A new NBER paper by Berg et al (2023) estimates how temperature affects real GDP per capita for 137 countries that have at least 30 consecutive years of data. They find that the relationship is negative for about half of countries, but interestingly that the response is positive for the other half of economies. The authors also find that ‘a country’s GDP is more likely to respond positively to a global temperature shock if it is poorer, has grown less rapidly, is more open to trade, more educated, and more authoritarian.’
The estimates for South Africa show that global (i.e. common) temperature changes and idiosyncratic (i.e. South Africa-specific temporary changes) have a negative but statistically significant impact on real GDP per capita. At a five year horizon, idiosyncratic temperature increases are estimated to have a positive (but insignificant) impact.
The authors acknowledge that the estimated relationships may change if global temperature increases are larger than been observed in their sample.
The results suggest that wealthy countries have a self-interest in investing in abatement policies, but will no doubt generate a lot of debate and critique.