This is the most clarifying exercise you'll do with cash flows. Take ten years of cash flow statements, which is one afternoon with screener.in for an Indian company or the 10-K archive/stockanalysis.com for a US one, and build the decade's sources and uses of cash:
SOURCES USES
Cumulative CFO Capex (PP&E + intangibles)
+ Debt raised (net, if positive) + Acquisitions (net of cash acquired)
+ Equity issued + Dividends paid
+ Disposals / stake sales + Buybacks
+ Decrease in cash pile + Interest paid (if in CFF) + lease principal
+ Debt repaid (net, if negative)
+ Increase in cash pile
TOTAL SOURCES = TOTAL USES (small residual = FX/other — chase anything > 2–3%)The identity has to tie. It's the decade-length version of the "net change in cash" reconciliation you met with the cash flow statement, and the tie-out is your error check, just as it was in the two-sided invested-capital build. Then convert the uses into paise per rupee of CFO (cents per dollar), and read the company's revealed priorities. Its stated ones can wait.
Advertisement
What does a ten-year cash walk look like for TCS?
Start with India and TCS, FY16 to FY25. It's illustrative and modeled on the reported scale, so pull the actual annual reports. You've already read one year of this company's CFS, so now assemble the decade:
| Decade item | ≈ ₹ crore | Per ₹1 of CFO |
|---|---|---|
| Cumulative CFO | ~3,40,000 | ₹1.00 |
| Capex | ~30,000 | ₹0.09 |
| Acquisitions | ~5,000 | ₹0.01–0.02 |
| Dividends | ~2,00,000 | ₹0.59 |
| Buybacks (five tranches, ₹16–18,000 cr each) | ~83,000 | ₹0.24 |
| Net debt raised | ~nil | — |
| Cash/investments build + other residual | ~20,000 | ₹0.06 |
Reading: nine paise of every rupee stayed in the business, and revenue still roughly doubled, which is capital-lightness. Eighty-three paise went straight to owners. Nothing was borrowed to do it, and the share count fell ~8% (bonus-adjusted). Decade FCF/PAT ≈ 90%. This is what "the business needs almost nothing and management knows it" looks like in a table. It leaves one critical question, the compounder's standing follow-up: was returning 83% right, or were reinvestment options left unexplored?
What does the same walk show for Apple?
Now the US, and Apple, FY2015 to FY2024. Again it's illustrative and modeled on the 10-K scale, so pull the filings:
| Decade item | ≈ $bn | Per $1 of CFO |
|---|---|---|
| Cumulative CFO (source) | ~900 | $1.00 |
| Net debt raised over the decade (source) | ~+45 | — |
| Net cash-pile drawdown (source; from ~$140bn net cash toward ~$50bn) | ~+90 | — |
| Capex (use) | ~110 | $0.12 |
| Acquisitions (use) | ~10 | $0.01 |
| Dividends (use) | ~140 | $0.16 |
| Buybacks (use) | ~660 | $0.73 |
| Other uses: RSU net-share-settlement taxes (~$30–35bn), the 2017 Act's one-time repatriation tax, interest, FX | ~115 | $0.13 |
Tie-out: sources ≈ 900 + 45 + 90 = 1,035 ≈ uses ≈ 110 + 10 + 140 + 660 + 115. Now notice what chasing the residual taught you. The biggest piece of "other" is taxes paid settling employees' RSUs. That's stock-based compensation leaving the building dressed as cash, the SBC problem showing up live. Add a one-time tax bill for bringing the offshore pile home. The residual is never boring.
Reading: Apple returned roughly $800bn, about 89% of CFO, and in several years more than 100% of FCF. Two deliberate sources funded the excess. One was cheap bond issuance at 2 to 3% coupons, against an offshore cash pile that pre-2018 US tax law made expensive to repatriate. The other, later, was drawing that pile down. The share count fell ~33% (split-adjusted, from 23bn to ~15bn), so EPS grew meaningfully faster than net income all decade. This is benign financial engineering, with leverage arbitrage supplementing a torrential Tier-1 CFO and never standing in for it. Telling it apart from Personality 4's malignant variety is the judgment this skill builds. Decompose the CFO (Tier 1? yes). Check distributions against FCF plus stated funding (coherent? yes). Then ask whether the story survives without the engineering.
It does.
Which questions should the cash walk answer?
The walk's four verdict questions. Answer them in writing every time:
- What fraction of CFO went back in (capex + M&A), and what has it earned (ROIIC)?
- What fraction reached owners, and was it funded by operations or borrowings?
- Is the share count telling the same story as the buyback line, or is SBC leaking?
- Does the decade's revealed allocation match management's narrated strategy?
A gap between what the decade revealed and what management narrated is among the highest-signal governance reads you'll ever make, and it costs you one afternoon.