The Swiss Solvency Test,
built from the ground up
Six lessons that take you from “why does a regulator ask for capital at all?” to computing a reinsurer’s one-year risk capital yourself, under FINMA’s StandRe standard model. Every formula is derived, not asserted, and every lesson ends in three drills you work with a calculator.
Who this is for
A reinsurance or insurance professional who has never run an SST. You know what a reserve is, what a treaty is, and roughly what a distribution is. You do not need measure theory, and you do not need to have met expected shortfall, Bornhuetter-Ferguson or a generalized Pareto distribution — each is explained where it first appears, in brackets, the first time.
If you have run an SST and want the model specification, read FINMA's technical description of StandRe instead. This course explains why that document says what it says.
What you will be able to do
- Explain what the SST measures — and why it is a one-year change, not an ultimate loss.
- Read the SST ratio and say what a supervisor will do about it.
- Place any piece of a reinsurer's business into one of StandRe's five components, and say why the boundaries are drawn where they are.
- Parameterise the attritional components and aggregate them with the variance-covariance formula.
- Turn a set of scenarios into a frequency-severity model, and calibrate it.
- Compute the one-year risk capital and the market value margin, and reconcile them.
The six lessons
| Lesson | What it answers | |
|---|---|---|
| 1 | Why the SST exists | What is a regulator actually protecting, and what is the SST ratio? |
| 2 | The one-year change | What quantity is the capital computed on — and why expected shortfall? |
| 3 | Segmentation | Five different ways to cut the same book, and what each is for. |
| 4 | Attritional events | AER and AEP: log-normals, correlations, discounting. |
| 5 | Individual events | IE1 and IE2: scenarios, exceedance curves, Poisson-Pareto. |
| 6 | Aggregation and the MVM | Putting it together, and the cost of holding capital for decades. |
Every lesson ships three drills — 🟢 easy, 🟡 medium, 🔴 hard — each with a worked solution beside it.
See it running
The model this course teaches is implemented as a live MeshWeaver plugin: the SST plugin computes StandRe deterministically over a reinsurance data model, and the reinsurance demo files a worked SST for a demo reinsurer. The numbers in Lesson 6 are that filing's.