Key takeaways
Capital Cities delivered 19.9% annualized returns over 29 years by structurally separating capital allocation from operating authority. Murphy centralized only capital decisions while granting operating managers near-total autonomy, a design that outperformed the S&P 500 by 16.7 times.
3G Capital's Kraft Heinz failure shows what happens when centralization absorbs local operating judgment. Margin expansion masked organic sales declines for roughly 18 months before more than $57 billion in market value was destroyed.
The Murphy model had boundary conditions that matter for PE firms studying it today. It worked in regulated, asset-light media with strong underlying economics and carefully chosen managers, conditions that don't automatically transfer to every portfolio.
Allocators are beginning to treat operating model design as a structural signal of whether returns hold across fund cycles. GPs that can articulate explicit authority boundaries between the sponsor and portfolio company management stand out during diligence at Fund IV and beyond.
A 36-person headquarters and a $19 billion exit
When Disney acquired Capital Cities/ABC in 1995 for $19 billion, the company's corporate headquarters employed fewer than 40 people. No PR department. No M&A staff. Tom Murphy's personal secretary fielded press calls.
One dollar invested when Murphy became CEO in 1966 grew to $204 by 1996: 19.9% annualized returns over 29 years, against 10.1% for the S&P 500 and 13.2% for an index of leading media companies.
The CBS comparison puts that in context. In 1966, CBS's market capitalization was 16 times larger than Capital Cities'. CBS had the top-rated broadcast network, stations in every major market, and a sprawling portfolio that eventually included a toy company and the New York Yankees. Capital Cities had five TV stations and four radio stations, all in smaller markets. By 1995, Capital Cities was three times as valuable as CBS. CBS diversified aggressively; Capital Cities stayed in the media business it understood and designed a structure that could compound within it.
That structure, more than any content advantage or distribution edge, is what separated Capital Cities from competitors working with the same industry economics. And it maps cleanly onto a problem most multi-asset GPs face today: how to scale a portfolio without eroding the returns that attracted capital in the first place. How a firm governs its assets turns out to be at least as consequential as how it sources them.
Murphy allocated capital, Burke turned it into cash flow
The division of authority at Capital Cities was explicit and functional. Tom Murphy handled strategy, acquisitions, and capital allocation. Dan Burke ran operations. Burke described the arrangement simply: his job was to create the free cash flow, and Murphy's was to spend it.
Murphy never delegated capital allocation. He never used investment bankers for deal sourcing. He built relationships with acquisition targets over years and moved fast when conviction was high, knowing Burke could improve margins quickly after any close. Outside a single stock sale to Berkshire Hathaway to finance the ABC acquisition, Capital Cities issued no new equity over the 20 years prior to the Disney sale. Shares outstanding shrank 47% through buybacks at single-digit price-to-cash-flow ratios.
Operating managers, meanwhile, received near-total autonomy. Burke conducted line-by-line budget reviews once a year. Beyond those meetings, headquarters stayed out.
The ABC acquisition stress-tested the model at scale
The 1985 acquisition of ABC, a company four times Capital Cities' size, was the ultimate proof of concept. Burke's team cut the staff overseeing ABC's TV station group from 60 to eight. WABC New York headcount went from 600 to 400 (Thorndike, The Outsiders). The Manhattan headquarters sold for $175 million. Burke's team closed the margin gap between ABC's TV stations, which had been operating in the low 30s, and Capital Cities' industry-leading levels of over 50% within two years.
Warren Buffett, whose Berkshire Hathaway helped finance the deal with a $517.5 million investment for an 18% stake, later said that Murphy taught him more about running a business than any other person. What Buffett learned was structural: hire carefully, refuse to delegate capital allocation, give operating managers genuine authority, and you build what Charlie Munger called a "seamless web of deserved trust" that compounds at very low overhead.
Murphy operated in regulated, asset-light media with FCC license barriers to entry and advertising-driven economics. He acquired businesses with strong fundamentals and placed carefully chosen managers inside them. The architecture amplified those advantages. It didn't manufacture them from scratch.
3G Capital ran the same playbook and destroyed $57 billion
3G Capital built its reputation on a sequence that looked familiar: acquire companies, cut costs aggressively, generate cash, repeat. At AB InBev, the model worked. US EBIT margins rose from 23% to 36%. At Burger King, where 3G combined cost cuts with a shift to an asset-light franchise model, the company's valuation roughly tripled within four years of the 2010 acquisition.
What happened at Kraft Heinz was structurally different.
Eighteen months of strong margins, then a decade of erosion
Following the 2015 merger, EBITDA margins reached 30%, where even Nestlé, a best-in-class peer, was reaching only about 20%. Competitors began copying the playbook. The early metrics looked exceptional.
Then the brand decay that had been accumulating beneath those margins became visible. But the cuts went deep into the functions that sustain consumer brands. By 2017, Kraft Heinz was spending roughly 2% of sales on advertising, less than half the FMCG sector average, according to Guggenheim Securities. R&D fell to similar lows. Total ad spend had dropped 39% from what Kraft and Heinz had invested separately before the merger.
The long-term results: more than $57 billion in market value destroyed, a 60%+ stock decline, and a $15.4 billion impairment charge in Q4 2018, with further writedowns in 2019 and 2020. By April 2018, 3G Capital co-founder Jorge Paulo Lemann was describing himself as a "terrified dinosaur" at the Milken Institute. 3G Capital sold its entire 16.1% stake in late 2023.
Why the same logic worked at AB InBev and failed at Kraft Heinz
3G Capital's zero-based budgeting model didn't fail because cost-cutting is wrong. It failed at Kraft Heinz, the consumer goods company 3G managed alongside co-investor Berkshire Hathaway. Centralized budget authority stripped marketing, R&D, and brand investment of resources in a category where consumer preferences were shifting faster than efficiency gains could compensate.
At AB InBev, where centralized procurement and distribution scale drove value, centralizing those decisions made sense. At Kraft Heinz, centralizing them starved the business of the local judgment it needed to compete.
At Kraft Heinz, margin expansion and a rising stock price sustained confidence for roughly 18 months after the merger, even as organic sales were already declining within the first year. By the time the $15.4 billion writedown forced a reckoning in early 2019, the structural cause had been embedded for three years. Murphy centralized one function, 3G centralized all of them, and the outcomes diverged accordingly.
How allocators are starting to evaluate a GP's operating architecture
As operating partner teams expand and bolt-on integration layers accumulate, the question of where sponsor authority ends and portfolio company authority begins becomes harder to answer by default. The firms that answer it explicitly stand out during diligence.
Experienced allocators are beginning to read these signals:
Management tenure. High portfolio company CEO turnover can indicate that autonomy is nominal, not real.
HQ headcount relative to portfolio size. A 200-person centralized support team for 25 portfolio companies communicates a different model than a 15-person team for the same count.
Governance language in fundraising materials. Does the GP specify who makes which decisions? Or does the deck describe “operational support” without defining where the GP's authority ends and the portfolio company CEO's begins?
Consistency vs. flexibility. A GP that describes the same operating model for every portfolio company is describing a reflex. One that can articulate different governance for a performing asset vs. a turnaround is describing an architecture.
Board governance structure is increasingly recognized as a value driver, and competitive fundraising conditions are pushing this evaluation forward. This lens isn't yet standard across all LPs. For allocators evaluating re-up decisions at Fund IV and beyond, where scalability becomes the central concern, how a GP governs its portfolio is one of the few structural signals visible before returns either hold or fade.
Bottom line
The Murphy case study gets cited in value investing circles, from Buffett's shareholder letters to Thorndike's The Outsiders. What gets less attention is why so few organizations have successfully replicated it, even with the blueprint in plain sight.
One reason is that the model requires a willingness to leave value on the table at the center. Murphy could have built a larger headquarters, hired integration teams, standardized operations across subsidiaries. Every one of those decisions would have looked reasonable to a board or an LP base. He chose not to because each one would have moved a decision away from the person closest to the revenue.
As PE portfolios grow more complex, the next generation of fund formation and LP communication will increasingly reward GPs who can do something Murphy did instinctively: define the boundary of their own authority, articulate it clearly, and hold it under pressure. The GP that can explain, in concrete terms, which decisions it makes centrally, which it delegates, and why those boundaries exist for each specific asset is building a governance narrative that will survive allocator scrutiny across multiple fund cycles.
The firms that compound at scale in the next decade will likely share one trait with Capital Cities: they will have decided, deliberately and early, what they are choosing not to control.
