Title: Modelling emissions by saturation estimation
Authors: Jonas Kai Kurle - University of Oxford (United Kingdom) [presenting]
Abstract: Super saturation (impulse indicator saturation and step indicator saturation) for modelling outliers and location shifts in statistical processes is highly relevant for environmental and climate econometrics. The properties of this approach are examined through extensive Monte Carlo simulations. They show that for tight significance levels, the detection rate of these breaks is generally high while the retention rate of irrelevant indicators is well-controlled. Furthermore, a newly designed indicator to model smooth location shifts is introduced, which is called policy transition indicator saturation (PTIS). Compared to step indicator saturation, the potency tends to be lower, which is partly due to larger uncertainty of detecting the correct break date. An application to UK CO2 emissions shows that PTIS may improve empirical modelling.