J.P. Morgan launches DNA and looks to shape future of FX algo trading

Demand for J.P. Morgan’s set of algo execution and TCA tools continues to grow in reach and sophistication and the bank continues to innovate in this space with the recent launch of a new FX algorithm called DNA or Deep Neural Network for Algo Execution. This combines what a multitude of algos currently do, into a single strategy and allows the framework to judge how a client order should be executed. Utilising Machine Learning techniques the new algo determines how best to execute a transaction, based on results of past trades and it has already been deployed for trading G7 currencies such as the euro, dollar and sterling. We asked Andrew Cole, FX Algo Product Manager at J.P. Morgan to tell us more about the growth of algorithmic FX trading and how the bank is responding to it.

J.P. Morgan launches DNA and looks to shape future of FX algo trading
Andrew Cole

In the past year, what changes have you seen in the appetite for FX algos usage?

J.P. Morgan’s algo order volume as a percentage of spot has increased in six of the last seven quarters and the number of unique users trading algos has grown in every quarter since Q3 2017. Clients have also reacted positively to our more sophisticated algo order types, with Aqua volumes as percentage of total algo volume nearly doubling in Q1 vs. the 2018 average.

FX algo trading volumes are notably on the rise for the APAC region, with the fastest pace of growth in Q1 taking place in APAC currency pairs. We are matching this demand with the release of our NDF Algos, which currently support three APAC NDFs.

Algo Central launched recently on Execute and Bloomberg. What has the response been so far in terms of client reception?

The response has been very encouraging, with March marking another month of record volumes and unique clients trading. Algo Central has also triggered important conversations with our clients, particularly around their pre-trade workflows.

Clients can now place orders via the normal J.P. Morgan Execute ticket and monitor real time performance via the tool. In order to improve choice for our clients, the tool is available on the Bloomberg App Portal. Algo Central can also be opened via FXall and FXConnect, with other platforms being made available in the coming months.

How do you plan to further expand your FX algo trading functionality and TCA suite?

According to J.P. Morgan’s 2019 e-trading survey, 61% of active traders believe machine learning will shape the future of trading over the next three year – the biggest single technology driver. To that end we have some exciting developments in the pipeline this year, including deeper use of machine learning for more intelligent order placement. In addition, our FX Code Only liquidity pool (made into by liquidity providers who have attested to the FX Global Code) is now live. We are also launching Algo Insights, a unique tool to provide real-time assistance to clients on the progress of their execution.

This will become more customisable, proactive and supportive of other third party channels as the year progresses. We already support limit based algos for NDFs and will also be moving into market tracking for these currencies later in the year.

In what ways do you see FX algo usage evolving?

Rising to the challenge of our increasingly sophisticated client set is what keeps this space most exciting. With their deeper awareness of FX market structure, it becomes far easier to explain why internalisation is important for minimising mark-outs, the role of a well-managed non-firm liquidity pool and how fee-paying algos accessing external liquidity can save more spread and fill faster than simple limit orders.

Clients increasingly understand the advantage of principal algo execution (vs. agency) and at J.P. Morgan we can offer our clients the best of both with segregated algo execution alongside a truly internal liquidity pool.