Human Bias in Algorithmic Trading
28 Pages Posted: 8 Jan 2014
Date Written: December 30, 2013
Abstract
This paper documents a stark periodicity in intraday volume and in the number of trades. We find activity in both variables spikes by about 20% at regular intervals of 5 or 10 minutes throughout the trading day. We argue that this activity is the result of algorithmic trading influenced by human traders/programmers’ behavioral bias to transact on round time marks. An alternative explanation, that algorithms choose to concentrate their trades in time to take advantage of lower costs or to protect themselves from better informed traders, is not supported.
Keywords: Algorithmic Trading, Microstructure, Behavioral Bias
JEL Classification: G10, G19
Suggested Citation: Suggested Citation