An online guide to more efficent automated FX trade execuTion for buyside firms

2021 Algo Trends: Increasing volumes, focus on TCA and transparency

By Andrew Cole, Algo Product Manager J.P. Morgan

In preparing for this article it’s interesting to reread some FX market commentary from January a year ago. How incredible to think that a principal challenge for FX traders at that time was volatility, or the lack of it, with headlines dominated by record equalling levels of low FX price movements. How quickly these concerns dissipated in March. How appropriate a reminder that any trend dominating the zeitgeist one day can as quickly be turned on its head; and what a convenient way of reminding readers the challenge of product management in financial markets. No sooner has the technology resource been lined up, the ‘must have’ product has evolved 5 times. And so, against the backdrop of constant change, we outline some algo trends for 2021, based on observations from our algo trading data.

Increasing FX Algo Volumes

2020 was a watershed year for FX algo execution, a product historically used by an increasingly active but limited user base has entered the mainstream. Since 2016 algo order execution has overtaken click to trade (CTT) as the primary means of transacting Spot eFX in sizes >$10M amongst JPMorgan’s clients. In 2016 clients traded <35% of their eFX Spot volume via algo orders (in these sizes). In 2020, this increased to almost 60%. This compares with Spot eFX CTT (~60% 2016 vs. 22% 2020).

This is not just about March 9 – 20 2020, although a fortnight where the Algo ADV was 33% higher than we observed when the Brexit decision was announced (pre-2020 record) clearly pushed the total up (who said algos weren’t used in volatile markets?). We saw a stabilizing of daily volumes rather than a decline after Q1, with our second and third largest months being November and June.

It is not just about our largest algo volume accounts (whose share of total algo volume decreased 30%, while total clients trading algos grew over 100% vs. 2016). It’s about the 21% of algo clients in 2020 who were trading FX Algos for the first time, the client segments which weren’t early adopters of algos showing Quarter-on-Quarter increases in volume and of course the gradual shift in client’s trading strategies from rigid limit to more passive market tracking. It’s about education, familiarity and increasing product sophistication.

Increasing JP Morgan FX Algo ticket Flexibility

One challenge that is anticipated to outlast 2021 is diversifying client demands. On one end of the spectrum are early adopters (algo users who have around 5 years of experience with the product), and existing users of algo trading in Equities or F&O with a robust understanding of the product. In general, JP Morgan has observed this group of experienced users seek greater control, features and flexibility in FX Algos. On the other end of the spectrum are beginners in FX Algo trading who seek simplicity and ease of use. Catering for both use cases inside a 240×380 pixel ticket is hard. This is evident from the results of the 2020 JPMorgan FICC eTrading survey that underscored this challenge – with approximately equal number of respondents citing “too complicated” (24%) and “lacking sophistication” (27%) as reasons why they did not trade algorithmically.

Offering this flexibility can be easier on single dealer platforms where tickets are more customizable. Nonetheless, building out our algo trading capability to reasonably enable client customization (liquidity rooms, client defaults and controls) is a priority for 2021. Determining the appropriate number of order types and parameters is arguably less critical when diverging from a ticket design based on a one size fits all model, and so a perhaps necessary transition for a client base as diverse as ours.

Increased FX Algo sophisti-plicity

Besides flexibility on the ticket another key focus is the evolution within the algo; helping clients understand the sophisticated decision making of our algos without confusing clients in the design complexity is key.

The evolution of time based strategies demonstrates the distinction (sophistication vs. simplicity). Although VWAPs are more sophisticated than TWAPs, using volume time (based either on historic volumes or, in our case, a quantitatively developed estimate of future market volume), we simplified them by offering fewer parameters on the ticket than the TWAP. This is not to say VWAP usage should totally replace the TWAP, there are times when a TWAP might be a more appropriate strategy, but we have observed growth in the use of VWAP.

Another example is liquidity; configuring additional liquidity rooms to manage client concerns around mark-outs on non-firm liquidity complicates the ticket; introducing criteria (round trip times and percentage fill rates) for Non-Firm Liquidity Providers interacting with JPMorgan can reduce market impact in a way that is more sophisticated and simpler for our clients. In both scenarios the goal is not to limit what clients can do (via customization more things will be possible), but rather to simplify what a client needs to do in order to experience the best of JPMorgan’s algo trading capabilities.

As algo sophistication increases bounding the perceived complexity through greater transparency is vital. Delivering greater visibility of what the algo is doing via real-time metrics and post-trade transaction reports should be a key focus for dealers seeking to show clients why they might execute more flow algorithmically.

Increasing order volume beyond Spot

As Non-Deliverable Forward (NDF) orders launched, JPMorgan were early-movers introducing to clients a product that internal traders have been using to trade in large volumes over many years. A significant investment for our order’s business has been CORE; an overhaul of order execution capabilities which will reduce the technological integration costs of introducing orders to new asset classes. Alongside NDFs we welcome more interest in products like Swap orders, which we launched in early 2018. A key challenge with these products is multi-dealer platform integration (all are available via JP Morgan’s Execute trading platform), we expect technology upgrades will bolster interest in them through 2021. The growing competition in the NDF ECN space in particular could point to greater opportunity for NDF Algos as liquidity fragmentation increases.

Beyond FX altogether and for multi-asset traders reading this: increased demand for Algo execution in other areas (Rates and Precious Metals) is something we’d expect and the firm would welcome more algo inquiries in these asset classes.

Increasing FX Algo Execution volumes “globally”

JP Morgan’s FX Algo trading volumes grew in all 3 risk regions last year, the NA vs. EMEA gap reduced substantially. But “global” growth continues to evade APAC, where algo order volume vs. CTT remains low. As the afore-mentioned changes in NDF market dynamics take hold and more algos offer greater certainty / transparency around end time we hope to see growing confidence and more algo flow in APAC.

Increasing venue specialization for FX Algo Execution

As algo execution takes on a larger share of large transactions, ECNs continue to look into ways to capture a share of those passive flows. Mid pools, full amount pools and other innovations will increasingly compete in this space against the traditional liquidity venues as clients look to reduce market impact. One of the noteworthy components of JPMorgan’s algos is the internal liquidity offered to our clients, which can vary by strategy and parameter configuration. Those clients that consider minimizing market impact to be a key criteria might consider the capacity of a provider to internalize, this can lead to lower market impact owing to the lesser information dissemination vs. external trades on ECNs which are visible to market participants. In a highly fragmented market such as FX both sources of liquidity can be advantageous to clients depending on their trading objectives and with this in mind JPMorgan offers hybrid liquidity options which allow clients to access both.

Single Dealer Platform (SDP) resilience

The increasing number of multi-dealer platforms for orders impacts single dealer platforms as well (although we’re yet to witness their decline), postrebuild JPMorgan’s Execute numbers are again growing and there are still compelling reasons to consider an SDP:

• Product Quality: what SDPs lack in multidealer streaming, they can make up for in a highly customizable, richly designed end-user experience.

• Data Access: the rich array of bank’s FX datasets are most widely and easily accessed via SDP.

• Continued Investment: Single bank platforms are typically dual purpose; almost half our volumes are internal; voice sales and trading. Just as our FX traders use our algos, the features we develop for external clients are used internally. SDP integrations are a long term investment.

Increasingly mobile

Our FICC eTrading survey again highlighted mobile trading as a potential growth area; a third of those surveyed considering it the ‘most influential’ for shaping trading in 2021. Although external FX Algo volumes traded via Execute Mobile were low last year (with percentage growth in line with desktop), the app remains popular with internal traders and receives a surprisingly high proportion of total Single Dealer algo volumes as a result. 2021 will show whether the ubiquity of mobile phones in every other part of our daily lives can be matched in FX Algo trading, which in turn will largely be driven by changes in control processes from our clients. Clients might consider the use of mobile trading applications for alerting and monitoring, our readonly mode can be configured to provide useful updates without any trading functionality.

Increasing need for real-time execution transparency

As the number of multi-dealer platforms continued to increase, keeping our tickets consistent was a challenge. Providing clients access to real-time performance data in a standardized manner needed a different solution. FX Algo Providers attempted to solve this in recent years by offering the single dealer execution experience on multi-dealer bank channels, as a way of providing consistent algo execution transparency. JPMorgan’s Algo Central was the first bank owned FX algo trading application hosted within Bloomberg, where others have joined and extended what’s possible we’re sure more will follow. The tools we give clients to evaluate algo performance in real-time must evolve in line with the algos upon which they are based. For clients to trust our algo’s behaviour, they must be able to understand and monitor the same market conditions as the algo uses for its decision making. We’ll be upgrading Algo Central again this year, focussed on facilitating this seamless information exchange.

Increasing focus on TCA

Post-trade Transaction Cost Analysis is also vital for algo execution traders who wish not only to challenge their own FX Algo choices (Dealer, Algo, Parameters) but also to justify these decisions to their clients, portfolio managers and others within their business. Clients may continue to use this data to guide their execution algo selection. Two trends that we welcome in the TCA space are:

1. The increasing emphasis on aggregate TCA – clients should remember that single order TCA whilst informative can easily be affected by the specificities of the market in which they were executing, potentially skewing the results.

2. Increased adoption of independent TCA – JPMorgan was an earlier adopter of third party TCA, we welcome new products in the market which allow clients to evaluate FX Algorithms from different providers in an economically viable way.

Increasing Innovation (we hope!)

More products developed by our competitors, the execution venues, other third parties and ourselves reflect deep collaboration with clients that have a specific objective in mind. Moving forward as an industry necessitates the devotion of early adopters to providing specific feedback, helping evolve the product. Recent years have been complicated by mandatory technology upgrades and migrations putting the core component of FX Algo execution; new execution capabilities, further down the list. We’ll be starting 2021 with a number of improvements to the way we execute, most excitingly as part of our CORE framework releasing a new FX Algo. We’ve no doubt that our competitors, like us, have various new products mapped out for the next 12 months. We look forward to continuing to make our client’s lives easier through the more seamless delivery of FX Algo Execution products this year and beyond. Good luck for 2021!

Andrew Cole’s views do not necessarily represent the views of J.P. Morgan.

Related articles

Market Trends See all