Discounted cash flow, without the spreadsheet fear

A discounted cash flow model says something almost tautological: a business is worth all the cash it will ever produce, adjusted for the fact that cash arriving in ten years is worth less than cash arriving today. Everything difficult about a DCF is hidden inside the word "will".

Why a pound later is worth less than a pound now

Not because of inflation, though that contributes. Because of opportunity and risk. A pound today can be put to work immediately. A pound promised in ten years might never arrive. So we shrink future pounds to their present-day equivalent, and the rate at which we shrink them is called the discount rate.

At a 9% discount rate, £100 arriving in ten years is worth about £42 today. Push the rate to 12% and it drops to £32. That sensitivity is not a flaw in the model. It is the model telling you the truth: the further out the cash, the less you should be willing to pay for the promise of it.

The terminal value problem

Nobody can forecast a company ten years out, let alone forever. So a DCF forecasts explicitly for five or ten years, then bundles everything after that into a single "terminal value", usually by assuming the business grows at some modest rate in perpetuity.

There is also a mathematical landmine. The perpetuity formula divides by (discount rate minus growth rate). As the growth rate approaches the discount rate, the denominator approaches zero and the valuation approaches infinity. This is why a terminal growth rate above about 3% is almost always nonsense: you are implicitly claiming the company will eventually outgrow the entire world economy.

How to use a DCF honestly

• Run it backwards. Instead of asking "what is it worth?", ask "what growth rate does today's price already assume?" Then judge whether that is believable. This is the single most valuable use of a DCF.
• Always run three cases. Pessimistic, base, optimistic. A single number is false precision. A range is an argument.
• Watch the sensitivity. If a one-point change in the discount rate swings fair value by 40%, your model is not telling you much, and you should size the position accordingly.

A DCF is not a machine for producing the right answer. It is a machine for making your assumptions explicit enough that someone can disagree with them.