Transparency or Market Impact? The Quandary of the 4pm Fix
Posted by Colin Lambert. Last updated: March 30, 2021
It has cost banks billions of dollars in fines, dozens of people their jobs and remains the subject of ongoing litigation, but the London 4pm WMR Fix remains as popular as ever. Colin Lambert looks at how the flow is handled, and what attitudes are to the mechanism amidst debate over its future.
When the Global Foreign Exchange Committee (GFXC) released a statement in mid-March 2020 highlighting the challenging liquidity conditions likely to be faced by market participants using the quarter-end London 4pm WMR Benchmark Fix, it was dipping its toe into a running sore in the FX industry – market impact during fixing windows.
The Fix had regularly seen significant movement both before and during previous month end windows but with the world in chaos thanks to the full impact of the pandemic being unleashed, there were genuine fears about what could happen at the end of March. The GFXC’s warning is widely regarded as having alleviated conditions at the end of that month as the Fix passed off largely without incident with banks reporting many clients executing their hedges earlier than usual.
The continuing challenge, if it needed reiterating at all, was demonstrated at the end of April 2020, when 50-plus point moves were seen in some major currencies in the lead up to, and during, the five-minute window. Since then, while there have not been serious blow outs as were witnessed in April, month ends continue to be challenging – for example in January 2021 EUR/USD at one stage moved 25 pips before the usual reversion, while Cable saw a 30-plus point move. Slightly smaller moves, but moves nonetheless, were seen in the February month end window.
Proponents of the Fix remain steadfast in the benefits of a transparent methodology for establishing a revaluation rate – for that is in effect what the Fix represents for many users – and the argument that liquidity is at its deepest during the window thanks to the flows it attracts.
It is hard to argue with the transparent methodology that creates the rate, the fixing process is also subject to strict oversight by an independent body tasked with ensuring that methodology is followed to the letter. There may be a case to suggest that more venues’ data is included rather than just those of EBS, Matching and where needed, Currenex, however it is not clear exactly how much value that would add considering the amount of liquidity recycling in the FX market. As a 2018 paper by the BIS’ Market Committee observed, while their share of the market has dropped considerably, EBS Market and Refinitiv Matching are still the main sources of price discovery.
In terms of the proponents’ volume argument, however, there are grounds for dissent. It is most certainly the case that the 4pm Fix represents the busiest time of the day in FX markets, but as an e-FX trading manager at a bank in London points out “There is a big difference between volume and liquidity. If all the volume is the same way and you’re trading with the wind the liquidity available is pretty expensive. Too many people look at the top level numbers and see a deep, liquid market that can handle any amount of flow – it’s simply not like that, especially at 4pm because the flow isn’t balanced.”
The Magnification Effect
The success or otherwise of a Fix window is the level of netting achieved and here data is anecdotal at best. Sources familiar with individual institutional netting levels suggest it is very much a day-to-day matter with no real pattern being established. That said, peak netting levels appear to fall short of 50% of volume most of the time, with one banking source observing that if anything the ratio is in decline. “It could be more passive funds using the Fix that tend to be the same way, but those days when netting levels are higher are often characterised by higher order flow from other sources, especially corporate treasuries and regional banks.”
There are even questions as to the real benefit of allowing fixing trades to be netted ahead of time. In February, CME’s EBS revealed that flows on its eFix platform in 2020 were up 21% on 2019 and that across November it saw flows of $19.7 billion (single count). The platform also highlighted what it said was a record day in March 2020 of $5.775 billion, however back in 2016, EBS also claimed a record day, 29 January, of $6.5 billion. Notwithstanding the latter may have not been single counted, the growth of netting on the platform is likely the result of banks being more willing to seek matches outside of their own environment – and it is not lost on the defenders of the behaviour in Bloomberg chat rooms a decade and more ago, that this is exactly what a lot of the conversations were about, finding a match.
“By not considering how unbalanced order flows distort exchange rates, investors are skirting “best execution” obligations”
On the face of it, netting would appear to offer nothing but benefits to the end user through reduced market impact, however it very much depends upon where the order is handled at the bank or executing broker. “If the orders are handled by a separate team, away from the trading business, then netting is a good idea,” says an execution desk manager at a European asset manager. “If, however, the trading business handles the orders, there are information security issues. If no match emerges for the order then that business knows the direction of the Fix. This is not to say they will deliberately position ahead of the Fix, but their pricing engines and traders will be happy to, for example, hold onto dollars they are given in the minutes leading up to what they expect to be a dollar-positive fixing window.”
Banking sources also acknowledge some managers’ reluctance to allow their orders to be netted, but suggest the vast majority of customers are happy to minimise market impact. “It’s swings and roundabouts,” one bank FX trading source in London observes. “Yes a bank may build a position passively going into the Fix based upon information they have, but the impact will be minimal, especially when oput up against the material savings that can be achieved from netting off flow.”
Again though, this does not find universal support. As a London-based quant analyst points out, “If you’re getting Fix flows netted it means you’re likely trading against the market and would get a better result by waiting and executing at the end of the window.”
Australian fund manager QIC recently published a paper looking at the use of the Fix, highlighting the “illusory benefits of maximum liquidity”, specifically the magnification effect that is achieved at the Fix thanks to the imbalance of orders typically seen. As the paper notes, the reasons for using the Fix are all compelling, “Except for the very significant, largely unrecognised and wholly avoidable market impact costs accompanying rebalancing at the London 4pm Fix.”
Raising the stakes further, the QIC paper also states, “By not considering how unbalanced order flows distort exchange rates, investors are skirting “best execution” obligations and incurring unobserved costs which elicits a material dollar cost to the end investor.”
The QIC paper highlights events in March 2020, specifically on March 12 in the middle of the stock market mayhem that generated large fixing flows, when an investor using the 4pm Fix was selling Australian dollars 2% lower than where the market was just an hour before. The reason for the move is provided in the paper, “pre-hedging activity”.
Pre-hedging is, in reality, a relatively new term in the foreign exchange industry, one that has grown with the size of orders submitted to banks, not just at the Fix, but predominantly so. It remains a controversial activity, not least when look at by someone more versed in the exchange world. “So a broker gets an order, buys or sells in front of the price setting window, but doesn’t pass that onto the end client? That’s front running,” argues an Asia-based futures trader.
Of course, there are nuances in OTC markets that counter that view, especially on those occasions when the pre-hedging is done for the benefit of the client, but the optics are not great. “Personally I would be more comfortable with pre-hedging if the final price included that activity, but a lot of fund managers will not, or cannot, accept anything other than the fixing price,” says the e-FX trading manager. “In those circumstances it is better for the bank to earn a few pips pre-hedging than to ram the flow through the window and create market impact.”
An interesting question when looking at the Fix is why banks continue to offer a product that cost them billions in fines and, at face value, offers minimal opportunity for revenue? Yes, there is a fee associated with the Fix, but as was the case in the first decade of the century, this is being squeezed by competition, and it could be argued no amount of fees will recover either the fines or the outlay required for technology and oversight.
“There is heightened oversight and first line of defence teams monitoring this business, usually staffed by experienced ex-traders,” explains a senior banker in London closely involved with Fix business. “This oversight costs money, so why are banks still offering it? Because they are still making money from it – and not just from the fees.
“This can only be because banks are using the knowledge from the matching process and/or pre-hedging,” the senior banker continues. “They trade ahead of the window, under the guise of pre-hedging, and hope that it gets noticed by the HFTs in the market who will jump on it. They then execute the balance in the window using an algo calibrated to the WMR mechanism, and care less what happens to the market, as long as it keeps moving in their direction.”
“If you’re getting Fix flows netted it means you’re likely trading against the market and would get a better result by waiting and executing at the end of the window.”
The London based quant analyst also points out that “skilful” pre-hedging actually increases the risk to the executing party. “If you do a really good job and the market doesn’t move, you’re exposed to any reversion and a loss. It’s one of the problems with this whole issue – the better job you do of protecting the client’s interest, the more risk you assume. There’s simply not the incentive to do that, so naturally any pre-hedging will be a little more aggressive than it needs to be.”
Sources spoken to at several banks confirmed that their trading business often pre-hedged Fix order flow, and also held onto risk acquired passively in the run up to the Fix, sometimes for their own books. This undoubtedly creates a revenue stream (there are very few who buy one of the original claims by traders involved in the chat room scandal that they often lost by pre-hedging) and as the e-FX trading manager observes, “There’s a reason the market often reverses at the end of the window.”
The QIC paper also discusses this signalling risk aspect of trading at the Fix, identifying speculative trading by firms with no fixing orders as one influence on price action, stating, “Even without [information on equity market flows] our research has identified systematic patterns which can be exploited from observing momentum leading in to 4pm, and mean reversion of exchange rates after 4pm.”
Best Practice Confusion
It is argued in some quarters that pre-hedging is yet another ticking timebomb waiting to go off in the FX market, that the legal system will not understand the nuances and see it as front running – a typical argument being, “how can it be to the benefit of the client when any price improvement is not passed on?”
Others though, point to the guidelines established in the FX Global Code as providing a defence against such accusations, assuming all relevant disclosures are provided along with the appropriate level of transparency.
“There seems to be this illusion that pre-hedging is about everything but the Fix, whereas the reality is the exact opposite – it’s all about the Fix.”
The FX Global Code is very clear that pre-hedging (with disclosures etc) is appropriate if the executing party is acting as a principal, but not when it is an agent. “I think this confuses the issue a little,” argues the FX trading source in London. “You’re executing the order on behalf of the client, according to pre-established rules, and if the market runs away during the Fix, the client pays the price. That’s an agency model, so pre-hedging shouldn’t be allowed.”
A trader in the US, however, has a different point of view. “It’s all about the risk. By sending you a fixing order a customer is not necessarily transferring the risk, but the minute the cut off time for submission is reached, the risk is the bank’s. They know they are going to have to have a market exposure and at what time they are going to have it – the only thing they don’t know is the rate. At the cut off time, these are firm orders, not ‘expected’ or ‘anticipated’ as the Code terms them, so can be pre-hedged – often that is the most sensible approach anyway.”
There is also some debate as to the granularity required in a disclosure to adequately meet the requirements of the Code. Some people observe that Principle 11 of the Code doesn’t even refer to fixing orders beyond one example in the Annex and therefore their actions are covered in their general disclosures; others disagree. “I would prefer an environment where a bank has to explicitly state they are going pre-hedge an order; make it the exception, not the norm,” says the e-FX trading manager. “There seems to be this illusion that pre-hedging is about everything but the Fix, whereas the reality is the exact opposite – it’s all about the Fix.”
Fixing the Fix
Those that argue against using the Fix often point to the benefits of a longer window to dilute the flow, this would allow banks to make the handling of these orders an agency matter. Market impact would be diluted by the longer window and any associated speculative flow would have less certainty and, importantly, reduced profit opportunities. This would all lead to less directional moves during fixes.
Others, however, believe a longer window will still create the same issues. “While they may get less bang for their buck, speculative traders could still pick out the direction of the flow in a longer window, you still have issues of information leakage,” the London-based quant analyst says. “It would be less of an issue than now, but would not necessarily be the panacea some believe it to be – especially at month ends when flows are significantly larger.”
It can also be argued that there is little in need of repair, as several sources pointed out when discussing this article, banks often beat the Fix without pre-hedging and face the dilemma of what to do with the excess revenue generated. Customers often only want the Fix rate and will not accept rebates or price improvements, and cutting fees to compensate is likely to incur the wrath of the authorities who explicitly stated their desire to see fees maintained at a sensible level.
A lot of this price improvement comes from the fixing algos accessing internal axes and matching off quietly, in front of mid – and this dilutes the argument of those that believe fixing orders should be removed from the FX trading business entirely. To a degree, fixing orders are already removed from the trading businesses, the challenge for some is building an infrastructure where they are able to interact anonymously with internal flows, for while the technical issues are easy to overcome, as the FX trader notes, “You can be anonymous all you like, but if the pricing engines see a pattern of behaviour they will react whether it’s from internal or external sources.”
What is clear, is that the Fix remains popular, and that raises the question, are managers actually listening to the advice of bodies like the GFXC and assessing its relevance for their business? It is not obvious that they are, but as QIC notes in its latest report, “Systematically executing FX hedge rebalancing trades at the benchmark closing rate of London 4pm does not represent ‘best execution’.”
The fund adds, “Investors need to challenge the use of the Fix for execution purposes and instead pursue execution strategies which meet their objectives. It will require more rigour, but the reward is the conversion of very material and unobserved execution costs evolving into visible implementation ‘profits’ relative to the industry benchmark fix.”
The biggest challenge, bizarrely in the view of many, is actually convincing money managers they actually want to make this extra money.