Margin of safety = (Intrinsic value − Price) ÷ Intrinsic valueRun the DCF backwards: ask what growth today's price assumes, then ask honestly whether reality can deliver it.
Stress Testing (an Investment Thesis)
Deliberately breaking your own thesis on paper — before the market does it with your money.
What is stress testing a stock?
Stress testing an investment means deliberately pushing the key assumptions behind a valuation — growth, margins, the exit multiple — to pessimistic but plausible extremes, then watching what happens to intrinsic value and downside. Borrowed from bank regulation, it converts a single "fair value" into a range and an honest view of how much could go wrong.
- Build a base case, then a bad case where each key assumption moves against you at once.
- The gap between the bad case and today's price is your true margin of safety.
- The goal is to size adverse outcomes with plausible values, not assume doomsday.
Definition
Stress testing is the practice of pushing the key assumptions behind a valuation to pessimistic extremes and watching what happens to value and downside. Borrowed from bank regulation, it asks: if revenue, margins, or the multiple all moved against me at once, would this still be survivable?
It turns a single "fair value" estimate into a range, and replaces false precision with an honest view of how much could go wrong.
How to do it
Pick the three or four assumptions your thesis depends on most — usually growth rate, operating margin, and the exit multiple. Build a base case, then a "bad case" where each moves to a conservative-but-plausible level simultaneously. Compute intrinsic value in both. The gap between the bad case and today's price is your true margin of safety; if the bad case still sits near the price, the thesis is fragile.
Simple example
Your base case values a software company at $120 using 15% growth and a 30× exit multiple. Stress test: cut growth to 8% and the multiple to 18×. Value drops to $62. If the stock trades at $110, the bad case implies a ~45% loss — telling you the price leaves almost no room for disappointment.
Common misconception
"Stress testing is just being pessimistic." The goal is not to assume disaster, but to size it. A good stress test uses plausible adverse values, not doomsday ones, so you can decide whether the price already compensates you for realistic bad outcomes — and walk away calmly when it does not.
- 1Build a base case, then a bad case where each key assumption moves against you at once.
- 2The gap between the bad case and today's price is your true margin of safety.
- 3The goal is to size adverse outcomes with plausible values, not assume doomsday.
| Shortcut | Deliberately breaking your own thesis on paper — before the market does it with your money. |
|---|---|
| Disciplined workflow | Build a base case, then a bad case where each key assumption moves against you at once. → The gap between the bad case and today's price is your true margin of safety. → The goal is to size adverse outcomes with plausible values, not assume doomsday. |
Sharpen your investing with free email lessons
Practical guides on valuation, margin of safety, and reading businesses — straight to your inbox. No spam.
We respect your privacy. Unsubscribe anytime.
True story · The Nifty Fifty, 1972
When the price already assumed a miracle
In 1972, institutions convinced themselves that about fifty "one-decision" growth stocks could be bought at any price and never sold. Polaroid traded above 90 times earnings, McDonald's and Disney near 80, Avon around 65. Run the arithmetic backwards and those prices quietly assumed decades of flawless, ever-accelerating growth.
The 1973–74 bear market repriced the dream: the group fell 60–90%, and Polaroid eventually went bankrupt. The cruel detail is that many of the businesses kept growing exactly as promised — the investments still failed, because the starting price had already spent the future.
Run the DCF backwards: ask what growth today's price assumes, then ask honestly whether reality can deliver it.
90×Polaroid's P/E at the top
−60–90%drawdown in 1973–74
10+ yrsto recover, if ever
True story · US Air, 1989
A confession Buffett kept retelling
In 1989 Buffett put $358 million into US Air preferred stock — drawn by the security's attractive terms rather than the business itself. Deregulated airlines were a commodity industry: brutal capital needs, no pricing power, profits competed away the moment they appeared. Dividends were soon suspended and Berkshire wrote the position down by 75%.
Buffett turned the episode into a running confession, joking that he needed a hotline — "I'm Warren, and I'm an aeroholic." A lucky upswing in the cycle eventually let Berkshire exit whole, which he insisted made the decision no less wrong.
A good structure cannot fix a bad industry. The moat question comes before the price question.
$358Minvested in 1989
−75%written down
luckwhat saved the exit
True story · Cisco, March 2000
The company delivered. The price still failed.
In March 2000 Cisco briefly became the most valuable company on earth at roughly $550 billion — over 100 times earnings. Run the DCF backwards and the price required Cisco to grow flawlessly for decades and become a meaningful share of the entire economy. Analysts at the time called it conservative.
Here is the cruel part: the business largely delivered. Revenue and profits are several times larger today than in 2000. The stock still fell almost 90%, and did not reclaim its 2000 high for about 21 years — because the question was never whether Cisco would grow, but whether it could grow faster than a price that had already spent the future. It could not.
Reverse the DCF before you buy: ask what growth today's price already assumes. A great company and a great investment are only the same thing at the right price.
≈$550Bpeak value, March 2000
−≈89%the drawdown that followed
21 yrsto see the old high again
Related tools
Frequently asked questions
How is stress testing different from a regular DCF?
A regular DCF produces one value from one set of assumptions. Stress testing re-runs that DCF under deliberately adverse assumptions to reveal the downside range, so you judge the price against the bad case, not just the base case.
Which assumptions should I stress first?
Start with whichever assumption the value is most sensitive to — usually the long-term growth rate and the exit (terminal) multiple. A reverse DCF can quickly show which inputs the current price most depends on.
Stress-test every thesis before you commit capital:
Download Valuefocus