Forecasting carbon phase II moments for option pricing and risk management applications using stochastic volatility models

This paper applies the General Scientific Model methodology of Gallant and McCulloch implementing MCMC simulation methodologies to directly specify a multifactor stochastic volatility model for the NASDAQ OMX carbon front December forward contracts (phase II). From the LIBOR models in interest rate theory the idea is to construct a dynamics of the traded contracts. The main objective is to structure a scientific model specifying traded carbon contracts matching the observed volatility term structure as having its own stochastic process. Stochastic volatility is the main way time-varying volatility is modelled in financial markets. Appropriate model descriptions broaden the applications into derivative pricing purposes, risk management and asset allocation. The paper reports option prices, risk and portfolio measures, conditional one-step-ahead moments, particle filtering for one-step-ahead conditional volatility, conditional variance functions for evaluation of shocks, analysis of multi-step-ahead dynamics, and conditional persistence. The analysis adds market insight and enables forecasts to be made, thus building up methodologies for developing valid scientific models for carbon market applications.