Detecting speculative bubbles in GARCH processes: A time-series based test for predicting explosive behavior

This paper provides further evidence for the usefulness of the speculative bubble test proposed by Franses (2014). The test is proven to be effective in predicting bubbles in high-frequency data. Three real world examples as well as Monte Carlo simulations show that the test gives signals when a speculative bubble occurs. Furthermore, this paper has adapted the recursive residuals method for GARCH processes. For stable processes, this method is even more accurate, but it cannot be used for detecting explosive behavior.