The Risks of High-Frequency Trading

[MP-AI-BOT]

Between 2:30 and 3 p.m. on May 6, 2010, the Dow dropped 7 percent before partially restoring itself by the closing bell. In the days following that sudden market shock, regulators have investigated possible causes, looking at a single trade in the Standard & Poor’s e-mini futures contract and other causes for the brief panic. The plunge has telescoped focus on the role of “circuit breakers” and the way high-frequency traders function in response to irregular use of those safety mechanisms.

“I think that market-wide coordination of regulatory mechanisms such as circuit breakers is a very positive thing. The lack of such coordination seems to have had a significantly detrimental effect,” says Professor Ciamac Moallemi, who has studied behavior in financial markets. In his recent research, he created a quantitative model to value latency or the delay between decision and trade execution, finding that the higher frequency of trading, the more impact latency has on transaction costs. (Read more about this research in Ideas at Work.)

Moallemi said the events of May 6 raised a number of unanswered questions, including:

1. Did buy-side firms employing algorithmic trading strategies contribute to the crash? Many mutual funds, pension funds, etc., try to efficiently buy or sell large positions via computerized strategies that buy or sell at a predetermined rate over the course of the day despite market conditions. Such strategies often trade at a faster rate in periods of high volume. Did these strategies accelerate selling just as the market was crashing?
2. High-frequency liquidity providers implement computerized market-making strategies based on the statistical analysis of markets. In periods of market anomaly, where historical statistical relationships may not hold, these traders may withdraw from the market and hence remove liquidity at times when it is most needed. To what extent did high-frequency liquidity providers withdraw liquidity from the market immediately prior to the crash?
3. Given that there does not seem to be any trading “error” involved in the crash, is it a wise policy for the exchanges to cancel trades that occurred at anomalously low prices? This would seem to destroy any incentive for investors to provide liquidity by buying during a crash. What incentives can be created for liquidity provision in turbulent times?

“Equity markets have changed dramatically in recent years, with the proliferation of electronic trading and the decentralization of trading across many new venues,” Moallemi adds. “While these changes have offered investors many benefits, there may be unintended consequences such as the momentary breakdown that occurred on May 6. This event highlights exactly how little we know about the complex and highly interdependent systems that constitute the market.”

Image credit: Google Finance



More...