The transition away from LIBOR—which underpins a vast amount of contracts in the financial system as a benchmark—faces several challenges. All of the identified successor rates, known as risk-free rates, are overnight interest rates, but LIBOR is a forward-looking term reference rate. Term rates have a tenor beyond overnight, e.g. three months. A new working paper published by the Bank for International Settlements explains how to solve this dilemma without repeating previous mistakes. The authors have produced a summary of their paper for SOFR Academy.
For benchmark rate reform to be ultimately successful, the crucial step is the wide adoption of Risk Free Rates (RFRs) in cash products, especially in the loan market. For this purpose, term reference rates are needed. But the devil here is in the details. In London Interbank Offered Rate (LIBOR) based cash products, interest rate payments are known at the beginning of an interest rate period (see also Figure 1). Most variable rate cash products are based on quarterly payments, where three-month LIBOR allows to pre-determine the interest rate payment of the counterparties at the beginning of an interest period. These obligations are then paid at the end of the interest rate period. LIBOR has the desirable feature of being pre-determined, which in turn has had an influence on how some of the market’s ‘plumbing’ (e.g. how IT systems for cash flow management) has been set up.
The Figure shows the typical design of a variable-rate cash product. By using a term rate that is pre-determined such as LIBOR, the next interest payment is known at the beginning of each interest period. By contrast, all RFRs are overnight rates. There are two ways of constructing term reference rates based on realisations of RFRs. One method, known as compounding ‘in arrears’, relies on the realisation of overnight rates over the same period as the interest rate period, and hence is only known at the end of the period. The second method, known as compounding ‘in advance’, relies on compounded realisations of RFRs over the same horizon as the interest rate period, but lagged by one interest rate period so as to arrive at a pre-determined reference rate.
The misleading preference for a forward-looking term rate
However, pre-determinedness is not the only feature of LIBOR. LIBOR also reflects interest rate expectations for specific longer-term tenors and is therefore a so-called forward-looking term rate. This feature could be offered by RFRs in case the RFR-based derivatives market would be used to construct a forward-looking RFR term rate. Such an approach, which is discussed among market participants, faces several challenges.
RFR-based derivatives market are currently not yet very liquid. Although this market should become more mature, it is unlikely that the bulk of transactions will match the maturity of the forward-looking RFR term rates desired to be constructed. Such circumstances would thus require a rather complex methodology for term rate construction.
A forward-looking rate constructed from derivatives also tends to be more volatile on a day-to-day basis compared to a compounded rate, as expectations can shift quickly. What is more, having a liquid underlying market is a core requirement for benchmarks, as noted in the FSB (2014) and IOSCO (2013) publications.
“For benchmark rate reform to be ultimately successful, the crucial step is the wide adoption of RFRs in cash products, especially in the loan market”
While there has been a strong push in recent years to develop derivatives markets based on RFRs such as SOFR in the US and these markets have indeed grown over the past years, liquidity in RFR-linked derivatives is still not sufficient to support robust benchmarks. Another downside is that liquidity in derivatives markets can quickly evaporate during crises. The main problem of LIBOR is the lack of the underlying activity supporting the benchmark rate. To avoid repeating the same mistake, it is preferable that products reference robust and reliable benchmarks, i.e. the RFRs themselves and not the RFR-based derivative markets.
The solution can be simple, efficient and robust
RFRs can be used to construct term rates by compounding the (overnight) RFRs over the relevant period. By doing so, this term rate is known at the end of an interest period, which is typically referred to as compounded RFR ‘in arrears’. One could argue that in the ideal post-LIBOR world, all products should use a compounded RFR ‘in arrears’ as a term rate. In fact, some derivatives markets have been able to function well based on this approach for quite some time. Using compounded O/N interest rates ‘in arrears’ is standard to settle the payment obligations in Overnight Indexed Swaps (OIS).
Ideally, a compounded RFR ‘in arrears’ would be used for both derivatives and for variable rate cash products. This in turn would allow to perfectly hedge cash products and to price fixed rate products, e.g. based on OIS. Therefore, it would be ideal to also use a compounded RFR in cash products. Indeed, over the past two years several floating rate notes (FRNs) based on SOFR and SONIA have been issued. However, adoption of compounded RFRs ‘in arrears’ in the loan market (with the exception of SARON in Switzerland) has been sluggish so far. Various factors—some of them related to how market plumbing is set up—make a transition to an ‘in arrears’ approach difficult for certain groups of cash market participants.
For this reason, many market participants have adopted a “wait and see” approach until the existence of a forward-looking term rate. A cash market waiting for a forward-looking term rate leads to a chicken-and-egg problem, in the sense that it is hard for the derivatives market to become liquid in absence of a deep cash market (and vice versa). Yet, for many cash market applications, it is not so much the forward-looking element that is necessary, but rather the fact of being able to know the rate ahead of the interest rate period, i.e. to have a pre-determined rate.
The note proposes to solve this chicken-and-egg problem by using past RFRs known at the beginning of an interest rate period in order to define a pre-determined term rate. One can think of the approach—sometimes also referred to as a compounded RFR ‘in advance’—as a lagged version of the ‘in arrears’ approach. Just like LIBOR, the ‘in advance’ approach offers the benefit of pre- determinedness that certain participants in cash markets require. However, an important implication of their findings is that for a range of applications it is rather the pre-determinedness that matters and not so much the feature that the term rate is forward-looking. An approach based on past RFRs alone (even if lagged) can work reasonably well in many cash market applications, and the note provides some practical guidance that can help market participants with the use of such RFR term reference rates. It could also serve as a useful option in currency areas with less developed derivatives markets, in particular emerging market economies.
The lagged behavior of compounded RFR ‘in advance’ is manageable
The main counter-argument against the ‘in advance’ methodology is that the rate, while being pre-determined, will be sluggish to respond when interest rate expectations change. By construction, a backward-looking rate will not be as responsive as say a forward-looking term rate based on RFR-derivatives would be. The lagged behavior of a compounded RFR ‘in advance’ may at times create a mismatch to a compounded RFR ‘in arrears’. This ‘basis’ is simply the difference between two consecutive compounded rate and can be especially pronounced when the central bank adjusts its policy rates during an easing or tightening cycle.
The note analyses two main ways how market participants can manage the basis that stems from using an “in advance” term rate in cash products. As a first step, an ‘advance’ basis is derived and it is illustrated under what conditions it arises. It is then showed that the basis by using an ‘in advance’ rate can be priced. The estimated basis at the start of the financial contract in turn can be used as an adjustment factor to match the present value of an “ideal” contract using an RFR compounded ‘in arrears’. One can also think of this adjustment factor as a convenience premium to have the reference rate pre-determined. Furthermore, it is showed that this adjustment factor can be hedged by either using existing derivatives, such as an OIS, or through contractual features.
The second option to reduce the basis is to rely on shorter observation periods prior to the interest period when calculating the pre-determined term rate. In this way, the term rate becomes more responsive, as not the entire past period is used for the calculation. Their empirical analysis suggests that using a shortened observation period is indeed a sensible way to reduce the basis from using an “in advance” term rate. For instance, using the RFRs compounded over the week prior to the interest period to determine the interest payments for the next three months, leads to a basis which is low on average and also not too volatile. At the same time, the approach still allows for payments to be on a quarterly basis.
To investigate the various approaches, the authors rely on empirical data for the effective federal funds rate (EFFR), which is also an O/N rate like the identified RFRs. While EFFR is not the chosen RFR in the US, the authors use it in their examples due to the longer history than the secured overnight funding rate (SOFR). This is especially important for comparing backward-looking RFR term rates to forward-looking term rates based on derivatives (there is only a short history of OIS linked to SOFR, but there is a long history and deep OIS market linked to EFFR).
These key findings have several key practical implications. The approaches analyzed here could help foster the usage of a compounded RFR ‘in advance’ in case a pre-determined rate is needed. Hence, they may be especially relevant for smaller market participants in cash markets, that otherwise could occur high switching costs in their back-office operations towards an ‘in arrears’ reference rate. As the ‘in advance’ term rate already exists and is underpinned by the robust RFRs, usage of the rate in cash contracts may help to accelerate the transition away from IBOR-style benchmarks.
The approach in this note ensures a robust anchor for term rates, as the construction is solely based on the RFRs themselves that are underpinned by strong market activity and trading volume. Determining the ideal methodology for pre-determined interest payments, will play an important role in a post-LIBOR world. Experience with the current reform and similar past episodes suggests that once a tipping point has been reached (McCauley, 2001) and a benchmark is widely used, it is challenging to transition away from it. It is therefore important to identify the ideal methodology for benchmark rate construction from the start, i.e. before it is actually used.
Basil Guggenheim is a senior economist at the Swiss National Bank. Andreas Schrimpf is Head of Financial Markets in the Monetary and Economic Department of the Bank for International Settlements. A link to the BIS working paper can be found here. The views expressed in this paper are those of the authors and do not necessarily represent those of the SNB or BIS.