Insurance markets in California, Florida, and Louisiana have been repricing, contracting, or exiting for several years, and the pattern keeps spreading. Lenders ask harder questions about physical risk. Rating agencies pay closer attention. None of this is news. What remains unresolved is the arithmetic underneath it, specifically how to put a defensible number on what a low-probability event might cost a particular building in a particular place.

“People have a hard time calculating long-tail odds,” said John Macomber, a Senior Lecturer in the Finance Unit at Harvard Business School whose work focuses on climate adaptation and the future of cities. “Risk is not well understood across the entire built world ecosystem. Insurance companies, lenders, and asset owners are all approaching things in different ways.”

An insurer models expected loss on a one-year policy. A lender cares about the term of the loan. An owner holds for a decade or more and carries the residual value. Each is looking at a different slice of the same exposure with different tools and different time horizons.

The wildfire numbers show how fast the underlying assumptions have gone stale. Before 2015, wildfire made up roughly one to two percent of global insured natural catastrophe losses. It now runs closer to seven to ten percent, and eight of the ten costliest wildfire events on record have happened in the past decade. Global insured fire losses are growing around twelve percent a year. The January 2025 Los Angeles fires produced an estimated $40 billion in insured losses, by a wide margin the largest wildfire loss event ever recorded. Population in high wildfire risk zones has grown three times faster than the country as a whole since 1975. Billion-dollar weather disasters in the United States have climbed from three a year in the 1980s to 27 in 2024.

Risk that is not measured does not get priced and that is one of the things holding back much needed preventive measures. Adaptation spending has to be justified against the damage that would have happened without the investment. When expected damage cannot be credibly estimated, the case for spending money to avoid it falls apart before it reaches a budget committee.

Governments face a worse version of the same problem, because the political cycle is shorter than the recurrence interval. “Government might not come to the rescue,” Macomber said. “What mayor is going to invest in something that is going to happen down the road?” A seawall or a fire break may only pay off decades later, under someone else’s administration.

Macomber’s alternative is what he calls voluntary self-organized pools of capital, where parties with correlated exposure to the same hazard fund shared protection instead of waiting on a public program. Institutions concentrated in one disaster area face the same event, and none of them can solve it alone.

The Texas Medical Center in Houston offers a version of this. Tropical Storm Allison caused more than $2 billion in damage across the 700-acre campus in 2001, effectively shutting down the largest medical complex in the country. The institutions on that campus went on to invest collectively in flood protection for the shared tunnel system connecting their buildings, installing submarine doors rated to withstand twelve feet of water. The protection was only maid possible by shared resources and coordinated funding.

Insurance may not be able to protect the real estate industry from disasters alone. Insurers are already slowing pulling out of certain parts of the country. “It is really hard to have a policy that requires a private company to lose money, that is what we are asking of the insurance industry,” Macomber said. Regulators get to choose between premiums that reflect actual risk and premiums people can afford. Suppress rates and carriers leave. Price accurately and property values take the hit. Neither is politically survivable, which is why most states have settled on muddling through.

The federal backstop also looks less dependable than it used to. “FEMA has become political so people are starting to doubt that they will step in if needed,” Macomber said. Once that doubt enters underwriting assumptions, the math changes for lenders and investors regardless of what the agency actually does.

Harvard Business School is bringing these questions to practitioners directly. Its Institute for Business in Global Society and Business and Environment Initiative are convening a Hazard Adaptation Finance Workshop aimed at the measurement problem and the financing structures that might follow from solving it.

Private capital moves faster than legislatures, which is the argument for working this out among practitioners rather than waiting on policy. The pieces of the problem are at least visible now. Exposure is growing faster than the models tracking it. Insurers, lenders, and owners are each solving for a different time horizon, which leaves the full picture unowned. Public backstops look less reliable than they did a few years ago, and the political incentives around prevention have not improved. What follows from that is either better measurement and the financing structures it enables, or a slower repricing that happens one market exit at a time. Whether pooled adaptation funds can scale past a medical campus or an industrial district into something resembling a market is still an open question, and the answer probably depends on whether anyone can produce a number that all three parties will accept.

The post Until Real Estate Can Price Disaster Risk, It Cannot Fund the Fix appeared first on Propmodo.

​