The Last Look…
Posted by Colin Lambert. Last updated: January 25, 2022
The last few months have seen more crypto applications and ideas deployed into the FX market, most prominently in the payments and settlement space, and people are now starting to wonder if the really big change is coming in 2022 – in data.
I have spoken to several people over the past few days about the Pyth white paper and its idea for distributing market data and while many agree with my initial view, as published in the aforementioned story, that it has merit, the consensus seems to be that in FX the model would have to be tweaked.
As I noted in The Full FX View in that story, the model of packaging up data and selling it back to the LPs is one that frustrates many in that industry segment, but there has never really been an attempt to disrupt it. Some platforms provide market data for free, of course, but there are questions over its scale and, frankly, worth, given it is often made up of smaller trades that do little more than create white noise, and the pricing is often subject to last look.
Sources at some of the top LPs have expressed an interest in the model promoted by Pyth, for providers to be paid for delivering high quality and accurate data, and certainly it would be a radical shift in FX as several LPs could shift from paying to getting paid, but could it work?
Probably the first thing to note is that the technology upon which it would be based needs to be significantly better than it is now. High profile outages in the crypto ecosystem only serve to highlight how inadequate such technology is for a market that deals in milliseconds and at very tight spreads.
There are, however, plenty of consumers of market data that don’t need a real-time or 25 millisecond feed, and therefore they could be open to a challenger. For the LPs such a solution would become a potential revenue source to offset at least part of the data costs they will continue to pay to their existing suppliers.
There is an assumption that the platforms could themselves contribute data to the new network – indeed that happens in the Pyth ecosystem – which introduces another interesting development, who actually owns the data?
There was a report from The Desk this week about how Citi is laying claim to the data from all credit and fixed income trades in which it is involved. The bank is reportedly targeting on-sellers of data and while there is less activity in these markets compared to FX (they also trade differently of course), the claim of ownership and control will be followed with interest by other traders. If the fixed income data is found to be owned by the price or liquidity provider, it will only be a matter of time before similar debates occur in FX.
The challenge would be how to get a platform to agree that it doesn’t own the data it is selling. The LPs would have to take a very tough stance to get such an agreement through, probably including the threat of withdrawing from the venue, and that then brings them into potential conflict with their customers.
The fairly unique role that last look plays in FX markets would also be a factor in any change in how data is distributed and paid for. If the LPs decided that taking on the platforms would be a step too far, they could easily distribute their proprietary data, which would include, of course, the huge amount of internalised flow, via a competitive mechanism. This data would prove very valuable to those users not in the pricing business.
What will be interesting, however, is when – and it does feel like a “when” – the technology improves to the extent that it suitable for use in the FX environment. In those circumstances a genuine alternative emerges and there is not only increased competition on price, but the revenue stream suddenly goes two ways rather than one.
Again though, there is a wrinkle. Specifically looking at the Pyth concept, the provider is responsible for the quality and accuracy of their data, and can be effectively sanctioned and fined for data that skews the mid. If a consumer (and this can be another LP) feels the data is being skewed by a provider’s stream (which is last looked perhaps and not as accurate as a firm price would be) they can put in a claim for compensation.
The challenge will be what constitutes skewing. The Pyth idea is a user forum with independent back-up, but as anyone who has experienced a committee will tell you, getting consensus can be very difficult in the best of circumstances – in the OTC FX market where skews abound for a variety of reasons, this could be a very busy, and contentious, committee!
I was speaking to a proponent of the Pyth idea late last week who believes we could see a bifurcated data business where the existing channels remain in place, but the business model is challenged for those providing data for post-trade analytics applications. That is fair enough, but it could be argued, as I did of course, that such a model already exists in the numerous independent TCA and analytics providers?
In reality, the only change that is being proposed is the technology used to deliver the data. For a buy side firm looking to keep a check on their LPs and conduct the appropriate checks on execution quality, solutions already exist, so a Pyth-type model is neither here nor there to them, because they would pay either way.
What will be interesting, however, is when – and it does feel like a “when” – the challenger technology improves to the extent that it suitable for use in the FX environment. In those circumstances, the big consumers of data, the major LPs, have a genuine alternative and there is not only increased competition on price, but the revenue stream suddenly goes two ways rather than one.
Ultimately, the technological shortcomings of the DLT mean I don’t see things changing dramatically in 2022, but for 2023? That is a different story. As the major LPs look even more closely at their businesses this year, one senses the data story is going to build and gather increased momentum.
We may not see a major structural change, but I suspect that, yet again in FX, the incumbents need to prepare for a different world and, importantly, a different revenue model.