The Truth about High Frequency Trading

By: William Fenick

High Frequency Trading (HFT) is making headlines following the launch of “Flash Boys,” a new book from Michael Lewis that serves to illustrate how firms specializing in HFT are unfairly profiting from the practice. Subsequently, the New York Attorney General's office sent subpoenas to a number of firms, bringing the practice under further scrutiny and leaving many wondering whether those engaging in HFT activity will face criminal charges.

For many of us who’ve been in the industry for quite a few years, we can appreciate that this is not a new phenomenon.   This type of low-latency trading has been around for well over a decade, and those in the know are fully aware of it - however, perhaps it hasn’t been explained to the man on the street. This post aims to provide clarity around the practice.

The first thing to acknowledge is that, for pure high frequency trading, the physical locations of the servers are extremely important. Ultra low high frequency trading occurs where the matching engines of exchanges are sitting, e.g. the trading firm’s engine is cross-connected into the exchange’s matching engine. This can be juxtaposed with proximity high frequency trading, which occurs when trading firms are centrally situated among top exchanges and can aggregate the exchanges’ market data feeds. Interxion’s Finance Hub in London, as one example, enables proximity HFT in European equities because we host NYSE Euronext’s SFTI POP in our data centre, which sits in close physical proximity to the London Stock Exchange in the City and is 25 miles from BATS.

The goal in high frequency trading is to quickly aggregate the prices coming in from various sources, to substantially minimize the trader’s algorithm’s reaction time and execute buy or sell orders before competitors. Getting ahead of competition via getting the information first is nothing new. In fact, in the early 1900s the New York Stock Exchange used to plant a boy in front of the newspapers first thing in the morning so they could trade from the freshest news.

Through the evolution of computers and electronic trading, the time advantage has become paramount to trading.  A nice analogy to understand the speed at which high frequency traders operate is to put the times into a different configuration.  Trading firms speak about nanoseconds, microseconds and milliseconds. But if we were to apply the same methodology to a longer period of time, say one second is equivalent to a millennium.  Thus, one millisecond would be the equivalent of  1/1000 of a millennium, and subsequently 250 milliseconds equals 3 months.  We can see very clearly the advantage of arriving earlier to a venue to execute an order.  If we consider that ultra-low high frequency trading is taking place at 100 microseconds or less (that is to say 1/10 of a millisecond) – we can see where the large buy-side  investors, or indeed broker-dealers, who are most likely operating in the 50-250 millisecond range, are complaining about electronic front-running.

While the market today is so heavily dominated by electronic algorithmic trading, there are a few important questions to think about with this recent focus on high frequency trading – should exchanges offer a different order type, e.g. not first order in,  first order matched - for instance? Should the multiple Dark Pools be obligated to publish how an order is executed?  But as history continues to show us, keen companies always find a way to get a leg up on their competitors, and we don’t have any reason to think now is any different.