Understanding volatility dynamics in the EU-ETS market

We study the short-term price behavior of Phase 2 EU emission allowances. We model returns and volatility dynamics, and we demonstrate that a standard ARMAX-GARCH framework is inadequate for this modeling and that the gaussianity assumption is rejected due to a number of outliers. To improve the fitness of the model, we combine the underlying price process with an additive stochastic jump process. We improve the model's performance by introducing a time-varying jump probability that is explained by two variables: the daily relative change in the volume of transactions and the European Commission's announcements regarding the supply of permits. We show that (i) sharp increases in volume have led to increased volatility during the April 2005-December 2007 period but not for the period beginning in January 2008, and (ii) announcements induce jumps in the process that tend to increase volatility across both periods. Thus, authorities face a trade off between disseminating information effectively and promoting market stability.