Risk-free rates: Sonia and Sofr at the forefront of new era
The bond market is well ahead of schedule in its adoption of potential Libor replacements, with several issuers having printed notes linked to the Secured Overnight Financing Rate (Sofr) in the dollar market, and to the Sterling Overnight Index Average (Sonia) in sterling. Borrowers are setting strong standards for other participants to take up, as well as adjusting structures to ensure the eventual market is optimal. That does not mean the job is finished, of course. GlobalCapital spoke to some of the pioneers in the Sonia and Sofr markets about their work so far — and the challenges ahead.
Participants in the roundtable were:
Nadine Bates, treasurer, Fannie Mae
John Gerli, chief capital markets officer, FHLBanks Office of Finance
Peter Green, head of public senior funding and covered bonds, Lloyds Banking Group
Xavier Leroy, capital markets officer in charge of sterling, European Investment Bank
Gregg Mazzaro, director, syndicate, TD Securities
Donald Sinclair, head of asset liability management, World Bank Treasury
Craig McGlashan, moderator, GlobalCapital
GlobalCapital: EIB started the ball rolling for Sonia issuance in the post-Libor scandal world with its £1bn five year FRN in June. Xavier, could you talk us through all the work that went into the trade?
Xavier Leroy, EIB: EIB’s Sonia journey started in early 2010 when we issued the first Sonia floater, before the Libor scandal. My colleague Thomas Schröder realised Libor was an insufficient benchmark, was too disconnected from the market and could be prone to manipulation. He considered that Sonia was a much stronger alternative as it was more robust and subsequently developed a structure close to what we’re using now, so this year’s Sonia trade is very much his legacy. This was extremely visionary at the time as it predicted what the market was going to look like eight years before everybody else.
However being right too early sometimes doesn’t work and despite the very solid structure proposed the trade did not sell as expected, mainly because investors were not ready for it.
What has changed since then is the regulatory environment. We had all the Libor manipulation, which probably drove the regulator to come up with a new solution. Investors will only adapt when they are under pressure to change, they are human beings after all.
Given our initial transaction, we’d been approached over the years by various bodies —banks, public entities and so on — to get our view on what the replacement of Libor should be and if Sonia would be appropriate.
We said yes, so EIB had a responsibility in making sure that, when the timing was right, we should do our part in building the new benchmark. The UK is one of our shareholders until at least March 2019 so we considered it our duty to help them set up their future risk free rate.
In March 2017 we started working faster on that project. First, we had to check our internal systems to make sure we could handle the product in various forms. It was mostly OK, given we’d issued that floater almost 10 years ago and the structure was similar. We had to discuss with the leads, however, what would be the optimum structure, because the market had developed over that time.
The main discussion was whether it should be backward or forward looking, and the clear answer was backward, because that was what the regulator wanted. The goal of Sonia is to get rid of all possible manipulation that came with Libor. You’re back to Libor if you use forward looking because it’ll need to be priced off a forward or derivative. That input will come from banks. Even if it’s traded it can still be manipulated. However, if it’s backward looking, it’s from the Bank of England and while it relies on market data people would have to manipulate every single data point in history to have an impact on the bond. In this sense, the backward looking approach is way less corruptible.
We then had to fine-tune it. On the previous transaction the spread was compounded, whereas on the new one it isn’t — we had the pure spread component and the pure Sonia component. We had to ensure the structure was bulletproof, so we ran it through all the entities — the UK DMO, the FCA, the Bank — to make sure it would match what the market wanted.
Then the technical work really started, which was probably the most complex part. Investors need to be able to value bonds. The simplest way is Bloomberg, so you need a page on Bloomberg that can load and price the instrument. But the structure we were going to use — backward looking with a five day observation period — was not available. So Bloomberg had to set up such a page and ensure the calculation was accurate, which took some time.
Then it became more complex as we had to deal with the booking systems for each bank. While Bloomberg is fairly standardised, it’s not the case for dealers. Some use Calypso, others Murex. Then they have different versions of each that are tailor-made for each bank.
Then things started to get messy, because the valuation or booking didn’t work, depending on the system. That was compounded by Sonia’s ticker being ‘Sonia/n’ — the forward slash was not recognised by most of these systems, which took months to solve.
We knew going first was going to be super-complicated because we’d have to do all the heavy lifting — with the risk of failing. But we wanted to be first to ensure the final product would be in line with what would make sense for the market.
That might not have been the case with every issuer. A dealer that was not so aware of these products could convince a smaller issuer to go first to capture more demand. That could potentially have ended up being a messy situation because we have a herd mentality in markets. People copy the last deal that worked and that becomes the market standard.
GlobalCapital: Did Fannie Mae have a similar experience to EIB as it brought the first Sofr linked bond, a $6bn three trancher?
Nadine Bates, Fannie Mae: At the beginning of the year we were asked to become a participant in the Alternative Reference Rates Committee (ARRC). That helped us to jumpstart our thinking internally. Prior to the first publication of Sofr in April, we started internal systems work to capture the rate in our database, then we brought the first transaction in July.
We had the benefit of having issued Fed funds floaters, which are similar in terms of the daily reset, so we could look at our booking and accounting systems and make any changes necessary to deal with the new rate.
We also had to work with our operations team because they needed to execute the reset on the floating-rate note (FRN), plus our risk, communications, legal and tech teams who all had roles to play in successful issuance. We had to span the enterprise to make sure we could book the trade, execute the daily reset, make sure it fit into our risk systems, be able to account for it, ensure the legal language was correct in the pricing supplements and so on. It took many months.
From an external perspective, we wanted to make sure our dealers could support a new security in the secondary market. We also thought about it from the Fed’s perspective, as they settle all of our debt securities. We didn’t want to assume that their systems could accommodate a Sofr issue, so we reached out to make sure they could.
We meet investors on an ongoing basis. Because the Fed announced Sofr as the new index in 2017, we were able to talk with investors about it, see how they were progressing and assess if they would be able to buy a Sofr-linked bond if we were to issue.
We thought about the maturities we were going to bring. We had a little information that the 2a-7 money funds that regularly buy FRNs would be the main target audience, so we knew there’d be a benefit of staying shorter, inside two years.
Again with an eye on not knowing precisely which investors would buy, we decided it would be a good strategy to offer a few different maturity points. That would help to ensure broader participation, as well as a starting point for a potential curve in the front end.
GlobalCapital: FHLBanks added to the growing Sofr supply in November. Can you talk us through the trade?
John Gerli, FHLB: The Federal Home Loan Banks worked diligently in the last few quarters to be operationally ready to issue our first Sofr-linked securities. The issuance in November completed our initial development phase and was a learning opportunity in the development of the Sofr market.
We had ongoing discussions with investors and dealers throughout the development process to understand their structural preferences and readiness to participate that included feedback on prior Sofr-linked debt issuance.
We chose to use a syndicated format to promote transparency for investors and to aid in the development of liquidity to this new index. The securities drew broad investor demand from what we’d define as a strong and diverse investor base.
We are very active in issuing many different securities in tenors of one year and less and felt that a dual tranche Sofr transaction of six and 12 months was relevant both to our funding needs and the development of the market.
GlobalCapital: World Bank is in the rare position of having issued both in Sofr and Sonia. What differences have you found in the two markets, in terms of the work done (and still needed) to achieve market adoption?
Donald Sinclair, World Bank: The biggest difference between the two reference rates is investors’ exposure to them. Sofr is brand new, while Sonia has been around for a while, so investors are more familiar with it. That also means that the infrastructure is more complete in the Sonia market.
The biggest part of the work still need, particularly for Sofr, is educating and attracting the investor base, building liquidity in cash and derivatives and establishing term rates.
Gregg Mazzaro, TD Securities: It’s really about the difference between the two markets, trades and structures.
As Don says, Sonia has been around since 1997 and thus the UK investor base is familiar with Sonia-based derivatives. Comparing EIB’s inaugural trade to the most recent Sonia transaction we were part of, namely from Yorkshire Building Society, the take-up from the investor base has been quick. Investors are mostly comfortable with the rate, so it was about getting them comfortable with the structure, getting the plumbing up to speed, working on coupon calculations, accrued interest and so on.
Sofr, on the other hand, is a brand new rate but these transactions are using a structure that is far more familiar to the investor base. Ultimately, what we’ve seen since Fannie Mae’s inaugural transaction, is issuers playing around with the structure a little more, when compared with what we’ve seen with Sonia transaction evolution. The four day lockout period has been reduced to two days, while recent deals from JP Morgan and Wells Fargo have defined reset dates slightly differently — their weekend resets are run off Thursday’s Sofr fixing, matching derivative market conventions.
GlobalCapital: Lloyds was the first FIG issuer to sell a Sonia-linked FRN. Did you follow the same structure as the supranational trades before you?
Peter Green, Lloyds: The transition from Libor has been very high on the radar since the start of 2018 in terms of the interaction between the various Bank of England risk-free rate working groups. Lloyds is a member of some of those.
We were keenly interested to be close to the front of the queue for bank issuers in this product. EIB’s trade was hugely successful and acted as a spur for us to push ahead with some of the system testing work required.
The derivatives market for Sonia was fairly well established, so adapting systems for bond issuance was not too problematic.
The UK Listing Authority (UKLA) made the amendments to programme documentation, to accommodate Sonia, slightly easier than would otherwise have been the case.
We did no particular investor work around the transaction. The only investor dialogue we had on the trade was when we put out the mandate announcement in early September. That clearly gave our dealer banks the opportunity to speak to investors and gauge interest.
GlobalCapital: How useful was the earlier issuance from EIB and others in structuring your note?
Green, Lloyds: The EIB trade helped in establishing a market convention that seemed to work. The key factor for the development of the Sonia market is that its conventions are as closely aligned to swap market conventions as possible. The trades that have come to market have been compounding with a slightly longer lookback period than in the swap market, but it was entirely logical how EIB brought its trades and others have since followed.
Mazzaro, TD Securities: EIB made the initial push of getting the UK investor universe focused on Sonia structures. It was the kick-start to the process of alternative rates and played a big part. As Peter points out, the structure versus the derivatives market was a logical choice. But it was EIB’s initial engagement with investors that led to the Lloyds’ trade and those that have followed since.
GlobalCapital: What were the biggest challenges when bringing these first Sonia and/or Sofr linked notes? With hindsight, would you have done anything differently?
Leroy, EIB: We did a test trade, which we weren’t so keen to do at first as it’s still a use of resources. But we’re glad we did, because it meant we could discover a lot of the issues before coming along with the final product, which in the end was fairly smooth. The biggest challenge was all the IT issues.
If we had to do something differently, it would maybe be in terms of trade management. We had two structuring advisors and it was the first time managing such a product for all of us. Sometimes the split of responsibility was not as clear as it should have been, or there was no seniority between leads.
If we were doing it again we’d only have one bank, or several banks with one senior on responsibility. That way we could have someone cracking the whip to make sure all the interests aligned, rather than people pulling in different directions.
That being said, given the difficulties that we had to face as a pioneer issuer the result we achieved is outstanding.
Green, Lloyds: We wouldn’t have done anything differently. The trade worked very well. EIB’s was over a five day period while we were over a three day period, so it’s encouraging that the time from marketing to deal execution is shortening.
For future trades it will be worth testing the willingness of settlement agents or investors over having the lookback period shortened slightly so it’s more aligned to swap market conventions. Generally though, the evolution has been logical and at a decent pace.
Bates, Fannie Mae: One of the largest challenges was understanding whether or not investors would be ready for the issuance and whether or not they would be willing to buy given that it was a new transaction with no track record of liquidity. In addition to that primary challenge, we had to determine the appropriate pricing points to bring the deal to market.
We remediated the first challenge by working with a dealer to sound out investors. On the pricing points, investors have a pool of money and choices they can make, so we tried to look at the other instruments they had and tried to price relative to those. That was hard, but based on the demand we saw we were able to tighten from our price indications, which was helpful.
Sinclair, World Bank: The biggest challenge, as with any new market, was finding investors that can take the exposure. That’s where our friends at TD really helped out. Investors have to understand the rate, have guidelines that allow them to buy it and systems that can evaluate it for their books. We found that doing the trade got some of our investors motivated to get themselves set up in time for the next one.
GlobalCapital: How responsive were investors to your plans in the build up to these first Sofr issuances? Did you find a variety of views on potential structures?
Bates, Fannie Mae: A handful of larger investors were very responsive to the dealer we had designated to sound out investors leading up to the transaction. We also had the benefit of looking at Fed fund floaters and the investors were very comfortable with that convention, so we took that as the starting point.
Fed fund floater interest is calculated using a simple average so we had to decide between using that or compounding the interest. Based on our understanding of how some of the 2a-7 funds are set up, operationally it was going to be easier for them to deal with the simple average, so we decided on that.
We also had a four day lockout period that we had to accommodate from an operational perspective. We did get investor feedback that if that number could shorten then it would be much appreciated. We couldn’t change at the time but took it under consideration and were able to update our second issuance to incorporate this feature.
Leroy, EIB: It was immediately very clear that investors endorsed the structure. We had interest from investors that do not necessarily buy us in sterling but wanted to buy this to make sure they were Sonia ready.
By buying the bond they could force changes internally. Their legal teams would have to look at it, their back offices would have to configure it. If they’d asked for that work they’d have been told it would take six months.
Mazzaro, TD Securities: EIB made a small change compared to its 2010 Sonia trade with regards to coupon compounding. The new approach mimicked the swap market, which played a big part in getting investors comfortable around the structure.
From a syndicate perspective, one of the things we were slightly concerned about was investors leaving only ‘test’ type orders in small size. An investor that normally may leave a £50m order might say, ‘It’s a new structure/ new concept; we’ll put in for £10m instead of our usual size’. Ultimately this was not the case and we were pleasantly surprised with investor participation.
Green, Lloyds: Several investors said they’d prefer a term rate. That was easy to counter because there’s no term Sonia market. For the market to evolve, investors need to be flexible. There is a debate to be had as to whether the bond market actually needs a term reference rate for Sonia. The trades that have come have all been in the same daily compounding structure with the lookback and there hasn’t been any particular pushback from investors.
The key element is for investors to accommodate the backward looking structure into their plumbing. But clearly the alignment between bond and swap is vitally important and the evolution of a term rate will be interesting to watch. For now, the convention seems to be set.
Mazzaro, TD Securities: It really was the appeal of the Lloyds name: attracting the widest possible spread of investors, that allowed the syndicate to gather as much info as possible at the IOI stage of the transaction. The inaugural Sonia-linked covered bond needed to be done by one of the bulge-bracket UK issuers who stood the best chance of getting clarity from investors.
GlobalCapital: What sort of investors have been buying Sofr and Sonia bonds?
Mazzaro, TD Securities: It’s difficult to provide a blanket statement on which investors are able to buy Sonia and which aren’t. We made great strides during the Lloyds/Santander UK CB transactions in our work on identifying which investors can buy Sonia.
I’d say, looking at the three larger investor sectors, namely bank treasuries, building societies and UK real money/asset managers, it’s fair to say 85%-90% of investors can now participate in a Sonia transaction. There’s still a bit of work to do for some of the overseas bank treasuries in Europe and Asia.
The sector that outperformed in terms of getting Sonia ready was the building society sector. We thought they’d be a bit slower moving due to issues around internal constraints, but since EIB’s inaugural deal versus the most recent YBS transaction, the sector has been heavily involved.
Green, Lloyds: We had around 75% of bank treasuries and around 25% of asset managers. In covered bonds, we expect treasuries to be the highest component. Central banks and official institutions have maybe slightly lagged the move to Sonia when I look at the breakdown versus a Libor bond, but that’s expected.
We were pleased with the reception of our trade. Our Sonia trade had 63 investors, while our last three year Libor floater in covered bond format — which we did in January — didn’t have dramatically more investors. That proves investors are engaged.
There were a couple of bank treasuries that couldn’t buy the Sonia bond, but the feedback and evidence we’ve seen is that those treasuries are doing the work to get the product approved for future issuance.
Where I’d most like to see investor base expansion is in real money rather than bank treasuries.
Leroy, EIB: The bulk of Sonia investors [for SSA issuance] have been bank treasuries, which is typical for any floating rate product. It will expand a bit, although not necessarily to other types of investors because bank treasuries are the main buyers of floating rate notes. Very few central banks or UK real money would buy sterling floaters.
The number of bank treasuries able to look at Sonia trades, though, has expanded already since our bond and could expand slightly more. We missed a few because, although they wanted to buy, they were not ready as their systems could not yet cope. Most of those are now ready because they tried to fix their systems as fast as possible. One or two may still not be ready but this number will diminish.
Bates, Fannie Mae: We were confident that 2a-7 money funds would be the primary investors, but were pleased on both our first and second Sofr issuance to see it expand beyond that. In addition to 2a-7 funds, we saw interest from asset managers, state and local government, corporate, insurance companies, banks, and even some other GSEs.
Sinclair, World Bank: The investors for our Sofr and Sonia bonds were the same sort of investors that already bought our Libor floating rate notes — central banks and non-money market funds. So no surprises there. The bonds were outside the money market realm as the Sofr was a two year and the Sonia a five year.
With time and more issuance the investor base will widen and deepen. More investors will have the authority to invest in these rates and develop the market.
GlobalCapital: How confident are you that the Sonia and Sofr markets will be fully functional by Libor’s scheduled shutdown at the end of 2021?
Sinclair, World Bank: In many ways, the transition plans for both the US and UK are ahead of schedule so there’s no reason to think the markets won’t be ready on some level. The key will be getting liquidity into the markets. Liquidity will be driven, by an extent, to investors, issuers and dealers adopting the alternative rate.
You could argue the market is functioning now because trades are being done. Is it ready to absorb everything going through the Libor market? No. But it’s on track and we still have time. It’s fair to say we don’t know when Libor will no longer be functional, but at least the alternative is being established.
Gerli, FHLB: We feel the Sofr market developments up to today are quite encouraging and favourable. That said, there is still a tremendous amount of education, financial products development, documentation creation and other transitional activities that need to occur in the coming years to ensure a smooth transition.
Ultimately, this is a large market with $200tr of financial transactions linked to Libor and no roadmap or predictable course of action.
Each institution will need to perform its own readiness assessment, and create its own independent transition plan from the known market to an unknown one. Those activities have accelerated in 2018 and are on track but there remains a significant amount of work to do.
FHLB recognises we are in a unique position as a leading issuer of Libor-linked securities to participate in the cash side of the market. We’re thinking about that in terms of our members’ needs as well as our general corporate funding requirements from an asset/liability management perspective.
However, it’s a learning experience for all of us. Until we completed our first transaction, we were contacting intermediaries — both on the sell side and buy side — to provide us their views on developments.
We now have better market knowledge having completed our first transaction and were encouraged by the outcome and support of dealers and investors to our first transaction.
Bates, Fannie Mae: We’re very supportive of what ARRC’s doing to provide an alternative if Libor goes away in 2021. We’re not taking a stance that it will or won’t. We’re accelerating the use of Sofr internally to make sure we’re operationally capable if it does.
A lot of progress has been made and the market is up to $23bn in Sofr debt issuance debt, which is amazing. It’s way ahead of where people would have expected it to be, but there’s a lot more work to be done to get to a term Sofr product by 2021. The Libor-based market is over $200tr and it took time to build up to that. But Sofr derivatives transactions are happening that will help build this market.
Leroy, EIB: I’m fairly confident. We’ve seen really little Libor bond issuance over the last few months and most floater issuance has switched to Sonia. The switch was much faster than we anticipated. But, although the bond switch is quite easy, the trickier part will be to handle long dated legacy bonds and the loans market.
Green, Lloyds: We should look at the development of RFRs in two buckets — front book and back book. From a front book perspective, we can say without too much fear of contradiction that the market is as close to fully functioning as we can get. We’ve seen a couple of Libor linked notes since the EIB’s Sonia trade in June, but the amount of Libor linked issuance has decreased drastically since then. Subject to investors and issuers developing systems and overcoming some of the practical elements, it will be the exception rather than the rule to see Libor issuance in primary.
The challenges are more for the back book, given some of the complications in changing terms in documentation. There’s clearly a wide variety of fall-back provisions in legacy bond contracts that makes transitioning the back book trades from Libor to Sonia slightly more challenging. That’s something that will get far more focus in 2019 and beyond as we move closer to the end of 2021.
GlobalCapital: Are you planning to stop issuing Libor floaters?
Gerli, FHLB: We’ll evaluate when to stop issuing Libor floaters between now and 2021. Our decision will depend on the evolution of the Sofr market, the relative liquidity that continues in the Libor market, and ultimately the pace of our member banks’ adoption of Sofr. It’s really dependent on these three factors and no one can predict the pace of these changes.
Sinclair, World Bank: We are, at the heart of it, still a Libor-based institution, so if there are investors that want to take a Libor floater we will do a Libor floater. We’re still hedging ourselves with Libor-linked instruments. The market’s not ready for us to stop Libor issuance, so we’ll still do it.
Leroy, EIB: We are in a good situation because our longest Libor bond matures in early 2022, so the last fixing will be in December 2021. We got lucky. And we’re not going to issue in Libor anymore.
GlobalCapital: Do you expect the pace of Sofr issuance to increase rapidly in 2019? What about your own issuance?
Gerli, FHLB: We believe those market participants that buy Libor-based securities either have capabilities or are working towards operational readiness to engage with Sofr bonds. Look at 2018 — there was really no activity in first two quarters, then good initial issuance in the third and fourth quarters. I would anticipate an active trend to continue throughout 2019.
But this is a material, structural market development and there’s more work to be done. We’ve developed operational and product capabilities to issue Sofr and they must be matched with investor preferences and appetite.
We saw that in our first transaction, but we also saw some investors are still working towards operational readiness but not there yet.
We intend to support the market through regular issuance in 2019. As demonstrated by our syndicated issuance in November, we anticipate a systematic and programmatic approach in 2019. As the market continues to develop in 2019 and beyond, we’ll assess our members’ funding needs and investor feedback, and adjust our methodology accordingly to support the market and our funding needs.
There’s really two drivers for our funding. First is the individual and collective needs of the FHLBanks for Sofr-linked funding, which will ultimately be driven by our member’s advance demand and internal ALM needs. Second, for the FHLBanks, issuance is a collaborative process and market access and low cost funding is paramount to our business model. Cost-effective funding and prudent risk management factors are always considered and can influence issuance activity.
The pace of activity in 2018 was very constructive for the market’s development and the increased focus by the issuers, investors, regulators and dealers has brought Sofr forward in the timeline in a very effective way.
However, as we approach 2019 there remains a tremendous amount of market development activities and participation from a broad group of market participants that needs to continue.
Bates, Fannie Mae: Our funding needs are always dependent on market conditions. We’re using short dated Sofr issuance as a substitute for other short dated products. Our balance sheet has continued to shrink, so our funding needs are limited. We’ll evaluate opportunities to support the market’s development and try to be present to help it evolve, but we’re not putting out a specific target.
Sinclair, World Bank: We would expect to do more Sofr-linked issuance in 2019. How much and how fast really is going to depend on how the market develops. We’d be open to doing Sonia as well, although we have a more natural exposure for Sofr than Sonia.
GlobalCapital: What about the other Sonia issuers?
Green, Lloyds: In a global context the sterling market is reasonably small and the majority of sterling market transactions are likely to come in fixed rate format. So we may not see a dramatic increase. Sonia floating issuance from UK issuers will continue, but it will be the evolution of the cross-currency swap markets referencing RFRs that drive how much issuance there is from borrowers without sterling balance sheet needs.
It’s different in the US, where a higher proportion of the overall market is issued in FRN format.
Mazzaro, TD Securities: Once UK Inc is fully up to speed with Sonia — which we’re starting to see with TSB announcing a mandate for a possible Sonia-linked covered bond transaction, focus will shift to other big players in the sterling covered bond market. Attention will be paid to the Scandinavian, Canadian and possibly the Australian issuer base. That said, as these issuers tend to swap proceeds into their home currencies, until there is a functioning cross-currency Sonia/Libor swap market, the perception is that they will have to pay extra execution costs.
In an ideal world you’d like to see a non-UK bank issue a Sonia-linked covered bond transaction by summer 2019.
GlobalCapital: All the Sonia issuance so far has had a similar structure, compounding each day’s Sonia rate to calculate a number each quarter. Is the market set on this model, while leaving room for a few tweaks? Will we see competing approaches?
Leroy, EIB: The structure is set in stone. In SSAs, EIB and World Bank have picked this structure, so the largest issuers have endorsed it. Banks have been following the exact same structure. Although corporates have not printed yet, I don’t see them going crazy on changing the structure.
Green, Lloyds: The lookback is the main tweak needed. The RFR working group is debating whether there’s a view on an alignment in different market conventions on the lookback in the UK versus the lookback in the US.
Issuers will have the ability to structure trades in any way. The difficulty is, do investors respond to a wide variety of different structures? Uniformity of approach is the easiest way to encourage deeper engagement from both investors and issuers, and build liquidity. Efficiency between bonds and swap markets is the way we think the market should go.
GlobalCapital: Has there been any variance in structure in the Sofr market?
Gerli, FHLB: Product development is an interesting question. Consistency in issuance structure is important to promote market development initially, and we believe it’s in our best interest to be a participant in that manner. In fact, many investors indicate that consistency of structure is important in developing secondary market liquidity as the transition away from Libor takes place.
That being said, as the market develops further in the coming years, there may be opportunities for other structures, including term Sofr. There is a lot of work to be done by issuers to understand the needs of investors before we see that. We are focused on regular discussions with our investors and dealers to gauge the markets appetite for debt product evolution.
Bates, Fannie Mae: We were very thoughtful in our approach of how we structured the bond so were pleased to see other issuers adopt a similar structure. But I don’t know if it’s set in stone — it’s too early to say.
The market still has to evolve. I wouldn’t be surprised to see a variation around the Sofr bond conventions. We did a simple averaging on the interest. We selected that over compounding because we knew some investors we expected to buy the first bond could handle simple averaging consistent with Fed funds floaters. But we remind investors that systems should be built to handle not only simple averaging, but also compounding. That’s something I could see evolve over time, because that’s how the derivatives market works, although there’s no need now. But beyond that no other obvious changes come to mind.
GlobalCapital: Fannie sold its second Sofr FRN in October. What differences did you notice compared with the first trade, in terms of investor demand, pricing, secondary performance, investor questions and so on? Was there any difference in structure?
Bates, Fannie Mae: After the first transaction, we had a lot of questions from investors and dealers as to whether this was something we’d do regularly. Our response was that we wanted to be supportive of what the ARRC is trying to do in providing an alternative to Libor. We didn’t promise when our next issuance would be, but said that if at some point we felt there had been sufficient progress made then we’d entertain another issue.
Examples of progress could be a demonstration that additional investors and dealers had made progress on getting Sofr as an approved index, or if there were changes that we could make to the structure that investors would desire, such as shortening the four day lockout. Also, because our first issuance provided three maturity points, if we waited a couple of months for those to roll down the curve, then there might be another opportunity to bring those same points and further define the curve.
We saw other issuers come with similar and different maturities out to two years, so decided to bring our second deal, which had many new investors. The demand for the second was over $18bn compared to over $8bn for the first. The number of investors from the 2a-7 community expanded significantly. S&P’s recognising Sofr as an anchor 2a-7 reference rate for rated money market funds immediately after our first issuance allowed more of them to come in. We also added another dealer, so more dealers could support Sofr in the secondary market. That was another principle in demonstrating progress.
The spreads on our first issue had tightened in secondary, so we could price the second issuance three months later at a tighter spread. The six month had tightened by 4bp, the 12 month by 5bp and the 18 month by 6bp. There was about 3bps between each maturity, which provided a nice curve.
The only change in the structure was taking the four day lockout down to two days. We were able to improve our operational processes and the Fed was able to support it as well.
GlobalCapital: The Sonia bond market definitely seems to be ahead of schedule. But is the swaps market where it should be?
Leroy, EIB: It’s going to grow. More issuers are going to issue and they will need to manage some legacy portfolios by swapping some of that back to Libor or vice versa. The vast majority of issuers can still decide to ignore Sonia if they want to.
They can issue in fixed rate, swap into Libor, and probably still lend in Libor. But at some stage they will end up with some Sonia assets or liabilities on their balance sheet and have to deal with them. When those investors have to deal with more and more Sonia products it will help develop the market as there will be more and more participants.
Green, Lloyds: Sonia is not a new benchmark, which is probably one of the advantages the UK has over other jurisdictions. The Libor/Sonia basis swap market is functioning. We’d expect more liquidity to come, but I don’t think it’s a constraint on the bond market development. The challenge for swap markets versus the underlying is in basis differences between how some bond markets are developing versus swaps markets.
In the UK it’s fairly well aligned on a market convention perspective, with a small difference on the lookback length. There’s more efficiency than in the US, where Sofr swap quotes are on a compounding basis, whereas Sofr bonds are on a daily averaging basis.
We need to see a bit more development or alignments in the US between bonds and swaps. That would feed into the evolution of the cross-currency swap market.
GlobalCapital: Is the swaps market sufficient to meet both legs when lending Libor but issuing Sonia?
Leroy, EIB: It depends on the size. The swap market is developing and is going in the right direction. Participants realise there’s no way out, so Sonia will catch up and the market will develop. Ultimately, for really large sizes, you’ll be able to find the swap.
Mazzaro, TD Securities: The most liquid basis market before Sonia/Libor came into effect was the 6v3 Libor basis swap. Most now say that Sonia/Libor basis is as liquid as the 6v3 basis market. There are a few more natural flows going through the Sonia/Libor basis market due to UK bank ringfencing operations. Our trading desk says the main difference with Sonia/Libor basis is that it can be more volatile, which is more of a consideration for issuers that have to move proceeds from Sonia. It’s all about size and how you manage that, but it’s quite impressive that Sonia/Libor basis has caught up in liquidity terms.
GlobalCapital: How is the Sofr swaps market progressing?
Sinclair, World Bank: It is definitely still in its infancy. There is a need for a basis so issuers can sell Sofr bonds but without needing a Sofr exposure if it doesn’t fit naturally.
Finding Sofr-linked assets and the lack of Sofr-linked derivatives is a limiting factor. There are some available and deepening the market will be good for the cash products and the derivatives products that go along too. Sofr has only been around since April. It’s continuing to develop and I’d expect it to continue doing so as people really start digging in and realising that this alternative rate needs to be established.
Swap markets for Sofr and Sonia will develop as there’s a need to transform and offset risks. That will continue as more cash products go into the market and participants need to use the new risk-free rates. It will be evolutionary and somewhat depend on the work in building liquidity.
There really is a need for longer dated cash products to spur some of the need to go long to take the futures market deeper and longer.
Mazzaro, TD Securities: I totally agree with Don. Longer dated cash product is a necessity. If we look at the transition plan for Libor published by the ARRC specifically around Sofr as the alternative rate, we’re ahead of schedule. From the perspective of LCH and CME, both exchanges are now set up to value these new Sofr-based transactions. We also have one and three month Sofr futures, which are becoming more liquid on a daily basis, while interbank trades continue to go through the system via LCH and CME which are of value in terms of creating a credible Sofr/Libor curve.
The difficulty is how to get the market trading not only 18 month or two year tenors, but seven to 10 year tenors. The derivative market may disagree, but it seems the cash market has to drive the evolution of the swap and futures market forward from here.
Sinclair, World Bank: ARRC has realised that most of the work it has done in terms of consultations over the last few months has focused on cash products. Next, it will be about dealing with legacy issues and contract language for trades.
Bates, Fannie Mae: We did not swap our first two Sofr floaters to anything, although some issuers have swapped Sofr into Libor. But we have had Sofr futures and derivatives approved internally.
We swapped a portion of a benchmark note in November to Sofr using derivatives, which I think is a first for any issuer. We generally don’t talk about our hedging strategies, but felt it was important to be transparent for this transaction, to further help the derivatives market. It was very successful and we hope others start using Sofr derivatives as part of their hedging strategies.
But even though derivative activity has been increasing, a lot more transactions need to go through to support a more liquid structure.
GlobalCapital: What work has to be done to create a term structure for the new RFRs? Can we rely solely on the short term market providing the basis for the longer term market, or does a longer term market require a term structure, which will need more liquidity and depth of the futures and swaps market for the RFRs? Are we in a ‘chicken and egg’ situation where that liquidity and depth can only come when more longer term bonds are issued linked to those RFRs?
Bates, Fannie Mae: To increase liquidity we need more transactions referencing these new rates and see these deals move out the curve. We can’t solely rely on the short term market as the basis for the long term market.
Some market participants have said that if more loan originations linked to Sofr that could help develop the term structure. The more Sofr on bank balance sheets, the more important for them to be able to hedge. But it’s complicated because we’re dealing with consumers and Sofr is an overnight rate, that doesn’t yet easily lend itself to a term-rate product. A lot more work needs to be done, but part of that is just transactions going through the market.
Sinclair, World Bank: It’s the same answer for Sofr and Sonia. The tasks laid out by the working groups need to continue for there to be a complete term structure. As Gregg said, we need longer dated cash products spur that along.
The market needs to agree on various methodologies when it comes to the term rates, for instance.
The working groups initially underappreciated the need for a term rate for banks. Maybe for some asset managers it’s not a big deal, but it is for institutions like World Bank that rely on setting in advance our lending products. We’re not atypical of those dealing with big cash products for this. There has been some commitment on term rates from the working groups, but we need more.
We also have to find the right methodology for coming up with a term rate. Is it the OIS market or the futures market? That’s a very much an open question.
I don’t have an immediate opinion as to which is the right way — I just need a decision. There’s pluses and minuses to all these approaches. There isn’t the liquidity in the futures market to fully support it and the OIS market seems to be more tailor-made for the purpose. But in some ways using the OIS market isn’t the way end users are set up to think about things. But that’s part of the process too — to see where there really is demand and where liquidity grows.
The evolution we’ve had so far just needs to continue. The regulatory agents say the pace needs to be faster. To get to their internal deadlines on this, they need to keep pushing market participants to adopt these new rates.
One of the beauties of Libor was the term rate structure and being able to come up with term rates. It made it infinitely easier in terms of valuation and so on. But obviously Libor had other problems.
Green, Lloyds: Does the bond market desperately need a term Sonia rate? Probably not. There’d also be a question whether the swap market would provide any liquidity to a term reference rate. The loan market would probably answer slightly differently. I’ve heard on the loans side the term rate is seen as more important.
In the UK there has been a consultation on term rates and the expectation is a term rate will be published sometime in late 2019, subject to the feedback of that consultation. But given the swap market is on a compounded basis with a lookback, there’s less of an imperative for bond markets to move to term rates. It would probably make calculations of accrued interest for trading and tapping slightly easier if you knew the rate up front, but I don’t think any difficulties in having a lookback rate are insurmountable in the bond market.
Mazzaro, TD Securities: From a purely derivatives standpoint, the market has yet to get to grips with the idea of a term rate. At some stage looking ahead, we’ll have a very liquid Sonia futures curve; we’ve got the official reformed Sonia daily resets as well as a more liquid fixed Sonia swaps curve that may be more than enough for a derivatives desk moving forward. It goes without saying, but it’s not going to be a rate linked to a future that can be moved around by someone sitting on a trading floor.
Leroy, EIB: We might end up having a term structure, but we don’t need one. The term structure creates new problems. Where do the rates come from? Derivatives, which means dealers, which means a potential source of manipulation.
That doesn’t happen if you stick to the backward looking approach. It’s also neat as the structure is almost exactly the same for the bonds and derivatives, so you have one pool of liquidity for both.
Some people claim the loan market needs a term structure, but I disagree. The argument is that people taking a loan need to know in advance how much interest they will need to pay. That’s not possible with a backward structure because you’d only know the payment near the end of the period.
That doesn’t hold for me at all. If you have a loan in Sonia with an interest payment in three months, you compound the last known rate for three months and have a super-good approximation for what the interest payment is going to be. I don’t know anyone that needs to budget to the cent when it comes to loans. If the rate were to change by, say, 25bp two weeks into the period, then you can redo your calculation and still have 2.5 months to budget. If the increase happens at the end, it will have almost no impact. That’s the beauty of compounding.
Some argue that smaller issuers need more certainty. But these clients will have small loans so the impact will be even more limited. And they are mostly non-rated or poorly rated. If you pay Sonia flat and Sonia moves 25bp it has an impact, but if you pay Sonia plus 250bp you effectively have a fixed rate bond.
GlobalCapital: Are any other issuers worried about a term structure being a backdoor for manipulation?
Sinclair, World Bank: There are ways to be reasonable about it without being worried about it. The strength of approaching this through a futures market would be that it was transaction-based and we could come up with methodologies on how to fill gaps in it.
There might be more work to do if using the OIS market, but the distinction is that the market and regulators have made it very clear that the risk-free rate that everything is based off is the overnight Sofr rate. So if a term structure exists, that’s going to be the one used for discounting and talking about the risk-free rate.
GlobalCapital: How far along are issuers in identifying your exposures to Libor? How progressed are you in dealing with these exposures?
Gerli, FHLB: As we’ve discussed, we have developed operational capabilities to issue Sofr-linked securities and the FHLBanks are evaluating other Libor exposures and preparing transition plans. Several combined FHLBank exposures are available through our public disclosures. As of September 30, the variable rate advances were about $305bn, primarily indexed to Libor. Libor-linked debt was approximately $340bn with the vast majority maturing before 2021. In terms of notional derivatives linked to Libor, the figure is approximately $515bn. In addition, the FHLBanks hold certain floating rate MBS securities that use Libor as the coupon benchmark.
Bates, Fannie Mae: We have worked in 2018 to develop an enterprise-wide approach and as part of that we’ve done an impact analysis, so I feel pretty good as far as identifying the exposures across the company in terms of the products. We’ve looked at systems and models that are either directly or indirectly impacted and feel good that we’ve identified those.
The next step is to be able to see what changes we need to make based on those identifications. We’re doing as others are, having an impact assessment and exploring the timing going into 2019 and 2020, and what changes we need to make if Libor does go away in 2021.
Green, Lloyds: We’ll have more information on that as the responses come to the letters from the Bank of England, FCA and so on. But clearly there’s lots of work being done across the industry to look at back book exposures.
Gerli, FHLB: We are developing strategies to successfully manage and mitigate risks associated with the transition to the Sofr index. We are in the beginning phases of that process and with each passing quarter, based on market developments, member needs, and our ALM practices and system liquidity requirements in the short end, transitional goals and milestones will be developed and managed internally. Part of the process will be assisting our members in understanding market developments as they prepare and execute their transition plans. We expect those transition activities will lead to our need to source Sofr-based funding.
As the leading issuer of Libor-indexed debt, a significant provider of Libor-based liquidity to members, and a significant user of Libor-based derivatives for interest-rate risk management, the FHLBanks recognize our responsibility as market leaders and our opportunity to participate in a successful transition.
The FHLBanks are doing the work to assess legacy Libor language issues and have specifically updated our debt offering documentation to address future issuance of Libor-linked securities. As to the broader issues around legacy documentation issues, we’re active members in the ARRC and monitoring the development and evolution of fall-back language.
GlobalCapital: Has there been any progress on fall-back language for legacy Libor issues and for the new RFRs?
Leroy, EIB: I can’t speak for the legacy Libor issues, but for the new risk-free rates we decided to have fairly strict fall-back language. Typically, you have an expert that sets the rate if the reference rate doesn’t exist anymore — usually the paying agent or a rate set by a panel of banks. We decided to have no mention of that, and only mention the BoE. We didn’t want Libor to return through the backdoor.
Bates, Fannie Mae: We’ve done a lot of work since as far back as 2016, looking at all the documents and seeing the references to Libor and the fall-back language that exists. On a lot of our legal documents, Fannie Mae would select another comparable rate if the prevailing rate ceases to exist, but even that sort of flexible language could present challenges under certain scenarios.
We’re trying to stay in line with the guidance that’s been coming out, participate in the working groups associated with the ARRC and industry groups, and work with our regulator and conservator.
Mazzaro, TD Securities: The US money centre banks have been doing longer dated Libor FRN tranches this year. One of the challenges when looking at the fall-back language used was around who determines the alternative rate post-trigger event. Some banks used the calculation agent on their programme, while others were the calculation agent — and therefore the bank that determines the alternative rate to be used. There isn’t ‘a one template fits all’ approach.
Since then, the ARRC released a consultation paper to come up with a template for a draft fall-back language on new Libor FRNs. It revolves around three sections — trigger events; the successor rate waterfall; and the spread adjustment waterfall. The end date on that consultation has now passed, so next steps would be a final version that will pave the way for issuers to tackle fall-back language going forward. In the UK, the vast majority of longer dated issuance comes in fixed versus FRN format, so it’s hard to compare versus the US examples. The only recent longer dated FRN issuance has come from the high grade SSA space, where it feels there is less investor concern over fall-back language.
In terms of alternative rate fall-back language, in the UK, the EIB came up with a very strong template which leaned on the BoE. Since EIB, Lloyds, Santander UK, Yorkshire and Coventry BS all used a very similar fall-back template, with some minor tweaks.
In the US, we have just seen a break from recent convention. The generic template was a four step Sofr waterfall that leant on the prior day’s Sofr rate plus a rate recommended by the Fed. Freddie Mac, however, adopted the template to a simpler three step process where in the absence of Sofr and a rate recommended by the Fed, the new rate will be determined at the issuer’s sole discretion.