Human intervention leads to rising trading costs – GLG
US buy-side traders could be adding unnecessary costs by micromanaging trades that could be better handled by
algorithms, according to Nik Wislang, head of electronic trading at global
investment management firm GLG Partners.
“People focus too narrowly on execution sometimes,” he says.
“The typical US desk likes to use direct market access and interact with the
microstructure of the market, even splitting up child orders manually. But the
cost of trading is not just execution – that’s only a small part of the total.”
GLG was purchased by Man Group, known for its black box,
systematic trading strategies, at the end of 2010. GLG’s core trading style used to be purely based on stock selection by human beings. Yet Wislang now believes that algorithms
are indeed better suited to certain tasks than humans – and that ignoring the
impact of trader meddling at the execution level would be a serious mistake.
“When a trader decides not to use an algorithm and interacts
with the marketplace, that causes a lot of tickets and a lot of costs
downstream in settling all these smaller trades,” he says. “I’m of the view an
algo will do better than a human. Using an algo, you only have one order to the
broker, reducing the settlement costs as well as the chance of things going
Taking into account the cost of a trade in terms of
settlement, risk management, confirmation, sending trade files to administrators
and maintaining sufficient staff in the back office to deal with the trade data, might cause asset managers to view the value of their execution choices
differently, Wislang suggests.
Back-office costs have long been a contentious issue in
Europe, with post-trade costs thought to be up
to ten times the equivalent in the US. The greater cost of post-trade services in
Europe could make the kind of direct intervention and micromanagement Wislang
observes in the US prohibitively expensive this side of the Atlantic –
suggesting that European traders are already more willing to use algorithms for
basic tasks such as splitting up child orders.
“Let the machines deal with the orders that are appropriate
to them,” says Wislang. “Human traders should deal when there are factors that
require the subtleties that humans can detect, such as the intonation in the
broker’s voice over the phone. That’s when a human being can really add value
to a trade.”