Next steps for new reference-rate adoption

BRISCHETTO In Australia, we are in a very different position regarding benchmark reform from jurisdictions using LIBOR. The BBSW benchmark remains robust after a lot of work to strengthen the methodology underlining it.

We are pursuing a multiple-rate approach, which means we are not advocating for wholesale transition to referencing risk-free rates (RFRs) as other jurisdictions are doing. What we are asking is for businesses to think about which reference rate makes sense for their products.

In some cases this will continue to be BBSW although, as we have mentioned previously, users of one-month BBSW should be preparing to use alternative benchmarks given the lack of liquidity in that market. In other situations, too, referencing AONIA will make more sense.

It is also worth remembering that the rest of the world transitioning to using RFRs may have an important influence on the benchmarks used here. It is important that we see progress on the use of RFRs in the Australian market.

LIBOR transition is a very different matter. It is expected to cease after the end of 2021 so continued reliance on it in contracts is just not an option. We have seen progress on LIBOR transition in Australia but it is clear continued focus and effort is required if institutions are going to be in the position they need to be by the end of 2021.

If institutions have not transitioned to ARRs by the end of 2021, and put robust fallback language in legacy contracts, they will face significant risks individually and will also put the wider financial industry at risk of disruption. This remains a key focus for us.

BRISCHETTO Conditions in the cash market have changed considerably due to the policy package implemented by the RBA in March. There has been a large increase in exchange-settlement (ES) balances that financial institutions hold with the reserve bank. This increased liquidity is an intended result of the policy measures.

Increased volume of ES balances reduces banks’ need to borrow from each other. This means transaction volume in the cash market has declined and the cash rate has fallen below 25 basis points.

This is different from what we have seen in the past. But it was expected and is consistent with the experience in other economies where there have been large increases in balances at the central bank. This very low level of the cash rate is also consistent with the aim of ensuring funding costs in the Australian economy remain very low.

There have been a number of days in recent months on which there has been insufficient volume to calculate the cash rate based on market transactions. The RBA has instead used expert judgement to determine the cash rate, as set out in its fallback procedures.

Users of AONIA will of course want to know the fallbacks are robust, and they can be assured that they are. On most occasions when we relied on fallbacks, the cash rate was recorded as the last published rate based on sufficient transaction volumes.

There was just one day when we published another rate, which was judged better to reflect market conditions. On this occasion, we took into account information about the rates on trades that occurred over a number of preceding days – even though the volume on any one of those days was not sufficient to publish a reliable measure based on market transactions.

We have been seeing transactions in the cash market each day, they are just not always above the A$500 million (US$357.9 million) threshold required to provide a reliable transaction-based measure of the cash rate. The fallback measures allow us to continue publishing a reliable cash rate even in the absence of market transactions.

In this case, the cash rate could be determined to be the last published cash rate based on sufficient transaction volume, the new cash-rate target if one is announced by the RBA board, or potentially another rate related to unsecured overnight funds for cash-market participants chosen by the RBA in its expert judgement and based on market conditions.

These procedures ensure the cash rate continues to reflect the interest rate relevant to overnight unsecured funds. Robust fallbacks are built into the cash-rate procedures themselves, so participants can be assured of its reliability as a benchmark.

KENNEDY It is worth noting that there are differences between the credit benchmark and the RBA published rate. One of our biggest concerns in relation to the credit benchmark was that the fallbacks did not necessarily align with delivering what we thought was sustainable, robust and definable under the IOSCO [International Organization of Securities Commissions] principles.

What we are seeing now with the RBA cash rate is that it is a victim of its own successes from the deployment of unconventional monetary policy. While it is obvious that the fallbacks in place are very strong, there is still a small amount of concern around investors’ and issuers’ trust in its sustainability as a benchmark if it is going for days at a time without sufficient transaction volume.

This is something we will not know until we test the water. However, SAFA [South Australian Government Financing Authority], as an issuer and investor in this landscape, has confidence the fallbacks that exist in the structure are robust according to the IOSCO principles, and can be used and relied upon for continued issuance in whichever format borrowers choose and for investors across the spectrum.

Everything that happens in markets has consequences and we are seeing some of these play out. We remain an observer but we are a strong supporter of the AONIA benchmark, because of the diversification and the risk-mitigation tool it provides and the robustness of the fallback arrangements in place.

BROWN There is a strange situation at present in that cash markets are relatively illiquid and there are very few cash-market transactions. There has also been a significant fall in transactions of bank bills that underlie BBSW calculations.

We are in a situation of flux where it is not entirely clear how the cash market will operate in three months, six months or two years. There is not immediate risk of a negative rate in Australia but New Zealand is making clear moves in that direction and it is a possibility in the US. The RBA might get to a negative rate if the rest of the world does, and if it happens in Australia there will need to be another rebalancing of how the cash system works.

None of the cash benchmarks are covering themselves in glory at the moment because the cash market is in such flux. It would be difficult to say that any particular approach is better than another because it is hard to say what the structure of the system will be, even in the near term.

MIALL BBSW has traded well inside the cash-rate target as a function of increased ES balances and the fact that the banks are very well funded. This is an interesting dynamic. From an investor perspective, it makes us look at alternatives – not necessarily for benchmarks but for yield – and to consider where our allocations are.

MIALL It’s possible. But T-note outstandings are still a relatively small component of the total market for short-dated paper – volume on issue needs to increase meaningfully to support a viable risk-free benchmark. We are certainly headed in that direction, though, so it could have greater merit at some point in the future.

Investors already have other alternative benchmarks to BBSW. As an example, we recently implemented a segregated mandate with a floating-rate benchmark that has moved away from referencing the bank-bill index. It uses the RBA cash-rate total-return index as its floating-rate benchmark.

BLACKSTOCK We noted repeated signals coming from the RBA on its concern around liquidity of one-month BBSW and therefore its validity as an appropriate benchmark. When [RBA assistant governor] Christopher Kent spoke about issuers being prepared to use alternative benchmarks in the securitisation market in particular, we believed we had a leadership role to play as the largest issuer of RMBS and the largest mortgage provider in Australia.

Separately, we were conscious that we were responding primarily to an RBA direction. We noted that SAFA had done some work engaging with investors, though, and thanks to Andrew Kennedy for his strong leadership position.

Ultimately, we were aware that we would be the first to issue RMBS linked to AONIA and there was a reasonable chance we would also be the second. Our intention would have been to bring another AONIA-linked Medallion transaction in this financial year. However, given the context and the extra liquidity we now have through additional deposits and the RBA’s term funding facility, we are less likely to be a securitisation issuer in the near term.

KENNEDY The main reason SAFA considered issuing with a non-credit benchmark was appropriateness. We are a government business and by issuing against a credit benchmark we are embedding credit risk in our portfolios – and those of investors, when they should just be taking a view on SAFA credit rather than second-derivative credit.

It was also about appropriate risk management for SAFA. The entities we lend to are government clients with their own specific financing requirements. By issuing off a credit benchmark and lending off a non-credit benchmark we are essentially creating balance-sheet risk.

We modified our internal practice in 2015 so state-government clients that borrow in floating-rate format do so based off the RBA cash rate rather than a credit benchmark. We had been doing internal work on this for a long time before we began issuing linked to AONIA.

It is about appropriate risk management as well as being able to issue against a one-month benchmark that we thought met the IOSCO principles. Aligning with this is an important risk-management tool for SAFA.

The way we have articulated this to investors is that we want them to understand this is a useful tool for SAFA but also for investors, especially if they view it as a short-term investment product. Taking daily cash risk on a one-year product significantly reduces their portfolio risk.

The ARR and RFR story has developed over a period of time and there is now acceptance about having a suite of products from which to choose. There is confidence now that the hurdles are not insurmountable to being able to transact and manage risk in any particular product. This was a significant stepping stone for a lot of investors to be able to participate.

Work is required in order to undertake a transaction linked to AONIA, which possibly explains why other issuers have not yet come to the table. It requires having a different mindset from risk management and the confidence to sell this externally.

MARSDEN We would like to follow the path SAFA and CBA have taken in issuing AONIA-linked transactions. But, before we do, there are two main challenges for us to consider. First is having comfort with the implied spread for credit and duration in the pricing of an RMBS, and second is the preparedness of investors, trustees and custodians to buy and hold an AONIA-linked transaction.

We have been doing work in the background but first need fully to understand the pricing implications of an AONIA-linked liability, given we have a lot of internal asset-pricing models calibrated to assume margins from BBSW to the cash rate. This is not insurmountable, and market preparedness for a new type of issuance is probably our biggest challenge.

MIALL Operationally, we can invest in AONIA-linked securities and in fact we did invest in SAFA’s most recent deal. Our risk system accommodates AONIA.

There are other considerations, though. QIC is very focused on LIBOR transition and we’ve looked at and continue to monitor the fallback rates of LIBOR-referenced securities for our global credit portfolios. We’re also doing this in the securitisation market in Australia as we want to understand the fallbacks for one-month BBSW in existing RMBS deals.

There is quite an array of these, and we think the securitisation market needs some further evolution to have robust fallbacks – though I note that fallback language in some recent deals has been more robust, in our view.

Liquidity is always a consideration for the bonds we buy. We can invest in AONIA-linked securitisation, though, if it is at the right level. The AONIA-BBSW basis is negligible at the moment but we may require some compensation for liquidity while the AONIA market builds out.

MINEEFF There are valid concerns and queries. In the work we undertook for the AONIA-linked Medallion deal there was a particular workstream around investor readiness. This included operational processes, systems, booking, and engaging with service providers such as trustees and custodians.

We found some positives in this process. The first is that work undertaken in offshore markets allowed a degree of replication in the Australian market for some global platforms. The AONIA calculation we used replicated the cash leg under OIS [overnight index swap], for instance, which allowed a lot of systems and booking processes to use it.

There are still challenges, but we found that working through these was the best form of engagement. We worked through investors’ requirements to understand where they were capturing their data and what systems they were using. This allowed us to dig into the details as we went through the project.

It was a long process and the work SAFA had undertaken certainly helped with our engagement. In RMBS, though, it is a different level of complexity for reporting and other information provided. This was something we had been working towards for a while with the CBA funding team. By the time we got to launching the transaction, we were at the point where the information provided for the AONIA deal replicated information provided for a BBSW deal.

The events of 2020 have probably delayed some other issuers coming to the market with AONIA-linked deals. There is a natural order of potential AONIA issuers in the Australian market. Semi-governments are first then major banks, which we have seen, and then it is probably nonmajor banks followed by nonbanks. The lack of funding need from some of these issuers this year is one reason for there being no further AONIA-linked issuers.

Also, as Phil Miall says, basis spread has contracted significantly. Borrowers were looking at a relatively wide cash-bills spread and the fact that it is now relatively flat has perhaps driven inactivity. The logic for why AONIA works as a benchmark for securitisation issuers holds, though, so I think it is a matter of continuing the development and engagement work regardless of the market backdrop.

CALDER I think the US is now as in line with its progress towards ARR adoption as the UK. Since the first SONIA [sterling overnight index average] bond from EIB [European Investment Bank] in mid-2018, to the end of Q2 2020 there had been £88 billion (US$114.9 billion) of issuance in 181 SONIA-linked transactions. This includes £21 billion in 56 transactions in the first half of 2020.This now is established as the only way borrowers are issuing sterling floating-rate bonds.

It is worth mentioning that very early in the COVID-19 crisis the BoE [Bank of England] and the RFR working groups came together and said there would not be any delay in transition due to the pandemic and that no-one should expect there to be an extension of LIBOR past 2021.

This has been backed up in the approach to legacy assets. These are often LIBOR-based contracts that, for instance in securitisation, are not easily transferable to a new rate. The FCA [UK Financial Conduct Authority] has been given extended powers, relatively in line with the ARRC [alternative reference rate committee] legislative proposal, which effectively means solutions can be enforced upon transactions that for whatever reason are difficult to move to a new rate.

Elsewhere, we have seen 500 SOFR [US secured overnight funding rate] transactions, with US$800 billion of issuance, which means this is a robust market. A lot of this issuance is coming from federal institutions.

This has not quite been the case in Europe, where the approach has been somewhat different. They are focused on replacing EONIA [euro overnight index average] with ESTR [euro short-term rate] and implementing fallbacks to ESTR for Euribor, rather than replacing it. As a result there have only been six ESTR-based transactions – two of these by EIB and one by KfW Bankengruppe.

The difference in momentum is quite striking. However, one thing that comes across very clearly is that there has not been any let-up in the impetus toward standardisation over the past six months.

The BoE has started to publish a SONIA compounded index on a daily basis. This follows on from March, when the Federal Reserve Bank of New York began doing the same with SOFR. It is much easier for bond documentation to reference this official index and calculate the compounded look-back.

CALDER There is a wrinkle in that both the BoE and ARRC have moved toward a shift calculation rather than lag. The lag was effectively overnight SONIA compounded in arrears with a five-day lag, whereas the shift methodology calculates interest according to the number and weighting of business days in the observation period. This gives a slightly truer rate than if it included holidays and weekends.

This is fairly recent. The way it has been approached is that existing issuers using the lag can continue to issue lag-based transactions. But in all likelihood issuance will itself shift over time as the official rate is being calculated using the shift methodology.

The slight hiccup is that the major banking-systems providers have not quite got around to updating the module for calculation yet, although this is imminent and does not necessarily present an insurmountable difficulty.

We are also seeing standardisation in fallbacks. There are two issues here: credit-spread adjustment between LIBOR and the RFR, and the actual underlying successor rate.

One helpful thing is there seems to be a widespread market move behind the ISDA [International Swaps and Derivatives Association] approach to spread adjustment. This means the five-year median difference between LIBOR and the RFR is used. This effectively provides a channelling effect as we move closer to cessation with a standardised add-on over the RFR.

This has been met with widespread approval. The one difference, I would say, is that the UK bond market wants to go to compounded overnight SONIA as the successor rate but there are still calls, in the US through ARRC and in the UK loan market, to move toward a term rate. This seems a more likely proposition in the US where a robust futures market already exists. We have not seen this in the UK.

If there is an overall wrap, it is that despite the hiatus in development of a lot of areas in the last six months, this is one issue where the message from regulators is that things are continuing and there is no reason not to have cessation of LIBOR at the end of 2021. Everyone needs to be moving toward widespread adoption.

CALDER The BoE only started publishing this at the beginning of August so there has not been a lot of issuance from it. In many ways, though, I see it more as a facilitation and clarification rather than an impetus for more transactions. It has become clear that the environment in the UK would not be favourable now to anyone attempting to issue in public markets based on anything other than SONIA.

We are now in the phase of fine-tuning. There has been SONIA-linked issuance since 2018. It has been widely adopted, and users have updated their systems and are clear on the spread adjustment between LIBOR and the RFR. This is often the discussion in pricing when there is a new transaction, as well as whether it is based off one-month or three-month because the bases are different.

There is wide familiarity with the mechanics in the market. These latest iterations of fine-tuning the fallbacks and the methodologies are therefore recognition of an established market and participants realising what could be done better.

BROWN One of the biggest challenges in the Australian market at present is that a lot of users are trying to design a perfect benchmark. We need to consider what the entire financial ecosystem will look like once there is a new benchmark embedded in it.

If a term reference rate is more beneficial for some users and an overnight compounding rate more beneficial for others, what is really needed is an OIS swap – because it allows both to operate together.

In Australia, where a lot of compounding cash transactions have been written in such a way that they mimic the OIS formulation, there is no reason why issuers cannot print in one format and have a swap with a bank to provide the hedging they need rather than everything having to come from the benchmark itself.

This would be the benchmark operating together with the rest of the financial system to provide what users need. I appreciate that in some ways this is a leap of faith. Clients would have to ask their banks to help and there would be small differences in methodologies. But it is not insurmountable.

The problems seem small enough that when there is a financial incentive to solve them – as there will be if there is a widespread movement to a new benchmark – it should be possible to create a financial-markets ecosystem that works better than a benchmark would work on its own.

At CBA research, we are confident we can get to a new benchmark that does not embed credit risk into transactions where it does not need to be there. There are good reasons to strip these credit risks out of benchmarks, but it is a complicated process to get there.

We hope market users are looking at what the new benchmark ecosystem could become rather than trying to pick holes in any particular benchmark.

KENNEDY A good way to reflect on this is through the last transaction we did, where we decided to go out on the curve to provide some term funding in FRN [floating-rate note] format using AONIA. We went from a monthly to a quarterly coupon, and out to three years on the curve. The change in appetite and engagement from the investor base was overwhelming and we ended printing volume in excess of our target.

Dialogue has increased substantially. One of the unintended consequences of lockdown is that we can have a lot more one-on-one conversations, rather than scheduling appointments to see people. This has given us a better sense of what some investors’ touch points are and what brings them to a transaction.

The investor base has changed from what we initially thought we would see in these transactions. It has certainly broadened. We have been surprised that the number of offshore investors has grown in each transaction, well past the couple of accounts we first targeted.

We have now raised more than A$2.6 billion in this format and have had around 80 unique investors, of which a pleasing component has been offshore. Investors’ risk appetite keeps changing and appetite for diversification keeps growing.

Developments in offshore markets have been helpful in driving these changing attitudes, as has the messaging of the reserve bank to reinforce the concept of AONIA as a sustainable benchmark. We plan to continue issuing FRN funding in AONIA format.

CALDER I certainly understand the motivations of the offshore investors that have participated in SAFA transactions. It comes with the mindset that this is the way the world is moving in developed markets.

I would counter this, though, with the relative lack of compulsion we have seen in the Australian market and the fear that there will end up being a bifurcated market with a knock-on effect on liquidity. Perhaps it is early days, but I have one AONIA deal in the portfolio and many more BBSW deals – some of which were issued after the AONIA transaction.

With the swathe of secondary-market liquidity that came into the market in late March and April, it was noteworthy that we saw plenty of high-quality issuance but not much AONIA-linked issuance.

An element that has driven the market forward offshore has been not so much the carrot but the stick. I see a danger of a halfway-house approach that creates an illiquid and bifurcated market.

MARSDEN I think the market transition will be led by the buy side, whether it is by stick or by carrot. In the last fortnight we have been engaging with our US dollar investors and there has been a noticeable mandate migration away from US dollar LIBOR. This is particularly noticeable in 144A accounts and the state-based retirement and pension funds.

We are now looking at two sets of investors. But this bifurcation is not necessarily a bad thing. Some accounts can still buy LIBOR FRNs and others are looking for SOFR-linked issuance. If we can take anything from the UK and the US, I think this is how our market should look to transition. But I think this needs to be led by the buy side.

MIALL The buy side needs to make sure it has the operational and systems capability to buy RFRs. This is certainly the buy side’s role to play – but I am not sure it should necessarily lead the transition.

It needs to be a market solution and it is about timeframes. As we stand now there is no compulsion. There is a directive that the market needs to find an alternative to one-month BBSW, but there is no drop-dead date as there is with LIBOR.

The market will probably move to a solution, but in the absence of a cessation date for one-month BBSW it will likely be over a longer timeframe and it falls on the buy side, the sell side and issuers to work on this.

The RBA’s work around fallback language for BBSW improves the fallbacks for RMBS, so this may be a solution that actually prolongs the use of that reference rate. If there needs to be a quicker transition, compulsion may be required.

BLACKSTOCK I take the point raised about liquidity. Ideally, we would have some company as an issuer of AONIA. But the reality is that others need to make their own decisions, so we have made a change in our syndicate structure to try to foster secondary-market liquidity.

FREILIKH For the Medallion deal issued last year, in addition to CBA – which was arranger and sole lead manager for the deal – we appointed three co-managers to support secondary-market trading and provide bond revaluation services for investors to foster secondary activity.

We also spent considerable time with ICE Data Services to make sure it could provide secondary-market revaluation on the AONIA bonds to investors and custodians.

BLACKSTOCK Potentially. We are not at that stage yet, though, and three-month BBSW remains a robust benchmark. While investors are keen to buy in this format we see no reason to move away from three-month BBSW.

MINEEFF In our role of arranging and bringing transactions to market, our advice always aims to offer issuers the best transaction to bring for broad distribution and to meet the client’s objectives.

The CBA funding team had a clear objective to deliver an AONIA-linked transaction with a very clear mindset and approach. It took a lot of work to get there but much of this work is replicable for other issuers.

Some issuers, as Andrew Marsden says, are positioning according to the feedback they receive in offshore markets and others are looking at whether it would work for them to bring an AONIA-linked deal in the Australian dollar market. We need issuers, investors and other stakeholders to bring this together.

We have seen, with the Medallion transaction, that it is possible and it makes a lot of sense. The deal was a capital-relief transaction, too, so we are not just talking about triple-A investors: it is the full capital stack.

I think we may have expected there to be a natural order of issuers we would see come to market with AONIA deals, but this has been disrupted over the last few months. The logic holds, though, and it is still about bringing the right parties and stakeholders together. If an issuer wants to issue linked to AONIA it absolutely can be done – we just need to do the work to bring the transaction to market.

CALDER In the US, ARRC believes a futures-based development of a forward-looking rate is the way forward. I think there is enough volume and robustness in that market to provide some degree of confidence. It is interesting, though, that UK market participants seem happy to stick with the compounded look-back rate for SONIA.

I think it will be a case of different markets deciding what is right for them. As an investor in structured finance, it would be nice to see a forward-looking rate. But having now lived with what we have for nearly two years I look at the positives and the differences and wonder whether we are gaining much by going through the process.

We are already getting a robust rate. Given the central banks’ attitude to interest-rate setting I think it may be better that we have a backward looking rate. When rates go negative there will at least be a short window where we are looking back to positive rates!

I have thought about forward-looking rates a lot and want to get excited, but I find it difficult. For corporate treasurers – who want to know how much they have going out each month in interest payments – I understand totally. But as a bond portfolio manager I am less concerned.

BROWN Australia can use an old-style market benchmark of OIS rates, but there has been pushback because rates are supposed to be risk free and based on real traded instruments rather than derivatives traded between bankers.

If you are not prepared to have this kind of forward-looking rate you are left with the need to transact public instruments against a backward-looking rate. But if you want to enter into a swap afterwards you can hedge the risk. The benchmark is not the only financial-management tool.

KENNEDY We spent a lot of time looking at all the different ‘risk-free’ options when considering the appropriate rate to use. The problem we had was that when we reviewed the data we could not guarantee what we would get was going to meet what we considered to be a best-practice benchmark.

We didn’t consider OIS or repo to be suitable, at the time, to the composition of transactions and there was a lack of progress for determining a rate-set process or fallback arrangements.

CALDER I think there is sensitivity to potential discrepancies that can emerge in forward-looking rates provided by whomever. I wonder whether the impetus will be for reliance on officially published compounded indices for official rates.