SECTION: CAPITOL HILL HEARING TESTIMONY

LENGTH: 3061 words

COMMITTEE: SENATE COMMERCE, SCIENCE AND TRANSPORTATION

HEADLINE: TERRORISM RISK

TESTIMONY-BY: DAVID A. MOSS, ASSOCIATE PROFESSOR

AFFILIATION: HARVARD BUSINESS SCHOOL

BODY:
Testimony before the U.S. Senate Committee on Commerce, Science, and Transportation

David A. Moss Associate Professor Harvard Business School

October 30, 2001

Thank you Senator Hollings and Senator McCain for the opportunity to address the committee today.

1. The Government's Role as a Risk Manager

In his letter of invitation, Senator Hollings asked me to consider "the role the Federal government should play, if any, in indemnifying terrorism related risks.

Since I have just finished writing a book that traces the history of government efforts to manage risk, let me start by saying just a little bit about what the historical record suggests.'

Contrary to popular wisdom, government involvement with private- sector risks is nothing new. American lawmakers have been managing all sorts of risks since the earliest days of the Republic. Many of these government policies are so firmly entrenched that we take them for granted. Limited liability, for example, was first enacted at the state level beginning in the early nineteenth century and has since emerged as one of the hallmarks of modern corporation law. Yet limited liability is really nothing more than a simple risk management device, shifting part of the risk of corporate default from shareholders to creditors. Federal deposit insurance is another major risk management policy. Inaugurated during the Great Depression, this federal program safeguards individual depositors by spreading the risk of bank failure across all depositors and potentially across all taxpayers as well. Other notable risk-management policies include federal bankruptcy law, workers' compensation, unemployment insurance, old-age insurance, product liability law, state insurance guaranty funds, federal foreign-investment insurance, and federal disaster relief. Still other examples involve federal caps on liability, such as the cap on nuclear power liability enacted in 1957 (Price- Anderson) and the cap on credit card liability established in 1970. One thing that these diverse policies have in common is that they all reallocate private-sector risks.

In a great many cases, such risk-management policies were introduced because lawmakers concluded that private markets for risk weren't functioning adequately on their own. As early as 1818, for example, Representative Ezekiel Whitman of Massachusetts focused on the problem of uninsurability as a reason for enacting a special bankruptcy law for merchants. "Every effort of the merchant is surrounded with danger..." he noted. "Gentlemen have said that the merchant may insure. ... He may insure against sea risks and capture. But are these all the risks to which the merchant is liable? Indeed they are not. The risks which overwhelm him are more frequently and almost always, those against which he can have no [private] insurance."-" Such logic - emphasizing the government's role as a risk manager in a world of imperfect markets for risk - has helped to produce some of our most enduring and vital public policies, from bankruptcy law to Social Security.

Based on this history, I think it is fair to say that the prospect of involving the federal government in the management of terror-related risks would in no way constitute a radical departure from the path of American policymaking.

What the historical record also makes clear, however, is that some risk-management policies have worked considerably better than others. Policies like limited liability law, federal deposit insurance, and even the cap on credit card liability have worked remarkably well. Some others have worked less well. Federal disaster policy, for instance, has probably proved unnecessarily costly by encouraging construction in hazard-prone areas.

There are many factors that serve to distinguish relatively successful risk-management policies from less successful ones. But perhaps the most important factor is the degree to which risk taking can be effectively monitored within the confines of the policy. Whenever risk is shifted from one party to another - whether through an insurance contract, some other financial transaction, or a government program - the party that sees its risk reduced may have an incentive to engage in additional risk taking. Economists call this moral hazard.

Private insurers have long recognized that controlling moral hazard requires that they carefully monitor their clients. In writing commercial fire insurance, for example, they may require the regular inspection of sprinklers and other safety devices. The same basic principle applies in the case of government risk management as well. Unless some sort of monitoring - either by public or private actors - is built into a risk management policy, moral hazard is liable to spin out of control.

Fortunately, this sort of monitoring is built into a large number of our risk management policies. In some cases, incentives are created for private actors to do the monitoring. Limited liability for corporate shareholders works well because, in most cases, corporate risk taking is effectively monitored by private creditors, such as bankers and bondholders. In other cases, the government itself does the monitoring. Federal deposit insurance provides a good example, since government regulators help to limit potential moral hazard by actively monitoring bank risk. Significantly, one of the most striking failures of federal banking policy came in the 1980s, when federal oversight of the S&Ls was relaxed at the same time that federal insurance coverage was actually increased. For the most part, however, federal monitoring of bank risk has proved reasonably effective over the years.

Unfortunately, effective risk monitoring has never been a major part of federal disaster policy, leaving it exceptionally vulnerable to moral hazard. To be sure, the emergency appropriations that Congress has consistently approved in the aftermath of major disasters have relieved - and even prevented - a great deal of suffering and distress; and they have helped facilitate and accelerate recovery in devastated areas. But there has been precious little success in fashioning a disaster policy in this country that would help to control reckless building and other risky behavior that ultimately compound disaster losses. Indeed, the disaster relief itself has probably increased this sort of behavior.

In the wake of the great Mississippi flood of 1993, which triggered over $6 billion in federal relief, Representative Fred Grandy of Iowa observed, "We're basically telling people, 'We want you to buy insurance, but if you don't, we'll bail you out anyway."' Similarly, Representative Patricia Schroeder of Colorado noted that as "we watch this tremendously awful flood scene unravel in the Midwest . . . we are going to have make some very difficult choices.One of the main choices will be: Do we help those who took responsibility, got flood insurance, put up levees, tried to do everything they could; or do we help those who did not do that, who risked it all and figured if all fails, the Federal Government will bail them out."' Sadly, federal disaster policy has never adequately addressed this challenge.

It is said that the path into the harbor is marked by sunken ships. My hope is that Congress and the President will successfully navigate around the defects of federal disaster policy in crafting a program that facilitates the efficient management of terror-related risks in the aftermath of September 11'". If the federal government is going to assume responsibility for bearing some part of the risk that is currently borne by private insurers, it is essential that the resulting public policy provides for the effective monitoring of risky behavior - either through outright regulation or, better yet, well structured incentives.

The history of risk management policy demonstrates unmistakably that government can serve an enormously constructive role as a risk manager. But the historical record also provides a warning about the problems that can ensue from open-ended risk-absorption policies that impose little in the way of discipline and ultimately look more like simple subsidies than anything else. This, it seems to me, is the proper context in which to take up the problem at hand.

II. Insurance After September 11th: Defining the Problem

One of the many ramifications of the horror of September 11th is that the market for insurance against terror-related risks has been severely disrupted. Some say it is on the verge of collapse. A report put out by Tillinghast-Towers Perrin estimates that insured losses stemming from the September 11th attack will be between $30 and $58 billion, making it "the largest single-event loss in history."'

An event of this magnitude affects the insurance market in two separate, though related, ways. First, our expectations about future losses stemming from terrorist attacks obviously increase substantially. And this implies that even if insurance and reinsurance providers were willing to continue covering terror- related risks, with no disruption in service, the cost of such coverage would rise sharply for many policyholders. There can be little doubt that the cost of insuring a skyscraper like the Sears Tower should increase substantially as a result of our new knowledge about the risks of terrorism. In some cases, insurance could prove prohibitively expensive, destroying the economic viability of certain business endeavors.

A second - and even more disturbing - consequence is that terror- related risks could be rendered uninsurable in the private marketplace. The magnitude of the September 11th catastrophe has forced insurers and reinsurers to think seriously about previously unimaginable events (or series of events), some of which could conceivably swamp their reserves. Although the combined resources of the insurance industry are obviously very large, they are nonetheless finite.

A closely related concern is that there is now enormous uncertainty about how to estimate the probabilities of future terror-related losses. According to most insurance textbooks, one of the preconditions for insurability is that expected losses can be estimated with a fair degree of confidence. "For an exposure to loss to be insurable," reads one prominent textbook, "the expected loss must be calculable. Ideally, this means that there is a determinable probability distribution for losses within a reasonable degree of accuracy. ...

When the probability distribution of losses for the exposure to be insured against cannot be accurately calculated, the risk is uninsurable."

In explaining their intention to withdraw from covering terror- related risks in the absence of government backing, many insurance and reinsurance executives have cited precisely this combination of factors: the extraordinary magnitudes of potential losses involved and the near impossibility of accurately estimating loss probabilities. Said the Chubb Corporation's Chairman and CEO Dean R. O'Hare at a recent congressional hearing, "The industry has a specific amount of capital and cannot insure risks that are infinite and impossible to price."'

The business community thus faces two distinct problems in obtaining coverage for terrorrelated risks in the wake of September 11th - high cost on the one hand and uninsurability on the other. Even under the best of circumstances, such coverage would likely become far more expensive than it was before the tragedy, raising costs for many businesses and potentially forcing some under water. Under the worst of circumstances, such coverage would be unobtainable at any price, severely disrupting numerous markets but especially real estate.

Assuming that federal lawmakers wished to address either one of these problems, the former (high cost of terror insurance) would require some sort of government subsidy, whereas the latter (uninsurability) would require the government to act as a risk manager, perhaps providing terror insurance itself or facilitating its provision in the private sector. Although these two problems - high cost and uninsurability - are obviously linked in the current crisis, I believe it is useful to think about them separately when contemplating a potential policy response.

III. Fashioning a Policy Response

Ideally, I believe we should work to fashion a public policy that ensures a working market for terror risk with as little subsidy as possible. That is, we should try to address the sources of uninsurability, to the extent that they exist, while working hard to avoid any action that would make the private costs of terror- related risks look smaller than they really are.

To return for a moment to the example of federal disaster policy: one of the consequences of repeatedly providing ad hoc disaster coverage (relief) that is not tied to any sort of premium is that it ends up encouraging construction in unsafe areas, such as flood zones. That is because those who live in these areas often do not have to bear anything like the full actuarial costs. Federal disaster policy, in other words, helps to manage a wide range of risks, some of which might otherwise be uninsurable in the private marketplace. But it also ends up subsidizing those in the highest risk areas, since federally covered losses are funded not out of experience-rated premiums but rather out of general government revenues. Although some degree of subsidy is probably inevitable in any public policy designed to address disaster losses, our own federal disaster policy seems to carry the practice to an undesirable extreme.

In constructing a federal policy to facilitate the coverage of terror-related risks, one way to avoid these pitfalls would be to follow a more explicit model of insurance or reinsurance, where the government demands risk-based premiums in return for the coverage it provides. Although private insurers and reinsurers have expressed doubts about their ability to cover future terror- related risks, the federal government is ideally suited to underwrite precisely this sort of risk. Unlike private entities, the federal government is well positioned to absorb even massive losses because it enjoys the power to tax as well as a near- perfect credit rating. If the premiums it had collected were not sufficient to cover a particular loss, whether because of simple bad luck or misestimation of the risk, it could always draw the needed funds from general revenues and then recalibrate its insurance (or reinsurance) program after the fact. As FDR's Secretary of Labor, Frances Perkins, once said of the Social Security old-age-insurance program, "we have the credit of the Government as the real underlying reserve. That is what gives this stability."'

One of the biggest challenges in constructing a federal program for insuring or reinsuring terror risk would be to figure out how best to set premiums, so as to avoid excessive cross subsidization and thus the distortion of traditional market incentives. Another challenge would be to structure the program so that the federal role would automatically shrink if private insurers and reinsurers ever demonstrated a willingness to reassume the burden.

One option, which I favor, would be to establish a new federal reinsurance program for terror-related risks. Primary insurers that met appropriate standards would be permitted to reinsure terror risk with the federal government. Under such a program, a primary insurer might be allowed to pass (at its discretion) between, say, 20 and 80 percent of its terrorrelated risk - along with the same percentage of the premiums it charged - to the federal government. Ideally, the primary insurer would also be able to purchase some sort of federal stop-loss protection on the portion of risk it retained!

There are three main advantages of this federal reinsurance approach:

First, by allowing insurers to cede a substantial portion of terror risk to the federal government and by setting an absolute ceiling on their terror-related losses (through stoploss protection), a federal reinsurance program would ensure that coverage against terrorism would continue to be written in the aftermath of September l It".

Second, this approach would exploit the inherent strengths of the private market.Since primary insurers would remain responsible for writing terrorism policies, setting premiums, and retaining at least a portion of the risk, a federal reinsurance program would make effective use of their unparalleled capabilities in risk assessment, risk monitoring, and policy administration. Perhaps most important, nearly all of the covered risk - even that portion ceded to the government reinsurer - would be appropriately priced in the private marketplace, thus minimizing any distortion of vital market incentives.Developers who wished to build skyscrapers or other structures that insurers deemed unusually vulnerable to terrorist attack might be deterred from doing so by the prospect of exceptionally high insurance premiums. As a result, there would be little need for additional federal regulation to monitor risk taking and control moral hazard. Risk-based premiums, combined with monitoring by private insurers, would likely prove sufficient.'

Finally, under a public reinsurance plan of this sort, the federal government's role would automatically recede if private insurers and reinsurers chose to assume more of the risk. Not only would participating insurers be free to increase their retention levels as desired (up to 80 percent), but they would also be free to withdraw from the program entirely or to obtain coverage from private reinsurers if the latter ever reentered the market. Competition, in other words, would be encouraged rather than stifled.

While no plan is perfect, I favor a program of federal reinsurance for terrorism-related risks because, in my view, it would combine the best of both public and private, drawing on the government's unique strengths as a risk manager without short- circuiting either the essential capabilities or the relentless discipline of the private market. It would come as close as is possible at the present time to making a broken market whole and restoring a precious source of security in our economic life.



LOAD-DATE: October 31, 2001