Using algos to find the jewel in the Nordic crowns

June 2023 in Previous Features

All the building blocks would seem to be in place for the take-up of algorithmic forex trading by the buy-side in the Nordic region, and yet it is still very much a work in progress as Adam Cox discovers.

The Nordic region boasts a clutch of large banks, liquid currencies, a track record in innovation in terms of electronic trading and a large and diverse community of buy-side players, many of whom are technologically sophisticated. Yet for now, algo trading by the buy-side remains more focused on equities, with funds often relying on their banks to slice and dice currency trades for them.

“It’s early days,” said Per Andersson, head of sales and marketing for leading regional platform provider Pantor Engineering. Andersson said that to date the adoption of algorithmic trading in the local FX market has been mostly confined to the sell side. Swedish banks, for instance, have been developing execution algos for themselves and are just beginning to talk to major corporates and asset managers about marketing them for use directly by the buy-side.

Andersson said: “The Swedish market hasn’t come that far yet. The first goal is to get the execution algos in place.”

Divide

Last year Sweden’s central bank, the Riksbank, published a study of algo trading in the FX market and divided up the activity into four main functions: executing orders, quoting prices, managing risks, and speculating or exploiting arbitrage opportunities. There was a sharp divide between buy-side and sell-side activity on the execution side. For instance, the study estimated that for market-makers, algorithmic trading constituted some 20% of pair trading in which the krona featured and 39% of trading in which only other currencies featured. But for some other market participants, the percentages were 4% and 2% respectively.

In the forex market, execution-based algos are among the most common. The large share of volume in the region that is based on genuine corporate activity and trade-related flows – as opposed to speculative activity – can help explain this. Sweden boasts a number of sector-leading companies such as Ericsson, Volvo, SKF and numerous others. Norway meanwhile is one of the world’s top energy producers and home to a world-leading sovereign wealth fund.

Yet the advantages for the buy-side in terms of understanding exactly how orders have been executed and later being able to conduct analysis have not to date been enough to sway the local participants, at least in Sweden.

The Riksbank’s report noted that: “The asset managers and non-financial companies’ treasury departments interviewed say that their usage of algorithms only amounts to at most 10% of the total foreign exchange transactions in Swedish krona.”

The authors, Maria Bergsten and Johannes Forss Sandahl, interviewed seven market participants, including Swedish and foreign banks, pension fund managers, hedge funds and non-financial companies. The banks included were responsible for about 50% of turnover on the krona FX market.

Heavily traded currencies

Per Andersson

“In general, the buy-side has not been pushing for algorithms to trade FX. On the equities side, algos are well established and it is now seen as a standard functionality. We believe FX and other asset classes will move in that direction too at a rapid pace.”

The Swedish crown is the biggest of the region’s currencies, although it has lost some ground in the rankings. Once a staple of the world’s top 10 currencies, it recently slipped to 11th place in the latest Bank for International Settlements survey (2013), with a 1.8% share of global volumes. That compared with 2.2% in the 2010 survey, when it was 9th most traded.

The Norwegian crown’s share has inched up to 1.4% from 1.3%, although its ranking actually slipped to 14th from 13th. Finland uses the euro, while Denmark’s crown is linked to the euro and generates comparatively little trade; it has been outside the top 20 currencies since 2007.

One banker we spoke to in Stockholm said there were some advanced buy-side players who have built algo systems themselves, and there were many others who have not yet realised that it would be a good idea to start trading their FX exposures algorithmically.

There was also a middle group, for instance large pension funds, who find they have too much volume to place in one order. “Basically, I would say any sophisticated equity derivatives or equity algo firm realises that it’s a good thing to have algos for FX as well. Those clients have already sorted it out for themselves,” he said.

Demand for customers to use algos provided by the banks has been growing, but it’s a starting from a low base. “We’ve done some checks and I don’t think there’s currently a great demand. But customers get more and more aware, so that is something that is about to change,” the Swedish banker said. He also stated that any firm trading multiple asset classes, in particular those other than equities, will want to do their FX risk hedging. “There are surprisingly few who have actually automated that, which is a bit strange, because now with the very tight margins that you could potentially make on the equities side if you do not manage your FX risk then you lose all that you potentially make. For instance, if a firm is using a TWAP algo strategy on the equity side, it may want to make sure its exposure on the FX side was automatically dealt with.”

A final thing to note about the Riksbank’s study was that it showed that daily turnover in the period measured was only about 14 billion kronor for the shares in the OMX 30 index of leading Swedish shares, versus about 72 billion in foreign exchange transactions between the Riksbank’s monetary policy counterparties and counterparties in FX transactions. Much of the algo trading, the study added, was in spot. There is some algo trading in the local market for forwards but the specific characteristics of forward deals, as opposed to the standardised nature of spot, made it more difficult for algo trading.

Pioneering tradition

The Nordic markets were actually early users of electronic trading in the 1990s, with OM – which later became OMX and ultimately was acquired by NASDAQ – pioneering electronic derivatives and later equities trading.

“The Swedish market, at least when it came to electronic trading of equities, was leading the pack,” Andersson said. “I think the Nordic market is currently lagging behind in terms of FX. It’s a bit of a catch-up situation.”

He added: “In general, the buy-side has not been pushing for algorithms to trade FX. On the equities side, algos are well established and it is now seen as a standard functionality. We believe FX and other asset classes will move in that direction too at a rapid pace.”

Another aspect that may have made for a slower start to the adoption of FX algorithms directly by the buy-side in the Nordics has been the traditional conservatism of the local financial industry. ”If you look at the buy-side in Stockholm, many have not yet fully adopted electronic trading in all asset classes,” says Andersson.

Pantor, whose platform integrates with various liquidity venues and consolidates the market data, provides a set of execution algos to its clients and also supports them in developing algos themselves using the platform. Andersson said that as new tools and systems come on board in the region the situation could change, particularly given the tradition of pioneering new trading techniques on the equity side. “It’s really about taking the knowledge from the equities side (and applying it to FX),” he concludes.