The Return to Adaptation in a Changing Climate

We develop a two-step econometric framework to quantify the economic return to adaptation in a changing climate. The approach exploits the time-inhomogeneity of disaster arrivals, which generates learning and time-varying incentives for adaptive investment. First, we estimate country-specific dynamics of perceived disaster risk using a flexible class of time-varying arrival processes. Second, we link these dynamics to the marginal impact of extreme-weather damage on growth. Damage is systematically lower when risk is high—i.e., when disasters arrive in clusters—consistent with risk-dependent adaptation. To identify the expectations mechanism, we show that the return to adaptation is highest in countries where disaster arrivals are most serially correlated or the signal of future strikes is strongest. Absent such adaptation, average national income today would be several percent lower, with the shortfall widening significantly as countries learn about their evolving climate risks. These estimates are robust across various perceived-risk processes with different martingale properties.