Robot trading on stock exchanges needs to be regulated

14 June 2012 21:50

(updated: 7 March 2016 21:50)

Robot trading on stock exchanges needs to be regulated

Lightning-fast share trading by stock trading robots is increasing strongly on the world's stock exchanges. Now, Professor of Finance Terje Lensberg believes it is time to consider the socio-economic costs of a millisecond.

Because it is all about the milliseconds. High frequency trading, popularly called stock trading robots, is the stock exchange equivalent of Formula 1. Here, it is all about speed and the best technology money can buy.

'Stock trading robots in themselves are nothing new. They have been used for years to carry our routine tasks that machines can do more efficiently than people can. What's new is the speed involved, which has increased hugely in the last few years,' says Terje Lensberg, Professor of Finance at the Norwegian School of Economics.

He is conducting research on the evolution of behaviour in financial markets and has himself engaged in algorithmic trading, another term for robot trading, for the Government Pension Fund of Norway and Storebrand.

Manipulation or a smokescreen?

This spring, the financial press has written a lot about high frequency trading, both about how stock trading robots are 'playing havoc' with Oslo Børs, and how they manipulate the market.

'Obscuring what you are doing is not the same as manipulating the market,' Lensberg comments.

Criticism has been raised of the way in which the new type of stock trading robots place buy or sell orders, then retract them so quickly that market players fail to latch on to what is happening. On an ordinary day's trading, as many as 50% of all orders can be retracted in less than a second, without a single trade having taken place.

'Manipulation can be one motive, but it could just as easily be a strategy to protect oneself against market players who want to exploit attempts at high volume trades. By quickly retracting an order, such players do not have time to exploit the position,' the finance professor explains. He believes that we should accept that investors wish to safeguard themselves against being tricked by the market.

The value of a millisecond

Lensberg believes that the debate is dominated by those who feel that their business model is being threatened by stock trading robots. He calls for a more nuanced debate, and believes that high frequency trading has both advantages and disadvantages.

Less incorrect pricing in the stock market, because the stock trading robots will quickly take advantage of it, and greater liquidity are two possible advantages of this type of share trading.

'However, much more research is needed before we can fully understand the consequences,' Lensberg emphasises.

Lightning fast stock trading robots are good business for those who use them, but the effect on the market as a whole is more uncertain.

'The question is how socio-economically profitable it is to spend so much money just to make it possible to trade a fraction of a second faster. Are the advantages great enough to weigh up for the costs this enormous use of resources involves?' Lensberg asks, and adds:

'Intuitively, I think the answer is no. I believe the focus on speed has gone too far.'

A direct line to the market

Important questions about regulation have followed in the wake of the increase in high frequency trading. Because stock trading robots fighting over milliseconds do not have the time to go via the stockbrokers, they have direct lines to the market. This worries the regulators, both in the EU and in Norway. In 2014, the EU will issue new regulations relating to share trading in Europe, which may put a stop to robot trading.

'An important control mechanism is lost because this form of trading does not go via stockbrokers. It makes it difficult to know whether those placing orders have enough equity to actually trade,' Lensberg explains.

According to the professor, another problem is that the speed can simply become too high, also for those who control and operate the robots.

'An algorithm cannot change its mind, and if those who look after the robots do not pay close enough attention, both the buyer and the seller can suffer great losses. There is a potential for great instability here,' the NHH professor believes.

One possible measure to avoid such problems is to test the stock trading robots before they are let loose on the market. But testing takes time, and a stock trading robot that is good this week might not be as good next week.

New expertise

After the brokerage houses started to use powerful computers and more technology in share trading, new expertise has also found its way into the world of finance. IT engineers, mathematicians and physicists are now teaming up with economists to develop quicker and smarter algorithms. Combined with deregulation of the financial markets, which has made it possible for the same security to be traded on several stock exchanges, this has created good conditions for high frequency trading.

'Most computer programmes and algorithms have "bugs". Giving them direct access to markets without any form of regulation can therefore be very unfortunate,' the finance professor explains.

There is talk of granting licences to individual players who will be permitted to engage in algorithmic trading, but Terje Lensberg is uncertain whether this would be an efficient solution.

'It would still mean that we have to trust those with licences to follow their own control procedures,' says Lensberg.