INSURANCE RISK

Risk transfer, defined as shifting the responsibility or burden for disaster loss to another party through legislation, contract, insurance or other means, can play a key role in helping to manage natural hazard risk and mitigate or minimise disaster losses. As the international community places increasing emphasis on disaster risk reduction, there is growing interest in the potential of risk financing solutions, of which risk transfer is a major component, as part of an overall disaster risk management strategy. Recent developments in this field include the use of a range of risk transfer mechanisms such as catastrophe bonds, catastrophe pools, index-based insurance and micro-insurance schemes. Social protection programmes such as safety nets and calamity funds can also provide effective financial instruments for managing risk and dealing with natural disaster shocks.

When Hannover Re consummated the first-ever insurance securitization in 1994, transferring $5 million of its risk to the capital markets in the form of a nebulous new product called a catastrophe bond, pundits hailed the birth of a dynamic—and potentially massive—new market for insurance risk. Reinsurance coverage following the Northridge, Calif., earthquake had all but dried up, and the capital markets were viewed by some as the successors to reinsurance in the insurance-risk-transfer business.

Predictably, prominent reinsurers weren’t pleased at the prospect of new competition. The reinsurance business managed to bounce back stronger than ever, but the securitizations kept on coming. Between 1994 and the end of 1997, 19 separate securitizations of one form or another were completed, providing billions of dollars of additional insurance coverage. Berkshire Hathaway, one of the nation’s biggest reinsurers, was clearly concerned about the trend when it issued this startling repudiation of the securitization business in its 1997 annual report:

“Many investors who are ‘innocents,’ meaning that they rely on representations from salespeople rather than on underwriting knowledge of their own, have come to the reinsurance business by means of purchasing pieces of paper that are called ‘cat bonds.’ The second word in this term, though, is an Orwellian misnomer. A true bond obliges the issuer to pay. These bonds, in effect, are contracts that lay a provisional promise to pay on the purchaser. This convoluted arrangement came into being because the owners of the contracts wished to circumvent laws that prohibit the writing of insurance by entities that haven’t been licensed by the state. A side benefit for the promoter is that calling the insurance contract a ‘bond’ may also cause unsophisticated buyers to assume that these instruments involve far less risk than is actually the case.”

Since this vituperation, however, much has changed. More than 15 transactions have taken place in the last 18 months alone, and reinsurers have come to accept—and in many cases embrace—securitization. “The existence of securitization is no longer seen by reinsurers as a threat to their business,” says Bryon Ehrhart, managing director of Aon Capital Markets. “They now think of it as their friend, as a tool they can use as part of their overall business strategy.”

Many believe the market is poised for explosive growth. While catastrophe securitization has dominated the market thus far, says Ehrhart, other types of insurance coverage, such as automobile and worker’s compensation, are ripe for securitization as well, once insurers take more of a capital markets approach to their balance sheets. “In automobile insurance, for example, the volatility over the last six years has been within 600 basis points, or 6 percent, yet insurance regulations force insurers to keep $1 in capital for every $2 or $3 they write in premium. If these companies were trading securities, they would not be allocating nearly as much capital as the current regulatory environment requires. So there’s a natural opportunity to take those risks off balance sheet, and we think there are unique ways to do that that haven’t yet been explored.”

Cooperation

Some reinsurers have become so bullish on securitization, in fact, that they’re now trying to help banks adapt the cat bond methodology to credit and other areas of the balance sheet. “The banking industry is almost in a state of crisis as a result of its large corporate presence,” says Michael Rulle, co-head of CIBC World Markets USA. He notes that, no matter which risk-adjusted-return models a bank uses, the actual returns in most bank credit portfolios are in the single digits. In fact, if the abnormally low default rates of the last five or six years are taken into consideration, one could argue that returns are close to zero on a risk-adjusted basis.

“A non-monoline insurer can have close to a

350:1 leverage ratio, while a bank can have a 10:1 ratio.

There are structures being explored to take advantage of

the extraordinary differences between these two regulatory regimes.”
—Michael Rulle

CIBC World Markets USA

For the last several years, of course, banks have employed various credit derivative strategies to take these exposures off the balance sheet. Lately, says Rulle, the reinsurance industry has tried to lend a hand as well. One reinsurer has guaranteed the subordinate tranches of certain collateralized loan obligation transactions around the single-A level, in order to make them more palatable to investors, improve liquidity and thereby improve the efficiency of the disintermediation process.

Another reinsurer has helped banks deal with unfunded assets such as revolvers, which are promises to lend when companies need to borrow. The reinsurer set up a Bermuda-based special-purpose vehicle to serve as a de facto derivative products company, providing a layer of credit insurance to offer an outlet for capital release from the banking industry. The SPV was a capital-arbitrage, tax-arbitrage and model-arbitrage strategy between the bank regulatory market and the insurance regulatory market.

Still another area ripe for exploitation, says Rulle, is regulatory arbitrage. “A non-monoline insurer can have close to a 350:1 leverage ratio for the same type of credit for which a commercial bank can have a 10:1 or 12:1 leverage ratio,” he says. “There are structures being explored that would use the capital-arbitrage advantage that a non-monoline insurance company has, along with some of the new credit value-at-risk modeling that is being put forth by a number of institutions, such that we could find a new slot in the capital structure of the marketplace to take advantage of the extraordinary differences between these two regulatory regimes.” Rulle notes that it’s only a matter of time before such a structure finds its way into the marketplace.

Meanwhile, the market recently witnessed a new area of securitization—insured credit risk. In April, Goldman Sachs arranged a deal in which Gerling Credit Insurance Group of Germany securitized 445 million euros of its insured credit risk. While this first deal was consummated with a primary insurer, some believe that trend won’t continue. “We think this type of risk-securitizing process will be the wave of the future,” says Alan Levin, managing director of ratings development and insurance ratings at Standard & Poor’s. “The only problem is that the insurance industry is, relatively speaking, capital rich, and doesn’t want to go through the months of effort, the legal documentation and all that’s necessary to create these transactions. That’s why reinsurance companies will probably be the vehicles through which transactions will be created in the future. Reinsurance companies will participate in this type of risk transfer, and then acquire their own capital by securitizing.”

False optimism?

Investor familiarity with these deals seems to be increasing dramatically. “In the recent Kemper deal, which was the first onshore transaction, we tripled the investor list,” says Aon Capital Markets’ Ehrhart. “We found a lot of people who had never invested in the asset class before who were willing to do at least $5 million in bonds. I don’t know if a securitization transaction has been done in the last year that hasn’t been massively oversubscribed.”

Some rating agencies, however, see the potential for big problems lurking just beneath the surface. The general trend for insurance securitizations has been toward less risky structures as the market matures, says Isaac Efrat, vice president and senior credit officer in structured finance at Moody’s Investors Service, but that pattern is starting to change. He notes that the June 1997 United States Automotive Association deal had extremely high-quality data, which made it easier for investors to understand the risks they were taking on. The July 1997 Swiss Re earthquake deal was indexed to the Property Claims Service index, offering even more transparency. The December 1997 Parametric Re transaction was tied to the magnitude of a Japanese earthquake rather than the losses the earthquake would create, removing moral hazard from the investment equation. This trend early on, says Efrat, was promising. But all the while, deals such as Georgetown Re, which securitized, among other things, the risk that a satellite would fall down, weren’t as attractive to investors, who were uncomfortable with that kind of risk. “My concern,” says Efrat, “is that, in recent times, we’ve seen some securitizations trying to lay off some unclear risks onto investors. That’s not something that anyone should get comfortable with, especially someone coming from the capital markets, where they don’t have reinsurance experience.”

“Investors are trying to become insurance companies,

but often they don’t know what’s being insured,

because nobody has told them what’s in the securitization package.”
—Isaac Efrat

Moody’s Investors Service

Efrat asserts that questionable modeling is the industry’s biggest problem today. Securitization structures are presented to rating agencies alongside a modeling agency’s model. But some deals are modeled improperly or incompletely. “Investors in these deals are trying to become, effectively, insurance companies, but often they don’t know precisely what’s being insured, because nobody has told them what’s in the securitization package.” Another problem, he says, lies in the allegiances of the modeling agencies, which are employed by the ceding insurers rather than investors. Ideally, investors would have their own sort of minimodel to supplement the modeling agency’s model. Moody’s, for one, tries to provide such a supplement for each structure it rates.

A more fundamental problem, says Efrat, is one of information imbalance. “Investors often say, ‘The sponsor knows everything about the risk that’s being ceded over, but I know nearly nothing about it.’ There’s a fundamental asymmetry there, which a lot of people are concerned about, and rightfully so. Something has to be done about this. For example, an insurer can go and take out things it doesn’t like and leave in things it does like. There is also the moral hazard issue—once an insurer’s securitization has been triggered, it can start underwriting recklessly, because it is protected from further default losses.”

Efrat’s solution to such problems? “Ultimately, we believe that going to parametric or indexed models is the best answer. Unfortunately, since the Parametric Re and Swiss Re earthquake deals, this has not happened. Most other deals have left investors exposed to what actually happens to their sponsor. That may not be something that can be sustained for too long.”

Others agree that modeling is far from an exact science, and that the insurance approach has fundamental advantages. “Many people think that capital markets groups understand how to price risk in a more sophisticated manner,” says Maureen Callahan, president at Callahan Co. “I would argue that capital markets groups can price only a tiny amount of the total business risk that global corporations have. There are a billion other kinds of risk that people have to deal with every day to stay in business, and insurers have a better grasp of that bigger picture.”

Berkshire Hathaway, for one, believes catastrophe modeling, at the very least, should be left to the insurers. Its 1997 annual report skewered the cat bond pricing system. “The lack of precise data, coupled with the rarity of such catastrophes, plays into the hands of promoters who typically employ an ‘expert’ to advise a potential bond buyer about the probability of losses. The expert puts no money on the table. Instead, he or she receives an up-front payment that is forever his or hers, no matter how inaccurate the predictions. Surprise! When the stakes are high, an expert can invariably be found who will affirm that the probability of rolling a 12 on two dice is not 1 in 36 but more like 1 in 100. In fairness, we should add that the expert probably believes that his or her odds are correct—a fact that makes him or her less reprehensible but in fact more dangerous.” The result, said the report: cat bonds were causing super cat reinsurance prices to sink far too low.

“The insurance industry is capital rich and doesn’t

want to go through the effort to create these transactions.

That’s why reinsurance companies will probably be

the vehicles through which transactions will be created.”
—Alan Levin

Standard & Poor’s

Some counter, however, that there’s little difference between the relative riskiness of cat bonds and other bonds. “Even though an investor continues to buy BB bonds, it’s conceivable that there will be defaults on a greater frequency than the experts have led the investor to believe,” says Morton Lane, president and CEO of Lane Financial LLC. “There’s nothing different with cat bonds. Certainly one ought to be careful about the choice of one’s experts, but everybody relies on experts, and everybody needs to do their homework.”

Most concede that it’s more difficult to do the homework in pricing catastrophes than corporate bonds, since there are rich historical data for corporate defaults but only a handful of meteorological catastrophe data points. In this area, reinsurers seem to have a natural advantage. But Lane isn’t so sure. “The fact that you don’t have many data points doesn’t necessarily lead to mispricing,” he says. “It’s as difficult for Warren Buffett to make those assessments as it is for the rest of us. Maybe he has better experts and more experience, but it’s still difficult for him.”

Other capital markets professionals argue, moreover, that in the long run the capital markets are better suited to absorb insurance losses—meaning that the capital markets securitization model, using risk-neutral pricing based only on expected losses, will ultimately prevail over the insurance pricing model, which includes expected loss plus a premium of some amount, known collectively as unexpected loss.

“Most diversification models say that the larger the span of capital, and the more diversified the nature of the risk within that span of capital, the closer you can price something to care only about expected loss, as opposed to unexpected loss,” says CIBC’s Rulle. “The more you condense something to a smaller pool of capital, and the less diversified that risk is, the more you have to price something for unexpected loss. So a single reinsurance entity that has a small amount of capital but is insuring an event that may have a high cat component to it, and low probability—but that could cause bankruptcy—might have to price its product at 5 percent of capital, simply because you can’t use a business model that says it’s OK to go out of business every 25 years. There’s no question that the capital markets are more equipped by definition. The process of getting there is different, of course, but in a macro sense the capital markets are equipped to take any capital risk, because they’re so much bigger.”

At some point, it seems, the capital markets model will prevail. At least that’s what some forward-looking reinsurers are hoping.