Valuation · August 3, 2026 · 8 min read

How a DCF actually works, step by step

"Walk me through a DCF" is the most reliably asked technical question in investment banking recruiting. It is asked because it is a single question that tests the three statements, cost of capital, and whether you understand what a valuation actually claims.

You should be able to do it in about ninety seconds. Here is the whole thing, in the order you should say it.

The one-sentence version

A discounted cash flow values a business as the present value of the cash it will produce for the rest of its life. You project cash flows for a forecast period, estimate everything after that as a terminal value, and discount both back at a rate that reflects their risk.

Step 1 — Project unlevered free cash flow

Usually five to ten years. The standard build:

EBIT
× (1 – tax rate) = NOPAT
+ Depreciation & amortization
– Capital expenditure
– Increase in net working capital
= Unlevered free cash flow

Unlevered means before any debt payments — cash available to every capital provider, lender and shareholder alike. You add back D&A because it was subtracted on the income statement but no cash left. You subtract capex and working capital because cash did leave, and the income statement did not show it.

The classic trap

Why tax-effect EBIT rather than use reported taxes? Because reported taxes are reduced by the interest deduction, and this is an unlevered figure — using them would smuggle the capital structure into a number that is supposed to be free of it. The benefit of that deduction belongs in the WACC, not here.

Step 2 — Pick a discount rate (WACC)

WACC = (E/V) × Cost of equity + (D/V) × Cost of debt × (1 – tax rate)

Because the cash flows are unlevered, they belong to everyone, so the rate must be a blend of what everyone requires. Cost of debt is tax-affected because interest is deductible — that is where the tax shield lives.

Cost of equity comes from CAPM:

Cost of equity = Risk-free rate + β × Equity risk premium

The risk-free rate is normally the 10-year Treasury yield. Beta usually comes from comparable companies — you unlever each comp's beta to strip out its capital structure, take a median, and relever it at your target's structure.

Step 3 — Terminal value

Two accepted methods, and you should name both.

Perpetuity growth (Gordon growth)

TV = Final year FCF × (1 + g) / (WACC – g)

g should be modest — typically between inflation and long-run GDP growth, call it 2–3%. Anything above long-run GDP growth implies the company eventually becomes the entire economy, which interviewers enjoy pointing out.

Exit multiple

TV = Final year EBITDA × Exit EV/EBITDA multiple

Usually set from where comparable companies trade today. More grounded in the market, but it imports the market's current mood into a long-run number.

Terminal value routinely accounts for 60–80% of total value. Say this out loud — it signals you know the forecast period is doing less work than it appears to.

Step 4 — Discount everything back

PV = Cash flow in year n / (1 + WACC)ⁿ

Sum the discounted forecast cash flows and the discounted terminal value. Use a mid-year convention if the model does, since cash arrives through the year rather than in a lump on 31 December.

Step 5 — Bridge to equity value

What you just built is enterprise value. To get to a share price:

Enterprise value
– Total debt
– Preferred stock
– Minority interest
+ Cash & equivalents
= Equity value
÷ Diluted shares outstanding = Implied share price

Diluted, not basic — options and convertibles become shares if the price rises, and treating them otherwise overstates the per-share result.

Step 6 — Sensitize it

No banker presents one number. You produce a table flexing WACC against terminal growth or exit multiple, because a DCF's output moves substantially on assumptions that are, honestly, estimates. Volunteering this is the difference between reciting a process and understanding one.

Follow-ups you should be ready for

"What happens to value if WACC rises?" It falls — you are dividing by a bigger number. "Which method gives a higher valuation?" It depends, but check whether the implied perpetuity growth from your exit multiple is sane; if it implies 6% forever, the multiple is too high. "When is a DCF inappropriate?" Early-stage companies with no positive cash flow, and banks, where interest is revenue and the unlevered concept breaks.

A DCF is the backbone question, and every part of it — WACC, beta, terminal value, the equity bridge — is its own follow-up. Our banking decks drill each of them separately.

Start studying →