High-frequency algorithmic trading (HFT) is responsible for many correlations
Comment of the Day

August 06 2012

Commentary by David Fuller

High-frequency algorithmic trading (HFT) is responsible for many correlations

Visitors to this site know a great deal about high-frequency trading, or if not, they can read about it by using the Search facility shown upper left, fourth item down. Here are two key paragraphs from an article on the subject: "Ruthless efficiency of high-speed trading might not be worth the costs", by Floyd Norris for the NYT and IHT last week:
High-frequency trading got its start in stocks but it has spread to commodities and changed the trading patterns there. The computers search for correlations between different markets, and try to take advantage of them before others can do so. The result is that markets that once tended to trade in the same direction much of the time now almost always do, particularly when one market starts to move.

That was on display Thursday. As stock markets fell in reaction to disappointing news from the European Central Bank, oil and copper prices immediately moved the same way. Bond prices rose. It was less that anyone had grown negative on copper and more that computers calculated that the two markets would move together, and assured that they did.

David Fuller's view I am not aware that the terms: 'risk-on' and 'risk-off' existed before HFT took over most of the daily trading volume in an increasing number of instruments a few years ago. I do not know how you feel about this but I am concerned about a market environment where the most profitable form of trading occurs when machines respond to a programmed signal and then drive most other markets in a risk-on or risk-off manner, before closing out positions at the end of the day.

The number of very short-term melt-ups or meltdowns is increasing, as I have illustrated with intraday charts on occasion, and we get the occasional derailing as seen recently with Knight Capital Group.

Back to top