Nordea FX algo: Learnings from BIS

Bank for International Settlement (BIS)  released their report on monitoring of fast-paced electronic markets in September 2018. The report takes an outset in the fact that Foreign Exchange markets have undergone significant structural changes in recent years.

Key trends according to BIS

1. Technological advances have changed the nature and location of trading

In spot FX, the share of trading volume executed electronically has almost doubled over the last decade. One result is that trading activity is increasingly fragmented across a range of new venues, while the proportion of trading on so-called primary venues with transparent central limit order books has dwindled over time. Another is that the frequency of activity and speed of information flows in FX markets has increased drastically.

This development has been noticed in Nordea. “We have seen a clear interest from FX customers who have been turning to algorithms to access multiple liquidity pools to reduce their trading costs and improve execution quality”, says Jesper Goor Pedersen, Algo Quants and Insights at Nordea.

2. As technology change, so does the nature of participation

BIS highlight that business models of traditional bank intermediaries have come under increasing pressure. That, liquidity provision has become more concentrated among the largest banks, which reap the benefits of a large electronic network of client relationships to internalise a large part of their customer flows.

“We have also found that internalization ratios for algos is a topic that many clients are interested in. There are some limits to the internalization ratios you can expect to achieve, if you are not willing to cannibalize best execution for the client. We have decided to be completely transparent about the model we use to internalize flow from algorithms, including the limits to internalization ratios that are implied by this model. It is important that customers understand that it is not always beneficial to internalize flow”, Jesper explains.

3. Greater electronification has led to the creation and commoditisation of large quantities of high-frequency data 

This can raise costs and erect barriers to entry the BIS report concludes. But it has also opened a range of opportunities for market participants. Liquidity providers can rely on a wider range of inputs to build a quoted price for clients, and end-users can utilise new data to benchmark and improve the quality of their trade execution.

Jesper explains; the choice that a client is faced with when choosing algo type and parameters is how to weigh market risk with the risk of causing market impact. In most cases, clients should aim to execute as fast as possible but not so aggressively that they move the market against themselves. Here, pre-trade analytics play a key role in ensuring that the choice of algo and the parameters of that algo make sense in comparison to the current market conditions. On the other hand, post trade analytics serve the purpose of validating the pre-trade input. It is important also to stress that doing only pre- or post-trade analysis alone is not enough. One must establish a feedback loop between pre- and post-trade analytics to be able to properly learn what is the best way to execute.

Technological developments in markets and FX algos

Technological advances that have changed the face of trading across a range of financial markets include greater processing power, lower storage costs and the ability to transfer data almost instantly.  The cost of trading, as measured by bid-ask spreads, has fallen notably as a result, the BIS report outlines.

Reduced costs are generally something all clients are looking from algos. However, we also see that transparency and control are some of the most important features that algos provide to clients when compared to a risk transfer Jesper explains. “By using algos, clients also have a possibility to select what kind of market impact their trading behaviour will have. We see control over market impact as a major differentiator because clients may have a very different execution horizon then a traditional trading desk that would serve them with a risk transfer. When the long execution horizon is combined on the client side with proper pre- and post-trade analytics the result can be very different from just doing a risk transfer”.

Less manual trading and more FX algos

BIS Survey data suggest that, since 2013, over 70% of spot trading is executed electronically while an estimated 70% of orders on EBS, a primary central limit order book (CLOB) and a major inter-dealer platform for spot FX, are now submitted by algorithms, rather than manually.

The BIS report further make note that greater electronification of trading can make it harder to keep track of and understand trading activity and market moves. Algorithmic execution designed to minimise market signature (e.g. by decomposing orders into multiple smaller blocks) can mask trading patterns and volume while prime brokerage on electronic communication networks (ECNs) masks the identity of the economic counterparty to the trade. Traditional market contacts (such as voice dealers) may have less visibility over flows that are driving short-term price formation.

“We see ourselves as providing value to customers by managing the complexity behind trading on the fragmented FX spot market. This means in practice management of infrastructure, personnel, algorithms and liquidity connections that make it possible to execute trades efficiently in the FX spot market. Generally, we are very open about how our whole value chain works when talking to clients to make sure that they understand enough to feel comfortable about our offering”. Jesper concludes.

The full BIS report on fast-paced eletronic markets, please click here