A0743
Title: Disentangling drivers of EU allowance prices: A mixed causal and non-causal time series approach
Authors: Eduardo Serrubeco Marques - University Paris Dauphine (France) [presenting]
Arthur Thomas - Paris Dauphine University - PSL (France)
Olivier Massol - Centrale Supelec (France)
Abstract: The purpose is to analyze the drivers of EU Allowance (EUA) price fluctuations within the EU emissions trading system (EU ETS), focusing on a methodology tailored to the carbon markets' unique characteristics. An extended autoregressive distributed lag (ARDL) model is employed, following prior studies, which allows for estimation of both short- and long-run effects irrespective of the integration order of the variables. The model captures the impact of key fundamental drivers' natural gas, coal, and Brent crude oil prices on EUA prices, accounting for nonlinearities using partial sum decompositions as proposed by another study. Fuel-switching dynamics between coal and gas are modeled based on deviations from the theoretical switch price. To address heteroscedasticity in the residuals, a mixed causal and non-causal framework is incorporated, and importance sampling-based partial filtering is applied to decompose EUA prices into fundamental (causal) and speculative (non-causal) components. The causal component is regressed on fundamentals using the NARDL model, successfully mitigating heteroscedasticity. The non-causal component is further analyzed via a VAR model to separate the effects of rational expectations from market sentiment, providing a comprehensive view of carbon price formation mechanisms.
