1
Estimating the Losses of Natural Disasters

The United States has experienced natural disasters since the nation's founding over 200 years ago. During the 1980's and 1990's the United States experienced numerous, expensive—and frequently deadly—natural disasters. These disasters and their impacts have become part of our history: the drought in the central United States in 1988, Hurricane Andrew in 1992, and the Northridge California earthquake of 1994 were the nation's three most costly to date. Through the course of this committee's study, the nation experienced a severe ice storm in the northeast United States, devastating floods along the Rio Grande and in southeast Texas, heavy rains and landslides in southern California, a five-week succession of wildfires in Florida, a severe drought in south and east Texas, and three major hurricanes which impacted the southeastern United States. The committee also conducted its study during the fabled El Niño of 1997–1998 (see Box 1-1). This warming of surface water in the central and South Pacific and attendant effects on global weather patterns was—rightly or wrongly—blamed for many of the nation's and planet's weather-related hazards during this period, and ''some forecasters have warned the nation to brace itself for much more expensive disasters in the future" (Larson, 1998).

Natural Disasters: A Brief Introduction to their Effects

Anyone who has lived through a major hurricane or an earthquake does not have to be told of a natural disaster's destructive impacts. People are often killed and many others are injured. Homes, office buildings, shopping centers, highways, and other physical facilities are destroyed.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 3
--> 1 Estimating the Losses of Natural Disasters The United States has experienced natural disasters since the nation's founding over 200 years ago. During the 1980's and 1990's the United States experienced numerous, expensive—and frequently deadly—natural disasters. These disasters and their impacts have become part of our history: the drought in the central United States in 1988, Hurricane Andrew in 1992, and the Northridge California earthquake of 1994 were the nation's three most costly to date. Through the course of this committee's study, the nation experienced a severe ice storm in the northeast United States, devastating floods along the Rio Grande and in southeast Texas, heavy rains and landslides in southern California, a five-week succession of wildfires in Florida, a severe drought in south and east Texas, and three major hurricanes which impacted the southeastern United States. The committee also conducted its study during the fabled El Niño of 1997–1998 (see Box 1-1). This warming of surface water in the central and South Pacific and attendant effects on global weather patterns was—rightly or wrongly—blamed for many of the nation's and planet's weather-related hazards during this period, and ''some forecasters have warned the nation to brace itself for much more expensive disasters in the future" (Larson, 1998). Natural Disasters: A Brief Introduction to their Effects Anyone who has lived through a major hurricane or an earthquake does not have to be told of a natural disaster's destructive impacts. People are often killed and many others are injured. Homes, office buildings, shopping centers, highways, and other physical facilities are destroyed.

OCR for page 3
--> Box 1-1 El Niño Caused Natural Disasters of 1997–1998 A recent series of weather disasters were caused by the record largest El Niño of 1997–1998, an event to which several intense storms across the United States were attributed. The term El Niño ("the child" or "the Christ child") refers to a warming of surface waters in the eastern tropical Pacific Ocean. Through history, this pool of warm surface water periodically appeared off coastal Ecuador, Peru, and Chile near Christmas. The tremendous amount of energy contained within this water has the ability to disrupt atmospheric patterns throughout the region and across the planet. The El Niño which developed in the eastern tropical Pacific during the summer of 1997, and lasted through much of 1998, was exceptionally warm and ultimately "blamed'' for several weather-related disasters in the United States during that period. As forecasted when the El Niño developed during June–August 1997, California was assaulted by a series of coastal storms and heavy rains causing floods, numerous landslides, and damages to the state's valuable agriculture, with losses totaling nearly $1 billion statewide. Florida, Texas, and other southern states were struck by a large number of severe storms and numerous tornadoes. Tornadoes led to more than 60 deaths, and El Niño-caused losses in Florida exceeded $500 million. A record early snowstorm swept across the High Plains in October 1997, and severe ice storms struck the Northeast in January 1998, creating losses in excess of $300 million and 28 deaths. The effects of El Niño on storm activity occurred from September 1997 through April 1998. The property insurance industry identified 15 catastrophes (events each causing greater than $25 million) during the 8-month period ending in May 1998, when El Niño's influence on the weather had largely ended. The total insured losses by these catastrophes reached $1.7 billion. Even after the storm activity ended, more damages occurred. Widespread fires broke out in Florida during June, fueled by a heavy growth of underbrush caused by the unusually heavy El Niño-caused winter rains. FEMA relief payments reached $300 million by the end of March 1998 and 18 presidentially declared disasters occurred from the fall of 1997 through April 1998, and all were partly attributed to El Niño's influence on the atmosphere. El Niño events have become stronger and more frequent since 1980, certainly one reason for the increased losses from weather-related natural disasters over the past 15 years.

OCR for page 3
--> In true megadisasters—such as the Northridge earthquake in California or Hurricane Andrew in Florida—the destruction can severely interrupt work, traffic, and the daily routine of a large area for months and, to some extent, years after the event. Experts in the field of disaster cost estimation use various terms to describe the effects of disasters, not always consistently. It is therefore important to define at the outset how these terms are used in this report: The impacts of a disaster is the broadest term, and includes both market-based and nonmarket effects. For example, market-based impacts include destruction to property and a reduction in income and sales. Nonmarket effects include environmental consequences and psychological effects suffered by individuals involved in a disaster. In principle, individual impacts can be either negative or positive, though obviously the impacts of disasters are predominantly undesirable. The losses of disasters represents market-based negative economic impacts. These consist of direct losses that result from the physical destruction of buildings, crops, and natural resources and indirect losses that represent the consequences of that destruction, such as temporary unemployment and business interruption. The costs of disasters, as the term is conventionally used, typically refers to cash payouts by insurers and governments to reimburse some (and in certain cases all) of the losses suffered by individuals and businesses. Losses suffered by those who are uninsured, those whose losses do not make them eligible for insurance payments, and those who do not receive government relief should be counted in any complete compilation of the impacts of a disaster—but these losses are not included as "costs," as that term is used in this report. The damages caused by disasters refers to physical destruction, measured by physical indicators, such as the numbers of deaths and injuries or the number of buildings destroyed. When valued in monetary terms, damages become direct losses. The formal charge to this committee was to "identify the cost components that, when combined, would most accurately reflect the total cost of a natural disaster event. To the extent possible, the committee will identify the relative importance of the components for accurate characterization of an individual event and the significance of the different components across the spectrum of hazards. The committee will also suggest possible sources for accurate cost information, regardless of whether data are generally available from these sources at present."

OCR for page 3
--> Given the formal distinction made between losses and costs, the committee felt that its deliberations and report should focus upon the losses and human impacts of natural disasters (which include costs), rather than the costs alone. One important request of this committee was to identify those data which should be consistently collected in compiling estimates of a natural disaster's impact. The committee felt that these data should include not only the cash payouts from governments and insurance companies (costs), but also a wider range of impacts. This report concentrates on how best to measure the economic losses in disasters, as they are either the most easily measured or the best understood. Nonetheless, it was recognized that noneconomic consequences of disasters, such as environmental impacts, can be very important and, in some instances, may exceed direct economic losses. Accordingly, a discussion on the environmental effects of disasters is included in Appendix A. Studies of the impacts of natural disasters vary in their coverage of losses. Some measure only direct losses whereas others purport to include indirect losses. Despite the difficulty of comparing different loss estimates, it is useful to be aware of the significant impacts disasters—especially megadisasters—can have. Based on available studies, Table 1-1 summarizes the four most expensive natural disasters in recent American history. TABLE 1-1 The Four Most Expensive Natural Disasters in the United States (in current dollars at the time of the disaster) Year Event Reported Loss Source 1988 Droughta $39 billion Riebsame et al., 1991 1992 Hurricane Andrew $30 billion Pielke, 1995 1993 Midwest floodsa $19 billion Changnon, 1996 1994 Northridge earthquake $44 billion Eguchi et al., 1998 a Direct and indirect losses combined. Other estimates refer only to direct losses. Table 1-1 is interesting, not only because of the magnitude of the costs of these large disasters, but also because it highlights their diverse nature: no single type of disaster dominates the list. It bears emphasis that the loss estimates illustrated in Table 1-1 should be viewed as best guesses, for there is no official disaster cost accounting system in the United States. What information we have must be compiled from different studies that use different estimation techniques and raw data sources. In addition, although insured property losses tend to be estimated reasonably well, there is

OCR for page 3
--> much greater uncertainty about uninsured losses, which can be substantial. This point is underscored in Table 1-2, which shows the top disasters ranked by their property insurance claims. Notice that on this list Hurricane Andrew ranks first and Northridge second. Neither the 1993 Midwest floods nor the 1988 drought even make the list because most of the losses in those events were not privately insured or were indirect. TABLE 1-2 Ten Costliest Natural Catastrophes, Ranked By Insurance Claims Paid Month/Year Catastrophe Estimated insured loss ($ Millions) August 1992 Hurricane Andrew: wind, flooding, tornadoes 15,500 January 1994 Northridge, California, earthquake 12,500 September 1989 Hurricane Hugo 4,195 October 1995 Hurricane Opal 2,100 March 1993 20-state winter storm 1,750 October 1991 Oakland, California fire 1,700 September 1996 Hurricane Fran 1,600 September 1992 Hurricane Iniki 1,600 May 1995 Texas and New Mexico, Wind, hail, flooding 1,135 October 1989 Loma Prieta, California earthquake 960 NOTE: Dollar losses are stated in the year of the event. SOURCE: Insurance Information Institute, 1998. The United States has borne significant costs not just from the major events shown in Tables 1-1 and 1-2 but also from the combination of the hundreds of smaller natural disasters that occur every year. Catastrophes in the United States are identified, tracked, and reported by Property Claims Services (PCS), an insurance industry organization specializing in the property insurance business. PCS is best known for its work in catastrophes and is recognized internationally as the singular source of information concerning insured damage resulting from most major natural disasters in the United States. PCS also serves as the property insurance industry's liaison with federal and state government officials, consumer

OCR for page 3
--> organizations, and the media regarding insurance issues associated with catastrophes. Property Claims Services was originally a division of the National Board of Fire Underwriters, which initiated the catastrophe identification system in 1949. This system was developed to address the insurers' needs to quantify the impact of catastrophes on insurance coverages. As more Americans bought homes and moved to suburban areas after World War II, there was little information about severe weather events across the country. Since 1949, PCS has identified over 1,200 catastrophes affecting U.S. insurers. PCS currently identifies a catastrophe as an event which causes $25 million or more in insured, direct property damage, and affects a significant number of property owners. This threshold has been adjusted through time: when PCS was founded in 1949, $1 million of insured, direct property damage constituted a catastrophe, in 1983 it was revised to $5 million, and was revised for a third time in 1997 to $25 million. For those events that exceeded the PCS definition of a catastrophe, the costs have increased as follows: 933 major events resulted in costs of $22 billion from 1949 to 1988; 312 major events resulted in costs of $79 billion from 1989 to 1997 (Kunreuther and Roth, 1998).1 The causes of the mounting costs of disasters in the United States (Figure 1-1) will continue to be debated. But certainly a major contributing factor has been the increased exposure of property and human beings to disasters. The nation's population has grown significantly since World War II, and more people than ever before live and work in disaster-prone areas—especially coastlines, floodplains, and seismically active regions. Various forms of economic development have also driven up the costs of natural disasters. For example, the destruction of wetlands, the clearing of forests for a range of human activities, and the paving of roads and parking lots all have increased the peaks of runoff from heavy rainfall. Furthermore, greater wealth means we simply have more to lose in disasters. Policymakers and citizens should be aware of these trends as they make decisions which affect the location of future housing and business developments. These societal trends help explain predictions of continuing increases in the losses from future natural hazards. 1   These costs have not been adjusted for inflation and for changing insurance coverage.

OCR for page 3
--> Figure 1-1. The increasing trends—based on catastrophes that caused losses between $10 million and $100 million—for five-year periods of the number of catastrophes, the amount of loss from these catastrophes, and the U.S. population (Changnon et al., 1997). Why Loss Estimates Matter Who should care about disaster loss estimates, and why? There are several reasons which matter to different constituencies. The federal government has a strong interest, on behalf of taxpayers, in accurate and comprehensive direct loss data. Largely through the Federal Emergency Management Agency, but also through several other agencies, the federal government covers a substantial portion of disaster-induced losses that are suffered by local and state governments, as well as by individuals and businesses, that are otherwise privately uninsured. Figures 1-2 and 1-3 provide data on federal government payouts from 1988 to 1997 and 1986 to 1997. As these figures demonstrate, FEMA provides only part of the federal disaster assistance effort. Other key federal agencies with disaster assistance programs include the Small Business Administration (SBA), U.S. Department of Agriculture (USDA), U.S. Department of Transportation (DOT), U.S. Department of Interior (DOI), U.S. Army Corps of Engineers (USACE),

OCR for page 3
--> Department of Health and Human Services (HHS), National Weather Service (NWS), Department of Energy (DOE), and Department of Labor (DOL). 2 Although extreme geophysical events cannot be avoided, their impacts can be reduced by government policies, or hazard mitigation initiatives such as reinforcing structures to enable them to better withstand the shock of an earthquake, elevating structures to reduce flood damages, land use planning to decrease structural exposures to natural hazards, and other measures. Because mitigation can be costly, however, it is important for policymakers at all levels of government also to be aware of the total losses of disasters—and ideally of the extent to which those losses can be reduced by various mitigation strategies—so that cost-effective mitigation strategies can be designed and implemented. 3 The same is also true for the private sector, where cost-effective mitigation measures can and should be used to reduce losses in future disasters. Figure 1-2. Major federal disaster assistance programs (Rhinesmith, 1997). 2   For a list of the most costly disasters (to FEMA), see Table 1-3. 3   For a recent discussion of various mitigation techniques, see FEMA, 1997. In pursuing mitigation, policymakers should take account not only of the direct costs that such measures may avoid but also of the total disaster-related losses—that is, the claims paid by private insurers and government plus the uninsured losses that victims of disasters suffer (including, to the extent practicable, estimates of the indirect losses, which are discussed in Chapter 3)—that may be prevented.

OCR for page 3
--> Figure 1-3. FEMA, SBA, and DOT disaster obligations, 1986–1997 (Rhinesmith, 1997). Policymakers should have an interest in accurate direct loss data for reasons that extend beyond mitigation. At this writing, Congress is considering proposals to establish a federal reinsurance program to support the private sector's supply of homeowners insurance for claims related to natural disasters (although critics of these proposals argue, among other things, that the expansion of reinsurance capacity in recent years makes federal intervention less necessary). If such a program is enacted, historical loss data will be essential both to potential buyers and the government (as the seller of the reinsurance, a type of insurance the government makes available to insurers to help protect them in the event of catastrophic events) so that all parties can be informed when bidding for or negotiating reinsurance contracts. An understanding of losses and their growth has been the basis for enacting foresighted legislation. Ex post analysis of the 1993 Mississippi floods, for example, led to needed revisions in crop insurance, flood insurance premiums, as well as broader programmatic changes in flood insurance. Insurers clearly have a commercial interest in accurate data on costs covered under their policies: claims for property damage and related economic

OCR for page 3
--> costs and for loss of life and injuries. Insurers also maintain their own payout data in order to operate their businesses. But to set actuarially sound rates, each insurer needs comprehensive data on payouts made by all insurers. Historical claims data must be combined with models that project the probabilities of future events over a range of possible magnitudes—as uncertain as these probabilities are—so that insurers can estimate expected future payouts. State insurance regulators must know and be able to project disaster related costs in order to properly regulate or monitor insurance premiums. Individuals and businesses have an interest in knowing the costs and losses of different types of disasters to help inform decisions about insurance, mitigation, and government policies for compensation of victims. As mentioned, however, there is no comprehensive disaster loss information available from either public or private sources. Imagine if that were the case for statistics about the overall U.S. economy: no indicators of national or regional unemployment, output, or inflation. Policymakers would be severely hampered in their efforts at setting sound economic and monetary policies. Private actors in the economy would not know whether it was in their interest to increase or reduce their investments, consumption, or other economic activities. Finally, researchers and experts in disaster loss estimation could benefit enormously from having a standardized data base of losses and other impact information that would permit them to improve their models that estimate both direct and indirect losses of disasters. Such improvements, in turn, would assist policymakers in designing cost-effective mitigation policies by simulating benefits and costs of alternative policies. Fortunately, some cost data for natural disasters are collected in the public and private sectors. As discussed in more detail in Chapter 2, individual agencies of the federal government generally know how much they spend on compensating victims of disasters. So do insurance companies. But data are not now generally available for uninsured and indirect losses. More broadly, there is no comprehensive and reasonably accurate data base that pulls all of the relevant information together in a way that can be easily accessed and used by a variety of professionals. In making this observation, we do not mean to suggest that the federal government should create a data base containing loss information for every adverse natural event. Instead, the objective should be to compile data on disasters that cross some threshold. This issue is also explored in Chapter 2.

OCR for page 3
--> Who Bears the Loss? It is important not only to assemble the total losses of disasters but to apportion those losses among all those who bear them—at least initially. Citizens and policymakers should know to what extent the losses are privately borne, especially through insurance, and what portion is paid by local, state, and federal governments. From the national taxpayer perspective, policies should aim to internalize to the maximum extent the losses of disasters—that is, induce individuals, businesses, and local and state governments to recognize and accept responsibility to purchase insurance (or to establish the functional equivalent of self-insurance pools) or mitigate such losses rather than to have the federal government come in after the fact and pay for the expense of reconstruction and cleanup. But of course there are limits to the losses insurers bear. For example, although the typical homeowners' insurance policy covers damages to the physical structure of a home, some policies may not cover the entire replacement cost (unless such protection is specifically purchased). Homeowners' policies also may cover some, but not all, of the contents of a residence. In addition, private homeowners' insurance typically does not cover flood loss (flood insurance must be purchased separately through the National Flood Insurance Program, managed by FEMA), and it may not cover earthquake damage unless specifically purchased (and even then, coverage may be limited). A common dilemma that disaster managers and policymakers face involves the issue of extending federal financial assistance to homeowners who refuse (or who are unable to do so due to low income) to adequately insure their property against disasters, despite warnings or previous experience with the same type of disaster. Extending federal assistance to those who did not insure and are financially capable of doing so discourages the purchase of insurance and the adoption of appropriate mitigation measures. Starting in the late 1980's, FEMA received a regular annual appropriation of $320 million to cover the costs of its relief efforts. For fiscal year 1999, Congress reduced this figure to $308 million. FEMA's regular annual appropriations, however, seldom cover all the costs of federal disaster relief it is responsible for paying in a single fiscal year. Table 1-3 lists the ten costliest disasters measured by FEMA relief payments. Clearly, the sums shown in the table have greatly exceeded FEMA's relatively meager annual appropriations. FEMA has been able to make its payments only by receiving supplemental appropriations at various times. Until the Northridge earthquake, these supplementals were treated as emergencies for budget purposes and simply added to the deficit. Beginning with the Northridge earthquake supplemental

OCR for page 3
--> request, Congress required all disaster supplementals (for FEMA and other federal agencies) to be offset with cuts in expenditures elsewhere. TABLE 1-3 Top Ten Disasters Requiring FEMA Assistance (Through Mid-1998) a   Event Amount of Assistance (in millions of dollars) 1994 Northridge, California, earthquake 5,997 1992 Hurricane Andrew (Florida, Louisiana) 1,772 1989 Hurricane Hugo (North Carolina, South Carolina, Puerto Rico, Virgin Islands) 1,318 1993 Midwest floods l,147 1989 Loma Prieta, California, earthquake 837 1996 Hurricane Fran (Eastern states) 560 1995 Hurricane Marilyn (Puerto Rico, Virgin Islands) 524 1994 Tropical Storm Alberto (Alabama, Florida, Georgia) 438 1997 Upper Midwest floods (Minnesota, North Dakota, South Dakota) 425 1996 Mid-Atlantic and northeast floods 366 SOURCE: FEMA, 1998. a The table excludes Hurricane Georges because payout data were not available at the time the report was prepared. The dollar figures are in nominal terms (and thus have not been adjusted for inflation). It is important to keep in mind that the federal government is not the only level of government involved in providing disaster assistance. Local and state governments also shoulder significant responsibility for managing emergencies. Federal agencies, particularly FEMA, stimulate and guide emergency planning efforts, furnish substantial response and recovery funding, coordinate response efforts after (and sometimes before) a governor secures help from the president, and fund many disaster mitigation endeavors.4 In addition, significant costs are borne by volunteer disaster relief organizations, which are important sources of loss data. 4   For a discussion of the various sources of disaster relief see Kunreuther and Roth (1998), pp. 10–11.