Even experts sometimes mix different issues: The Economist used examples of computer trading errors to conclude that high frequency trading is harmful. The fact is: computer trading and high frequency trading are not the same thing. Computer trading errors should be addressed, like rogue human trading behaviour. "Computer Trading" and "high frequency finance" are only vehicles. We should not get rid of cars just because people are killed in car accidents. We should try to licence drivers and make cars safer.
In the article "Wait a second" on High-frequency trading, 11 August 2012, The Economist said: "high-frequency trading may make markets less volatile in normal times". They went on to say that "it may add to the turbulence at the worst possible moment". The Economist established the causal relationship between "high frequency trading" and "adding to the turbulence" with examples of computer trading errors, such as Knight Capital's recent loss of $440 million as a result of them sending errant buy-and-sell orders. The 2010 flash crash is also used frequently to show that high frequency trading is dangerous due to potential manipulation.
The Economist's view is shared by many. The argument appears to be:
Computer trading can lead to problems fast. But computer programs are just implementing algorithms designed by people. If computer trading should cause any problem, the problem is caused by the algorithms that the people use. Computerisation itself does not cause the problem. The trader could be using the same algorithm in manual trading, causing the same problem.
Speed in computer trading just amplifies the problems. Problems can be built up in low speed trading. The banking crisis in 2007-08 was built up over decades, though problems were not spotted until the system collapsed.
Are programs more unreliable than human traders? Programming bugs have led to damaging orders by Knight Capital. But if computer trading is dangerous, so is human trading! At least computer programs don't have personal interests. Computer programs can be debugged, not human traders. Nick Leeson bankrupted the Baring Bank in 1995.
If reliability is the problem, which it is, then more robust testing is the solution. Lawyers, architects, pharmacists, car drivers all have to be licensed. Why not trading programs?
What's wrong with requiring all trading programs to go through a robust test in an artificial market which resembles all adverse market conditions that one can think of? Then only licence those whose behaviour look acceptable to trade in volume. This is by no means fool-proof, but neither do driving licences prevent accidents.
High frequency finance is not the cause of volatility. Lack of liquidity is. High frequency trading are likely to improve liquidity, which could reduce volatility. Reducing speed itself could cause other problems. If waiting a second is good, then why shouldn't we allow trading per minute? or trading per hour?
[End]
Related: "High-frequency trading harms the economy"
All Rights Reserved