The Last Look…

Is it time to declare the end of the era of the primary FX venue? I have been commenting on the inexorable decline in the fortunes of Reuters/Refinitiv Matching and EBS Market for more years that I care to remember – in fact I can state it was in 2010 because I still have the scars to show from a “robust” exchange of views with a then-head of one of the venues!

What seems to be different now is the number of people in the foreign exchange industry that I talk to who are voluntarily declaring the era to be at an end. Last week in three separate conversations with people from different areas of the industry (not with competitors I should add) the opinion was expressed that they have had their day, as one senior figure observed, “how long can they carry on like this?”

I suppose the short answer is, just as several other platforms manage to do, they can carry on regardless – it’s just the numbers being generated by the business won’t be as impressive. Whether their exchange group owners are willing to support that is another matter, but given both primaries are re-platforming, the sense is, at the moment at least, the investment support is there.

I also think there is a question of relative scale to take into account. Yes, the primary venues’ ADV have been on an inexorable decline (and I was on the receiving end in 2011 when, after predicting their decline, volumes picked up), but everything is relative. I would argue that what has actually been in decline in the CLOB model itself – thanks to the usual blend of internalisation, technology improvements around aggregation and growing awareness of market impact.

This not a good thing for the industry in some ways. Looking at firm liquidity, I estimate the two primaries hover around, often under, the $30 billion per day mark, but where is the model really growing? CME’s FX volume has declined slightly from the mid $80 billion per day to the high $70 billions and that includes options.

Elsewhere, LMAX Exchange has been on a steady incline and last year reported total trading volumes of $4.5 trillion, taking out a little piece for crypto (which has become a bigger piece of the pie in 2021), and allowing for 250 trading days, that around the $18 billion per day mark. Decent growth, yes, but not earth-shattering in the big picture, it’s probably $5-6 billion up over a two-year period – good for the company, but in the wider scheme of things, with spot FX volumes around $2 trillion per day, less so.

Likewise, CboeFX has been publishing data on its no-last look firm liquidity since the start of 2020 and the needle hasn’t moved. In fact, it has ticked downwards, but to be fair the comparison includes March 2020 when there was a spike in activity – overall the business looks steady, but not showing signs of serious growth.

So yes, the primary venues are in decline, but it has to be assessed against a backdrop of a general decline in the CLOB model – which itself has implications for the industry, especially around data.

Talking to quants, they have adjusted their models to be less reliant upon the primaries, notwithstanding that, however, they are not particularly using other venues’ data that much more – internal flows play a much bigger role. The challenge facing the industry is the still heavy reliance upon these venues thanks to their no-last look policies. Yes, data from CboeFX’ firm pool and LMAX can bolster the pricing framework, but it becomes an expensive business to have multiple data services, especially as they provide smaller incremental benefits as they are amassed (no to mention how, in a consolidating world, the major LPs are paying for data of which they are the major generators, therefore are effectively buying their own data!)

I wonder whether there are deals to be done in the data business, whereby two, preferably more, of the firm venues, agree to a joint data product?

It has become fashionable to term spot FX a commoditised business, and to a degree it is, especially when markets are relatively quiet. I happen to think that the trading environment on the primaries would be a lot better with higher minimum quote amounts – both have tinkered with this, but to my mind in the wrong direction. There is probably more value in a firm liquidity pool with, for example, three million units as a minimum and a few less “LPs”, than there is in one with a million or less and dozens of recyclers.

That’s unlikely to happen, so what can be done?

I wonder whether there are deals to be done in the data business, whereby two, preferably more, of the firm venues, agree to a joint data product – one connection, multiple no-last looked data, split revenue. It would be hard to get across the line and the technological capabilities of the various platforms are, currently, at very different levels, but there might be something there that would benefit the industry and, down the road, the firm venues.

If that was to happen then we would most definitely declare the primary venue to be over, but I am not sure how bothered those venues would be. There is a challenge in that status because so many people key off moves on these venues, often players shy away due to signalling risk.

Another possible way to bolster the business, especially the crucial data aspect, would be for Refintiv and EBS to change the structure of FXall and EBS Direct respectively. If certain streams to these platforms were to be firm, no last look, then that pricing and trade data could be added, making the data source much more valuable. Because don’t get me wrong, the big problem here is a balance of too many LPs not willing to stand behind their price (even to decent customers with reasonable mark out), and the platforms not being brave enough to call them out on it.

If FXall and EBS Direct had more no-last look streams, the real value of their offering would improve dramatically for those LPs at the top end of the market (and could be built into a structure whereby trading volumes are linked to availability of the data, as tried previously by EBS).

There are a lot of platforms out there trying to push their data business, but if the vast majority (or all) of it is based upon indicative pricing – for that is what last look pricing really is – then where’s the value? Data based upon actionable, truly tradeable, quotes, is the firm venues’ differentiator.

Another strategy could be to hang on long enough for the re-platforming of both venues and/or sustained volatility to return to FX markets. The latter will happen, the world is cyclical and while the length of the cycles varies, I am convinced that the nature of markets will change again, especially when interest rates start doing something quite a few traders in the market have never seen, and start rising.

When conditions are tricky enough that superfluous LPs are squeezed out thanks to a combination of rising costs and growing awareness amongst consumers of how a good LP minimises market impact, the firm venues could be in a very good place. We are often concerned about conduct in FX markets, but the fact is firm liquidity pools are generally going to be characterised by better behaviour, and that means losses have to be realised on occasion – that should take care of more than a few so-called LPs!

The primary venues have diversified their businesses, EBS more actively so than Refinitiv, but then it had catching up to do thanks to the latter’s purchase of FXall. I recall predicting to a senior executive at EBS around launch that Direct could become an important, perhaps even the dominant, liquidity pool for the firm, and the NDF business has been a beacon of light in recent years.

The non-firm space is the usual bunfight with last look allegedly policing “good” and “bad” customer behaviour, but how far does this segment grow without a healthy primary source – the firm venues?

Where both platforms have suffered during this period of diversification is in technology comparisons to some competitors. Assuming the re-platforming goes according to plan and is completed early enough, there is optimism there for both.

So, is the primary era as we know it over? Yes, it is. The glory days will never be revived and no two platforms are likely to dominate to the degree these two did for a decade and more. That said, I believe that the multi-dealer spot FX market should be seen along two lines – firm and non-firm. In the former LMAX and CboeFX have done enough for their firm venues to be aligned with EBS Market and Refinitiv Matching, and together the quality and quantity of data could support a good business.

The non-firm space is the usual bunfight with last look allegedly policing “good” and “bad” customer behaviour, but how far does this segment grow without a healthy primary source – the firm venues?

Ultimately this is about the quality of liquidity and data derived therefrom, and if sustained volatility does return, which I have said it will (and if that isn’t a “sell vol” signal I don’t know what is), good (firm) liquidity will attract a premium. If that happens then I would ask one more question regarding the make-up of the multi-dealer market – which model offers the most value to an LP’s business and is therefore the one they would be happy to commit revenues to (and let us not forget that generally the LPs sustain these businesses)?

Non-firm venues often offer limited (if any) yield and charge the LP brokerage; firm venues offer a fair, equal environment for trading (and charge brokerage), but crucially, also provide (at a price) really good data.

The primaries are dead; long live the firm venues!

colin@thefullfx.com

@colinlambertFX