Version numérique PDF

We discuss the economics of high-frequency trading (hereafter HFT), survey empirical findings and offer policy recommendations. HFT involves high speed connections to exchanges, computerized trading, and very short-term positions. Beyond these common features, HFT strategies are heterogeneous. They can involve market-making, directional trade, arbitrage, and possibly manipulation. The empirical literature finds that HFT market-making is profitable only because of favorable exchange fees. HFT market orders predict future short-term market movements and correspondingly earn profits, at the expense of other market participants. Yet, there is no empirical evidence of adverse effects of HFT on liquidity. HFT could generate negative externalities, by inducing adverse selection for slower traders, or enhancing the risk of trading firms? failure waves. To cope with these problems, slow-traders? friendly market mechanisms should be available, and minimum capital requirements and stress tests should be implemented.Keywords: High Frequency Trading; Algorithmic Trading; Liquidity; Price Discovery.JEL codes: G1; G2; G14.

Auteurs :Biais Bruno
Extrait de la revue BMI 128