Well, Monday’s column (that, to confuse everyone, went out last Thursday such was my race to publish), certainly poked the hornets’ nest. It’s been a while since I have been involved in such a spirited debate over something and I suppose it was inevitable if it wasn’t last look then it was, after more than a decade of banging this particular drum, going to be the Fix.
So, let’s keep it moving by asking the question, is it time to revisit the recommendations of the 2014 FX Benchmark Group that, under the auspices of the Financial Stability Board, created the current framework?
Before getting into that I want to reiterate, for this was misunderstood in some circles, that I understand the WM Fix plays a vitally important role in one area of the financial markets’ community. While I am still a little averse to the herding that trying to hit a benchmark creates, however, as well as the paradox of managers charging fees for proving themselves “average” (for that is what a benchmark is), I understand there is a need for a measure of sorts.
So it is not the existence of a benchmark process that I so object to, it is the calculation methodology. I need to stress that while I argued for a longer window of 10 minutes at the time, I backed the recommendations of the 2014 review, they were a step in the right direction and did indeed solve the problem of order handling and information security. They also made the 4pm Fix a little less volatile thanks to the five-minute window. What it did not solve, in fact what it created, was an increase in signalling risk as everything went to the agency model.
I would also respectfully suggest that what worked in 2014 (including my 10-minute window), does not work in the current foreign exchange market (putting aside the pandemic even) thanks to changes in the market structure and business attitudes. On the former we have much larger data sets and the analytical tools with which to dissect them, meaning algos can spot flow much easier than before and react accordingly (a pricing engine is not going to stand in front of a runaway steamroller), and on the latter we have fewer and fewer banks actually willing to handle the Fix flows.
The secondary point actually feeds into the first, because what we are achieving is not only signalling risk of the flow from a group of TWAPs entering the market at the same time, but they are coming from an even smaller group of players, therefore the “non-playing algos” can narrow their search down.
I also think it needs to be stressed, because I continue to hear this from certain areas, that no-one is behaving badly here. The customers are being naïve in my opinion, and perhaps, as I suggested last week, negligent, but they are using a Fix that is approved by their oversight and which has a widely-recognised methodology for calculating a benchmark rate. The banks are handling the orders as agents and, having checked with a small number, are all telling their clients if they pre-hedge the Fix and get approval for them to do so. Third parties are using publicly available information to take a trading decision using market data and, finally, WM Company is taking a snapshot of the market that accurately reflects events therein, according to a robust framework.
Who is doing anything wrong there that requires an investigation into behaviour? Nobody of course, but what does require an investigation in my view, is the length of the window. If we can lengthen the window and create an environment in which the flow has much reduced, if any, market impact, that may actually entice a few more players back to the table. There is no small irony in the fact that what has actually happened thanks to those market structure changes is that we have a situation wherein orders are channelled to a small group of banks for them to execute in the window. In the criminal case against the “Cartel” this was seen as “building” and claimed as illegal by the prosecution. Putting aside the obvious differences around how the orders were accumulated, is the outcome that different?
It was pointed out to me, quite correctly, that what we saw last Thursday at month end was a natural function of supply and demand. Yes, it was and we cannot and will never, escape that – it’s why markets exist after all, to reflect that. My argument is that if the window is longer then this supply and demand has more time in which to interact with other flow. We are all repeatedly reassured by the big banks that their internalisation rates for G3 spot FX are in the 80-90% range, what could be more natural that interacting with that flow? Clearly it is hard to do so in a five-minute window as EUR/USD revealed last week.
At this point I will point people in the direction of Saturday’s In the FICC of It podcast on which my guest is Jamie Walton, co-founder of Raidne, who has created a different – and regulated – benchmark. I really enjoyed the conversation and hope you will too, it contains some pretty staggering numbers in terms of what can be achieved by lengthening the window by just 15 minutes.
Personally speaking, I would go longer than 20 minutes, I would take a benchmark measured over hours, it’s not as though it can’t be done for in the 2007 Digital FX Awards I highlighted Lehman Brothers’ Algo Workbench that allowed clients to execute an order over a certain timed window (minutes or hours long) and receive the benchmarked mid-rate from EBS or Matching plus a fee. If it could be done that well in 2007 (and that remains one of my favourite product rollouts in my almost 19 years at P&L), it can certainly be done now.
Equally, I have to be fair in this debate and stress how I don’t think it matters who calculates the benchmark as long as their methodology is sound, robust…and regulated. It can be WM Company or it can be a challenger firm like Raidne. The likelihood is that asset managers, who often regale me with reasons (I prefer the word excuses) why they cannot switch and cite the legal work necessary, would find it easier to have WM change its window, but it doesn’t have to be that way (and incidentally, I would be fascinated to know the difference in legal charges between changing documentation and defending a class action).
I also wonder if, assuming a longer window does gain recognition (and the Raidne numbers in our podcast would suggest that it should) I wonder if that will force WM into a change? It will be in a tricky position because many of its customers seem content, but if the firm’s commercial position is eroded, who knows?
So, let’s get this clear. I am not accusing anyone of acting incorrectly, if anything I am trying to highlight a problem that I think has evolved with the market structure – a problem that needs resolving. I am obviously not privy to conversations in regulatory or benchmark oversight circles, but I would hope that there is at least some debate around the issue – notwithstanding the extant challenges we currently face.
We should not listen to the excuses of users that it is “too much trouble” to change, or that “it works fine”, because it isn’t and it doesn’t. There are times when we need to enforce change and I believe this is one of them, for the reality is the market has been telling the world for some time that the window needs changing. After all, if it worked fine, then why would banks feel the need to pre-hedge and their customers acquiesce to it?