At 5:13 in the morning on April 18, 1906, San Francisco shook. Then it burned for days.

Inside the policies sitting in the city’s safes and its wrecked broker offices was a contradiction that adjusters were about to argue over: fire was covered, earthquake was not, and the fires had been started by the earthquake.

In London, one underwriter sent a cable that refused to have the argument.

The clause that split the city

Fire insurance was the standard product of the era, and it was written to pay for fire.

Earthquake was carved out. One San Francisco policy, written by the Swiss insurer Helvetia, excluded fires “caused directly or indirectly by earthquake.” Many policies also carried what was known as a fallen building clause. A standard version read: “if a building or any part thereof fall, except as the result of fire, all insurance by this Policy on such buildings or its contents shall immediately cease.”

Read strictly, the two clauses together closed both doors on a city that had just been shaken flat and then set alight. If the building fell first, the policy had already ended before the flames arrived. If the fire came first, the fire was still an earthquake consequence.

Every claim in San Francisco therefore turned into a question about sequence and cause: what happened to this specific address, in what order, on a morning when nobody was taking notes.

The British consul general in the city, Walter Courtney Bennett, put the stakes in one line: “If the insurance is not paid, the city is ruined. If it is paid, many of the insurance companies will break.”

Both happened in part. By the Insurance Information Institute’s count, at least 12 American insurers went bankrupt, along with one Austrian and one German company, and of $235 million in insured losses only about $180 million was paid.

The man in London

Cuthbert Heath was a Lloyd’s underwriter, described in Lloyd’s own history as a leading one, and at the time of the earthquake a leading earthquake underwriter in the market.

Lloyd’s is not an insurance company. It is a market where syndicates, meaning groups of backers who accept risk together, write business under their own names. Heath’s syndicate of twenty Names led Lloyd’s international earthquake policies.

He priced that book from his own handwritten Earthquake Book, a ledger of rates that is still held in Lloyd’s vaults.

So Heath was not a bystander in San Francisco. He was heavily exposed to what had just happened, and he knew it before the cables from California finished arriving.

What he actually sent

Heath cabled his San Francisco agent with an instruction that Lloyd’s own history records in a single clause.

“Pay all of our policyholders in full, irrespective of the terms of their policies.”

That is Lloyd’s rendering. The Insurance Museum drops the comma and the “of” and hyphenates policy-holders; a 2006 Lloyd’s spokesman gave it as two words with no comma. No telegram form is reproduced in any of them; what survives in print is the instruction, not the paper it went out on.

Seven words at the end of that sentence are the whole story. Irrespective of the terms of their policies.

He did not instruct the agent to interpret the fire clause generously. He did not instruct him to settle at a discount to avoid litigation. He removed the policy wording from the decision, in writing, for his whole book, before anyone had counted the claims.

The bill

Lloyd’s says the San Francisco earthquake cost the market over $50 million in 1906 money, and describes that as more than $1 billion in today’s dollars without naming a base year. Run through MeasuringWorth’s consumer price index comparison, $50 million in 1906 is about $1.85 billion in 2025 dollars.

That was not a gesture.

It did the opposite of what a bill that size usually does.

Lloyd’s dates its American reputation to that cable. The city rebuilt on money that arrived without an argument about sequence and cause, and American brokers who had never heard of a London syndicate learned in one season what the name on the policy was worth when the worst case arrived. Heath’s Names, by the Insurance Museum’s account, still shared an average profit of £431 on the year.

The syndicate that paid claims it could have legally contested bought something the contested claims could never have bought back.

Why it counts as rulebook, not charity

It is tempting to file the Heath cable under generosity and move on. That reading misses what it demonstrated.

An insurance policy is a promise that only gets tested once, on the worst day, by a customer who is in no condition to litigate.

San Francisco in 1906 was an early large public demonstration that the wording and the promise can come apart. A carrier can be entirely within its contract and still hand a burned-out family a denial letter that ends the relationship, and the town’s, forever.

Heath’s instruction priced that. He treated the reputational loss from a technically correct denial as a real cost on the ledger, and he paid cash to avoid it.

The fire also changed the form. Lloyd’s notes that earthquake itself is still excluded from most standard US homeowner and business policies, and that fire following an earthquake has been included in the majority of policies since 1906. The argument Heath refused to have was written out of the wording.

The city also gave the industry a lesson in accumulation, meaning the risk of writing too many policies in one place, and it gave it in one morning.

What the cable leaves out

There is no reservation of rights, no request for documentation, no instruction to review each file on its merits. There is no committee.

One underwriter, holding a handwritten rate book, made a decision for his own book and turned out to be right on a horizon nobody was measuring in 1906.

The Earthquake Book is still in the vault. The city is still there. The clause about the terms of the policies is the sentence that gets quoted 120 years later, because it is the rare case of an insurer being remembered for the argument it declined to make.

Lloyd’s own archive account of the 1906 San Francisco earthquake holds its record of the Heath instruction.