The rise of the high-frequency trader in the US and Europe has left many traditional buy-side market participants wondering whether their execution performance is suffering from trading with counterparts that have superior technology.
One of the main concerns is that by reaching the exchange faster than regular traders, high-frequency firms gain an unfair advantage and can leave institutional investors a distant second in the race to optimise liquidity. Given the level of high-frequency trading – currently estimated to account for 60-70% of trading volume in the US and 35-40% in Europe – the potential ramifications are significant.
But first we must understand the firepower that long-only and other non-high-frequency firms are up against.
While the details of an individual firm’s high-frequency strategies are closely-kept secrets, many centre on automated or virtual market making, i.e. spotting opportunities to make markets by posting multiple orders – usually comprising a small number of shares – to buy and sell the same stock, hoping to profit from capturing the spread as well as the rebates offered by trading venues for resting liquidity on them.
To do this effectively and win the race to exploit these opportunities, a low-latency infrastructure is key.
High-frequency traders need to be co-located, i.e. place their trading engines as close to the exchange’s data centre as possible, thereby reducing the time it takes to reach the exchange’s matching engine. They also need a fast and reliable high-capacity execution platform so they can send ‘bursts’ – rather than constant streams – of orders to the exchange as quickly as possible.
To support bursts of orders, some high-frequency firms may need to balance their orders across multiple connection ports to a single exchange so they can handle the high volume of messages sent and received. If a stock price moves adversely, high-frequency traders also need the ability to cancel orders as quickly as possible.
According to Valerie Bannert-Thurner, executive director, FTEN Europe, a supplier of high-frequency execution and risk solutions for both the buy- and sell-side, the latency incurred by a high-frequency firm’s execution infrastructure can vary depending on its measurement approach, making straight comparisons difficult. For example, firms may or may not include the time added by risk checks, real-time management of open orders, or the symbology and message translations that are needed across venues.
“For the execution piece, latency can vary from under 100 microseconds to a millisecond. Consistency of execution performance over time and the ability to handle short ‘micro bursts’ are equally important. For example, some strategies need to burst 100 messages in one millisecond while other strategies need to send more than 250,000 orders in the first four seconds of market opening,” says Bannert-Thurner.
High-frequency traders also require the freshest and fastest market data to spot market-making opportunities and feed their algorithms. According to Bannert-Thurner, most market data can be delivered in below 100 microseconds to virtual market makers, but again, this depends on how the data is delivered.
“The speed at which market data is disseminated can depend on whether it is raw or normalised, the depth of book you want to use, and in Europe whether you have to consolidate order books across multiple exchanges,” she says.
Algorithms also need to be specifically tuned to high-frequency strategies so that they can spot opportunities from the market data and post orders on trading venues before any expected flurry of market activity.
Unlike high-frequency firms, most buy-side market participants simply do not trade in and out of positions on an intraday basis and so do not need the fastest possible market data delivery or co-location capabilities. Nevertheless latency is a factor for institutional investors because speed to market is now so fundamental to the ability of executing brokers to achieve best execution.
If they are dealing in markets where other market participants are able to trade many times faster, should long-only firms just reap the available benefits – such as higher levels of liquidity – or should they have concerns about sharing the field of play with counterparts that have vastly superior equipment?
Mark Howarth, CEO, Chi-X Europe, says the diversity to be found in his trading venue benefits all.
“High-frequency traders are an important part of our marketplace. As electronic market makers they have a different objective to other participants on our platform and can sell when everyone wants to buy and vice versa, and ensure that trades happen,” he comments. “High-frequency traders have evolved from market makers who are necessary in maintaining efficient and orderly markets.”
To vote in the new poll on high-frequency trading, please click here.