Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 216
Working Paper G
Economic Aspects of
Earthquake Loss Estimation
The economic consequences of an earthquake are presented in
most loss studies only as direct property losses, usually estimated as
a percentage of replacement cost. While these estimates do provide
some indication of the financial resources needed for reconstruction,
another reason for often quoting direct losses in dollar terms is one
of convenience. For planning, preparedness, and recovery purposes,
one could just as easily use only estunates of the numbers and the
types of structures with varying qualitative degrees of damage.
The study of natural hazards has long been dominated by engi-
neers, sociologists, geographers, and social psychologists (Cochrane,
1984~. Few economists have engaged in this field of study, leaving
a large gap in knowledge of the overall economic accounting of the
consequences of catastrophic earthquakes or other natural hazards.
This does not mean, however, that these consequences are insignif-
icant. Rather, it reflects the difficulties involved in conducting a
comprehensive economic accounting of the effects of an earthquake.
The preceding working papers clearly demonstrate the complex-
ities surrounding procedures for estimating direct earthquake losses.
Efforts to estimate the indirect economic effects can complicate the
study procedures significantly, particularly with respect to collecting
additional information about structures and identifying the interrela-
tionships among sectors in the economy and how they would change
after the event. It is unlikely that these extended analyses will soon
216
OCR for page 217
217
be incorporated into large-scale, general-purpose loss estimate stud-
ies. Interest in better understanding the economic consequences of
earthquakes, however, led FEMA to sponsor an ambitious study,
ATC-13 (Applied Technology Council, 1985), to lay the groundwork
for estimating these impacts in a comprehensive fashion. This paper
attempts to place a number of ATC-13's procedures in perspective,
to discuss the current feasibility of doing such comprehensive eco-
nomic analyses, and to outline briefly a research agenda that might
enhance the feasibility of future studies.
CHARACTE1tIZING ECONOMIC [OSSElS
The economic consequences of an earthquake can be classified in
several ways, but for purposes here, three types are delineated: (~)
direct losses due to damage; (2) losses due to premature death or
injury, and (3) indirect losses due to business disruption. Estimates
of the direct property losses follow in a straightforward fashion from
damage estimates, but the other two types of losses warrant some
further discussion.
As a first approximation to losses from premature mortality,
Sorkin (1982) suggests multiplying the expected number of deaths
by the present value of expected future earnings foregone, considering
the likely age, sex, and occupational profiles of the victims and their
effects on expected future earnings. The indirect costs of injuries are
reflected in foregone earnings and medical costs. In extremely severe
earthquake events, these economic losses could be substantial and
certainly tragic for the victims' families. Estimates of this kind may
also be important for insurance purposes or other questions of legal
liability. However, the majority of these losses are in the form of
foregone future earnings, rather than immediate out-of-pocket costs.
For this reason, in addition to the tre~nendous uncertainties
surrounding casualty estimates, these losses should not be a major
focus of economic loss studies. The public concern should be with
the casualties themselves and efforts to reduce them, rather than
foregone future earnings.
However, the same conclusion cannot be applied to indirect busi-
ness losses stemming from physical damage and disruptions clue to
the earthquake. These indirect losses are immediate and can persist
throughout the recovery effort. They can affect the entire region and
spill over to other states and regions of the country. For a variety
of reasons, ranging from hazard mitigation and recovery to concerns
OCR for page 218
218
about national security and increased vulnerability after the event,
these indirect losses are potentially of major concern to the local
economy and to the federal government.
MEASURING INDIRECT ECONOMIC IMPACTS
FEMA's ambitious study to identify in a comprehensive fashion
the economic consequences of a catastrophic earthquake has two ma-
jor components. The first, ATC-13, involves a damage estimation
technique that integrates geocoded seismic intensity simulations and
inventories of buildings and other facilities with damage functions,
relating seismic intensity and construction characteristics to dam-
age estimates. The second component is designed to determine the
overall economic impact by using the results from the damage evalu-
ation methodology in conjunction with recently developed economic
interindustry modeling capacities.
ATC-13 describes only the first component of FEMA's study
design. Its loss estunates are confined to the direct effects of the
earthquake (e.g., damage from ground motion and collateral hazards)
along with estimates of casualties, property loss (measured as a
percentage of replacement cost), and loss of function. From this
standpoint, its objectives are not that much different from those of
other studies or approaches.
However, the procedures by which estimates of these losses are
produced differ significantly from what others have done. One major
difference is the level of detail attempted in terms of the number
of construction classes and the classification of economic and social
function. The attempt to add detail to the damage relationships
by consulting a number of experts was unique, as was the attempt
to generate a comprehensive inventory from socioeconomic data in
automated form available from FEMA. The rationale for the inven-
tory procedures was in part due to a desire for consistency in studies
throughout the country.
Shortcomings of FEMA's methodology stem from the large num-
ber of construction and use classifications and the fact that the it-
erative process used with the experts led to distributions that may
underestimate the true variability in damages. The accuracy of infer-
ring structural information from the social and economic functions
of buildings Is questionable and has not been empirically verified.
In terms of the damage relationships, it is probably true that
little would be lost by considering a smaller number of separate
OCR for page 219
219
damage curves or matrices. If this were done and the estunates were
not revised through this iterative process, the damage relationships
would probably not be too much different from those used in other
studies. The real shortcorn~ng of the method is in having to relate
economic and social function to structure type at such a disaggregate
level at the level of each individual building or other facility.
One way that the procedures could be improved ~ to invest more
time and money in collecting more detailed information about the
use of structures in the inventory. Art alternative might be to conduct
some general field research to determine if there is any systematic
relationship concerning economic function, geographic location, and
age and type of structure.
Why was such a high level of disaggregation needed in the ATC-
13 study? The answer derives from FEMA's interest (or that of the
National Security Council, which requested the study) in identifying
the impact of an earthquake on any one of up to 470 economic
sectors identified by the Standard Industrial Classification (SIC)
code used by the Bureau of Economic Analysis (Executive Office
of the President, 1972~. This motivation ~ probably related more
to the national security implications of loss of function to specific
defense or related high-technology industries than it is to education,
mitigation, and planning efforts.
If the first phase of the ATC-13 methodology could be imple-
mented at this level of detail, then some initial estunates of loss of
function to defense related or other ~critical" industries might be
possible. However, these direct damage and loss estimates ignore im-
portant secondary effects throughout the economy after catastrophic
events. (This is true regardless of the level of disaggregation in the
analysis.)
These secondary impacts are due to a variety of things. Probably
most important is the loss of productive capacity from damage to
physical plant and equipment. This reduces the capacity of the
economic sectors to produce goods for final consumption as well as
for use as intermediate inputs (some of which might have strategic
value) in other productive activities. Because of the damage to the
area's productive capacity, a larger fraction of the area's continuing
demands for goods and services need to be imported from other
regions of the country, at least during the recovery period.
Employment and income in those sectors damaged by the event
are reduced also, and this in turn reduces the demand for goods and
services in many of the region's economic sectors. However, recovery
OCR for page 220
220
activities bring with them an influx of financial resources (e.g., from
government recovery and relief efforts, and insurance cIaims) that
increases the demand for the output of certain sectors, particularly
construction. These new demands are either met by the remaining
productive capacity of the area or through interregional imports.
The purpose of the second phase of FEMA's study is to attempt
estimates of these secondary impacts at the four-digit SIC level. In
theory, this is possible by using an interregional interindustry mode!
of the U.S. economy. The most complete description of the mode!
intended for use In conjunction with ATC-13 ~ In a paper by Wilson
(1982~.
The methods to be used in this phase of FEMA's study can
be described in abbreviated fashion through simple equations. The
basic interindustry, input-output (~-O) mode! developed initially by
Leontief (1951) is described In numerous economic books and in a
summary by Wilson (1982~. The mode} ~ developed essentially from
a double-entry bookkeeping description of an area's economy that
records purchases and sales of goods from one sector to another, as
well as imports and sales to final users (e.g., to final demand).
Total sales or output of any sector (e.g., agriculture, manufac-
turing, and services) of an e-sector model are recorded along the
rows of the transactions table and are expressed as
~Xij+Yi=Xi(i=l' ...),
j=1
(1)
where xij is the value of the output of sector i purchased by sector j,
pi is the final demand for the output of sector i, and xi is the value
of the total output of sector i.
To complete this set of balance equations, the entries down the
columns of the table also add to the value of a sector's output.
I`
xij + pj = Xj(j = 1, · ~ n)'
i=1
(2)
where pj is the final payments (purchases of imports and primary
factors of production by sector j], Xj is total outlay (purchases) of
sector j, and ~ equals Xj for all i= j.
From this transactions table, a matrix, A, of direct input re-
quirements from sector i (in dollars) per dollar of sector jets output is
given by
OCR for page 221
221
A = As = { $i] } (i, j = 1, . . ., n). (3)
Substituting (4) into (~) yields
n
Xi = ~ aijxj + yi(i = 1, . . ., n), (4)
j=1
which may be expressed more compactly as
X= AX+Y,
where
Ax
x2
X= .
-On -
I all al2
a21
a22 · · ~
... al"
a2"
, A= .
·
·
a"1 at .·· ann"
(s)
IY1 -
, and Y =
An
. (6)
Rearranging tints set of equations, it is easy to see that gross
output minus intermediate use equate the net output or final use of
the system
X—AX= (I—ALEX= Y.
(7)
In the econorn~cs literature, much of the policy analysm that uses
interindustry models is focused on the fact that this set of equations
can be used to estimate the total output in the economic system
required to meet any given set of final exogenous demands (e.g.,
consumer demand, government purchases, and exports). That is, if
one knows the specific values for the components of Y. one can solve
for required output by
X= (I—A`J-iY.
(8)
In the planned second phase of the economic study, FEMA
would make use of the direct damages and loss estimates coming
out of the AT~13 method. The first task would be to estimate the
interindustry mode} for the geographic area of interest (e.g., estimate
the predisaster A matrix). Historically, this has been done either
through extensive questioning of a sample of local businesses (Bills
and Barr, 1968), or through systematic adjustments to the national
OCR for page 222
222
interindustry table based on some measure of the region's economic
activity in a particular sector to that of the nation (Boisvert and
Bills, 1976; Hwang and Maki, 1979; Lofting and Davis, 1973~.
To estimate economic losses from natural hazards, these nonsur-
vey techniques are the only feasible approach, and FEMA chose to
use the procedures developed by Lofting and Davis (1973), which are
based on a biproportional matrix-balancing technique (RAS) devel-
oped by Stone (referred to in Wilson, 1982 and Boisvert and Bills,
1976~. The procedures by Lofting and Davis, and Hwang and Maki,
accommodate the development of integrated interindustry models
that account directly for trade flows across more than one region and
can trace the impact to other regions in the country. Wilson (1982)
discusses this extension of the model.
Once the interindustry mode] is in place, on the basis of the initial
direct loss estimates, procedures would be developed to estimate new
levels of fin e] demand, Y. in the postdisaster situation. This would
require establishing estimates of the loss in income due to the event
and the projected influx of resources due to recovery efforts, as well as
estimates of how these changes affect final demand for each sector's
output. Projecting changes in final demand as a result of disruptions
in an economy (be they due to economic or other factors) is not an
easy task, but it is something that is done frequently in interindustry
studies.
The third task would be to modify the interindustry tables for
the region. That is, in most interindustry studies, it is assumed that
the intermediate input requirements, the A matrix, is invariant to
the initial change in economic activity. This, of course, could not
be assumed after an earthquake because of the damage to plant and
equipment and the corresponding reduction in productive capacity.
In general, this would mean that many of the components of the
matrix A would be reduced indicating that more of a sector's in-
termediate input requirements from other sectors would be imported
from outside the region. There has been very little, if any, work
attempting to modify interindustry models to account for an imme-
diate structural change in intermediate input flows caused by a major
disaster.
SUMMARY AND CONCLUSION
Little comprehensive analysis of the overall economic impact of
earthquakes on a regional economy exists but an economist's general
OCR for page 223
223
knowledge of a region's interindustry relations would suggest that
the secondary (or indirect) eEects stemming from the initial damage
are likely to be substantial. It would be useful to link our estimates of
damage to buildings and other facilities with their economic function.
This information could assist recovery by helping to set priorities
for reconstruction of essential services and perhaps to identify the
location of industries that use toxic or other hazardous substances
that could be released during the earthquake.
The key question is, however, At what cost? Data to imple-
ment the procedures do not exist, en c! if the inventory of facilities
had to include data on economic function, the costs of this phase
would increase substantially (by as much as 40 percent by one esti-
mate). furthermore, even if there were reliable estimates of direct
losses to structures by economic function, serious problems remain
in trying to relate direct losses to changes in final demand and other
interindustry relationships. These difficulties can only be resolved
through additional research.
Regardless of how rapidly some of the research problems are
resolved, it is unlikely that comprehensive economic analysis will
be viewed in the near future as an integral part of what has been
called Type ~ studies (general purpose, large scale) in Working Pa-
per A. This does not mean that the procedures used in future loss
estunation studies should be insensitive to the data requirements of
more complete economic analysis of the consequences of catastrophic
earthquakes. At a minimum, researchers should collect inventory
information that relates construction class to economic and social
function or undertake specific research to establish any systematic
relationships that Knight exist.
Furthermore, to be useful for hazard reduction, emergency plan-
ning, and recovery planning efforts, the level of detail in terms of
econorn~c and social function does not need to be fine enough to
differentiate all 470 sectors. A reasonable objective would be to look
initially at the 25 to 30 major economic classifications defined by the
SIC, with the expectation that there might be a Landfill of impor-
tant individual industries in any region that could be exaTruned in
greater detail. These would depend on the location being studied
and the purpose of the study. Major defense contractor plants and
military bases could be studied in greater detail if the purpose is
defense-related, as In the case of ATC-13.
Representative terms from entire chapter:
direct losses
