Reading bear, base and bull cases

Every valuation model is a chain of assumptions — growth, margin, discount rate, terminal growth — and every one of those assumptions could reasonably come in higher or lower than expected. Reporting a single "fair value" number buries that uncertainty rather than resolving it. A bear, base and bull range is an attempt to show the uncertainty instead of hiding it: three scenarios, not three predictions.

How the cases are built

The standard approach — and the one we use — moves growth, margin, the discount rate and terminal growth together, in the same direction, for each case. The bear case lowers growth, compresses the margin, raises the discount rate, and trims terminal growth, all at once. The bull case does the opposite. The shifts are deliberately moderate: a full percentage point on the discount rate alone would swamp every other assumption, so no single lever is allowed to dominate the story.

Moving every lever together, rather than varying one at a time, matters because real-world outcomes don't arrive one variable at a time either. A genuinely bad year for a company tends to mean slower growth and tighter margins and a higher cost of capital (since markets reprice risk together), not just one of those in isolation. A case that only moved the growth rate would understate how bad a real bear scenario actually looks.

What the range is telling you — and what it isn't

Bear and bull are stress cases, not probabilities. They're not a 10th-percentile and 90th-percentile outcome from some underlying statistical distribution; they're "what if the business's prospects were meaningfully worse" and "what if they were meaningfully better," each self-consistent. A wide gap between bear and bull means the fair value estimate is sensitive to the underlying assumptions — not necessarily that the company is a bad investment, just that confidence in any single point estimate should be lower.

A genuinely enormous spread — the bull case several multiples of the bear case — is itself a useful signal, but about the model's confidence, not the company's quality: it usually means at least one input (often the discount rate or the terminal growth assumption) is doing an outsized share of the work, which is exactly the kind of thing worth checking before trusting the headline number.