July 14, 2005
js-2633

Testimony of Treasury Secretary John W. Snow
Committee on Banking
U.S. Senate
on the Terrorism Risk Reinsurance Program

Thank you, Chairman Shelby, Ranking Member Sarbanes, and other members of the Committee. I appreciate the opportunity to discuss the Treasury Department report on the Terrorism Risk Insurance Act (TRIA).

As you know, President Bush signed TRIA into law in November of 2002 to help safeguard America's economy following the terrorist attacks of September 11, 2001. The September 11 losses led the insurance industry to reduce its exposure to future losses largely by excluding coverage of terrorism risk in many policies. The pullbacks in terrorism coverage and the quotations of rapidly increasing premiums raised concerns that this period of adjustment to the reality of global terrorism risk in the insurance market could have a negative effect on the economy.

In response, TRIA was passed. It was meant to address any market disruptions and ensure the continued widespread availability and affordability of property and casualty insurance for terrorism risk, and to allow a transitional period for the private markets to stabilize, resume pricing of such insurance, and build capacity to absorb any future losses, while preserving State insurance regulation and consumer protections.

TRIA required the Treasury Department to assess the effectiveness of the Terrorism Risk Insurance Program. It also required Treasury to assess the likely capacity of the property and casualty insurance industry to offer insurance for terrorism risk after termination of the Program.

The report finds that TRIA has been effective in meeting its goals of supporting the industry during a transitional period and stabilizing the private insurance market. Consistent with TRIA's design to encourage the development of the private market, the Administration opposes extension of the program in its current form. Extending TRIA in its current form is likely to discourage the private market development needed to deal with the risk of terrorism. The Administration has outlined principles that any extension should recognize and we look forward to discussions with the Congress on them. Before I review the main findings of the report, however, I would like to discuss the approach that the Treasury Department took in the course of evaluating TRIA.

Treasury contracted with an outside survey research firm to conduct two independent, nationally representative surveys. One survey sampled insurers in the commercial property and casualty line, which is eligible for the federal reinsurance provided under the Program. The other survey sampled policyholders, businesses and other organizations that purchased commercial property and casualty insurance in TRIA-eligible lines. Respondents were asked to provide information on an annual basis from 2002 (prior to passage of TRIA), to the first two months of 2005. The data therefore give a unique, comprehensive overview of the availability and affordability of terrorism risk insurance coverage in the private market.

From insurers, the surveys collected information on the amount of terrorism coverage written, the cost of terrorism coverage, terms and conditions on terrorism coverage and the purchase of reinsurance. From policyholders, we collected information on take-up and cost of terrorism coverage, the characteristics of firms and other organizations that purchase terrorism coverage, special terms and conditions associated with that coverage, reasons why the mandatory coverage offer was declined, and loss-mitigation efforts.

To safeguard the confidentiality of the business information requested in these surveys, Treasury took great care to ensure that the data were assembled at arm's length from the government. All identifying information was removed or masked prior to analysis by Treasury staff and officials.

I have been insistent throughout this process, consistent with Congress' direction to us in TRIA, that we draw upon as many sources of information and input as possible. Treasury has in fact consulted with a broad range of experts representing the insurance industry, the National Association of Insurance Commissioners (NAIC), policyholders, and taxpayer groups in developing the survey instruments. Preliminary survey instruments were reviewed by insurance industry representatives, NAIC representatives and others experts, including the American Insurance Association (AIA), and the Alliance of American Insurers (AAI) after consultation with its members. Members of the Coalition to Insure Against Terrorism (CIAT) also met with Treasury staff to review the policyholder survey.

We are very pleased with the extensive collaborative process that Treasury undertook to conduct this assessment, and believe that it reflects fully the extensive input of the industry and other groups. The completed survey results, and information derived from these other sources forms the basis of the Report to Congress.

TRIA established a temporary federal program of shared public and private compensation for insured commercial property and casualty losses resulting from foreign acts of terrorism. TRIA represents a form of publicly-provided and subsidized terrorism risk reinsurance, which essentially transfers risks associated with terrorism losses from the private to the public sector (taxpayers).

Under TRIA, companies that provide commercial property and casualty insurance are required to offer ("make available") terrorism coverage on the same terms and conditions as offered in their non-TRIA coverage. To be eligible for TRIA reinsurance, an act of terrorism must be certified by the Secretary of the Treasury, with the concurrence of the Secretary of State and the Attorney General., and must have resulted in aggregate property and casualty losses of $5 million or more. TRIA defines an act of terrorism as (1) a violent act or act that is dangerous to human life, property or infrastructure, that (2) has resulted in damage within the United States or outside of the United States in the case of an air carrier or vessel (as defined by TRIA) or on the premises of a United States mission, and (3) has been committed by an individual or individuals acting on behalf of any foreign person or interest, (4) as part of an effort to coerce the U.S. civilian population or influence the policy or affect the conduct of the U.S. government by coercion.

The federal government would have to cover 90 percent of insured losses beyond an insurer deductible, up to $100 billion per year. In the first full Program Year (2003) the deductible was 7 percent of 2002 premiums, in 2004 the deductible was 10 percent of 2003 premiums, and in 2005 the deductible is 15 percent of 2004 premiums. The purpose of the graduated deductible amounts was to encourage development of private market capacity over time. Insurers are also liable for 10 percent of losses above the deductible threshold.

In the event that the federal government provides compensation for insured losses for an act of terrorism under the Program, TRIA requires recoupment of at least a portion of the federal compensation through policyholder premium surcharges. Recoupment is mandatory in cases where the aggregate industry insured terrorism losses (deductibles and co-pays) are below a specified aggregate retention amount. The annual aggregate retention amount was $10 billion for 2003, $12.5 billion for 2004 and is $15 billion for 2005. The government is required to collect the difference between these recoupment amounts and the aggregate industry insured terrorism losses through an industry-wide surcharge, not to exceed 3 percent of the premium paid on a policy. If the aggregate industry insured terrorism losses exceed the aggregate retention amount, the federal government may require recoupment at the discretion of the Secretary of the Treasury, but the statute does not require recoupment.

To encourage the development of private market capacity over time, provisions in TRIA have gradually shifted more of the risk to the private sector.

The Treasury Department report finds that the Program provided support in a transitional period, during which the capacity of the insurance industry to write terrorism risk insurance has improved.

I will elaborate on four main findings in the report:

• Industry capacity to provide coverage for terrorism risk has improved, as has take-up of such coverage.
  
• Insurers are increasingly pricing terrorism risk insurance, and the price of coverage with an explicit charge has decreased.
  
• Industry surplus has improved.
  
• Many insurers reinsure a substantial portion of their retained risk under TRIA, but overall reinsurance purchases have not increased substantially.

Results from both the survey of insurers and the survey of policyholders show that the availability and the take-up (purchase) of terrorism insurance increased while TRIA has been in effect.

Insurers now provide terrorism coverage on a greater share of commercial property and casualty insurance policies than in 2002 (the year before TRIA). While 60 percent of policies written in 2002 included terrorism insurance coverage, fully two-thirds of such policies included such coverage in 2004. Terrorism insurance was also more widely available in the market, as the share of insurers providing any terrorism coverage rose from 73 percent to 91 percent over the period.

Policyholders as well are now more likely to purchase terrorism risk insurance than in 2002. The data show a doubling in the take-up rate of terrorism risk coverage: from 27 percent of policyholders in 2002 to 54 percent of policyholders by 2004. The finding that just under half of policyholders do not take-up such coverage does not necessarily reflect a problem in the market. The decision to purchase terrorism insurance reflects a tradeoff between the benefits and cost of the coverage. Firms that perceive a low risk of terrorism attacks or that have some form of self insurance (for example, through diversified portfolios) may simply not place a high value on terrorism insurance. It is useful to note that TRIA did not mandate the purchase of terrorism insurance, but rather that such coverage be made available.

Both insurer and policyholder surveys show that insurers increasingly began pricing terrorism risk insurance during the time TRIA was in effect. More than 75 percent of insurers providing coverage for terrorism risk in 2002 did not charge for it, but only 40 percent in 2004 provided coverage for free. These numbers are very similar to those reported by policyholders.

The average cost of terrorism insurance (measured as the share of total premiums paid for terrorism coverage) generally rose during this period. Overall, insurers reported costs ranging from 0.9 percent of premiums in 2002 to 1.8 percent by 2004. Among insurers who charged for terrorism insurance, the share of premiums charged for terrorism coverage first declined from 3.7 to 2.4 percent of premiums between 2002 and 2003, but then increased to 3.1 percent of premiums by 2004.

The average costs reported by policyholders increased from 1.2 percent of premium in 2002 to 1.6 percent in 2003, and further to 1.7 percent of premium by 2004. Among policyholders who reported paying for terrorism coverage, cost declined steadily over the period: from 4.0 percent of premium in 2002 to 2.8 percent in 2003 and further to 2.7 percent of premium in 2004.

Policyholders located in high-risk cities faced overall declining costs for terrorism risk coverage that varied from 2.8 percent of premiums in 2002, 3 percent in 2003 and 1.9 percent in 2004.

Industry surplus, a key source of insurer capacity, has returned to pre- September 11th levels. Insurers are financially stronger and more able to bear unexpected losses than they were prior to the enactment of TRIA. Reinsurance, another important component of an insurer's capacity to absorb losses, has not increased substantially, however. Seventy percent of insurers reported purchasing reinsurance for terrorism risk in 2003, but only 65 percent in 2004 reported purchasing reinsurance in 2004. Preliminary data from the first months of 2005 are encouraging and suggest a rebound to 75 percent. Smaller and medium-sized insurers generally reported greater use of reinsurance for terrorism risk exposure (TRIA deductibles and co-payments) between 2003 and 2005. During this same period, however, larger insurers reported less use of reinsurance for terrorism risk exposure.

The findings from the surveys of insurers and policyholders point to the success of TRIA in achieving its short term goals. TRIA effectively "addressed market disruptions and ensur[ed] the continued widespread availability and affordability of property and casualty insurance for terrorism risk." While we don't ascribe a direct causal effect to TRIA, we note that insurer financial strength has improved substantially over this period. More generally, TRIA allowed both insurers and policyholders time to adjust to the post-September 11th view of terrorism risk.

TRIA provisions shifted an increasing share of expected terrorism losses back to the private sector, as the deductible was increased from 7 percent of premiums in 2002 to 15 percent of premiums in 2004. Had there been no improvement in capacity, we should observe a pullback of terrorism coverage in response to this shift in cost. The expansion of terrorism risk coverage availability and take-up, and the decline in cost even as the TRIA deductible has increased therefore highlights the improvement in the industry's ability to cover terrorism risk

Congress also directed Treasury to assess the likely capacity of the property and casualty insurance industry to offer insurance for terrorism risk after expiration of the program. TRIA provided a federal backstop for terrorism losses that effectively subsidized terrorism risk insurance. It is reasonable to expect that the removal of the subsidy will result in adjustments in coverage and pricing. In the Treasury report, we present a framework to evaluate the impact of a TRIA sunset in more detail, and provide evidence from our surveys and from insurance industry statistics, data, and discussions with industry and other experts. Two important determinants of insurers' ability to effectively write coverage for terrorism in the near-term are the ability to model terrorism risk and the industry's financial capacity – including both surplus and access to reinsurance - to cover terrorism losses.

To provide and price insurance efficiently, insurers should be able to quantify their exposure to losses from terrorism risk. The primary tool available for quantifying loss exposure is modeling terrorism risk. Our assessment of developments in risk modeling over the past few years is positive, but we note that challenges do remain.

Modeling terrorism risk has two critical components: the ability to identify and quantify the severity of an event in terms of insurers' losses, and the probability of the loss occurring. Our study concludes that insurers' ability to identify and quantify the severity of an event in terms of insurers' losses has improved greatly. In particular, insurers are much better able to assess their exposures or accumulations of risk for a given terrorist event on an overall and individual customer basis. The industry – particularly the primary insurance industry - has made great progress in tracking aggregate exposure by location to estimate exposure to losses from physical damage and considerable progress in tracking aggregates of employees down to the level of individual locations to estimate exposure to workers' compensation losses. Modelers have created and implemented sophisticated probabilistic loss estimates that are said to take account of terrorists' shifting goals and strategies. Insurers writing coverage for high risk exposures are able to use multiple methods of assessing terrorism risk. This is important because it allows insurers to more effectively underwrite coverage. We acknowledge that the industry faces some difficulty in assessing the probability of the loss from terrorism. The uncertainty surrounding their predictions reduces the usefulness of these models.

An insurer's capacity to write coverage is limited to the maximum coverage it could provide, while retaining its ability to meet current and future obligations to its base of policyholders. An important determinant of insurers' capacity to cover terrorism losses is financial strength, which incorporates both balance sheet strength and operating performance. The financial health of the insurance industry, especially surplus, has improved greatly in the past three years. Among insurer groups providing coverage in TRIA-eligible property and casualty lines, surplus was higher in the third quarter of 2004 than it was in the third quarters of 2001, 2002 and 2003. Between the 3rd quarter of 2001 and the 3rd quarter of 2004, surplus increased from $256 billion to $341 billion. Measures of the industry's capacity to cover terrorism risk, including the ratio of net premiums to surplus, the return on surplus, and the capital adequacy ratio (accounts for underwriting, investment and credit risk) have all improved since the losses following the September 11 attacks.

By purchasing reinsurance, insurers can write additional coverage without increasing their financial holdings. Our survey results show that reinsurance is available, and purchased, for a sizable portion of the retained risk under TRIA. Seventy percent of insurers purchased reinsurance for TRIA-eligible risks in 2003. The results also indicate, however, that over the time period covered by our study, purchases of reinsurance have not increased substantially.

The expiration of TRIA will change the business environment in which insurers operate and will therefore change their behavior. Insurers will likely consider factors such as the possibility of insolvency from terrorism losses given the levels of surplus available and the effect on credit ratings. Experience with natural catastrophe risk underwriting and assignment of agency ratings suggests that in order to avoid ratings downgrades, insurers may significantly alter their approach to terrorism risk insurance after TRIA's expiration. Among the changes insurers may institute are increasing the use of private reinsurance, building surplus by tapping into capital markets, and raising premiums or placing exclusions on some policies.

Our surveys included direct responses on the availability of coverage after the expiration of TRIA. Responding to questions about policies written in early 2005 that continue into 2006, nearly 50 percent of insurers reported that they are not writing coverage for terrorism risks in 2006 (after the scheduled expiration of TRIA) that is similar to the coverage they write under TRIA. One-quarter of policyholders with terrorism risk coverage indicated that their coverage excludes terrorism coverage after the expiration of TRIA.

TRIA's expiration will conclude the transitional assistance first provided to the insurance markets in the uncertain economic environment of 2002. While the immediate effect of the removal of the TRIA subsidy is likely to be less terrorism insurance and, higher prices, we expect that over time the private market will develop additional terrorism insurance capacity. We anticipate that the initial response of premiums in the market will spur the buildup of surplus as insurers tap into capital markets; and the development of additional private reinsurance and other risk shifting mechanisms.

We do not believe that the elimination of the federal terrorism risk reinsurance subsidy is likely to have a discernable macroeconomic effect. In late 2001 and 2002, there was concern that there could be macroeconomic effects associated with the transition between a world in which terrorism coverage was provided for a negligible price and one where terrorism risk was considered a non-negligible risk. The economic climate during the discussion of TRIA and its enactment was highly uncertain. Industrial production had peaked in mid-2000, and by September 2001 had already fallen more than 5 percent. The terrorist attacks of September 11 created macroeconomic uncertainties that most analysts believed would translate into a further sharp downturn in economic activity that would last at least two additional quarters. Nonresidential building activity tumbled about 33 percent at an annual rate in the fourth quarter of 2001, and continued to experience declines well in excess of 15 percent in the subsequent three quarters. It was difficult at the time to assess whether the substantial declines in nonresidential building were due to the chilling effect of terrorist activity, terrorism insurance issues or the result of a cumulative unwinding of activity more typical of a recession and even the excesses of the late 1990s.

Helped by tax cuts and monetary stimulus, the economy has since improved substantially. GDP growth rose from just 2.3 percent in 2002 to 3.9 percent in 2004 (fourth quarter over fourth quarter). The unemployment rate, which was 6 percent in December 2002, fell to 5.1 percent in May 2005. However, despite the rising economy and the enactment of TRIA, nonresidential building has rebounded only slightly. Nonresidential building is currently 4.2 percent higher than the trough reached in the first quarter of 2003, but remains substantially below the previous peak. From our current perspective it appears that neither the potential lack of terror risk insurance nor a general economic downturn were responsible for weakness in nonresidential building activity.

The risk of terrorism changed fundamentally and permanently after the events of September 11, 2001. In the words of the President:

…Our country is safer than it was on September the 11th, 2001, yet, we're still not safe. . . . We are a Nation in danger. We're doing everything we can in our power to confront the danger. We're making good progress in protecting our people and bringing our enemies to account. But one thing is for certain: We'll keep our focus and we'll keep our resolve and we will do our duty to best secure our country."

It is our view that continuation of the program in its current form is likely to hinder the further development of the insurance market by crowding out innovation and capacity building. Consistent with TRIA's original purpose as a temporary program scheduled to end on December 31, 2005, and the need to encourage further development of the private market, the Administration cannot support a straight extension of TRIA.

Any reform of TRIA should be consistent with several principles. It is the Administration's view that extension of the program should recognize the temporary nature of the program, the rapid expansion of private market development (particularly for insurers and reinsurers to grow capacity), and the need to significantly reduce taxpayer exposure. The Administration would accept an extension only if it includes a significant increase to $500 million of the event size that triggers coverage, increases the dollar deductibles and percentage co-payments, and eliminates from the program certain lines of insurance, such as Commercial Auto, General Liability, and other smaller lines, that are far less subject to aggregation risks and should be left to the private market.

It is also important to keep in mind that the program would cover damages awarded in litigation against policyholders following a terrorist attack. Current litigation rules would allow unscrupulous trial lawyers to profit from a terrorist attack and would expose the American taxpayer to excessive and inappropriate costs. The Administration supports reasonable reforms to ensure that injured plaintiffs can recover against negligent defendants, but that no person is able to exploit the litigation system.

The events of the past week in London have been an unwelcome reminder that the risk of terrorism is real and that the war on terrorism is one that will be waged over a long period of time on many fronts. Some believe the fact that terrorism risk is real suggests the need for a permanent and obtrusive federal role in the market for terrorism risk insurance. I agree that the risk of terrorism is likely to remain a part of our lives for some time to come, but that is precisely why the federal government needs to encourage the development of the most creative and cost effective means of covering terrorism risks. The Administration looks forward to working with the Congress to achieve this end.

-30-