Historical weather-related insolvencies illustrate the vulnerability of large and small insurers to the types of natural disasters that potentially are associated with climate change (Stipp, 1997; Swiss Re, 2000a; Mills et al., 2001). Nearly 650 U.S. insurers became insolvent between 1969 and 1998 (Figure 8-4b) (Matthews et al., 1999).
Of the 36 of 426 specifically attributed insolvencies occurring primarily as a result of natural catastrophes, more than half occurred between 1989 and 1993the period of Hurricanes Hugo, Iniki, and Andrewdespite increased insurer capacity (Davidson, 1997; Doherty, 1997; Matthews et al., 1999; Swiss Re, 2000b). Given the multi-factorial nature of most insolvencies, weather-related losses were no doubt a contributing factor in other cases as well. Although small or geographically specialized firms are most vulnerable, insolvencies of larger and more regionally diversified companies have occurred in the European Union (Swiss Re, 2000a) and in the United States (see Section 15.2.7). As a result of Hurricane Andrew, the largest U.S. home insurer, State Farm Fire and Casualty, was brought to the brink of insolvency by a US$4 billion loss and had to be rescued by its parent company (Stipp, 1997). The second largest U.S. home insurer, Allstate, paid out US$1.9 billion (which was US$500 million more than it had made in profits from its Florida operations from all types of insurance, including investment income, over the 53 years it had been in business) and also had to be rescued by its parent company (III, 2000a).
Little analysis of historic insolvencies in relation to natural catastrophes has been conducted outside the United States. Data for property/casualty firms in France, Germany, the UK, and the United States show that annual "baseline" insolvencies range from 0 to 0.5% of all firms to more than 2% in years with larger natural disasters (Swiss Re, 2000a). Natural disasters contributed to the severe difficulties experienced by the London market, including Lloyd's of London's near insolvency.
Future-oriented analyses of insurer solvency also have been conducted; these analyses show PMLs of US$45-100 billion, whichdespite several recognized conservatisms in the analyseswould result in serious levels of insolvency in the industry (ISO, 1996; Cummins et al., 1999; Kelly and Zeng, 1999). As many as 45% of insurers in the United States (representing 62% of the market share) could be placed in this position (GAO, 2000a,b). These findings are comparable to an earlier study showing that the rise in PMLs is stretching insurers' adaptive capacity (AIRAC, 1986).
Although much attention is focused on catastrophic loss events, "small" loss events are responsible for 50% of total economic and insured losses from weather-related events globally (Munich Re, 2000). If such events are closely spaced, they also can generate insolvencies or deplete surplus (Swiss Re, 1997; Ryland, 2000). Hybrid events involving multiple sources of insurance losses are of particular concern (White and Etkin, 1997; Francis and Hengeveld, 1998). This concern is exemplified in the case of ENSO events. A series of small events could be worse for insurers, in fact, than a single large event because individual losses per event often are capped (Stipp, 1997). Very little has been published on this subject since the mid-1980s (AIRAC, 1986).
Other reports in this collection