HFT – High Frequency Trading has had a rapid development during the past few years and recently has created some controversy in media and among politicians. The critics say it is an unfair practice and that it risks destabilizing the markets. The defenders claim it adds liquidity and makes market more efficient.
There is not any clear definition of HFT, but most tend to agree with the following characterization: A proprietary trading strategy that repetitively executes many relative small trades where the trades has a holding period from a maximum of a few minutes down to milliseconds.
There are no good statistics of how much of the trading volume is generated by HFT. In the U.S., the estimates vary from 40% of the equity volume on the low side to about 70% on the high side. Whatever the number, it is significant.
So is HFT a new predatory practice or does it benefit the market?
I argue that there is nothing new with HTF, it simply does what has always been done in trading, only faster. Moreover, striving to do things faster in order to gain a temporary competitive advantage has also always been part of the evolution of markets.
HFT strategies can be divided into four categories:
1. Arbitrage – cross-market, cross-asset class, pairs, index, statistical, volatility, etc.
2. Market Making – gaining from the (minimal) spread and today, more importantly, from passive order rebates
3. Technical/Pattern Analysis – striving to predict price movements by analyzing volumes, trends, volatility, momentum, data patterns, etc.
4. Order Detection – identifying whether there are large orders being worked in the market and trading ahead of them.
The first three categories are commonly accepted as beneficial to the market. The only strategy that could be questioned is the fourth. Where the Order Detection strategy is passive, in the sense that it is only based on actual market data, it is not questionable. In this case it should rather be characterized as a Technical/Pattern Analysis. However, if the strategy actively probes the market with tiny orders to gain non-public information on orders being worked in the market, and that information is used to front-run or purposely mislead others to trade, then the activity in most developed markets would be illegal also under current laws and regulation.
My conclusion is that HFT is nothing new per se, not even the questionable practices in category four. It simply does what always has been done on markets, only faster. Most of the effects of HFT are clearly positive for the market. It adds another source of liquidity and uses classic strategies, commonly accepted as making markets more efficient. However, when an old strategy is applied in a new way, regulators and exchanges need to review and align regulation, rules and systems.
The exchanges and regulators need to:
I. Clarify rules targeting the evolved form of front-running and intentionally misleading trading strategies.
II. Update market surveillance systems with pattern recognition and anti-gaming algorithms to alert on suspected trade abuse.
III. Update central matching systems to keep up with the increased volumes and introduce throttling mechanisms to avoid overload.
IV. Introduce rules for HFT systems, for example maximum order to trade ratio and minimum order size.
V. Charge the participants for advantageous market access. Allowing a participant to co-locate his trading system and use relatively large amount of exchange processing power to execute his business strategy is a clear benefit. This should not only render a fee, but also come with some obligations, just as market makers usually have certain obligations.