15 June 2020

Are Economists Getting Climate Dynamics Right and Does It Matter?

The assumptions of important economic climate models do not correspond to the actual dynamics of climate change. Because some of their forecasts overestimate warming, they give the false impression that there is no way the goals of the Paris Agreement could be met. For climate policy decisions to be correct, these economic models must be brought into line with the state of the art in climate science.

Key issue

Climate change is one of the major challenges facing our planet. Carbon dioxide remains in the atmosphere for centuries after it is emitted, and the climate system operates on timescales ranging from seconds to millennia. Clearly, there is a need to accurately represent climate dynamics in economic models of climate change if appropriate policy prescriptions are to be made. We pose the question that the key integrated assessment models of the economy and the climate used by economists fail to get the climate dynamics right and therefore their policy messages are flawed.

Approach and methodology

We select six economic models – the three most influential integrated assessment models and three analytical models from prominent economists – and test how their climate modules respond in two experiments compared with the best climate science. We show that the climate science models uniformly heat up very quickly to a constant, steady-state level, whereas the economic models heat up slowly, and in some cases very slowly, and do not attain a steady-state temperature within two centuries. We show that Earth System models exhibit positive feedbacks in the carbon cycle due to diminishing marginal uptake of CO₂ by carbon sinks with respect to atmospheric CO₂, which in turn is proportional both to cumulative CO₂ uptake by carbon sinks and to temperature. By contrast, most of the economic models exhibit increasing marginal uptake.

Key findings and conclusions

The unsupported assumption made in economic models of a long lag between emitting carbon and resulting warming is behind policy recommendations of (too) low carbon prices. The same assumption is also behind the debate on discounting, which has taken a prominent place within the academic debate. Without a long delay, optimal carbon prices are much less sensitive to the discount rate, since this means the costs of global warming are materializing much faster. A failure to simulate positive carbon cycle feedbacks also leads to optimal carbon prices that are too low. The effect is larger when cumulative CO₂ uptake and temperature are high; overall it is of comparable size to a long delay. Lastly, we specifically find that the economic module of Nobel laureate William Nordhaus’s DICE model heats up too much in the long run; this contributes to the false impression that it is infeasible to limit warming to 2 degrees Celsius as mandated by the Paris Agreement. Hence, it is important to bring economic models into line with the state of the art in climate science.