If you ever wanted an example of how an enhanced market structure – specifically around electronic trading – helps to build volumes, one need look no further than NDFs. Volumes have been climbing steadily over the past five years, indeed I am starting to think they may eclipse FX options soon, and how people trade them also appears to be shifting, to the degree that I wonder if the NDF market is giving us a glimpse into the market structure of the foreseeable future?
NDFs have been undergoing a rite of passage for a currency pair – voice pricing through the brokers and direct between dealers, evolves to customers being able to access single dealer pricing electronically (initially RFQ of course), which then shifts to RFS and ultimately stays there but with tighter spreads. The driver of all of this is the availability of better data thanks to the public markets seeing greater depth.
This is what happened in G10 markets of course, but as the e-space become even more competitive, the major dealers started getting a little choosier over where they streamed. Will this happen in the NDFs?
That is a hard one to answer because in spite of the tremendous growth in NDF volumes, the percentage of trading through the electronic brokers and multi-dealer platforms has barely changed over the past five years. While NDF daily volumes have increased in the UK and US markets combined by 50% in just two years (from when the FXC in New York started breaking out NDF data), in the UK (the only centre to break out NDF trading styles) the percentage of volume handled by ECNs and multi-dealer platforms has actually dropped slightly, to 30.3% of volume from 31.9%.
Look back further in the UK data and one sees that those two channels have consistently been handling very close to 30% of volume for the past five years. The data, when looking at the multi-dealer platforms alone is much more volatile, but the latest survey actually has the lowest percentage of turnover traded via this channel, 18.25%, than at any time in the previous five years. Generally speaking, this channel has handled in the region of 21-24% of daily volume.
This all suggests to me that NDFs are still very much in the growth phase of their existence, a time when price discovery remains important and volume on the public platforms beget volume. How long this lasts, however, is an interesting question, because talking to sources in the market they report more emphasis on the bilateral relationships in NDFs, with dealers willing to make tighter prices over the single bank channel than elsewhere.
We should not actually be surprised by this, I discussed something similar in this columnlast month, and where the G10 markets go in terms of automation, the NDF market generally follows. Actually that is wrong, because long gone are the days when banks distinguished between the two in terms of how they build and stream a price – so NDFs move very much in step with G10.
It is important to understand that the multi-dealer channel remains the most popular, but given this apparent shift towards bilateral, disclosed trading, I am a little surprised it appears to be at historical lows in terms of its share of volume. The single dealer channel is at historical highs – and that is very much to be expected – but the ECN model in the UK is also hitting new highs in terms of percentage of volume, so what does this signify?
Of course it is very early days to be making assumptions on this stuff – perhaps another five years’ worth of data would improve matters – but I wonder if what we are seeing in front of us is the result of easier aggregation? It used to be that the multi-dealer platform represented the cheapest way to aggregate liquidity, but is it now?
Access to a good aggregator has become cheaper as competition has grown – is there anyone out there not trying to push some piece of aggregation kit? – and that means that the multi-dealers need to be aware of their pricing.
They also need to be aware of how many dealers they are putting in competition because having too many creates the “winner’s curse” involved in some models, wherein the dealer only “wins” the deal because they are likely to lose money on it. Of course, the deal may suit that dealer’s book, but there is still a growing awareness amongst the dealers that not all multi-dealer channels are equal and that means they price differently accordingly. After all, if a platform has 50 or more dealers streaming a price in a popular NDF the value in being part of the stream has to be limited surely?
For a customer with a disclosed, limited supply, liquidity pool, however, the NDF world – as in G10 – would appear to be their oyster. Maybe because the market has by-passed some of the angst felt in G10 circles over the past 10 years when we spun through the ‘volume is everything’ cycle, NDF is moving to a different model much quicker? Certainly talking to people indicates a much more comfortable feel with how the NDF market operates.
All of which leads me to wonder whether, after a decade of following the G10 model, NDFs are actually showing us the market structure of the foreseeable future? Will we see the disclosed but limited, liquidity pool dominate for customer trading, and the ECN model thrive thanks to the dealer hedging?
Suggesting that the multi-dealer platforms could be challenged is a brave, possibly foolhardy call, because this segment has such a strong hold, but within those providers I wonder if there will be a further shift towards the bespoke liquidity pool model? The broader model will not go away, but it may see its dominance eroded to the level that it remains the number one channel, but not by as much as it once was.
The key will be cost. Competition is growing and firms are keen to grab any edge they can (on which note two or three have told me they are eyeing the Fastmatch customer base keenly since Dmitri Galinov left), and that inevitably means pressure on prices.
Historically the multi-dealer platform industry has been able to withstand calls for price cuts because it represents a great source of customer flow and, more importantly, the level of competition simply wasn’t there. Smarter aggregation, better liquidity management (on the part of both provider and consumer) and easier connectivity suggests to me these firms will be less able to knock back questions over cheaper pricing going forward.