The "Ripple Effect"
A great deal of concern has been raised about the economic consequences of a catastrophic earthquake extending beyond the region of impact to disrupt state, regional, or even the national economy. Of particular relevance is the concern raised about the ability of the insurance industry to cover losses from a catastrophic earthquake without having to sell off investments—stocks, bonds, real estate—an action that might have negative financial implications for regions outside of the disaster area. How farreaching could the economic impacts extend from such an event? What conditions are likely to impede or exacerbate this effect? What research exists to support the possibility of such an effect?
The two presentations in this session will explore the methodological and theoretical bases for projecting such an effect. Barbara D. Stewart has undergraduate degrees in economics and business administration from Beaver College. Ms. Stewart has worked for Stewart Economics, a consulting firm specializing in insurance issues. Her presentation will address the national economic impact of a catastrophic earthquake.
The second presentation is made by Professor Leonard K. Cheng, an associate professor of economics at the University of Florida. Dr. Cheng received a Ph.D. in economics from the University of California at Berkeley, where he specialized in international trade and trade policies under conditions of uncertainty. Dr. Cheng will assess the theoretical and empirical evidence related to economic "ripple effects."
PRESENTATION OF BARBARA D. STEWART
A catastrophic earthquake will have a national impact, and there will be national damage. The United States has a highly developed, specialized, interdependent, money economy. While those features make our economy productive and resilient, they also mean that an earthquake of the magnitude that we are contemplating in this forum will not be just a regional event.
There are three ways to view national economic damage: (1) disruptions to supply lines, (2) shocks to financial markets, and (3) drain on the insurance system.
There has been very little study of these consequences for obvious, very understandable reasons. It is quite human to focus on the human suffering and the physical damage that will occur immediately after an earthquake. The problem is that it is unknown, other than estimates of the physical damage, just how bad the general economic damage might be, and that uncertainty is a problem in itself.
The disruption in the earthquake area could easily break critical supply lines in the economy. It is very well known that gas and oil pipelines run through the New Madrid region and supply many businesses and individuals throughout the Northeast. We are also very critically aware that the important semiconductor industry is concentrated in California, and a catastrophic earthquake there would affect a wide array of other businesses, because suppliers there would shut down.
In general, shutting down most activity in the earthquake area will spread beyond that area as customers or suppliers are hit by the shutdown, and many businesses that are far away will suffer. There is a good chance that many of those, which might have been marginal to begin with, will just never start up again.
The inability to supply, the inability to sell, and the multiplier effects that will spread from the area are what many analysts are calling the ripple effect. Now, ripple connotes a less and less noticeable effect as the earthquake is spread over time and over space. But it is important for us to keep in mind that there are going to be more than just ripple effects, and they are not going to be orderly, spread over time and space. There are going to be some immediate and large impacts on the national economy, and those are going to come through the financial markets and through the insurance system.
The U.S. financial system is vulnerable to the physical damage of an earthquake. Just consider banks. Major banks in cities across the country are important switches in a complex financial network. They serve not only their regional economies, but are part of a national payments system. The money-center banks in particular transfer billions of dollars every day by wire around the country and around the world. One of the large California banks has estimated that if its central data processing were inoperable for 3 days, it would affect the entire state. If it were for 5 days, it would disrupt the whole U.S. economy, and if it were for 7 days, the world would feel it.
Having lived with the earthquake threat for a long time, California banks have some very sophisticated emergency planning systems, but whether those emergency plans contemplate everything in the area being shut down is something else. Another question is about banks outside of California that have not lived with this threat. Are they prepared for more than isolated events like fire or terrorist attacks?
Physical damage would also bring loan defaults. There have already been experiences with mortgage defaults after the San Fernando earthquake. Many people walked away from their homes if they had very little equity in them. The same thing is happening right now in the Northeast as property values fall and a lot of people walk away from their mortgages. It would happen again.
Commercial loan defaults would be even more serious, simply because the damage and the effects on other people would make it impossible for many businesses just to service their debt. The national question here is whether this country can deal with another round of bank failures, or call it another round of rescues.
How would our securities markets respond? I would not presume to predict what stock and bond markets would do. That is a folly for those who want to do it, but let us just think about it for a minute. Everything else being equal, if the assets that are underlying the securities that are being traded have been damaged, surely the prices of those assets would want to fall. But the greatest threat to security markets is not so much the damage but the uncertainty coming out of the damage. Financial markets work because of continuity and because of confidence. As we heard yesterday, our whole financial system is based on nonearthquakes. The United States economy does not have any experience, or in any other developed economy, with a catastrophic earthquake in a major business center. We just do not know what it means.
There are big questions of recovery. When? How? Where are the funds coming from? What else is going to be affected? What are the third-, fourth-, and fifth-order effects of this sort of thing? These are tremendous uncertainties, and if there is anything financial markets cannot stand, it is uncertainty. They can deal with good news, they can deal with bad news, but uncertainty is the worst. They are not going to be helped, either, by the fact that at the same time there are going to be tremendous demands for funds from the earthquake area in order to rebuild, and at the same time, the insurance industry is going to be dumping stocks and bonds on the market in order to raise cash to pay for claims.
The congestion in financial markets could have two effects. First, it would cause many businesses and governments to postpone financing for probably critical projects, for how long we just do not know. And second, it could have a liquidity effect on others who needed to sell securities to raise cash just for ongoing business needs.
When measuring systems are put together, specific things must be kept in mind. What is the cost of the things that are not done, the projects that are not built, the activities that are not undertaken?
After a catastrophic earthquake, what the insurance system would have to do to raise the money to pay claims and keep sound books of account would send another kind of aftershock throughout the country. It would happen in three ways in an increasing order of seriousness. First, as mentioned earlier, insurers would have to sell billions of dollars of securities to raise cash to pay claims. They just could not borrow. The needs would be too great and the debt service would be too much relative to the capital and earnings they have. These massive sales would depress prices in markets that already would be in turmoil because of the uncertainties over the earthquake.
Most affected would be the municipal bond market. Property and casualty insurance companies hold 20 percent of the municipal bonds outstanding in this country. In some years, they are enormous purchasers. They may take as much as 100 percent of the new supply coming on the market; in other years, they may not take much at all. But they are big players in the municipal market. Municipalities that wanted to raise funds, even if
they were far away from the earthquake, may not be able to get into the market.
The second way the insurance system would disrupt the national economy would be from the magnitude of claims. Although the industry has more capital than the anticipated claims that we would get from a catastrophic earthquakes, that capital is not evenly spread throughout the industry, so we can expect a number of insolvencies to come out of the earthquake. The states have set up funds for dealing with the unpaid claims of insolvent insurers, but they typically do not pay in full, or else they pay over a very long period of time. And insolvent insurers would not just be those involved with earthquake claims. They would include some large, diversified national companies, so not just earthquake claims would not get paid. All over the country other kinds of claims (e.g., automobile, workers' compensation, and fire) would not get paid.
The third and most serious way that the insurance system would affect the national economy is what would happen after the insurance system had finished paying the claims. This is the real question. Earthquake claims would wipe out at least half of the industry's capital. Capital is the industry's ability to absorb and take on risk. Therefore there would be less capital, which would lead to shortages and higher prices of all kinds of insurance, regardless of location.
Why is this so serious? Well, insurance, like banking, is so interwoven in our everyday transactions that it is difficult to imagine an economy without it. We tend not to think about it, but it is very much integrated in everything we do. Insurance is an essential facilitating mechanism. By shifting and spreading risk, it lets an individual, business, or government pursue activities without fear of jeopardizing the household, the enterprise, or the institution.
How important it is for others that risk be shifted and spread is reflected in the fact that three-quarters of the property and casualty insurance sold in the United States is required by someone else. State statutes require automobile and workers compensation coverage. Commercial and residential mortgage lenders require borrowers to have property damage and sometimes liability insurance.
Without insurance, many entities would have to bear their own risks or else curtail their activities. The economic and social costs, as well as the uncertainties, would be enormous. What would we do to replace insurance? Would we change our law? Would recovery in torts be permitted? Would we import it? There are tremendous uncertainties.
In conclusion, a catastrophic earthquake will do more than ripple through the economy. There is going to be an immediate impact on financial markets and the insurance system, both of which are national and international in scope. This is more than a measurement problem. It is also a conceptual problem. There are no models to go on. The United States today is not San Francisco in 1906. Nor are our major business centers Armenia, Iran, or Managua. We are dealing with a completely different order of magnitude, quantitatively and qualitatively.
There is a reluctance of participants in our financial markets to even discuss their vulnerability. Why should they undermine confidence if they do not have to? These uncertainties are useful to keep in mind when discussing the earthquake project and what to do with this kind of hazard, particularly when we get into issues of mitigation, because just as physical damage is not the whole story of catastrophic earthquake, mitigating physical damage is not the whole story either. It is going to be just as important to mitigate these uncertainties, that can have just as large an economic impact as the physical damage will.
PRESENTATION OF LEONARD K. CHENG
This presentation offers a few theoretical considerations which may be useful in assessing the economic ripple effects of a major earthquake (i.e., economic effects on areas other than the impacted region). In addition, some related empirical evidence will be provided to get an idea about the order of magnitude of these effects.
Since all regions of the United States are part of an integrated national economy, undoubtedly the economic effects of a major earthquake (taken to be in the order of magnitude 8 to 8.5 on the Richter scale) will not be limited to the impacted region alone. Hence, there will be ripple effects. The real question, however, is about their scope, intensity, and duration.
To see how the rest of the country will be affected by a major earthquake in the long-run, imagine an artificial economy which is completely diversified in the sense that every household in the country (say, n of them) owns 1/n of everything. In this fictitious world, if an earthquake hits, every household will be affected in the same way independent of the position of the epicenter. Suppose 5 percent of the country's total assets are destroyed; then every household will lose 5 percent of its wealth. If the loss in assets and people is permanent, then there will also be an additional loss in real income due to the reduced scope of specialization. If the earthquake causes a total economic loss of $50 billion, then this loss will be borne by all of the households in the country, which has a GNP of about $5 trillion per year and total wealth well in excess of $15 trillion. The impact per household outside of the impacted region is very small.
Obviously, the pattern of asset ownership is far from completely diversified. Assets in a particular region tend to be owned mostly by households residing in the same region. This implies that most of the long-run losses will be borne by the impacted region, and the losses to the rest of the country, perhaps in the order of billions of dollars, will be even smaller. To many they will be negligible. An important theoretical conclusion is that if the ripple effects are transmitted to a larger area through economic linkages, then the average effect per household outside of the impacted region will be smaller.
In the short run (anywhere from weeks to months, or even a few years), however, the effects outside of the impacted area are likely to be larger than the long-run effects, since adjusting to a changed environment takes time and can be costly. The severity of the short-run ripple effects depends on the degree of substitutability in the economy. In one extreme, if (1) the products produced in the impacted region are vital to production or consumption in the rest of the country, and (2) substitutes are not available or producible outside of the impacted region (not even in foreign countries), then the short-run ripple effects would be disastrous. In the other extreme, if there is a great deal of substitutability in both production and consumption, then the short-run effects would be approximately equal to the long-run effects.
As an example, if a certain computer chip produced in the Silicon Valley became unavailable after a major earthquake hit California, it can be substituted with identical or similar chips produced in other parts of the United States or in foreign countries. Even if substitutable chips are not available (which is extremely unlikely), so that products, incorporating the chip cannot be produced, consumers may still substitute these products with other products because the same need can be met in different ways. Finally, consumers can also substitute current consumption with consumption in the future, when the chip becomes available again.
An advanced economy like the United States is not only highly integrated (which may give rise to enormous adjustment costs under extremely adverse conditions), but also exhibits a great deal of flexibility and substitutability. As a result of this second characteristic, it is unlikely that the short-run ripple effects would be much higher than the long-run effects, which would be very small in relative terms. For instance, there is no evidence that the San Francisco earthquake on October 17, 1989, has had much of an economic impact outside of the Bay Area.
Without the economic linkages, the different regions of the country would be like isolated island economies. A region which is hit by an earthquake, large or small, will be the only party to bear the entire burden. In contrast, an advanced and highly integrated economy allows risks to be shared through trade and credit relations as well as ownership diversification. It is analogous to numerous small nets knitted into a big net. When an object falls on any small net, its impact will be partly borne by other nets. To ignore the benefits provided by the economic linkages, including a greater capacity to deal with disasters, and to view them primarily as the nodes of a network by which diseases are spread would be theoretically wrong and very misleading.
On the one hand, the loss of capital and equipment due to an earthquake implies the loss of income in the impacted region derived from the original economic activities. On the other hand, the need for reconstruction would increase employment. Provided that there are enough past savings, credit, insurance payments, and investment by outside investors, it is likely that the level of economic activities in the impacted region could quickly exceed that before the earthquake. It is also very likely that producers outside the impacted region would benefit from the reconstruction effort in the region. Of
course, any surge in economic activities for the purpose of replacement and repair will taper off once recovery is more or less completed. These predictions about the economic effects of a major earthquake on the impacted area are consistent with simulation results.49
The economic data presented by Dr. Tapan Munroe in connection with the recent San Francisco earthquake have revealed the remarkable resiliency of the impacted region's economy. The data suggest that, at the county level, the quake has had little economic effect. To find drastic, localized effects which are marred by aggregates, one has to look at data for cities or even blocks within cities.
Full and just insurance payments to the insured victims of earthquakes will speed up recovery in the impacted region. An effort by the insurance industry to do its best to settle claims in full and speedily not only is not a disruptive force to the national economy, it would lead to good business for the insurance industry in the future.
If the effect on future demand for insurance, is ignored then any shortfall in insurance payment for validated claims amounts to a transfer of wealth from the insured victims to the shareholders of the insurance companies. If the insured victims are able to undertake the same needed repair and replacement without the full payment from the insurance industry, and if their marginal propensity to consume out of wealth is identical to that of the insurance companies' shareholders, then after an earthquake hits, how much the insurance industry ends up paying will only affect the wealth distribution of these two groups. However, since how much and how fast the insured victims can replace and repair their damaged properties are likely to depend on the payments they receive from the insurance companies, and provided that the earthquake victims' marginal propensity to consume should be no less than that of the insurance industry's shareholders, we can conclude that full (and justified) payments to the insured victims will lead to a faster recovery in the impacted region and greater output and income for the country as a whole.
The effect on the financial markets of the insurance industry's need to liquidate some of their bonds to settle claims (say $30 billion) would be minimal because of the size of the securities market. The outstanding Treasury bills and bonds alone are in the order of trillions of dollars, and recently the federal government has been borrowing hundreds of billions of dollars each year. For similar reasons, a major earthquake's destruction will have a very small effect on the stock market. For example, even though the San Francisco earthquake last October inflicted about $8 billion of damages to man-made facilities (which includes $2 billion of damages to the transportation system), it did not have much of an adverse effect on the stock market. Indeed, the stock market has since been achieving record highs.
To get an idea about how the insurance industry may be affected by a major earthquake, we can look at the performance of the industry in 1989, when it was hit by several catastrophes. According to the insurance industry's trade journal, Best Review,50 insured losses caused by (1) Hurricane Hugo, (2) the San Francisco earthquake, and (3) storms, floods, tornadoes, and the
Phillips Petroleum plant explosion were $4.1 billion, $1.1 billion, and $1.8 billion, respectively. Despite these major catastrophes, insurance stocks performed favorably, especially in the area of property/casualty. Stock values surged to their highest record levels. The overall A.M. Best Insurance Industry Stock Index went up by 37 percent in 1989, compared with a 25 percent growth of both the Dow-Jones Industrial Average and S&P 500.50 Moreover, the ability of the insurance industry to provide service to the economy as a whole is not limited by their current capital and surplus. Since capital is mobile across industries and will move to whichever industries are profitable, the capital base of the insurance industry is not a constant equal to their current capital less insured losses. Instead, in the long-run it will be determined by demand for insurance, including insurance against earthquakes.
Reinsurance is a device for spreading risks among the insurance companies, including foreign companies. Like all economic linkages discussed above, it spreads the burden of shocks to a large number of participants, and therefore is definitely not a cause of disruption to the industry. Furthermore, participation by foreign companies in insurance and reinsurance has significantly helped the U.S. insurance industry deal with the insured losses.51
On the basis of the theoretical considerations and available evidence from the San Francisco earthquake last October, it can be concluded that the economic ripple effects of a major earthquake would likely be small. Moreover, due to the scope of substitution in production, consumption, and investment, any significant ripple effect will be dampened quite quickly. Of course, rigorous research is needed to obtain precise estimates of the magnitude and duration of the ripple effects under different scenarios.
One useful approach is simulation, with a carefully constructed model of a regional economy embedded in a national economy which incorporates real-world economic linkages between them. For example, an extension of the regional econometric model constructed by Ellson, Milliman, and Roberts, which incorporated supply-side constraints and spatial disaggregation, would be a fruitful avenue. Since the linkages between the impacted region and the rest of the country need to be modeled in detail (such as spatial disaggregation), the level of disaggregation within the impacted region can be reduced accordingly to retain tractability.
A report prepared by Japan's Tokai Bank attempted to estimate quantitatively the economic ripple effects of a major earthquake in the Tokyo area on the rest of the world. Unfortunately, the estimates were derived from highly questionable assumptions. For example, it assumed that the earthquake in Japan would raise the U.S. Treasury bill rate by 5 percent. Given that Japan's total lending to the United States is only a trivial portion of the U.S. capital market, it is implausible for the real interest rate to rise by such a great magnitude. Since the world capital market is highly integrated, an appropriate comparison is the total reduction in lending by Japan due to the earthquake against the size of the world capital market. From this perspective, the effect on the interest rate in the United States must be quite small.52 My colleague, Dr. David Denslow, pointed out that, with the beginning of
World War I in July 1914, long-term corporate bond yields in the United States increased from the prewar (June 1914) rate of 4.06 to 4.22 percent in December of the same year. By January 1916 they were back down to 4.06 percent.
The methodologies for estimating the economic ripple effects of a major earthquake in the United States exist, but care must be exercised in constructing an appropriate model which can be used to generate meaningful and reasonable results.
Acknowledgment is given to Dr. Jerry Milliman for helpful guidance and support.
GENERAL DISCUSSION OF CHAPTER 6
QUESTION: Dr. Cheng, you said that there were no effects on the economy from many disasters, or the economic effects were minimal. What level of a disaster, what level of a catastrophe, would cause effects on the insurance industry? Loma Prieta did not. What is the catastrophic disaster that would trigger some effects?
DR. CHENG: Of course, if the quake gets bigger, it is going to have more of an effect. That is obvious. But how big is the big one that we are going to prepare for? Are we talking about magnitude 10 or 9.5 on the Richter scale? And I am sure that if the quake is big enough, available surpluses could be wiped out and, therefore, in the short run there would be a serious effect. But still I do not believe that is what we are talking about here. We are talking about the effect which is great but not of magnitude 9.5. In the Blue Book published by the Earthquake Project, they estimated a $50 billion loss.53 Now, of course, if it is $5 trillion, then I am sure it is going to have a terrible effect. But is that likely to be from an earthquake?
I think we had better talk to the loss-estimation experts. How likely we are going to suffer a $500 billion loss from one single earthquake?
QUESTION: I thought I heard one of you say that when a great earthquake happens, it is going to be really big from an economic impact point of view, and another one of you say it is going to be really small. What is it really going to be?
DR. CHENG: The answer lies in the following. The key to the different predictions lies in the degree of substitutability. Do you believe that this economy has no scope for substitution? If that is what you believe, and if you feel that in the real world, everything produced in every corner of the world is vital to our lives and without it we die, then there is no way you can substitute. If that is the case, then you get the first scenario—that everything will be terrible. I do not believe that is the world we live in.
In the world we live in, we have a great deal of substitutability and a great deal of flexibility. Therefore, I tend to believe on a theoretical basis and from empirical observation that the effect will be very small outside the impacted region.
MS. STEWART: I certainly agree that we have a very resilient economy, and it is possible to substitute. It is important to keep in mind the short-run dislocations and the long-run readjustments. In time, definitely we will adjust. We will find substitutes, we will arrange new sources of supply. Our financial markets will recover. Both Germany and Japan rose out of the ashes of World War II.
The short term is what counts. It is important to look at the short-term dislocations, where we disagree as to the magnitude of some of the problems.
Part of this disagreement may be related to this one issue that is really not emphasized much in the Blue Book—that is, the uncertainty that is going to be facing many parts of our economy. This is something that is very difficult to deal with on a theoretical basis. We can model all we want to and crunch the numbers all we want to, but this is going to be a tough thing to deal with because of the lack of prior experience. It is something we should advisedly consider and not ignore as we get on with our abstract models and our measuring and other devices to deal with trying to judge the magnitude of this thing.
DR. CHENG: I would like to add one point. I am not denying at all that there will be serious economic consequences. What I am saying is that most of that will be in the impacted region. If you are going to look for devices to address the problem, I think that is where you look, not from the national economy, given the usual sizes of the effect of an earthquake that we are talking about.
QUESTION: The All Industry Research Advisory Council was interested in the ripple effect on the insurance, reinsurance industry for a large, catastrophic event. We chose to look at the simulated effect of a $7 billion insured hurricane on the Gulf or East coast, and then a combination of two of those events in the same year. We simulated the effect, looking at the various possibilities where you have clusters of properties along the Gulf Coast—Houston, Miami—and then made estimates from each insurance company where their losses were reinsured. We tried to follow those dollars—the ripple effect—to Europe and back to the United States.
The insurance industry can handle a $7 billion loss without much problem. When you get up into $14 billion, you get into a marginal area. If you get much above $14 billion, there is a large internal ripple effect, which is much larger than was implied in your model. Could you comment on that?
DR. CHENG: I think you know better than I do about the internal workings of the insurance industry. But if you are talking about the need to raise even $30 billion from the security markets, that is nothing compared with the overall size of the security markets. Let us not lose perspective on that. We are talking about a huge capital market out there. I am not denying that the insurance industry would have more difficulty when it has to make more payments. That is to be expected. But I believe that is an industry-speciffic effect. Going back to the ripple effect, I think it is negligible.
QUESTION: I would like to refer to an earlier point regarding the difficulty of assessing economic impacts, even on a nine-county basis, even on
a single-county basis. The resiliency, the substitutability, and the flexibility of the economy of the San Francisco Bay Area was really, really impressive. It is not just silicon chips we are talking about. We are talking about petroleum refining, smokestack industries, about a very diversified economy.
I know of five plants that were out of commission—four of them got back into production in 3 days. One of them was totally ruined, but they found a new facility within 2 weeks and got back on their feet. This is something we have to remember. The evidence that we have so far on earthquakes in this country would suggest that the damages, or rather the losses, are certainly on the lower end of the spectrum rather than on the higher end. So far, the conclusions that have been made about overwhelming losses from catastrophic earthquakes suggest that these losses may not be all that great.
Now, we always go back to 1906; but the U.S. economy is not on a 1906 basis in terms of infrastructure, in terms of institutions, in terms of technology. It is absolutely incorrect to extrapolate damages and losses on the basis of 1906, 80 years hence. Since it is unrealistic and incorrect, I would suggest that we try to run simulations on other data. For example, in a study done by Jerry Milliman in 1983, a county-by-county model of the California economy was used. A catastrophic earthquake did not result in catastrophic losses for the state. We need to be very careful in making basic assumptions, because that is what makes all the difference.
I am really concerned about the highly pessimistic nature of the base case that I have heard.
QUESTION: There has been a prediction, I am sure you know, that the first of December we are going to have a major event in the New Madrid area, so probably we may learn whether we are right in our estimates or not. The insurance industry came up with an estimate of insured losses—something in the area of $50 billion. Those are the insured losses, but insured losses are only a small fraction of a total loss to the economy. Public properties, for example, are not insured. If you add those, a $300 billion top estimate is necessarily wrong.
The second point I would like to make is that being in the insurance industry, we have to provide for, if not the worst possible case, fairly near it. If anything can happen, it will. I want to look at the liquidation of securities. For technical reasons, the property and casualty insurance industry is very largely invested in municipal bonds, because they receive favorable tax treatment. In any event, the insurance industry is so invested and it is not a liquid market. We are large purchasers of municipal bonds, and we know that the market could not handle an order to liquidate $30 billion of municipal bonds. Because this could not be handled, we would be obliged to liquidate stocks. Even though the stock market is a very wide market, I think that the specialized effect of a demand by the industry to raise $30 billion would have a very material effect on the security markets. That is a short-term effect, obviously.
In the long-term, the market is huge. It will absorb it, but if we wanted to liquidate it overnight, relatively speaking we would have a problem.
DR. CHENG: First of all, the way to handle the prospect of huge losses is to accumulate surpluses. I know very little about the insurance industry, but I think what the insurance industry can perhaps do is to ask the government to change the law so that it can accumulate surpluses before tax.
My second comment is related to your suggestion about the liquidity of the municipal bond market. All bonds, to different degrees, are substitutable. If the yield on the municipal bonds is very high because of a ''fire sale,'' then a lot of people will sell other bonds and buy municipal bonds. All bonds, while not perfect substitutes, are close substitutes in the eyes of the investors. I am not trying to deny that there will be difficulties in the short run. But I believe that the ripple effects will be minimal.
QUESTION: We are not allowed to accumulate any reserve taxfree, whatsoever, for an event that has not yet happened. If we go a year without an earthquake, the IRS considers that we have made a profit on whatever premiums we have collected during that year, and taxes it as such. It is not very easy to build a reserve for future events. Now, of course, we do build surpluses to the extent we are allowed to by various insurance commissioners around the country. But no doubt that a carefully constructed provision allowing insurers to accumulate reserves specifically earmarked for a major event, not necessarily only an earthquake, would be a very helpful thing. For example, in Japan, we are allowed to accumulate such a reserve. It gives you a sense of comfort.
However, the industry's chances of getting such a change through the Congress at this time are very limited. We would welcome anybody to speak up very loudly and clearly outside of our own industry that this is a desirable thing to be done, and it could be, provided that be totally invested in United States securities.
DR. CHENG: The current rules applied by the IRS only work if the annual expenses and payments reflect the expected payments for events that recur on an annual basis. But when you are talking about something that might not occur for 100 years, the actual payment in one year would bear little relationship to what you expect to pay in the future.
QUESTION: When we do have this major 8.3 event, it will likely take place within the context of increased overall seismic activity in that area. There will be a number of earthquakes, not just one. An event of that magnitude will probably be followed by some very large aftershocks in the ensuing months, and maybe even years. For example, the New Madrid earthquake was not really a single earthquake but a series of events, including three very large ones. Would this likely circumstance change your predictions or your thinking about the economic impact? How do you think the markets will react to this?
DR. CHENG: Well, if the aftershocks come immediately after the big quake, then the reactions would be as though they all occurred at the same time. If the aftershocks come only after a long period of time, that will give room for the market and people to adjust and revise their expectations, therefore revising their preventive measures accordingly.
MS. STEWART: I would tend to disagree with that. I think if you have a second a year after the first one, it would have a very serious effect, that it would certainly question any efforts to rebuild in the area. Individuals who lived in the area and businesses that were there would question whether it made any sense to stay. There would be some real decision-making going on as to whether they should pick up and go somewhere else.
The same thing would hold true for the reaction of some of the financial markets. After they had absorbed the first shock and gotten back to working more or less normally again, the second one—although it would not have the impact of the first—would certainly undermine a lot of the normality that had been reestablished in the markets.
It is a very good question, and one worth thinking about. It also raises the earlier question about how much capital the insurance industry is going to have left. This is the real question. I do not think the industry is worried about paying that first round of losses. What happens with the next round and the next round and the next round? Capital does not move that quickly between industries and sectors in the economy. Certainly in the insurance industry, the way it has been losing money on its own without any catastrophes, it has not been able to attract new capital. I imagine it could attract capital but at a very, very high price and perhaps in limited amounts, so that we would also be talking about very large increases in insurance premiums and perhaps in all kinds of insurance premiums.
In other words, there are going to be effects upon effects upon effects. We are dealing with a very complex problem. These things will adjust over time, that is true. But in the short term, there are going to be some very serious problems.
QUESTION: I have a couple of short comments. First, you know, economics is often referred to as the "dismal science." It is good to have a cheerful economist like Dr. Cheng. I think there are some conceptual problems here in the way that we are looking at this problem. The first conceptual problem is the loaded language. For example, the term "catastrophic" earthquake. This term begs the question of what a very large earthquake will do. The real concern is, we need some very cool and careful studies of the impact of certain kinds of very, very large earthquakes—economically, physically, socially, and politically. A catastrophic earthquake has been defined as one which disrupts the national economy, etc., etc., etc. From the very beginning that definition set the wrong conceptual tone.
The second problem concerns the question of uncertainty. If you look forward on a short-term basis, there is great uncertainty. If you look back on a long-term basis, there is no uncertainty. If you look back, for instance, at the history of cities—which I am interested in doing—you find that cities have survived for hundreds and thousands of years in highly earthquake-prone areas. The occasional earthquake is, in effect, a natural phenomenon which is accommodated by the city, and the city continues because the reasons for its continuing are much, much stronger than the effect of the earthquake. So what do we really mean by short-term/long-term?
As you look forward, I would suggest there is much less uncertainty on the long-term basis than on the short-term basis. We are a very short term society, and this affects us in many ways other than earthquakes. You know, we have a 1-year business orientation to the bottom line. We have a 4-year political orientation to what is going to happen while in an office; this cuts against the way in which nature works, which is on a very long-term basis. We have got to adjust our social and economic systems to the pattern which exists and to taking care of ourselves on a long-term basis, whatever that may be.
For you, 1 year may be short-term. For me, that may be long-term. Fifty years may be long-term. For nature or geology, that is a gnat's eyebrow.
QUESTION: Dr. Cheng, in your closing remarks, you mentioned that the Tokai Bank estimates it were based on highly questionable assumptions. Also in your closing remarks, you mentioned that methodologies do exist which can be used to make the precise estimates that you feel are needed. Were the assumptions the only problem with the Tokai Bank study? Was there a solid methodology used there or not?
DR. CHENG: I only read the abstract, and I really have not seen the details of the methodology. But given the fact that some regional economic models have been developed, the next step is to put them in the context of a national economy model and build in realistic economic linkages, that is, linkages that come from the real world as opposed to what we assume.
Now, talking about the Tokai study, it makes several unrealistic and wrong assumptions. For instance, it assumes that the impact would hurt the United States. That would be opposite to my prediction. If Tokyo were wiped out, the U.S. producers are in good shape; we are going to supply stuff to the world market because they are not coming from Japan. We are also going to supply goods to Japan because they need them for reconstruction.
My specialty is international trade, which is concerned about interdependence all the time. The Tokai study was an attempt to come up with quantitative estimates, but I think it failed because it made very unrealistic and wrong assumptions.
MS. STEWART: I do not think any of us would disagree that reconstruction after a catastrophe, be it an earthquake or a war, would be stimulating to the economies that would help in the reconstruction, just as we benefitted after World War II. This forum focuses on what to do to relieve the damage that is going to occur in the devastated area, rather than who is going to benefit from someone else's loss. I do not think there is any disagreement as to the stimulative effect of reconstruction.
QUESTION: Is there an analogous case that may lead to some testing of your models? For example, the reunification of East and West Germany may have an equivalent effect of a great earthquake. They have to rebuild the infrastructure, and I understand that estimates are something like $60 billion a year for the next 5 years, which is equivalent to $300 billion.
DR. CHENG: I think that is a very good point, and I would like to relate that to what Professor Hal Cochrane presented in his summary. There are many big events, and earthquakes are by no means the largest events in terms
of economic losses. As you know, the German reunification would require hundreds of billions of dollars to reinvigorate East Germany. That, of course, is going to have an impact in terms of the interest rate in the same way that the Tokyo earthquake will have. That will drive the interest rate up, and resources will be diverted away from other borrowing countries to East Germany. That is a very good case, and a great test of what we predict about these ripple effects. If that is the only event, I do not think that will affect the capital market very much. But if you combine that with, let us say, the economic reform in the Soviet Union, you may be talking about several trillion dollars.
Let us also not lose sight of the fact that right now, the less-developed countries already owe $1 trillion. When comparing $50, $60 billion with the overall size of the world capital market, of which $1 trillion is only a small portion, that is not going to have any major effect. Like the figures that I gave about World War I, the impact on the interest rates is a very good indirect test of the predictions. The effects will be small unless the events all occur at the same time.
QUESTION: Apparently, the markets at this point seem to agree with that; the markets are remaining stable.