Long-Term Swings and Seasonality in Energy Markets

This paper introduces a continuous-time model for commodity pricing under the assumption that logged prices converge to a mean level that experiences smooth, periodic fluctuations over long periods of time. Our model incorporates that assumption by modelling the mean reversion level through a Fourier series. To validate the model, we perform an empirical study of futures prices on Natural Gas, Crude Oil, and Heating Oil. We provide evidence that such long-term fluctuations are present in the price of these energy commodities, possibly together with standard seasonal and cyclical components. We analyse the empirical performance of our pricing model versus two alternative competitors, namely, those proposed in Schwartz (1997) and Lucia and Schwartz (2002). Our findings show that our model outperforms both benchmarks, providing a simple and powerful tool for portfolio management, risk management and derivative pricing.