In July 2017, the U.K. Financial Conduct Authority (FCA) announced that the existence of the London Interbank Offered Rate (LIBOR) could not be guaranteed after 2021. Most European structured finance transactions across all sectors will likely be exposed to benchmark replacements if LIBOR and similar IBORs are phased-out, since notes are typically floating-rate, there are hedges for interest rate and currency risk, and maturities extend beyond 2021. As a result, the LIBOR phase-out may require changes to various aspects of transactions, potentially affecting asset yields, bond coupons, and swap payments, which could have cash flow implications. In S&P Global Ratings' view, structured finance issuers will be more sensitive than other issuers to these benchmark rate replacements. Structured finance transactions are typically established as special purpose entities with recourse limited to the assets backing the rated debt, with no extra support available post-closing, i.e., no additional liquidity facilities or reserves would be available to cover any expenses or cash flow leakages related to the benchmark replacement. Other issuers, such as financial institutions or corporates, might have other sources to cover these costs.

We have observed increased use of alternative benchmarks among structured finance transactions issued in late 2018 to avoid having to manage benchmark replacements if LIBOR is discontinued. Particularly in the U.K., some recent covered bond and residential mortgage-backed securities (RMBS) issuances have referenced the Sterling Overnight Index Average (SONIA). Based on feedback from market participants, we understand that other structured finance asset classes will likely follow those precedents going forward. This is consistent with the U.K. Risk-Free Rate Working Group's selection of SONIA as the preferred alternative to LIBOR. This Credit FAQ provides background on the differences between LIBOR and SONIA, and the implications the use of this alternative benchmark could have for U.K. structured finance transactions.

Frequently Asked Questions

How does SONIA differ from LIBOR?

LIBOR is a bank-to-bank lending rate based on expert judgment, and includes various risk premiums such as credit, liquidity, and other risks. Meanwhile, SONIA is a measure of the interest rate on short-term wholesale funds derived from transaction data where credit, liquidity, and other risks are minimal.

Another key difference is that LIBOR is a term rate, while SONIA is an overnight rate. This creates challenges for structured finance transactions, which generally have monthly or quarterly interest payment dates. In recent issuances, to get from the overnight rate to a term rate, the daily SONIA rates are compounded over the applicable interest period (commonly referred to as the 'compounded setting in arrears' rate). However, this means that the applicable interest rate would only be known at the end of the interest period, whereas with LIBOR these amounts are known at the start. To allow for an operationally feasible calculation of the applicable compounded SONIA rate, a lookback period, typically five days prior to the payment date, is applied to provide time for the transaction parties to calculate the applicable rates and determine the required distribution amounts. Compared to LIBOR, the daily compounding of SONIA reduces the rate's volatility, and the applicable interest rate for the period is therefore less sensitive to the prevailing rate on any given day.

There has been a close relationship between backward-looking compounded SONIA and forward looking LIBOR determined for the same period. However, since SONIA does not include the credit risk or term structure characteristics of LIBOR, the former has generally been lower, particularly during the global financial crisis in 2008-2009. There is also the limitation that the SONIA time series only goes back to 1997, and therefore does not include the more stressful interest rate environments from the 1980s and 1990s (see chart 1). Below, we have plotted the compounded backward-looking three-month SONIA rates at the end of the interest period and the corresponding forward-looking LIBOR rate determined at the beginning of the same period.

Chart 1

How will the market shift to SONIA from LIBOR?

To switch to SONIA based interest rates from LIBOR based interest rates, a mean/median spread adjustment approach is the preferred alternative, according to preliminary market feedback. This approach accounts for the systematic differences between LIBOR and SONIA (i.e., LIBOR rates would become SONIA plus the applicable spread adjustment). However, there is no clear consensus at this time on what the applicable length of the historical period should be to determine the spread adjustment if a benchmark replacement event occurs. The below table provides an overview of the rate differences between SONIA and three-, six-, and 12-month LIBOR based on the past five years, the past 10-years, and all historical SONIA data. Since SONIA does not include the various risk premiums reflected in LIBOR, the rate differential becomes larger the longer the LIBOR tenor is. Among these three alternatives, the five-year lookback period has the lowest rate differential, while the 10-year lookback period has the highest.

Table 1

How Do We Assess Basis Risk in EMEA RMBS and ABS transactions?

Transactions with floating-rate liabilities referencing a different index or resetting in different periods than the floating-rate assets contain basis risk. For example, in RMBS transactions with liabilities referencing three-month LIBOR and assets referencing the Bank of England Base Rate (BBR), we quantify the effect of the basis risk and apply it as an input to our cash flow modeling. To measure the basis risk in Europe, the Middle East, And Africa (EMEA) RMBS transactions, we use our basis risk percentile stresses model (see "Credit Rating Model: Basis Risk Percentile Stresses," published on Oct. 13, 2016, and "Methodology And Assumptions: Assessing Pools Of European Residential Loans," published on Aug. 4, 2017).

Our model uses historical data from two base interest rate indices to calculate a set of rating-dependent spread haircuts. These haircuts then reduce the income earned on the assets in the cash flow model. Firstly, we use the historical data for the paying index (i.e., the index on which the liabilities are paid) to calculate a daily moving maximum value for that index. We then use the historical data for the receiving index (i.e., the index on which the assets are earning) to calculate a daily moving minimum value for that index, considering resetting dates. We then arrange the daily difference between these two values in an ascending order, and select a predefined rating-dependent percentile value under our criteria as the basis risk haircut.

How could shifting to SONIA affect basis risk in structured finance transactions?

If the underlying rates for notes, assets, and swaps in structured finance transactions do not all change benchmarks at the same time and/or to the same rate, transactions could be affected by different basis risk and levels of excess spread compared to pre-replacement.

While we expect to see new RMBS and asset-backed securities (ABS) notes referencing compounded SONIA rates soon, we do not expect legacy retail credit contracts migrating to SONIA in the short term, as most loan documentation would require individual amendments. More recent contracts are more likely to be drafted in a way that would facilitate a smoother transition to an index other than LIBOR. In the current and still early stage of the transition, we expect to see transactions with SONIA only affecting the liability side in this sector.

Transactions with fixed-rate assets and LIBOR-based floating-rate notes, as is typical in the U.K. ABS market, usually have a fixed-to-floating hedge in place. As a result, basis risk is not a key credit factor in our analysis of those transactions.

In the U.K., retail mortgages are mainly referenced to BBR, Standard Variable Rates (SVRs), or LIBOR. SVRs are typically linked or correlated to BBR or LIBOR. While SONIA, the BBR, and LIBOR are positively correlated, our preliminary findings suggest less basis risk when SONIA is referenced in U.K. RMBS notes than when LIBOR is used. The overall economic impact on transactions remains uncertain as higher note spreads could offset this positive effect.

In the commercial credit sector, the migration to SONIA might arrive earlier than in the retail markets. Consequently, we envision a preliminary period where assets could transfer faster to SONIA than would the notes. Under this scenario, basis risk may be more stressful as it will capture the historical lower value of SONIA over LIBOR. While the basis risk would be higher, the expected higher spreads on assets could mitigate the overall economic impact on transactions. However, the exact magnitude of both effects is also uncertain at this time.

In the final stage of the migration to SONIA, we could see transactions in the market referencing SONIA on both the asset and liability side. Under this scenario, we expect to see minimal basis risk embedded in structured finance transactions since the compounded daily SONIA for different terms is based on the same overnight rates. For example, for assets earning one-month daily compounded SONIA and liabilities paying three-month daily compounded SONIA, both would generally be the same at the three-month point, assuming cash flows are reinvested at SONIA rates. Under this scenario, the basis risk would be limited by how both legs are executed in the market, which can be affected by reset day, lags, and other factors.

What are the most common benchmark replacement provisions in transaction documents?

Robust fallback provisions around benchmark replacement or discontinuation is a relatively new concept. We consider that strong provisions in transaction documentation ahead of LIBOR's discontinuation are an important step to limit the number of legacy structured finance transactions.

Legacy transactions   There is usually a fallback sequence of events that first references an electronic screen quote for LIBOR for the relevant payment date, and if that is unavailable, it references an average bank-level LIBOR quote. If a bank-level LIBOR quote cannot be obtained, then some documents mention using the previous period's rate or another overnight deposit rate with or without a margin spread. However, using the previous period's rate creates the risk that the rate could be perpetually locked for what is supposed to be floating-rate benchmark (i.e., the note becomes fixed for the remaining life of the transaction).

The documentation for some existing contracts never envisioned a LIBOR reference would not be available, and therefore have no provisions for an alternative should LIBOR disappear. Furthermore, switching to fixed-rate interest might not be feasible for various transaction participants. Modifying the fallbacks in existing transactions could equally be challenging. The wording usually requires a majority of lenders, borrowers, or a bondholder class to consent, which creates dispute risks, and an actual transition would create room for disputes and market disruption. Additionally, given that structured finance notes usually have multiclass structures, disputes are likely because it could be difficult to achieve a consensus in selecting a new benchmark. That said, this is not an exhaustive list of the language that could be present in the documentation, and there may be other options that we have not mentioned (see "With A LIBOR Phase-Out Likely After 2021, How Will Structured Finance Ratings Be Affected?," published on Oct. 19, 2017).

New issue   Most recent transactions are still referencing LIBOR, and many issuers have elected to use the Association for Financial Markets in Europe's (AFME) proposed Benchmark Rate Modification model or variations of it. While it is still evolving, the model strives to lessen the gravity of the risks indicated above. The wording we have seen provides the concept of an alternative benchmark (without specifying it) along with the how-to process for switching to it. At the same time, the model uses the idea of "negative consent", meaning noteholders are only being asked to actively participate if they disagree with the change. The wording also introduces a possible adjustment to the effective margin paid to noteholders.

Are there special analytical considerations for derivative contracts?

In September 2018, the International Swaps and Derivatives Association (ISDA) released a benchmarks supplement that can be included in ISDA-based derivative contracts as a framework for implementing the transition to the new benchmark rates. According to the supplement's terms, by switching to a new index the parties may elect to apply an adjustment payment in order to compensate for any transfer of economic value as a result of the benchmark rate replacement. Alternatively, the parties may decide that the calculation agent apply an adjustment spread on future periodic payments during the swap's life to compensate for the same. Therefore, transitioning to the new benchmark may expose the issuing special-purpose entity to the risk of making either a large single payment or a series of regular payments to the counterparty to reflect the spread differential. These payments would rank senior in the issuer's priority of payments, affecting the amount of funds available to the noteholders. Depending on the outgoing payments' size, they may affect our cash flow analysis and ultimately our ratings on the notes.

We would also analyze the opposite scenario, where the issuer is entitled to receive a large adjustment payment from the counterparty, to assess whether this cash would be trapped in the transaction and applied for future coupon payments. If the cash is released from the structure, it could constrain our ratings on the notes given the issuer may face a higher effective cost of funds (i.e. after accounting for swap receipts) if note margins have been increased following the benchmark replacement. Meanwhile, if the parties elect to apply the periodic adjustment spread, this additional cash would be available to the issuer over the swap's life, and would therefore not likely affect our ratings on the notes.

The ISDA benchmark supplement provides that if the parties do not agree on the new index, they can all exercise their right to terminate the swap contract. In this situation, a transaction risks losing the hedge, which could result in the issuer owing termination payments. This would have a negative ratings implication. That said, we currently believe that swap terminations due to benchmark replacements would be ratings-remote events and therefore do not include this in our rating analysis. Our current view is that the benchmarks will be replaced and transactions will continue to rely on the derivative contracts.

How would changing benchmarks in a transaction affect our cash flow analysis?

Market participants have indicated that a change in benchmark from LIBOR to an alternative rate should not result in large value transfers. Therefore, our current working assumption is that a change in the benchmark rate should be relatively neutral for noteholders such that they are in an economically equivalent position to current LIBOR benchmarks. Nevertheless, we expect that rating implications will vary on a case-by-case basis, and will be transaction-specific. They will depend not only on the timing and the nature of benchmark rate replacements for notes, assets, and swaps, but also on the magnitude of the spread adjustment and how operational challenges are addressed. The available credit enhancement and liquidity in transactions could also mitigate these risks.

Related Criteria

• Methodology And Assumptions: Assessing Pools Of European Residential Loans, Aug. 4, 2017
• Global Framework For Cash Flow Analysis Of Structured Finance Securities, Oct. 9, 2014
• Global Derivative Agreement Criteria, June 24, 2013

Related Research

• With A LIBOR Phase-Out Likely After 2021, How Will Structured Finance Ratings Be Affected?, Oct. 19, 2017
• Credit Rating Model: Basis Risk Percentile Stresses, Oct. 13, 2016