We should not be blind-sided by the relative calm at this week’s quarter end WM Fix, the fact is FX market conditions are as bad as I can remember them in my almost 43 years in the industry – and that includes countless multi-big figure days (or hours sometimes). The difference when those conditions hit back in the voice days was that while prices were wide, the events were fleeting by comparison and liquidity was there in reasonable size. As such, normal deals could be executed, admittedly with slippage, but it was reasonable.
Compare that to this week, which started off with shenanigans in USD/CAD, these then rolled into the AUD, the madness of the Tokyo Fix on Tuesday, and then, as I write this, AUD has just ramped up almost 100 points in less than a minute as someone apparently tried to buy what would normally be called a modest amount (over the next hour the ramp was retraced back to original levels – welcome to FX 2020).
I suspect that the numerous warnings about the quarter end – and the Global FX Committee should be praised for its timely intervention last week – led sufficient buy side firms to adjust how they hedged by either going earlier or over a longer window than the WM 5 minute 4pm Fix. Sanity, it seems, won the day, if it only, at this stage, one day.
The same cannot be said for the Tokyo fiscal year-end Fix earlier where, as our report noted, it was mayhem as markets moved 100 points or more up and down in less than a couple of minutes. Clearly the Japanese market hadn’t got the memo as it chased the rate up and down leading into the Fix, which, a source tells me was pointed out to them by a local player, is a ‘reference rate’ and not a ‘benchmark’. I don’t know about you but that sounds like someone knew what was going to happen and they spanked the market in the minutes leading up to the rate set? Normally this would lead me to call for an investigation of sorts, but to be frank, such things are going on so often it’s hard to keep up – witness Cable just before the quarter end WM Fix.
There is one thing worth pointing out about that Fix while we are here, however, and that is with all the positivity around how there was no catastrophe, it still wasn’t a great Fix. Now readers would expect me, a sceptic over the mechanism, to say that, but I am able to access something much more authoritative in the form of comparative analysis I have been able to attain that was conducted by BestX. This looks at execution performance this quarter end compared to the last and finds that overall spread costs (measured as spread to mid at benchmark time) were approximately five times compared to end Q4, thus it was clearly a much more expensive roll (the costs include both spot and forward components).
The analysis also found that overall slippage to the 4pm London WMR Fix this week was approx. three times worse on average compared to end Q4, although it was interesting to see that in terms of currency pairs it looks like liquid G10 in many cases had slightly better performance relative to WMR compared to Q4, but less liquid pairs suffered more. This is to be expected of course, because liquidity in these pairs is even worse than the “majors”, indeed as the analysis highlights, for example, slippage for Scandi currencies approximately doubled, and many crosses experienced significant slippage increases.
Further investigation finds, to me, evidence of the deteriorating liquidity picture in that the February month-end performance in terms of slippage was flat (and Feb 28 was probably the first really hectic day we had in this current spell) but by the week before the quarter end, slippage to WM was up to two times before hitting three times on Tuesday.
It is worth reiterating, in my view, that what we are talking about here is a comparison with what is normally the worst quarter end of them all, December 31 – a day when, until recent events, staffing levels were historically at their lowest, along with risk appetite. So the end of March Fix was even worse than the one that takes place while the world prepares for the biggest party of the year!
It is also important to note that while the basic numbers are not huge here, the amounts being transacted and subject to the slippage are – we are talking tens of billions of dollars (or equivalents).
Earlier this year BestX released a white paper – you can access it here – which formed the work behind its latest fill position metric which was released at the weekend. The paper took the fill position logic and applied it to the WM Fix, and, in a nutshell, it expected the fill position score (which is ranked by comparing actual fill rates with tick data to see if the algo hit the best price available) to be close to 50, which would suggest executing at the Fix is the same as at any other time of day.
This is indeed what the research found, however what interested me was how BestX also found that the variation of the fixing price was approximately three times larger than that of a TWAP over a longer period. As the paper notes, “This suggests there is opportunity cost in waiting for the fixing window to execute.”
Not for the first time I may be mis-reading something, but it strikes me this is yet more evidence in the argument that benchmark fixes – and by association fixing transactions – should be measured and executed over a much longer window. We need to break this obsession with the five minute window which, while it is immeasurably better than the previous one-minute iteration, is still flawed.
I have mentioned repeatedly in this column a concern over what I perceive to be a lack of responsibility among the buy side when it comes to executing their FX hedges, especially at the Fix. As I noted earlier this week in my Monday column, WM is undoubtedly capable of calculating using a longer window, so why are we not considering it? After all, it’s not as though there are no independent firms capable of measuring and assessing execution quality – this is not 2015, it’s 2020 and there is a TCA cottage industry out there!
Analysis such as that described above tends to dispute the assertion from some buy siders that the risk involved in executing over a longer window is not worth it – clearly they prefer to have market impact over a short window than benefit from better overall liquidity.
Equally, I find it hard to accept the argument that there is no real alternative to the Fix because net asset values are established using WM rates and any difference could hurt investors. This kind of misses the point I am trying to make – I have no problem with managers using a benchmark to establish NAV, nor do I have problem with them using algorithms to match, as near as possible, that Fix.
My problem is that they insist on using a highly-publicised five minute window at one of the less liquid times of the day with Europe going home and London preparing to do so. Surely with all this independent TCA available (of which WM Company can easily be one) they can instigate, manage and check, a more flexible Fix? One that operates, for example over an hour, or even eight as someone suggested to me today, so that the whole world does not see them coming? In this fashion they can have a benchmarked rate for the NAV, but crucially it will be one that by executing to match, they will not send the market into a potential tailspin.
Due to unreasonable and unethical, some argue criminal, behaviour on the sell side several years ago, the FX industry has got much more serious about how it handles Fix flows in recent years. By doing so, however, it has highlighted a flaw that has been exposed by the changing FX market structure, one that needs addressing. All of which leads me to ask, if one part of the industry is taking the FX Fix seriously then why can’t another – the people actually using it?