Support does not give the parent's rating to the subsidiary: it closes a share of the default probability gap, and the share is a judgement.
A BB subsidiary has an illustrative 5.38% five-year default probability on its own and its BBB parent 1.24%. With a 70% probability that the parent steps in, the subsidiary’s probability is 1.24% + 0.30 × 4.14% = 2.48%, or 0.50% a year, in the BBB zone but not the parent’s own figure. At 40% it is 3.73%, and if the parent is also BB it is unchanged.
| Item | Value |
|---|---|
| Annual default probability, BBB | 0.25% |
| Annual default probability, BB | 1.10% |
| Annual default probability, B | 4.00% |
| Parent, rated BBB: five-year default probability | 1.24% |
| Subsidiary on its own, BB: five-year default probability | 5.38% |
| Probability that the parent steps in if the subsidiary is in distress | 70% |
The five-year figures come from 1 − (1 − p)5 on the annual rates. The 70% is a judgement about the parent’s willingness and legal ability to support, not a statistic.
If the parent supports the subsidiary, the subsidiary defaults only when the parent defaults too. If it does not, the subsidiary defaults at its own stand-alone rate. With a support probability s, the default probability of the subsidiary is a weighted average of the two.
PDsub with support = s × PDparent + (1 − s) × PDstand-alone
= PDparent + (1 − s) × (PDstand-alone − PDparent)
Excel: =Pp+(1-s)*(Ps-Pp)
The gap between the two is 5.38% − 1.24% = 4.14%. At s = 70% the subsidiary keeps 30% of it: 1.24% + 0.30 × 4.14% = 2.48%.
An annual rate is easier to compare with the scale: 1 − (1 − 2.48%)1/5 = 0.50% a year. That sits between BBB (0.25%) and BB (1.10%). Using the geometric midpoint of the two, 0.52%, as the dividing line, it is closer to BBB.
| Support probability | Five-year PD | Equivalent annual rate | Closest to |
|---|---|---|---|
| 0% | 5.38% | 1.10% | BB |
| 25% | 4.35% | 0.88% | BB |
| 50% | 3.31% | 0.67% | BB |
| 70% | 2.48% | 0.50% | BBB |
| 90% | 1.66% | 0.33% | BBB |
| 100% | 1.24% | 0.25% | BBB |
Each 10 points of support probability removes 0.41 points of five-year default probability, which is 10% of the 4.14% gap.
A parent that is only partly committed, with no guarantee and a minority stake, deserves a lower figure. At s = 40% the subsidiary’s five-year PD is 1.24% + 0.60 × 4.14% = 3.73%, or 0.76% a year, which is closer to BB (1.10%) than to BBB. The same subsidiary is treated as investment grade at 70% and as speculative grade at 40%; the rating depends on a number that cannot be observed.
If the parent falls to BB, its five-year PD rises to 5.38%, equal to the subsidiary’s stand-alone figure. The gap is zero, so 5.38% + 0.30 × 0 = 5.38% whatever the support probability. Support from a parent that is no stronger than the subsidiary adds nothing.
It closes a share s of the gap between the stand-alone and the parent default probability: Pp plus (1 minus s) times (Ps minus Pp). With s of 70% the example moves from 5.38% to 2.48% over five years.
It cannot be observed and the result depends on it directly: 70% gives 2.48% and 40% gives 3.73% in the same example, the difference between a BBB and a BB profile.
When the parent is no stronger than the subsidiary. If both have a 5.38% five-year default probability the gap is zero and the uplift is zero for any support probability.
The free companion files for The Credit Rating Analyst rebuild the Thornvale case, with the adjusted metrics year by year and the scorecard from leverage to the issuer and issue ratings. The support arithmetic in this article uses invented rates and a judgemental support probability. No account and no email address.
Not on Amazon: 36 paperbacks by Julian R. Sterling on how to read the documents, printed on demand and shipped worldwide. Browse the bookshop →
Reading guide: credit, private credit and distressed debt → · All 1908 articles →