The test runs on risk-weighted amounts, not notional, so the tranche a bank keeps matters more than how much of the mezzanine it keeps.
The mezzanine test is passed when the risk-weighted exposure of the mezzanine positions the originating bank keeps is no more than 50 per cent of the risk-weighted exposure of all mezzanine positions in the deal. It is measured in risk-weighted amounts, not notional, and that changes which tranche a bank can afford to keep. On an illustrative three-tranche synthetic deal, a bank that sells the riskier mezzanine and keeps the safer one holds 57.1 per cent of the mezzanine notional but only 19.5 per cent of its risk-weighted exposure, and passes; keep the riskier one instead and it holds 42.9 per cent of the notional and 80.5 per cent of the risk, and fails.
Worked in full in Significant Risk Transfer by Julian R. Sterling, with every figure reproduced in a free workbook.See the book on Amazon →
Article 245(2) of the EU Capital Requirements Regulation gives a synthetic securitisation two mechanical routes to significant risk transfer (Article 244(2) gives the same two for a traditional one). Where the structure has mezzanine positions, the test is the mezzanine test. Where it has none, it is the first-loss test, worked separately in how to pass the first-loss test. Passing either is necessary, not sufficient: the supervisor can still refuse recognition if the capital reduction is not justified by a commensurate transfer of credit risk.
As the CRR now defines it, a mezzanine securitisation position is one that ranks below the senior position and above the first-loss tranche and carries a risk weight lower than 1250 per cent and higher than 25 per cent. That excludes the senior tranche at the top and anything at the bottom weighted at 1250 per cent. Older versions of the text and some supervisory proposals frame the boundary differently, for example by reference to KIRB, so read the definition that applies to the transaction before relying on a classification. Under SEC-IRBA, any tranche detaching at or below KIRB is weighted 1250 per cent, so the mezzanine sits at or above KIRB, or straddles it.
| Input | Value |
|---|---|
| Reference portfolio | 2,500 |
| KIRB | 4.00% |
| Exposure-weighted LGD | 45% |
| Effective number of exposures, N | 150 |
| Tranche maturity, MT, years | 2.5 |
| Junior tranche, kept by the bank | 0 to 4.00% |
| Mezzanine B | 4.00 to 7.00% |
| Mezzanine C | 7.00 to 11.00% |
| Senior tranche, kept by the bank | 11.00 to 100% |
Protection is bought on the two mezzanine tranches, and the question is how much of each the bank may keep. Risk retention under the Securitisation Regulation is a separate rule and is not tested here.
Step 1: weight each tranche under SEC-IRBA. Both mezzanine tranches are non-senior, so the supervisory parameter is p = max(0.3; 0.11 + 2.61/N − 2.91 × KIRB + 0.68 × LGD + 0.07 × MT) = 0.492. The formula itself, with KSSFA and the 15 per cent floor, is set out in how to calculate the SEC-IRBA risk weight.
| Tranche | Size | Risk weight | Risk-weighted exposure | Mezzanine? |
|---|---|---|---|---|
| Junior, 0 to 4.00% | 100.0 | 1250% | 1,250.0 | No, 1250% |
| Mezzanine B, 4.00 to 7.00% | 75.0 | 641.4% | 481.1 | Yes |
| Mezzanine C, 7.00 to 11.00% | 100.0 | 116.4% | 116.4 | Yes |
| Senior, 11.00 to 100% | 2,225.0 | 15.0% | 333.8 | No, most senior |
| All mezzanine positions | 175.0 | 597.5 |
Step 2: the denominator. The risk-weighted exposure of all mezzanine positions is 481.1 + 116.4 = 597.5, so the bank may keep at most 298.7 of it.
Step 3: the numerator, for each way of placing the deal.
Retained share = (share of B kept × RWEAB + share of C kept × RWEAC) ÷ (RWEAB + RWEAC) ≤ 50%
Sell B, keep C: 116.4 ÷ 597.5 = 19.5%, pass
Sell C, keep B: 481.1 ÷ 597.5 = 80.5%, fail
Excel: =SUMPRODUCT(Kept,RWEA)/SUM(RWEA)<=50%
| Placement | Mezzanine notional kept | Share of notional | RWEA kept | Share of RWEA | Test |
|---|---|---|---|---|---|
| Sell B and C | 0.0 | 0.0% | 0.0 | 0.0% | Pass |
| Sell B, keep C | 100.0 | 57.1% | 116.4 | 19.5% | Pass |
| Sell C, keep B | 75.0 | 42.9% | 481.1 | 80.5% | Fail |
| Keep 50% of each | 87.5 | 50.0% | 298.7 | 50.0% | Pass, at the limit |
The bank can keep all of tranche C and still keep up to 37.9 per cent of tranche B, 28.4 of notional, before the test fails. Going the other way, it cannot keep all of B under any placement of C, because B alone is 80.5 per cent of the mezzanine risk. The economic logic is the right one: the test asks whether the bank has handed over most of the risk sitting between the first loss and the senior, and most of that risk sits in the lower tranche.
Keep the mezzanine running from 4.00 to 11.00 per cent, sell the lower tranche, keep the upper one, and move the point where B stops and C starts.
| B detaches at | B size | B weight | C size | C weight | Notional kept | RWEA kept | Test |
|---|---|---|---|---|---|---|---|
| 5.0% | 25 | 980% | 150 | 235% | 85.7% | 59.0% | Fail |
| 6.0% | 50 | 785% | 125 | 164% | 71.4% | 34.3% | Pass |
| 7.0% | 75 | 641% | 100 | 116% | 57.1% | 19.5% | Pass |
| 8.0% | 100 | 534% | 75 | 84% | 42.9% | 10.5% | Pass |
| 9.0% | 125 | 453% | 50 | 62% | 28.6% | 5.2% | Pass |
Selling a sliver of mezzanine directly above KIRB is not enough. With B only 25 thick, the bank keeps 85.7 per cent of the notional and 59.0 per cent of the risk, and fails. Thicken B to 50 and the bank passes with 71.4 per cent of the notional still on its book. Somewhere between those two splits the structure crosses the line, and a structurer should locate it before fixing the tranche sizes. A bank that wants to sell as little notional as possible will therefore push B as thin as the test allows; a supervisor looking at commensurate transfer will look hardest at exactly that structure.
The common mistake is to run the test on notional. In the base case the notional answer and the risk-weighted answer point in opposite directions: by notional, keeping B (42.9 per cent) looks safer than keeping C (57.1 per cent); by risk weight it is the reverse, 80.5 against 19.5. A second mistake is to count the 1250 per cent junior or the senior tranche as mezzanine. Add the junior's 1,250.0 to the denominator and almost any retention would appear to pass.
Compute each tranche's SEC-IRBA weight first, keep only the positions weighted between 25 and 1250 per cent that rank below the senior, and divide what the bank keeps by the total in risk-weighted terms. On Fernhill 2026-2 the bank sells 75.0 of mezzanine B, keeps the 100.0 of C, and passes at 19.5 per cent. The weights depend on KIRB, so a portfolio whose KIRB rises moves the weights, and with them the shares. The companion files include the free workbooks for this case, where the SEC-IRBA sheet weights any tranche you enter.
On risk-weighted exposure amounts. The bank's retained mezzanine RWEA must not exceed 50 per cent of the RWEA of all mezzanine positions. On an illustrative deal, keeping a 100.0 upper mezzanine weighted 116.4 per cent is 19.5 per cent of the mezzanine risk, although it is 57.1 per cent of the mezzanine notional.
No. A mezzanine position ranks below the senior position and above the first-loss tranche, with a risk weight lower than 1250 per cent and higher than 25 per cent, so the senior tranche and a junior weighted at 1250 per cent are both outside the test. Including an illustrative junior's 1,250.0 of RWEA in the denominator would make almost any retention look compliant.
When the structure has no mezzanine positions to measure, typically a two-tranche deal whose junior sits around KIRB. The bank then holds no more than 20 per cent of the first-loss tranche, and shows it exceeds expected loss by a substantial margin. A three-tranche deal like the illustrative one, with 175.0 of mezzanine, uses the mezzanine test.
This article is one calculation from Significant Risk Transfer. The book takes the same case from first principles to the decision, chapter by chapter, and every figure it prints is a live formula in the free companion workbooks.
Get the book on Amazon →Free companion files
Also on Amazon UK · Amazon Germany · Amazon France · Amazon Canada
Reading guide: credit, private credit and distressed debt → · All 1614 articles →
If this book helped, or didn’t, a few lines on Amazon are worth more than they look: they are what the next reader goes on. Write a review. The workbook stays free either way.