THAT 2017 suffered from more than its fair share of natural catastrophes was known at the time. In the wake of Hurricane Harvey, the streets of Houston, Texas, were submerged under brown floodwater; Hurricane Irma razed buildings to the ground on some Caribbean islands. That the destruction was great enough for insurance losses to reach record levels has only just been confirmed. According to figures released on January 4th by Munich Re, a reinsurer, global, inflation-adjusted insured catastrophe losses reached an all-time high of $135bn in 2017 (see chart). Total losses (including uninsured ones) reached $330bn, second only to losses of $354bn in 2011.

A large portion of the losses in 2011 was caused by one catastrophe: the earthquake and tsunami in Japan. Losses in 2017 were largely traceable to extreme weather. Fully 97% were weather-related, well above the average since 1980 of 85%. If climate change brings more frequent extreme weather, as Munich Re and others expect, last year’s loss levels may become depressingly familiar. Already, the data show many more frequent high-loss events since 2000—lots of them weather-related—than in the two preceding decades.

Last year’s disasters were particularly concentrated in North America (including the Caribbean), with 83% of global losses; half of those were in America alone, hitting that country’s insurers particularly hard. Fitch, a ratings agency, expects the “combined ratio” for American property-and-casualty insurers to rise from 100.7% in 2016, meaning costs and claim payouts just exceeded premium revenue, to 104.4% in 2017. That implies a substantial underwriting loss for the industry. Even Warren Buffett’s Berkshire Hathaway looks poised for its first full-year underwriting loss in 15 years. It took a $3bn hit from the three hurricanes and an earthquake in Mexico.

For all the gloom, the 2017 losses were also proof of the resilience of the reinsurance industry. Insurers have long spread catastrophe risk by taking out reinsurance policies. This time, reinsurers had such ample capital buffers that they are expected to suffer only a small dent, of around 5-7% of capital. And 2017 was also the biggest test so far of reinsurance provided directly by investors, whether through catastrophe bonds or “collateralised reinsurance”, where a fund manager puts up collateral to cover potential claims. These forms of “alternative capital”, which reached $89bn in mid-2017, now make up around 14% of total reinsurance capital, up from 4% in 2006, according to Aon, a broker.

Their performance has been remarkably smooth. Investor demand has held up; many asset managers in the field have raised new money since the losses. Demand may yet grow further, says Paul Schultz, head of Aon’s capital-markets arm, since the yields on alternative capital are poised to rise because of growth in reinsurance premiums. Mr Schultz’s concerns lie elsewhere: he laments that the proportion of all losses covered by insurance “is still too small”. Much risk is retained by governments, or uninsured. Offloading more to private markets would benefit governments, property owners and the insurance industry alike.