The Bank of England (BoE) has recently published a consultation paper on reforming the Sterling Overnight Index Average (SONIA) benchmark. It had conducted a similar exercise, albeit exploratory, as a public consultation in Q3-Q4 of 2015 with the objective of monitoring the sterling overnight money market and reforming the SONIA benchmark. The outcome of the previous iteration of the consultation resulted in the BoE issuing a four-step process of reforming the £46 trillion benchmark.
In a nut-shell, the four-step process included (i) Taking over of SONIA administration by the BoE (previously handled by WMBA), (ii) expanding the coverage of SONIA input data, (iii) reviewing of SONIA methodology itself and, (iv) reviewing the need for change in the publication of benchmark in the light of proposed changes. While the first step of the process came to fruition in April-2016, the current consultation is focused on fulfilling the remaining steps.
Why change now? The reform process is primarily triggered by the need for strengthening the benchmark by broadening the market coverage so that it is robust and representative of the underlying market. Also, the BoE’s quest for finding ‘near risk-free reference rates’ as alternatives to existing critical benchmarks coupled with the benchmark regulations such as the EU Benchmarks Regulation in the offing, it only becomes imperative to take up this exercise. The reform process of SONIA is multilayered. Let’s look at it briefly.
Article 27 of the EU benchmarks regulation mandates the administrator to provide a ‘statement of the benchmark’ which includes the definition, underlying market represented and the details of the benchmark methodology. In this spirit, BoE has enhanced the definition of SONIA by making it more explicit. Based on the main uses of the SONIA such as being reference rate for sterling overnight wholesale funds and as a proxy to near risk-free interest rates, the definition would now read as a ‘measure of the rate at which interest is paid on sterling overnight wholesale funds in circumstances where credit, liquidity and other risks are minimal.’
Let’s talk about data. The stimulus to the SONIA reforms are based on significant decline in the volume of transactions underpinning the benchmark in the recent years. This reflects the reduced money market activity as well as a decline in brokers’ share of that activity. Broadening the type of input data is an obvious choice for strengthening the robustness of the benchmark. While the market activity captured by SONIA will remain as unsecured overnight deposits in the wholesale sterling money markets, it will include bilaterally agreed as well as brokered transactions. The Sterling Money Market Data collection (SMM) exercise provides the necessary data for this endeavor which includes daily wholesale transaction data for all sterling negotiated cash deposit transactions. This widened data scope would also result in expanded geographical scope with SMM data collection including all transactions in European Economic Area.
So what’s done with this data? Currently, SONIA is determined as a volume-weighted mean of the underlying transactions. The drawback of this method is that it is prone to be affected by the transaction rate outliers. In effect, skewed (too high or too low rates) transactions would result in skewed rates. While such transactions are not too uncommon, it is important to protect the benchmark from such anomalies. A better method, proposed to be adopted by the BoE, isvolume-weighted median. This involves ordering the eligible transactions from lowest to the highest and choosing the rate associated with the middle (50th percentile) of the cumulative volume. This ensures that the benchmark is sufficiently insensitive to the marginal trades at plausible volumes.
OK. So when will it all pan out? The switch over to the new reformed SONIA will probably happen in Q4 2017, following a 9-month notice period, contingent to the responses to this consultation and other considerations. A 'point-in-time' switch-over strategy is being considered currently. This involves, assuming Ts as the day of switchover: discontinuation of current method on Ts-1; collection of data and Non-publication of rate on Ts and new methodology rate publication on Ts+1. This is consistent with the methodology of publishing the rate on 9:00 am every London business day following to that which the rate refers.
IS THAT IT?:
The reform activity of SONIA is a necessary activity in line with international benchmarks recommendations. To ensure a successful adoption, certain open questions must be addressed. From an input data perspective, how are contributors ensuring that the input data is accurate? Whether there are effective controls to monitor and ensure that the data is accurate. This would involve implementation of an in-depth governance and control mechanism in addition to the BoE efforts and in line with the global benchmarks regulations and recommendations.
While adopting the point-in-time switch-over strategy, how does the BoE ensure that current contracts and transactions are not affected causing market disruptions? This is in the light of the strategy where there is a ‘deferral’ of the rate publication to the next day. The implications for the central counterparties and payment on interest margins will be high given that the SONIA cannot be used as input for calculation for that day.
Another key question to be pondered in this process is what are the contingency thresholds and strategies in case of lack of sufficient input data? Though BoE’s analysis indicates that there is sufficient input data trend under the revised methodology, it is imperative to have contingency measures that might be necessary in case of certain situations such as extreme market events when the benchmark inputs might be affected.