You can measure a company's cash flows from two altitudes, and everything in DCF-land follows from keeping them straight.
What is FCFF, and how do you calculate it?
Altitude 1 is the whole firm. Imagine you own the entire capital structure: every share, every bond, every loan. What cash does the business generate for all of you together, after paying for its own operations and its own growth but before anyone decides how to split it between lenders and owners? That's free cash flow to the firm (FCFF), the cash flow to all claimholders before financing.
Altitude 2 is the equity slice. Now you own only the shares. What cash is left for you after operations, after growth, and after the lenders are serviced (interest paid, debt repaid, new debt raised)? That's free cash flow to equity (FCFE).
Build FCFF from the income statement down, one line at a time, with the reason for every line:
FCFF = EBIT × (1 − t) … start from operating profit, tax it as if unlevered
+ D&A … add back the non-cash charges inside EBIT
− Capex … subtract cash actually spent on long-lived assets
− ΔWC … subtract the cash absorbed by growing working capitalAdvertisement
What does each line of the FCFF formula do?
Line 1 is EBIT × (1 − t), and not EBIT minus actual taxes. EBIT is profit from operations, before any financing effect, which is the pre-split altitude you want. So why tax it at the marginal rate t instead of subtracting the taxes the company actually paid? Because actual taxes are contaminated by financing. Interest is deductible, so a levered company pays less tax than the same business would unlevered. FCFF wants the operating cash flow as if all-equity financed, which is EBIT(1−t), the number you know from Phase 2 as NOPAT. The tax benefit of debt isn't thrown away. It's in the discount rate, inside WACC's Rd(1−t) leg. Put the shield in the cash flows and the rate and you've counted it twice, which is one of the classic errors in the common mistakes list. For the long run, use the marginal statutory rate: 25% for most Indian companies under the new regime, ~25% US federal+state. If the company's effective rate is temporarily lower because of tax holidays or loss carryforwards, you may start at the effective rate and drift to marginal by the terminal year. Write it down if you do.
Line 2 adds back D&A. Depreciation and amortization were subtracted inside EBIT, but no cash left the building this year. The cash left years ago, when the asset was bought, so add them back. Any other non-cash charge inside EBIT, such as impairments or write-offs, gets the same treatment. For one-offs, though, the better fix is to normalize the base year, which you'll do before any forecast.
Line 3 subtracts capex, meaning all investment in long-lived operating assets. That's the cash-flow statement's "purchase of PP&E," plus purchased intangibles, plus acquisitions, which analysts forget constantly. A company that "grows" by buying other companies is reinvesting as surely as one building factories. Average lumpy M&A spend over several years instead of taking one year's zero. Netting out asset-sale proceeds is standard for routine disposals.
Line 4 subtracts the increase in working capital, and specifically non-cash working capital. A growing business locks up cash on the way up. There's more inventory on the shelves and more receivables outstanding, partly offset by more payables owed. Here's the definition:
Non-cash WC = (current assets − cash & equivalents) − (current liabilities − short-term debt)
ΔWC = NCWC_end − NCWC_begin (an increase CONSUMES cash → subtract)Leave cash out. It's what you're measuring, and surplus cash is a non-operating asset that the equity bridge handles. Leave short-term debt out too. It's financing, and it belongs to the FCFE story. Phase 2's cash-conversion-cycle work told you why WC grows with revenue. Here you charge for it. And note the delicious exception. Businesses with negative working capital get paid to grow: FMCG distributors' advances, subscription deferred revenue, retailers who sell inventory before paying for it. For them ΔWC is a cash inflow. It's one of the quiet reasons the market pays premium multiples for them, as you'll see when you get to multiples.
FCFF deliberately ignores interest and debt raised or repaid, since both are financing. It ignores dividends and buybacks too, because they distribute free cash flow and don't determine it. And, worth saying out loud, FCFF isn't EBITDA. EBITDA is pre-tax and pre-reinvestment. Calling it "cash flow" is the most flattering lie in finance, and the cash-conversion work showed you companies whose EBITDA grew for a decade while FCFF never appeared. FCFF is what an owner of the whole firm could actually take out without impairing the growth you are forecasting. Watch that last clause. It's why reinvestment forecasting takes up so much of what follows.
FCFF worked example: an Indian cement leader
Here's the base year for the Indian cement leader from last week's WACC build. All figures are illustrative, modeled on ~FY24/FY25 scale, so pull the actual annual report. Revenue is ₹80,000 cr. EBIT is ₹12,000 cr, a 15.0% margin. Marginal tax is 25%. D&A is ₹4,000 cr, capex ₹6,000 cr. Non-cash WC rose from ₹3,600 cr to ₹4,000 cr:
NOPAT = 12,000 × 0.75 = ₹9,000 cr
+ D&A = +4,000
− Capex = −6,000
− ΔWC = −(4,000 − 3,600) = −400
FCFF = ₹6,600 crRead it like an analyst. The business earned ₹9,000 cr after tax from operations and plowed ₹2,400 cr net back into itself. That's ₹2,000 cr of capex beyond depreciation plus ₹400 cr into working capital, which leaves ₹6,600 cr for all its financiers. Two of those numbers will keep coming back once you start forecasting reinvestment. One is the reinvestment, capex − D&A + ΔWC = ₹2,400 cr. The other is the reinvestment rate, 2,400 / 9,000 = 27%. So learn the compact form now:
FCFF = NOPAT − Reinvestment where Reinvestment = capex − D&A + ΔWCSame number, two lenses. You use the long form to extract FCFF from filings and the compact form to forecast it.
Three routes to the same FCFF: and the classification traps
Real filings don't hand you EBIT-route inputs on one line, so analysts cross-check by computing FCFF more than one way. All three routes must land on the same number. When they don't, you've found either your own error or the company's classification choice.
Route 1, from EBIT (the workhorse): as above.
Route 2, from net income (undo the financing):
FCFF = Net income + Interest expense × (1 − t) + D&A − Capex − ΔWCNet income is struck below interest, so you add interest back. Only add it at (1−t), though, because the deduction saved real tax that net income already enjoys. Adding back pre-tax interest would smuggle the shield into the cash flow. Check it on the cement base year. NI = (12,000 − 1,000) × 0.75 = ₹8,250 cr, so FCFF = 8,250 + 1,000×0.75 + 4,000 − 6,000 − 400 = 8,250 + 750 − 2,400 = ₹6,600 cr. That's the same answer as before.
How do you get FCFF from CFO when interest is classified differently?
Route 3, from CFO (fastest from a filing, most traps): CFO already contains net income, D&A, and ΔWC. So:
FCFF = CFO + Interest × (1 − t) − Capex … IF interest paid is classified inside CFONow for the classification trap. It differs between India and the US, so learn this one cold. Under US GAAP, interest paid must go in CFO, so Route 3 works as written. Under IFRS and Ind AS, the company may classify interest paid in CFO or CFF, and most Indian non-financial companies put it in CFF. If interest paid is in CFF, the reported CFO is higher than its US-GAAP twin by the full interest amount. Blindly adding back interest(1−t) then double-counts. Memorize the correct adjustments as a pair:
Interest paid inside CFO (US GAAP; some IFRS): FCFF = CFO + interest × (1−t) − capex
Interest paid inside CFF (common Ind AS/IFRS): FCFF = CFO − interest × t − capexThe second line surprises people. Why subtract anything if interest never touched CFO? Because CFO still reflects taxes actually paid, and those are lower than all-equity taxes by the shield interest × t. FCFF wants unlevered taxes, so you strip the shield out. The first line follows the same logic. Its (1−t) add-back, in place of the full interest, leaves the shield out for the same reason. You can check where a company puts interest in ten seconds. Look at the CFF block of the cash-flow statement, or at the accounting-policies note. Worked example 1 and P3 in the practice set drill both cases until the reflex is automatic.
What other rules apply before you trust an FCFF figure?
Three more extraction rules apply before you trust any FCFF:
- Stock-based compensation is added back inside CFO as "non-cash," but it's real compensation, paid in ownership instead of cash. For valuation, treat SBC as a cash expense, and don't add it back. The alternative is to add it back and then fully count the dilution in shares outstanding and future option overhang. That's defensible, but it's harder to execute well, so the course standardizes on expensing. For a big US tech firm SBC can be 10 to 20% of revenue. "FCF" figures that add it back overstate owner cash flow by that much. Indian companies use ESOPs more modestly, but the rule is identical.
- Leases follow a consistency rule. Last week's WACC build counted lease liabilities as debt in the weights, so the cash flows have to be measured before lease financing. Under IFRS 16 / Ind AS 116 this happens naturally. Lease interest is in finance costs below EBIT, lease depreciation is inside D&A, and new ROU assets belong in capex. Under US GAAP (ASC 842), operating-lease rent is still one operating expense inside EBIT. So if you counted those lease liabilities as debt, add the imputed lease interest back to EBIT. The lease note discloses it. The numbers are usually small. If you ignore them, though, you'll be wrong in the same direction every time. One rule covers it: whatever you call debt in the denominator's weights must be financing in the numerator, and never an operating cost.
- The FCFF you compute is only as clean as the base year beneath it, and cleaning that base year is the first step of every forecast below.