Should the Federal Government Backstop Insurance Industry Terrorism Coverage?
December 1, 2003Judyth W. Pendell, Warren B. Azano
Judyth W. Pendell, Director, Center for Legal Policy, Manhattan Institute
Warren B. Azano, Vice President, Hartford Financial Services Group
**The Federalist Society takes no position on particular legal or public policy initiatives. All expressions of opinion are those of the author or authors.**
``…the viability of free markets may, on occasion, when you are dealing with a degree of violence, require that the costs of insurance are basically reinsured by the taxpayer ...''
- Alan Greenspan, October 17, 2001, in comments before the Joint Economic Committee of Congress.
The question addressed by this paper is whether or not this particular occasion, as shaped by the events on and after September 11, is one that requires direct federal involvement in the private insurance/-reinsurance marketplace.
This paper discusses whether and to some extent how the federal government should be involved in the providing of insurance protection against terrorist acts, particularly for businesses and properties that are perceived to be especially at risk. Beginning with a description of why insurers cannot currently provide such coverage, it then discusses: the impact on those left unprotected and on the economy; whether insurers will return to the market or other risk spreading mechanisms will step in to provide protection; existing federal (re)insurance programs; and public policy options regarding terrorism insurance.
It appears that the Congress will be driven by the current politics of terrorism and the economy to take some action to ensure that businesses and individuals get insurance protection against terrorist acts. Accordingly, the paper concludes by recommending that the legislative relief include significant private marketplace incentives and an exit strategy in the event the market for this insurance recovers.
Terrorism: Insuring the Uninsurable
Prior to September 11th insurers did not price for or reserve for losses from terrorism. Terrorism was simply not considered a significant peril (cause of loss) in the pre-September 11 rating and underwriting of insurance in the United States. While the 1993 World Trade Center bombing and the 1995 Oklahoma City bombing certainly illustrated the potential for significant loss of life and property, insurers were no more prescient than anyone else in or outside of government in foreseeing the magnitude of the losses visited upon America on September 11.
Accordingly, terrorism was not identified as a separate peril in insurance policies, nor was any significant premium charged for terrorism risk (losses from prior events would have been included in rate calculations, but they were de minimis relative to the total losses over time for the insurance coverages in question). Finally, since pre-event reserving is not allowed under current insurance industry accounting rules, there were certainly no reserves specifically set aside for terrorism.
The losses incurred on September 11th were unprecedented - substantially larger than any prior man-made or natural disaster losses. Immediately after the attacks, estimates were that the combined loss would be the most expensive man-made loss in history, perhaps reaching $15 billion (by comparison, insured losses for the 1993 World Trade Center bombing were $510 million; for the 1995 Oklahoma City bombing, $125 million; and for the 1992 Los Angeles riots, $775 million). Since then, as insurers have refined their estimates, it appears that insured losses will be in the $30-$50+ billion range, with some estimates as high as $70-$80 billion. This far exceeds the costliest single event on record, man-made or otherwise, which was Hurricane Andrew, in 1992, with inflation-adjusted losses of about $20 billion. To put this into perspective: The insurance industry's costs for Superfund and other environmental liabilities have been estimated at roughly $40-$50 billion - costs that have been incurred over decades, not on a single day; and, if the September 11 losses are at the high end of the estimates, they would exceed the industry's total losses for commercial property & liability insurance, general liability insurance, and workers' compensation insurance combined for the entire 2000 year.
The new understanding of terrorism, combined with loss of capital and reinsurance, precludes insurers from providing terrorism coverage.
The insurance industry will simply not be able to provide coverage for terrorist acts within its standard insurance policies as it has done in the past.
First, the costs from the attacks themselves significantly eroded the capital base of the companies that are the principal underwriters of commercial insurance in this country. As of year-end 2000, the property-casualty insurance industry collectively had about $300 billion in "statutory surplus," the insurance regulatory measure of capital. However, this total included the surplus of personal lines insurers, i.e., companies that write primarily automobile and homeowners policies for individuals. These companies were not significantly affected by the September 11 attacks and would not be major participants in the commercial lines markets going forward.
Tillinghast has estimated the statutory surplus of the commercial lines writers that will be most affected at only about $80 to $100 billion. Since these primary insurers - as opposed to the reinsurers - are expected to absorb about one-third of the insured losses, the impact is significant.
The import of this may not be readily apparent to those who are unfamiliar with how insurers and their regulators determine how much business companies can write. The term "surplus" is itself somewhat misleading, in that it might be read as meaning "excess" or "extra" capital. It does not. A company's surplus is, in effect, the capital cushion necessary to write insurance, and the amount of insurance a company may sell is directly tied to its surplus by its regulators (if surplus is too low relative to premiums, the insurer may be flagged for additional regulatory scrutiny). Thus, the multi-billion dollar reductions in surplus from the September 11 losses have already reduced the affected companies' ability to write future business.
More importantly, the affected companies will be less willing to further jeopardize their earnings and their surplus (read "ability to write business") by intentionally writing insurance on a peril for which losses cannot reasonably be estimated but which could be staggering. This is the second major difficulty arising from the September 11 attacks, and it is arguably more significant than the loss of capital itself. Specifically, those attacks - and the resulting war on terrorism - have completely transformed how insurance companies (and the public generally) view the risk of future terrorist attacks. Whereas insurers did not previously regard terrorism as a significant peril, it is now universally acknowledged that no one knows what lies ahead and that there is at least a potential for losses far worse than even those resulting from September 11 hijackings. Even nuclear attacks against major U.S. cities have suddenly become a very real concern. As noted, for example, in the Wall St. Journal (October 17, p.1):
The means for carrying out nuclear attacks are scattered around the globe - in the form of hundreds of commercial nuclear plants, tens of thousands of nuclear weapons and tons of stored uranium and plutonium that could be fashioned into bombs. Efforts to make nuclear materials more secure have been hampered by tight budgets, geopolitical squabbling and inertia.
Unfortunately, with an almost 4000-mile northern border that is essentially unguarded, and with thousands of ship containers arriving in our ports daily without careful inspection, the U.S. appears to have insufficient protection against the delivery into this country of nuclear or other weapons of mass destruction.
Although some policies have exclusions for losses caused by nuclear incidents or acts of war, the uncertainty and the potential for staggering losses are unprecedented, and this has a number of adverse ramifications. First, there is no actuarially credible way of pricing for such uncertainty; and, second, having already incurred unprecedented losses that in some cases decimated 2001 earnings to shareholders, managements cannot "bet the company," or at least the next few years' earnings, on the chance that there will be no more mega-events.
Finally, the last nail in the underwriting coffin for primary insurers is that the one protection against "bet the company" losses - reinsurance - is disappearing for acts of terrorism. Insurers spread their risk by purchasing reinsurance, sometimes referred to as "insurance for insurance companies." Unfortunately, the signals to date from the domestic and international reinsurance market indicate the almost certain loss of reinsurance protection as of the next round of renewals. This is being accomplished by the addition to reinsurance agreements of exclusions for acts of terrorism. Since the majority of the treaties expire at the end of the calendar year, it appears that as of January 1 the primary insurance market will no longer be protected by reinsurance treaties that cover terrorist acts. As if that were not bad enough, the excess of loss reinsurance treaties typically apply to losses occurring during the year, not to policies written. Consequently, once the new treaties are written with terrorism exclusions, insurers will be unprotected for those primary policies already written during the previous year and expiring over the course of the new year.
This triple whammy (loss of capital, unmeasurable but exponentially higher risk, and loss of reinsurance protection) precludes responsible managements from providing coverage against acts of terrorism without significant backup from the federal government. To do otherwise would be to jeopardize their ongoing financial stability - and that is what insurers trade on and their customers rely on. This has been recognized by Standard & Poors, the insurance industry rating firm, that has stated flatly that maintenance of insurance ratings "depends on mitigating terrorism risks," and that insurers offering terrorism in the current environment without some form of government protection would "invite rating downgrades" (Oct. 19 S&P Commentary).
The Damage Done By the Threat of Terrorism Reaches Through and Past the Insurance Industry to Directly Impact Millions of Businesses that Depend on Insurance to Thrive.
The insurance industry will not be able to completely exclude terrorism, but it will take steps to withdraw from difficult markets. Insurance company policy forms are regulated in each state, most of which require that forms cannot be changed without the approval of the state insurance departments. Absent very comprehensive federal backstop protection, and assuming loss of reinsurance, most insurers will presumably submit their own terrorism exclusions to the insurance department in each state in which they do business; some have already done so (note: unlike primary companies, reinsurers are not required to submit their forms for insurance department approval, so they did not need regulatory approval for their terrorism exclusions).
How insurance departments will react to filed terrorism exclusions remains to be determined. It is very unlikely, though, that such submissions will be universally approved, particularly in states with cities that might be seen as likely terrorist targets. In such event, failure to gain approval for such terrorism exclusions will likely cause insurers to decline to issue policies to businesses presenting high risk of terrorism losses, e.g., high buildings or public arenas, so as to lower their exposure to large terrorist losses.
Thus, whether through terrorist exclusions or withdrawal from markets, insurers will almost certainly reduce coverage provided, particularly to those businesses that pose a higher risk, although they will not be able to eliminate it entirely.
Higher prices and/or reductions in insurance availability will be disabling to many business, even lethal to some. Insurance is essential to an expanding and healthy economy. In his October 17 comments to the Joint Economic Committee, Alan Greenspan noted that hostile terrorist environments are detrimental to the functioning of free and open markets and expansionary economies, because they cause people to withdraw rather than to reach out and take risks. Further, he indicated that the insurance requirements relating to terrorist attacks might be so large that the resulting premiums would be unaffordable, leading to the "very unusual conclusion" that the viability of free markets might require taxpayer-funded insurance; and that "in this very unusual circumstance" free markets and government insurance "are indeed compatible."
Unfortunately, it is already generally acknowledged that the country is in a recession. Over 400,000 Americans lost jobs in October, as the unemployment rate increased to 5.4%, the highest level in over five years. The four-week average of first-time claims for unemployment benefits rose to 463,000 in early October, its highest level in nearly 10 years. The two best-known gauges of U.S. consumer confidence have posted their largest one-month drops since 1990, and retailers reported that September was their worst month since 1991.
While no one can accurately predict the ultimate effects on the economy of significant reductions in insurance for terrorist acts, it is clear that the effects on existing businesses, and particularly on new enterprises and projects requiring capital funding, would be significant - impacting across the marketplace, from the smallest shopkeeper to the largest multi-national, as insurers seek to reduce concentrations of exposure in perceived target areas. Businesses that are already struggling to survive in the economic slowdown since September 11 can ill afford higher insurance premiums, let alone withdrawal of coverage and inability to obtain loans or perhaps even retain existing borrowing.
The real estate industry is especially at risk, as are investment products backed by real estate, e.g., commercial mortgage-backed securities. An October 22 letter to the President, signed by nine real estate trade associations including the Mortgage Bankers Association of America and the National Association of Real Estate Investment Trusts, noted that the real estate industry accounts for over a quarter of the nation's gross domestic product, and that it would face "severe economic dislocation in the coming months" if insurance issues were not resolved.
In a letter dated November 6, the Risk and Insurance Management Society, whose members consist of the risk management professionals in over 80% of the Fortune 500 companies, as well as many smaller companies, urged the Administration and the Congress to "act now to protect the economy by assuring a feasible property and casualty reinsurance source for U.S. policyholders …"
In addition, over 60 national trade associations, including the Chamber of Commerce, the Business Roundtable, and trades representing such diverse businesses as auto manufacturers, banks, chemical companies, resort developers, general contractors, electronic retailers, hotels, amusement parks, shopping centers, home builders, and florists co-signed a letter noting that, "… without insurance, financiers cannot lend, … companies are reluctant to invest … new construction would cease… cargo would be stuck in place ..." and urging immediate enactment of federal backstop legislation.
In sum, withdrawal of terrorist coverage would affect a broad range of businesses, and this in turn would adversely impact an already weakened economy.
Insurance markets are unlikely to stabilize of their own accord for a very long time. Based on the predictions from the Bush Administration as to the likely duration of the war on terrorism, it is very unlikely that the private insurance market will offer anything approaching full coverage against losses from terrorist acts for a number of years. It will certainly take years to eliminate or neutralize a substantial part of the terrorist network; and, since even single terrorist cells can bring about major loss of life and property damage, particularly if its members are willing to undertake suicide missions, there is no reason to assume that even a multi-year respite from major attacks would mean that the threat had disappeared.
As a point of reference, note also that a comprehensive private market has still not returned for environmental and asbestos liabilities, which would appear to be more insurable than terrorism.
Other risk spreading mechanisms might arise, but they would address narrow aspects of the problem, and terrorism coverage is unlikely to be an attractive business venture to even the most entrepreneurial, risk-prone investor. A variety of risk-spreading mechanisms might develop if insurance were no longer available, at least for some of the risks presented. The banking industry, for example, could simply assume the risk on commercial mortgages, particularly in areas and/or for buildings not considered terrorist targets. For properties considered more vulnerable to terrorist attack, banks might spread the risk pursuant to programs in which existing mortgages were defaulted, because no insurance at all was available or the terms were unsatisfactory. In such event, individual mortgagees might enter into new agreements, demanding higher interest, points, or additional security, and then might assume the risk themselves or securitize the mortgages.
Alternatively, real estate investment trusts might develop below-investment-grade high risk/high return products. If owners who had lost their insurance and had their mortgages defaulted were in effect forced to sell their properties for whatever they could get, perhaps REITs could then use the presumably still-existing rental income as backing for high risk/high return investments.
However, to date we have seen no evidence that banks or other capital providers are prepared to take on this risk. Even if they were, it does not appear that any risk-spreading mechanism would or could protect existing property owners and businesses in potential target areas from significant loss in property or income in the short run.
Federal Reinsurance Programs Now Exist For Other Perils, and They Provide Valuable Guidance
Federal Programs. In addition to various financial product insurance programs (e.g., FDIC, FSLIC, PBGC), the federal government is directly involved in insuring or reinsuring a number of property-casualty insurance perils or risks, including flood, crop, and liability for nuclear reactors. These programs were put into effect because the exposures were considered uninsurable in the private insurance marketplace.
It is worth noting, however, that each of these would appear substantially more insurable in the private marketplace than terrorism. First, they cover "fortuitous" (accidental) risks - unlike terrorism, in which the peril is man-made and the perpetrators have the avowed intention of killing as many people and destroying as much property as possible. Second, the dollar exposures are far less. The average annual losses for all three of these programs combined are only about $5 billion per year, a full order of magnitude lower than the $30-$50+ billion loss suffered on one day, September 11.
The federal involvement varies from program to program. For flood insurance, the government is the true underwriter of the risk, with the insurance industry acting only as an administrator. For crop insurance, the government and the participating insurers share the exposure through proportional insurance agreements. For nuclear liability insurance, the participating insurers underwrite only $200 million of liability in the event of a nuclear event. Liability above that is covered via assessments against nuclear reactor owners, up to a level of $9+ billion. Only if that is insufficient does the government step in pursuant to a statutory requirement to "consider" additional payments.
Problems. The flood and crop programs have been criticized for a variety of problems and abuses, including subsidizing loss-increasing behaviors (e.g., building in flood plains), poor management, and inefficiency.
On a more conceptual basis, any federal insurance program inherently involves forced assumption and allocation of risk on the funding (taxpayer) side and almost certainly involves some insured to insured subsidization. Insureds that can get government insurance at less than market costs have reduced incentives to take steps to minimize their exposure to loss - and it is the taxpayers who pay for this.
Public Policy Options
A range of options are available, from doing absolutely nothing and letting the market sort things out to establishing a federal insurance mechanism that would be the sole underwriter of terrorism and would market and service the policies. Below are brief descriptions of four options that might be considered along this continuum: (1) no federal backstop - let the markets work; (2) "barebones" federal backstop - under which insurers retain a very significant exposure; (3) substantial federal backstop - under which insurers retain some exposure, but most is transferred to the federal government; (4) total federal assumption of responsibility - with insurers merely servicing the policies. All of these except the first contemplate that the federal government would bear the risk of a mega-event, since insurers would otherwise be very unlikely to return to the market.
- No federal backstop - let the market work. In this scenario, insurers would withdraw terrorism coverage to the extent they could, with a particular focus on areas and properties considered to be potential targets. If they could not get regulatory approval to exclude the terrorism peril, they would eliminate or substantially reduce their commercial lines business, at least for high risk insureds, and raise rates to the extent they could.
Loss of insurance would likely cause some banks and other lenders to stop lending in certain areas and possibly to declare existing loans in default, due to failure to maintain satisfactory insurance coverage. Other real estate investors would likely redirect their capital. It is highly probable that properties and businesses in perceived target areas would lose significant amounts of income and value.
Small and medium size companies in urban areas could well be hit the hardest, facing loss of coverage and/or higher rates for various policies (e.g., workers' compensation, property, liability). Some might choose to close down their businesses rather than to continue to operate with no coverage and/or much higher insurance costs. Others might choose to "go bare," risking property or liability losses that would put them out of business anyway. And, of course, in that eventuality, any third parties with legitimate claims would in turn be unable to recover.
- Barebones federal backstop - under which insurers retain a very significant exposure. Congress might, for example, enact "promise to pay" language evidencing a commitment to step in and backstop the insurance industry in the event that terrorist losses exceed a specified amount, e.g., $10 - $20 billion. A time frame would presumably be specified, e.g., one or two years, but there would be little or no other specific involvement in the insurance marketplace or the regulation thereof.
While this option might be intuitively attractive to some in Congress and the Administration, insurers would still probably be unwilling to use scarce capacity to provide insurance to those that are perceived most at risk, particularly without reinsurance.
First, because most policies are written on an annual basis, a one-year time frame covering 2002 would clearly be too short. Insurers could not issue a policy in March, 2002, for example, if the backstop program did not provide protection for that part of the policy period that continued into 2003, when there would be no federal backstop. Even a two-year (2002-2003) time frame would not be much of an improvement in terms of sending a signal of stability. In less than a year, by the 4th quarter of 2002, insurers would be considering whether or not to renew policies written in early 2003. Since these policies would of course continue into 2004, i.e., into the time when there would be no federal backstop, insurers would again face uncertainty. Realistically, this would translate into uncertainty almost immediately for the business community, which would have to wonder whether Congress would revisit the backstop issue soon enough to ensure an ongoing market for terrorism coverage.
Second, assuming no preemption of state regulation, insurers would probably not be able to tailor their coverage or obtain the significant rate increases needed to offer coverage in terrorist target areas. Without reinsurance, and with federal involvement only at high-loss levels, insurers would still be significantly exposed (e.g., for a $20 billion event, an insurer with 5% of the loss would suffer a $1 billion pre-tax loss).
- Substantial federal backstop - under which the insurance industry retains a significant exposure, but most is transferred to the federal government, and individual insurers have some means of avoiding break the company losses within the industry retention amount. This might be structured in a number of ways. For example, insurers might be responsible for losses up to a specified amount, e.g., $10 billion, with government/insurers splitting losses 90-10 over that amount up to a second specified amount, e.g., $100 billion, with the government then assuming full responsibility for all losses. State insurance regulation of rates and forms would be overridden to provide more appropriate pricing for higher risk insureds (less cross-insured subsidization) and so terrorism coverage could be tailored and made uniform throughout the country and for insurers, reinsurers, and government coverage alike. Finally, some mechanism would be provided to facilitate risk spreading within the industry's retention level, e.g., a national reinsurance pool or a post-loss financing mechanism, so individual insurers need not risk crippling hits to surplus if they incurred a disproportionate share of loss.
This would facilitate a market in which insurers and reinsurers might be able to work together to spread the more manageable risk, while the multi-billion dollar exposure they retained would still encourage aggressive action to assess and reduce risk.
- Total federal assumption of responsibility - with insurers merely servicing the policies. Under this scenario, the insurance industry might continue to provide full coverage without terrorism exclusions but then have the government reinsure the insurers for terrorist acts.
This would be the most attractive option for insurers, but it would also present the most potential for government inefficiencies and inappropriate subsidization across insureds. Also, insurers would presumably be less incented to undertake risk reduction activities, e.g., ensuring that buildings are well maintained and that there are appropriate security and safety measures in place. While underwriting standards might be required, they are less likely to be as effective as they would be if private market insurers with "skin in the game" were involved.
A final comment is in order regarding competing public policy objectives. In the October 15 background briefing on terrorism insurance, the Administration noted that they did not want "any symbol of the nation to have exceptionally high insurance rates during this period of disruption," and that there was a role for the federal government to make sure that national symbols such as the Empire State Building did not have to pay exceptionally high rates. This is understandable in the current environment, and a comparable concern presumably applies to properties generally in New York and other high risk target cities - but it does run contrary to the operation of private markets.
In the short term, at least, certain properties and businesses, particularly those in large cities, will be perceived as inherently riskier than they used to be and inherently riskier than other properties and businesses still are. No investor will willingly commit capital to these riskier properties and businesses without the promise of a higher return - and probably a significantly higher return.
From an insurance standpoint, the riskier the building, business, or city, the more difficult it will be to get affordable insurance. Assuming insurance is not available, the same kind of analysis would inevitably apply for other risk-spreading mechanisms, and this presumably would result in losses to existing businesses and properties in likely target areas. The Congress and the Administration need to balance very carefully the interests of the owners and employees of those businesses and properties - many of whom in New York City at least have already suffered egregious economic and psychological injury - and the interest in reducing government involvement in private markets and inappropriate insured/insured and insured/taxpayer cross-subsidizations.
There are no easy answers. While private solutions are preferable to governmental, they are very unlikely to be forthcoming in the near term. This is also an issue that transcends pure economics and gets into the realm of what government should or should not do for the people (some guidance on this point may be found in the preamble to the U.S. Constitution, which includes as objectives the insuring of domestic tranquility, providing for the general defense, promoting the general welfare, and securing the blessings of liberty). Certainly, to the extent the insurance in question relates to terrorist attacks, which arguably might be considered to reflect failures in our security/defense system, the government should be more inclined to get involved than they would be for other kinds of coverage - and they will likely be driven to do so in light of the shaky economy and the political downside of being seen as abandoning the people and businesses in New York and other commercial centers at this time.
In that event, however, it is important to craft legislation that will: (1) commit the federal government to sufficient backstop protection that insurers can return to the market at some level and provide the kinds of risk reduction programs that the government does not do as well; and (2) include an exit strategy as part of the designated time frame for the program.
We recommend a federal backstop arrangement under which insurers would retain the first level of coverage and perhaps a percentage share above that, commensurate with their relatively small capital base; state regulation would be appropriately preempted with respect to rates and forms for terrorist coverages only; and risk-spreading would be facilitated. Insurers should have a stake in the system, so they can implement underwriting and prevention programs to reduce the risk of loss, e.g., to ensure the policyholders tighten security measures, develop and implement communications and emergency plans, and have access to appropriate medical resources.
The system should be put in place for at least three and preferably five years to send a clear signal that the government will ensure a stable market, subject to review going forward as markets and the war on terrorism evolve. Just as the federal riot reinsurance program put in place three decades ago was eliminated when the private marketplace returned, there must be an opportunity to phase out the federal backstop for terrorism insurance, but this is unlikely in just two to three years.
Such a partnering approach would encourage insurers to stay in the market and facilitate ultimate government withdrawal if and when terrorism became more manageable from a private marketplace standpoint.