Abstract
This paper considers the problem of detecting short explosive bubbles in financial data. Based on the backward sup Dickey-Fuller (BSDF), we propose a modified version of BSDF, namely mBSDF. We define the dating statistics by adding a modified term to the BSDF, enhancing the bubble emergence signal. We demonstrate the asymptotic distribution and the bubble duration estimates. A series of Monte Carlo simulations show that mBSDF significantly outperforms BSDF for shorter bubbles, and mBSDF can improve the bubble detection rate by up to 22.7%. As an empirical application, we apply the methods to the Brent crude oil futures price and the results confirm that mBSDF detects the latest oil spike shocked by the Russia-Ukraine conflict and almost all periods of explosive bubbles the oil market has experienced in recent years in contrast to BSDF, which further supports the superiority of mBSDF.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
