Klaus Desmet & Esteban Rossi-Hansberg
Climate change is an unintended consequence of the industrialization of the world economy. The evidence that human activity has released large amounts of CO2 into the atmosphere, leading to rising global temperatures, is by now uncontroversial. However, so far, the scientific and political recognition of this reality has not translated into a commitment to emissions reductions sufficient to stop further global warming. As a result, economists are tasked with evaluating the economic costs of climate change and designing policies to address them. These evaluations are essential: the world cannot embark on ambitious attempts to reduce carbon emissions if we are not reasonably confident that the benefits of these actions will outweigh their costs.
Evaluating the economic impact of climate change is difficult. First, there is the natural science. Models that map carbon emissions to changes in global and local temperatures are readily available, but the mapping of many other physical impacts, such as sea level rise, extreme weather events, or nonlinearities in the climate system, is more complex. While our understanding of these effects is rapidly improving, as shown by the recent Intergovernmental Panel on Climate Change report, there are still no good off-the-shelf models that we can easily plug into our economic analysis.
Second, climate change evolves relatively slowly, unfolding over decades and centuries rather than over months and years. While anthropogenic temperature change is already affecting our present-day reality, many of its more pernicious effects will only be felt in the distant future. Evaluating the implications of warmer temperatures in the far-off future requires dynamic models, as recognized since the pioneering work of William Nordhaus. These protracted effects limit the usefulness of reduced-form empirical studies: extrapolating so far out of sample is undesirable and does not recognize the capacity of humans to react, respond, and adapt to changing circumstances. The Lucas critique — that historical data on the results of economic policy cannot be used to accurately predict the consequences of future policy because people’s behavioral responses also change over time — bites hard here.
Third, CO2 emissions are a global externality with local economic impact. Because CO2 mixes rapidly in the atmosphere, emissions from anywhere on the planet lead to changing temperatures across the globe. As a result, any attempt to evaluate the economic impact of climate change needs to be global in nature. At the same time, an aggregate dynamic model of the world economy is not sufficient if it ignores spatial heterogeneity. How can we discuss the impact of coastal flooding without recognizing the difference between Miami and Dallas, or without considering that people can move inland to escape inundation? And how can we evaluate the cost of a 1°C increase in global temperatures without recognizing that this will result in a more than 2°C increase in the most northern latitudes but only a 0.5°C increase in some equatorial regions? Perhaps more importantly, how can we do a comprehensive evaluation if we ignore the fact that higher temperatures are bound to have very different economic effects in the world’s coldest and warmest areas? Recognizing this spatial heterogeneity is essential for an accurate assessment of not just the aggregate impact of climate change, but also the spatial inequality that it might generate.