For centuries, bond investing has boiled down to forecasting two things: which way interest rates are going to move and how likely a borrower is to repay its debts. A handful of startups are betting that to predict repayments in the future, bond analysts will need better data on something they’ve long overlooked—climate risk.
The new firms are competing to design algorithms that can predict the likelihood of natural disasters hitting specific towns, industrial parks, even individual buildings, and how much damage they could do. That could become more relevant if wildfires, floods, storms and drought strike more frequently and with greater severity, creating potential new losses for holders of municipal, corporate and mortgage-backed debt.