Norway’s oil fund goals to avoid wasting $400mn of buying and selling prices utilizing AI | USAEMALL.com

Unlock the Editor’s Digest for free

Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.

Norway’s giant $1.8tn oil fund expects to save a fifth of its annual trading costs by using artificial intelligence to trade less frequently and more efficiently, in a sign of how the technology could upend global finance.

Nicolai Tangen, chief executive of the world’s largest sovereign wealth fund, told the Financial Times that it was targeting a saving of $400mn from its trading costs of $2bn each year, and it had already achieved close to $100mn.

“We have these AI programmes where you can predict [our buying and selling]. The goal is to save $400mn in trading costs a year — it’s huge numbers,” Tangen said. “Sometimes we buy a company on a Monday and sell it on a Friday because of index changes. Now you can predict this kind of things much better and you can internalise these flows,” he added.

The Norwegian fund is one of the biggest equity investors in the world, owning on average 1.5 per cent of every listed company globally, and it makes more than 46mn trades each year. By handling more of these internally it can save both on transaction costs and the gap between the prices at which stocks are bought and sold.

Nine out of 10 investment managers surveyed by Mercer last year said that they already used or planned to use AI, with two-thirds highlighting the reduction of trading costs as one of the motivations. If big investors are able to make meaningful savings on their trading costs it could crimp the earnings of Wall Street banks and trading firms that act as intermediaries.

Parts of Wall Street have long embraced AI, with big trading firms such as Citadel Securities, Tower Research and Hudson River Trading using fields such as “machine learning” to quickly and cheaply match buyers and sellers on stock exchanges — and in their own internal systems.

But large institutional investors have tended to be slower to adopt these techniques. Many AI projects to date have failed to target concrete, measurable improvements, executives say, but there is now growing optimism over what they can achieve.

Tangen, a former hedge fund manager who recently received a second five-year period as chief executive, said that the fund started the AI trading initiative two years ago. “There are so many ways that we can tweak this organisation, and the numbers are so big that even small improvements [have a big effect],” he said.

The main ways that the Norwegian fund is seeking to crimp its trading costs is by trading “smarter” — timing and sizing its purchases and sales better to minimise transaction costs and the impact it can have on the market — and by simply trading less, identifying when it might be better to wait or swap securities between in-house investing teams.

Tangen added that to implement AI fully in an organisation it needed “a total maniac at the top who is just talking about it all the time and who’s really going through the organisation properly, and then everybody you recruit needs to be super IT literate and able to programme . . . You just need to carpet bomb the organisation with these initiatives.”