What it is
A discounted cash flow (DCF) valuation is a method of estimating the intrinsic value of an asset by projecting its future cash flows and discounting them back to the present using a rate that reflects their risk. The theory is sound: the value of any asset is the present value of all future cash flows.
The problem is that in practice, DCF outputs are extraordinarily sensitive to small changes in the input assumptions — particularly the terminal growth rate and the discount rate. Changes that seem like rounding errors produce equity value swings of 30, 50, or 100%.
The sensitivity problem illustrated
Consider a simple DCF for a company generating $100 million in free cash flow today, growing at 3% per year forever, discounted at 10%. The terminal value formula gives:
Value = FCF / (r - g) = $100M / (0.10 - 0.03) = $1,429M
Now change the growth rate by half a percentage point:
At g = 3.5%: Value = $100M / 0.065 = $1,538M (+7.6%) At g = 2.5%: Value = $100M / 0.075 = $1,333M (-6.7%)
Change the discount rate instead:
At r = 9%: Value = $100M / 0.06 = $1,667M (+16.6%) At r = 11%: Value = $100M / 0.09 = $1,111M (-22.2%)
Combine both in the "optimistic" direction (lower discount rate, higher growth):
At r = 9%, g = 3.5%: Value = $100M / 0.055 = $1,818M (+27%)
Combine both in the "pessimistic" direction:
At r = 11%, g = 2.5%: Value = $100M / 0.085 = $1,176M (-18%)
The difference between the optimistic and pessimistic scenario — each using inputs that are individually defensible — is 55%. The model is not measuring value; it is translating assumptions into a number.
How sell-side analysts use DCFs
Investment bank analysts are frequently observed to work backward: they know (or suspect) what conclusion management wants, or they anchor to the current market price, and they adjust their inputs until the model produces the desired output. The model then provides the justification for a conclusion already reached.
This is not always dishonest — it is often unconscious. The analyst genuinely believes the company is worth the current price (because markets are assumed to be approximately correct), builds a model to support that view, and publishes a price target derived from the model. The DCF provides false precision: a specific number that implies a level of accuracy the underlying process cannot support.
The correct use of DCFs
The DCF is most valuable not for the point estimate it produces but for the explicit mapping of assumptions it requires. Building a DCF forces you to commit to specific views about revenue growth, margin progression, capital intensity, and risk. These views can then be stress-tested.
A sensitivity table — showing how equity value changes across a range of discount rates and growth rates — is more informative than any single point estimate. The range reveals how much the conclusion depends on small assumption changes. If a company is worth $800M in pessimistic scenarios and $2,000M in optimistic ones, the "right" price depends heavily on which scenarios are actually plausible.
The discipline is to use the model as a framework for structured uncertainty rather than a machine for producing false precision.
One thing most people get wrong
DCF assumptions are often treated as facts dressed in financial language. The WACC (weighted average cost of capital) used as the discount rate is itself an estimate built from other estimates: the risk-free rate (observable), the equity risk premium (estimated from historical data, with large standard errors), beta (estimated from historical regression, highly unstable), and capital structure weights (observable but changing). The discount rate used in most DCFs has a plausible range of 3-4 percentage points. Over that range, equity values typically move 30-60%. The output of the model cannot be more precise than the noisiest input — and the inputs are very noisy.