Discounted Cash Flow: What a Valuation Model Can and Cannot Tell You

Discounted cash flow is the standard method for valuing a business from first principles. It underlies most equity research, most acquisition pricing, and most of what a financial analyst is actually paid to do.

It is also the model most often misused, and the misuse follows directly from how it is taught: as a calculation, when it is really a structured argument.

The idea

A business is worth the cash it will produce for its owners, adjusted for the fact that cash arriving later is worth less than cash arriving now.

That adjustment is the discounting. Money you receive in ten years is worth less than money today, because today's money can be put to work in the meantime and because the future is uncertain. A DCF makes that trade-off explicit rather than leaving it to instinct.

Three components, in order:

1. Forecast the cash flows

Not profit — free cash flow, the cash left after the business has paid its costs, its taxes, and the capital spending needed to keep running.

Profit and cash diverge, sometimes for years. A company can report healthy profits while consuming cash, because profit includes non-cash items and excludes the capital spending that keeps the lights on. The distinction is the entire reason cash flow, not earnings, sits at the centre of the model.

You forecast these year by year, typically over a period long enough for the business to reach something like a steady state.

2. Choose a discount rate

The rate reflects what a supplier of capital requires to fund this business rather than something else — driven by the risk of the cash flows and the return available elsewhere.

Higher risk, higher required return, higher discount rate, lower present value. That relationship is the model's core intuition, and it is why the same forecast can be worth wildly different amounts to different buyers.

3. Add a terminal value

Your forecast stops, the business does not. Terminal value represents everything beyond the forecast horizon, usually by assuming a modest perpetual growth rate or applying an exit multiple.

Here is the fact that governs the honest use of a DCF: terminal value routinely accounts for the majority of the total. Most of the number you produce comes from the part you did not forecast, derived from an assumption about the indefinite future.

Anyone presenting a DCF without saying what proportion sits in terminal value is presenting a conclusion, not an analysis.

Why two honest analysts disagree wildly

Because the output is extremely sensitive to inputs that cannot be known.

Move the discount rate slightly and the valuation moves a lot. Move the perpetual growth assumption slightly and it moves more — the terminal value calculation divides by the gap between the discount rate and the growth rate, so as those two converge the result runs away.

That sensitivity is not a flaw to be engineered out. It is the truthful statement that the value of a business genuinely does depend enormously on things nobody knows. The model is not adding uncertainty; it is exposing it.

Which is why a DCF that produces a single number is being used wrongly. A DCF used well produces a range, and a clear statement of which assumption the range is most hostage to.

The failure mode: reverse-engineering

The most common misuse is starting from a desired answer.

You believe a company is undervalued. You build a DCF. It says the company is worth less than the market price. So you revisit the growth rate, or the margin assumption, or the discount rate — each adjustment individually defensible — until the model agrees with you.

The output then carries the authority of a model while containing only your original opinion. This happens constantly, including in professional research, and it is very difficult to detect from outside because every individual input looks reasonable.

The defence is procedural, not mathematical: write down the assumptions and the reasoning before you compute anything, and treat any subsequent change to them as requiring a new reason, not a new result.

The genuinely useful way to use one

Turn it around. Instead of asking "what is this worth", ask "what would have to be true for today's price to make sense?"

Take the market price as given and solve for the growth rate or margin it implies. Now you have something you can actually evaluate: is that implied growth plausible for this industry, this competitive position, this size of company?

This reframing removes most of the false precision. You are no longer claiming to know the value. You are testing whether the market's implied assumptions are believable — a question you may genuinely have an edge on.

Where DCF does not work

What this means if you are not a professional

Two honest takeaways for an ordinary investor.

First, understanding the mechanism inoculates you against a specific kind of persuasion. When research states a price target, you now know it rests on a discount rate and a terminal assumption that could have been otherwise, and you know to ask what they were.

Second, building one yourself is a genuinely good way to learn how a business works — you cannot forecast cash flow without understanding where the cash comes from. It is a poor way to decide what to buy. The gap between "I built a model" and "my model is right" is the whole of the profession.

For most people, the practical conclusion is the reason index investing exists: valuing individual companies well is difficult, competitive, full-time work. Our guide to ETFs versus individual stocks covers the alternative to doing it at all.

Nothing here is financial advice, and no valuation method identifies a good investment on its own. Capital is at risk.