What we’re forecasting, and why it isn’t profit

We have a discount rate. We need something to discount. That something is free cash flow to the firm (FCFF) — the cash a business generates that’s genuinely available to everyone who funded it, lenders and shareholders both, after paying for everything the business needs to keep running and growing.

Two words in there are doing real work.

Cash, not profit. Profit is an accounting opinion about which period a transaction belongs to. Depreciation reduces profit without any money leaving; a factory purchase drains the bank account without touching profit much at all. A discounted cash flow (DCF) model values cash because cash is what you can actually pay out. This blog’s post on OCF/PAT is the same argument in ratio form.

To the firm, not to shareholders. FCFF is measured before interest payments, which is why it pairs with WACC — the blended cost of all capital. The debt gets accounted for on the other side, when we subtract net debt at the end. Handle it in both places and you’ve charged for the borrowing twice.

The formula

FCFF = EBIT × (1 − tax rate)          ← operating profit, after notional tax
     + Depreciation & Amortisation    ← add back: reduced profit, cost no cash
     − Capital Expenditure            ← subtract: real cash, never hit profit
     − Increase in Net Working Capital ← subtract: cash tied up in the business

Line by line:

  • EBIT × (1 − t), sometimes called NOPAT — net operating profit after tax. Start from operating profit, before interest, and tax it as if the company had no debt at all. The tax benefit of the debt has already been handled inside WACC.
  • Add back depreciation. It was subtracted to get to EBIT, but no money moved. This is the classic reason EBITDA exists as a concept.
  • Subtract capex. Machines wear out and factories need building. Skip this and you’ve valued a business that never reinvests, which is a business that eventually stops existing.
  • Subtract the increase in net working capital (NWC). Growth swallows cash. Selling more means holding more inventory and waiting on more receivables, and that money is tied up until the business shrinks again. The cash conversion cycle post covers the mechanics.

Note that working capital enters as the change, not the level. A company with steady working capital consumes no incremental cash even if the balance is large; a fast-growing one bleeds cash into it every year.

Building the forecast

Forecasting isn’t guessing a revenue number for FY30. It’s choosing a small set of drivers and letting the arithmetic carry them forward. For Desi Bites, five drivers do the whole job:

Driver Assumption Reasoning
Revenue growth 18% falling to 10% by FY30 FY24 and FY25 both grew ~20%; growth rates fade as a base gets larger
EBITDA margin 17.2% rising to 17.5% FY25 was 17.0% after three years of steady expansion; assumes the trend flattens
Depreciation 4.5% of revenue FY25 actual: 115 / 2,592 = 4.4%
Capex 8% of revenue, easing to 5% The IPO raised money for expansion, so capex runs above FY25’s 3.1% before normalising
Working capital 8.68% of revenue FY25: (inventory 211 + receivables 199 − payables 185) / 2,592

The fading growth rate deserves a defence, since it’s the assumption people most often get wrong. Extrapolating 20% growth indefinitely produces absurdities quickly — a company growing 20% a year for 30 years becomes larger than its entire addressable market. High growth attracts competition, large bases are harder to grow, and the fade is the norm. Assume otherwise and you should be able to say why.

Also note what the capex assumption is doing. Desi Bites raised ₹1,600 lakh in its IPO explicitly to expand. Modelling flat capex while that cash sits on the balance sheet would give you the cash and the growth for free. If the money is being spent, the model has to spend it.

Worked example: Desi Bites Foods Ltd, FY26–FY30

All figures ₹ lakh, built off the FY25 actuals in the case study.

  FY26 FY27 FY28 FY29 FY30
Revenue growth 18.0% 16.0% 14.0% 12.0% 10.0%
Revenue 3,058.6 3,547.9 4,044.6 4,530 4,983
EBITDA margin 17.2% 17.4% 17.5% 17.5% 17.5%
EBITDA 526.1 617.3 707.8 792.7 872.0
Less: Depreciation 137.6 159.7 182.0 203.8 224.2
EBIT 388.4 457.7 525.8 588.9 647.8
NOPAT (EBIT × 0.75) 291.3 343.3 394.4 441.7 485.8
Add: Depreciation 137.6 159.7 182.0 203.8 224.2
Less: Capex 244.7 283.8 242.7 226.5 249.1
Less: Increase in NWC 40.5 42.5 43.1 42.1 39.3
FCFF 143.8 176.6 290.6 376.9 421.6

Now read the FCFF row, because it tells a story the revenue row hides. Revenue climbs smoothly every single year. Free cash flow does not — it sits at ₹143.8 lakh in FY26 and ₹176.6 lakh in FY27, then jumps to ₹290.6 lakh in FY28 and keeps climbing.

Operating profit kept growing in FY28, but most of the jump is capex. Intensity dropped from 8% to 6% as the expansion programme wound down; at FY27’s 8%, FY28 capex would have been ₹81 lakh higher. The business was generating plenty of operating cash in FY26 and FY27; it was spending it on factories.

This is worth dwelling on, because the same effect appears in real accounts. The free cash flow post noted that Britannia’s FCF rose between FY24 and FY25 mainly because capex fell from 3.3% to 2.1% of revenue, not because operations got better. Weak free cash flow during a build-out phase is not the same thing as a weak business — and strong free cash flow from a capex holiday is not the same thing as a strong one.

Doing it in Python

The whole forecast, in a form you can change one number in and re-run:

FY25_REVENUE = 2592.0
TAX, DEP_PCT, NWC_PCT = 0.25, 0.045, 0.0868

growth = [0.18, 0.16, 0.14, 0.12, 0.10]
margin = [0.172, 0.174, 0.175, 0.175, 0.175]
capex_pct = [0.08, 0.08, 0.06, 0.05, 0.05]

revenue = FY25_REVENUE
for year, (g, m, cx) in enumerate(zip(growth, margin, capex_pct), start=2026):
    prev, revenue = revenue, revenue * (1 + g)
    dep = revenue * DEP_PCT
    ebit = revenue * m - dep
    fcff = (ebit * (1 - TAX) + dep
            - revenue * cx
            - (revenue - prev) * NWC_PCT)
    print(f"FY{year}  revenue {revenue:7.1f}  EBIT {ebit:6.1f}  FCFF {fcff:6.1f}")

Run it and you get the FCFF row above. Change growth to a flat 25% and watch what happens to the answer — that’s the exercise the sensitivity post later in this series is built on.

Common mistakes

  • Subtracting interest. The single most common FCFF error. Interest is excluded here by design, because WACC already accounts for the cost of debt. Subtract it as well and you’ve penalised the company twice for the same borrowing.
  • Ignoring working capital entirely. It’s the least visible line and it can be the difference between a business that funds its own growth and one that needs a rights issue every three years.
  • Forecasting capex below depreciation forever. That’s a company whose asset base is shrinking. Fine for a few years of a capex holiday, incoherent as a permanent assumption — over the long run, maintenance capex and depreciation should converge.
  • Building a ten-year forecast because it looks more thorough. Nobody can forecast year eight of an Indian mid-cap. Longer forecasts don’t add accuracy, they add false precision — and they quietly shift value out of the explicit forecast and into the terminal value, where it’s harder to scrutinise. Five years is the usual compromise.
  • Assuming margins expand indefinitely. Every percentage point of margin expansion needs a reason — pricing power, scale, a better mix. “It went up last year” isn’t one.
  • Forecasting backwards from the answer. If you know what the share price is and you nudge growth rates until the model agrees with it, you haven’t valued anything. You’ve decorated a number you already had.

Takeaway: Free cash flow to the firm is operating profit after notional tax, plus depreciation, minus capex and the cash growth swallows into working capital — the money genuinely available to everyone who funded the business. Forecast it from a handful of defensible drivers rather than a revenue hunch, and expect the cash flow line to be lumpier than the revenue line, because reinvestment is lumpy.