The Month-End Fix in 2023 – Lower Numbers but Still Costly

Analysis of the 12 month-end fixes in 2023 suggests that market impact was slightly reduced from the 12 months previously, however both in terms of average and absolute savings, there is still a considerable difference in execution outcomes between the WM and Siren FX methodologies.

The headline numbers remain stark, for those funds aligned with the month-end net flows, there were absolute savings of $6,904 per million by using the 20-minute window compared to the more-widely used WM five-minute window across nine currency pairs. The average over 2023 was $575.33 per million for the same portfolio of currencies.

While there may be some who can dismiss the larger savings as coming in the less-liquid Scandinavian currencies, more concerning will be both the month-end average saving in USD/JPY and Cable of $562.92 and $519.58 per million as both are much more liquid. On a gross basis, the longer window produced potential savings of $6,755 and $6,325 per million respectively.

It is also notable that even at 60% aligned with the fixing flows, the longer calculation provided gross savings of $1,382 per million over the 12 month-ends.

There is some evidence that the market may be adjusting how it handles the flow, or more pertinently that managers are seeking to hedge more flow away from the WM 4pm Fix, in that the average savings in 2023 were – with the exception of USD/CHF – all lower than the 33-month average since The Full FX started tracking and publishing the data, provided by Siren FX. In particular it was a better year for EUR/USD with a monthly average of just over $411 per million, compared to the 33-month average of just over $601 per million.

At no point did the data indicate that users would have been better off with the WM window – that has happened in just one occasion in the 33 months – however in September, the difference between the two methodologies was just $8 per million in USD/CAD, this dropped to $2 per million at 60% alignment.

This was the lowest potential saving in 2023 from the longer calculation methodology, the highest was recorded by $2,318 per million in USD/CHF, in November. In the three major pairs, the lowest saving of the year was $47 per million in USD/JPY – again in September – while the highest was also in USD/JPY, at $1,487 per million in December.

Are Things Changing?

It is hard to pin down an answer to this question because of the lack of hard data, but anecdotally, sources say that demand for the 4pm Fix continues to grow steadily. There are, however, indications that some customers are thinking more carefully about how they execute around month-ends.

Multiple banking sources say that funds are delivering “month-end” fixing orders for execution in the days before the month-end fix, in the hope of avoiding some of the excesses of market impact. These funds are, according to the sources, estimating their month-end flows and executing ahead of time, with a small “adjustment” trade actually on the last day of the month to balance the books.

The problem remains, however, that only managers with a degree of flexibility over how they hedge are able to do this, more generally, funds are still, as one source puts it, “blindly pumping their orders in on the last day of the month”.

It is a small sample, but the reduced impact from funds spreading their business (or indeed using other fixes or methodologies) can be seen in the fact that in 2023 there were four months when none of the nine currency pairs tracked by The Full FX delivered a saving per million above $1,000. This compares to just one month when this happened in 2022 – the first full year in which the data was tracked. Equally, as noted, the averages compared to 2022 are still lower, and there were only two instances of savings above $2,000 per million, compared to five instances in 2022.

Room for Improvement

Notwithstanding the lower numbers in 2023, there remains little doubt that the shorter five-minute window throws up challenges for the FX industry. On multiple occasions during the year the evidence pointed to the amount of flow directed to the Fix being way in excess of what the market could reasonably handle in a five-minute window.

It should also be reiterated that in spite of the numbers being lower, the savings are still significant – a fund at 70% correlation would still have realised $2,761 per million in savings from using the 20-minute calculation methodology. This translates to almost $1.4 million dollars in savings for a fund executing just $500 million per month – the actual numbers are much bigger than that.

Savings from comparison of Siren FX and WM month-end fixes

The ultimate challenge remains, however, how to better capture the actual hedging flows, rather than the theoretical flows. A question that has yet to be fully answered – usually the answer hides in a field of ambiguity over what is “hedging” and “pre-hedging” – is why, when orders executed over, for example, a 20-minute window around, again for example, 2pm London time are subject to a TCA analysis that incorporates every trade, are those orders executed over the same time horizon around 4pm not checked in the same way?

There is little doubt that a number of executing parties, with full disclosure, execute their fixing flows over a longer window than five minutes to avoid excessive market impact, but currently there is no TCA provided for those trades – therefore, while they do provide a benchmark of sorts, it is very far from best, or even a fair reflection of, best execution.

The impact of the (pre)hedging is to be found in the data between the Siren FX and WM fixes, because the former captures a large part of that hedging activity. It is reasonable to point out, as some do, that the very timing of the two fixes means Siren will always outperform because it leads into the WM window, but that does not negate the argument that it is a more accurate Fix because it captures more of the billions of dollars actually traded for that mechanism on behalf of customers.

Ultimately it is hard not to maintain the suspicion, that has been around for many years, that the end investor is not best served by there being a five-minute window. For while there may be other factors that would dilute the outcomes, at the very best the likely difference is in the hundreds of dollars per million – and that represents a significant efficiency gain that would benefit the end investor.

Empirically, the data suggests that a better reflection of the actual fixing flows is found in the longer window, and therefore that represents a better benchmark. Heading into 2024, however, we are yet to see anyone really seek to change, or even question the current thinking. Change for change’s sake is rarely a good thing, but using a longer calculation window would not only seem to be a positive change for investors, it would also better reflect the reality of the FX market functioning.