Six days after the earthquake, the real estate men held a meeting.
The city was still smoking. The ground had moved at 5:12 on the morning of April 18, 1906, along nearly 270 miles of the San Andreas Fault, and the fires that came after it burned for three days. On April 24 the San Francisco Real Estate Board convened at Calvary Presbyterian Church, and the Chronicle called it the largest meeting the organization had ever held. They talked about relocating Chinatown. They talked about forming a joint rebuilding committee with the architects and the merchants. And they agreed, as the Chronicle reported the next morning, "that the calamity should be spoken of as 'the great fire,' and not as 'the great earthquake.'"
The record here is thin. No resolution survives, and if there were minutes they are gone. What we have is the newspaper account, and every historian who has cited it since traces back to that one paragraph. But the thing didn't need a vote to hold. A room full of men who owned property, sold property and wrote insurance on property understood something that goes without saying in a room like that: the name you give a disaster decides who pays for it.
The Contract Problem
Most fire policies in San Francisco carried no express earthquake exclusion. What a good many of them carried was a fallen-building clause. If a building or any part of it came down for any reason other than fire, the insurance on that building ceased at that moment. Something like sixteen companies used explicit earthquake language. The rest leaned on the fallen-building provision.
So everything turned on sequence. Building still standing when the fire reached it, the policy paid. Building already down when the fire reached it, the policy was waste paper. Adjusters had to reconstruct the order of destruction in a city where the evidence had burned, the witnesses had scattered, and every company's San Francisco office was gone. On April 21 the insurance men crossed the bay and met at Reed's Hall in Oakland, where they set up a General Adjusting Bureau to handle claims involving more than five companies. The bureau's committees were to determine the actual fire loss, after which each company remained free to negotiate according to its own reading of its own policy. The claims machinery was built around the fire question from the first week: adjusters working from across the water, reconstructing sequences in buildings they often could not inspect, out of records that had burned along with everything else.
Property owners needed a fire because fire was what they had bought coverage against. The boosters needed a fire because an earthquake implies the ground itself is unreliable, which is bad for investment, bad for immigration and bad for the railroad business. Southern Pacific's general passenger agent put it in writing to the Stockton Chamber of Commerce, in correspondence that survives:
"We do not believe in advertising the earthquake. The real calamity was undoubtedly the fire."
The Stockton Chamber wrote back that it was contacting eastern lecturers to stress fire over earthquake, and would try to "control lecturers already in the field."
The insurers themselves were split, which is what keeps this from being a conspiracy. A company looking to minimize payouts had every reason to call it an earthquake, since prior earthquake damage voided the fire policy. A company that meant to pay, out of honor or out of market calculation and the two are not always separable, had reason to accept the fire framing and settle fast. The Chamber of Commerce hired an insurance specialist from the University of California named Albert W. Whitney to find out who was paying and who wasn't, and his November 1906 report became a public shaming document.
Queen and Royal paid at 100 percent. Scottish Union and National paid in full less a two-percent cash discount. Rhine and Moselle invoked its earthquake clause, denied liability, pulled out of California and offered fifty cents on the dollar for claims of five hundred dollars or less. Trans-Atlantic denied liability, pulled out, and paid nothing.
The market rendered its opinion promptly. Companies that landed on the contemporary "Roll of Honor", the published list of firms that had settled square, increased their Pacific Coast premium business by 85.7 percent between 1905 and 1907. The others fell 4.4 percent. Paying fire claims turned out to be good business. Calling it a fire was good for everybody who wanted to be paid and everybody who wanted to keep selling policies on the coast. In the end insurers covered roughly 80 percent of about $250 million in insured losses.
The only people the arrangement was bad for were the ones who would need the next building to survive the next earthquake. Those people were in the future, and the future does not attend real estate board meetings.
A Building Code That Couldn't Say the Word
San Francisco's post-disaster Building Law was approved on July 5, 1906, less than three months after the ground stopped moving. Read it now and you find a document organized, with real thoroughness, around fire. Frame buildings were banned inside the fire limits and roof coverings had to be fireproof. Class A construction, the top fireproof classification, required a steel frame with incombustible structural members, the steel protected by at least three inches of concrete or by brick, terra cotta or approved metal lath and plaster. Masonry fire walls and parapets were required, thicker for commercial and industrial work.
There is a lateral-force provision in there. Buildings over 100 feet tall, or more than three times as tall as their narrowest dimension, needed a steel frame designed for 30 pounds per square foot of force from any direction, with diagonal or knee bracing carried from roof to basement.
That is a wind load. The word "earthquake" appears nowhere in the operative provisions of the law. No seismic coefficient, no fault setback, nothing about building on filled ground.
Governor Pardee had convened a State Earthquake Investigation Commission three days after the shock. Its preliminary report went in on May 31, five weeks before the building law was approved, and it already said that damage ran worst on artificial fill, reclaimed marsh and loose sand, and lightest on solid rock. That was on paper and available to anyone in city government who wanted it. The wind provision let everybody treat the earthquake problem as handled without having to name it. Engineering opinion of the day held that 30 pounds was adequate for shaking on the order of 1906, and for the tall steel buildings it covered, it may have been. It did not cover the wood-frame apartment houses, the unreinforced brick, or the buildings going up on fill where the shaking was worst.
I have spent enough years around job sites and around the men who write the rules for them to know that what a code leaves out gets as much thought as what it puts in. Nothing in the surviving record explains why San Francisco chose a wind number over an earthquake number. The choice fit a city that had already settled on what to call the thing.
The Lawson Report and Its Irrelevance
Andrew Lawson's full report, published by the Carnegie Institution in 1908, was a landmark of earthquake science. More than twenty researchers documented fault rupture, ground failure and structural damage across the region. The central urban finding was unambiguous: damage tracked soil. Buildings on rock came through the shaking with relatively minor structural harm. Buildings on fill and reclaimed marsh suffered severe ground and foundation failure before the fire ever got to them. The South of Market marshes, the Mission Creek corridor, the filled-in former Yerba Buena Cove. You could stand in the street and see the pattern in what was still upright, and then go look at what was underneath.
The report was not suppressed, whatever gets repeated on that score. Lawson's correspondence shows copies going out through the Carnegie Institution. But the print run was small, the readership was scientific, and the effect on policy was about zero. Philip Fradkin, in his history of the disaster, quotes the engineer John R. Freeman writing decades afterward that there "seems to have been a local attempt to suppress information about earthquakes." Freeman hedges, and he names no directive and no one who issued it.
What is documented is a civic habit of keeping earthquakes out of the conversation: the Real Estate Board agreement, the Southern Pacific letters, the booster framing running through the coverage. The Lawson Report lived inside that habit. Nobody burned it. It simply had no bearing on anyone deciding what got built and where, and it was not broadly reprinted until 1969.
You don't have to suppress a document when nobody with authority over construction has any reason to open it.
The Marina
Beginning in 1912, hydraulic filling prepared the Marina cove for the 1915 Panama-Pacific International Exposition. The fill was mainly loose sand, pumped in over beach deposits, soft bay sediment and older dense material. A popular version of this story has the Marina built on 1906 rubble. The geologic record does not support that as a description of the principal fill. The material responsible for most of the later liquefaction was hydraulic sand, put there six years after the earthquake, on ground the scientists had already identified as the most dangerous kind there is.
After the exposition the Marina became a residential neighborhood, mostly 1920s work: three- and four-story wood-frame apartment buildings, a great many of them with open first stories for garages. Good location. Common building type. And the ground under it was precisely what the Lawson Report had flagged as amplifying earthquake damage.
It wouldn't have taken a prophet to see it. It would have taken somebody reading a report that the city's business establishment had decided was beside the point.
Santa Barbara Figured It Out in 1925
Santa Barbara adopted explicit seismic building provisions in 1925, after its own earthquake. Palo Alto followed in 1926. After the 1933 Long Beach earthquake, magnitude 6.4, with 120 schools damaged and 70 destroyed and the children alive only because it hit at 5:54 in the evening, California passed the Field Act for school construction and the Riley Act, which put minimum seismic requirements on the whole state.
San Francisco stayed with its wind bracing. By then the surrogate had become self-confirming. Engineers could point at steel-frame buildings still standing from 1906 as proof the 30-pound number worked, and for tall steel buildings it may well have. Whether it did anything for the rest of the housing stock went unasked, because asking meant reopening the question of what the 1906 disaster had actually been. The 1927 Uniform Building Code offered optional lateral forces calculated from the weight of the building, which is a seismic approach rather than a wind proxy. San Francisco passed. The city revised its building law in 1921, 1928, 1930 and 1934 without adopting a seismic coefficient in any of them.
The city didn't adopt height- and soil-sensitive seismic coefficients until 1947 — forty-one years after the earthquake it wouldn't name.
Nothing in the available record ties that particular delay to the Real Estate Board's afternoon at Calvary Presbyterian. But a city that had organized its rebuilding around fire resistance, folded its earthquake provisions into wind design, and built a public identity as a place that burned rather than a place that shook was never going to be first in line when the seismologists came around with their coefficients.
What the Ground Did in 1989
On the evening of October 17, 1989, the Loma Prieta earthquake tested what the rebranding had built.
In the Marina, seven buildings collapsed and sixty-five took moderate or severe damage out of more than 1,400 structures. The losses concentrated in the type the neighborhood was full of, four-story wood-frame corner apartment buildings. Seventy-four of those accounted for six of the seven collapses. The weak first stories, the garage openings, went. The loose fill under them liquefied. The same filled and marshy areas recorded as heavily damaged in 1906 were heavily damaged again.
San Francisco's mandatory soft-story retrofit ordinance was signed on April 18, 2013, the 107th anniversary of the earthquake. It covers qualifying pre-1978 multifamily buildings with five or more units and two or more stories over a weak or soft first story. Which is to say it targets the construction type that failed in 1989, which was the construction type that went up on fill in the 1920s, on ground the scientists had described in 1906 and mapped in 1908, in a city whose business leaders had agreed that the problem was fire.
What the Real Estate Men Knew
The rebranding worked well enough to outlast everybody who did it. "The Great Fire" is still what people call it. The shaking lasted under a minute and the fire ran three days, and the fire makes better copy, and has for a hundred and twenty years.
What those men understood, sitting in that church six days after the ground moved, took no cynicism to arrive at. They were property owners in a wrecked city. Their coverage was fire coverage. Their livelihoods ran on money from somewhere else. The name they put on the disaster would decide whether that money came west or stayed home. Calling it a fire was rational, and self-interested in the way most consequential decisions are self-interested, which is not through villainy but through the ordinary human gift for believing whatever it pays to believe.
I have watched that gift operate in every industry I ever covered. A man looks at a thing and sees what his mortgage requires him to see. He isn't lying. He has organized his perception around his obligations and he will defend the arrangement with perfect sincerity. Multiply him by a city, hand him a newspaper and a Chamber of Commerce and an insurance adjuster, and you get a catastrophe reclassified by contract language.
The bill came later and it came to other people — a building code that took careful care of the disaster the city was willing to name, decades of construction on ground already mapped as dangerous, forty-one years of waiting for standards Santa Barbara had adopted in 1925. And then on an October evening in 1989, when the fill turned to liquid under the Marina and the garage-level apartments came down on themselves, the city found out that the rebranding had never been filed away in any archive. It was in the dirt.
- Johnstown's "Flood Free" promise: A federal channel project completed in 1943 became a civic identity, and a GAO survey of 67 flood victims found that 80 percent had believed the city was floodproof before the 1977 disaster — a belief some said influenced their decision not to buy flood insurance.
- Florida's code-enforcement gap after Andrew: The 1992 hurricane destroyed 49,000 homes, and the Florida Building Commission's investigation found the problem was not merely weak code text but compliance and enforcement across more than 400 local jurisdictions — a gap between a formal standard and a building that actually embodies it.
- Tulsa's stormwater fee as quiet adaptation: After a 1984 flood killed 14 people and damaged nearly 7,000 buildings, Tulsa created a dedicated stormwater utility whose monthly charge was raised to $14.65 in June 2026 — an adaptation that persisted after disaster memory faded by becoming an ordinary bill.
- Valmeyer's relocation arithmetic: The celebrated 1993 move to higher ground preserved the town's name and municipal government, but later research found that roughly 40 percent of original inhabitants did not move to the new site and only about 25 percent of businesses survived the transition.