Conference Agenda

The existing literature on the macroeconomic impacts of climate change has established links between climatic shifts and economic growth. However, impact projections vary considerably across studies and are difficult to validate in hindsight, as the counterfactual growth path is unknown. This paper addresses this issue by leveraging the IMF's semi-annual World Economic Outlook forecasts from 1990-2022 as a baseline counterfactual. We first validate that these forecasts do not systematically capture climate impacts by demonstrating that the identifying variation for regression models in the literature stems from growth variation not predicted by the IMF. We then produce GDP nowcasts for different damage functions by adding climate impacts to IMF forecasts for recent years and validate these nowcasts against actual observed growth rates. We find that most damage functions modestly improve IMF growth forecasts, reducing forecast errors by 0.1-0.2 percentage points on average. However, the most severe damage functions predict substantial economic contractions in high-income countries between 2014-2019 that are inconsistent with observed growth patterns. Our methodology provides a novel validation approach for climate damage functions that complements existing methods and suggests that the most extreme damage estimates may overstate near-term climate impacts.