By: Michael Fowlkes
A new study puts data behind the contention that high-frequency trading is bad for regular investors.
A new study conducted at the University of Michigan took a look at a type of high-frequency trading known as high-volume arbitrage activity. This is the practice of finding and trading on price differences between different exchanges.
Under Regulation NMS, prices across exchanges are supposed to be the same. To do this, a central system, the Security Information Processor, compiles information from the various exchanges and calculates the National Best Bid and Offer prices.
High-frequency traders who have ultra-fast computers and network connections can take advantage of the time it takes to calculate those best bid and offer prices and trade on the difference in price between venues before the Security Information Processor can publish its results.
Consider a stock being sold at one price on the New York Stock Exchange and a slightly different price DirectEdge. In the simplest terms, high-frequency traders could buy a security on one exchange and sell it on another exchange quicker than both exchanges adjust their bid/ask prices, making an easy profit for the traders, but coming at a cost for normal traders.
Arbitrage traders contend that what they are doing adds liquidity to the markets, which is beneficial to average investors, but the study reached a different conclusion. It found that when arbitrage traders get involved in a security, the bid/ask spreads on securities tend to widen, which means that neither buyers nor sellers are able to get the best price for their trades.
The reason is that it takes time, albeit a small fraction of a second, for bid/ask prices to be computed. It may be a very small amount of time, but in those few milliseconds of time, high-frequency traders are able to take advantage of the price discrepancy.
The effect that this has on each trade the average investor makes is minimal, but over time the pennies add up. If you are the sort of investor that makes just a couple of trades a year, then the impact is negligible, but for active traders it can result in a material impact on returns over the course of time.
While some changes need to be made, it is hard to imagine a system that will ever completely result in a fair playing field. There will always be people with enough computing power and enough knowledge of the system to find holes, and take advantage of them.