Modeling volatility and risk in the CO2 emissions market

Abstract

In this master thesis various GARCH models for volatility and Value at Risk for ECX CO2 futures contracts have been examined. This is of great importance for market participants such as CO2 emitting companies, traders and risk managers. The first chapters of the thesis introduce the carbon market and the different futures contracts, which are later on tested. From the descriptive analysis, we found the futures contracts to be suitable for GARCH modeling, and we also found which specifications we had to include to model the returns and Value at Risk. The applied methods are different univariate and multivariate GARCH models, which have been tested based on evaluation criteria such as Akaike, Log likelihood and numbers of significant parameters. Problem for discussion: Analysis of time-varying volatility in the carbon market and timevarying binary correlation between the carbon market and other energy markets (electricity-, oil-, gas-, and coal market). There are several different CO2 futures contacts being traded, but for this analysis we found the 2009 and 2010 EUA futures contracts to be most interesting and suitable. To find good models for volatility and Value at Risk, we compared the GARCH results with the results from the descriptive statistics of raw data. The estimation of Value at Risk is tested for each contract, with the Kupiec test. Monte Carlo simulations have also been implemented to support our findings. The datasets have been tested within the whole trading period, also after the structural break in 2006. Finally we have presented analysis of CO2 futures contracts in bivariate portfolios with other different energy commodities. The main focus of this thesis has been to find suitable GARCH models for volatility of returns and Value at Risk.