The Bank of England, the FCA and the Working Group on Sterling Risk-Free Reference Rates (the Working Group) have asked that, by the end of Q3 2020, all new sterling loans should reference SONIA rather than LIBOR. A related expectation is that these new loans will not use forward-looking term SONIA1. Instead, 90%2 (in effect) of these loans are to compound average in arrear all daily SONIA quotations related to each interest period with a five-day lag between each interest period and its "observation period".
We recap the function of lag periods and their length in more detail below. However, among the challenges the above request and expectation present are that:
In this situation, it is tempting to explore how to reduce the complexity of compounded SONIA loans and resolve related issues around adequate notice of the amount of interest payments plus mid-interest period prepayments, loan trades or refinancings. In theory, if appropriate associated SONIA hedging is available, these goals are achievable by making each lag period equal to the entire interest period to which it relates. Accordingly, this note explores whether whole interest period lags are a serious option for simplifying compounded in arrear SONIA loans where the lenders have appropriate associated hedging.
The Working Group has noted that certain Islamic, trade, receivables and working capital financings will not be able to work with compounded in arrear SONIA with a five-day lag. Consequently, the Working Group suggests that parties to such financings may wish to work with, for example, fixed rates or the Bank of England's bank rate instead. The entire interest period lag proposal that this note explores is not confined to those types of financings, but might be used in some mid-market or other loans where there is a desire to reduce complexity.
SONIA is an overnight interest rate that is published daily on each London banking day. It does not price in bank credit or term risk. As such, it is not a term rate that is appropriate in its raw state to produce an interest rate for an entire interest period. The bilateral SONIA loans that have been done as pilot projects, and the LMA Exposure Drafts3 that broadly illustrate how such loans could be done on a syndicated basis when market conventions are more developed, both produce their SONIA rates for interest periods by compound averaging all daily SONIA quotations relating to each interest period in arrear.
When daily SONIA quotations are compounded in arrear, if interest is to be paid on the last day of each interest period, it is necessary to find a way to compound and average all relevant daily SONIA quotations before the end of each interest period. This is done by taking the interest rate quotations from a lagging "observation period".
Under this method, all daily SONIA quotations relating to an interest period are taken from a corresponding "observation period" which lags (i.e. begins and ends X banking days before the beginning and end of) the interest period to which it relates. This lag enables all interest rate calculations and notifications to be done:
If a lender were to adopt a whole interest period lag then, using a three-month interest period running from October to December 2020 as an example;
The lag method originates in the sterling floating rate notes market. In that market, and for SONIA loans, the standard lag period is five banking days. In the loan markets, shorter lag periods may be adopted for shorter interest periods. Longer lag periods might sometimes be necessary and possible, for example, for loans into those few developing markets whose exchange control regulations require significant advance notice of the amount of each interest payment. However, the Working Group recommend that parties not resort to lag periods of more than 10 days without taking specialist advice.
The FCA, the Bank of England and the Working Group have detailed and prescriptive expectations as to how and when the loan markets should move from LIBOR to SONIA. These expectations do not currently clearly appear to include using whole interest period lags. For example: while an August 2019 Working Group paper appeared to consider whole interest period lags a viable option for compounded in arrear SONIA loans, a January 2020 Sterling Working Group paper:
However, while the Working Group is clearly prudent to be cautious on this issue, this latter paper:
So it is not out of the question that the UK regulators will have further thoughts if market participants see longer lags as a useful option.
If and when appropriate SONIA hedging is available, we would suggest not. We would also suggest that residual concerns some lenders might have in this regard could be priced into the margin on the loan or alleviated by capping the length of interest periods. We discuss this and related issues in more detail just below.
For the UK regulators, the main benefits of compound averaging SONIA in arrear appear to flow from the robustness, transparency and freedom from manipulation that comes with the greatly increased number of real underlying transactions upon which the resulting SONIA interest rates will be based. For example, a January 2020 Sterling Working Group paper points out that: "… a three-month compounded SONIA contract is built on overnight market transactions worth on average ~£42 billion per day. Over 60 trading days (the number of trading days in 3 months), this means that ~£2.65 trillion of transactions underlie three-month compounded SONIA". Importantly, these benefits would not be lost by using whole interest period lags.
Where the lag between the observation period and the interest period is just five days, another benefit is that the daily SONIA quotations are very recent, mainly coming from days that fall within the current interest period to which the observation period relates.
If a whole interest period lag is used instead, the daily interest rate quotations will typically5 come from the previous interest period. Should market participants be concerned that such rates will be "stale"? Related to this, should lenders be wary given that when interest is fixed on this basis at the start of an interest period, that rate will not have been calculated using large volumes of SONIA index swaps or future contracts, as would be the case if the parties were using a forward-looking term SONIA rate?
The treasury function of a lender using a long lag period may wish to hedge the risk of significant differences in interest rates between the observation period and the interest period itself. Subject to such hedging being available:
Last month, the FCA and the Bank of England announced that they were encouraging the derivatives markets to change the market convention for sterling interest rate swaps from LIBOR to SONIA in Q1 2020. In their words, this change "is intended to move the greater part of new sterling swaps trading to SONIA and reduce the risks from creating new LIBOR exposures and is also reflected in the timeline set out by the Working Group. Following FCA discussions with market makers, the authorities have identified 2 March 2020 as an appropriate date for this change to happen".
Once this change has taken place, it may be easier in time for lenders to obtain hedging for the differences in SONIA rates prevailing during interest periods and their observation periods. If so, lenders who were able regularly to enter into portfolio basis swaps on SONIA rates, for example, from month to month or quarter to quarter, might be willing to lend under simplified compounded SONIA loans with entire interest period lags of one or three months.