Time-variations in commodity price jumps

• The number of price jumps in commodity markets exhibits seasonality.
• We propose a stochastic volatility jump–diffusion model that captures this seasonality.
• We estimate the model using the Markov Chain Monte Carlo methodology for four markets.
• The results show that our model is superior to models with constant jump frequency.

In this paper, we study jumps in commodity prices. Unlike assumed in existing models of commodity price dynamics, a simple analysis of the data reveals that the probability of tail events is not constant but depends on the time of the year, i.e. exhibits seasonality. We propose a stochastic volatility jump–diffusion model to capture this seasonal variation. Applying the Markov Chain Monte Carlo (MCMC) methodology, we estimate our model using 20 years of futures data from four different commodity markets. We find strong statistical evidence to suggest that our model with seasonal jump intensity outperforms models featuring a constant jump intensity. To demonstrate the practical relevance of our findings, we show that our model typically improves Value-at-Risk (VaR) forecasts.