Programming note: As a reminder, paid subscribers now have access to my new podcast, Net Interest Extra. In the next episode, I’m joined by Andrew M. Dresner, an expert on payments who has advised JPMorgan Chase and others on strategy. To receive the link when it airs, sign up here.

“The thing to remember is the earthquake does not know the premium that you receive. I mean the earthquake happens regardless … you know, you don’t have somebody out there on the San Andreas Fault that says, ‘Well, he only charged a 1 percent premium so we’re only going to do this once every 100 years.’” — Warren Buffett

In August 1992, a tropical wave formed in the air off the west coast of Africa. By itself, this was nothing unusual: Similar troughs of low atmospheric pressure develop in this part of the world at a rate of around sixty a year, with little annual variance. But this one was different. As it moved west over the Atlantic, it picked up intensity. Within two days, it had gained a sufficiently clear circulation pattern to be called a depression. A day later, picking up wind, it became one of only around a fifth of tropical waves to morph into a storm.

Initially, the storm struggled to intensify due to wind shear, which disrupted its development and nearly caused it to dissipate. However, as wind shear decreased, it began to strengthen rapidly. On August 22, the storm reached hurricane status southeast of the Bahamas. As the first of the season, it was named with the first letter of the alphabet. The prior year, there’d been Ana; this one was Andrew.

Over the next 36 hours, Andrew underwent explosive intensification. It reached peak winds of 170 mph close to the Bahamas, where it caused extensive damage. At around 4:00 am on August 24, it entered Biscayne Bay causing a storm surge that lashed the coast of Florida.

Andrew's eye made landfall at about 5:00 am. Just before they were destroyed, instruments at the National Hurricane Center in Coral Gables, at the northern edge of the eyewall, recorded a maximum sustained wind of 138 mph, with a peak gust of 164 mph. The hurricane ripped through Dade County, destroying more than 25,000 homes and damaging 100,000 more. After briefly re-intensifying over the Gulf of Mexico (as it was known then) Andrew made a second landfall in Louisiana, where storm tides, tornadoes, and winds up to 105 mph damaged crops and property. It was one of the most powerful hurricanes in US history – only the third Category 5 to strike the continental US since 1900.

Before Hurricane Andrew, insurers had assessed risk by looking backward. No hurricane had ever caused insured losses of more than $1 billion until Hurricane Hugo in 1989, which resulted in $4.2 billion in claims, so experts estimated a worst-case scenario of $7-8 billion. It turned out to be a gross mis-estimate. Insurers failed to account for shifts in population density, increased construction and property values along the coastline, and the possibility of a bigger storm. On final settlement, losses amounted to $15.5 billion which several insurance companies could ill-afford. A number of them became insolvent and others were severely financially challenged. Allstate alone – one of the largest insurance companies in America – paid out $1.9 billion, $500 million more than it had made in profits from its Florida operations across all lines, including investment income, over the 53 years it had been in business.

The hurricane transformed the way insurers think about catastrophe risk. Many primary insurers had inadequate reinsurance cover, a point Warren Buffett was referencing when he first made his quip about seeing who’s swimming naked only when the tide goes out. In the immediate aftermath, pricing for catastrophe risk spiked, benefitting some companies like Buffett’s Berkshire Hathaway (whose reinsurance business we discussed in Insuring the Unknown in January 2023). Increased demand led to more capital being pulled in, including from alternative sources in capital markets. For almost 20 years, finance professionals had experimented with ways to capture insurance tail risk in derivative structures. One innovation was cat bonds – securities that pay a high rate of interest and draw down on principal only when a specified catastrophic event occurs. Following Hurricane Andrew, issuance of these securities began to rise.

Today, catastrophe losses are materially higher. Adjusting for inflation, Andrew would rank only seventh among large insured US hurricane losses, replaced in top spot by Katrina, whose losses would be equivalent to $100 billion in 2023 terms, versus $38 billion for Andrew. Reinsurance capacity has grown commensurately. According to Aon, the industry has $715 billion of capital including $47 billion from cat bonds. And these cat bonds are growing in popularity. Even against the backdrop of rising catastrophe frequency, they have returned 41% – with very little volatility – since the beginning of 2022, based on Swiss Re’s total return index. Now, they are opening up to retail: An exchange-traded fund based on a portfolio of 75 of the 250 cat bonds outstanding is due to start trading on the New York Stock Exchange next month.

An occasional theme here at Net Interest is the emergence of new asset classes. In the past, we’ve talked about litigation finance, infrastructure, private lending, general partner stakes, subprime mortgages, supply chain finance, and more. I’m interested in how they develop, why some blossom and others don’t, whether they cause instability in underlying markets. Cat bonds look interesting – I mean, check out the chart – but what lies beneath? To explore the evolution of cat bonds and whether the cat bond ETF might be for you, read on.