The response of CO2 emissions to the business cycle: New evidence for the U.S

This paper investigates the response of CO2 emissions to the business cycle for the U.S. on a monthly basis between 1973 and 2015. Using a rolling-regression approach, we find that the emissions elasticity with respect to GDP is not constant over time, irrespective which filtering method, such as the Hodrick-Prescott, the Baxter-King, the Christiano-Fitzgerald or the Butterworth filter has been employed. In order to check whether or not emissions react differently during normal and recession times, next, we employ a Markov-switching approach. We find, first, that emissions are significantly more elastic during recessions than in normal times. Second, depending on the filtering method, we also obtain parameter estimates of the emissions elasticity above one in recession times and below one in normal times. The results are also robust against including monetary policy also in times of the zero lower bound. Thus, environmental policy instruments not turning out to be sub-optimal should account for this asymmetric response of emissions due to changes in GDP.