Forecast horizon
DCFHow many years a DCF forecasts explicitly before capitalising the rest, chosen by how long the business needs to reach a steady state.
Also written: explicit forecast period, forecast period length
Five years is a convention, not a rule, and the rule underneath it is that the explicit period must run until the final year can honestly be repeated forever. A company still growing twenty per cent with expanding margins in the last modelled year cannot be capitalised, whatever the convention says.
So the horizon is set by distance from steady state. A mature industrial may genuinely be there in five years. A business still scaling toward its addressable market, or one part way through a capital cycle, may need ten or more, and a cyclical needs enough years to span a full cycle rather than ending on a peak or a trough.
Extending is not free of cost, and it is worth being honest about which cost. It does not usually change the value, since the same cash flows get discounted either way. What it changes is how much of the answer has been written out and can be challenged, against how much sits inside a single capitalised assumption.
The counterweight is false precision. Years eight, nine and ten of a build are not more knowable than a terminal assumption, they merely look more rigorous because they have been broken into line items. A long horizon buys scrutiny, not accuracy, and it is worth having that distinction ready because interviewers press on it.
Worked example
A forecast ends in year five with revenue growing 15%, margins still expanding and capex at twice depreciation.
Capitalising that year assumes all three conditions persist forever. Extending to year ten with growth fading to 2%, margins settling and capex converging toward depreciation makes the terminal year capitalisable.
Enterprise value need not move much. What moves is the share of the answer resting on one assumption, and whether the final year describes a steady state at all.