loading page

Market Making High Frequency Traders
  • Andreas Park
Andreas Park
University of Toronto

Corresponding Author:[email protected]

Author Profile

Abstract

We study the order submission and trading behavior of market-making high frequency traders. Using proprietary (but masked) trader-level data, we classify trader IDs as high-frequency based on fast reaction speeds. Since a single HFT firm may use multiple trader IDs, we further group traders IDs with a cluster analysis based on similarities in trading characteristics. We then classify traders as (voluntary) market-makers if their submission of passive buy and sell order volume is balanced. HFTs account for around 80% of all orders, around 65% of all orders that improve the best price and 70-80% of all orders that match the prevailing best prices. We pay particular attention to (HFT) market makers behavior after trades. Namely, we observe that their order cancellation rate is highest in the first few milliseconds after a trade compared to half or a full second after the trade, which documents the quote-fade phenomenon. Similarly, in the first few milliseconds after a trade, there is a significant increase in aggressive order submissions that are in the same direction as the trade. Taken together this behavior suggests that HFT market makers move the price right after a trade, that they trade aggressively against stale orders. Since trading fees are positive, such behavior indicates that they either aim to close positions quickly to avoid a loss or that they attempt to trade ahead of future orders. Both order the cancellation and the aggressive submission rates are positively related to the volume of the order and the directional volume that the aggressive trader has already traded on the day.