CSL (ASX: CSL) Follow-Up: The Probability Model Behind My Investment Thesis (Challenges Welcome)
This is the math behind my earlier CSL post and how I sized my position. I run this through a Bayesian split rather than a single-point forecast, and it comes out to a 69% probability of a full recovery case: plasma volumes keep growing, cost-per-litre pressure eases as the network expansion and efficiency programs land, ROIC climbs back toward 14–15% (still under the historical 20%+ peak, but strong for a healthcare oligopoly), earnings grow \~9–10% a year, and the stock delivers 7–10% annual returns from here. Against that sits a 31% probability of a slower recovery — plasma/donor cost pressure lingers longer than expected, ROIC settles at a lower-but-still-decent 11–13%, and returns come in closer to 5–6% a year. Neither branch of that tree is a capital-loss scenario; the "bear case" here isn't CSL breaking, it's CSL compounding more slowly than it used to.
Zooming into what actually drives the downside branch: a genuinely damaging donor/labour cost spiral only shows up in about 14% of my modeled scenarios (the worst-case combination of wage inflation and demand), regulatory tightening has a >50% chance of hitting the sector over the next decade but CSL has a 100-year unbroken track record navigating it (\~80% modeled chance of success this cycle too), and even a full write-off of the underperforming Vifor segment wouldn't be enough to seriously damage group financials — it would just be a bad year. Blend the two scenarios and you get a probability-weighted expected return of roughly 6.5–8.5%, versus \~3–5% for bonds, which on my risk/reward scoring comes out around a 3 (risk) against a 4 (reward) — a favourable asymmetry, not a coin flip. That's the actual reasoning behind the staged entry I laid out last time (40% now, 35% on a dip toward support, 25% on trend confirmation): it's not a conviction that CSL definitely recovers, it's that the odds and the payoff both lean my way even in the slower scenario. Not financial advice, just showing my work — DYOR.