Essentially, the reason why FX algo trading has grown in popularity is because they deliver an execution strategy akin to a slight improvement on the average price over the time of the order. That’s according to one UK-based corporate treasurer, who relayed their views to Michelle Price, associate policy and technical director at the Association of Corporate Treasurers. “In this way, they are similar to how VWAP algos are deployed in the equity market,” the treasurer added. “Intuitively beating an average price has appeal for treasurers who should not be trading FX, but rather seeking to hedge effectively – and with minimum execution costs.”
But according to Price, in her experience it is still only a small number of sophisticated corporates who actually use algorithms at present. “It’s just not a relevant tool for the vast majority,” she says. Survey data collated by NeuGroup in 2017 also found that 62% of members across its two FX managers’ peer groups (FXMPG) said they were not regular algo users. Of the 34 members – consisting of very large multi-nationals with sophisticated FX hedging requirements – who responded to the question on FX algo use, some 29% said they were ad hoc users, while 9% said they were regular users.
Anecdotally, usage among members was roughly the same a year later, except one of the ad-hoc users from the previous year has since increased their use to become a regular user – and a regular user has increased their volume through algo trading, says Anne Friberg, senior director, peer knowledge exchange, NeuGroup. “Directionally, that will tell you something,” she adds. “And whether they are regular users or ad hoc users, they tend to use those for the larger currency trades, or sometimes in situations when they are not targeting a fixing rate in combination with significant volume, large capex or M&A transactions.”
Knowing the benefits
One large manufacturer who responded to the NeuGroup survey explained that over the last three years, the bulk of their corporation’s FX trade flow has moved to algo execution from multi-bank RFQ. This was driven by a need to reduce market impact on large orders, transaction costs and the fact that their counterparties are increasingly offering algos where they didn’t before. In addition, FX algos were said to offer greater trading possibilities and, as more algos become available, there is greater potential to negotiate the fees down, according to the respondents who used them. Another responder said that from doing no algo trading at all in 2016, it has now become a major part of routine trading.
A different set of survey data published by Greenwich Associates last year indicated a slight decline in the percentage of the largest corporates trading in excess of $50 billion annually who used FX algos, while their use rose more significantly among smaller corporates. Yet this year there has been a one-third increase in algo use by the largest corporates, with no further increase reported among small buy-side firms. Alfred Schorno, global head of sales and Managing Director, at 360T says the platform has also observed a significant increase in bank algo usage among its largest corporate clients – but no significant increase to date from smaller firms. “However, their interest in ‘fact finding’ and questions around usage, details of workflow, benefits etc. has risen notably,” Schorno adds.
But in comparison benchmark orders, which still tend to be prevalent among corporate users, are a much blunter tool for risk transfer and with more limited flexibility, warns Schorno. “Algos are about empowering the user to make more granular decisions,” he explains. “The power to reduce cost of trading through ‘sell-side like’ participation, to use alternative liquidity pools and reduce market impact and carefully selecting liquidity pool/providers allow clients to achieve goals previously not attainable.”
Recognising limitations
Algo adoption among corporate treasurers is still in its early stages, says Curtis Pfeiffer, Chief Business Officer, at Pragma Securities. However, bank clients are now reporting more demand from their corporate clients to provide them with algo trading tools, he adds. “There are some who are quite sophisticated users who have now been using algos for a number of years,” Pfeiffer observes. “But the overall trend for leveraging algo execution trading tools has started to pick up – and will continue to be strong over the coming years.”
“The ability to specify if a corporate wants the order executed with the bank acting as a principal or an agent is high on their wishlists when it comes to improving the adoption of FX algos,,”
Michelle Price
Liquidity fragmentation is also becoming more pronounced in the FX market, while the speed of quote, market and price updates are happening much faster than a human can trade, according to Pfeiffer. “You need some kind of automated tool to help respond to that frequency of quote updates,” he warns. “FX algos can consolidate multiple pools of liquidity, help reduce market impact, improve efficiency and free the trader to focus on some of their more difficult orders.” In comparison, corporates who are not currently using algos are potentially risking a poorer level of execution quality, Pfeiffer warns.
Selecting the right algo for the trading environment is a growing science
But Price reports that one of the main risks or worries about using FX algos, was highlighted by one corporate treasurer, in relation to how they may perform in a “flash-crash” event. This was seen in October 2016 when sterling fell sharply against the dollar, with algos cited as one of the potential causes by the Bank for International Settlements (BIS). “This could cause the algo to follow the market lower if it was executing a sell order,” the treasurer explained.
Getting it right
According to one of NeuGroup’s FXMPG members, benchmarking orders is also more appropriate than using algos in certain circumstances, such as when the fixing rate and the accounting rate are the same or for when dealing spot FX currencies from countries with capital controls. Many of the groups’ members are also passive users who employ algos to become a liquidity provider in the interbank market and collect spreads, which “to their mind helps offset some of the algo fees that they have to pay”, Friberg explains. As a result, TWAP tends to be the strategy mentioned by members more than others, with one treasurer stating that it also “easier to understand”. Others which have been mentioned include VWAP and an algo called Silent Partner, Friberg adds.
“FX algos can consolidate multiple pools of liquidity, help reduce market impact, improve efficiency and free the trader to focus on some of the more difficult orders.”
Curtis Pfeiffer
However, if they need quick execution they might choose a hybrid strategy instead, which runs passively but will agress if the market needs to be completed by a certain time – or as one member said, if the market was starting to go against them. While entry level algos tend to be largely Time Slice algos or TWAPs, Schorno explains that larger transactions might benefit from a strategy averages-in over time, participating on the bid or offer and creating a small footprint during execution. “As people get more comfortable, more complex strategies are being engaged,” he adds. “Adaptive strategies are growing in sophistication, coupled with greater access to empirical data gives clients greater understanding of market micro-structure at the point of trade.”
“Adaptive strategies are growing in sophistication, coupled with greater access to empirical data gives clients greater understanding of market micro-structure at the point of trade.”
Alfred Schorno
Algo users should also be clear that market risk remains with them until completion, Schorno warns. “The longer you run an algo, the more event risk you introduce,” he says. “Of course over time, the power of averaging helps you introduce both upside and downside potential of such risk.” So selecting the right algo for the trading environment is a growing science – the balance is market impact versus time, according to Schorno. To get this right, clients should be clear about their goals and work closely with their providers to help understand which strategies are suited for their needs.
Compare and contrast
Banks should also be aware that if they are not transparent enough with corporates as to when they are acting on an agency or principal basis, then that is another serious limitation, warns Price. “To the extent that the algo can access internal flows within the banks, then the bank will be acting as a principal,” one treasurer told Price. “To the extent that they are utilising third party liquidity from the likes of an electronic platform such as EBS, then they are acting as an agent.”As a result, says Price, “the ability to specify if a corporate wants the order executed with the bank acting as a principal or an agent is high on their wishlists when it comes to improving the adoption of FX algos in this space.”
However, some bank algo providers do already provide their clients with a choice of agency only or principle only execution, or a combination of the two, according to Pfeiffer. “This shouldn’t be a concern, because any bank that has signed the FX Global Code of Conduct will be transparent about whether it is an agency or a principle trade – and give their clients the choice,” he adds. “Principle nine of the FX Global Code re-enforces this point.” Corporates also want a better understanding of how algos compare to their other execution methods in terms of transaction costs and pricing,” claims Friberg, FXMPG members said they would also like FX algo providers and platforms to offer better reporting tools and transaction cost analysis (TCA), adds Friberg. “Being able to see what you are getting out of algo trading is the key to getting more people to start using it,” she adds.
According to one FXMPG member, TCA should be provided by all the major trading platforms as well as better levels of pre-trade optimisation. Another commented that they were disappointed with the commercially available TCA solutions and found the bank provided TCA’s to be of varying quality. As a result, this particular member has developed their own TCA in house to analyse their algo use. Historically, TCA solutions were impractical to the extent that deal entry was manual, adds Price, quoting another corporate treasurer. “Consequently, executing 500 separate trades in order to fill a purchase of €100 million was prohibitively time consuming,” the corporate explained. “Now, with automatic and secure uploads of executed trades into a Treasury Management System (TMS), this has become more practical and straightforward.”
Algo users should also be clear that market risk remains with them until completion
However, according to Price, the corporate treasurer still did not expect FX algos to become an indispensable tool in the near future, with order based execution remaining their preferred option – particularly when it comes to trading the less liquid currency pairs. But the last few years has also seen a strong surge in choice – growing from just a handful of liquidity providers offering an algo suite to something covered by most major FX franchises, argues Schorno. “The onus on us has been to keep pace with this supply and we invested heavily over the last six months to ensure the biggest players are available to our clients,” he explains. “Fully integrated solutions are key for corporate clients. Algos should be another ‘tool in the tool belt’ and it is important that customers can get access to this in one stop, as part of their daily workflow.”
Moving to the mainstream
Although he expects to see strong growth in FX algo usage over the coming years, Schorno also emphasises that this remains just one of a number of alternatives. “Certain size trades, depending on currency pair, incur such a small risk premium to transfer that an algo may not be appropriate,” he says. “The analysis of the market is what will be indispensable – those positioned to help guide clients to the most appropriate solution at that point in time will be the winners.”
Pfeiffer also predicts the uptake of FX algos will continue to grow, particularly as more corporates become comfortable with the algos, how they work and will start demanding them to be better customized to their needs. Another key benefit of using algos is that it allows you to perform TCA, Pfeiffer adds. “More and more people want to be able to benefit from TCA in order to improve their execution quality,” he says. “With MiFID II this year, and the FX Global Code last year, corporates are taking best execution to heart and will turn to TCA to help improve their execution quality.”
And according to the NeuGroup’s FXMPG member who took up algo trading in 2016 to become a regular user, FX algos have since become an indispensable tool. Another ad hoc member said that although it wasn’t indispensable, the algos did serve a purpose and do achieve cost-lowering results. However, Friberg adds that corporate teams are fairly lean so they are mainly driven towards efficiency, risk management and “as little administration as possible” – they are not trading for profit. “Until execution quality can be easily compared through reports, it is going to be difficult for the great majority of corporates to justify their use,” she concludes. “It definitely needs to become easier. To become really indispensable for corporate treasurers, more mid-market banks need to offer algo trading capabilities over electronic trading platforms to truly enable algo penetration in the corporate market overall.”
Key benefits of algorithmic execution for corporate treasurers
- Breaking up a big order into multiple smaller parts can lead to paying less than the cost of one large trade (or principle trade) where there is a premium paid for immediate execution and risk transfer
- Algorithmic trading aids transparency by offering time stamps on orders and an audit trail that makes it easier to measure execution quality against a corporate treasurer’s benchmark
- Building algorithms on top of an aggregated liquidity pool effectively narrows the spreads being traded on
- Automation means more orders are handled more reliably by fewer people which frees up treasurers and dealing staff to focus more time on other important areas of the treasury function
- There is reduced information leakage as the wider market is not aware of the full size of an order being traded which is important for larger orders which could impact the price of the currency
