The latest Bank for International Settlements (BIS) Quarterly Review provides execution method data relating to the recent Triennial Survey of FX Turnover and finds that while e-ratios generally are stable there have been some significant shifts in the channels handling the volume. Perhaps significantly, the report also picks up on what the authors says are “signs that fragmentation may be reaching its peak” as previously hinted at by Profit & Loss, it also warns over the FX market’s capacity to handle stressful conditions.
In their conclusion the authors highlight how the changed market configuration, whereby more risk has moved to the buy side and non-bank firms are grabbing a bigger piece of market volume at the expense of regional banks (while the big players internalise more and more), has emerged largely during a prolonged period of low volatility. “[The FX market’s] resilience might be tested if the volatility regime were to change,” the report states. “For example, during periods of stress, FX dealers might ration liquidity and favour clients with whom they have a strong relationship, such as those using their single-bank platform. Thus, customers who spread execution across venues could face a sharp evaporation of liquidity.
“The question of whose risk-bearing capacity to rely on under such circumstances could become a pertinent one,” it adds, voicing one of the louder debates in FX markets currently – who will stand in the market and risk warehouse during volatile markets.
In terms of how participants are trading there is the usual frustration over slight changes in methodology between reports, meaning that direct comparisons are difficult, however generally speaking e-ratios went up slightly from April 2016 in all areas except for FX swaps. The latter could be significant because interest rate volatility was certainly a feature of markets leading into this year’s turnover survey and the fact that voice trading made a (slight) comeback might suggest the same could happen in spot markets hit a prolonged period of volatility or illiquidity.
In terms of numbers, the spot e-ratio globally rose to 69% from 65% in 2016 – that translates to a notional increase just shy of $300 billion per day at $1.37 trillion per day. There was a greater percentage climb in outright forward e-ratios, from 52% to 57% (for a notional value of $569.4 billion), although that is largely likely to be a result of the increased electronification in NDF markets which come under this banner. FX options remain the least automated of product segments at 31% but this is up from 26% in the 2016 survey and a notional value of $91.1 billion.
As noted, the only product to see a decline was FX swaps, although even here it was a drop of just 1% to 52% for a notional volume of $1.67 trillion.
Within the e-FX spot volume data there was a significant shift, however, away from “direct” channels to “indirect”. The percentage of spot volume executed electronically direct, via a single dealer portal or direct API connection, fell from 43% of volume in 2016 to 32% in the latest survey. This represents in notional terms, a fall in volume from $710 billion per day to $635 billion. Conversely, volume traded indirectly via e-channels rose to 37% from 23% three years previously, a notional change from $380 billion to $735 billion.
While this data would appear to fly in the face of anecdotal evidence in recent years that volume is increasing on disclosed channels, and again this is where the different reporting methods across surveys hinders comparison, it could well be that aggregation channel volume is now being allocated to indirect channels, hence the change in the data, however anecdotally market participants suggest that even though the volume is indirect it is still on a disclosed basis.
The BIS report finds that of the indirect volume, 26% is traded anonymously, which equates to $516 billion per day. The BIS report stylises indirect anonymous flow as being from “Multi-bank dealing systems that facilitate trading on a disclosed basis or that allow for liquidity partitioning using customised tags (eg 360T, EBS Direct, Currenex FXTrades, Fastmatch, Refinitiv FXall Order Book, Hotspot Link); while indirect disclosed venues are “Electronic trading platforms geared towards the non-disclosed inter-dealer market (eg, EBS Market, Hotspot FX ECN, Refinitiv Matching).”
Again, because of the change in reporting style no direct comparison can be made with the 2016 survey, however it seems as though the survey has confirmed that more customers are putting their LPs into competition via aggregation.
In terms of other FX products, the share of direct channels went down in all except for FX options. In outright forwards it fell from 32% to 29% ($290 billion per day in 2019), while in FX swaps it fell from 29% to 26% ($832 billion per day). FX options activity via direct channels rose to 21% of turnover ($61.7 billion) from 16% in 2016.
Again, there was a shift to indirect venues, with the share of outrights traded via this channel rising to 28% ($280 billion) from 20%; in FX swaps it rose to 26% ($832 billion) from 24% and in FX options it was unchanged at 10% ($29.4 billion).
Overall the picture painted by the execution data is one of slow, but steady progress. Inevitably given the complexities around trading such products online the FX swaps segment remains heavily voice driven, however with several initiatives in the pipeline to further automate these markets (not to mention the potential from clearing and compression), this may change by the time of the next survey.
For the other product sets, NDFs will continue to drive the outright forward segment and FX options, due to the bespoke nature of the product, will likely remain a manual process for the foreseeable future. The interesting factor to look for in the spot market could be the impact (if any) of LPs reducing the number of venues they support (if indeed they do make the cuts).
There seems little doubt that e-ratios in spot have hit a natural pause and progress will inevitably be slow from here, that does not mean, however, that it will stop entirely. It is hard to escape the view that it would be better for all concerned if the local FX committees and the BIS could agree on a single methodology for measuring execution channels – not least what actually represents anonymous and disclosed trading, however until then the message can best be discerned as “steady but unspectacular progress”.