#818 1997 · USAA · Insurance / capital markets
USAA couldn't buy enough hurricane insurance for itself, so it sold hurricane risk to bond investors instead
the problem
Hurricane Andrew wiped out so much reinsurance capital nobody could buy enough coverage
background
Hurricane Andrew tore through Florida in 1992 and produced roughly $17 billion in insured losses — about double what catastrophe models had predicted — bankrupting several insurers outright and burning through so much of the reinsurance industry's own capital that the reinsurers many insurers depended on for worst-case protection simply stopped writing new hurricane coverage, or priced what little capacity remained far beyond what insurers could afford.
USAA, the insurer for U.S. military members and their families and heavily concentrated along the hurricane-exposed Southeast coast, faced exactly this squeeze: a single storm bad enough to breach its reserves could threaten the company, and the reinsurance market it would normally turn to for that protection no longer had the capital to sell it at any reasonable price. The insurance industry's own capital base — on the order of $250 billion in total equity across the whole U.S. sector — was itself simply too small relative to the largest imaginable hurricane loss to remain the only place that risk could live.
what everyone would do
Buy more reinsurance — the standard, decades-old way an insurer protects itself against a loss too large to absorb alone. After Hurricane Andrew burned through the reinsurance industry's own capital, that market simply no longer had enough left to sell, at any price USAA could pay.
what they saw
The reinsurance industry's entire capital base, a few hundred billion dollars, was itself too small to reliably absorb the largest possible hurricane loss. The broader capital markets, by contrast, held on the order of $19 trillion — a pool roughly seventy times larger, and one that would treat a rare, uncorrelated hurricane risk not as a threat to avoid but as valuable diversification. The reinsurance industry didn't need to grow; USAA just needed to stop asking only the reinsurance industry.
the move
In June 1997, USAA worked with Goldman Sachs and Merrill Lynch to issue a $477 million bond, Residential Re, that passed hurricane risk directly to bond investors instead of a reinsurer: buyers earned a high coupon for the life of the bond, but if a qualifying East Coast hurricane caused USAA more than $1 billion in losses, investors began losing their principal, forfeiting it entirely above $1.5 billion — in effect renting investors' own capital as reinsurance, with the bond market standing in for the reinsurer.
why it works
Structuring the risk as a bond let USAA pay a high coupon in exchange for investors accepting a small, well-modeled chance of losing principal — a trade reinsurers, whose whole business concentrates in exactly this risk, could no longer offer at scale after Andrew, but that a diversified bond investor could accept easily because one hurricane's effect on their broader portfolio was negligible. Because the bond's payout depended on hurricane losses rather than stock or interest-rate movements, it behaved as a nearly uncorrelated asset — precisely what made investors eager to hold it rather than merely willing to.
the payoff
Demand for the bond ran to roughly $1 billion against the roughly $150 million USAA had originally planned to raise, and the company ultimately placed over $477 million to meet it — because a catastrophe bond behaved as a nearly zero-beta asset, moving independently of stocks and other bonds and offering real diversification alongside a high yield; that year's tranche returned about 11 percent. The deal, the largest cat bond issued to that point, demonstrated at scale what smaller bonds pioneered a year earlier by underwriters like St. Paul Re UK had only hinted at: capital markets, not just reinsurers, would price and hold hurricane risk.
where it breaks
It depends on investors trusting the models that translate a storm's physical severity into the bond's payout trigger — when the model and the insurer's actual losses diverge (basis risk), investors either overpay for protection the insurer never needed or the insurer discovers its 'coverage' didn't pay out when a real loss hit. It also needs a capital market genuinely receptive to esoteric, hard-to-price risk at real scale, which requires investor sophistication and depth that a smaller or less developed market simply doesn't have.
what came after
Catastrophe bonds grew into a standing multi-billion-dollar annual market that insurers, reinsurers and even governments now use routinely to offload earthquake, hurricane and flood risk directly to capital-markets investors; USAA alone has since sponsored dozens more, including bonds covering risks the original 1997 deal never touched, such as flood losses on auto policies.
references
- [1]Catastrophe and the Capital MarketsWharton Magazine, University of Pennsylvania, 1998magazine.wharton.upenn.edu
- [2]Catastrophe Bonds – A to ZInsurance Journal, 2018insurancejournal.com