Major Natural Catastrophes, 1950–2001

The organizations with most interest in recording worldwide disaster statistics are humani‐tarian agencies such as the International Red Cross (see its annual World Disasters Report) and the large reinsurance companies. The latter are likely to be more meticulous. The data and the brief text below come from an annual report on natural disasters, Topics 2001, issued in March 2002 by Munich Reinsurance Company (more familiarly known as Munich Re), and are reproduced by permission. The trend in economic losses from “great catastrophes” (those requiring interregional or international assistance) over the last 50 years is strongly upward, as shown in the annual statistics and decadal comparisons. Both total losses and insured losses have been rising, the latter more sharply. An implication is that insurance premiums calculated on the basis of historical experience will underestimate future risks. The trend in disasters mainly results from greater exposure to risk through thegrowth of economies and populations rather than from changes in natural hazards themselves. One exception noted in the report is an increasing likelihood of “extreme precipitation” during hurricanes and other windstorms, which may be associated with global warming‐a warmer atmosphere can hold more water vapor. (There is no consensus on whether the frequency or intensity of storms is increasing, but there is evidence that they are getting wetter.) Greater allowance will have to be made for flood damage: Hurricane Andrew in the United States in 1992, causing a record $30 billion in estimated total losses, was a relatively “dry” storm‐ and it missed Miami and New Orleans. Munich Re describes 2001 as an “average year” for natural disasters, with an estimated 25,000 fatalities worldwide and total economic losses of some $36 billion. The four events during the year classed as great catastrophes were an earthquake and landslides in El Salvador in January, killing 845 persons; a magnitude 7.7 earthquake in Gujarat in the same month, killing more than 14,000; a hailstorm in Kansas City in April, killing no one but costing the insurance industry $2 billion; and Tropical Storm Allison in June, in which 72 m of rain over 12 hoursjlooded Houston, Texas, causing about $6 billion in total losses‐ “the costliest non‐hurricane of all time.” A recurrent concern of Topics 2001 is the problem of what are termed unidentified loss potentials. These are low‐risk but high‐loss events, exemplified by the 1999 Taiwan earth‐quake, which had an estimated return period of 10,000–100.000 years (or, more exotically, by the remote chance but catastrophic eflect o f a large meteorite impact‐the subject of one section of the report). Other kinds of hard‐to‐calculate loss potentials are related to unanticipated chains of events leading to or following from a disaster. The September terrorist attack on New York's World Trade Center, although unambiguously man‐made and thus not treated in the report, has been a further stimulus to wide‐angled thinking on loss potentials. It demonstrates the broad scope of worst‐case scenarios that now have to be considered by under‐writers who must seek to eliminate “the ‘bare patches’ on the ‘risk landscape.’” The full report is available online at http://www.munichre.com/pdf/topics~2001‐e.pdf