The terminal value formula
In a discounted cash flow (DCF) valuation model, we can’t forecast forever. Even an Excel spreadsheet has a limit to the number of columns it contains! The terminal value solves this problem by answering the question: “What’s the business worth at the end of the forecast period?” In the final year of an Excel financial model you will usually see a big lump of cash (the terminal value). What we are doing in the model is trying to work out what the company we are valuing is worth at the end of the forecast period, or what it might be able to be sold for.
Many analysts are used to seeing the following formula used to calculate terminal valuation in financial modelling: terminal value = (next year’s forecast cash flow) divided by (discount rate minus expected long term growth rate). But where does the formula itself come from?
The formula for terminal value is an application of an old valuation formula
The formula is an application of an old valuation methodology called “the dividend discount model” or the “Gordon growth model”, where a business is valued as a stream of its dividends. The Gordon growth model holds that a company’s valuation is the sum of that company’s discounted forecast dividend payments. For more detail see any good corporate finance textbook or Gordon, M. (1959) “Dividends, earnings and stock prices”, Review of Economics and Statistics, Vol. 41, pp. 99-105. This model pre-dates discounted cash flow valuation, and the capital asset pricing model on which DCF is based. What we are doing at the back end of our financial model is applying a very old methodology to determine the valuation of the company at the end of the cash flow forecast period.
Big sensitivities in financial modelling for valuation
Any analyst who has spent a little bit of time modelling will be able to tell you that big sensitivities on terminal value and hence valuation are the final year cash flow, the long term growth rate, and the discount rate. Making small changes to any of these can result in a very different terminal value. And terminal value can end up accounting for a large proportion of total valuation in a financial model.
So, to summarise, the terminal value formula used in DCF valuation modelling is an application of a very old (and otherwise now regarded as outdated) valuation methodology, predating the DCF methodology itself. We can put an awful lot of work into modelling intermediate cash flows as accurately as we can, and then try to be very precise about how we discount those cash flows in valuation. But when terminal values (where we are essentially just dividing one quite rough number by another rough number) account for a large proportion of overall valuation, perhaps we should be answering another question. Instead of discounted cash flows, should DCF stand for “Deceit by Computer Fraud”?