Revisiting the temperature-economic growth relationship using global subnational data

Previous studies have used national data to demonstrate that higher annual temperatures negatively affect economic output and growth. Yet, annual temperatures and productivity can also vary greatly across space within countries. With this in mind, we revisit the relationship between temperature and economic growth using subnational short panel data for 10,597 grid cells across the terrestrial Earth. Our estimates from fitting a quadratic model to the data imply that cell-level economic growth in countries with below-median per-capita incomes is concave in temperature, with a maximum at about 16 °C. Our findings suggest that even with similar economic development within a country, climate vulnerability can vary at the regional level. Furthermore, as soon as we take into account the nonlinear relationship between temperatures and economic growth within countries, the impacts of temperature increases are found to be larger, compared to those that disregard such within-country heterogeneity.