Published in London & New York
10 Queen Street Place, London
1345 Avenue of the Americas, New York
Creditflux is an
company
© Creditflux Ltd 2025. All rights reserved. Available by subscription only.
Analysis CLOs
SRTs aren’t exactly a new idea
by Oussama Nasr
In the first of a two-part look at regulatory arbitrage, we look at balance sheet synthetic CLOs — deals which were the forerunners to today’s significant risk transfer transactions
Barely a week goes by without the publication of some research report or media article exalting the virtues and ingenuity behind significant risk transfer (SRT) transactions.
The occasional wag points out that these types of transaction first became commonplace as far back as the late 1990s, under the name “balance sheet synthetic collateralised loan obligations” (SCLOs). But are SRTs truly a reincarnation of a financial instrument that saw the light of day more than 25 years ago? Or are they meaningfully different?
1: Bank XYZ’s USD 1bn loan portfolio transaction structure and pricing
Source: Deal offering circulars/information memorandums
In these two articles, we compare SRTs to SCLOs. We argue in summary that although the two are similar in many respects, SRTs reflect a significant evolution from SCLOs — one that has been driven partly by a wholesale repricing since the 2007-09 Global Financial Crisis; partly by the breathtaking growth of private credit funds with appetite for high risk/high yield tranches; and partly by regulatory developments that have brought about important tweaks to the typical structure of these transactions.
This part of the article concentrates on the credit rating and pricing aspects of SCLOs and their regulatory capital benefits. Part two will flesh out the core features of SRTs and compare them to SCLOs.
Source: Deal offering circulars/information memorandums
Structuring an SCLO
We base this description on a number of SCLOs implemented in the early-to-mid-2000s, when Basel I and Basel II were still (successively) in force. Since several of these transactions were issued publicly, greater information is available about them than would be the case today with private SRTs. In particular, portfolio characteristics and tranche credit ratings were often announced publicly, which is rarely the case now.
The initial steps in a typical SCLO are summarised in figure 1 (Bank XYZ’s USD 1bn loan portfolio transaction structure and pricing). All tenors are three years.
2: Bank XYZ’s USD 1bn loan portfolio quarterly cashflows in absence of credit events
Source: Deal offering circulars/information memorandums
The following steps occur during the closing of the transaction:
- Bank XYZ identifies a USD 1bn portfolio on its balance sheet consisting of 200 USD 5m investment-grade corporate loans.
- XYZ enters into a CDS contract with SPV (the “mezzanine CDS”), protecting it against USD 70m of losses on those loans, but only after the portfolio has incurred USD 10m of losses. XYZ retains this USD 10m “first-loss” risk. The annualised fee under this contract is 108 basis points and is paid quarterly.
The credit rating agencies gave the notes high ratings
- XYZ enters into another CDS contract with a double A-rated bank (the “senior CDS”), protecting it against USD 920m of losses on the loan portfolio, but only after the portfolio has incurred USD 80m of losses (the USD 10m retained by XYZ and the USD 70m assumed by the SPV). The annualised fee under this contract is 10bps and is paid quarterly.
- An SPV issues the four tranches of notes to a variety of noteholders with different risk-return appetites. The class A notes are the most senior, while the class D notes are the most subordinated. The SPV raises USD 70m. It invests in three-year US government agency securities that pay a coupon of three-month Libor flat.
- The CDS contracts define “credit event” to mean either a bankruptcy filing or a failure to honour debts.
3: Bank XYZ’s USD 1bn loan portfolio final cashflows in absence of credit events
Source: Author’s illustrative example
Cashflows in absence of credit events
Absent any credit events, the cashflows during the three-year life of the transaction are as shown in figure 2 (Bank XYZ’s USD 1bn loan portfolio quarterly cashflows in absence of credit events).
The following payments and cashflows occur on each quarterly payment date:
- XYZ pays USD 230,000 (USD 920m × 0.1% × 0.25) to AA-rated bank.
- XYZ pays USD 189,000 (USD 70m × 1.08% × 0.25) to the SPV. This is just sufficient to cover the spread over Libor of the SPV notes. (This means the annual fee under the two CDS contracts is USD 1,676,000, a number we revisit later).
- The SPV receives Libor coupon payments from the US government agency securities.
- The SPV pays the coupons on the notes it has issued.
- The SPV receives a USD 70m principal payment from the maturing US government agency securities.
- The SPV pays the USD 70m aggregate principal on the notes it has issued. The transaction then terminates.
4: Cash flows following credit events
Source: Author’s illustrative example
Cashflows following credit events
What happens if credit events occur is quite different and is summarised in figure 4 (Cashflows following credit events, above).
Following a credit event, the calculation agent seeks bids from five dealers for the loan that has defaulted. A predefined arithmetical process is used to determine a percentage recovery value (RV) and the loss attributed to this credit event equals USD 5m × (1 – RV). This is added to any previous losses.
Once aggregate losses exceed USD 10m, the SPV liquidates some of its government agency bonds and passes on the proceeds to XYZ in satisfaction of its obligations under the mezzanine CDS. This process is repeated for all subsequent credit events.
Simultaneously, the face value of the affected SPV notes is reduced by the amount of loss recovered by XYZ, starting with the class D notes and working all the way up until the class A notes are fully written down.
Once the SPV notes, the mezzanine CDS and the government agency bonds are exhausted, XYZ claims further losses from the AA-rated bank.
Risk assessment of SPV notes
The credit rating agencies clearly considered the A, B and C notes to be of excellent credit quality since they typically gave them ratings of triple A, double A and single A, respectively. The best way to understand the basis for these ratings is through tables 5 to 7 (below).
Table 5 calculates the portfolio’s aggregate expected loss given (i) the portfolio loans’ ratings, (ii) cumulative historical default rates for these ratings and (iii) a 40% recovery assumption. The note to this table reveals that the portfolio’s cumulative three-year expected loss is only 0.202% (which we will round to 0.2%) or USD 2m, given the USD 1bn portfolio size. This figure may appear minuscule, but a single credit event results in losses of USD 3m (see figure 1), which is 1.5 times the portfolio expected loss.
Table 6 calculates for each class of SPV notes (i) the aggregate dollar amount of subordination, (ii) subordination as a percentage of the portfolio, and (iii) subordination as a multiple of the portfolio’s expected loss.
For example, for the class C notes: (i) aggregate subordination is USD 25m (consisting of the USD 10m first-loss piece retained by XYZ and the USD 15m class D notes), (ii) this amounts to 2.5% of the total portfolio, and (iii) this 2.5% is 12.5 times larger than the portfolio expected loss of 0.2%. All this emphasises how badly the fortunes of the portfolio obligors must turn before the class C notes incur a loss. The same logic, but to an even greater extent, applies to the class A and B notes and AA-rated bank.
The savings continue in perpetuity
Finally, table 7 calculates for each layer in the structure the first credit event that triggers a loss on that layer. For example, the class C notes suffer a partial loss when the ninth credit event occurs. At this point 4.5% (nine of 200 credits) of the portfolio has defaulted, causing aggregate losses of 2.7% of the portfolio (4.5% × (1 – 40%)), which exceeds the 2.5% tranche subordination and thus triggers a partial loss.
Again, this emphasises how badly the fortunes of the portfolio obligors must turn before the class C notes incur a loss. And again, the same logic, but to an even greater extent, applies to the class A and B notes and AA-rated bank.
Note that the class D notes did not typically receive an investment-grade rating. And the first-loss layer was not even rated, since it incurs a loss as soon as a single credit event occurs. But with respect to the class C, B and A notes and the AA-rated bank layer, the risk of loss is minimal, justifying the stellar credit ratings on these layers and their narrow spreads over Libor at issuance.
Regulatory capital arbitrage
At the time of these transactions, the Basel-based risk weights for relevant risk positions were as shown in figure 8.
Before the SCLO was entered into, the portfolio capital requirement was USD 80m (USD 1 billion × 100% × 8%). After the transaction is completed, XYZ’s capital requirement can be broken down into three components: (i) the capital requirement for the senior CDS with the AA-rated bank, (ii) the capital requirement for the mezzanine CDS with the SPV, collateralised by US government agency securities, and (iii) the capital requirement for the retained first-loss layer. This brings the capital requirement to USD 24.72m, computed as shown in the box (below).
The three-year weighted-average default rate for the portfolio is 0.34%.
The three-year weighted-average loss (if recovery is 40%) is 0.34% × 60% = 0.20%.
Source: Deal offering circulars/information memorandums
Remember, total portfolio expected loss = 0.2%. Source: Deal offering circulars/information memorandums
Remember, each exposure is 0.5% of portfolio and recovery is 40%. For example, nine defaults is 4.5% of portfolio in default and loss of 2.7%.
Source: Deal offering circulars/information memorandums
*This risk weight is unique to first-loss securitisation tranches and gives rise to a 100% capital requirement when multiplied by the 8% minimum capital requirement. In essence, this reflects the regulators’ view that such tranches are so risky that only a 100% capital requirement is appropriate for them.
Source: Bank for International Settlements and US Federal Reserve
Assuming a risk-free rate of 5%, an equity market return of 12%, and a beta for XYZ of 1.00 (reasonable for a large US-based bank), we obtain a 12% cost of equity (again computed as shown in the box).
We can summarise the regulatory capital arbitrage of this transaction as follows: in exchange for the payment of USD 1.676m annually in CDS fees (which is tax-deductible in most jurisdictions), XYZ reaps annual savings of USD 6.63m (USD 55.28m × 12%) on its cost of equity capital. So the net annual saving is USD 4.96m. And these savings continue in perpetuity since equity has no redemption date. The present value of a perpetuity of USD 4.96m, discounted at a 5% rate, is USD 99.2m!
Risk weight calculations
($920m × 20% × 8%) + ($70m × zero × 8%) + ($10m × 1,250% × 8%) = $24.72m
This capital amount is down from the $80m mentioned above – a capital saving of $55.28m.
We now apply the capital asset pricing model (CAPM) to estimate XYZ’s cost of equity capital:
Cost of equity = risk-free rate + beta × (equity market return – risk-free rate)
