Investing Classics: Acquisition Case StudiesWhat Berkshire's Deals Teach

This was the watershed of Buffett's philosophy. Early on he was a Graham disciple, picking up "cigar butts" — one last free puff left in a dying company. But a mediocre company's value doesn't grow; you earn a one-time reversion to fair value, then must hunt for the next butt. Munger forced him to see that the real money comes from the long compounding of a wonderful business, not a one-time discount. A company that can keep reinvesting profits at a high ROIC will thicken your return over time even if you paid a fair price — that is the essence of "quality."

Nebraska Furniture Mart, 1983. Buffett bought roughly 90% from the 89-year-old Rose Blumkin ("Mrs. B") for about $55 million — no audit, no due-diligence team, closed on a single handshake. This single-store furniture seller, through relentlessly low-cost operations, generated sales far beyond its peers, and almost no one could beat it on price; four decades later it remains one of the highest-volume single furniture stores in the U.S. This is the template for a fair price paid for a great business: a moat that is real and durable, not a one-time bargain on paper.

Limit: the biggest misuse of this rule is treating it as a permit to overpay for an overvalued name — in 1999, countless investors paid unrecoverable prices for tech stocks in the name of "great companies." A fair price is not "any price"; overpay for a great company and it still takes a decade to break even. The flip side: a mediocre company, even cheap, is often a value trap whose discount never closes.

• Is my "great" backed by a specific moat mechanism, or just a recent good chart?
• Are the growth assumptions embedded in today's price realistic?
• If this company's ROIC were only 10% for the next decade, would I still hold it?
• Am I persuaded by quality, or rationalizing a price that has already risen?

Turnaround buying is about diagnosis, not bottom-fishing. You have to pull it apart and ask: is the core economics of the business broken, or just the balance sheet and management? The former is unsolvable; the latter is fixable. GEICO's low-cost model — bypassing agents to sell auto insurance directly — gave it a structural cost advantage, and that core was never damaged in 1976; the real problem was grossly inadequate reserves and reckless expansion. This "good business, bad operations" combination has enormous rebound power once the right management arrives.

GEICO's stock fell from about $61 in 1972 to about $2 in 1976, near bankruptcy over a reserve shortfall. Buffett invested about $4.1 million in 1976, backed the reforms of new CEO John Byrne, and kept adding. It closed a full circle: Buffett had first bought GEICO in 1951 at age 22 (his teacher Graham sat on its board); in 1996 Berkshire bought the remaining ~49% it didn't own for about $2.3 billion, taking full ownership — from a young man's first investment to owning it whole, across 45 years.

Limit: the biggest trap in turnarounds is mistaking structural decline for temporary distress. Kodak, Sears, and Nokia's phone unit were all "cheap bad-news stocks" — catching any of them was catching a knife, because their core economics were collapsing. GEICO could rebound precisely because its low-cost moat was intact; if the moat itself is crumbling, no price is cheap enough. Never use "turnaround" as an excuse to deny a decline.

• Is the core profit engine (the moat) itself damaged, or just a temporary operating error?
• Is there a clear, executable repair path (new management / capital / cutting units)?
• If the turnaround never comes, is my downside limited, or could it go to zero?
• Am I diagnosing fundamentals, or merely lured by "it fell a lot"?

Railroads are the textbook of the "heavy-asset + irreplaceable" pole. The moat comes from three things: a replacement cost so high no one can build a parallel line; energy efficiency that crushes trucking (one gallon of diesel can pull a ton of freight hundreds of miles); and a near-duopoly in the west (BNSF vs. Union Pacific). Such a business has low ROIC and slow growth, but is extremely predictable and a natural inflation defense — rates can rise with inflation and physical assets don't depreciate away. It is the opposite pole from "asset-light, high-ROIC" in Buffett's framework.

In November 2009 Berkshire announced it would buy the ~77.4% of BNSF it didn't own for about $34 billion in cash and stock, valuing the deal at roughly $44 billion including assumed debt, closing in February 2010 — the largest acquisition in Berkshire's history at the time. For over a decade since, BNSF has delivered vast, steady cash flow. This was not a deal chasing a high rate of return, but one chasing "predictable + irreplaceable + enduring."

Limit: the price of heavy-asset predictability is a low ROIC ceiling and capex that swallows cash; overpay, and returns get ground down by ongoing maintenance investment. More fundamentally, "betting on an economy" only holds when you're confident that economy trends up over the long run — bet on the wrong country or era (say, Japan at its 1989 peak) and the same heavy assets become a long-term drag. Note too: BNSF enjoys Berkshire's ultra-low-cost capital, which individuals cannot replicate.

• Is this asset truly irreplaceable, or just temporarily unchallenged?
• Can its ROIC and capex needs support the long-term return I require?
• Do I have real conviction in the twenty-year direction of the economy / industry I'm betting on?
• Does my entry price already reflect the reality of "stable but low-growth"?

Dexter holds two stacked lessons. First, moats are dynamic. Dexter's cost advantage was wiped out in the 1990s by low-cost imported shoes; statically it looked like a cheap good company, dynamically its moat was already dead. Second, and more lethal — the method of payment. Buffett paid not in cash but in 25,203 shares of Berkshire A stock (about $433 million then); those shares kept compounding wildly for two decades, worth roughly $5.7 billion by 2014. He not only bought a company that ended up worthless, he permanently gave away equity that would have kept compounding. A mistake paid in cash is bounded; a mistake paid in stock compounds.

Dexter Shoe, 1993: an all-stock acquisition for about $433 million, crushed within a few years by import competition, its value falling to near zero. Two decades later history nearly repeated: Precision Castparts, acquired in 2016 for about $32 billion (Berkshire's largest ever), had its aerospace-parts demand gutted by the 2020 pandemic; Berkshire took a roughly $9.8 billion writedown, with Buffett publicly admitting "I paid too much." A generation apart, the errors had the same shape: overestimating a moat's durability.

Limit: not every writedown is a mistake — economies are cyclical, aerospace always recovers, and a temporary headwind differs from permanent damage. The real error is permanently overestimating the durability of a competitive edge. The opposite misuse is just as dangerous: using "long-term thinking" as a shield to refuse to admit an already-evaporated moat and stubbornly hold on. Honestly separating "cyclical headwind" from "structural collapse" is the hardest and most crucial work here.

• Can this moat survive a decade of shifting technology and cost structure?
• Am I paying in cash or equity? If equity, have I counted the forgone compounding as real cost?
• If I overpaid, is it a temporary cyclical issue, or did I read the business wrong?
• Am I still holding on new facts, or because selling means admitting my original error?