The Bank for International Settlements (BIS) has released one of the most detailed reports produced to date exploring the drivers and implications of the rising use of Execution Algorithms (EAs) in FX markets. Established by the Markets Committee in mid-2019, the findings also cover the performance of FX algorithms during the pandemic-related period of market volatility earlier in the year.
The key takeaway of the report is that EAs support price discovery and market functioning in an increasingly fragmented market. However, the report warns that they also contribute to the ongoing changes in market structure and with increasing scale of use, give rise to new risks and challenges that warrant close monitoring.
In order to gain insight into the usage and provision of FX EAs, the study group conducted a survey of large buy-side and sell-side market participants and a series of interviews with sophisticated market participants on preferences, offerings, motivations and challenges related to EA usage and provision. Among these large global institutions, roughly half were found to already be using FX algorithms, which the report found to be relatively high compared with other metrics on EAs’ FX market share, but similar to results from a survey conducted by the GFXC, where 40% of respondents indicated that they use algorithms in their trading. Meanwhile, 55% of the potential providers (ie banks and principal trading firms) in this group of large institutions were already offering FX algorithms to their clients, and half of the remaining 45% noted they were considering offering these tools.
Reproduced from BIS report: FX execution algorithms and market functioning
Reasons for using FX algos
The survey also found that respondents were predominantly using EAs to trade FX spot, while EA usage for FX derivatives such as non-deliverable forwards (NDFs) – where electronification did not gain traction until 2008 – was nascent. FX algo usage was most prevalent in liquid advanced economy currencies and used less for emerging market economy currencies. Users’ average ticket sizes for advanced economy currencies were also about 50–100% larger than in EME currencies.
Survey respondents indicated that they use EAs primarily to reduce trading costs or market impact. According to the report, others may also be using FX algorithms to ensure execution within some allocated period of time, or to trade close to or better than particular benchmarks. Finally, the report found that FX algos are often used by market participants to achieve greater operational efficiencies related to trade entry, trade monitoring and trade settlement, as algorithms automate what is otherwise a manual process. The report adds that FX algo transactions are generally settled as one single ticket regardless of the number of individual child orders, significantly reducing operational costs.
Use by participant group
The study group’s survey and interviews with FX algo providers suggest that institutional investors, such as asset managers and hedge funds, are among the more common users of EAs. These market participants have the ability to accept some market risk in order to reduce execution costs by minimising the bid-ask spread paid, and reducing the market impact of their trades, especially for large-sized orders.
In addition banks, typically small to mid-size regional banks that do not offer FX algorithms, were also found to be users of EAs. Yet penetration rates and volumes transacted were smaller than those of the buy side. These entities generally used EAs alongside other modes of execution, and mainly targeted swift execution. Many provider banks also allow EA access inside their own organisation by desks that need to transact FX – so-called in-house users. Notably, feedback from providers suggests that in-house users typically have access to the same suite of EAs as are offered to clients, suggesting that tools for strictly proprietary use do not seem widespread among providers.
The report found that non-financial corporations represented the smallest segment of algo users among FX market participants, in line with their generally smaller FX footprint. Some corporations deliberately stay away from EAs, according to the report, which require relatively deep market expertise and understanding of technical aspects of FX transactions. These institutions prefer traditional execution styles such as voice trading for risk transfer or use of fixing orders. Others, however, find they have the required level of sophistication in their organisation. These included several large multinational corporations for which EAs accounted for about 20–25% of their FX transaction volumes.
Central bank usage of FX execution algorithms
The survey also covered central banks’ use of EAs, particularly amongst 15 central banks from Asia-Pacific, Europe, North America and South America. Of the central banks that were surveyed, five reported using EAs, with the extent of usage varying widely. One central bank indicated that it used EAs for almost 90% of overall volumes, two used EAs for about 25–30% of overall volumes and the other two central banks used EAs for less than 10% of their overall FX volumes. Central banks’ use of algorithms for FX trading was concentrated in developed market currencies. More than half of them used time-sliced algorithms for more than 70% of their transaction volume. Most central banks indicated that they have three or more FX EA providers while ten of the 15 central banks reported that they were not currently using EAs. Of these, two are currently considering the future use of EAs.
In addition, six highlighted that they did not consider the use of EAs necessary due to the relatively low volumes of their activities in the FX market, while two highlighted internal restrictions on the use of such systems. Overall, central banks were relatively cautious when using EAs, and would not use them outside their main trading hours. When using EAs, all of them had people overseeing the execution. Most of the users attached limits to the orders, mainly limit price controls. For central banks that have adopted EAs, the main motivations were to reduce trading costs, improve desk productivity, and access multiple liquidity pools in order to reduce market impact or footprint. Some central banks also cited improving execution consistency, transaction cost analysis and confidentiality as reasons for using EAs. None of the central banks highlighted pre-trade analytics or best execution requirements as key motivations for their use of EAs.
The report concluded by saying that Execution Algorithms have become a well established execution channel in the FX market in recent years, which was mostly characterised by subdued volatility. And while many observers and market participants expected them to be less widely used during a crisis, it appears that the opposite has happened: EAs look to have retained an important role during the turbulence caused by the Covid-19 pandemic. And they
are likely to be here to stay.