The impacT of nonlineariTies for carbon markeTs analyses

This paper examines empirically whether nonlinearities play a significant role in the modeling of the carbon price. We highlight the limits of previous carbon markets analyses based essentially on a linear econometric framework. Instead, we propose to revisit the main results on carbon pricing and the inter-relationships with energy markets and CERs based on nonlinear techniques (threshold vector autoregressive and Markov regime-switching models). While we are able to confirm most of the findings present in the literature, we show interestingly that these results seem to vary (and may even fade) depending on the regimes considered. Thereby, we illustrate the importance of including threshold effects in future studies of the relationships between CO2 and energy prices, which have been neglected so far. JEL Classification: C32; C51; Q40; Q54.