Wharton’s Benjamin Keys was running into a problem with one of his research projects. The Rowan Family Foundation Professor and professor of real estate and finance had been hearing about sharply rising home insurance premiums across the U.S., especially in places vulnerable to natural disasters. But when he went to insurance companies for hard data, he came up empty. “We thought, ‘If the insurance industry doesn’t want to share geographically granular data, where can we find it?’” says Keys.
To that end, he and Philip Mulder GRW22 — assistant professor of risk and insurance at the University of Wisconsin-Madison — got creative. “The mortgage industry collects data on escrow accounts,” Keys says. “If you own a home, you usually submit your monthly payment all in one — including principal, interest, taxes, and insurance. We realized that from that data, we could back out how much people are paying in insurance.”
From there, they built a massive dataset with more than 84 million inferred instances of property insurance payments from 2014 through 2024, uncovering sharp increases in premiums starting in 2020, particularly in disaster-prone areas. “In high-risk zip codes that are exposed to things like hurricanes and wildfires, there’s been a big jump in premiums,” says Keys.
Much of the differential rise in high-risk zip codes, they found, was driven by mounting costs of reinsurance — a.k.a. insurance for insurance companies. Insurers buy reinsurance to spread some of the risk they take on and in turn pass some of those costs to homeowners themselves. “We project that if the reinsurance shock persists, growing disaster risk will lead climate-exposed households to face $700 higher annual premiums by 2053,” Keys and Mulder write in their paper, “Property Insurance and Disaster Risk: New Evidence from Mortgage Escrow Data.”
To help students think about using real estate data in their own work, Keys launched a new course this past spring: Real Estate Data Analytics. In it, students were tasked with a series of projects to compare U.S. cities and figure out where it might make sense to invest in apartment buildings. “We looked at things like rents, vacancies, population and employment growth, and where properties are being built,” says Keys. Once they narrowed down cities, they dug into the details — including neighborhood safety, commuting options, and schools — to find more precisely where to invest in their chosen cities.
In many ways, the course mirrored Keys’s recent work. “I wanted students to think about how they chose which data to analyze, how they prepared that data, and how they analyzed it,” he says — just as he did with the insurance data.
Published as “At the Whiteboard With Benjamin Keys” in the Fall/Winter 2025 issue of Wharton Magazine.
