I was not surprised to hear that another two FX platform providers are working on the delivery of a mid-market matching, or ‘dark’ mechanism to add to their suite of services. Inevitably given the (apparent) success of BGC’s dark pool MidFX, this is an avenue that they should explore, just as FXall and Hotspot have over the past few years.
Given that few platforms rarely launch anything without extensive customer feedback and demand, one has to assume these launches have been driven from the customer base, but that leaves me with one question. If the customers are demanding ‘dark’ trading, why has there been no serious challenge to BGC?
It’s more than a decade since Dave Ogg at LavaFX introduced the concept of dark trading to FX, (although, as some will tell you, FX is predominantly dark because it takes place on a bilateral, closing system basis), and in that time the only mid-market mechanism to gain real traction appears to be MidFX. Volumes have come and gone on the others, but anecdotal evidence (funnily enough in the era of total transparency ‘dark’ volumes are not published!) suggests they have mainly gone.
I remain a little mystified at the lack of success of these mechanisms – I was a very public fan of the concept back in 2007 – but perhaps the increasing awareness of market impact is leading more participants down the road to dark matching? Certainly the original intention of the concept was to allow banks to match off large risk with each other – and that remains the ethos of MidFX – but with more customers looking at market impact I suspect demand is growing, and this is where the other platforms can gain an edge.
I have reached out to both firms that I have heard are planning a new product, but neither is ready to comment on the plans so it is clearly early days, but the likelihood is that both will differ from MidFX is some fashion, be it the user base, the matching methodology, the cost perhaps, and/or the broader functionality around the trade. I think both have to face up to the fact that it will need to be slightly different, because the world doesn’t need another MidFX – it needs an innovative approach to reshaping what was an innovative idea.
Anytime I talk to someone in the banking world about it, I hear that MidFX volumes are good, they’re not published of course (hence the “apparent” in the first paragraph) but BGC’s latest regulatory filing indicates the firm handled just over $36 billion per day in “fully electronic FX”. Given how it operates a hybrid model for most of its core FX swaps and derivatives business, I think it’s a fair assumption to day that the vast majority of that 36 yards is on MidFX, which validates the concept.
Although there are one or two instances, the history of electronic FX tells us that it is very difficult taking volume away from an incumbent. There was a period, fairly recently, when the influx of tech-savvy, largely non-bank firms meant that some volume shifted away from the incumbents and helped boost new entrants who were perceived to have better tech, but generally speaking it’s hard. And it’s not just in the multi-dealer world, for there are multiple instances of a bank halting investment in its proprietary pricing and risk technology only to see no discernible impact on business.
This means then, that the new models have to be different – institutions are embedded with MidFX and unlikely to move away. Why should they when it works well?
There is, of course, one very obvious way to differentiate from MidFX – allow non-bank firms – including buy side customers – on. Some other providers have tried this with the aforementioned limited success and the challenge here has to be that banks will be unwilling to meet with many of their customers in a dark environment and therefore support from that sector will be lukewarm. There is also the factor that for all the aspirations of client-to-client matching in the industry, the reality is very different. Customers do not trade at the same time of day in opposite directions and rarely are they content to hold market risk while waiting for a match.
The new entrants, therefore, will have to overcome this not insignificant challenge and succeed where others have failed. There are, however, reasons for optimism.
Firstly, we are talking slightly different customer bases and cost models. It could be the cost factor that has been holding people back and it could also be, as I noted earlier, that more customers are interested in the concept to help minimise market impact.
For some banks also, there is a reason to be on such a mechanism because for all the chatter about servicing the client, for many now, risk transfer has become “use my algo” because the bank’s FX trading business is not allowed, or built for, holding large risk.
It is interesting to talk to people on the user side about their choice of LPs because as market impact and signalling risk has become more of an issue for the buy side, so they have discovered – through analysis – that some pretty major LPs are actually pretty poor at handling larger risk. These LPs are quick to exit the position or skew their pricing on public platforms – all while the customer is still trying to finish off the order.
For these banks, who are handcuffed by a lack of investment or support from very senior levels, a mid-market mechanism on which they can meet some of their clients could make sense. They may not take all the risk on board, but they will take what they can – and that means they stay in touch with the customer (very loosely I accept) and the latter can execute in a more efficient fashion by matching off to multiple LPs at mid.
One other factor in favour of a dark mechanism is how the concept is already being validated in the industry – and not just by BGC. The increasing interest from clients in accessing the internal pools of the major LPs is a signal the buy side wants this. A key to minimising market impact is (and has actually always been) executing the minimum amount in a public environment. If an LP can internalise 95 million of your 100, that’s a compelling proposition.
So the model is being validated by the increasing number of clients seeking to play in the LPs’ internal pools, so it is natural that the next step is to democratise the mechanism by allowing anyone to match off with anyone.
I am not sure how well this will go down with the regulators or some “LPs” who rely upon public market data to compete – after all we are constantly told the drive is for more transparency, not less. But the customers are voting for this mechanism with their business – much as they did in equities markets when dark pools started there – so should we, as an industry (including regulators) tell them they can’t?
I don’t believe there should be an issue for regulators anyway, because as I have long stated, since the first dark pool was mooted, the key to it is the matching rules. Make the rules totally transparent and auditable and I don’t see why there should be a problem. Yet again, it’s an example of what we should really be striving for – transparency of action.
So at face value when I first heard this chatter in February I was sceptical – and I have to say there are algo execution providers and some LPs who rely more on the public market who are unhappy at the prospect – but as the chatter has increased in volume, and the whisper underpinning it say that the models are slightly different, I have changed my tune.
I don’t think more dark mechanisms will bring new players to the FX market, but I do think, if they are done right, then existing customers will have more confidence in executing larger tickets, and do more volume.
There have been fears that the apparent retreat of some banks from the top end of the FX table would exacerbate issues such as market impact, but if these dark mechanisms can become popular enough to mitigate that risk, the industry as a whole will be in a better place.