The Subnational Effect of Temperature on Economic Production: A Disaggregated Analysis in European Regions

In order to develop efficient strategies to counter the adverse economic consequences of climate change, accurate and spatially detailed assessments of economic damage are required. Estimates to assess the impact of temperature variations on macroeconomic output are usually based on country-level weather aggregates, neglecting that weather realizations tend to vary significantly within countries. Using data from multiple decades for spatially small-scaled European regions, we conduct a disaggregated analysis to mitigate the potential bias arising from spatial aggregation. We examine the economic impacts of temperature by analysing annual variations in two different weather indicators, namely yearly averages representing rise in temperature levels and standardized deviations from the region-specific climate norm representing unusual warm and cold periods. Our spatially explicit approach considers spatial dynamics and the spatial distribution of temperature effects as it captures spatial dependence via spillovers and allows for potential heterogeneous effects sizes for distinct spatial regimes. We find that regional-level growth reacts non-linearly to a rise in temperature levels, with a concave response curve similar to those estimated in earlier country-level studies. Interestingly, baseline temperature levels also moderate the effects of temperature deviations as unusually hot years adversely affect warm regions, whereas overly cold years foster growth. In contrast to most of the literature, we disclose that the relationship between economic growth and temperature variations is not generalizable. The uniform temperature-growth relationship found in the literature for countries at a global scale does not hold at the subnational level. The “world city” regions at the top of the urban hierarchy are not prone to any form of tested temperature variation. The resilience of these city regions can be explained, inter alia, by the prevalence of invulnerable sectors. The uneven effect sizes suggest that spatially ...