Ask what the price assumes.

The problem with fair values

The standard output of equity research is a number: this stock is worth so much. The trouble with that format is not that the numbers are wrong, though they usually are. It is that a fair value hides its inputs. Behind every price target sits a growth rate, a margin path, and a discount rate, and by the time they have been compressed into a single dollar figure, nobody can argue with any of them. You can only accept the number or dismiss it, and both responses teach you nothing.

There is a better question, and it inverts the machine. Instead of asking what the business is worth, ask what the current price already requires to be true. Solve for the assumption instead of the value. The output is not an opinion to defend. It is a claim the market is making, extracted and held up to the light, and claims can be checked.

The mechanics, in one worked example

A company carries an enterprise value of $1,000 million against $50 million of owner earnings, which is 20 times. Is that expensive? The multiple alone cannot say. So run the valuation backwards: at a 9% discount rate, with growth fading to 3% over a decade, solve for the starting growth rate that makes the discounted cash flows equal the price. The answer here is 6.7% a year.

Now the argument has a shape. If the business is growing at 4% with no particular reason to accelerate, the same machine says it is worth about $895 million and the price is ahead of the facts. If it is compounding at 12% with room to run, the machine says $1,245 million and the price is behind them. Same company, same multiple, opposite conclusions, and the entire debate now lives in one checkable question: is 6.7% a reasonable decade for this business? That is a question evidence can answer. Whether 20 times is too much was never one.

The method decided our last two calls, in opposite directions

This is not a framework we admire from a distance. It is the machine our recent research actually ran on, and it is worth showing that it pushes both ways, because a method that only ever says buy is a sales technique.

In our Uber initiation, the price implied roughly 8.5% annual growth in owner earnings, fading, at the very moment the company was compounding gross bookings at 22% with expanding margins. The market's embedded claim was pessimistic against observable evidence, and the call was BUY. We had, candidly, walked in leaning the other way, anchored on the multiple. The reverse framing is what changed the answer, which is the point: a good method should sometimes overrule its operator.

In our Progressive initiation, the same inversion ran on an insurer, solving for the combined ratio instead of a growth rate. The price implied underwriting margins normalizing to roughly 91 to 92, almost exactly where an honest mid-cycle estimate lands. The market's embedded claim was reasonable, so there was nothing to disagree with, and the call was HOLD. The method does not manufacture conviction. It locates whether there is anything to have conviction about.

Why this beats the forward version

Running valuation forward stacks your errors: you guess growth, you guess margins, you guess a discount rate, and the output inherits every guess while looking precise. Running it backwards spends its precision differently. The price is a fact. The arithmetic is mechanical. The single extracted assumption is where all the uncertainty is deliberately concentrated, sitting in plain view where it can be compared against base rates, guidance, and history. The forward version asks you to be right about several things at once. The backward version asks you to judge one claim someone else has already made.

It also disciplines the temptation every analyst knows: deciding the conclusion first and selecting assumptions to reach it. A model run forward will produce whatever fair value you quietly want. A model run backwards confronts you with what the market wants, before you have chosen anything.

The honest limits

The machine is only as honest as its fixed inputs. The discount rate you hold constant while solving is itself a judgment, and moving it moves the implied assumption, which is why our reports publish the sensitivity rather than a single figure. The fade shape and horizon are conventions. And an implied assumption that looks absurd is not an automatic trade: prices can embed absurd claims for years, and knowing the market is wrong says nothing about when it stops being wrong. The method tells you what you are being asked to believe. Whether to believe it, and what to do about it, remain your job. The free valuation workbook on this site includes a reverse mode for exactly this reason: the most useful tab is the one that asks the question backwards.