A worked example of the capital charge on committed excess spread, and the coupon cut an investor must give for it to pay.
Synthetic excess spread sits beneath the first-loss tranche, so the originator carries it as a securitisation position weighted at 1,250 per cent. On an illustrative 2,000 million portfolio, committing 0.40 per cent a year puts 8.0 million of excess spread in the structure, adds 100.0 million of risk-weighted assets and cuts the capital relief by 21.0 per cent, from 475.5 million to 375.5 million. Commit it for four years instead of one and the relief falls to 75.5 million.
Worked in full in Significant Risk Transfer by Julian R. Sterling, with every figure reproduced in a free workbook.See the book on Amazon →
In a synthetic securitisation the bank keeps the loans and buys protection on a tranche of their losses. Synthetic excess spread is an amount the bank promises, out of the portfolio's own income, to absorb losses before the investor's tranche is touched. The investor likes it because it is a cushion it does not pay for. The bank likes it because a thinner effective first loss buys a lower coupon.
The catch is capital. Since the 2021 amendments to the CRR (Regulation (EU) 2021/558, which inserted point (e) into Article 248(1)), the originator must treat contractually designated excess spread as a securitisation position and include it in exposure value. The rules set no special weight for it; it is weighted like any other position, and because it is the most junior one in the waterfall, attaching at zero and detaching far below KIRB, that weight is in practice the 1,250 per cent cap. At an 8 per cent minimum that is 100 per cent of the amount: every euro of excess spread is a euro of capital, and at a bank's actual operating ratio it is more. Article 248(1)(e) captures excess spread designated for the current year, any designated in earlier years that is still available to absorb losses, and any designated for future years, so a multi-year commitment counts in the exposure too. That single term, how many years are committed, decides whether the feature pays.
| Input | Value |
|---|---|
| Reference portfolio | 2,000 |
| Average risk weight before the trade | 50% |
| Retained first loss, 0% to 1.0% | 20 |
| Mezzanine sold, 1.0% to 8.5%, funded | 150 |
| Retained senior, 8.5% to 100% | 1,830 |
| Risk weight: first loss / senior | 1,250% / 15% |
| Capital held per unit of RWA | 10.0% |
| Cost of equity | 12% |
| Synthetic excess spread committed, per year | 0.40% |
| Investor coupon without / with excess spread | 9.0% / 8.0% |
Step 1: relief without excess spread. Before the trade the portfolio carries 2,000 × 50% = 1,000 million of RWA. After it, the bank holds the first loss at 1,250 per cent, 20 × 12.5 = 250.0, and the senior at 15 per cent, 1,830 × 15% = 274.5. The funded mezzanine is collateralised in cash and carries nothing. RWA after the trade are 524.5 million, so the relief is 1,000 − 524.5 = 475.5 million, or 47.55 million of capital at 10.0 per cent.
Step 2: the excess spread exposure. One year of excess spread at 0.40 per cent on 2,000 is 8.0 million. If the deal commits four years, the exposure is 32.0 million.
Step 3: its RWA. Multiply by 12.5. One year adds 100.0 million of RWA and four years add 400.0 million.
Net relief = RWA before − RWA after − SES × 12.5
One year: 1,000 − 524.5 − 8.0 × 12.5 = 375.5
Four years: 1,000 − 524.5 − 32.0 × 12.5 = 75.5
In Excel: =B1*B2-(B3*B4+B5*B6)-B7*B8*12.5, with B7 the annual SES amount and B8 the years committed.
Step 4: what the capital costs. At 10.0 per cent, one year of excess spread ties up 10.00 million of capital, 125 per cent of the excess spread itself. At a 12 per cent cost of equity that is 1.20 million a year. Four years tie up 40.00 million, costing 4.80 million a year.
Step 5: the coupon cut that pays for it. Spread the annual capital cost over the 150 million tranche. One year needs the investor to cut its coupon by at least 0.80 points; four years need 3.20 points.
| Line | No SES | SES, 1 year | SES, 4 years |
|---|---|---|---|
| SES exposure | 0 | 8.0 | 32.0 |
| RWA relief | 475.5 | 375.5 | 75.5 |
| Capital released | 47.55 | 37.55 | 7.55 |
| Protection premium at 9.0% / 8.0% | 13.50 | 12.00 | 12.00 |
| Cost of the SES capital | 0 | 1.20 | 4.80 |
| Net gain against no SES | 0 | 0.30 | −3.30 |
With one year committed on a use-it-or-lose-it basis, the investor's one-point coupon cut saves 1.50 million against a capital cost of 1.20 million, and the bank is 0.30 million a year ahead while keeping 79.0 per cent of its relief. Commit four years and the same coupon cut leaves the bank 3.30 million a year worse off, with only 15.9 per cent of the relief left. The transaction would still be called an SRT; it would no longer be doing much.
| SES a year | Years | SES exposure | Added RWA | Net relief | Break-even cut, pts |
|---|---|---|---|---|---|
| 0.20% | 1 | 4.0 | 50.0 | 425.5 | 0.40 |
| 0.20% | 4 | 16.0 | 200.0 | 275.5 | 1.60 |
| 0.40% | 1 | 8.0 | 100.0 | 375.5 | 0.80 |
| 0.40% | 2 | 16.0 | 200.0 | 275.5 | 1.60 |
| 0.40% | 4 | 32.0 | 400.0 | 75.5 | 3.20 |
| 0.60% | 1 | 12.0 | 150.0 | 325.5 | 1.20 |
| 0.60% | 4 | 48.0 | 600.0 | −124.5 | 4.80 |
The table is linear in the product of rate and years, which is the point: 0.20 per cent for four years costs exactly what 0.40 per cent for two years costs. At 0.60 per cent for four years the structure adds 124.5 million of RWA rather than releasing any. No coupon cut an investor would offer makes that worthwhile.
Pricing excess spread as if it were free because it is paid out of income the bank would earn anyway. The income is real, but the capital is the cost, and the capital scales with every year written into the contract. The second mistake is comparing the coupon cut with the expected loss the excess spread absorbs, rather than with the capital it consumes. A desk that asks only “how much loss does the cushion take?” will agree to multi-year commitments the treasury then has to fund at 1,250 per cent.
The rules on how committed excess spread is measured are detailed and have been refined by technical standards. The figures here are illustrative; check the treatment that applies to your transaction and supervisor before relying on any number.
Synthetic excess spread pays only when it is short and small: the investor's coupon cut must exceed the capital cost, which on these assumptions is 0.80 points for each 8.0 million committed. Model it as a fourth tranche, not as a side term. The free workbooks for this case carry the same portfolio, KIRB and tranche risk weights, and the arithmetic of the bank's ceiling is set out in the most a bank can pay for SRT protection. For the thickness of the cushion beneath the investor, see how thick the first-loss tranche should be.
Because synthetic excess spread absorbs losses before every tranche, it attaches at zero and detaches far below KIRB, so its risk weight is in practice the 1,250 per cent cap. At an 8 per cent minimum that equals 100 per cent of the amount. In the worked case, 8.0 million of excess spread adds 100.0 million of risk-weighted assets.
Yes, because only the current year is exposed. On a 2,000 million portfolio, 0.40 per cent use-it-or-lose-it is 8.0 million of exposure and cuts relief to 375.5 million. The same rate committed for four years is 32.0 million and leaves only 75.5 million of the original 475.5 million of relief.
Divide the annual cost of the capital it consumes by the tranche size. With 10.0 per cent capital and a 12 per cent cost of equity, 8.0 million of excess spread costs 1.20 million a year, so the investor must cut its coupon on a 150 million tranche by at least 0.80 points. Four committed years need 3.20 points.
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 1501 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.