this study considers a wild bootstrap approach to existent tests for bubble identification to account for the possibility of heteroskdastic residuals that can attributed to breaks in volatility in order to study market efficiency in the commodity markets.
The simulation results suggest that the wild bootstrap test offers improved size control while offering significant power gains as the series of interest has been modelled as a moving average process rather than a unit root process since under the null hypothesis of market efficiency the difference between the future spot price and the futures contract price will be a moving average process of order specified by the length of the futures contracts.
In the empirical application, the proposed wild bootstrap test identifies periods of market inefficiency prior to a bubble episode that existent bubble detection tests do not detect during two periods of oil crises, acting as an early warning mechanism of non-stationary behaviour in the market that could, potentially, lead to a bubble episode.
This online seminar is free to attend and there is no need to book in advance.
Dr Ioannis Korkos is a Lecturer in Economics at the University of Essex.
His research examines explosive time series and macroprudential policy.