Actionable insights — I'm Buying The Whole Business

1. Value a whole-business purchase by its recoup period, not by its multiple

1. Take the business's annual free cash flow and divide the purchase price by it. That number is the recoup period in years — the multiple restated as time.
2. Assume the business merely survives and grows with inflation. Do not model improvement.
3. From the recoup year onwards, treat the full annual cash flow as new investable capital, deployed at your normal market return.
4. Compare that path against putting the same equity into the market on day one and leaving it there.
5. Judge the deal on the gap between the two paths over your actual holding period, not on the entry multiple in isolation.

• The recoup period assumes the cash flow is genuinely distributable. Maintenance capex, working-capital growth and regulatory capital in an insurer are all claims on it before you see it.
• A recoup period only beats the market if the business survives the recoup years. The shorter the payback, the less that assumption matters — which is the whole reason to insist on a low multiple.

2. Run the deal twice — once on EBITDA, once on earnings — and let the gap tell you how much is leverage

1. Frame one: enterprise value = EBITDA × entry multiple. This is what the whole business costs including its debt, before interest and tax.
2. Frame two: price = net earnings × P/E. This is what the equity costs, after interest and tax.
3. Run the identical assumptions through both and record the 20-year value, the CAGR and the equity actually invested.
4. Compare the two answers. A wide gap is not noise — it is the size of the benefit your model is attributing to the capital structure and the tax shield.
5. Report both; publishing only the flattering framing is the tell.

• The real reason for the gap: EBITDA is measured before interest and tax, so a levered buyer counts the benefit of debt twice — once in the cheap entry multiple, once in the returns. The P/E framing is the more honest of the two.
• Headlining the EBITDA number, as this issue does ("almost 7 times as much"), while the earnings framing says 3.5x.

3. Test whether an insurance business gives you float you can actually invest

1. Establish the gap between premium collection and claim payment — the longer and more predictable, the more usable the float.
2. Check underwriting profitability. Float is only free if the insurance itself at least breaks even; an underwriting loss is the interest rate on that borrowing.
3. Size the float against the equity you are putting up. That ratio is your true leverage, and it does not appear on any multiple.
4. Ask what the float may be invested in — reserve rules, duration matching and regulator constraints decide whether "invest it in equities" is available at all.
5. Only then add the float's investment return to the deal's expected return.

• The conditional the article states once and then drops: "If you do well as an insurance company and you are profitable." Unprofitable float is expensive borrowing, and most insurers are not Berkshire.
• Life-insurance float is not P&C float. SNFCA's liabilities are long-dated and reserved against; the example silently borrows Berkshire's flexibility for a very different balance sheet.

4. Put the disaster in the model as an input, not as a caveat

1. Add three explicit fields: the year the bad thing happens, the multiplier applied to earnings/EBITDA that year, and the interest rate you refinance at afterwards.
2. Set the shock mid-hold, where debt is still outstanding and compounding has not yet done its work — that is the vulnerable window.
3. Refinance at a punitive rate from the shock year onward, not just for one year.
4. Re-run and check that the deal still clears the do-nothing alternative. If it only works shock-free, the leverage is too high.
5. Report the shocked figure as the headline, not the clean one.

• A single shock year. Real distress is consecutive bad years plus a covenant breach, not one halved year followed by a clean recovery.
• The debt strategy toggle set to "pay down first" — deleveraging into the shock is what saves this run; the "keep debt" mode would not be as forgiving.

5. Use the "cheap enough that a bad business still pays" test

1. Value the assets you would own outright — working capital, cash, receivables, inventories — per share.
2. Express the price you would pay as a fraction of that. Below ~0.5x, the asset base alone is a floor.
3. Then, separately, forecast nothing: sum what the operating business plausibly generates over the next decade even in decline.
4. Ask whether the second number alone exceeds the price. If it does, the business quality is a bonus rather than the thesis.
5. Recognise the trade-off explicitly — this buys you protection, not compounding; you still need somewhere to redeploy the cash.

• What Buffett himself concluded from the same facts: he called it his dumbest purchase because the opportunity cost of a decade in textiles was enormous. Recovering your money is a low bar.
• The asset floor only holds if the assets are liquid and unencumbered. Book value at an insurer or a mortgage originator is a very different thing from inventory and receivables at a mill.

6. Before believing any of it, price the thing the model leaves out: control

1. Establish whether a controlling stake is actually purchasable — founder and insider holdings, dual-class shares, and any poison pill.
2. Add the control premium to the entry multiple. A quoted 6.5x P/E is a minority price; a control price is typically materially higher.
3. Add transaction costs, financing fees and the time cost of a contested process.
4. Re-run the model at the control price, not the screen price. If the edge disappears, the deal was an artefact of the quote.
5. Ask separately whether you can run the business, and who does if you cannot.

• The word "listed" doing a lot of work: a company being quoted at 6.5x does not mean anyone will sell you the whole thing at 6.5x.
• The circular risk — a small illiquid quote is cheap partly because control is not available.

7. Keep your quality gate in front of the arithmetic, not behind it

1. Run the qualitative screen first — moat, returns on capital, durability of cash flow, management — exactly as you would for a share purchase.
2. Only then compute the acquisition arithmetic. A model applied to a business that fails the gate produces a precise wrong answer.
3. Compare the levered whole-business case against a levered market alternative, not an unlevered one, so the debt is not credited to the strategy.
4. Show a sensitivity on the debt share — 0%, 50%, 80% — since the assumptions tab itself labels 80% "max leverage."
5. Re-read your own most recent stated rule before making an exception to it.

• A model whose edge is proportional to the debt input. Halve the leverage and the "6.96x" shrinks; the article shows no such run.
• The commercial context — the issue closes with a survey soliciting interest at a $250,000 minimum to "buy a company together" or invest in a fund doing it. Method published alongside a raise deserves the same scepticism you would apply to a broker's model.