The decision that compounds

Everything in the first module was about measuring a business. This post is about the one decision management makes that changes what the measurements will say in five years: what to do with the cash the business throws off.

A company that generates free cash flow has six options. It can reinvest in the existing business, buy another one, pay down debt, pay a dividend, buy back its own shares, or sit on the cash. Every rupee goes to one of the six. Over a decade, the pattern of those choices — not the margin, not the growth rate — decides whether a good business made its shareholders rich or merely made its managers busy.

There is one test that ranks the six, and this blog already built it.

The test

Is the expected return on this use of cash  >  the cost of that cash (WACC)?

The WACC post worked out that Desi Bites’ capital costs about 13.91% a year. Any use of cash that earns less than that destroys value — even if it makes profit go up, even if it makes the company bigger. The ROCE post made the same point from the other side, with one adjustment: WACC is an after-tax cost, so compare it with an after-tax return, ROCE × (1 − tax rate). Every return in the table below is after tax for that reason. Capital allocation is that comparison, applied one decision at a time.

Desi Bites’ ₹1,600 lakh

In June 2025 Desi Bites raised ₹1,600 lakh in its IPO. Twelve months later, here is where it went and what each use is earning:

Use of cash Done in FY26 (₹ lakh) Still open (₹ lakh) Expected return, after tax Clears 13.91%?
Hold as cash / fixed deposits 1,031 — 4.9% — FD rate 6.5% pre-tax, ~4.9% post-tax No
Repay the term loan 60 340 8.2% — saves after-tax interest; only the scheduled repayment was made No
Organic capex (new line) 245 455 23.4% — if it earns the existing business’s ROCE (31.2% pre-tax), after tax Yes
Acquisition (Chatpata Foods) 600 — 2.1% — year-one, on H2 numbers annualised No
Dividend 112 — n/a — returns capital; the shareholder decides —
Buyback — — n/a — returns capital; earnings yield at the price paid —

“Done” is money actually put to that use in FY26 (for cash, the rise in the bank balance). “Still open” is how much more could have gone there: the term loan still outstanding, and the part of the prospectus’s capex budget not yet spent.

Actually deployed in FY26: ₹600 lakh on the acquisition and ₹245 lakh on capex; ₹1,031 lakh more sits in the bank than a year ago. Each row is worth a paragraph.

Cash. The safest choice and, judged by the test, the worst. Fixed deposits earn about 6.5% before tax, roughly 4.9% after — some 9 percentage points below the hurdle, every year the money sits there. Desi Bites ended FY26 with ₹1,311 lakh of cash, 41.5% of total assets. The visible cost is in the ROE: FY25’s 30.8% (34% on average equity) became 12.2% on closing equity, not because the business got worse but because the denominator filled up with idle money. Cash has a real option value — the ability to act when a competitor stumbles — but “we’re keeping our powder dry” is also the most common excuse for not having a plan.

Repaying debt. Earns exactly the after-tax interest saved, 8.2% here. Below the hurdle, so on the test it’s a poor use — for this company. For a company whose debt is expensive or whose lenders are nervous, retiring it is the best available return, and the peace of mind doesn’t show in the arithmetic.

Organic capex. The strongest case on paper: if a new production line earns what the existing business earns — a ROCE of 31.2%, or 23.4% after tax — it clears the hurdle by a mile. The catch is the if. The first line was built into demand the founders knew personally. The second line needs demand that doesn’t yet exist, in states the brand hasn’t reached. Incremental returns are usually lower than average returns, and the prospectus promised ₹700 lakh of this; ₹245 lakh was spent.

The acquisition. Judged on year one, 2.1%. Judged on management’s case — Chatpata’s revenue growing 15% with margins converging toward Desi Bites’ own — about 9.3% (₹552 lakh of revenue × 13.5% EBIT margin × (1 − 25% tax) ÷ ₹600 lakh). Still below the hurdle. The goodwill post explained why acquisitions look like this: the seller knows the business better than the buyer and sets the price accordingly, and the buyer justifies the premium with synergies that often fail to arrive in full. Studies of acquirer share-price returns generally find an average around zero or negative, with wide dispersion: some deals work very well, many don’t. It doesn’t mean never; it means the burden of proof sits with the deal.

Dividends. A dividend earns the shareholder whatever they do with it, so it can’t be scored on the company’s hurdle. Its logic is the reverse of the test: if the company can’t find uses that clear the hurdle, it should hand the money back. Desi Bites paying ₹112 lakh while holding ₹1,311 lakh idle is a small gesture in that direction. The dividend yield post covered what the shareholder sees; this is what the board sees.

Buybacks. Also a return of capital, with a twist: the company buys shares, so the return depends entirely on the price paid. Buying back stock at the IPO price of ₹640 — a P/E of about 26.5 on FY26 earnings per share of ₹24.12 — is accepting an earnings yield of 3.77% (1 ÷ 26.5), below the hurdle again. Buying back at a P/E of 10 is a 10% yield on money the company knows better than any other investment. Buybacks are excellent or terrible depending on one number, and companies tend to do them when the stock is expensive and cash is plentiful, which is the wrong end. (How a buyback is taxed in the shareholder’s hands has changed twice since 2024; the tax series covers it.)

🧒 Explain it like I'm 10 (optional — skip if this is already clear)

You’ve saved ₹1,000 from your lemonade stand. What do you do with it?

You could buy a second table and jug (more lemonade, if there are more customers). You could buy your friend’s stand (but he wants ₹600 for something that only earns ₹10 a summer). You could pay back the ₹200 you borrowed from your sister. You could give some to the people who lent you money at the start. Or leave it in a jar.

The jar is easy and earns nothing. Each of the others is a bet. The question for every rupee is the same: will this earn more than I’d get just by not doing anything silly with it? A company’s version of “not doing anything silly” is called the cost of capital, and it’s about 14% a year for Desi Bites.

How to read a company’s allocation record

You can’t see the decision being made, but you can see the pattern over five to ten years, in three places:

The cash flow statement. Investing activities show capex and acquisitions; financing shows debt repaid, dividends and buybacks. Add up ten years and you have the company’s revealed preferences.

Incremental ROCE. Take the change in operating profit over five years and divide by the change in capital employed. That’s roughly what the new money earned. A company with a 30% ROCE and a 9% incremental ROCE is living off an old plant and wasting new cash.

Goodwill and impairments. A balance sheet where goodwill grows every year is a serial acquirer. Impairments in the notes are the acquisitions that didn’t work, admitted late.

The best allocators are rare and look boring: they reinvest when returns are high, buy back when the stock is cheap, hand cash back when neither is true, and say so in plain language in the annual report. The worst look exciting — a deal a year, a rising cash pile that’s “strategic,” and a chairman’s letter about transformation.

Common mistakes

  • Judging a decision by whether profit went up. Buying a fixed deposit raises profit. So does buying a 5% business with 14% money. The test is return versus cost of capital, not more versus less.
  • Applauding cash on the balance sheet. Safety, yes. But a cash pile earning 5% inside a 14% company is a drag the ROE line will show, and the ROE post is where beginners misread it as the business weakening.
  • Treating dividends as generosity. A dividend is an admission that the company has nothing better to do with the money. Often that’s the right admission. It isn’t a gift.
  • Cheering buybacks without checking the price. A buyback at a high multiple transfers value from remaining shareholders to selling ones.
  • Taking acquisition synergies at face value. The seller priced them in. Judge the deal on the target’s margins two years later.
  • Confusing average ROCE with incremental ROCE. The average is history. The incremental figure is what your money, invested today, is actually earning.

Takeaway: Every rupee a company keeps goes to one of six uses — reinvest, acquire, repay, pay out, buy back, or hold — and one test ranks them: does the expected return beat the cost of capital? Idle cash and a pricey acquisition can both fail it quietly while profit still goes up. Capital allocation is the decision the ratios can’t see happening, and the one they’ll be reporting on for the next decade.