Jumps in High-Frequency Data: Spurious Detections, Dynamics, and News

We propose a technique to avoid spurious detections of jumps in highfrequencydata via an explicit thresholding on available test statistics. Weprove that it eliminates asymptotically all spurious detections. MonteCarlo results show that it performs also well in finite samples. In DowJones stocks, spurious detections represent up to 50% of the jumps detectedinitially between 2006 and 2008. For the majority of stocks, jumpsdo not cluster in time and no cojump affects all stocks simultaneously,suggesting jump risk is diversifiable. We relate the remaining jumps tomacroeconomic news, prescheduled company-specific announcements, andstories from news agencies which include a variety of unscheduled anduncategarized events. The majority of news do not cause jumps. Oneexception are share buybacks announcements. Fed rate news have an importantimpact but rarely cause jumps. Another finding is that 60% ofjumps occur without any news event. For one third of the jumps with nonews we observe an unusual behavior in the volume of transactions. Hence,liquidity pressures are probably another important factor of jumps.[...]