Projecting and Forecasting Stochastic Volatility Characteristics for the Nasdaq OMX Nordic/Baltic Financial Electricity Markets

Abstract

This paper builds, implements and interprets multifactor stochastic volatility models for the Nordic/Baltic electricity markets. The main objective is step ahead volatility projections followed by volatility forecasts and market implications. From conditional moments and a long-simulated state vector realization the paper establishes a functional form of the conditional distribution (non-linear Kalman Filter), which is evaluated on observed data convenient for step ahead volatility projections. For the front year and quarter financial electricity contracts, the SV model report one persistent and slowly moving factor and one choppy mean reverting factor. From these factors, static volatility forecasts using optimal and generous lags, report a Theil covariance portion well above 98% for the front year and 92% for the front quarter contracts.