Summary

Terrorists using unconventional weapons, also known as nuclear, biological, chemical, or radiological (NBCR) weapons, could cause devastating losses. The Terrorism Risk Insurance Act (TRIA) of 2002, as well as the extension passed in 2005, will cover losses from a certified act of terrorism, irrespective of the weapon used, if those types of losses are included in the coverage. Because of a lack of information about the willingness of insurers to cover NBCR risks and uncertainties about the extent to which these risks can be and are being insured by private insurers across various lines of insurance, GAO was asked to study these issues. This report discusses (1) commonly accepted principles of insurability and whether NBCR risks are measurable and predictable, and (2) whether private insurers currently are exposed to NBCR risks and the challenges they face in pricing such risks. GAO collected information from and met with some of the largest insurers in each line of insurance, associations representing a broader cross section of the industry and state insurance regulators. GAO makes no recommendations in this report.

Insuring NBCR risks is distinctly different from insuring other risks because of the potential for catastrophic losses, a lack of understanding or knowledge about the long-term consequences, and a lack of historical experience with NBCR attacks in the United States. Measuring and predicting NBCR risks present distinct challenges to insurers because the characteristics of the risks largely diverge from commonly accepted principles used in determining insurability. According to these common principles, when assessing insurability, the risk generally must (1) have past occurrences sufficient in number and homogeneous enough (invoking the "law of large numbers") to enable insurers to accurately predict future losses, (2) be definite and measurable in terms of dollar value, (3) occur by chance, and (4) not result in catastrophic losses for the insurer. While the condition of insurability or uninsurability is not an absolute, NBCR risks generally fail to meet most or all of these principles of an insurable risk. Indeed, insurance experts GAO interviewed said that the potential severity of NBCR risks alone could diminish the willingness of some insurers to insure NBCR risks. Although NBCR risks may not fully satisfy the principles of insurability, there are enough variations in exposure across lines of insurance that some insurers or some lines of insurance may have no willingness to offer coverage for NBCR, while others may choose to offer coverage for some or all of the risks. For example, even with TRIA, property/casualty insurers generally have attempted to limit their exposure to NBCR risks by excluding nearly all NBCR events from coverage, both for commercial property/casualty and homeowners. According to industry representatives, property/casualty insurers believe they have excluded NBCR coverage by interpreting existing exclusions in their policies to apply to NBCR risks, but some of the exclusions could be challenged in courts. Unlike property/casualty insurers, however, workers' compensation, life, and health insurers are exposed to NBCR risks and generally have not excluded them from coverage for a variety of reasons. Specifically, workers' compensation insurers generally offer NBCR coverage because many states limit the exclusion of perils for workers' compensation. Conversely, while life and health insurers may not always be required to insure NBCR risks, they generally face other challenges in segregating and excluding NBCR risks. However, representatives of workers' compensation, life, and health insurers expressed concerns that the prices they currently charge may not cover their potential exposures to NBCR risks, sometimes because of regulatory limitations, and generally because of difficulties in measuring and pricing for NBCR losses. Given the challenges faced by insurers in providing coverage for, and pricing, NBCR risks, any purely market-driven expansion of coverage is highly unlikely in the foreseeable future.