Underestimation of climate damage

Estimating how much damage climate change will do to the global economy is a difficult task. There are many non-linear effects at play that create positive feedback loops once a tipping point is crossed. Models that simply look at the average increase in annual temperatures in a region significantly underestimate the likely damage.

Models that estimate the economic impact of rising temperatures are extremely complex. But economists continue to increase their complexity to capture more nuances and details. Romain Fillon and his associates have shown that these details sometimes have major consequences for the economy. In their analysis, they used advanced models to estimate the effect of rising temperatures on global GDP, broken down by climate region. The additional complexity came from their focus on intra-year variations in temperatures.

Rising average temperatures mean that extreme temperatures on individual days become much more likely, while the average temperature on any given day changes very little. But it is these extreme heat waves and days of extreme precipitation that cause the most economic damage. What happens if you use the expected change in temperature variation within a year to predict economic damage?

The graph below shows the results for each major climate region in the world (the share of global GDP generated in each region as shown in the graph) and the current, most likely shared socio-economic trajectory. As you can see, climate models that ignore intraannual temperature variations underestimate economic damage by about 13% in temperate regions (e.g. Western Europe and most of the US) and by 47% in continental climates (e.g. Canada, Scandinavia, Russia). Not really what I wanted to learn on a Monday morning, but there you go.

Underestimation of climate damage to the economy

Source: Fillon et al. (2025)