Are algorithmic trades informed? An Empirical analysis of algorithmic trading around earnings announcements

Alex Frino, Tina Viljoen, George H. K. Wang, Joakim Westerholm, Hui Zheng

Research output: Chapter in Book/Report/Conference proceedingConference proceeding contributionpeer-review

Abstract

This study examines the impact of corporate earnings announcements on trading activity and speed of price adjustment, analyzing algorithmic and non–algorithmic trades during the immediate period pre– and post– corporate earnings announcements. We confirm that algorithms react faster and more correctly to announcements than non–algorithmic traders. During the initial surge in trading activity in the first 90 seconds after the announcement, algorithms time their trades better than non–algorithmic traders, hence algorithms tend to be profitable, while non–algorithmic traders make losing trades over the same time period. During the pre announcement period, non–algorithmic volume imbalance leads algorithmic volume imbalance, however, in the post announcement period, the direction of the lead–lag relationship is exactly reversed. Our results suggest that as algorithms are the fastest traders, their trading accelerates the information incorporation process.
Original languageEnglish
Title of host publicationProceedings of the the fifth annual Asian Conference of the Financial Management Association International (FMA)
Place of PublicationUnited States
PublisherFinancial Management Association International
Pages1-38
Number of pages38
Publication statusPublished - 2013
Externally publishedYes
EventFMA Asian Conference (5th : 2013) - Shanghai, China
Duration: 17 Apr 201319 Apr 2013

Conference

ConferenceFMA Asian Conference (5th : 2013)
CityShanghai, China
Period17/04/1319/04/13

Keywords

  • Algorithmic Trading
  • Earnings Announcements
  • Market Efficiency

Fingerprint Dive into the research topics of 'Are algorithmic trades informed? An Empirical analysis of algorithmic trading around earnings announcements'. Together they form a unique fingerprint.

Cite this