New evidence for explosive behavior of commodity prices

Over the past decade, the academic literature has engaged in a lively discussion about speculative bubbles in commodity markets. A number of papers have empirically investigated explosive behavior, albeit employing different econometric approaches that examine univariate time series for the existence of bubbles. Largely, these individual tests have been applied on a set of univariate time series such as stock indices or commodity prices, ignoring the multiple testing nature of the problem. Departing from there, we examine explosive behavior in commodity futures markets, by employing a panel-data set containing the ten most liquid agricultural futures contracts traded in the US to conduct the Generalised Sub-ADF test by Phillips et al. (2015). We aggregate individual test-results using a stagewise rejective multiple test procedure. Overall, our results show evidence for several periods of explosive behavior in the markets for wheat, cattle, cocoa, coffee and cotton over the past 35 years. Our paper is the first to employ this approach on a panel data set, thereby solving econometric shortcomings of previously published work in the extant literature.