Extreme price moves: an INGARCH approach to model coexceedances in commodity markets

Abstract

Abstract This study presents a set of integer-valued generalised autoregressive conditional heteroskedastic models to identify possible transmission channels of joint extreme price moves (coexceedances) across a group of agricultural commodities. These models are very useful to identify factors affecting joint tail events and they are superior in terms of goodness of fit to models without autoregressive components. Emerging market demand, crude oil, exchange rate, stock market conditions and credit spread explain extreme joint returns. Psychological factors and the Monday effect play a role in affecting extreme events, while weather anomalies (El Niño and La Niña episodes) do not have explanatory power.