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

This chapter focuses on inter- and intraregional shifts in resources following a catastrophic earthquake. What is known about the conditions under which, and the likelihood that, those kinds of shifts in resources will take place? What kinds of problems are caused by those shifts or a lack of them? What types of theoretical and empirical models exist to anticipate whether such shifts will take place?

The presentations in this chapter will raise the issues of substitutability of products and the redundancy of services as they relate to different economic sectors. Although major consideration should be given to the commercial sector, some consideration should also be given to the energy, fuels, and utilities sector. A major disruption in "industrial-strength" power, for example, may have major consequences for the recovery of heavy industry across a distribution system.

Ronald Eguchi is a civil engineer and an associate with Dames and Moore in Los Angeles. Mr. Eguchi has broad research experience in risk analysis, earthquake engineering, and natural-hazards engineering in general. He is a member of various professional societies' technical committees on lifeline systems and performance in earthquakes. His presentation focuses on lifeline systems and the major issues associated with regional shifts of resources following a catastrophic earthquake.

Dr. Tapan Munroe has a doctoral degree in economics from the University of Colorado. Since 1984, Dr. Munroe has been the chief economist for Pacific Gas and Electric Company, headquartered in San Francisco. His presentation will address the economic impacts of the 1989 Loma Prieta earthquake on the Bay Area.

This presentation focuses on the lifeline problem and some of the major issues that are important from the standpoint of the recovery of these systems after earthquakes. Direct and indirect losses have been topics of discussion in previous chapters, but nowhere does the problem of indirect losses become more significant than when lifelines are examined. For example, the direct economic effects associated with lifelines (i.e., direct damage) are small when compared with the effects associated with the disruption of lifeline services. Therefore, focusing on indirect impacts of lifeline failures is very important.

This presentation will focus on the resource problem and more specifically, how intra- and interregional shifts in resources can improve the

recovery of a community after a catastrophic earthquake. Very little attention has been given to this area. Some of the major problems associated with resource shifts will be identified and some positive aspects of resource sharing will be discussed. An example of how one might incorporate seismic vulnerability studies into looking at this problem of postearthquake recovery will then be provided.

It is important to view the current state-of-the-practice for lifeline seismic design in relation to current standards. From the standpoint of seismic vulnerability studies, scientists have come very far and have actually done quite well in assessing the seismic vulnerability of various kinds of lifeline systems. A lot of work has been done on water, natural gas, and electric power systems and, to some extent, communication systems. These kinds of methodologies have been applied in different regions of the country. For example, in California systems, vulnerability studies have been applied very heavily. These kinds of studies are also being performed in Seattle. In addition, earthquake awareness is increasing in the Midwest. These studies are funded by both government research projects and some commercial clients. Several businesses axe performing these kinds of studies to evaluate what their vulnerabilities might be during a catastrophic earthquake.

Planning efforts in the postearthquake response and recovery area have been limited. The primary problem is that the results from these seismic-vulnerability studies are not being integrated directly with response-planning efforts. Once this kind of integration is emphasized, certain shortfalls in inventory and response capability will become evident. Based on these shortfalls, other resources can be examined to handle the problem.

In terms of resources, this is important at three different levels:

1. supply—for example, suppose natural gas or oil supplies are disrupted, or local production facilities impacted; are there other sources that can be used to provide this supply?

2. response—in the event of a disaster, do resources exist that are needed in order to detect damage and finally isolate this damage so that it does not become a problem?

3. recovery procedures—are there repair inventories or the manpower and equipment to make these repairs? Do resources exist to bring damaged lifeline systems back up so that they become functioning systems within the community?

What are some of the major problems associated with intraregional transfer of resources? First, if a very large earthquake occurs it is very likely that similar utilities (i.e., utilities that may be relied on for resources) would also be affected, and this has happened in past earthquakes. In order for this type of lifeline (i.e., underground) to function or to operate, whatever damage has occurred to the distribution system must be repaired. In most modern earthquakes that have been examined, there is some type of distribution

pipeline damage, and this damage must be repaired before the system can be operated.

There is the interutility or interaction problem that we are beginning to recognize as being more important. That is, lifelines by themselves are not independent. Lifeline systems will depend upon other lifeline systems. As in Loma Prieta, many of the water systems experienced minimal damage to the piping systems. However, because there was no electric power to run the water pumps, the operation of these water facilities was essentially curtailed. Thus, the interaction problem is important.

The interaction problem also becomes important when failures in common utility corridors are examined. It is quite possible that damage from one lifeline system may actually impact the operation of others that are adjacent to it.

The demands on local contractors and those who can supply equipment and manpower are great after a major earthquake. That is, there may be a limited number of suppliers who provide this service. The following example suggests that supplies available during normal accidents are not necessarily adequate for earthquake conditions. The neighboring utilities that may also be damaged will require immediate inventories, resources, and manpower. The fact that they require the same services will impact the repair and recovery period for the region.

So these are some of the major problems seen with intraregional transfer: specifically water, gas, and oil systems. Transportation systems have other very special problems, and interregional transfer issues become very important. If there is a major disruption to a transportation system, importing supplies from out of region or out of state will be very difficult. In addition, a lot of the equipment that is needed to make these repairs are not small items or simple things to transport, so there should be concern over transportation disruption.

Major disruption to local lifeline systems may also have an impact on other regions. For example, this may be particularly true for the Midwest when gas and oil systems are considered. A major disruption mused by an earthquake in the Midwest will likely cause shortages of supplies to the Northeast. This question of economic impact, in fact, is being addressed by the National Center for Earthquake Engineering Research in one of the studies that focuses on oil pipelines i.e., what are some of the socioeconomic impacts of a major disruption of oil supply?

Importation of interregional supplies must use existing pipeline systems. It is not as though you can change the routes or incorporate new pipeline components into the system. In order to transport these bulk supplies, they must be transported through the existing lines. If the existing lines are damaged, the supply from these different regions will be hindered.

Finally, this may not be a problem, but in the Loma Prieta earthquake last year, manpower resources were transferred quite effectively. However, if manpower resources are received they must be effectively managed and there must be places to put these people and things for them to do. If these tasks

are not performed, then this may create significant logistical problems that we may not want to deal with.

There are, of course, some very positive aspects associated with the transfer of resources, particularly if a small or moderate earthquake occurs. In the intraregional case, neighboring utilities will provide a very effective resource for equipment, manpower, and possibly supplies. There were examples of this during the San Fernando earthquake, where temporary waterlines were laid out for the city of San Fernando.

Very important agreements, like mutual aid agreements that would be useful in response and recovery efforts, will likely be in effect for utilities that are located very close together. This type of planning exists, at least in California, and these agreements should facilitate any response or recovery efforts during the event. Some of these utilities will have already been involved in exercises or planning efforts related to response and recovery. Therefore, at least a vehicle or mechanism is in place to examine this problem.

Then finally, similar utilities will generally be regulated by the same agency, and to the extent that regulations and standards play a role in terms of design, construction, or repair of these types of systems, this may help facilitate a recovery effort.

There are positive aspects associated with the interregional transfer of resources. The obvious benefit is that there is a much larger supply of inventory, manpower, and equipment to pull from. The second item may not be as important or as significant as the first, but when out-of-town or out-of-state utilities become involved with a disaster, they are not so much involved with the response efforts but are more involved with the recovery aspect, and to the extent that the recovery effort is more easily handled, this may be an advantage.

Focusing on supply during a moderate earthquake, both inter- and intraregional resources may provide a reasonable source. When the magnitude of the earthquake is increased, however, to the extent that larger regions are affected, these resources become less effective.

Interregional resources can be reasonably effective, however, they will be very effective during moderate earthquakes because of their close locations and commonality of the systems. The effectiveness of intraregional resources stays about the same with large earthquakes, but the interregional resources become less effective. From the standpoint of recovery, the same kind of pattern may result. Again, the important point here is that when the earthquake becomes too large, it beans to impact the immediate resources that would be very effective during a small or moderate-size event.

Vulnerability studies have been integrated into the response/recovery planning that is being performed for a water company in Southern California. The types of facilities being considered are water wells. If a very large San Andreas event exists—an event that would break three of the major segments on the fault—it is likely that the major aqueducts coming into the area will be severed. It is estimated that the outage may be as long as 4 to 6 months. The seismic vulnerability of these facilities reveals that they will also be impacted.

The kinds of models that were developed for lifeline studies are very similar to the kinds of models discussed by Don Friedman. A hazard model is needed, that is, some way of representing what the shaking effects might be throughout the region.

It is very important to consider all the effects that might impact the performance of underground pipeline systems. Ground-failure effects (i.e., fault rupture, liquefaction, and landslide) are very important in assessing the performance of underground systems.

There are various ways that water wells, including local water-production facilities, can be made inoperable and in terms of modeling vulnerability, these various failure modes were examined. A number of these failure modes can be combined because of the similar impact that they would have in terms of repair or the failure of the facility. For example, sewage contamination and chemical contamination basically lead to the same impact or result. Sanding, casing damage, and pump motor damage all imply that the casing will be damaged and that something needs to be done about the casing to repair that facility. Wellhead damage and connecting piping damage would be things that would affect aboveground components. Finally, failure of the building enclosure may also be a reason why this facility may not operate after an earthquake.

Other factors, such as power outage, also are very critical. In our analysis, each of these different failure modes is taken and the important parameters are identified. Some of the important parameters included the type of drilling method that was employed to construct a well, for example. In certain cases, in terms of contamination—which is what we are looking at here—the voids on the sides of the casing may allow contaminants to penetrate the soil. Of course, the depth of the well is also important. The important point to focus on is the impact these parameters will have on estimating the probability of contamination at a sate. The loss, if it did occur, would be 50 percent of the replacement cost of the well, and the outage time for repair could be as long as 6 months or greater.

Out of 3,000 wells about 500 of them would be expected to experience moderate to major damage. Out of the 500 approximately 166 would suffer major down-hole damage, 41 would experience minor damage to the casing, and over 400 would suffer damage to the pump motor.

In terms of impact to the region, production shortfalls would be created. When the time required to repair each of these facilities, and the expected loss are considered a time line can be developed to show what the immediate production shortfall would be and how the system or facilities would be restored as a function of time.

Based on a survey conducted, the repair resources are quite limited. Drill rigs—which are important for making these down-hole repairs—only number about 30 to 40 within the basin. Therefore, considering the total number of damaged wells—about 166—only a quarter of them can be repaired at a time. This means that a 6-week time frame may be extended by a factor of 4 when addressing the resource problem.

The same problem with resources also occurs if one looks at minor casing damage, replacement of pumps, etc. This type of information and this kind of analysis can be used to determine whether there are adequate resource capabilities. This analysis also gives a good idea of what to expect in terms of damage or vulnerabilities to lifelines. This type of analysis has been done for pipelines and should be performed for other lifeline facilities that are critical for regional response and recovery.

A lot has been presented about various types of economic impact projections. This section aims to reveal some of the specific effects observed in the San Francisco Bay area following the Loma Prieta earthquake. Some of these facts are important from an economist's point of view.

This report contains a lot of discussion about the differences between national, state, and regional perspectives. California accounts for about 14 percent of the national economy. If a major earthquake occurs in California, this could certainly have a major impact on the U.S. national economy.

The city of San Francisco was well prepared to cope with the immediate effects of the earthquake, as was Pacific Gas and Electric Company (PG&E). As a matter of fact, several months before this earthquake, PG&E went through a simulated exercise involving recovery and restoration of services in San Francisco. That experience was absolutely providential, because the speed at which PG&E could respond to the disaster, particularly in the Santa Cruz area and in the Marina District of San Francisco, was remarkable. Gas pipelines were repaired—that would have usually taken 6 weeks to complete under crisis conditions—in 2 weeks.

Pacific Gas and Electric Company was interested in assessing the impacts of this earthquake from several different perspectives: regional, intraregional, and specific industries or economic sectors. Damage estimates prepared by analysts immediately following the earthquake were problematic, with losses ranging anywhere from $5 billion to $7 billion. The initial estimates were much larger, about $10 billion. ''New'' damages are still being discovered from the earthquake in what has become a long, drawn-out process. Our initial impression was that the loss and damage estimates were on the high side.

If one looks at the spatial distribution of the damage, it was fairly concentrated on the Marina District in San Francisco, downtown Oakland, the Bay Bridge, the Cypress section of the I-880 freeway, the Embarcadero Freeway, and Santa Cruz. The Cypress structure does not exist any longer. The off-ramps of the San Francisco Bay Bridge, the Main Street exit, and several other exits are still not repaired. It is very likely that the Embarcadero Freeway is not going to be there much longer. Most likely, there will be a sunken freeway in its place. Some of the infrastructural effects are long lasting and very likely permanent.

Significant structural recovery has occurred in a short time in the Marina District, but the effect lingers on in its economy. In Santa Cruz, there are major impacts in terms of infrastructure, buildings, and the economy. On the whole, the damage and the distribution of damage was fairly focused and concentrated in San Francisco, East Bay, and the Santa Cruz area.

In assessing economic impact in the 7-county Bay Area—using employment or income data—we find that the impact has been minimal and barely observable in the indicators (Table 5-1). Within 6 to 9 months after the Loma Prieta earthquake, economic effects appear to be nearly nonexistent. The data in Figure 5-1 is a composite index number involving six different variables, including real estate, the services sector, the retail sales sector, tourism, trade, and manufacturing, and it remained remarkably fiat in the entire period. The tourism sector shows a noticeable effect. After a decline, followed by recovery,

the indicator has leveled off. The service sector is the most stable part of the index number. The real estate component was on a decline between June and December. Of course, one has to keep in mind what was happening in the Bay Area in terms of housing affordability in this period. Only 10 percent of households in San Francisco could afford a median-priced home in the region, which partially accounts for the decline in home prices and slowdown in the housing sector. The earthquake was very likely the other factor underlying the slowdown in the housing sector.

The only sectors where some impact from the earthquake is seen are in retail sales and tourism; surprisingly, there was very little impact on manufacturing. Few production facilities were disrupted for more than 3 to 5 days. There were two or three production facilities companies in East Bay where production stopped completely, but these companies found other facilities within a matter of weeks. The manufacturing sector of the economy recovered shortly after the earthquake.

With respect to joblessness, four Bay Area counties (Alameda, San Francisco, San Benito, Santa Cruz) and two counties outside the Bay Area had most of the unemployment claims resulting from the earthquake. Comparison of the same 5-to 6-week period from 1988 to 1989 reveals that this was a short-lived phenomenon. Examination of monthly unemployment figures between October and December 1989 in three Bay area counties (Marin, San Francisco, and San Mateo) indicates that the rates are surprisingly low and about the same for all counties: October = 2.9 percent, November = 2.6 percent, and December = 2.8 percent.

Unemployment levels in the San Francisco Bay area are usually among the lowest in the country. We did not see a rise in unemployment resulting from the earthquake. If anything, there was a dip in the unemployment level in November, immediately following the earthquake. This was true throughout much of the region, even in Santa Cruz County, which sustained significant physical damage.

There has been a great deal of discussion about the impact of the earthquake on the hotel/motel occupancy rates. Table 5-2 shows some impact from the event. San Francisco showed a 68 percent occupancy rate for rooms over the $110 price range for the first 6 months of 1989, through June. However, for the same period in 1990, the occupancy rate was 63 percent. If June 1990 is compared with June 1989, significant recovery in occupancy rates can be seen in the aftermath of the earthquake.

Therefore, there are basically two types of economic impacts. One is induced by a breakdown in the infrastructure. For example, the Bay Bridge was probably the single biggest factor that created temporary job and income losses. The second one is related to the economics of fear, which, for example, played a significant part in the decline in hotel/motel occupancy rates in the Bay Area. Tourism figures in the aggregate for the entire region indicate that there is no effect; but in San Francisco, there has been some decline resulting from the earthquake. Fear and anticipation of a recurrence of an earthquake has resulted in a decline of the tourism industry in the city.

For example, eight major conventions were canceled in November 1989. Some have been rescheduled, but some have not. Also, fewer U.S. tourists are visiting the Bay Area; but there does not seem to be an impact in terms of tourism from the Far East or Europe. There is even an increase in tourism from Japan, perhaps because of good bargains and relative familiarity with living in seismic regions.

Much smaller regional areas must be examined in order to really assess economic impacts. For example, there appear to have been major economic consequences in Santa Cruz, the Marina District, and the Oakland area.

In the aftermath of the Loma Prieta earthquake, enormous traffic congestion occurred as the Bay Bridge went out of commission. Ridership on BART (Bay Area Rapid Transit) increased tremendously; but 6 months after the disaster, BART appears to have gained only about 2,000 permanent passengers. Immediately following the earthquake, BART ridership increased from 60,000 to 70,000 passengers per day.

A great deal of discretionary travel from East Bay to San Francisco appears to have stopped; people are staying home in East Bay. They are not crossing the Bay Bridge. The bridge seems to have become a psychological barrier.

The earthquake resulted in a significant decline in the supply of low-cost housing for the poor and for members of various ethnic groups because of the damaging structural impacts in the Oakland area, the area south of Mission Street in San Francisco, and the Tenderloin District (a region with a large Indochinese refugee population). The issue of housing for these impacted populations has become an important social and political concern in the Bay Area.

An interesting phenomenon may now be occurring with respect to real estate and land values. Although there has been a general decline in real

estate values in the San Francisco area, the prices of land and buildings on bedrock have risen, while the value of areas involving higher seismic risk has declined in general.

On the whole, it is somewhat frustrating to try to assess the economic impact of an earthquake, particularly if large regional areas are examined. The various effects wash out. Part of this washout is explained by relief and rehabilitation money that has flowed into the region for repair and reconstruction. This inflow offsets some job losses and income losses.

With regard to the flow of funds into the Bay Area, only about half of what was initially expected has been received. This is a matter of much concern in the region. Often, initially there is a high level of optimism in the region about this "windfall" inflow of funds.

The relief money flowing into a region is not a benefit in a larger sense. Somebody is paying for it. The cost of this aid is being passed on to the next generation or to another region. People who live in safe areas subsidize those who live in dangerous areas. This is a matter of interregional transfer. But from a regional point of view, net transfers are good in that they minimize regional losses.

There are some important public policy issues that must be mentioned. One is that there has been a tremendous awareness of regional linkages in regional policy initiatives as a result of this earthquake. For example, we became very aware of transportation vulnerability issues in the Bay Area, particularly because of the damage to the Bay Bridge. This may have been a preview of transportation problems that the Bay area could experience, perhaps in 1998. But, as the memory of the disaster fades, some of these lessons and experiences are already being forgotten. The window of opportunity for instituting traffic-congestion-mitigation policies may already be closed.

In summary, if we look at the nine-county Bay Area, the economic impact has been minimal. If we look at the state as the basis for our analysis of economic impact, there is no longer even a blip. But there are certainly major economic impacts within the region, especially at the subregional and neighborhood levels. More attention needs to be paid to some of these issues also because good public policy must look at distributional effects, not just economic aggregate effects.

QUESTION: It was interesting that you did not mention that there were few insurance dollars that flowed into the San Francisco area after Loma Prieta. There was quite a bit of money during the course of the recovery.

DR. MUNROE: That's true. It certainly helped the recovery, no question about it. The inflow of funds has created incomes and jobs in the region.

QUESTION: It appears that after the Loma Prieta earthquake, PG&E coped very well with repairing damage and restoring services. They did this by

calling upon all of their personnel and also by using mutual aid from all around the western region from other utility companies. Now, this was a moderate event. What would be the special challenges that would be presented by a catastrophic event? What is the difference between the kinds of things that we have been looking at in terms of earthquake events and the major catastrophic events?

MR. EGUCHI: As I tried to indicate, it depends very much on the size of the event. In California we have a very special situation where there really are two major gas suppliers or distribution companies—PG&E and Southern California Gas. The sort of scenarios in California are really kind of independent events, so a lot of mutual aid resources and so forth can be expected to be to be there. Now, for situations like water companies, it is quite different. They are much smaller and, in terms of maintenance and supplies, it is much more limited. There would be a tremendous impact on those kinds of facilities.

DR. MUNROE: Let me add an example. PG&E had to shut off gas supplies to about 50,000 homes. Within about 2 or 3 days, we had to relight most of the pilot lights. We were able to do this with nearly 1,100 people from PG&E's and 350 people from Socal Gas from southern California. The latter went on our payroll for several days. If we are talking about a more significant earthquake, we could certainly obtain greater assistance from Socal Gas and elsewhere; we could also get assistance from Nevada and Oregon. The regional cooperation was absolutely first rate, and the inter-regional cooperation between utilities was impressive. We, as a society, do so well in times of crisis, it is just incredible.

QUESTION: But as Mr. Eguchi also pointed out in his presentation, when more and more are invited people in, that many more human resources must be managed. Where you are going to put these people; how you are going to house them, and so on?

MR. EGUCHI: In some of the smaller areas down south, there was a real problem, because there were no places to house the crews, and feeding them was very difficult. They had to go in and out of the region each day. Personnel management can be a significant problem.

DR. MUNROE: PG&E had a drill for a major earthquake, actually much more than a 7.1 earthquake; and that drill was extremely useful. I think preparedness and employee dedication was a key to our successful response, and I think it will have to be key to anything in the future.

QUESTION: Commenting on the previous question. I do not think you can extrapolate linearly. First of all, this was a remote earthquake. This was not an essential part of the Bay Area. The [PG&E] network was not damaged, its major facilities were not damaged. The distribution system in some localized places was lost because of the geology. A larger event, at the core of the Bay Area, would be qualitatively significantly different, not just taking more time to make repairs but actually rebuilding some facilities.

DR. MUNROE: Another factor was favorable for a fast recovery. For example, the Chevron refinery, one of the largest in the West, had no damage.

The water system was intact. Most of the utilities, including telecommunications, were really in very good shape. I guess this is what occurs in a moderate earthquake. The real impact on PG&E from the impact of the event was $100 million, and most of that was due to structural damage at two fossil-fired power plants and two or three major substations.

MR. EGUCHI: I think we had a very special situation during that event with regard to lifelines. The unique feature was that the epicenter occurred very much south of, say, the San Francisco Bay area, where you have a lot of urbanized systems and so forth. If you look at the areas where you had damage and where you did not have damage, there was a very strong distinction between those areas that were located on fall or were affected by some type of ground-failure amplification, which was actually very limited. If a large earthquake occurred closer to that area, the shaking effects would cause significant damage to facilities. Damage may not be so limited as what we saw in this event. I think we were very fortunate in that respect.

DR. MUNROE: Yes, there were many fortunate circumstances. The fact that a game of the World Series was being played that day may have saved many lives, because people left work early, before the earthquake occurred. Only 60 people died on the Cypress Overpass, and not 200 as initially estimated. The reason only 60 people were there instead of 260 is because all of those people were watching television at home and were not on that freeway.