Case Study 1 — Andrew, Katrina, and the Making of the Property-Catastrophe Reinsurance Market

A real, public sequence of events. All magnitudes are described qualitatively; no precise loss, price, or market-share figure is invented. The point is the structural lesson, not a number.

Background

For most of the twentieth century, property-catastrophe reinsurance was an established but relatively sleepy corner of the industry. Primary insurers wrote homeowners and commercial property along the U.S. coasts, bought catastrophe cover from a familiar set of reinsurers, and — because the country had not been struck by a truly enormous insured hurricane in living memory — broadly assumed the cover they held was sufficient. The catastrophe models that would later become central to the business were in their infancy or did not yet exist in commercial form. Pricing for hurricane risk leaned heavily on the recent historical loss record, which, in the absence of a major landfall for many years, looked deceptively benign. This is the setup the whole book warns about: a low-frequency, high-severity peril, priced on a quiet stretch of history, in a market that had not recently been forced to confront the correlated event.

Then, in 1992, Hurricane Andrew made landfall in South Florida. It was, at the time, one of the most destructive insured catastrophes the United States had ever experienced. For the reinsurance market the significance was not merely the size of the loss but what the loss revealed: that the industry's understanding of its own catastrophe exposure had been badly incomplete. A number of primary insurers discovered that their hurricane exposure — the accumulation of all the coastal policies they had written in one storm's path — was far larger than they had appreciated, and that the catastrophe reinsurance they had purchased attached too high, ran out too low, or simply did not exist in sufficient depth. Several insurers failed outright. Many survivors found their reinsurance recoveries, while real, left them holding a retained loss that strained or exhausted their capital.

Thirteen years later, in 2005, Hurricane Katrina struck the Gulf Coast and became another defining catastrophe for the industry, compounded in the same season by additional major storms. Katrina tested the market again — this time a market that had supposedly learned Andrew's lessons — and exposed further complexities: the interaction of wind and flood coverage, the adequacy of the catastrophe models a decade into their commercial life, the concentration of risk in vulnerable coastal zones, and once again the question of whether primary insurers and their reinsurers truly understood the magnitude of a single event's correlated loss.

The insurance / underwriting issue

This case sits at the exact intersection of everything Chapter 27 teaches: it is the story of how the catastrophe XOL market — the cover that stands between one storm and a primary insurer's solvency — was forced into its modern form by two events that proved the old form inadequate. Several underwriting issues braid together:

• The failure of the independence assumption, made concrete. Andrew and Katrina are the textbook demonstration of the point from Chapter 1 (§1.2): a hurricane does not strike one policy in the pool, it strikes thousands at once. The losses that overwhelmed insurers were not a frequency of independent claims but a single correlated event. This is why catastrophe reinsurance exists and why per-risk cover is no substitute for event cover.
• Accumulation that no one had measured. Primary insurers had underwritten coastal property one file at a time, each looking acceptable, without an adequate view of the aggregate exposure those files created in a single storm's footprint — the accumulation-management problem that Chapter 30 addresses in full. Andrew turned an unmeasured accumulation into an existential loss.
• A reinsurance program sized to the wrong event. Insurers who had bought catastrophe cover up to what they believed was a severe event discovered that the real severe event sat well above the top of their tower. The first job of the cat program — to define a retention and buy cover up to a chosen return period (the "1-in-100-year" thinking of Chapter 30) — failed when the chosen return period turned out to have been estimated against an inadequate model and a too-short history.
• The market's cyclical, capital-driven response. Each event destroyed reinsurer capital, and the reinsurance market reacted exactly as §27.6 describes: prices for catastrophe cover spiked, capacity contracted, terms tightened — and then new capital flowed in to take advantage of the hardened prices.

What it shows

Three structural lessons emerge, and each is a Chapter 27 principle written in the largest possible letters.

First, catastrophe is the peril that justifies the entire reinsurance edifice, and the cat XOL treaty is the instrument the property insurance industry cannot live without. Before Andrew, the market underestimated how much cover it needed because it had under-modeled the correlated event. The events did not change the logic of catastrophe reinsurance — that logic was always sound — they changed the quantity the industry understood it had to buy and the seriousness with which it measured its own accumulation.

Second, the reinsurance cycle is real, violent, and consequential for primary underwriting. After each event, the cost of catastrophe cover rose sharply. Because that cost flows straight into primary pricing (§27.5), the hard reinsurance market produced hard primary markets in coastal property: higher premiums, higher deductibles (the percentage wind deductibles of Chapter 12 spread precisely because reinsurance forced insurers to share more of the loss with insureds), and in some zones, insurers retrenching or exiting. The primary underwriter who did not understand that their cat-cover cost had repriced would have kept writing coastal business at rates that were adequate gross but ruinous net — the central trap of §27.5.

Third, and most enduringly, these events created the modern catastrophe-modeling and Bermuda-reinsurance infrastructure. Andrew accelerated the adoption of probabilistic catastrophe models as the basis for pricing and buying cat cover — the industry could no longer price a 1-in-100-year hurricane off thirty quiet years of history. And the post-event capital that flowed into Bermuda (§27.6) became a permanent, repeating feature of the market: when a catastrophe destroys capital and hardens prices, fresh capital enters to write the now-expensive cover. The market that exists today — model-driven, Bermuda-heavy, increasingly funded by insurance-linked securities — is in large part the residue of learning, expensively, what these storms taught.

Outcome

The property-catastrophe reinsurance market did not collapse; it reformed, and in reforming it became one of the most sophisticated risk-transfer markets in the world. Catastrophe models moved from novelty to the indispensable foundation of cat pricing and accumulation management. Primary insurers built genuine accumulation-management discipline, measuring their exposure by peril zone rather than discovering it after the storm. Reinsurance programs were rebuilt to attach and exhaust against modeled return periods rather than against an inadequate recent history. New reinsurance capital — much of it in Bermuda, and later much of it from the capital markets via catastrophe bonds — entered to provide the depth the old market had lacked. The percentage wind deductible became standard in catastrophe-exposed property, sharing the first layer of every storm loss with the insured. And the cost of catastrophe cover became a central, explicit, and repriced input into primary coastal underwriting, exactly as the net-versus-gross logic of this chapter demands. The events were tragedies and the insurer failures were real; the industry's response was to build the machinery that makes catastrophe-exposed property insurable at all today.

Lesson

The transferable lesson for the underwriter is the spine of Part V. Catastrophe reinsurance is not a back-office detail; it is the precondition for writing catastrophe-exposed property, and its cost and adequacy must live in the primary underwriting decision. A risk like Harbor Steel — a coastal plant in a named-windstorm zone — is writable only because a catastrophe XOL treaty stands behind the whole coastal book, and it is priced correctly only if its net price carries the cost of the cat capacity it consumes. Andrew and Katrina are the proof, at industry scale, of three things this chapter asserts at desk scale: that correlated catastrophe defeats the law of large numbers and requires its own cover; that the cost of that cover repriced sharply and flowed into primary rates; and that an insurer which mismeasures its accumulation or under-buys its cat cover is one storm away from insolvency. The underwriter who internalizes this stops seeing the cat treaty as someone else's concern and starts pricing every coastal file as a piece of a reinsured, accumulation-managed book.

Discussion questions

1. Explain precisely how Andrew and Katrina demonstrate the failure of the independence assumption from Chapter 1 (§1.2). Why can a primary insurer not diversify away hurricane risk by simply writing more coastal policies?
2. After each event, catastrophe reinsurance prices rose sharply. Trace the path by which that increase reaches the premium an ordinary coastal homeowner or commercial property owner pays. (§27.5, §27.6)
3. Andrew is credited with accelerating the adoption of probabilistic catastrophe models. Why was the historical loss record an inadequate basis for pricing hurricane cover, and what does a model add that history cannot? (Preview of Chapter 30.)
4. The percentage wind deductible spread widely after these events. Explain how a percentage wind deductible is, in effect, a way of making the insured share the first layer of a catastrophe — and how that relates to the cedent's retention on its cat XOL treaty. (§27.4; Ch. 12)
5. "A primary underwriter who prices coastal property gross will write business that is unprofitable net." Using this case, construct the argument for that claim, and state what the disciplined underwriter does instead. (§27.5)
6. New capital flowed into Bermuda after the events, when prices had hardened. Why does fresh reinsurance capital tend to arrive after a catastrophe rather than before, and what does that say about the reinsurance underwriting cycle? (§27.6; Ch. 3)