Identifying events in financial time series - A new approach with bipower variation

- An extension of the standard bipower variation based jump detection test is proposed.
- The new test significantly outperforms the standard method in a Monte Carlo setting.
- Empirical analysis of a million trading days shows improved detection capability.

We present a statistical test to identify significant events in financial price time series. In contrast to “jumps,” we define “events” as non-instantaneous, but nevertheless unusually fast and large, price changes. We show that non-parametric tests perform badly in detecting events so defined. We propose a new approach to explore the dependence of jump detection statistics on the sampling method used and find that our method improves the event detection rate of the standard test by a factor of three.