureaucracy to support any program that is created.

• Creating a program that is budget neutral according to the standard budget practices employed by the Congressional Budget Office and the Office of Management & Budget. (Of course, according to current budget practice, any program that is priced as described in the previous bullet would exceed the requirements currently imposed on any other program in the budget).

• Avoiding the creation of a direct federal insurance or reinsurance mechanism that is subject to risk-shifting (e.g., moral hazard or adverse selection) by the insured.

Conclusion

In conclusion, there is a strong economic and financial argument for the federal government to play a role in the financing of catastrophic risk. However, the challenge for the Committee is how to correctly define this role so as to encourage cost-effective mitigation and the development of a mechanism that would allow the insurance industry to smooth large disaster claims over time.

At this point, I would be happy to answer questions that the Committee may have on my perspectives on the role of the Federal Government in financing catastrophic risk.

_____________________
¹Lewis, Christopher and Kevin Murdock, "Alternative Means of Redistributing Catastrophic Risk in a National Risk Management System," in ed. Kenneth Froot, The Financing of Property Casualty Risks, National Bureau of Economic Research, (In Press).