Speech FX Benchmarks

Today I will talk about benchmark rates in the foreign exchange (FX) market.

FX benchmarks are an important part of the financial market infrastructure. They are referenced in many financial contracts and are particularly prevalent in global bond and equity market indices where they are used to aggregate markets into a common currency. The longstanding approach to do this has used a set of foreign exchange rates calculated at a particular time of day, generally 4pm London, the London ‘fix’, which is published by WM/Reuters (WMR).[1]

Attention on this arcane part of the financial market infrastructure has increased significantly over the past couple of years as concerns mounted about the integrity of benchmark rates and the possibility of market misconduct around the benchmark fixing process. Following investigations by a number of authorities, some of these concerns have proven to be well founded, and investigations are ongoing.

Reflecting the emerging concerns about the FX benchmark process, in early 2013 the Financial Stability Board (FSB) tasked a group chaired by Paul Fisher of the Bank of England and me to formulate a set of proposals to improve the benchmark process and reduce the scope for manipulation. Our work was conducted completely separately from the investigations into allegations of FX manipulation and we did not have access to any of the evidence gathered by those investigations.

The Foreign Exchange Benchmark Group (FXBG) was comprised of members from all the major financial centres. We talked extensively with most parts of the FX industry to get their views, including asset managers, index providers, FX platforms, corporates and banks.

A preliminary report was released for public comment in July last year and we received a good amount of constructive feedback on it.[2] The final report of the FXBG was presented to, and approved by, the Financial Stability Board last September.[3] I have had a number of people ask what status the report has. The answer is that it has been approved by the FSB, a body which consists of heads of central banks, securities regulators and ministries of finance/treasuries of all the major jurisdictions.

The report contained 15 recommendations aimed at improving the benchmark process and reducing the incentives for manipulation.The recommendations garnered broad support across the industry. We are now several months on from the release of the report, so it is useful to see what has been achieved in that time. My general assessment is that there has been significant progress on some recommendations but little on others.

The recommendations addressed both the structure of the benchmark process as well as the conduct of participants in the process and can be divided into four main areas: benchmark methodology; execution infrastructure; market conduct; and guidance on central bank reference rates (which I won't cover today).

The group believed that the recommendations it made would be implemented by the market participants concerned and indeed, as I just said, they were supported by market participants at the time and endorsed by the FSB. There is a strong expectation they will be implemented. The recommendations are, however, not explicitly embodied in regulation.

Nevertheless, the group stated that if these recommendations were not acted on, authorities could conclude that a regulatory response was necessary to generate the desired improvement in market structure and conduct. Moreover, as a result of the investigations into the alleged misconduct, a regulatory response may be required and indeed, in the UK, FX benchmarks are now a regulated activity. The UK Fair and Effective Markets Review (FEMR) has also actively sought feedback on such issues.[4]

The Calculation of the Fix

The first set of recommendations concerned the calculation of the London 4pm fix. For a number of years, WM has calculated the fix over a one minute window centred on 4pm. WM calculated the fix using data from either Thomson Reuters or EBS (depending on the currency pair). It took snapshots of trade and order rates at one second intervals and calculated the (unweighted) median rate over the minute window. That is, the fix is calculated based on executed prices, in contrast to LIBOR which was a submission-based process. Hence the issues which we needed to address in the case of FX benchmarks differed from those associated with the old calculation methodology for LIBOR.

While we concluded that the WM benchmark calculation methodology was essentially sound, the group recommended widening the window from one minute to enhance the robustness of the fix calculation. The wider window should reduce the capacity for manipulation while still generating a replicable market price.

The group also recommended that rather than use generally only one price feed, WM incorporate feeds and transactions data from more sources into its calculation, provided the additional sources were of sufficient quality and representative of the market. While arbitrage should ensure that the single data sources used by WM should be representative of the market, we thought it relatively costless to use a wider range of sources and this would be beneficial in getting a ‘truer’ picture of the market over a short time frame.

WM has subsequently announced that it will widen the fix window to five minutes commencing on 15 February.[5] It will also utilise additional price feeds from that time.

In addition, the group recommended that WM form a user group to consider future changes to the fix calculation methodology. In announcing its recent changes to the calculation of the fix, WM sought feedback from the industry and is in the process of establishing such a user group. Finally, WM has agreed to follow through on a number of recommendations made by IOSCO which assessed WM's fixing processes against the IOSCO principles for benchmarks.[6]

Infrastructure of the Fix

The next set of the group's recommendations addressed the market infrastructure around fixing trades, that is, how fixing trades are executed in the market.

The aim here was to reduce the incentives and opportunity for misbehavior generated by the market structure. The FXBG report documented the concentration of activity through the fixing window. It noted that trading volume at that time was up to 10 times larger than at other times in the trading day. A major source of the large volume of these trades is likely to be asset managers seeking to rebalance their portfolios in line with the indices they track that incorporate the London fix in their calculation. One way to minimise this tracking error between their portfolio and the foreign exchange component of the index they are tracking is obviously to execute at the fixing price itself.

Orders for these trades and other fixing trades are provided by the client to the FX dealers ahead of the fix (often this time is quite short). Importantly, the dealer typically agrees to execute these orders at the, as yet unknown, fix price and often does not charge either a fee or spread on the trade.

These arrangements differ from standard principal-based risk transfer services offered by dealers, where the risk to the dealer is mitigated by using the current market price and applying a spread. The arrangements for fix-related orders mean that the price risk on the transaction is transferred entirely from the client to the dealer without any compensation. The dealers then have to manage the risk associated with this flow.

The trades conducted as a result of this process can generate optics of dealers trading ahead of the fix, even if the dealer is simply seeking to manage the risk. But these dynamics can also create an incentive for dealers to manipulate the fix rate to generate profit from what would otherwise be potentially a loss-making exercise.

To address this problem, the FXBG supported the development of industry initiatives to create independent netting and execution facilities for transacting fix orders. A number of such products have been launched on the market in recent months. They generally have the feature of maximising netting opportunities and then executing the order in a way that delineates the separation between the dealer trading on its own account and the dealer transacting on behalf of the customer. It remains to be seen how much traction these products will get.

In our preliminary report, the group had sought market feedback on the development of a central netting utility to maximise netting opportunities and reduce the need for customers to provide orders in advance of the fix to dealers. While some of the feedback was supportive, others confirmed some of the potential practical problems that such a utility might confront, particularly around the execution of the residual positions once all netting opportunities had been exhausted, which would determine the fixing price.

Market Conduct

The above recommendations address the structure of the foreign exchange market around the fix. The FXBG believes the recommended changes will improve outcomes but at least as importantly, it provided a set of recommendations about the behavior of market participants on all sides of the market: dealers, asset managers, index providers, to address the possible incentives for manipulation.

First, the group recommended that the FX dealers charge for fixing transactions in a more transparent manner (rather than for example, relying on the business being subsidised by activities in other parts of the dealers' institution). This charge might be implemented either through a fixed fee or a spread. A fixed fee would be consistent with an agency mode of fixed execution. A spread is consistent with compensation for the risk transfer where the bank is acting as principal. The shift to a wider window for the fix calculation means that the risk transfer is likely to be larger for the bank in being able to replicate the fixing price.

While a number of banks have started to discuss this with their customers and are in the process of moving to charge for this service, this recommendation is not being adopted universally at this point. Not surprisingly, there has been push back from some customers against paying for something that had been previously offered for free (at least notionally).

So long as some dealers remain willing to not charge directly for this service, competitive forces mean that it is difficult for others to charge. There is clearly a first-mover disadvantage. More progress is required here. When WM implements the wider window later this month, this may prove a focal point for a shift in pricing models. There have been various media reports in recent days suggesting that a number of banks are doing this. However, if the fixing service is not directly and transparently charged for, the incentive for inappropriate behaviour remains.

Second, the group recommended that banks and other FX dealers separate their fixing business from their regular business. The primary aim of this recommendation was to reduce the likelihood of inappropriate use of the information obtained from the fixing orders. Clearly, such a separation is not costless and it does reduce risk absorption opportunities. Nevertheless, the group's assessment was that the benefits would outweigh the costs. Some banks have implemented this recommendation over recent months but adoption has been somewhat mixed.

I have received a number of questions about what exactly separation entails. We did not intend to be too prescriptive here. The main principle is the appropriate segregation of the information flow. There may be a number of different ways of achieving this. An issue which arises here is that the separation of business is clearly more complicated in smaller currencies than it is in the more heavily traded currencies. In response to these issues, I will consult with my fellow central banks and securities regulators on these questions and determine if more guidance can be given around this.

Third, the group made a number of recommendations around the appropriate sharing of information. These recommendations focused on general principles around information sharing. The group's assessment was that outlining a set of principles was more beneficial in ensuring appropriate behavior than more prescriptive statements. Detailed prescriptions run the risk of being incomplete with the potential for market participants to make the assessment that if something was not in the prescriptions then it must be allowable. A set of principles require market participants to think more carefully about their general approach to information sharing.

Codes of Conduct

A similar set of principles are to be embodied in a code of best market practice and shared global principles, which is currently in the final stage of being written jointly by the various foreign exchange committees. The aim is to promote a set of shared global principles common across all the FX committees, notwithstanding the fact that the various jurisdictions employ different codes of conduct. A final version of the statement of shared principles should be published following the next global meeting of foreign exchange committees in Tokyo in March.

The FXBG also recommended that market participants should demonstrate stronger compliance with the various codes of conduct than has been the case in the past, as well as with their own internal codes. Various initiatives are under way on this recommendation at both the institutional level as well as with the market as a whole. Appropriate enforcement of these codes remains a major challenge. In the UK, the FEMR highlights this issue as a significant opportunity for reform.

Much of this comes down to ensuring that the institution, and the FX market more generally, has the appropriate culture. But the same problem arises with codes of conduct as with the principles for information sharing I talked about a minute ago. If the codes of conduct are too prescriptive, they are very difficult to draft, they quickly become unreadable and they run the risk that if something is not explicitly prohibited, it must be okay.

Matt Levine articulates this problem very succinctly: ‘If you don't define “culture,” it's hard to enforce it. If you do, it's going to be gamed.’[7]

Index Providers

During its work, the group discovered that there was considerable variation in the understanding of index providers about the way the FX rates they used in their indices were calculated. Hence the group recommended that index providers consider whether the FX fixes they used in their calculation are fit for purpose.

Asset Managers

Similarly, in the course of our consultation with the market, the group found there was considerable variation among asset managers in their understanding of the FX rates used in the indices they were tracking and, perhaps more importantly, considerable variation in understanding how their FX fixing business was being executed in the market.

While asset managers should be able to assume that the market infrastructure was adequately robust and that their counterparties conducted themselves with some degree of appropriate behaviour, at the same time, greater scrutiny and enforcement from market participants is desirable.

Hence the group recommended that asset managers conduct appropriate due diligence around the FX execution and be able to demonstrate that to their clients if requested. Clearly, such due diligence would be consistent with the asset managers' regular fiduciary responsibility to their clients.

While some asset managers have done this for many years as part of their standard operating procedure, there remains scope for further adoption of this across the asset management industry and, more generally, a greater understanding of the way their FX business is executed in the market, as it becomes an increasingly important component of asset managers' portfolios.

The Way Forward

The 15 recommendations of the group have been adopted to varying degrees, but there is still substantial scope for further progress in some areas. The recommendations to enhance WM's calculation of the London fix are well in train. There are a number of initiatives in the marketplace to provide independent fix execution and netting. Codes of conduct remain a work in progress, particularly in terms of their enforcement, though the various foreign exchange committees are working with a common purpose on this currently and I expect that we will have something concrete to show for this very shortly.

However, there has yet to be significant progress in terms of the pricing and execution of the fix business within institutions. On pricing, while there may be more than one way of charging for the fix business that could be adopted, there needs to be a widespread adoption across the industry. On the separation of the fixing business, the primary objective is separation of the information flow. There are, again, a number of potential solutions to this and I accept there may be some scope to provide more guidance around this recommendation, particularly in the case of less-traded currencies.

To conclude, there is a strong expectation that these recommendations will be implemented to deliver an improvement in the execution of foreign exchange transactions referencing FX benchmarks and the integrity of the benchmarks themselves. The FSB has approved the recommendations contained in the FX Benchmark Group's report. If these recommendations are not implemented, then the likelihood of a regulatory response will increase.


Similar remarks are also contained in an article in the Euromoney Foreign Exchange & Treasury Management Handbook 2015 (forthcoming). [*]

WM publishes reference rates throughout the course of the day. The London 4pm rate is by far the most heavily used. There are a number of other reference rates, with the 2.15pm CET rates published by the European Central Bank quite widely used. [1]

FSB (2014), ‘Public Responses to July 2014 Consultative Document Foreign Exchange Benchmarks’, 20 August. Available at <http://www.financialstabilityboard.org/2014/08/c_140819/>. [2]

FSB (2014), ‘Final Report on Foreign Exchange Benchmarks’, 30 September. Available at <http://www.financialstabilityboard.org/2014/09/r_140930/>. [3]

Details of the FEMR are available on the Bank of England website at <http://www.bankofengland.co.uk/markets/Pages/fmreview.aspx>. Submissions closed at the end of January. [4]

See <http://wmcompany.com/wmr/index.htm>. [5]

IOSCO (2013), ‘Principles for Financial Benchmarks: Final Report’, July. Available at <http://www.iosco.org/library/pubdocs/pdf/IOSCOPD415.pdf>. [6]

Levine M (2015), ‘Levine on Wall Street: Culture and Retaliation’, BloombergView, 2 February. Available at <http://www.bloombergview.com/articles/2015-02-02/levine-on-wall-street-culture-and-retaliation>. [7]