News

Floods rank among the costliest natural disasters worldwide, and European regulators are pushing banks to assess climate-related losses more rigorously. A new study by WorldQuant University instructor Marek Folprecht offers a practical way to quantify flood risk in real estate holdings without relying on crude assumptions that every property floods at once.

“I have been working on climate risk topics for a while at KBC Bank & Verzekering and wanted to explore these methods further and perhaps find a different perspective,” says Marek, who lives and works in Czechia.

His paper, published in the September 2026 issue of the Journal of Climate Finance, presents a framework that combines flood maps with financial loss modeling. The approach treats flood risk much like market risk, building continuous loss estimates for individual properties and then linking them through dependence structures that reflect how floods spread across regions.

This captures diversification effects that simpler stress tests often ignore. Two practical applications are demonstrated on a large sample of Czech house listings.

In stress testing, the model replays historical peak water-flow data from hundreds of monitoring stations and converts those events into portfolio losses. It successfully ranks past major floods—such as devastating 2002 and 1997 events—in roughly the same order as inflation-adjusted damage figures, confirming that the method tracks real-world severity.

A second approach calculates Value at Risk by fitting extreme-value distributions to water flows and linking them with a nested Gumbel copula. This structure accounts for stronger dependence within river basins and weaker ties between them. The model can generate loss scenarios more severe than anything seen in recent decades, giving risk managers a forward-looking view.

The framework is deliberately modular. Risk managers can swap in climate-adjusted flood maps, adjust correlation assumptions, or apply the outputs to stress loss-given-default figures for mortgage collateral.

Marek graduated from the Master of Science in Financial Engineering program at WorldQuant University in 2021. He is pursuing a PhD in finance at the Prague University of Economics and Business, where he also teaches.