Stress-testing for portfolios of commodity futures

In this paper, we perform stress-testing for a portfolio of commodity futures which mimics the dynamics of the DJ–UBS index. We identify extreme events that impacted commodity prices over time and look at correlation structures in a dynamic way, with copula functions. In line with Basel III financial regulations, we derive baseline, historical, and hybrid scenarios and discussed their advantages and shortfalls. We find that the financialization of commodity markets led to an increase in correlations and in the probability for joint extremes. However, we identify structural breaks in commodity markets that temporarily led to a breakdown of expected statistical patterns and of traditional dependence structures among commodities. This fact shows the need for forward-looking stress testing techniques, like hybrid and hypothetical scenarios, as encouraged by financial regulators.