A Production Function Approach to the Gdp-Temperature Relationship

Abstract

Hotter countries tend to be poorer. In this paper, we estimate the income-temperature relationship for a cross-section of 97 countries. The relationship is both distinct and powerful. A simple cross-section regression of the log of per capita GDP against the log of average temperature in the capital city shows that temperature explains more than forty-five percent of the variance in income. A one percent increase in temperature is associated with a decrease in per capita GDP of between 2.0 and 3.5 percent. Figure 1 in the paper shows the data, which exhibit a clear downward slope between long-run average temperature and income per capita in 1985.

We investigate the income-temperature relationship using a Cobb-Douglas production function with temperature added as an input along with physical and human capital. A property of this model is that temperature lowers the marginal products of physical and human capital, which further implies that hotter countries will accumulate lower levels of these forms of capital. We discuss the evidence for these propositions. Estimation is based on the steady-state implications of the Solow-Swann model as amended by Mankiw, Romer, and Weil.

We then discuss different explanations of the income-temperature relationship, which are shown to have different inferences for what would happen if temperatures got warmer as a result of global warming. Under the interpretation with the strongest implications, we predict that a 2 degree Fahrenheit increase in average temperatures will lead to a 7.7 percent decrease in U.S. GDP and a 7.4 percent decrease in total GDP among the 97 countries in our data. Our estimates provide rare econometric evidence about possible economic consequences of global warming.

Keywords: global warming, income and temperature, new geography, Solow-Swann model

JEL Classification: Q0