Testing for Co-bubble Behaviour in Financial Time Series

Seminar abstract

This Seminar presents the paper, joint with Andria Evripidou, David Harvey and Stephen Leybourne, a test to determine if two price series that each contain an explosive autoregressive regime consistent with the presence of a bubble, are related in the sense that a linear combination of them is integrated of order zero.

The paper refers to such phenomenon as "co-bubble behaviour" and purpose a test based on a stationarity testing framework. The test allows the bubble episode in one series to lead (or lag) that in the other by a number of time periods.

It is established that the asymptotic properties of the test statistics and propose a wild-bootstrap procedure for obtaining critical values that are robust to heteroskedasticity.

Simulations show that the proposed test has good finite sample size and power performance. An empirical application to detect whether co-bubble behaviour exists between the price of gold and of other precious and non-ferrous metals is presented.

Booking

This event is free to attend. There is no need to book in advance. Please feel free to bring along your friends, colleagues and classmates.

Speaker bio

Professor Rob Sollis is a Professor in Financial Econometrics at Newcastle University.

He has degrees from the University of Liverpool, University College London, and the University of Nottingham. His PhD focused on detecting and modelling structural changes in financial and economic time series and was written under the supervision of Professor Steve Leybourne and the late Professor Paul Newbold.

His current research focuses mainly on forecasting financial time series and developing econometric methods to detect asset price bubbles. He also has a general interest in computer programming in finance (particularly with MATLAB and more recently Python)