Economic specialization and heterogeneous temperature-economy relationships suggest net costs of climate change in...

Econometric studies of temperature and GDP imply that warming harms hot countries, benefits cool ones, and that a single globally optimal temperature exists. We show that such aggregate relationships mask substantial spatial and sectoral heterogeneity and can mislead mitigation and adaptation policy. Using administrative district-level data for Europe on Gross Value Added (GVA) and GDP growth, we estimate the contemporaneous effects of temperature at national, district, and industry scales. In contrast to earlier global studies, warmer-than-average years reduce growth in relatively cold districts (0-14°C) and raise it in warmer regions (>14°C), with the pattern reversing at the extremes (<0°C and >20°C). This U-shaped relationship implies an average effect across Europe of -0.19 percentage points on annual growth, rather than the +0.18 benefit reported previously. Under RCP4.5, annual growth falls by 0.20 to 1.24 percentage points by 2070-2099, highlighting local temperature optima and heterogeneous vulnerabilities both within countries and across regions and sectors.