How Do High-Frequency Traders Make Their Money? Prediction, Shared Knowledge and Competition in Automated Trading

This paper will examine the practices of high-frequency trading (HFT) of US equities, drawing upon 20 interviews with founders, managers and traders in HFT firms, and also upon a wider set of 44 interviews with exchange staff, suppliers of algorithms, etc.  The paper will in addition cross-check interview-based conclusions against the findings of recent work on HFT in financial economics.

After a discussion of the methodological difficulties of qualitative research on HFT, the paper will focus on the role in HFT of price prediction based on phenomena such as order-book imbalances and ‘futures lag’ (the tendency for information relevant to the overall stock market to be impounded first in index futures). Predictions of this kind are typically crystallised by HFTs in notions of ‘fair value’ or ‘theoretical value’, and the main techniques of prediction are shared knowledge amongst competing firms: there are ‘secrets but no secrets’ as an interviewee put it. Given that, competition largely takes two forms: technological pursuit of speed; and sophisticated use of order types, the latter structured largely by regulatory requirements on exchanges, especially the prohibition on price quotations that ‘lock’ another market (eg a bid that has the same price as an offer on a different market).

Although for reasons of time they cannot be tackled in the paper, a number of other topics are left for discussion, in particular: the question of ‘algo sniffing’ and the unequal moral valence of different forms of prediction; and the question of how profitable HFT actually is. Although it will be largely implicit in the paper, relevant social-science literature drawn on will include Knorr Cetina and Preda on the role of technology in financial markets, and Burchell on competition under neoliberalism. 

The Business History Seminar is organised by the Business History Centre and has been made possible by financial support from the Erasmus Research Institute of Management (ERIM) and the Erasmus School of History, Culture and Communication.