Quantifying climate damages when regions trade: a structural gravity approach

The world has warmed considerably in recent decades, a trend that is expected to continue. It is therefore important to measure the economic consequences of this warming, both to design effective policies and to understand why the economies of different countries are increasingly moving along different paths.

Much of the existing literature on climate damages estimates the effects of climate change by looking at past links between local weather shocks — such as unusually hot or cold periods — and local economic outcomes. However, when regions are connected through trade, a weather shock in one place can also affect economic activity in other regions, not just where the shock occurs.

The main contribution of this paper is a new way to measure how sensitive productivity is to climate conditions. The authors use the same basic sources of variation that are common in standard empirical approaches, but instead of relying on hidden or ad hoc assumptions, they make explicit assumptions grounded in economic theory. This framework is flexible enough to capture differences across regions and the trade-related spillover effects between them. Compared with simpler methods, the only additional data required are records of trade flows between regions, which are widely available at the international level.

This approach is used to re‑examine earlier studies on how climate change affects total economic output. Once the spillover effects created by trade between regions are taken into account, projected climate damages are found to be much larger than previously estimated.

Key points for decision-makers

• The method developed in the paper can also be used to measure the impact of other country‑level productivity shocks, not just climate change, while still accounting for complex spillover effects between countries.
• Extreme heat — measured as the number of days when temperatures exceed 30^oC — mainly reduces agricultural productivity, with smaller effects on manufacturing.
• These productivity losses led to reductions of about 5% of global output and real wages in 2019, compared with a scenario in which temperatures had not risen since 1991.
• These estimated losses are at least 50% larger than those found using earlier methods.
• Looking ahead to the end of the century, the authors project substantially larger climate damages than previous studies, as well as considerably greater uncertainty driven by where and how warming occurs across regions.