«Thomson Reuters response to the European Commission Consultation Document on the Regulation of Indices Thomson Reuters is the world‟s leading ...»
Interest Representative Register ID: 51334142356-37
28th November 2012
Thomson Reuters response to the European Commission Consultation Document
on the Regulation of Indices
Thomson Reuters is the world‟s leading source of intelligent information for businesses
We combine industry expertise with innovative technology to deliver critical information
to leading decision makers in the financial, legal, tax and accounting, healthcare,
science and media markets, powered by the world‟s most trusted news organisation.
Thomson Reuters welcomes the opportunity to respond to this consultation.
Thomson Reuters and its predecessor companies that now form part of Thomson Reuters have been involved in the calculation and distribution of benchmarks, such as LIBOR, Euribor, plus many other similar third-party sponsored rates, since their inception. For each of these rates, the sponsoring association provides the specification for the rate and Thomson Reuters acts in various roles which include, as an agent to collect, calculate and disseminate information to the market.
In addition Thomson Reuters Indices business is responsible for the calculation of several equity, fixed income and commodity indices which we consider to be benchmarks.
Thomson Reuters is well-known as a neutral and independent source of information; our interests lie in ensuring that markets are transparent and trusted.
Should you have any queries on our response, or wish to discuss the wider issues around the provision of market data, we would be very happy to discuss further at your convenience.
Below we address the questions from the Consultation Document on the Regulation of Indices which we feel qualified to answer.
Peter Moss Managing Director, Trading, Financial & Risk Chapter 1.
Indices and Benchmarks: What they are, who produces them and for which purposes (1) Which benchmarks does your organisation produce or contribute data to?
For the purposes of this consultation, our responses are based on the following definition of “Benchmark”: an instrument whose numeric value (‘value’ being in relation to some specific starting amount on a specific date) is derived on a regular basis and where that instrument’s value is used as a standard of measure/point of reference against which the performance of another instrument is being measured in a contract between two or more parties.
Thomson Reuters Indices (a) produces more than 10,000 proprietary indices in different asset classes, including, for example, equities, commodities and fixed income, which are not currently being used as benchmarks, but could potentially be used as such, and (b) is responsible for the calculation of several equity, fixed income, commodity and multiasset indexes which we believe could to be considered benchmarks according to our
definition. These indices are publicly available on our website:
In addition our role in the calculation of LIBOR is set out on the following website:
www.thomsonreuters.com/libor Some market participants use commodities benchmarks derived by Thomson Reuters (e.g. for Bullion and some grades of Palm Oil, Crude Oil and Vegetable Oil). In addition, Thomson Reuters derives a set of Asian intraday fixings for swaps curves for Crude Oil and Refined Products, which some market participants choose to use to mark-to-market.
(3) Have you recently launched a new benchmark or discontinued existing ones?
We will shortly be discontinuing several benchmarks. A recommendation from the Wheatley Review was to discontinue LIBOR calculations in five of the 10 currencies calculated. The currencies to be discontinued are Australian Dollars, Canadian Dollars, Danish Kroner, New Zealand Dollars and Swedish Kronor. In addition for remaining currencies, publication of LIBOR for 4 months, 5 months, 7 months, 8 months, 10 months and 11 months tenors will be discontinued. The decision to discontinue these was taken by HM Treasury, on the recommendation of the Wheatley Review. We will discontinue these calculations as and when we are instructed by the owner of the data, which is currently the British Bankers‟ Association until a successor is appointed.
During July 2012, Thomson Reuters discontinued the daily publication of a series of snapshot prices for physical BFOE (Brent/ Forties/ Oseberg/ Ekofisk) Crude Oil.
We also offer benchmark calculation as a service to the markets. There are a number of models for this. We may act as a distribution agent for organisations that calculate benchmarks in-house or we may calculate benchmarks on behalf of the data owners.
We also produce certain proprietary benchmarks. We were recently nominated by JSCC (Japan Securities Clearing Corporation) as an official calculation vendor to calculate and publish new swap settlement rates.
We have not recently launched or withdrawn any Thomson Reuters Indices benchmarks in or from the marketplace. Thomson Reuters Indices does continue to offer index calculation solutions to clients that design an index and require another party to calculate that index and many of these Thomson Reuters client indices are then licensed by the client to 3rd parties to act as benchmarks.
(4) How many contracts are referenced to benchmarks in your sector? Which persons or entities use these contracts? And for which purposes?
It is not possible to answer this question with certainty. A number of the benchmarks we calculate and/or distribute are used as the underlying in derivative contracts. Some of these are traded on exchange, for example LIBOR, Euribor and Palm Oil futures, and the exchanges that offer trading in these (principally CME Group and NYSE Liffe) will have information on the volumes and outstanding notional value of these contracts.
However, these types of derivatives are more commonly traded over-the-counter. In either case it is not necessary to inform us as the calculator / distributor of trading activity in these products. There are also multiple exchange and OTC products listed against the Thomson Reuters/Jefferies CRB index. These products allow buyside exchange traded funds (ETFs) and sellside structuring desks to provide commodities exposure to both retail and institutional clients.
Because of the reasons above, we cannot provide any estimates of the number of contracts outstanding against these benchmarks.
In Table 2A of his initial discussion paper, Martin Wheatley provides estimates on the usage of LIBOR, and concludes around USD 300 trillion of contracts are linked to LIBOR. In their enforcement order on Barclays in June 2012, the CFTC stated that, “U.S.
Dollar LIBOR is the basis for the settlement of the three-month Eurodollar futures contract traded on the Chicago Mercantile Exchange (CME), which had a traded volume in 2011 with a notional value exceeding $564 trillion. In addition, according to the BBA, swaps with a notional value of approximately $350 trillion and loans amounting to $10 trillion are indexed to LIBOR. Euribor is also used internationally in derivatives contracts.
In 2011, over-the-counter interest rate derivatives referenced to Euro rates had a notional value in excess of $220 trillion, according to the Bank for International Settlements.” LIBOR and Euribor are likely to be the benchmarks that underlie the largest number of interest rate contracts, but contracts are traded on a wide range of benchmarks. The CFTC also noted that “LIBOR and Euribor are relied upon by countless large and small businesses and individuals.” Contracts linked to interest rate benchmarks are usually used by firms to hedge their interest rate exposure, although it is also possible to use these contracts to speculate on future interest rate movement.
In commodities, a few benchmarks are believed to be used as the basis for settlement of OTC contracts and, in Asia oil swaps, to mark to market.
(5) To what extent are these benchmarks used to price financial instruments?
Please provide a list of benchmarks which are used for pricing financial instruments and if possible estimates of the notional value of financial instruments referenced to them.
One of the principal uses of benchmarks is to price financial instruments. It is not possible to provide a comprehensive list of these, for the same reasons as stated in the answer to question 4. However, we have provided a sample list of investable products that are based on Thomson Reuters Indices in Appendix A of this document..
Well known benchmarks such as NASDAQ or FTSE equity indices are used as the underlying for instruments such as ETFs, futures and options. The owners of these indices or the venues on which the financial instruments linked to them trade may be able to provide information on the outstanding notional value.
However, financial instruments can be highly bespoke products including many inputs that are created for individual clients by banks. The existence of these may well not be apparent to anyone but the bank and its client, so it is not possible to even begin to estimate the notional value of such products or even provide a list of the benchmarks used for these purposes.
(6) How are benchmarks in your sector set? Are they based on real transactions, offered rates or quotes, tradable prices, panel submissions, samples? Please provide a description of the benchmark setting methodology.
All these methodologies are used across the benchmarks we provide, the nature of which depends on the characteristic and set up of each.
With specific regard to LIBOR, Thomson Reuters calculates and distributes LIBOR on behalf of the BBA. The BBA names the panel banks that submit the rates to Thomson Reuters each day and also sets the methodology Thomson Reuters follows. Every morning Thomson Reuters receives the contributions electronically from the panel banks. We perform a series of pre-defined validity checks following the criteria set by the BBA. Some of the checks are aimed at spotting any obvious “fat finger syndrome” errors, numbers that were clearly keyed in incorrectly, ie decimals in the wrong place, transposed numbers.
However some other validity checks require us to question numbers from individual contributors if they differ greatly from the number submitted the previous day. We will call the contributor and ask them to confirm their number. Clearly we will never tell them what other banks have contributed but just give them the opportunity to inform us if the electronic number we received is the rate intended.
Thomson Reuters operational role, up until now, is to ensure that we publish/use the numbers as intended by the banks, following the guidelines set up by the BBA.
Once the checks are complete, we follow the calculation formula; essentially discount the top and bottom quartiles of the figures and average the two quartiles in the middle.
We then publish LIBOR prior to 12pm each day. All individual bank contributions are published – even the figures that were discounted in the top and bottom quartiles – making the process and contributions all very transparent to the market so it can place its own value/judgement on LIBOR.
(7) What factors do you consider to be the most important in choosing a reliable benchmark? Could you provide examples of benchmarks which incorporate these factors?
There are a number of factors that are important in choosing a benchmark. The first of these is transparency which we address as part of our response to question 12. There are a further four factors that are listed below. However, we do not consider that we are in a position to offer judgement on the various calculation methodologies employed by the owners of benchmarks Reliability The data must be available in a scheduled and timely fashion, whether this is in real-time for benchmarks such as Foreign Exchange rates, or at the scheduled time of day/month in the case of benchmarks such as interest rates.
The data must be calculated accurately every time. As many financial products have a lifespan of decades, users should also be able to rely on the benchmark being available for the duration of their product.
Predictability Users must be able to anticipate how the benchmark will take account of market events.
This might be in the form of a disruptive event that affects the ability to calculate the benchmark. Typical responses to this are to use the previous day‟s rates, declare a “nofix”, or move to a back-up methodology. Other examples of market events are corporate actions in equity benchmarks. Users tracking these indices must be able to predict how the operator of the benchmark will address mergers between constituents, stock splits and other events. Thomson Reuters has a fair and transparent solution to equity Corporate Actions and believes that a system should be as predictable as possible, with human discretion used as infrequently as possible.
Appropriateness Benchmarks should be an appropriate metric for the product they underlie. For example, it may not be appropriate to base a product designed to reflect the performance of the US equity market on the Dow Jones Industrial Average, as this index does not have sufficient constituents to reflect the performance of the US market as a whole.
Benchmarks should also be appropriate to the intended user. For example, a complex benchmark with many inputs, whose behaviour is difficult to model may not be appropriate for products for use in the retail sector.
Conflict of interest Benchmarks should be calculated by independent bodies that have no interest in the levels at which the benchmark sets, and do not offer for investment products that reference the benchmarks set by that body.
In certain circumstances, contributors to benchmarks may use those benchmarks in the course of their business, or offer products for trading/investment based on the benchmarks to which they contribute. In these cases there must be rigorous controls and governance in place to ensure that there is separation between those parts of the contributing business that submit data to benchmarks, and the parts of the business that create, buy or sell products based upon them.
(8) What kinds of data are used for the construction of the main indices used in your sector? Which benchmarks use actual data and which use a mixture of actual and estimated data?