National Academies Press: OpenBook

The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum (1992)

Chapter: 2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?

« Previous: 1 WHAT IS KNOWN ABOUT SEISMIC RISK NATIONALLY?
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

2
What are Likely Categories of Loss and Damage?

Chapter 2 provides a basis for understanding how loss estimates are generated for different categories of losses and damages. Specific issues that will be addressed include: What loss-estimation methodologies currently exist? How satisfactory are inventories for these approaches? In relation to actual losses, how good are these estimates? Are there methodologies for estimating nonstructural losses? What data bases are available from which projections of nonstructural losses can be made?

Mr. Christopher Arnold is an architect and the president of Building Systems Development in California. He is a Fellow of the American Institute of Architects, elected for his contributions to research in the architectural aspects of seismic design. Mr. Arnold is also a member of the NAS/NRC Committee on Earthquake Engineering and has served on that committee's panel on loss-estimation methodologies. Mr. Arnold will present an overview of loss-estimation approaches.

Dr. Don G. Friedman is the director of the Natural Hazards Research Program at the Travelers Insurance Company. Dr. Friedman has 35 years of experience in the assessment of casualty and damage potentials of natural disasters for the insurance industry, federal agencies, and most recently for the All-Industry Research Advisory Council. His presentation focuses on risk-assessment models and the types of data necessary to estimate casualty and property loss potentials from a catastrophic earthquake.

Professor Kathleen Tierney is an associate professor of sociology and the research director of the Disaster Research Center at the University of Delaware. She is the author of a number of monographs, articles, and book chapters focusing on various hazard- and disaster-related topics, induding socioeconomic consequences of earthquakes. She is a member of the Advisory Committee for the National Earthquake Hazard-Reduction Program. The topic of Professor Tierney's presentation is on loss estimation and public policy from a social science perspective.

Professor Robert W. Kling is from Colorado State University. Dr. Kling has a doctoral degree in economics from the University of Kansas and has recently been involved in three National Science Foundation (NSF) projects that have addressed different aspects of social and economic effects of different types of natural hazards. The result of one of these projects is a work entitled Natural Hazards Damage Handbook: A Guide to the Uniform Definition, Identification, and Measurement of Damages from Natural Hazard Events. In his presentation, Dr. Kling focuses on the loss of the cultural environment from a catastrophic earthquake.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

PRESENTATION OF CHRISTOPHER ARNOLD

This section provides an outline of the methods that are currently used for developing estimates of loss, and an outline of the nature of loss focusing on one particular building in the Loma Prieta earthquake. The information on loss-estimation methods is based on the Academy panel study on estimating losses from future earthquakes conducted under the chairmanship of Bob Whitman a few years ago. But the following information does not necessarily represent the views of the panel or the Academy.

The typical parameters of a loss-estimation study shows that the following things have to be determined before developing a specific study:

  • the type of loss (e.g., casualties, the number of homeless, the functionality of essential facilities, and economic impact);

  • the kinds of facilities (i.e., all buildings or structures and essential facilities like hospitals and lifelines);

  • the degree of certainty, and some feel for the degree of detail which is necessary—these obviously have very large cost and time implications; and

  • the time span and the kind of seismic risk.

Perhaps the number of earthquakes which might occur in a time span is of particular interest. Either a predicted earthquake or an actual historic earthquake may be used. Studies have been done, for instance, that show the impact of the 1923 Kwanto earthquake on today's Tokyo. Also, before developing a loss-estimation study, questions of geographic scale—local, regional, state, or national—must be addressed.

Typically, the kind of loss-estimation studies that have been done, other than those which may have been done for very specific purposes such as the inventory of a large corporation's set of buildings, are studies such as the NOAA studies of the San Francisco Bay area in the 1970s; the FEMA studies in selected cities, such as St. Louis and Boston; or the midwestern six-city study. These have provided damage estimates that are expressed in dollar terms and estimates of casualties. They tended to deal with all buildings, although some of them have focused somewhat on essential facilities such as hospitals. The degree of certainty is probably low. The degree of detail, because of the cost of producing the study, is also low. The estimates of seismic risk typically use the modified Mercalli scale, for better or for worse. It is at least generally understood and accepted, and has been commonly used as a way of defining the seismic risk. These studies typically have been at a regional or large-city level.

The use of these studies has primarily been political. They have been used to assist the politics of earthquake-hazard mitigation and the process of consciousness raising, and they have also been used to support some of the commercial aspects of the "earthquake industry," under the new circumstances by which the earthquake problem is starting to become a recognized industry.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

Recognizing that these studies generally have a large amount of scientific qualification on the way in which they have been done and the authority of the statements, the political use of these studies has been to look for the large numbers" and to use those to attempt to increase earthquake awareness and hazard-mitigation activities.

The methodology of loss estimation is extremely simple and obvious. Some time ago Mr. Arnold wrote a paper on this, which stated that methodologies are cheap, but data is very, very expensive. All studies follow the same basic methodology:

  • develop an inventory of building stock;

  • define the seismic risk that applies to those buildings; and

  • utilize a mechanism for relating motion, damage, and loss so that damage to the inventory can be estimated.

This is then typically converted into the dollar loss that is received from exposure to a given seismic risk, as applied to that inventory, and out of that comes a loss estimate definition. That is the general methodology. To focus on one particular aspect of it, the inventory and the motion-damage-loss mechanism must use the same building classification, therefore a classification for defining the inventory of buildings must be developed that has to be the same classification used to apply the motion-damage-loss mechanism to the inventory. That sounds obvious and simple but, in fact, it has proven to be rather difficult.

Table 2-1 shows a classification system for buildings which is very widely used; it was developed originally by Karl Steinbrugge for the Insurance Service Offices in the Bay Area and has been slightly modified since. It is a simple and broad classification, with 21 different building types. One who is unfamiliar with buildings may think that is a lot. In fact, it is a very small classification because when a building inventory is inspected, every building must be assigned to one of these 21 types. Within each type there will be a great deal of variation; that presents difficulties, but nevertheless this is a very widely used system.

Another system was developed somewhat later under a program called ATC-13 (Applied Technology Council Study Number 13). This was a FEMA-sponsored study which was intended to bring the loss-estimation technique to a more advanced level. It was also intended originally to go right through to loss-estimation studies which would be used to estimate economic losses, industrial capacity losses, and so on. That was never quite achieved, but a lot of the study intent was accomplished. The study uses a rather more complete classification system than the Insurance Services method. The classification system has 40 building types instead of 21, so that it is a slightly freer-grained classification system.

For any study, the inventory is critical, because this is the whole basis upon which you are going to assign your loss estimation. Unlike other aspects of loss estimation, an actual inventory exists. In other words, there is a finite

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

TABLE 2-1 Construction Classes Used in the ISO and NOAA/USGS Methods

Building Class

Brief Description of Building Subclasses

1A-1

Wood and stuccoed frame dwellings regardless of area and height

1A-2

Wood and stuccoed frame buildings, other than dwellings not exceeding three stories in height or 3,000 square feet in ground floor area

1A-3

Wood and stuccoed frame structures not exceeding three stories in height regardless of area

1B

Wood frame and stuccoed frame buildings not qualifying under class 1A

2A

One-story, all metal; floor area less than 20,000 square feet

2B

All metal buildings not under 2A

3A

Steel frame, superior damage control features

3B

Steel frame, ordinary damage control features

3C

Steel frame, intermediate damage control features (between 3A and 3B)

3D

Steel frame, floors and roofs not concrete

4A

Reinforced concrete, superior damage control features

4B

Reinforced concrete, ordinary damage control features

4C

Reinforced concrete, intermediate damage control features (between 4A and 4B)

4D

Reinforced concrete, precast reinforced concrete, lift slab

4E

Reinforced concrete, floors and roofs not concrete

5A

Mixed construction, small buildings and dwellings

5B

Mixed construction, superior damage control features

5C

Mixed construction, ordinary damage control features

5D

Mixed construction, intermediate damage control features

5E

Mixed construction, unreinforced masonry

6

Buildings specifically, designed to be earthquake resistant

number of buildings of certain kinds. The problem is that it almost never exists in any published form, and the costs of achieving that are astronomical; when one talks about defining an inventory, he is really defining some kind of simulation or subterfuge for the actual inventory that exists.

The Academy panel spent 2 days discussing this inventory question: how is it determined? and how is an inventory defined? For instance, the census does not reveal the things about buildings needed to determine loss estimation. Assessor's records do not indicate the things needed to know about loss estimation. The top half of Figure 2-1 shows an idealized version of achieving

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

FIGURE 2-1 Earthquake-damage-loss estimation.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

an inventory, in which the buildings would have some sort of computerized system that would send out their accurate vital statistics. The lower picture in Figure 2-1 shows the way inventories are actually done: A small number of people gather together in a room and assign buildings to their correct classification. You will find that the smoke-filled room reoccurs rather often in the loss-estimation methodology.

The loss and damage mechanism, developed primarily by Karl Steinbrugge, is critical and is shown in Figure 2-2. It is simple and relatively easy to apply. The designations of the building classification system are shown as the curves: the modified Mercalli intensity is on one axis and the per

FIGURE 2-2 Loss ratio versus modified Mercalli intensity (mean damage ratio curves). From: Estimating Losses from Future Earthquakes, p. 29, National Research Council, 1989; Source: Algermissen and Steinbrugge, 1984.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

centage loss is on the other axis. Select a given type of building, such as unreinforced masonry, and find the curve for it, then find the modified Mercalli intensity, and establish a percentage loss. These are obviously very gross figures; as they are studied, all kinds of reasons why they may be inaccurate can be though of. This, however, is currently the level at which these estimates are done.

The ATC-13 process developed a somewhat more elaborate methodology, and the team was also interested in making the methodology more transparent. Rather than the smoke-filled room, the idea was to have a methodology in which its method of development was clear and could continue as more information came in. The project developed a general matrix, called a damage probability matrix, which defines damage states in words. These are defined numerically and can in turn be developed into a dollar-loss estimate related to the modified Mercalli scale. The estimates, of course, vary according to the building type.

The actual estimates were developed by expert committees in a delphi process with a number of experts filling in forms and, in effect, voting on their estimates of the correct numbers. Figure 2-3 shows the scatter: for one particular class of building, the low estimate, the best estimate, and the high estimate. The symbols show the estimates of specific people. There were two rounds, with a fairly small number of people involved. The object of the two rounds was to try and reduce the aberrations in the estimating, although sometimes the aberrations may be correct and the ''enforced'' agreement may not be correct.

Nevertheless, the process arrives at a set of numbers that can be used in the actual matrices. Figure 2-4 shows the ATC-13 matrix for facility class 18, which is a low-rise, concrete, movement-resistant, frame-building type. It can be seen that the matrix shows fairly small amounts of damage at even the high modified Mercalli figures, and 100 percent damage would be expected in this class of building for any modified Mercalli figure. One can agree or disagree with that, but this procedure enables a dollar figure to be arrived at for a given building type related to a range of modified Mercalli figures.

There is another system which has been used. This is the "fragility curve," which was developed by a consultant and used for one of the FEMA studies (Figure 2-5). This is really a rearrangement of the same basic material, in which the peak ground acceleration or Mercalli intensity is used. The numbers 1, 2, 3, 4, and 5 represent different damage states. All these systems are pushing around the rather small amount of real data that there is about the effects of earthquakes on actual buildings.

The above is the essence of how loss estimating is done. The question then comes up as to how accurate is the information that is received from this sort of process. Figure 2-6 shows one of the studies that was done for the Academy report, in which the curves represent different estimates of damage to wood-frame buildings based on various research studies which people have done at different times. The black spots are actual recorded damage, so the range of variation between different consultants' estimates and how those

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

FIGURE 2-3 Expert responses to round one damage factor questionnaire for Facility Class 18—low-rise moment-resisting ductile concrete-frame buildings. Note: Each symbol represents the estimates of one specific person.

FIGURE 2-4 Expert responses to round two damage factor questionnaire for Facility Class 18—low-rise moment-resisting ductile concrete-frame buildings. Note: Each symbol represents the estimates of one specific person.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

FIGURE 2-5 Fragility curves for wood-frame buildings. FROM: Estimating Losses from Future Earthquakes, p. 38, National Research Council, 1989;

Source: Kircher and McCann, 1983.

estimates relate to actual damage is clear. Not nearly enough of this sort of validation exercise is being done. So far, a number of loss estimates have been done, but very little in the way of validation. Validation is rather expensive and does not seem to have the sort of appeal to the research community that other aspects of the earthquake problem have, but it is very critical and very important. As earthquake occurrences, such as Whittier and Loma Prieta, continue more information is developed but methodologies are not being reviewed and validated like they should be.

A frequent subject of interest in loss estimation is deaths and injuries. This is perhaps even more vague than the dollar-loss aspect, but there is a table in the ATC-13 study which, again, was developed in a smoke-filled room by a small number of people (Table 2-2). By applying this table, depending on the damage state, some estimate of injuries and deaths can be calculated. This may not be very accurate, but it is certainly much more useful than just speculating about the number of injuries and deaths.

Some of the Academy committee members pushed to try and get some numerical estimate of accuracy. The engineers were rather reluctant to do this, but some numbers were published that are interesting. One was that estimates for single-family wood-frame houses, where there is a lot of experience, might perhaps be accurate to within a factor of about 1.5. For run-of-the-mill

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

FIGURE 2-6 Intensity-damage relationships for unreinforced masonry buildings. Source: R. Reitherman, 1988.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

TABLE 2-2 Injury and Death Rates in Relation to Damage

Damage State

CDF(S) Percent

Fraction Injured

Fraction Dead

 

 

Minor

Serious

 

1

0.0

0

0

0

2

0.5

3/100,000

1/250,000

1/1,000,000

3

5.0

3/10,000

1/25,000

1/100,000

4

20.0

3/1,000

1/2,000

1/10,000

5

45.0

3/100

1/250

1/1,000

6

80.0

3/10

1/25

1/100

7

100.0

2/5

2/5

1/5

NOTE: Estimates are for all types of construction except light steel construction and woodframe construction. For light steel construction and wood-frame construction, multiply all numerators by 0.1. SOURCE: Applied Technology Council, 1985

commercial-type buildings, the accuracy might be within a factor of 3, and for buildings in areas where there is not a lot of seismic activity, outside California, the accuracy was perhaps an order of magnitude—a factor of 10. In scientific terms this is very inaccurate. In other terms, however, it may be useful. It is useful to know whether there will be hundreds of houses down or thousands of houses down.

Table 2-3 shows the result of a study that was part of the Academy study, which compared the ATC-13 and the Steinbrugge figures. This shows that for wood-frame buildings, for instance, the ATC-13 study has a damage ratio of 8.8. The Steinbrugge study has a ratio of 8 or 12, depending on the kind of building, so that is fairly close. If one looks at tilt-up, ATC-13 shows a damage ratio of 16 percent; the Steinbrugge curve shows a damage ratio of 30. Again, depending on the viewpoint, this is a wild spread, or it is quite useful in terms of what it is used for.

A new development in loss estimation has been its entry into the commercial area. An example of a commercial project, to some extent sponsored by the insurance industry, and really directed at providing information of specific value to the insurance industry, is a project based on research done originally at Stanford University. The Insurance Investment Risk Assessment System (IRAS) Project is a computerized system in which for a given site or region damage to an individual building or an inventory of

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

TABLE 2-3 Comparison of Some Building Damage Ratios (D/R)

U.S. Geological Survey (USGS) and ATC-13 @ MM IX

ATC-13 Name

No.

D/R

USGS Code

D/R

wood frame

1

8.8

1A

12 old

8 new

light metal

2

5.6

2A small

6

 

 

 

2B large

8

VRM

 

 

 

 

low

75

42.0

5E

35

medium

76

52.9

 

 

braced steel frame

 

 

 

 

medium

13

11.3

3A

10

high

14

14.0

 

 

concrete DMRF

 

 

 

 

low

18

8.7

4A

13

medium

19

10.3

 

 

high

20

12.5

 

 

Tilt-up

21

15.8

4D

30

 

SOURCE: H. Degenkolb

buildings can be estimated and the cost/benefit of retrofitting can be provided. Losses can be related to any valid earthquake expressed as a maximum risk or averaged to determine the probable expected losses in a given time frame.

This type of system is a new development which represents a new evolution. Such systems still essentially follow the same methodological basis and use the same information, but output is provided in a much more usable form. At the moment, this particular system applies only to California, in which the hazard is fairly well defined.

What, in fact, is a loss? These studies end with a dollar figure which represents the damage to the building, converted into a dollar figure for replacement. But, there is much more to it than that. The loss does not stop at the damage.

The Hyatt Regency Hotel in Burlingame, near San Francisco Airport, is a new, 400-bed hotel that was completed about a year before the Loma Prieta earthquake, and is a 400-bed hotel. It is a rather nice building architecturally; like most of the Hyatt Regency Hotels, it is built around an

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

atrium, which is a 10-story lobby. It has an interesting translucent fabric roof, which is supported on light steel frames. The restaurants and bars in the center of the atrium, underneath which are the main meeting areas within the hotel, can be seen looking down from the galleries. It is an interesting architectural plan.

The hotel suffered some damage in the Loma Prieta earthquake. It suffered some nonstructural damage, but this was symptomatic of some structural damage within the building. This building may be subject to litigation, so little information is available. The following summary is based on published information, which may or may not be correct. The structural damage consisted of some severe but repairable damage to concrete beams and shear walls in the lower portions of the building.

The hotel was fully occupied, and nobody was killed or injured, but it was decided that the damage was severe enough to justify closing the building, so the building was closed, repaired, and reopened on July 20, 1989, nine months after the earthquake.

The loss has been quoted as $12 million in damage repair costs and $12 million in lost revenues between October and July, while the hotel was dosed. This might be described as a $12 million direct loss and a $12 million indirect loss. The cost of construction was about $54 million, so perhaps the direct loss was on the order of 25 percent (which probably would not have shown up using typical loss-estimation methods).

Because the loss to the Hyatt Company was insured, the loss was to the insurance company. The building was rebuilt, so this represented a gain to the architects and engineers; the figure of $5 million in fees has been quoted, though this may not be true. Two separate engineering firms have been involved in the hotel's rehabilitation work. There have been arguments and controversy, so perhaps $5 million may be correct. The building contractors have gained money, and the local construction labor force in Burlingame has gained money. Attorneys have gained, to date, something on the order of a few hundred thousand dollars. In fact, this situation represents a new redevelopment project, an unexpected redevelopment project for the contracting industry in that area.

Some people lose, some people gain. It could be argued that the insurance companies have not lost. The insurance companies are simply prodding their services. When an architect designs a building for someone, the work should not be regarded as a loss; it is a service that is the architect's job to provide. It is only a loss if an error is made in calculating what it will cost to provide the plans. Thus, if the insurance companies make up Hyatt's loss, they are simply providing a similar service. The Hyatt company obviously had not paid an equivalent sum in premiums in the year in which the hotel existed, but the insurance industry is not based on that sort of arithmetic.

Hyatt's loss of revenue was balanced out by other airport hotels, which gained revenue. It has been quoted that, because of Loma Prieta, Bay Area hotels lost about 10 percent of new business. A lot of tourists did not show up, but business travel continued. No study has yet been done, for instance, that

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

relates the loss of tourist trade to the increase in researchers' trade after the Loma Prieta earthquake; there may be a net gain in trade.

Other airport hotels, downtown hotels, and possibly other cities gained (some people may have gone to Los Angeles rather than San Francisco). Hyatt closed the hotel, but they also reduced their costs. They lost revenue, but they also did not have to support the hotel, so their costs were reduced. On the loss side, Hyatt lost profit, and some employees lost their jobs. Vendors who provided services to the hotel, such as food and laundry suppliers, lost. It is also unknown how much they could make up their losses from the other hotels. In addition, there were some additional insurance losses in terms of paying off the employees.

The city of Burlingame lost $120,000 in business tax, but Burlingame gained in building permits for the new work that was done, so that was probably a standoff. Therefore, the question of loss is not easy to evaluate. There are gainers and losers, and under certain circumstances, what happens is not a loss but a redistribution of resources; there seems to be a kind of conservation-of-energy principle at work.

Clearly, it is only under certain circumstances that this situation would apply. In this case the loss is really a redistribution, because the general business continues; the people continue to come to San Francisco and the San Francisco airport; the hotel is rebuilt, so that there was a reconstruction project for the construction industry; and the hotel reopened. If the general economy of the region or the country breaks down, obviously this situation does not apply. Also, there are other cases in Loma Prieta where clearly the situation did not apply and where there were significant losses. But, if this sort of individual case is multiplied by 100,000 to 200,000 times, a picture of the complexity of the total economic situation is revealed.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

PRESENTATION OF DON G. FRIEDMAN

Three topics will be addressed in this presentation. First, a brief discussion is given of changes in loss-estimation procedures that Travelers Insurance Company has used over the past three decades. There has been significant progress in the development of sophisticated methodologies. Unfortunately, the successful, practical application of these procedures is severely hampered by the lack of appropriate input information, such as geographical inventories of buildings of various types, their damage susceptibilities to earthquakes, and consequent casualty-producing potentialities.

The second topic outlines the need to gain a better general understanding of the major factors that occasionally combine to produce a natural disaster and, subsequently, determine its severity. The need to know more about the disaster-producing mechanism was necessary so that this information could be used as a supplement to, or a replacement for, the often inadequate or inaccurate results obtained from specific applications of the numerical models when appropriate data was not available as input to these computerized procedures.

The last topic is an illustration of the use of natural-disaster knowledge in making risk assessments when sufficient input data is not available for the numerical models. This illustration attempts to answer the question of whether a useful estimate currently can be made of the casualty- and damage-producing potentials of low- and medium-rise buildings (insured and uninsured) due to a catastrophic earthquake in the central or eastern United States. It also describes the various types of information that would be needed if a large-scale effort were made to develop a more credible estimate.

Useful recent information includes results of a 1990 Federal Emergency Management Agency (FEMA)-sponsored study on Estimated Future Earthquake Losses for St. Louis City and County, Missouri;1 a 1985 FEMA study on An Assessment of Damage and Casualties for Six Cities in the Central United States Resulting from Earthquakes in the New Madrid Seismic Zone;2 a United States Geological Survey (USGS) 1983 workshop on The 1886 Charleston, South Carolina, Earthquake and its Implications for Today;3 and a 1990 expert group review of a Metropolitan Boston Earthquake Loss Study.4

To carry out this illustration, there is a need to clarify the meaning of "catastrophic earthquake." Should it be defined in terms of the earthquake's magnitude, its epicenter location, its probability of occurrence, or the casualties and damages that it could produce? In order to examine the use of alternative definitions of a catastrophic earthquake, the fatality and building-damage potentials of a 1990 recurrence of the 1811 New Madrid, 1886 Charleston, and 1755 Cape Ann (near Boston) earthquakes have been estimated, along with a number of lesser-magnitude events with epicenters at the locations of the three major events. The implications of defining a catastrophic earthquake in terms of its physical characteristics such as its magnitude and location rather than the losses that it might produce are examined with the use of a catastrophe index. The use of this index denotes

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

the wide range of uncertainty in loss estimates when sufficient input information is not available.

Uses of Natural-Disaster-Loss Estimations in an Insurance Operation

A large, multiline insurer can have hundreds of millions to billions of dollars of exposure, which is spread haphazardly across hazard-prone areas. Insurance companies make decisions with regard to these risks using whatever data are available. To attempt to answer these questions, a better understanding was needed of the major factors that combine to produce a natural disaster and determine its severity. Numerical modeling and computer simulation techniques have been used to provide this understanding.

In the following discussion, reference will be made to natural disasters caused by hurricanes. Disasters caused by intense hurricanes occur more frequently than high-magnitude earthquakes, but these hurricanes have many similarities with destructive earthquakes, including loss-estimation methodologies.

Loss-estimation procedures for earthquake-caused disasters depend upon the interaction of the geographical pattern of ground motion with the spatial array of the population or properties at risk (elements-at-risk) and their loss vulnerabilities. What happens when an element-at-risk is exposed to ground motion of a given severity is defined as its vulnerability. A hurricane, with its accompanying high-wind pattern, can affect large segments of the population and the built environment. A hurricane-loss-estimation methodology evaluates the interaction of the wind-speed pattern with this geographical distribution of the elements-at-risk and their loss vulnerabilities. Currently, earthquake-hazard evaluations of insured exposures of an insurance company are made using, when possible, ZIP code areas as the geographic designator for locating element-at-risk data. In this way, effects of local influences such as local ground conditions or areas of potential liquefaction can be included in the analysis. However, the elements-at-risk location is not always available within a ZIP code area or even within a county. In these situations, the geographical distribution of the statewide totals must be approximated. How useful is the output of these estimation procedures? It depends on the problem at hand and the accuracy of input information data on the elements-at-risk locations and their other characteristics.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

Numerical Modeling of Earthquake-Caused Natural Disasters

In the past, when an estimate was needed of the damage-producing potential of earthquakes for the total building inventory (insured and uninsured) in California and elsewhere, very little credible information could be found on buildings, by type and loss vulnerability, even on a statewide basis. To attempt to answer damage-potential questions on the overall inventory of buildings in spite of the lack of specific element-at-risk information, these losses have been numerically approximated through ''what-if'' analysis procedures.5,6

Early computer simulations modeled the geographical pattern of ground motion on bedrock and then superimposed the effects of local ground conditions, which were approximated on a 0.1 degree latitude and longitude grid system, using broad definitions taken from a geology map of California. This approach was encouraging because of the similarities of the simulated ground-motion patterns and the actual isoseismal patterns of past California earthquakes.

The type and quality of available input information for defining an earthquake's physical characteristics, the resulting ground-motion patterns, and the effects of local influences have vastly improved over the past quarter century. The USGS now has computerized estimates of local ground conditions on a much finer scale. In addition, research seismologists and engineers have developed a much better understanding of the earthquake mechanisms and the response characteristics of various types of buildings to a range of possible ground-motion frequencies and durations. The development of new physical measures of ground-motion severity may lead to the replacement of the qualitative modified Mercalli intensity (MMI) scale as a primary measure of ground-motion intensity of future earthquakes. However, for loss-potential evaluations of the recurrence of earlier events, the MMI scale is the only measure that is presently available for estimating ground-motion patterns of these past earthquakes.

The geographical pattern of ground motion of earthquakes can be expressed in terms of physical measures that are specific to various types of buildings—for example, pseudo-acceleration to evaluate damage to low-rise buildings. As a result of these improvements, the numerical modeling and simulation of the disaster-producing mechanism of earthquakes is much more sophisticated than it was in the past. However, a major problem is still the lack of appropriate input information on the elements-at-risk to effectively utilize them.

An additional improvement is in the specification of damage vulnerabilities of various types of buildings to given levels of ground-motion severity. These vulnerabilities can be expressed in terms of structural and nonstructural damage potentials: statistical distributions of the degree of damage expectancy, including the percentage of buildings that might collapse. This latter information is useful in determining the casualty-producing

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

potentialities of these structures. The problem is that these vulnerability characteristics, with a few exceptions,1,2 have not been determined for building inventories in the earthquake-prone areas, especially in the central and eastern United States.

Another important development in earthquake-loss estimations is the awareness that a moderate-magnitude earthquake can produce the same severity of ground motion as a great earthquake. The difference is that in a moderate-magnitude event, the area affected by this strong motion is much smaller, and the average duration of significant shaking is shorter. The higher-magnitude event is assumed to produce a larger MMI than the moderate-magnitude earthquake, given the same ground-motion severity at a specific location.

Approximating the Natural-Disaster-Producing Mechanism

The second topic deals with attempts to satisfy some information needs by obtaining a better understanding of how and why natural disasters are produced. If one is willing to accept the numerical modeling and computer simulation concept, various what-if questions can be asked. For example, it can be hypothesized that a particular type of building with its characteristic vulnerability has a uniform geographical distribution of maximum possible density. The ground-motion patterns of each of a series of earthquakes of different magnitudes then can be mathematically superimposed upon this building-at-risk pattern, and the overall damage-producing potential of these earthquakes can be simulated. When the estimate of the total loss-producing potential of each earthquake is plotted versus its Richter magnitude, a nonlinear relationship is obtained. This result suggests that a great-magnitude earthquake has a much greater overall damage-producing potential than a moderate- or minimal-magnitude one. Reasons for the nonlinearity include: the size of the area affected, the severity and duration of the strong ground motion, and the mix of ground-motion frequencies.

Fortunately, this overall damage-producing potential of an earthquake, related to its magnitude, is never fully realized, because the elements-at-risk do not have geographical distributions of maximum possible density over large enough areas to be encompassed by the entire ground-motion pattern of the quake. Therefore, the actual realized loss production of an earthquake depends on how the ground-motion pattern happens to overlap the geographical distribution of elements-at-risk. A moderate earthquake centered on the Newport-Inglewood fault in southern California under a large area of densely clustered elements-at-risk can have a much larger actual damage-producing potential than a high-magnitude earthquake on the Garlock fault along the edge of the Mojave Desert or a great earthquake on the San Andreas fault north of San Francisco, where there are fewer elements-at-risk.

The importance of this interaction of the severity pattern of the event with the spatial array of the elements-at-risk in determining actual damage

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

production was highlighted by last year's Hurricane Hugo. If the storm had taken a westward track across the Georgia coastline, where there are relatively fewer exposures, it probably would have been a $1 billion storm. Instead, Hugo directly hit Charleston, causing about $4 billion in insured losses. If it had moved northward across the North Carolina coastline, following the track of highly damaging Hurricane Hazel of 1954, it would have produced about $7 billion in losses. Finally, if Hugo had come up the East Coast with a landfall on western Long Island, with the same intensity that it had at its Charleston landfall, there could have been losses of $18 billion.

The same-strength storm with different tracks had a wide range of possible damage productions, depending upon its final interaction with the geographical distribution and density of the elements-at-risk.7 Hugo's status as a catastrophic event depended on this interaction. The same type of relationship holds for the earthquake hazard. An individual earthquake can have a much different actual damage-producing potential, depending on its magnitude and the location of its epicenter relative to the spatial array and density of vulnerable elements-at-risk. Consequently, the combination of an earthquake's magnitude and its epicentral location relative to the elements-at-risk is of utmost importance in determining its actual loss-producing potential.

This raises the question: what is a catastrophic earthquake or hurricane? Certainly, the physical magnitude of the event is an important factor, but perhaps of equal importance is how its severity pattern (ground motion of an earthquake or high wind of a hurricane) happens to overlay the usually haphazard spatial array and density of the exposed elements-at-risk that are susceptible to loss.8 A plot of the landfall location of the 247 hurricanes that have crossed the United States coastline since 1870, classified by their physical intensity at landfall as expressed in terms of the five-unit Saffir-Simpson scale, represents a hurricane climatology. If each of these storms recurred in 1990, would all of the Saffir-Simpson code 4 or code 5 storms be considered "catastrophic hurricanes" loss producers? Definitely not! The interaction of severity (high wind) patterns of these storms with coastal clusters of the elements-at-risk determines their "actual" damage production. Many code 3 storms, if they were to recur today, would produce greater losses than the code 4 or code 5 storms because of their particular paths relative to the geographical distribution of the elements-at-risk that are susceptible to damage.

The 1990 loss-producing potentials of all of the 247 landfalling hurricanes, when tabulated against their Saffir-Simpson intensity at landfall, produces a pattern of estimated loss productions of less than $100 million and greater than $1 billion (Table 2-4). There is not a close relationship between hurricane intensity and damage production. For example, 28 percent of the code 3 storms produced simulated losses of less than the $100 million, and 22 percent had a potential of exceeding $1 billion. Because of this analysis, the Saffir-Simpson intensity scale was deemed an inadequate measure of the actual damage-producing potential of hurricanes. Essentially, the scale is an

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

TABLE 2-4 Percentage of Past Hurricanes with a Simulated 1990 Recurrence that Produce Various Loss Potentials when Grouped by Storm Intensity

Saffir-Simpson Intensity at Landfall

Simulated Loss

Code 1 (minimal)

Code 2

Code 3

Code 4

Code 5 (maximal)

< $100 million

82

57

28

5

0

> $1 billion

0

10

22

59

67

indicator of losses to a hypothetically uniform distribution of properties of maximum density.

A "catastrophe index"9 was developed to provide a more realistic representation of the actual damage-producing potential of individual storms or earthquakes than the physical scales that are currently in use. Table 2-5 lists the computed catastrophe index versus the Saffir-Simpson intensity at landfall for the simulated 1990 recurrence of each past storm. A wide range of these damage potential indicies exists within the various hurricane intensity categories. There is a much closer correspondence between the intensity of a landfalling hurricane and its subsequent damage-producing potential, based on a worst case scenario, denoted by a "+" symbol in Table 2-5. Use of the catastrophe index carries the loss-estimation procedure an additional step by taking into account the effects of factors such as the hurricane's landfall location and inland track (or the magnitude and epicenter location of an earthquake) relative to the geographical distribution and damage susceptibilities of the elements-at-risk.

The geographical distribution of the catastrophe index, assuming a 1990 recurrence of each of the 247 landfalling hurricanes, is very different from the distribution of these storms grouped by their Saffir-Simpson intensity. The catastrophe index analysis also can be used to demonstrate how various combinations of an earthquake's magnitude and epicenter location, relative to the geographical distribution of elements-at-risk, can be utilized to better understand the pertinent characteristics of a catastrophic earthquake in the central or eastern United States.

Loss Potentials of a Catastrophic Earthquake in the Central and Eastern United States

The third topic to be covered is an illustration of some of the earthquake-loss-estimation problems that currently exist because of the lack of appropriate input data to the various loss-estimation models. To begin, an

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

TABLE 2-5 Catastrophe Index Resulting from the Simulated Present-Day Recurrence of 247 Past Hurricanes (1871–1990) Listed Versus Each Storm's Saffir-Simpson Intensity at Landfall

Hurricane Intensity at Landfall (Saffir-Simpson Scale)

Catastrophe Index

Damage potential (current dollars)

Code 1

Code 2

Code 3

Code 4

Code 5

Total

1L

1.00 – 2.49 million

7

0

0

0

0

7

1M

2.50 – 4.99 million

8

3

0

0

0

11

1H

5.00 – 9.99 million

13

2

0

0

0

15

2L

10.00 – 24.90 million

24

10

4

0

0

38

2M

25.00 – 49.90 million

17

10

5

0

0

32

2H

50.00 – 99.90 million

8

11

9

1

0

29

3L

100.00 – 249.9 million

9

14

14

2

0

39

3M

250.00 – 499.9 million

7

5

13

2

0

27

3H

500.00 – 999.9 million

1

2

6

4

1

14

4L

1.00 – 2.49 billion

0

6

10

5

2

23

4M

2.50 – 4.99 billion

0

0

4

5

0

9

4H

5.00 – 9.99 billion

0

0

0

3

0

3

5L

10.00 – 24.99 billion

+

0

0

0

0

0

5M

25.00 – 49.99 billion

 

+

0

0

0

0

5H

50.00 – 99.99 billion

 

 

+

0

0

0

6L

100.00 – 249.9 billion

 

 

 

+

0

0

Total

 

94

63

65

22

3

247

The "+" symbol represents the catastrophe index for a worst case scenario in which the insured properties have an unrealistic uniform geological distribution of maximum possible density across the entire area affected by the storm's spatial pattern of high winds.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

analysis has been made of the fatality and building (residential and commercial) damage-producing potential of a simulated 1990 recurrence of the 1811 New Madrid earthquake, the 1886 Charleston earthquake, and the 1755 Cape Ann earthquake near Boston. To account for the importance of the combination of an earthquake's magnitude and location relative to the elements-at-risk in determining its actual damage production (i.e., its catastrophic event status), these potentials also have been estimated for a number of quakes of successively lower magnitudes located at the New Madrid, Charleston, and Cape Ann epicenters. A catastrophe index, which has been determined for each of the scenario earthquakes, is used to define the characteristics of a catastrophic earthquake in the central or eastern United States.

Choice of the scenario earthquakes was made by considering the ten strongest events with epicenters in the central and eastern United States during historic times (Table 2-6). Different scales for representing the strength (magnitude) of the earthquakes are listed in the table. In the following discussions involving earthquake magnitude, the scales will be specified, because they are quite different for higher-magnitude events.

Of the seven largest-magnitude earthquakes with United States epicenters, five were located in the New Madrid seismic zone within a 100-mi strip that runs from northeastern Arkansas to southeastern Missouri. Four of these Richter magnitude 8+ events occurred within several months of one another between late 1811 and early 1812. The fifth earthquake, which occurred in 1895, had a lower-magnitude (Richter 6.7) and was located about 30 mi north of the February 1812 epicenter. The other two events had epicenters near Charleston, South Carolina, and Cape Ann, Massachusetts.

A composite map of the maximum ground motion resulting from the four Richter 8 + earthquakes in 1811 and 1812 is shown as the right-hand map in Figure 2-7. It covers a much larger area than the pattern of any one of the individual events that it represents, because the epicenters of the four quakes were not at the same location but were displaced northward along a 60-mi line. As a result, the ground-motion patterns of each event also were displaced northward, thereby overlapping one another. The composite map shows the largest ground-motion severity of the overlapped patterns in each affected locality. It has been assumed, without evidence, that a 1990 recurrence of the 1811 and 1812 seismic activity would be in the form of a single 8 + event and not the series that originally occurred. As a result of this assumption, it was not feasible to use this composite ground-motion mad of the 1811 and 1812 events, which was prepared by the USGS in 1985.10 Note that the USGS attempted to include the probable effects of local ground conditions as indicated by the distorted shape of the ground motion severity pattern in Figure 2-7.

The only immediately available estimate of the ground-motion pattern associated with any of the four 8+ New Madrid events in 1811 and 1812 is the Richter magnitude 8.6 earthquake that occurred at 2 a.m. on December 16, 1811.11 Because of a lack of observations of the effects of this event to the

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

TABLE 2-6 Occurrence Date, Location, and Magnitude of the Ten Largest Earthquakes that Affected the Central and Eastern United States and Southern Canada in Historic Times

Date

Epicenter

Location

Maximum

MMI

Io

Surface Wave

Magnitude

(Richter Scale)

Ms

Body

Wave

mb

United States Epicenters

Feb. 7, 1812

New Madrid zone

XI–XII

8.7

7.3

Dec. 16, 1811 (2am)

New Madrid zone

XI

8.6

7.2

Jan. 23, 1812

New Madrid zone

X–XI

8.4

7.1

Dec. 16, 1811 (8am)

New Madrid zone

X–XI

8.3

7.0

Aug. 31, 1886

Charleston, SC

X

7.6

6.7

Oct. 31, 1895

Charleston, MO

IX

6.7

6.3

Nov. 18, 1755

Cape Ann (Boston)

VIII

6.0

5.9

Southeastern Canadian Epicenters

Feb. 5, 1663

St. Lawrence River region

IX–X

7.2

6.5

Nov. 18, 1929

Grand Banks, Newfoundland

(IX–X)

7.2

6.5

Feb. 28, 1925

St. Lawrence River region

VIII

7.0

6.4

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

FIGURE 2-7 Composite map of the highest MMI that might be observed at each location if the magnitude of asimulated earthquake held constant at 8.6 and its epicenter were shifted in increments along the New Madrid seismic zone.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

west of the Mississippi River, only a partial ground-motion pattern could be constructed by researchers.

In order to utilize this pattern in the analysis, the ground-motion contours were extrapolated to the west of the Mississippi to provide an approximately symmetrical pattern about the epicenter. The smooth contours indicate that the effects of local ground conditions have not been included. This pattern was superimposed upon a population density map to simulate its overlapping with the current spatial array of the elements-at-risk (population and buildings) as shown in Figure 2-8. Note that the nearest dense cluster of exposures to the strongest ground motion of this particular earthquake is in the Memphis metropolitan area. A Richter 8 + earthquake with an epicenter farther north in the New Madrid seismic zone would generate very strong ground motion closer to the large duster of exposures in the St. Louis metropolitan area.

Because of the importance of the combination of an earthquake's location and magnitude in determining its loss-producing potential, it was necessary also to estimate the ground motion patterns associated with lesser-magnitude events that have the same epicenter location as the December 16 (2 a.m.) earthquake. As demonstrated in Table 2-6, there is an empirical relationship among the measures of magnitude. The USGS10 used this relationship to approximate the composite severity patterns of earthquakes of lower magnitude (Richter 7.6 and 6.7) that have the same epicenters as the Richter 8+ event (Figure 2-7). In the conversion procedure, the USGS assumed that the shape of the ground-motion pattern would not change with a reduction in the earthquake's magnitude and that the modified Mercalli intensity could be reduced in the overall pattern by one unit in order to obtain an approximation for a Richter 7.6 event, and reduced by two units overall to denote a Richter 6.7 event.

Similar combinations of earthquake magnitudes and locations; were needed to represent earthquake-prone sections of the eastern United States-Charleston, South Carolina, and Boston, Massachusetts. Using the same procedure, the 1886 Charleston earthquake and the 1755 Cape Ann earthquake listed in Table 2-6 were modeled to estimate the loss-producing damage potentials if such events were to recur in 1990. Figures2-9 and 2-10 show the superposition of these ground-motion patterns on the elements-at-risk spatial array.

Some earthquake experts suggest that a Cape Ann earthquake of magnitude 6.0 is not necessarily the largest possible earthquake in this seismic zone.3 To estimate the possible loss-producing effects of a higher-magnitude earthquake with a Cape Ann epicenter, the USGS10 procedure was reversed to estimate the ground-motion pattern of a stronger, Richter magnitude 6.7, quake at the Cape Ann location. One unit of modified Mercalli intensity was added to each of the ground-motion categories defined by Street and LaCroix.12 The enlarged pattern of the effects of a Richter 6.7 event, superimposed on the 1990 elements-at-risk density map (Figure 2-11) indicates that, although the area affected by the VII-or-above intensities is small

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

FIGURE 2-8 Loss-producing potential of a recurrence of the December 16, 1811, New Madrid earthquake.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

FIGURE 2-9 Loss-producing potential of a recurrence of the 1886 Charleston, South Carolina earthquake.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

FIGURE 2-10 Loss-producing potential of a recurrence of the 1755 Cape Ann (Boston), Massachusetts earthquake.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

FIGURE 2-11 Loss-producing potential of a recurrence of a stronger (magnitude 6.7) Cape Ann (Boston), Massachusetts earthquake.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

compared with that of the New Madrid and Charleston quakes, the strong ground-motion area overlaps one of the largest dusters of densely packed elements-at-risk in the United States.

One of the important information needs for making estimates of losses caused by earthquakes in the central and eastern United States is a mapping of local ground conditions and possible liquefaction areas, as has been done by the USGS in California. Another need, of equal importance, is for the development of a more realistic, physical, measure of ground motion that can be used as a replacement for the qualitative (and in many instances unsatisfactory) modified Mercalli intensity scale. Ideally, this new measure could be translated into an estimate of the ground motion of past earthquakes. However, for the purposes of this illustration, the only ground-motion measures that are available for three of the strongest past earthquakes are the modified Mercalli intensity patterns.

Estimation of Earthquake-Caused Fatalities

Because most earthquake-caused deaths and injuries result from damaged buildings, the casualty-estimation procedure should be based in some manner upon building damage, especially with respect to the percentage of structures that might have serious structural and nonstructural problems. Ideally, an estimation procedure for determining the casualty and damage potentials of these scenario earthquakes should have detailed information on the number and spatial distribution of each type of building in affected areas, along with their characteristics relating to damage and casualty-producing potentialities. Important considerations would be such items as the type and quality of construction, age, condition of upkeep, local ground conditions, building code in effect at time of construction, contents, usage, and number of occupants at various times of the day. Unfortunately, a spatial inventory of buildings and their characteristics is not available in these earthquake-prone areas of the central and eastern United States. However, one of the purposes of this illustration is to attempt to obtain order-of-magnitude estimates of the casualty and damage potentials from scenario earthquakes based on information that is available. The approach taken is described below.

The number of persons within each of the ground-motion categories for four scenario earthquakes (Figures 2-8 through 2-11) plus four others of lesser magnitude were estimated by overlaying in turn the ground-motion patterns on a map of counties in the central and eastern United States. The number of persons within each ground-motion-severity pattern in each of the affected states was summed for each earthquake.

Table 2-7 lists the cumulative number of persons in various ground-motion categories from MMI V-or-more to IX-or-more for each of eight earthquakes. The largest number of persons that would be affected is estimated to exceed 120 million in areas with ground motions of MMI V-or-more during a Richter 8.6 New Madrid earthquake. A Richter 6.0 Cape

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

TABLE 2-7 Estimate of the Number of Persons Who Would be Exposed to Various Levels of Ground-Motion Severity Caused by Each of the Scenario Earthquakes

 

Magnitude

 

Number of Persons (thousands)

 

Richter

Body Wave

Ground-Motion Severity (Modified Mercalli Intensity Scale)

 

Ms  

mb  

V or More

VI or More

VII or More

VIII or More

IX or More

Earthquake

New Madrid

6.0

5.9

4,380

2,140

420

130

0

New Madrid

6.7

6.25

21,920

4,380

2,140

420

130

New Madrid

7.6

6.7

50,440

21,920

4,380

2,140

420

New Madrid

8.6

7.2

120,790

50,440

21,920

4,380

2,140

Charleston

6.7

6.25

23,330

11,060

2,830

620

290

Charleston

7.6

6.7

43,530

23,330

11,060

2,830

620

Boston (Cape Ann)

6.0

5.9

17,730

7,190

2,690

0

0

Boston (Cape Ann)

6.7

6.25

45,990

17,730

7,190

2,690

0

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

Ann earthquake could affect over 2.5 million persons within areas where strong ground motions equalled or exceeded MMI VII. The size of this exposure is much larger than the 400,000 persons that are estimated to be affected by a New Madrid earthquake of Richter magnitude 6.0. If the Cape Ann earthquake had a magnitude of Richter 6.7, over 7 million people would be subjected to ground motion of MMI VII-or-more compared with about 3 million in a Charleston (Richter 6.7) earthquake and about 2 million in a New Madrid (Richter 6.7) event.

Estimation of casualties resulting from each of the ground-motion categories was made by use of fatality rates versus modified Mercalli intensity relationships, which were applied to the number of persons exposed in each of the ground-motion-severity categories within each state for each of the eight scenario earthquakes. The rates were expressed in terms of the number of deaths per 100,000 exposures. Relationships between fatality rate and ground-motion severity were developed from three fatality scenarios:

Fatality Scenario 1: This relationship was developed using an estimate of the number of deaths that might be expected, by state, if the 1886 Charleston earthquake recurred, based on information given in a USGS workshop report.3 For this illustration, these fatality estimates were related to the assumed ground-motion-severity pattern of the earthquake from which a best-fitting, nonlinear curve was drawn.

Fatality Scenario 2: The scenario i relationship based on the South Carolina information was calibrated by use of results of a Boston study, which estimated the number of fatalities that might be expected due to building damage in Boston and some of its suburbs resulting from a present-day recurrence of the 1755 Cape Ann earthquake during working hours on a weekday.4

Fatality Scenario 3: This fatality-versus-ground-motion relationship was based on fatality rates that were estimated in FEMA's six-city and St. Louis studies. Estimates of the number of deaths that could occur as a result of building damage caused by a repeat of a New Madrid (Richter 8+) earthquake were made in the 1990 FEMA study of St. Louis city and St. Louis county1 and in six other Midwest cities in a 1985 FEMA study.2 Implied fatality rates were determined using the fatality estimates and the estimated numbers of persons in these towns and cities at the time of the simulated earthquake as reported in the FEMA studies.

These rates were then related to the ground-motion severity that was hypothesized by FEMA for each of the locations. Because a range of MMIs was mapped across the study areas, a single, ''weighted'' MMI was determined for each city by overlaying a grid system on the maps of the seven Midwest cities and St. Louis county. The average MMI for each location was obtained by assuming that the MMI scale is continuous and by weighing various MMI values by the percentages of the total town or city area that they represented. Because of a lack of information on the spatial distribution of buildings within these localities, it was necessary to assume that they were uniformly distributed.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

Ideally, fatality rates should be based on expected damage to various types of buildings in areas affected by each of the eight scenario earthquakes. Since this information is not available, it was necessary to assume that the death rates estimated by FEMA in the eight midwestern locations could be related to the attendant MMI and then used as a universal relationship between fatality rate and ground-motion severity for other values of MMI in the central and eastern United States that would be affected by each of the scenario earthquakes. In developing these vulnerability relationships, it was assumed that the fatality rate would be very small when the ground motion was MMI V (1 death per million exposures).

Table 2-8 summarizes state-by-state estimates of fatalities resulting from each of the eight scenario earthquakes using the three vulnerability scenarios. An inspection of this table indicates that there is a multiple of 4 or 5 in the estimated number of fatalities between the lowest values (using scenario 1) and the highest values (using the scenario 3 relationship). A 1990 repeat of the Richter 8.6 New Madrid earthquake could cause somewhere between 7,000 and 27,000 fatalities, depending on the scenario used and assuming that the sets of underlying assumptions are realistic.

Even though the available information cannot provide fatality estimates with a high degree of accuracy, the implied interactions between the earthquake's magnitude, its location relative to the spatial array of the elements-at-risk, and the fatality vulnerability relationships emphasize the importance of considering these particular factors when attempting to define the fatality-producing characteristics of a catastrophic earthquake in the central and eastern United States.

Estimation of Earthquake-caused Building Damage

Estimation of building damage resulting from each of the eight scenario earthquakes also was based solely on the use of immediately available data. Ideally, to estimate building damage due to ground motion, an analysis similar to that carried out in FEMA's six-city and St. Louis studies1,2 should be done for each city or town in the affected areas. At present, there is a discouraging lack of useful information on various types of buildings, their numbers, spatial distribution, and vulnerability characteristics in the central and eastern United States.

The only immediately available information on the spatial distribution of buildings was obtained by estimating the total value of residential and commercial buildings by county, using data given in a recent report prepared by the Insurance Research Council,13 which listed the total value of residential and commercial buildings insured against the wind peril in each of the coastal counties along the Gulf and East coasts. This represents a large percentage of the total building inventory. For the purposes of this study, these numbers were related to the population in the counties, permitting the development of a relationship between size of county population and the total value of insured

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

TABLE 2-8 Estimate of the Number of Fatalities Caused by Each of the Hypothetical Earthquakes and the Three Fatality-Vulnerability Scenarios

 

Magnitude

 

Number of Fatalities

 

Richter

 

Body Wave

Fatality-Vulnerability Scenario

Earthquake

Ms

mb

1

2

3

New Madrid

6.0

5.9

90

340

480

New Madrid

6.7

6.25

430

1,460

2,200

New Madrid

7.6

6.7

1,990

5,220

7,880

New Madrid

8.6

7.2

6,890

18,110

26,930

Charleston

6.7

6.25

690

2,430

3,620

Charleston

7.6

6.7

3,360

8,490

13,000

Boston (Cape Ann)

6.0

5.9

250

1,260

1,570

Boston (Cape Ann)

6.7

6.25

1,550

5,040

8,170

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

residential and commercial buildings in the county. The relationship changes with an increase in county population. The spatial extent of the county was not taken into account in the analysis. These relationships for coastal counties were assumed to be universally applicable and were used to convert the population of counties affected by ground-motion patterns of the scenario earthquakes into rough estimates of their residential and commercial building exposures. It was assumed that if county population rather than ZIP code population were used, the mix of the number and type of commercial buildings relative to residential structures as a function of town or city size (urban versus suburban and rural conditions) would be minimized.

Based on this analysis, about 50 percent of the total United States insured residential building values of $6.3 trillion would be affected by ground motions of MMI V-or-more intensity if there were a recurrence of the Richter 8.6 New Madrid earthquake. Strong ground motion (VII-or-greater) would affect about $213 billion of residential buildings caused by a hypothetical Richter 6.7 Cape Ann earthquake as compared with an exposure of $69 billion for a Richter 6.7 Charleston earthquake or $54 billion with a Richter 6.7 event located at the New Madrid epicenter zone. About $6 trillion of the $13 trillion total insured commercial building values in the United States would be affected by ground motions of MMI V-or-more during a repeat of the 1811 (Richter 8.6) earthquake. The overall accuracy of these estimates (based on a conversion from county population to a measure of insured building values) is not known.

Translation of this building-exposure information into a measure of the damage-producing potential of various ground-motion severities was accomplished by constructing and applying three damage-vulnerability relationships similar in form to the ones used for estimating fatalities. These relationships were based on three damage scenarios.

Damage Scenario 1: To obtain an estimate of the minimum damage-producing potential, it was assumed that all of the residential buildings in the central or eastern United States that are affected by the scenario earthquakes would be of frame construction, which has one of the lowest damage susceptibilities to earthquake-caused losses. A relationship given in FEMA's 1990 St. Louis report1 between frame-building damage expectancy and modified Mercalli ground-motion intensity was used.

Damage Scenario 2: The relationship between ground-motion severity and residential building damage was based on the best-fitting curve through a plot of the implied percentages of value lost for residential-type buildings (insured and uninsured) in the six cities analyzed in the 1985 FEMA study2 and St. Louis city and St. Louis county in the 1990 FEMA report.1 The estimated total value of residential buildings in each of these six cities was obtained using the population-versus-insured-residential-building values that were obtained from the coastal county information. These estimates of insured building values were used as an index for approximating the value of all residential buildings (insured plus uninsured as defined in the FEMA studies).

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

The spatially weighted MMIs for each of the eight areas were used in plotting these implied percentages-of-value-lost data.

Damage Scenario 3: The relationship between residential building damage and modified Mercalli intensity was based solely on information given in FEMA's St. Louis report on expected damage versus ground-motion severity (MMI) weighted by the mixture of residential building types in St. Louis county.

By applying each of these three damage-vulnerability scenarios in the simulated recurrence of the Richter 8.6 New Madrid earthquake, a range of estimated residential damages between $38 billion and $65 billion was found. This range is consistent with an upper bound estimate of about $50 billion for residential building damage made by Algermissen in a 1990 USGS paper.14 A repeat of the Richter 7.6 Charleston earthquake would cause between $19 billion and $32 billion in residential building (insured and uninsured); between $5 billion and $9 billion of residential damage would occur due to a recurrence of the Richter 6.0 Cape Ann earthquake near Boston.

This same procedure was used to estimate the ground-motion-caused damage potential to commercial buildings. Loss estimates were based on a set of three damage-vulnerability scenarios similar in form to those for residential-type structures. A 1990 recurrence of the Richter 8.6 New Madrid earthquake would cause between $37 billion and $105 billion in commercial building damage. A repeat of the Charleston (Richter 7.6) quake would produce commercial building damages somewhere between $18 billion and $52 billion, and a 1990 recurrence of the Richter 6.0 Cape Ann event would cause between $5 billion and $15 billion in damage.

Estimation of Building Damages by Fire Following an Earthquake

Significant building damages also can be caused by fire that follows some high-magnitude earthquakes. No quantitative estimates or estimating procedures were found in the literature regarding the damage potential of this peril in the central or eastern United States. Therefore, to provide at least an order-of-magnitude estimate for this possibility, an approach used by the Insurance Research Council (formerly the All-Industry Research Advisory Council) in California15 was adapted for conditions east of the Rockies. It was assumed that the major contribution to fire-caused damage would be from individual buildings or small groups of adjacent structures.

Again, three vulnerability scenarios were constructed based on relationships that were derived from information in the AIRAC study of Los Angeles and San Francisco fire-following-earthquake susceptibilities. In that study, the fire-fighting capabilities in various communities were taken into account (e.g., the number of firetrucks available and the dependability of the water supply under earthquake conditions). Because of the lack of this type of information for the central and eastern United States, it was necessary to assume that a

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

first approximation to the fire-following-earthquake damage production could be obtained by an averaging across the range of individual community fire-fighting capabilities so that the damage-producing potential of this hazard could be related directly to ground-motion severity. It is assumed that, as the duration and severity of the ground motion increases, the number of fire ignitions caused by the earthquake increases, and the capability of the fighters to limit damage to individual structures or the spread of the fire to adjacent buildings decreases, given the need for high-priority search-and-rescue activities for firemen, equipment and communication failures, broken waterlines, and debris and congestion in the streets.

The three fire scenarios were applied to the estimates of total (residential and commercial) building values, by ground-motion-severity categories, for each of eight scenario earthquakes. For purposes of this analysis, it was assumed that residential and commercial building damage caused by ground motion was independent of the fire losses and that the damage threshold is at MMI V, where the average fire-loss potential on nonearthquake days would occur. A repeat of the Richter 8.6 New Madrid earthquake in 1990 would cause fire damages ranging between $7 billion and $25 billion. A recurrence of the Richter 7.6 Charleston event would produce fire losses between $3 billion and $13 billion, and the Richter 6.0 Cape Ann quake could cause fire losses between $300 million and $2.5 billion.

Loss Expectancies Based on Various Combinations of Earthquake Magnitude and Epicenter Locations

Table 2-9 is a state-by-state tabulation of earthquake-caused building damages using the middle (scenario 2) vulnerability relationship for ground motion and fire damage to residential and commercial buildings resulting from a repeat of the Richter 8.6 New Madrid earthquake. Because of the likely low degree of accuracy of these estimates, the relative ranking of the states by damage expectancy is probably more realistic than the absolute values of the loss estimates.

Table 2-10 shows the probability of earthquake occurrence for two magnitudes of earthquakes in the New Madrid zone, the southern United States, and the New England region during the 1990s. Tables 2-11a and 2-11b list fatality and building-damage potentials implied by various combinations of earthquake magnitude and location in the central and eastern United States, using the scenario 2 vulnerability relationships. These potentials increase at different rates as the simulated earthquake's magnitude is increased to its maximum likely value at the New Madrid, Charleston, and Cape Ann epicenters. The rate differences are caused by interactions of the geographical pattern of ground-motion severity and duration with the particular spatial array, density, and vulnerability of the elements-at-risk near each of these three seismic sources.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

TABLE 2-9 Estimated Building-Damage Losses by State Resulting from a 1990 Recurrence of the December 16, 1811, New Madrid Earthquake with a Richter Magnitude 8.6, Based on Damage Vulnerability Scenario 2

 

 

Damage (millions of dollars)

 

Fire

Ground-Motion Damage

Total

State

Damage

Residential

Commercial

Damage

Alabama

291

1,152

956

2,399

Arkansas

2,646

8,978

23,726

35,350

Georgia

29

114

112

255

Illinois

853

3,427

3,875

8,155

Indiana

688

2,696

2,318

5,702

Iowa

1

4

4

9

Kansas

1

5

4

10

Kentucky

1,222

4,795

6,082

12,099

Louisiana

206

804

681

1,691

Maryland

0

1

1

2

Michigan

1

3

3

7

Minnesota

0

0

0

0

Mississippi

650

2,683

2,551

5,884

Missouri

1,581

5,854

9,806

17,241

Nebraska

0

0

0

0

New York

0

1

1

2

North Carolina

17

67

67

151

Ohio

462

1,744

1,593

3,799

Oklahoma

38

154

125

317

Pennsylvania

0

2

2

4

South Carolina

9

35

34

78

Tennessee

3,619

13,185

28,055

44,859

Texas

35

138

116

289

Virginia

4

17

16

37

West Virginia

8

30

28

66

Wisconsin

0

1

1

2

Total

12,361

45,890

80,157

138,408

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

TABLE 2-10 Probability of Earthquake Occurrence in the Decade Before the Year 2001 (in Percentages)

 

Earthquake magnitude

Region

Richter 6.25 Body wave 6.00

Richter 8.25 Body wave 7.00

New Madrid seismic zone

13

2

Southeastern United States

11

2

New England

8

1

The catastrophe index (Table 2-5), which incorporates the effects of these interactions, can be utilized to more clearly denote important characteristics of a catastrophic earthquake in the central or eastern United States. The lowest and highest damage estimates (scenarios 1 and 3) for various combinations of earthquake magnitude and location have been converted to the catastrophe index scale and plotted in Figure 2-12. Hatched areas define the range of estimates for the three epicenter locations, caused by the choice of vulnerability scenario.

If the catastrophe criterion is defined in terms of magnitude of the geophysical occurrence, such as a Richter 8+ event, then the New Madrid seismic zone would be the primary potential producer of catastrophic earthquakes because of its unique capability for generating these great earthquakes. Earthquakes of this magnitude are not likely to occur in either the Charleston or Cape Ann source region, based on current information.

On the other hand, if the catastrophe criterion is based on the relative size of the earthquake's damage-producing potential, then a moderately severe Cape Ann event could qualify as a catastrophic earthquake. Even though its magnitude might be less than that of an earthquake centered either in the Charleston or New Madrid areas, damage production could be greater because of its proximity to large clusters of vulnerable elements-at-risk. Results of the analysis suggest that a moderately severe or high-magnitude earthquake centered at any of these three epicenter locations could cause thousands of fatalities and billions of dollars in building damage.

The purpose of this estimation exercise was to determine the implied casualty- and damage-producing potentials of earthquakes in selected sections of the central and eastern United States as implied from immediately available information. This analysis produced unacceptably wide ranges of estimates because of the lack of pertinent data. It stresses the need for development of the many types of data that are required to efficiently apply the available sophisticated numerical loss-estimation models. The information needs include: the physical characteristics and ground-motion patterns of earthquakes with various combinations of magnitude and epicenter location in the

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

FIGURE 2-12 Estimated damage to buildings caused by ground motion and fire following an earthquake, versus earthquake magnitude. Damage, expressed in terms of the catastrophe index (Table 2-5), is based on vulnerability scenarios 1 and 3.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

TABLE 2-11a Estimated 1990 Fatality and Building Damage Potentials in the Central and Eastern United States Resulting from Simulated Earthquakes of Various Magnitudes Centered at the Location of the 1811 New Madrid, 1886 Charleston, and 1755 Cape Ann Events, Based on the Scenario 2 Vulnerability Relationship: Number of Fatalities

Earthquake Magnitude

Epicenter Location

Surface Wave (Richter) Ms

Body Wave

mb

New Madrid

Charleston

Cape Ann

4.00

4.90

0

0

0

4.50

5.15

0

0

1

5.00

5.40

4

6

13

5.50

5.65

48

70

230

6.00

5.90

340

410

1,300

6.50

6.15

1,000

1,700

3,700

7.00

6.40

2,350

3,800

——

7.50

6.65

4,700

7,600

——

8.00

6.90

9,000

——

——

8.50

7.15

16,500

——

——

central and eastern United States; the effects of local ground conditions on ground-motion severity and its duration; inventories of various types of buildings, their number, and spatial distribution; and their damage susceptibilities to various combinations of ground-wave frequency, severity, and duration.

A final consideration in defining the characteristics of a catastrophic earthquake is probability of occurrence. Nishenko and Ballinger16 have recently made estimates of the probability of occurrence of major earthquakes in three regions (Table 2-10). Given that an earthquake occurred in one of these broad areas, a conditional probability would have to be applied to determine its chances of being located in one of the three specific epicenter areas discussed above. Nevertheless, based on these estimates, any one of the three source areas is capable of producing an event which could be classed as a catastrophic earthquake in the current decade if loss-producing potential is the criterion.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

TABLE 2-11b Estimated 1990 Fatality and Building Damage Potentials in the Central and Eastern United States Resulting from Simulated Earthquakes of Various Magnitudes Centered at the Location of the 1811 New Madrid, 1886 Charleston, and 1755 Cape Ann Events, Based on the Scenario 2 Vulnerability Relationship: Building Damage (Millions of Dollars)

Earthquake Magnitude

 

 

Epicenter Location

Surface Wave (Richter) Ms

Body Wave

mb

New Madrid

Charleston

Cape Ann

4.00

4.90

0

0

0

4.50

5.15

5

5

8

5.00

5.40

65

100

230

5.50

5.65

500

950

3,100

6.00

5.90

2,300

4,300

14,000

6.50

6.15

6,500

12,500

38,000

7.00

6.40

15,000

30,000

——

7.50

6.65

34,000

61,000

——

8.00

6.90

65,000

——

——

8.50

7.15

120,000

——

——

The focus in this presentation has been on fatalities and damages to buildings caused by a high-magnitude earthquake. There are many other sources of loss potential that also have to be considered, such as damage to utilities, roads, and bridges; medical cost for the injured; the cost of debris removal; damages to automobiles and other personal property; business-interruption costs; and liability-loss potentials. Many of these loss potentials are of significant size.17

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

PRESENTATION OF KATHLEEN TIERNEY

Loss-estimation methodologies seek to determine how one or more earthquakes in a given geographic area will affect people, property, and social and economic activity. Approaches vary considerably. Theoretically, the unit of analysis for loss projections can range from the macro- to the micro-level, but most studies done to date focus on communities or slightly larger units. Ground shaking is the most frequently used hazard in these studies, but other primary effects, such as fault rupture and liquefaction, as well as secondary effects, can also be taken into account. Loss-estimation studies can vary in the number of effects they take into consideration; they may, for example, deal only with physical earthquake effects to specific types of structures, or they may be quite broad, taking into account a range of impacts and both direct and indirect earthquake effects. Different strategies for developing inventories of what is at risk and for modeling losses constitute another source of variation in the loss-estimation literature. Nevertheless, although a broad range of potential approaches exists, the loss-estimation studies that have been conducted to date have been relatively limited in the impacts and variables they have considered.

The National Research Council's (NRC) Committee on Earthquake Engineering report, entitled Estimating Losses from Future Earthquakes18 contains a concise review of loss-estimation methods and approaches in engineering and related disciplines that focuses particularly on those methods that are used by local and state governments for hazard mitigation and emergency planning. That report briefly discusses the basic components of loss-estimation studies (seismic-hazard analysis and vulnerability analysis), outlines the elements of deterministic and probabilistic approaches to seismic-hazard analysis, and identifies the types of direct losses and indirect impacts such studies have attempted to quantify. Major categories of losses discussed include building damage, fatalities and injuries, homelessness, and loss to special facilities and lifeline systems. Secondary losses discussed in the report include fire, hazardous materials releases, and indirect economic impacts.

The NRC report and the working papers on which it is based constitute a good overview of available methods and approaches, and they also contain several important discussions on the usefulness and limitations of earthquakeloss estimates. Using these ideas and other insights from the literature as a basis, the following are some general observations about the utility of the work that has been done to date for projecting structural and nonstructural losses, indirect earthquake impacts, and long-term socioeconomic effects, as well as some observations about their relationship to policy. The messages that will probably come across are that existing findings from loss-estimation methodologies have certain inherent limitations as policy tools and that there are currently significant gaps in our understanding of the range of probable earthquake impacts, particularly impacts not related to building damage.

The first topic to consider is how much we know about various types of earthquake losses. There appears to be general agreement among practitioners

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

that loss-assessment methodologies are well developed and validated only for some categories of losses. A taxonomy developed by Petak and Atkisson19 classifies natural-hazards effects into three categories: (1) primary, such as injuries and damage to buildings; (2) secondary, which follow almost immediately as a consequence of those basic impacts, such as homelessness and disruption of utility services; and (3) ''higher-order,'' which manifest themselves later, such as long-term unemployment and changes in the tax burden associated with disaster recovery. There is considerably more known about impacts in the first category than about those in the other categories. In some subareas, there is virtually nothing known.

There are at least two reasons why this has been the case. The first, of course, is that a considerable amount of loss-estimation research has been done mainly to address the concerns of particular clienteles. The pioneering work in the field20 was undertaken for the property insurance industry and the government departments responsible for insurance regulation. The field developed in such a way that estimating probable direct damage to property-particularly buildings—thus became a major focus. Ironically, we know considerably less today about the estimation of earthquake-related deaths and injuries than we do about potential building damage. Fortunately, a number of very capable scientists have recently begun efforts to close this gap. When it comes to other social impacts, such as earthquake-caused homelessness, almost no systematic work has been done.21,22

Second, while presenting enormous challenges by anyone's standards, estimating direct earthquake effects is in many respects easier than attempting to take longer-term, higher-order impacts into account. When that line of research was developed, building inventories, construction classifications, and other necessary elements for loss assessment already existed, as did a useful historical record on how earthquakes affect buildings. From the standpoint of available data, the picture becomes much murkier as broader impacts are considered. For example, while the task of developing building inventories is difficult, obtaining up-to-date information on building occupancies and uses-which is important for casualty estimation and projections of losses to building contents—is even more difficult. To do this type of work on anything but the most modest scale would require considerably more funding than appears to be currently available.

For some categories of losses, systematic empirical data on events comparable to those of interest do not exist, so there is little in the way of an empirical basis from which to extrapolate. In other cases (e.g., damage to some industrial and to nuclear power plants and defense-related facilities), the data may exist, but access is restricted because of organizational concerns, and the information does not enter the public domain.

Projects such as the major, pioneering loss-estimation work undertaken by the Applied Technology Council,23 attempt to compensate for the lack of a data base by developing estimates based on the judgments of expert panels, with impressive results. Ron Eguchi, a colleague of Professor Tierney's, developed some ingenious ways of getting around data limitations when

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

attempting to model the likelihood and risk to local residents of earthquake-induced hazardous materials releases.24,25 We are currently attempting to do something roughly comparable for potential earthquake-generated oil pipeline failures in the New Madrid fault zone, with socioeconomic impacts being the main variables of concern. However, efforts like these represent only preliminary steps in assessing secondary losses, and any method is only as good as the data and the assumptions on which it is based. It is also useful to keep in mind that the catastrophic event with which we are concerned here is unprecedented in the United States. A catastrophic earthquake is likely to be qualitatively different from less serious events; the effects generated may be orders-of-magnitude different from those that have served as a basis for experts' projections.

Eminent sociologists who study risk, such as Charles Perrow,26 have done an excellent job of outlining the pitfalls involved in estimating the effects of accidents and failures—particularly those analyses that focus on infrequent and catastrophic events. Work by people such as Allan Mazur27 has highlighted problems inherent in trying to base public policy on expert analyses. However, even if disbelief is suspended and a great deal of faith is placed in the methods that have been developed for estimating losses, the problem remains that, for some losses and impacts, there are often simply not enough data from which to extrapolate.

Estimating long-term, regional, or systemwide economic impacts is also complicated by the fact that social systems are so complex. Steinbrugge, in discussing the problem of compiling loss statistics, notes that dollar loss estimates can vary widely, depending on whether losses are considered as "personal" or "impersonal."20 In other words, the notion of who is likely to bear the costs of damage is interwoven with the cost figures themselves. This is apparent even in the most straightforward cases of physical damage to buildings, and it is even more significant when higher-order effects are considered. In recent U.S. history, the Chrysler Corporation and the savings and loan bailouts show how flexible social systems can be with respect to distributing losses. New public and private initiatives in the earthquake insurance area will likely result in changes in loss projections (and, when the earthquake occurs, changes in how losses are distributed) for various societal sectors. Earthquake investigators may have a reasonably good idea about how a particular earthquake, with a particular intensity, will affect a particular kind of structure and, generally speaking, how much that could cost. Socioeconomic effects are inherently more difficult to model.

Discussions on how to estimate certain sets of economic effects18 using input-output (I–O) modeling seem to proceed on the implicit assumption that social systems are "closed systems." Even dynamic I–O models are relatively insensitive to changes in the larger environment in which economic subsystems are embedded. Treating societies, regions, and communities as "open systems" could lead to very different loss estimates. In short, without having a clearer idea of what policy options and programs might come about to contain the

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

economic costs of future earthquakes, those attempting to make analyses are at a distinct disadvantage in trying to calculate those costs.

The second topic that needs a closer look is the relationship between loss estimates and public policy. Loss-assessment methodologies appear to be based on the assumption that, since losses can be expressed in terms of dollar figures, persons killed or injured, or other standard scales, the quantitative results of analyses will be immediately relevant for planning and policy purposes. The assumption is that mitigation and planning priorities can be set rationally, on the basis of the numbers the estimates contain. However, the results of quantitative loss estimates will invariably be judged in light of qualitative and value judgments. To use an extreme example, a statistic indicating that 50 persons were killed (or will be killed) due to earthquake damage will be judged very differently by society, depending on whether these are 50 isolated individuals who died from a range of causes, 50 prisoners who died because of the collapse of a correctional facility, or 50 children crushed in a collapsed private school building. To give another, slightly different illustration, the loss of a certain number of housing units of low value—say 9,000 inexpensive units—may constitute a relatively small economic loss for a community in sheer dollar terms, but a very large loss when the community's low-cost housing needs are taken into account. Dollar estimates and raw numbers do not convey a sense of the social meaning of losses.

The literature clearly indicates that risks and losses are not judged in straightforward cost-benefit terms, but rather are assessed by members of society according to a range of criteria. For example, potential losses borne involuntarily are not perceived in the same way as those that are voluntarily assumed. Unfamiliar, uncontrollable, and catastrophic effects are seen as particularly undesirable.28 If one "loses" $50, that loss will be evaluated in context. It is an entirely different matter whether the money fell out of a pocket, whether it was stolen from a wallet, or whether the $50 was lost because the money was bet on a losing horse at the racetrack.

Damage estimates for geographic areas and categories of assets also tend to mask the fact that losses are typically not distributed evenly in society. Harold Cochrane was among the first to call our attention to what he terms the "distributive" effects of natural disasters. In his paper, Cochrane notes that in disasters, lower-income groups consistently bear a disproportionate share of the losses; they receive, in most instances, the smallest proportion of disaster relief; they are the least likely to be insured (for either health, life, or property); and they live in dwellings which are of the poorest construction and the most subject to damage.29

Aggregate statistics on the number of dwellings that will be lost, of persons who will be killed and injured, or of jobs that will be destroyed do not address these disproportionate impacts, which are of great importance when the earthquake problem is viewed from the standpoint of policy. The point here is that similar dollar figures and casualty figures deriving from loss-estimation studies may not in fact be equivalent, in the "social" sense. As an

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

attempt is made to estimate future earthquake losses, and to develop and implement policies to reduce those losses, this is a point that merits emphasis.

A third related issue involves the audiences to which loss estimates are directed and the utility of such reports. Scientists have expended considerable effort in developing and refining loss-estimation methods, but, unfortunately, the outcomes of loss-estimation studies may not be relevant to some potential users, for a variety of reasons. One problem with the existing research is the limited nature of most studies. For example, in attempting to be comprehensive, large-scale loss-estimation studies focus on entire regions. In so doing, they sacrifice applicability to smaller units of analysis and the ability to generate specific estimates. Estimates derived for categories of structures apply only in the aggregate, not to smaller units of analysis. Additionally, budget and data limitations typically constrain the work, so that studies usually focus on shaking only, rather than other seismic hazards, and only on certain categories of outcomes, such as direct losses to buildings. The practical result is that planners and policy makers receive at best only a partial picture of potential losses.

Of course, it can be argued that we have to start somewhere in projecting potential losses, and that any data obtained, however incomplete, is certainly better than none. Nevertheless, policy based on existing estimates is likely to be effective only to the extent that estimates do in fact approximate future losses. What if it is actually the harder-to-measure, poorly understood, secondary earthquake effects that end up actually costing more in the long-run? If the very creative loss-estimation work being done by is accurate,30 earthquake-generated fires are an extremely important area of concern. However, except for his work on San Francisco and Los Angeles, the fire problem has not been studied in depth in the United States. Thus, policy initiatives based on studies of the most obvious, direct effects of earthquakes may not be reducing losses as much as anticipated. At present, perhaps the best that can be done is to be explicit about the limitations of the methods used and about what loss estimates do not reveal about overall costs—while at the same time try to put as much emphasis as possible on understanding potential higher-order impacts.

Finally, the Panel on Earthquake Loss Estimation noted several ways in which the scope of loss-estimation studies sometimes limits their usefulness.18 State officials, understandably, want state or regional estimates, while local officials find most useful those studies that focus on their local areas. Other feedback provided to the panel gave additional insights on why loss-estimation studies have been of limited use. Among the problems cited were: (1) insufficient effort on the part of analysts to involve local officials and policy makers in the loss-estimation process, (2) conflicts and disagreements among experts that undermined their credibility, and (3) the highly technical style in which many reports were written. As has frequently been noted, policy makers and planners are much more comfortable with clearcut decision making criteria than with seemingly vague probabilistic projections. Ironically, the state-of-the-art probabilistic loss-estimation methodologies so favored by

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

professionals may be less "accessible" to users, and less likely to be used, than cruder but more clearcut projections.

Researchers invariably end their papers by talking about the need for more research, and this presentation will not be an exception. It is obvious that, in order to reduce earthquake losses, we need more empirical work to better understand what those losses might be. Presumably, a loss-estimation research agenda will be one of the outcomes of this forum. However, while researchers seek to obtain new data on probable losses, techniques also need to be developed that are sensitive to societal concerns about losses and useful products need to be developed that put different categories of loss into perspective.

PRESENTATION OF ROBERT W. KLING

Types of potential disaster losses can be categorized as follows: (1) direct economic loss, (2) indirect economic impacts, (3) loss of cultural environment, and (4) loss of natural environment. The third and fourth categories are actually components of the first and second categories; however, they are different enough in character that it is useful to consider them separately. This presentation will focus on the third category, loss of the cultural environment.

What is cultural environment? This is a broad notion, but first, a society's cultural environment includes its stock of historic and cultural assets, a broad category of historic monuments, human artifacts, and works of art that are important in providing identity and continuity to a society's culture.

Second, cultural environment includes intangible assets such as human relationship networks and an individual's sense of place, which, together, can be called social capital. The goal of this presentation is to give some indication of how the value of these kinds of assets can be incorporated into economic analyses of the value at risk from natural hazards like earthquakes.

What is a historic monument or cultural asset? Any definition is likely to be overly broad in some ways and too narrow in others. Yet certainly there is a universal sense of what is meant. As a starting point, think of a house. If it were destroyed, how would the loss be assessed? For most houses, the market value would be fairly easy to estimate, and a good measure of the loss. Alternatively, the replacement cost could be used. Often, these two numbers would give about the same measure.

But what about the most famous house in Fort Collins, Colorado: the Avery House? It is more than 100 years old, and almost everyone would agree that it is worth more than whatever might be implied by its ability simply to supply housing space. What about the White House? Is it worth more to the United States than its value as a luxury residence and an office building? What about a 2,000-year-old hut? It may not look like much, but what is it worth?

Going beyond houses, consider a pottery, image of the god Quetzalcoatl, more than 1,000 years old. What is it worth? In the Loma Prieta earthquake, the Museum of Asian Art in San Francisco lost two ancient Chinese vases

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

when they tumbled off their stands. The museum used market-price appraisals to make insurance claims, but what would be done if the Statue of Liberty tumbled off its base?

There are many pictures of the section of the Oakland Bay Bridge that dropped last fall. In that case, no one would assess the loss as anything but damaged transportation infrastructure. But what about the Golden Gate? Is it valuable to society only as transportation infrastructure?

In large measure, the value of such assets lies in their ability either (1) to give a sense of place in history and an identity shared by other people, times, and places, or (2) to provide examples of great aesthetic traditions. These functions are epitomized by the cultural roles played by the Ark of the Covenant or Michelangelo's statue of David. For the purpose of economic valuation, assets like these are differentiated by two key characteristics: they are irreplaceable goods, and they are public goods. Imagine the importance of this.

A common element among examples of cultural assets is that they are in some sense unique and irreproducible. Of course, in this case market price observations or other value bases will be few and will be subject to doubt about the effects of elapsed time and transaction irregularities. The surprising prices at which some Impressionist works have sold recently highlight the difficulty. Irreproducibility compounds valuation difficulties. The value of an easily replaced object normally would be near the cost of its replacement production, but what if no replacement can be produced, as for an Egyptian mummy?

A historic or cultural asset is also a public good, even though some are privately owned. Economists traditionally define a public good as one for which consumption is both nonrivalrous and nonexcludable . Cultural assets typically have both characteristics, and both raise important issues for valuation.

Nonrivalrous or joint consumption occurs when consumption of the object by one person does not at all diminish the object's ability to provide satisfaction to others at the same time. Such is the case with a work of art, for viewing the work does not use it up. Joint consumption affects economic valuation because the value of the object must be measured, not as the value to a single person who would consume it, but as the sum of the values to all persons who would partake in its benefits.

Nonexcludability in consumption occurs when the benefits from a good cannot be limited to those who help pay for it, when the benefits automatically extend to everyone in the community. In the case of a work of art, individuals may be excluded from actual viewing, but certainly not from their cultural stake in the object. In other cases of cultural assets, excludability even from viewing may be impractical.

Nonexcludability affects economic valuation, because it normally prevents public goods from being provided via markets. Therefore, market prices for such goods are seldom observed. The Statue of Liberty, for example, provides a political and cultural symbol for hundreds of millions of people. Though an

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

admission fee may be charged for an actual visit, no private owner of such a monument could ever charge every citizen for the benefits they gain from the monument.

Even when cultural assets such as works of art are privately owned and traded in markets, they retain many public good aspects. Therefore, a market price is unlikely to capture all the asset's value to all the individuals with a stake in art as part of culture.

Thus, cultural assets tend to be unique, irreplaceable, and enjoyed jointly and freely. For all these reasons, market prices are often inappropriate or unavailable as standards of value, and special assessment methods must be employed. For goods traded in normal markets, prices usually reflect either replacement value (i.e., reproduction cost) or benefit value, or both. What about for cultural assets?

Replacement Value

The first candidate as a measure of value is replacement value, or replacement cost. Many cultural assets are irreplaceable assets; consequently, reproduction cost rarely will be an appropriate standard of value for these assets, and one must turn to benefit value.

Benefit Value

While replacement cost relates to the supply price of an asset, benefit value focuses on its utility to demanders. People who act as demanders or beneficiaries of cultural assets will value them for several reasons. These include the benefit they derive from actually enjoying the asset themselves, the benefit from preserving the option to enjoy it in the future, and the benefit they wish to bequeath to future generations. In addition, there may be indirect benefits from the asset's existence, even to those who are not demanders.

Use Value. The most obvious component of benefit value is value in use. This is the dollar value of benefits that accrue directly to people who take the opportunity to enjoy the asset by visiting the site and is the same kind of benefit that normal consumption goods provide their users. Since cultural experiences often are unpriced or are priced on a nonmarket basis, people's willingness to pay for this kind of experience cannot be observed accurately in markets.

Option Value is the dollar value placed on the potential consumption services offered by an asset, services that may or may not ever be actualized. People are likely to be willing to pay to have the asset available, in case circumstances lead them to want to see it. Option value is thus a type of conservation value.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

Bequest Value is another type of conservation value people attach to assets they are not presently using. In this case, the value comes from the opportunity to pass on potential benefits to future generations. For private assets, willingness to pay for such an opportunity is widely evidenced within families. For irreplaceable public assets, the bequest is necessarily communitywide; individuals' willingness to participate in those bequests is hard to measure, but is important nonetheless.

Tradition Value. In many cases, historic assets and certain other artifacts are instrumental in contributing to the cultural identity of a community or a nation, in ways beyond the apparent value to individuals. This can be called tradition value. The preservation of social continuity and community identity enhance the well-being even of those individuals who would claim no appreciation for the asset based on use, option, or bequest value. In these cases, there is a value in the very existence of the asset, independent of actual or potential personal experience of it.

How can cultural asset values be estimated? Current practices for the evaluation of such goods are few, even nonexistent for many types of assets. For insurance purposes, museums typically value their collections at estimated market prices, yet many museum personnel express the view that some objects are truly priceless, and valuation is pointless. Historic preservation experts indicate that economic value is discussed, if at all, in terms of the economic side-benefits that preservation can bring to a community's economy, and a satisfactory methodology for assessing a historic building's intrinsic value has yet to be developed.

The primary valuation method is market-price appraisals. But since actual prices often are not observed for cultural assets, or since prices often do not reflect the full social value of the asset, substitute methods of valuation are often needed. Some helpful valuation methods adapted from the field of environmental economics include:

  • opportunity-cost method;

  • contingent-valuation method; and

  • travel-cost method.

Market-Price Appraisal

Market price is a natural way to assess value to cultural objects that commonly are bought and sold. In a well-functioning market, the sales price of an object reflects both what a buyer is willing to pay to acquire it and what a seller demands as compensation for giving it up. From either perspective, the price provides a measure of the benefit the asset yields its owner.

Cultural assets are normally unique, making market-price appraisal difficult. However, expert appraisers have established methods of valuation that can usually give reasonably good estimates of market value.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

Market prices are thought to be the most objective measure available of how buyers and sellers value goods, and that is an advantage. Though partly subjective, appraisal of an asset's likely market value, based on the actual selling prices of comparable objects, is just one step removed from observation of an actual price for the asset.

An obvious limitation on the use of market-value appraisal is that it cannot be applied to assets for which there are no markets. For example, normal appraisal methods could not yield a market value for the Statue of Liberty. Another important limitation stems from the public goods nature of many cultural assets. The price at which an asset sells would understate its full social value. A museum that acts as buyer of art or artifacts on behalf of society may offer a price that reflects social value to some degree, but underfunding of such institutions guarantees that they will be unable to express society's full demand for such objects. The same is probably true in the area of historic preservation. Therefore, for the purpose of social valuation of cultural assets, private-market appraisals can at best serve as lower bounds for the actual values.

Opportunity-Cost Method

A second way to assess willingness to pay for an asset is to tally the opportunity costs of the resources currently dedicated to keeping it. For example, suppose a cathedral in downtown Manhattan is preserved at the sacrifice of $200 million that could be had by converting the land to high-rise office space. Then one can conclude that the landmark is valued implicitly for at least that much by the owner, normally a public entity or nonprofit group acting on behalf of the community.

There are other ways society has shown its willingness to make sacrifices to preserve an asset. The temple of Ramses II, threatened with inundation in the 1960s by the Aswan High Dam, was moved to higher ground at a cost of many millions of dollars, funded by 50 nations and UNESCO. Other assets were left to be submerged, though, apparently a signal that they were not valued so highly. The Acropolis is another monument that is being destroyed by a man-made disaster, air pollution. Apparently the cost of eliminating that threat is too high.

An advantage of the opportunity-cost method is that it uses observable, market-based information, such as land values, to infer a minimum value for a nonmarketed asset. So long as one can have confidence that the owner is rationally preserving the asset on behalf of the public, the inferred value would include option, bequest, and tradition values.

The main problem with the opportunity-cost method is that it often does not measure policy-relevant value. The method estimates what people appear willing to sacrifice to preserve the asset; ironically, though, it is measuring the alternative value those people could have if the asset were destroyed. In the case of the cathedral, for example, if an earthquake demolished the structure,

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

the land could be sold for $200 million. To value the cathedral at $200 million would therefore be to imply that the earthquake caused no net loss: a $200 million cathedral is lost but $200 million in commercial real estate is gained. There is real net loss, though, which is the mount by which the cathedral is valued beyond the $200 million land cost. Unfortunately, observation of opportunity costs yields no information about the net value of the asset over and above those opportunity costs. The usefulness of the opportunity-cost method is sometimes limited to prodding corroboration of values generated by other means. This is not true, however, when using sunk costs, like those involved in moving the Ramses Temple.

To infer the value of asset preservation from opportunity costs, one must have confidence that the owner's decision, whether deliberate or by default, is a rational reflection of public desire. Bureaucratic inertia, for example, might lead to preservation unjustified by real value, and opportunity costs might overvalue the asset. On the other hand, institutional blockages may give signals that undervalue the asset, as in the case of the Acropolis.

Contingent-Valuation Method

A third possible approach is the contingent-valuation method, which is based on direct surveys of a sample of the population (either the "user" population or the population at large). For example, households might be surveyed to discover their willingness to pay to have the Statue of Liberty restored to its original condition and reinforced for its long-term preservation.

In this case, the absence of an observable market value for the asset is handled by creating a hypothetical market in which respondents are asked to make hypothetical economic decisions. The accuracy with which these hypothetical decisions represent real economic valuations depends upon the care with which the survey instrument is constructed.

An important advantage of the contingent-valuation method is its ability to capture the value of the asset to the entire population, including option and bequest values. Furthermore, the measure of benefit given by the contingent-valuation method is a policy-relevant measure; it tells how much people value the asset beyond the opportunity cost they bear to obtain its benefits.

However, because contingent valuation involves surveying a sample of the population and posing hypothetical market scenarios, the method is subject to normal survey biases and to varying interpretations of the meaning of responses. For example, respondents may overstate their willingness to pay for an asset, particularly if they think their statements will affect public policy and yet they may not have to help finance the policy. In any case, the survey instrument must be very thoughtfully constructed and carefully interpreted.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

Travel-Cost Method

The travel-cost method is intended as a direct measure of what people are willing to sacrifice to benefit from the asset. For example, one measure of the cultural value of the French Quarter in New Orleans is the amount of time and money people are willing to spend to experience its historical, architectural, and cultural character. This is a variation of the opportunity-cost method. However, the travel-cost method is more useful, for it allows inference of the value people receive over and above what they pay for it; this net benefit is a policy-relevant measure of value.

A more extended example to illustrate this method is a fictional numerical case of a famous old lighthouse. Table 2-12 shows calculation of travel costs, including out-of-pocket expenses and opportunity costs of personal time, depending on how far a visitor comes. Then it relates frequency of visits to that cost.

One measure of the total value of the lighthouse might be suggested by summing the amounts people spend to visit it. In this case, total expenditure (including opportunity costs) amounts to over $1 million per year. One might take this level of expenditure to imply a lower bound for the benefits visitors receive per year, and then infer a value for the lighthouse. But this amount represents the opportunity cost of the lighthouse visits, and if the lighthouse were destroyed, the visitors could receive this same amount of value elsewhere. The real policy-relevant value is the benefits the lighthouse yields over and above the opportunity costs of its services. For instance, Zone A visitors pay an average of $6.30 per visit, though many would be willing to pay more, say $10 or $15.

Table 2-13 uses the relationship between costs and visits shown in Table 2-12, to estimate how much extra visitors would be willing to pay and still come. By estimating how many people would still visit at various cost levels in excess of actual cost, the so-called consumer surplus is calculated. In this case, a rough estimate of the annual net benefit to visitors is about $193,000 per year. A final step is to capitalize this yearly benefit, to infer an asset value. For example, at an interest rate of 9 percent, the lighthouse value would be about $2.1 million.

The main advantage of the travel-cost method is that it generates an asset value using concrete, market-based proxies to substitute for the unobserved user price. It is based on actual (not hypothetical) information about user decisions and implied costs. Also, although the travel-cost method is related to the opportunity-cost method, it is superior in its ability to give a policy-relevant measure of the value that would be lost if the asset were destroyed.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

TABLE 2-12 Per Trip Costs and Per Capita Visits

Travel Data

Calculation of visit cost per person

Calculation per capita visits

Zone of origin

Average round-trip distance from this zone

Round-trip travel hours from this zone

Site visit time

Direct travel cost at $0.15 per mile, 2 persons per car

Opportunity cost of time, at $1.20 per hour

Entry fee

Total visit cost from this zone

Visits from this zone

Population of this zone

Per capita visits from this zone

 

 

 

 

(D)

(O)

(E)

= D+O+E

(V)

(P)

= V÷P

A

20 miles

0.5 hours

1 hour

$1.50

$1.80

$3.00

$ 6.30

6,000

40,000

0.15

B

60 miles

1.5 hours

1 hour

$4.50

$3.00

$3.00

$10.50

80,000

800,000

0.10

C

100 miles

2.5 hours

1 hour

$7.50

$4.20

$3.00

$14.70

10,000

200,000

0.05

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

TABLE 2-13 Visits Demanded at Various Cost Increments

 

At actual cost

If costs were $4.20 higher

If costs were $8.40 higher

If costs were $12.60 higher

Zone of origin

Cost is:

Visits from this zone are:

Cost would be:

Per capita visits would be:

Visits from this zone would be:

Cost would be:

Per capita visits would be:

Visits from this zone would be:

Cost would be:

Per capita visits would be:

Visits from this zone would be:

A

$ 6.30

6,000

$10.50

0.10

4,000

$14.70

0.05

2,000

$18.90

0

0

B

$10.50

80,000

$14.70

0.05

40,000

$18.90

0

0

$22.40

0

0

C

$14.70

10,000

$18.20

0

0

$22.40

0

0

$26.60

0

0

Totals

 

96,000

 

 

44,000

 

 

2,000

 

 

0

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

The travel-cost method has several disadvantages and limitations. Because the value estimate is based on information only about actual users of the asset, it indicates only value in use, and not option, bequest, or tradition values. Those who value the option of visiting a site, but have not yet done so, have made no observed choices that will be incorporated into the data. Second, the method is difficult to apply to an asset that typically is visited in conjunction with several others during a single trip, because the total costs of the trip must somehow be allocated among the various sites visited. This complication would occur in applying the method to value each of the monuments in Washington, D.C., or each of the paintings in an art museum. For these reasons, the travel-cost approach would be used most easily when visiting the asset site requires a short trip dedicated to that asset alone, or when one wishes to value a group of assets collectively.

This discussion has been intended to give a flavor of the methods that might be considered to assess cultural asset values. None of these approaches is yet very refined. Analysts hope to explore their potential in the near future.

Loss of Social Capital

The term social capital has been used many different ways. For the purpose of this presentation, social capital is the set of human ties that a community's members find in that place. These ties can be of at least three types: friendships, professional relationships, and an internal sense of stability or home. Such relationships might at first be considered entirely noneconomic and beyond the scope of a tallying of economic damages. However, their economic value can indeed be assessed, because this social capital is an extremely valuable asset that people evidence a very high willingness to pay to preserve.

A piece of capital is an asset that yields income. We think of social ties as capital in the sense that they are assets that yield psychic income that would be important to a household along with pecuniary income. Now, psychic income is a concept rather well developed by economists, and it can have many sources other than social ties: job conditions, public amenities in a place, natural beauty, climate, etc. But this psychic income from social capital is different in at least one key way; it is nontransferable, since it is attached to specific individuals and is developed historically over time. If one family moves out of a community and another just like it moves in, the new family could take over enjoyment of the mountains, the town, the schools, etc. just as much. But they could not take over the human ties.

The destruction of social capital could be a natural consequence of the forced migration that could follow a major disaster; one might think of such capital as a special type of irreplaceable cultural asset, but one that is more personal and intangible than those discussed earlier.

The evidence of social capital in our society is widespread. The best evidence is the fact that people like to stay where they are, despite many

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

incentives to move elsewhere. There are all manner of natural disasters, as well as economic disasters like plant closings and agricultural depressions, that one might think would be a large inducement to relocate, . . . but people stay.

It is impossible to deny the existence of social capital. But what is the importance of this social capital from a public policy perspective? When there is a disruption, when a move is forced, a valuable asset is destroyed. Prevention of that destruction then has a value that might be included in the benefits side of various benefit-cost-based public policy decisions, just as the protection of material assets is included. This is an element that is rarely explicitly included in formal benefit-cost analyses.

In the area of natural-hazard management, we can see that hazard-mitigating types of investment decisions might be made to take into account the value of saving a town, where the town is viewed as more than just a collection of real estate and physical capital. Even programs normally denounced, for imposing the condition that the relief be applied to rebuilding on the original spot, may be seen as serving (perhaps to excess) the objective of preserving social capital.

Social capital is not bought or sold. So, as for the assets mentioned earlier, less-direct methods of assessing its value will be required. One might use a contingent-valuation approach, surveying people to find out what their social capital is worth. Alternatively, one might look for ways social capital affects economic decisions.

Two markets that will be affected by migration are housing markets and labor markets. Residents' hesitance to leave may be reflected in the prices at which they are willing to sell their homes Or, their desire to stay may show up in the wages they are willing to accept; it might take surprisingly large wage drops to induce people to leave their hometowns. In economists' terms, labor supply would respond inelastically to wage drops. This idea can be expanded upon as an example of how an economist might use objective market data to estimate intangible, subjective values. These ideas probably make more or less sense, depending on whether one is more or less familiar with the economists' concept of supply and demand curves, and their interpretation.

The phenomenon common to the three examples cited earlier is the fact that many communities appear to face a dual labor-supply curve, or what might be seen as a kinked curve. In growing, these communities find that a small wage differential relative to other regions will induce a sizable labor influx (Figure 2-13). Labor demand shifts out, wages rise a little, and the labor force grows a lot.

But a reversal of economic growth generally will not lead to a labor force shrinkage that is so elastic (Figure 2-14). If labor demand drops back down, wages and employment will move, not back along the original labor supply curve, but along the steeper, dashed curve. The result is that wages drop more significantly, but emigration is not significant. The clearest explanation for this dichotomy involves the stock of social capital that a region's inhabitants establish as they spend time there. Reluctant to abandon this social capital, which has a significant human value, workers are willing to accept more-

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

FIGURE 2-13 Effect of a local labor demand increase.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

FIGURE 2-14 Effect of a local labor demand decrease.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

severe wage cuts in order to remain in what has become their home community.

Figure 2-15 is a more elaborate version of Figure 2-14. For simplicity, the labor-supply curve is drawn horizontally. Labor demand is shown as falling by steps, with more workers leaving the labor force at each step. In each case, the heavy line segments indicate the wage gap that is required to induce that marginal worker to move somewhere else. It thus measures the money the worker is willing to accept elsewhere in compensation for giving up his or her ties locally. If labor demand falls all the way to the lowest level (L3), and if the social capital value is added up for all the workers who leave, the value of social capital lost would be the area of the shaded triangle in Figure 2-16, which could be measured if the labor supply and demand curves were known. This is a loss to those who were caused to move.

The shaded triangle, by the way, measures the wages lost by those who still choose to stay. That is an indirect economic loss of the sort that will be discussed in Chapter 3.

It is important to note that a disaster need not cause lower labor demand for this analysis to be relevant. Even if emigration were caused by destruction of housing stock, for example, estimation of labor supply and demand curves could indicate the amount of social capital at stake.

In measuring social capital in this way, adjustments would have to be made for at least two important factors. First, labor immobility could be due to other ties that are not social capital: equity in a house, for example. That would have to be accounted for. Second, some communities might be better able to afford wage drops, to preserve social capital, than others. By affecting how much they show a willingness to pay for social capital, that would affect the implied dollar value of that intangible asset, in a way that might not be justified on philosophical grounds. Economic values might be adjusted accordingly.

In conclusion, there are important intangible benefits that are at risk from natural hazards. Though they are intangible, there are promising methods available for assessing their value in economic terms, so that they can be included in economics-based policy analysis.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

FIGURE 2-15 Social capital lost from relocation.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

FIGURE 2-16 Losses to workers from lower labor demand.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

GENERAL DISCUSSION OF CHAPTER 2

QUESTION: A number of the issues you raised are quite rich, and all of them could be discussed at great length, but I have just a couple of points. First, some of these issues have been discussed at great length by NAPAP, The National Acid Precipitation Assessment Program. For example, Joel Shirago of EPA did a fairly careful analysis of some of these issues and the approaches. I think we have a long way to go before they come up with really useable results, but there will be an economic assessment done on the impact.

Second, it is amusing that the Cape Hatteras Lighthouse was referred to, because the Academy undertook a major study for the National Park Service of the cost associated with moving the Cape Hatteras Lighthouse or protecting it with copper gowns and so on. Many of the disaster scenarios on the erosion effect will go into the next study.

DR. KLING: That is interesting because the National Park Service has a duty to assess historic assets; yet their reports indicate that they are full of great assessments of the cost of protecting or restoring these assets, but there is nothing about benefit. Is it worth a million dollars? Is it worth $20 million? Is it worth $100 million?

QUESTION: This is for Don Friedman. Does Travelers Insurance currently use your models to determine insurance rates or what markets to participate in? In other words, are these models now influencing the actual insurance policy of the company?

DR. FRIEDMAN: For the last 35 years I have attempted to answer questions that management has on the various effects of natural disasters. This is one input that would go into whatever other considerations they might have in terms of competition or what have you. Yes, this is an input into the decision-making process.

QUESTION: Dr. Friedman, have you tried to test your method against the results of the Loma Prieta earthquake?

DR. FRIEDMAN: Yes. Simulation modeling is never finished because it is a continuing process. Each time you get a new event, you try to look at it in terms of verification, calibration, and so forth. Fortunately, it worked out well.

QUESTION: Maybe I should change my question. When you did that, how big a change did the Loma Prieta have on your model? I was asking whether your procedure would have given a similar loss estimate to what the actual losses were from the Loma Prieta earthquake? In other words, how good is your model?

DR. FRIEDMAN: Good is a relative thing. From the point of view of decision-making, we were very satisfied with the results. The insurance industry cannot afford to go to each structure and do an engineering analysis. What is going on below the surface is still unknown. But from the point of view of an insurance operation, the law of large numbers rules.

QUESTION: This is addressed to all four panel members from your various points of view. Each of you, in some way, addressed the secondary

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×

effects that occur from earthquakes. These effects are obviously very difficult to insure against. They are hard to measure, etc., and yet we as a society have to deal with them in some fashion. I wondered if there were some thoughts that the panel had about how we address these effects in terms of both the concept of loss mitigation and also with the question of the role of government and who should pay, who should be involved in this? I think these are questions we are going to have to address. All the remarks were very stimulating in terms of raising those issues.

DR. FRIEDMAN: One thing I did not get a chance to discuss was the payoff by various lines of insurance for a large insurance company. How many assets would a company have to sell off in order to settle, perhaps, hundreds of millions of dollars in claims? You would look at the various distributions in terms of types of insurance. This reduces the ability of these companies to sell more insurance because of the ratio between the amount of money they have on reserve and the amount of premiums they can put out into the field. There are obvious difficulties that must be implied from a catastrophic event—a great earthquake, a great hurricane—in terms of the ability of individual insurance companies to continue to sell insurance in other parts of the United States in addition to the area where the event occurred.

DR. ARNOLD: There is a balance sheet. The balance may not come out exactly, but one has to look at the balance sheet. I know there is going to be a lot of focus on losses. However, there are also gains. The insurance company tends to operate as a closed system and look only at its own balance sheet.

If you look at the whole universe, it is an open system, and we must look at the gains, even though they do not balance out. In terms of Dr. Kling's presentation, it is great to see an economist using architectural arguments; but even there, there are gains. For instance, I would argue that Mexico City is a much better city environmentally now because of the earthquake. There are two reasons: (1) it now has a lot of small parks which are environmentally very satisfactory in a city which was very short of open space; and (2) it now has 44,000 units of well-designed, affordable housing which it did not have before. Both of these things were forced by an unfortunate circumstance, but if you look back at the history of cities and the evolution of cities, you will see that earthquakes and catastrophes are part of the natural process. So, we should keep an eye on the balance sheet as we proceed.

DR. TIERNEY: We are continually hearing about how expensive it is and how difficult it is to study certain kinds of problems, to develop certain types of inventories, to look at certain kinds of variables. Wouldn't it be nice if there could be some modest efforts directed at trying to get a handle on some of the more elusive effects, just to balance off the tremendous emphasis on some other types of earthquake-damage effects that we see. This is a plea for more empirical research in this area.

Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 35
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 36
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 37
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 38
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 39
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 40
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 41
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 42
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 43
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 44
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 45
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 46
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 47
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 48
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 49
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 50
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 51
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 52
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 53
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 54
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 55
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 56
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 57
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 58
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 59
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 60
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 61
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 62
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 63
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 64
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 65
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 66
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 67
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 68
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 69
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 70
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 71
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 72
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 73
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 74
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 75
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 76
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 77
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 78
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 79
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 80
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 81
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 82
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 83
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 84
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 85
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 86
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 87
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 88
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 89
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 90
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 91
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 92
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 93
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 94
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 95
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 96
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 97
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 98
Suggested Citation:"2 WHAT ARE LIKELY CATEGORIES OF LOSS AND DAMAGE?." National Research Council. 1992. The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum. Washington, DC: The National Academies Press. doi: 10.17226/2027.
×
Page 99
Next: 3 OVERVIEW OF ECONOMIC RESEARCH ON EARTHQUAKE CONSEQUENCES »
The Economic Consequences of a Catastrophic Earthquake: Proceedings of a Forum Get This Book
×
Buy Paperback | $55.00
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

This book presents the proceedings of an August 1990 forum held at the National Academy of Sciences in Washington, D.C. Topics covered include the current and potential roles of the private sector and the various levels of government before, during, and after an earthquake occurs, and alternative strategies that could be implemented to reduce the economic impacts, with emphasis placed on the role of the insurance industry.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    Switch between the Original Pages, where you can read the report as it appeared in print, and Text Pages for the web version, where you can highlight and search the text.

    « Back Next »
  6. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  7. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  8. ×

    View our suggested citation for this chapter.

    « Back Next »
  9. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!