A Markov switching model of the conditional volatility of crude oil futures prices

This paper examines the temporal behaviour of volatility of daily returns on crude oil futures using a generalised regime switching model that allows for abrupt changes in mean and variance, GARCH dynamics, basis-driven time-varying transition probabilities and conditional leptokurtosis. This flexible model enables us to capture many complex features of conditional volatility within a relatively parsimonious set-up. We show that regime shifts are clearly present in the data and dominate GARCH effects. Within the high volatility state, a negative basis is more likely to increase regime persistence than a positive basis, a finding which is consistent with previous empirical research on the theory of storage, e.g. Fama and French (1988a,b) and Ng and Pirrong (1994). The volatility regimes identified by our model correlate well with major events affecting supply and demand for oil. Out-of-sample tests indicate that the regime switching model performs noticeably better than non-switching models regardless of evaluation criteria. We conclude that regime switching models provide a useful framework for the financial historian interested in studying factors behind the evolution of volatility and to oil futures traders interested short-term volatility forecasts.