High Frequency Trading: Do machines make fewer mistakes than humans?

High frequency trading (HFT) has been cited as the cause for increased volatility and systemic risk. However, HFT firms counter that it has brought down the cost of trading for retail and institutional investors
 
High frequency trading (HFT), or trading by machines based on complex algorithms, has increased exponentially over the past few years. According to a recent analysis only 16% of the buying and 13% of the selling on US exchanges is actually done by people. Ten years ago it was 35% and 25%, respectively. High frequency trading used to occur in milliseconds (it takes you 300-400 of these to blink), but the new hardware has increased that to microseconds, or millionths of a second. These trades are not only fast, they are also short. The holding period for a particular position if there is one is for only a few moments. This type of trading is supposed to be good for the markets. Academic research shows that since HFT has been introduced, there has been increased liquidity as measured by shrinking bid-ask spreads. So this wonderful technology really works. Except when it doesn’t.

Last year Knight Capital, the largest independent wholesale market maker in the US, was showered with awards for its pre-eminence in high-speed trading systems. This all changed in a mere 45 minutes. The company had developed a new piece of software. But it didn’t work. It sent out numerous orders for stocks and then buying them almost instantaneously, guaranteeing a loss based on the difference in price. The machine went on a buying spree that no trophy wife could even imagine. It purchased $7 billion worth of securities, an amount way beyond the limits allowed by the regulators. When the mistake was discovered, Knight’s traders stepped in and started selling. By the end of the day they were able to reduce the total to a mere $4.6 billion, but it was still too much. Fortunately for the firm, it had a saviour. Goldman Sachs kindly stepped in and bought the remaining shares for a discount of 5% to market value. Between the losses from the sales and the discount to Goldman, Knight had lost a total of $440 million. The only thing that enabled Knight to open the next day was a rescue package from six financial firms in exchange for securities that can be converted into 73% of the firm.

The story of Knight Capital would be bad enough except that it might have had greater repercussions. Knight Capital executes trades for some very large retail trading firms including Etrade, Fidelity, Scottrade, TD Ameritrade, and Vanguard. They use Knight Capital because it is cheaper than sending the trade directly to the New York Stock Exchange. A panic among retail investors could have been disastrous. It would demolish any remaining trust Americans have in their markets.

The other problem with the so called ‘Knightmare’ is that it is not a unique event. Most investors remember the Flash Crash on 6 May 2010 when the American equity markets dropped 10% in a few minutes. Technical glitches put a damper on the much-hyped debut of Facebook and dropped the shares of exchange operator BATS Global Markets to a fraction of the offering price within seconds of its IPO. According to one technology expert miniature flash crashes happen all the time.

Naturally as with any new technology there have been criticisms. HFT has been cited as the cause for increased volatility and systemic risk. Institutional investors are concerned with front running. Other problems include manipulative practices with rather esoteric names such as ‘spoofing’ and “quote stuffing”. There is the problem that when it is really needed, the liquidity dries up. The HFT firms counter that it has brought the cost of a trade for a retail investor down from $20 to $8 and a penny a share for institutional investors. But the central question is whether the present system is worse than what went before? Do the machines make fewer mistakes than humans? Of course with the controversy and the problems have come the regulators and the United States Congress, who are all in the process of investigating and, hopefully, creating solutions.

But a major meltdown on developed markets is not what worries me. Markets in the US are generally well regulated, transparent and liquid. The problem is that the vogue for high-speed trading has spread to emerging markets. Rapid electronic trading is now taking place in Brazil, Thailand, Turkey, Hong Kong, India and now China. Some are trying to protect their exchanges. The local regulators in India require approval of all algorithms before they are used. Hong Kong mandates that they be tested annually. Still these are small markets with inexperienced underfunded regulators. They are often dominated by a few players, who routinely engage in corrupt practices. These exchanges are not only subject to technical glitches, but outright fraud. When the debacle occurs it will affect far more that just one company.

You may also want to read: SEBI unveils norms to check systemic risk of algo trading

(William Gamble is president of Emerging Market Strategies. An international lawyer and economist, he developed his theories beginning with his first hand experience and business dealings in the Russia starting in 1993. Mr Gamble holds two graduate law degrees. He was educated at Institute D'Etudes Politique, Trinity College, University of Miami School of Law, and University of Virginia Darden Graduate School of Business Administration. He was a member of the bar in three states, over four different federal courts and has spoken four languages. Mr Gamble can be contacted at [email protected] or [email protected].)

Comments
Ashith Kampani
1 decade ago
Dear William,

Quoting a para from your article

"Rapid electronic trading is now taking place in Brazil, Thailand, Turkey, Hong Kong, India and now China. Some are trying to protect their exchanges. The local regulators in India require approval of all algorithms before they are used. Hong Kong mandates that they be tested annually. Still these are small markets with inexperienced underfunded regulators. They are often dominated by a few players, who routinely engage in corrupt practices. These exchanges are not only subject to technical glitches, but outright fraud. When the debacle occurs it will affect far more that just one company."

Need to get some clarity on a specific comment from the above para "They are often dominated by a few players, who routinely engage in corrupt practices."

In my view the emerging markets are evolving and within EM India seems to be doing its bit as regulator is pro-active but the above comment does not seem to fit in the Indian context as I am unable to comment on other EM's

Appreciate your response ..... Ashith Kampani
Array
Free Helpline
Legal Credit
Feedback