The Last Look…
Posted by Colin Lambert. Last updated: July 6, 2026
It’s been a while, I know, but we do really still need to talk about the London 4pm Fix, especially at month-end, for, in spite of seemingly a good number of users shifting their behaviour, end investors are still paying a hefty price.
Regular readers will know that unfortunately, due to circumstances beyond our control, we had to discontinue our regular analysis of the month-end Fix, but that does not mean it has disappeared off our radar, or as an issue. Take the end of June as an example, without doing any in-depth analysis, such as that we used to publish, it was obvious that the five-minute window cannot handle the amount of flow being put through – and that is in spite of sources telling me month-end flows have actually diminished slightly as some participants moved to different days/hours.
On 30 June, Cable had a 60-point range for the whole day, more than half of which occurred in the 20 minutes leading up to, but not in, the five-minute WM window either side of 4pm. In the 17 or so minutes to the window opening in Cable the market moved from around 1.3240 to 1.3275. It actually exited the window flat to slightly lower, before, 15 or so minutes after the window closed, it was back in the 1.3240s.
It wasn’t only Cable. EUR/USD, a much bigger market, had something like a 50-point range on 30 June, in the 20 minutes leading up to 4pm, it moved higher by 25 points, drifted in the actual window, before drifting lower in the 17 minutes after it closed, to be some 10 pips higher from 40 minutes earlier. AUD/USD did something similar, and even USD/JPY, while it happened later, moved some 20-points in the 5-10 minutes leading up to the window opening.
Best execution? Well, it might be the best way to do it, but the data being used to quantify execution quality is incomplete, flawed and frankly useless
There are some that believe my problem with how the industry is using the Fix (and to stress, LSEG, with its head in the sand admittedly, regularly reminds everyone it is a reference rate only) is one of conduct, leading back to the days of the chat rooms. It’s not, it is an entirely different issue. What is happening at the moment is not about misconduct – people are executing according to an empirically-developed strategy, while others are using publicly-available market data to pick a direction in the market and take a position accordingly.
The problem that I have is the absolute detachment from reality that dictates that we, as an industry, still try to put through tens of yards of volume in a five-minute window. Of course, the optimal word there is “try” because as previously noted, the analysis highlights how it needs to be – and often is – done over 10,15, 20 or even more minutes; not five.
So, to repeat a point I have been making for years, the execution takes place over a 15-20 minute window, but the TCA is only taken for the last five minutes. Best execution? Well, it might be the best way to do it, but the data being used to quantify execution quality is incomplete, flawed and frankly useless, because it has little connection to what actually happens.
We need to stop being “clever” about how we discuss this, and throwing obstacles in the path to reform by worrying about “events” that may influence the market in a longer window; we need to stop hiding behind, “the clients don’t see the need for change”. In the first case, that can happen in a five-minute window just as easily, but the real reason behind that argument is “I won’t be able to make as much money from the pre-hedginghedging ahead of the Fix as I do now”. If I am an executing party for a Fix order, do I really care if something happens in the calculation window? Do I care if this happens during a “regular” algo execution? No, I don’t, because the risk is with the user, not the provider. What I care about is if I pre-hedgehedge ahead of the Fix, I am running a risk (albeit one that will probably only hurt me once in a blue moon).
On my second observation, because the clients apparently don’t want change, well that is because the oversight function, in the form of trustees and/or senior management are either naïve in the extreme or wilfully negligent about their duty to look after the end-investors properly – because the latter are the only people really getting hurt by this nonsense.
Just because things have become quieter, doesn’t mean the problem has gone away – the Fix still needs fixing
We published the comparison data between a 20-minute and five-minute calculation for 52 months, which is a decent sample size I hope you agree. The average saving on the last day of the month across that period was $471.94 in EUR/USD; $562 in USD/JPY; $506.70 in Cable; and across our portfolio of nine pairs it was $555.41. To be clear, those numbers are cost per million.
I have little doubt that the good work of some on the buy side has convinced their oversight or management that there are better ways, or times to execute this business and still hit the precious benchmark, so I would confidently expect the numbers now to be lower than that average.
If they are even half, however, that is still way too much money being given away by laziness and a lack of care for the end-investor. The end of June was expected to be a fairly neutral day in terms of predicted flow from rebalancing etc, so the chances are the amounts were even further down, but the fact remains, even if you were rebalancing to the tune of a yard in Cable, that is $2 million your investors missed out on. It may not seem much, and in the big scheme of things it isn’t, but this happens repeatedly and is cumulative.
Just because things have become quieter, doesn’t mean the problem has gone away – the Fix still needs fixing.
