In its letter of 30 August 2013, the Ministry of Finance requests Finanstilsynet (Financial Supervisory Authority of Norway) – in consultation with Norges Bank and with contributions from the financial sector – to prepare a report on reference rate options in Norway. The Ministry refers in its letter to the potential advantage of developing new reference rates with different characteristics from those of NIBOR, such as reference rates based on actual transactions. The Ministry notes in particular that financial sector agents should work towards developing a market for overnight index swaps (OIS) in Norway.

To solicit contributions from the financial sector, Norges Bank sent a letter to Finance Norway containing a number of questions relating to alternative reference rates. In the same letter, Norges Bank also raised issues related to the existing NIBOR rules (these letters are in Norwegian only).

Norges Bank has assessed possible alternative reference rates in the Norwegian market. Our conclusion is that there are currently no realistic alternatives to NIBOR and that NIBOR will have to continue to be the most important reference rate ahead. The robustness of NIBOR must therefore be enhanced. There are clear deficiencies in the current rules requiring NIBOR to be quoted as an fx swap rate. As a result of this requirement, NIBOR is volatile and unpredictable without being more closely aligned with the market or more "correct" than other countries' benchmark rates. The basis for these conclusions is explained in Appendix I.

The financial sector should contribute to the development of markets with interest rates that could function as alternative reference rates or "cross-checks" for NIBOR. The most suitable candidates are the repo market and an OIS market. In a repo market, loans are raised against collateral. An OIS is an unsecured interest rate reflecting the future key rate as expected in the market. The OIS in particular would be suitable as a supplementary reference rate. Even if they were not used directly as reference rates, a repo rate and an OIS rate would enhance transparency in NIBOR quoting.

In principle, the interest rate on short-term government securities could be an alternative reference rate. Turnover in the Norwegian Treasury bill market is, however, low and prone to fluctuation. In periods, interest rates in this market can contain substantial premiums, positive or negative, depending on the prevailing market conditions. These interest rates are thus unsuitable as reference rates. It is worth noting that even countries with highly liquid government securities markets do not use rates from this market as reference rates.

Which properties should a reference rate have?

The properties a reliable reference rate should have depend on the purpose it is meant to serve.[1] In some contexts, some users may want the reference rate to reflect banks' credit risk.[2] In this case, the reference rate is an expression of banks' marginal borrowing costs. This is primarily a feature of unsecured interbank lending. In addition, the reference rate should preferably be traded in a liquid market with a large number of participants.[3] Activity in the unsecured interbank market, however, is very low for maturities of more than a few days.[4] The most well known international reference rates, LIBOR and EURIBOR, are therefore banks' judgement-based estimates of unsecured interbank rates.

Unsecured money market rates can be decomposed into a risk-free rate, close to the expected key rate, and a credit risk premium. Until the onset of the financial crisis in 2008, the risk premium, the spread between the reference rate and the expected key rate, was fairly stable. The reference rates were thus also suitable as a basis for calculating a risk-free interest rate. During the financial crisis, risk premiums in international reference rates rose considerably and volatility increased as a result of higher liquidity and credit premiums.

Since the financial crisis, initiatives have been taken in a number of countries to review reference rates. In an international context, the suggestion is not to replace the traditional IBOR rates. A very high volume of financial contracts have already been tied to these rates for many years ahead. And even though traded rates are preferable in principle, the possibility that these rates could be established for the most important maturities (3 and 6 months) also seems unlikely. International efforts have therefore to a large extent focused on improving the rules for existing reference rates.

International reports also discuss the need for supplementary reference rates. In the period following the financial crisis, substantial differences arose across banks in terms of credit risk. A reference rate – which reflects average risk for banks on the relevant panel – will then be a less accurate measure of the individual bank's short-term borrowing costs. Furthermore, high and volatile premiums reduce the reference rate's suitability as a basis for calculating a risk-free rate. In addition, more of banks' funding against collateral in specific assets takes place on their balance sheets, reducing lenders' exposure to banks' credit risk and reducing the need for an unsecured reference rate. The volume of derivatives trading has also increased. For the parties in a derivatives trade, the reference rate should not necessarily be linked to banks' credit risk. The market value of the derivatives is hedged by margin payments or the exchange of collateral between the parties. An interest rate with low or zero credit risk, such as OIS, is more suitable for discounting such payments, particularly in periods of financial turbulence and stress in the banking sector. A group of experts established by the BIS has conducted discussions with a number of market participants in Europe, the US and Asia concerning reference rates and the need for alternatives. The group reports that market participants are starting to diversify their use of reference rates. In particular, as a result of increased derivatives trading, greater interest is being shown in credit risk free reference rates such as OIS.[5]

The reference rate in Norway: NIBOR

NIBOR is the only reference rate in Norway. Since there is no realistic alternative, swift action must be taken to strengthen confidence in NIBOR and enhance robustness. Confidence is the result when non-panel participants can understand the factors behind a change in NIBOR from one day to the next. Confidence also presupposes that the rules will prevent NIBOR being set by employees who may have a personal interest in the rate moving in a specific direction. The robustness requirement has two dimensions: First, NIBOR's sensitivity to an individual bank's contribution should be as low as possible. This implies that the number of banks on the NIBOR panel should be increased. Second, NIBOR should be robust to changes in other countries' financial markets in periods of turmoil. The current rules and practices for NIBOR quoting do not satisfy the requirements related to confidence and robustness.

In accordance with the rules of NIBOR revised on 30 October 2013, the rates submitted by the individual panel bank must «[...] reflect the interest rates the bank would charge on lending in NOK to a leading bank that is active in the Norwegian money and foreign exchange markets». [6]The rules also state that NIBOR must be quoted as an fx swap rate, determined by a USD rate and the so-called forward premium between NOK and USD in the foreign exchange market.[7] The forward premium is the difference between the forward rate and the spot rate and expresses the price of swapping currencies today and reversing the swap in the same amount in the future, for example in three months (the forward premium reflects the interest rate differential between the currencies).[8] The forward premium is traded in the market. The USD rate is in principle regarded as reflecting the price banks have to pay for unsecured USD loans in the interbank market.[9]

NIBOR is considerably more volatile than benchmark rates in other countries. This is because NIBOR is quoted as an fx swap rate. High volatility in forward premiums results in corresponding volatility in NIBOR.[10] The NIBOR rules argue in favour of calculating NIBOR as an fx swap rate because it is in the fx swap market NOK liquidity is to be found.[11] High volatility in NIBOR via the forward premium is explained by fluctuations in the demand for NOK liquidity.

Finance Norway writes in its letter to Norges Bank of 31 January 2014 that «[...] using forward rates when calculating NIBOR results in a benchmark rate that is more closely aligned with the market. NIBOR will then to a greater extent reflect changes in liquidity demand in the Norwegian market, which must be considered one of NIBOR's strengths compared with some other IBOR rates.»

The defects in the current NIBOR construction are discussed in more detail in Appendix 1. The analysis in the Appendix raises doubts as to whether there is a basis for the claim that the swap construction results in a NIBOR rate that is more closely aligned with the market than other countries' benchmark rates. Based on data with a 15-minute sampling frequency, banks' individual contributions to NIBOR, particularly around the NIBOR fix at 12 noon, are analysed. We find several examples of banks whose NIBOR quoting changes from one day to the next in the opposite direction from the other panel banks' quotes. This can hardly be explained on the basis of general market movements and a liquidity premium on lending in NOK. In some cases, one bank's quote has changed to such an extent, in the opposite direction from the other banks' quotes, that this change in itself has been sufficient to affect the average that is fixed as NIBOR. The analysis shows that divergent NIBOR quotes sometimes reflect wide differences in forward premiums, while in other cases the divergence is the result of banks' USD rates moving in the opposite direction.

Because of the combination of (i) generally high volatility in NIBOR from the forward premium, (ii) that NIBOR quoting can be strongly influenced by differences in individual banks' quoted forward premiums around the time of the NIBOR fix at 12 noon, and (iii) wide variations in the USD rate on which quoting is based, there is a lack of transparency in NIBOR quoting and it may be difficult for non-panel market participants to understand the NIBOR quotation.

The swap construction may at first glance appear to be an objective framework, tying NIBOR to actual interest rates, i.e. the USD rate a NIBOR bank has to pay for unsecured interbank borrowing in USD and the forward premium traded in the market. Banks have, however, ample opportunity to influence both the forward premium and the USD rate quoted around the time of the fix at 12 noon. Thus banks have, in reality, considerable room for manoeuvre when they determine their contributions to NIBOR. In Norges Bank's view, banks' judgement should be made explicit. This will have to be the case if NIBOR is quoted as a pure NOK rate. Banks must then give other reasons for their contribution to NIBOR, and not just refer to the forward premium and the USD rate on which quoting is based. The unsecured interbank market for maturities of more than a week is also virtually absent in other countries, but it is nonetheless not common to use fx swap rates as reference rates. NIBOR banks should, as in other countries, submit their contributions to the benchmark rate based on their best estimate of the lending rate they would require as described in the definition. Banks should base their estimate on developments in all relevant markets. Changing the rules in this way, where banks exercise judgement and provide reasons for the estimate submitted, will, in the view of Norges Bank, enhance confidence in NIBOR as a reference rate.

Another argument used for the preservation of the current NIBOR construction is that liquidity in the forward market is strengthened by quoting NIBOR as an fx swap rate.[12] This argument is difficult to understand. Other countries have forward markets that are at least as liquid and well functioning as the Norwegian market without having an fx swap rate as the reference rate.

The current NIBOR rules include a requirement that a NIBOR panel bank is «[...] an active market maker in the market in which the redistribution of NOK liquidity takes place». The requirement implies that a NIBOR panel bank must be an active market maker in the fx swap market and must quote current forward premiums between NOK and USD. This requirement excludes banks from joining the NIBOR panel. For example, a large Norwegian savings bank may well be able to estimate the interest rate it would charge on lending in NOK to other banks, even if it does not quote forward premiums. In most other countries, reference rates are a subjective estimate of the rate required for interbank lending, without any explicit reference to the forward market. A comparison of NIBOR quoting with STIBOR quoting illustrates most clearly how unnecessary this requirement is. The STIBOR rules, published by Svenska Bankföreningen on 13 January 2013, state that the reference rate STIBOR should be calculated based on several rates, where swap rates can be included in the information basis. STIBOR is nonetheless considerably less volatile than the implied interest rate on SEK swapped from USD.[13] Thus, STIBOR panel banks are able to quote a judgement-based rate for SEK without allowing high forward premium volatility to have a direct effect on the benchmark rate. There is no reason why NIBOR panel banks could not do the same. Five of the six STIBOR panel banks are also members of the panel.[14]

Necessary changes in the NIBOR rules:

1. NIBOR must be quoted as a NOK interest rate. A bank's contribution to NIBOR should be based on its best estimate of the interest rate it would require on NOK lending to another bank as described in the definition. Banks should base their quoting on a broad range of price information, as is the procedure in Sweden and Denmark. For any material changes in the bank's contribution to NIBOR an explanation must be given beyond referring to changes in the forward premium and a USD rate.
2. The requirement that NIBOR banks need to be market makers in the fx forward market must be removed to allow other banks to join the NIBOR panel. Expanding the number of banks on the panel will enhance robustness by reducing NIBOR's sensitivity to a single bank's contributions.
3. The NIBOR rules should emphasise conflicts of interest more than at present. Employees involved in NIBOR submissions cannot hold individual risk mandates to take active positions in derivatives for which NIBOR is the benchmark (or whose own interests are in any other way directly dependent on the level of NIBOR).
4. NIBOR should be defined more precisely. The current rules state that NIBOR should reflect the interest rates banks charge on lending in NOK to "a leading bank that is active in the Norwegian money and foreign exchange markets". The assessment as to which banks are "leading" may vary across NIBOR panel banks. This may in turn result in different estimates of the lending rate for "leading banks". NIBOR will then be more difficult for non-panel participants to interpret (disregarding the NIBOR swap construction itself). It is Norges Bank's view that the definition should be formulated in the same way as for LIBOR. Alternatively, it should be changed to state that NIBOR should reflect the lending rate a NIBOR bank would charge on unsecured lending in NOK to other NIBOR banks. In both cases, the credit risk element of NIBOR would be far more clearly reflected than it is today.
5. The composition of the NIBOR steering committee should be reassessed. One of the steering committee's tasks is to propose changes in the rules and assess applications from new NIBOR banks. The committee currently includes several members who are themselves submitters or who are closely involved in NIBOR submissions. It is Norges Bank's view that the steering committee should comprise persons who are somewhat more detached from the daily process of contributing to NIBOR.

Alternative reference rates

There are currently no alternative or supplementary reference rates in Norway. As in other countries, the unsecured interbank market in Norway is concentrated on the shortest maturities (overnight and tomorrow-next), driven by banks' need for short-term liquidity management.[15] Interest rates on government securities (bills and bonds) are strongly influenced by conditions specific to the government securities market, such as time-varying liquidity and safe haven premiums. The repo market is somewhat immature and there is no OIS market. Norges Bank sees the need for developing the repo market and establishing an OIS market in Norway. The initiative for developing these markets must primarily come from market participants. It would not be appropriate for Norges Bank to be a trader or market maker in an OIS market. As the OIS rate is closely linked to the expected key rate, any such activity by Norges Bank would be perceived as signals concerning future monetary policy.

The repo market is currently small in Norway, with considerably lower daily turnover than in the unsecured overnight market and in the fx swap market. Activity in the repo market is highest for maturities of around one week.[16] A better developed repo market could relieve the pressure on the fx swap market. It could make the Norwegian money market more robust to turbulence in international financial markets. A better developed repo market could also stimulate activity in other markets. Markets for securities used as collateral in repo transactions, including government securities and covered bonds, could become more liquid.

Finance Norway assesses the need for repos in Norway in its letter to Norges Bank of 31 January 2014. Finance Norway expresses a clear interest in improving the repo market in Norway, noting that a well functioning repo market could stimulate the bond market, increasing liquidity and transparency. Finance Norway envisages active trading in maturities of up to one month. In addition, repo rates are suggested as possible reference rates in the bond market, as repo rates are closer than unsecured interbank rates to risk-free interest rates.

Finance Norway notes that it is important that Norges Bank begins to use repos in liquidity management in order for the repo market to develop. Norges Bank is familiar with market participants' view on this issue, but must assess whether repos are suitable for use in liquidity management.

An OIS is an interest rate swap where two parties exchange a floating rate for a fixed rate. The floating rate is referred to as the short or floating leg, while the long rate is referred to as the swap rate, or the OIS rate in the case of an OIS. In an OIS, the floating leg is a very short rate, usually an overnight rate.

The OIS is as close as possible to a risk-free yield curve. With a developed OIS market, market participants can hedge against changes in the general interest rate level, which is determined over time by the key rate. An OIS would also be suitable for market participants who need a risk-free rate to discount future cash flows. Furthermore, OIS will enable interest rate expectations and risk premiums in the money market to be calculated more directly than at present. NIBOR will be more transparent if risk premiums can be observed.

In its letter to Norges Bank of 31 January 2014, Finance Norway notes that the financial sector considers in principle that the establishment of an OIS market in Norway is of interest. However, Finance Norway also notes that general risk willingness in the Norwegian interest rate market has shown a falling trend over the past year and that it is therefore assumed that interest in an OIS market in Norway is limited. These objections are based on the Norwegian interest rate market as it is today. There are reasons to believe that market participants' willingness to trade in Norwegian interest rates depends on the instruments offered in this market. The introduction of new products can in itself boost demand and activity from banks' customers. At the same time, new products that enhance the transparency of NIBOR may increase participants' willingness to enter into derivative trades that are referenced to NIBOR.

The overnight rate, tomorrow-next and central banks key rates are used internationally as the floating leg in OIS contracts. In Norway, the tomorrow-next rate is not quoted; only a forward premium for this maturity is quoted. An OIS must therefore be based either on Norges Bank's key rate or on the overnight rate NOWA. Finance Norway refers to the high volatility in NOWA at quarter-ends, which may make it unsuitable as a reference rate in an OIS. In addition, Finance Norway holds the view that an OIS with the key rate as the floating leg does not provide a hedge against exposure to changes in liquidity in the money market.

OIS exists in almost all countries with reasonably developed financial markets. In most countries, the overnight rate is the floating leg, even though it can occasionally be highly volatile at month- and quarter-ends.[17] Other countries use the tomorrow-next rate, while a few countries use the key rate.

The need to enhance the transparency of NIBOR implies that the financial sector should make a thorough assessment of the prerequisites for an OIS market in Norway. Norway is now virtually the only advanced economy without an OIS market.

Nasdaq OMX has been in contact with Norges Bank concerning the establishment of an interest rate future on Norges Bank's key rate, corresponding to the RIBA future in Sweden. Although not quite the same as an OIS, the instrument has clearly similar characteristics and will enable the expected key rate (risk-free rate) to be observed. Nasdaq OMX is currently engaged in talks with various market participants concerning the establishment of this instrument. The transaction will, as in Sweden, be over-the-counter, i.e. that the parties wishing to engage in a trade will contact each other directly, while the trades will be reported to and cleared via Nasdaq OMX. Norges Bank calls on market participants in Norway to contribute to the establishment of such a market.

Summary and conclusion

NIBOR will continue to be the primary reference rate and the rules must be changed as soon as possible to enhance the robustness of and increase confidence in NIBOR. The most important change is to establish that NIBOR must be quoted as a NOK rate and not as an fx swap rate. NIBOR will then be less volatile and easier to understand, and more banks will be able to participate in the NIBOR panel. Norges Bank is of the view that the rules for STIBOR and CIBOR provide a sound basis for reassessing the rules for NIBOR. A point by point summary of the changes that are needed is provided above.

In parallel, efforts should be made to further develop the repo market and to establish an OIS market in Norway. Repo and OIS rates would be valuable cross-checks for NIBOR. Norway is one of the very few advanced economies without an OIS market. OIS rates with different maturities are a good proxy of a risk-free yield curve. With a developed OIS market, market participants can hedge against changes in the general interest rate level, which is determined over time by the key rate. An OIS could also be suitable for market participants who need a risk-free rate to discount future cash flows. Furthermore, OIS rates provide a basis for observing risk premiums in the money market. NIBOR will be more transparent if risk premiums can be observed. In the future, it is possible that repo and OIS rates can take over NIBOR's role as reference rates in some types of loan agreements and derivative contracts. Norges Bank assumes that the financial sector recognises that there is a need for new interest rate instruments in the Norwegian market and that it is in their interest to contribute to the development of markets for these rates.

 

Yours sincerely: Kristin Gulbrandsen and Olav Bø

 

• Appendix 1: Weaknesses in NIBOR (pdf, 895 kB)
• Appendix 2: The letter referred to in footnote 10 and 13, «Calculation of the NIBOR rate» (pdf, 1,2 Mb)

Note: In the Norwegian version of the letter, three more appendixes are enclosed, none of which will be translated into English: [Reference rates in Sweden, Denmark and other countries], Letter from Norges Bank to Finance Norway (3 December 2013) «Tilbudet av referanserenter i Norge» [Reference rate options in Norway] and Letter from Finance Norway to Norges Bank (31 January 2014) «Tilbudet av referanserenter i Norge» [Reference rate options in Norway].

 

Footnotes

[1] The discussion below draws to some extent on the BIS paper (2013): "Towards better reference rate practices: a central bank perspective". http://www.bis.org/publ/othp19.pdf.

[2] Suppose for example that an unsecured bank bond is priced at the reference rate plus a premium. When the reference rate reflects banks' credit risk, the lender will be automatically compensated for this risk.

[3] This ensures that new information and participants' preferences are quickly reflected on both the supply and the demand side, ensuring a (microeconomic) efficient price.

[4] Activity in the shortest maturities is driven by banks' need for short-term liquidity management, not by their funding needs. The majority of transactions are overnight and tomorrow/next trades.

[5] See BIS (2013), cf. reference in footnote 1.

[6] There are six banks on the NIBOR panel (DNB, Nordea, Danske Bank, Handelsbanken, SEB and Swedbank). NIBOR is calculated as the average of the submitted rates after the highest and lowest rate has been discarded.

[7] This implies that NIBOR can be written as follows: NIBOR = USD rate + forward premium

[8] The forward premium reflects the difference between the expected key rate in the two relevant countries. In countries with an OIS market, the forward premium with a three-month horizon will reflect the three-month OIS spread. The forward premium can deviate from the OIS spread if there are different liquidity premiums in the two currencies. For more details, see Norges Bank Staff Memo 21/2012.

[9] It is important to distinguish between an interest rate swap and an fx swap. In an interest rate swap, a short rate is swapped for a long rate in the same currency. An OIS is a particular kind of interest rate swap where the short rate is an overnight rate. In an fx swap, the parties swap currencies at the current spot rate in the foreign exchange market and agree to reverse the swap on an agreed date in the future at a rate agreed on at the initiation of the swap. The future rate is the forward rate. The difference between the spot rate and the forward rate, known as the forward premium, expresses the interest rate differential between the two currencies during the life of the swap.

[10] Norges Bank has on several occasions expressed the opinion that the swap construction for NIBOR is not appropriate: in a letter to Finanstilsynet of 16 August 2012 («Meeting follow-up»), in a letter to Finanstilsynet of 20 March 2013 («Calculation of the NIBOR rate») and in a letter to Finance Norway of 3 December 2013 («Reference rate options in Norway»). The effect on NIBOR of general volatility in the forward premium is discussed in particular in the letter to Finanstilsynet of 20 March 2013. This letter is enclosed below (the other two letters are in Norwegian only).

[11] See appendix «Additional guidelines for panel banks' NIBOR submissions», in «Rules for the calculation and publication of Norwegian money market rates – NIBOR».

[12] See letter from Finance Norway to Norges Bank 31 January 2014 (top of first paragraph on page 2, in Norwegian only).

[13] For more details, see the letter from Norges Bank to Finanstilsynet of 20 March 2013 (the letter is enclosed below).

[14] With the exception of DNB, all the NIBOR panel banks are also members of the STIBOR panel.

[15] The overnight rate in the interbank market is the interest rate on a loan provided by one bank to another with a maturity from the current trading day to the next. Tomorrow-next is the interest rate on a loan from the following trading day until the subsequent day.

[16] The unsecured market is the dominant market for overnight, while the fx swap market is the dominant market for longer maturities. For further details, see (in Norwegian only) http://www.norges-bank.no/pages/98279/aktuell_kommentar_2013_06.pdf

[17] The overnight rate EONIA, for example, is used as the floating leg in the EUR OIS market. EONIA is calculated in a similar way to NOWA in Norway. EONIA is often highly volatile at month- and quarter-ends.