That’s where I left it last time. Every empire in Part 1 is real, every number checks out, and laid end to end it looks like the world is being quietly carved into private kingdoms. But my hunch was that we’ve seen this movie before, more than once, and that the interesting part isn’t that the empires rose. It’s how they came down. Or whether they came down at all.

So let’s run the tape. Not as a list of villains who got their comeuppance, because that’s the comforting version and it’s mostly wrong. Run it sorted by how each one ended, because the ending is the whole argument. Some got broken. Some got broken and then quietly reassembled. Some got a stern talking-to and nothing else. One needed an armed rebellion. And the pattern that falls out of that sorting is, to be honest, not the one I expected when I started pulling on this.

The clean break that made the loser richer

Start with the case everyone reaches for, because it’s the one that actually worked. Standard Oil. Rockefeller didn’t out-drill anyone. He out-structured them: secret railroad rebates, a scheme where the railroads paid him a kickback on every barrel his competitors shipped, and then in 1882 the trust device itself, which is literally where the word “trust” as a synonym for monopoly comes from.¹ At its peak Standard ran around 90% of American refining.¹

Then the correction came, and notice where it came from. Not from inside the company. It came from a journalist, Ida Tarbell, whose father had been ruined by one of Rockefeller’s schemes, writing a serial exposé that turned public opinion.² From a law, the Sherman Act, passed in 1890. And finally from the Supreme Court, which in 1911 ordered Standard dissolved into 34 successor companies.¹ Those successors are the oil majors you still know: Exxon, Mobil, Chevron, Amoco, Marathon. The system worked. The state broke the monopoly.

Here’s the twist, and it’s the most important thing in this whole post. The breakup made Rockefeller richer. He owned roughly a quarter of every successor, and once they were separate companies the shares roughly doubled.¹ The breakup nearly made him the world’s first billionaire. So even the textbook success story comes with a warning label: a structural remedy is not a wealth remedy. You can shatter the empire and leave the emperor wealthier than before.

And before this gets too triumphant, the honest counter. A revisionist strand, Dominick Armentano and the Chicago-school crowd, argues that Standard was already losing market share before 1911, its share falling below roughly 70% by the time the Court ruled, that prices had been falling under it, and that the breakup was therefore more politics than economics, punishing a giant that competition was already eroding.³ I don’t fully buy it. The rebate machinery was genuinely predatory. But it’s a fair point, and it sharpens the real question. If markets were already correcting Standard, then the most celebrated antitrust win in history might have been the state taking credit for what the market was doing anyway. Hold that thought, because it’s exactly the argument the optimists will make in Part 3.

Broken, then quietly put back together

Now the case that should bother you more than any breakup that failed: the breakup that worked and then came undone.

AT&T. In 1984, after a decade-long government case, the Bell System was split apart: seven regional “Baby Bells” spun out, AT&T kept long distance.⁴ A real, court-ordered dispersal of one of the largest concentrations of corporate power America had ever produced. Clean.

Then the Telecommunications Act of 1996 changed the rules and let them merge again.⁴ So watch what happened. SBC, one of the seven, ate two of its siblings, then in 2005 bought old AT&T itself and took its name. Bell Atlantic ate another sibling and became Verizon. Within about 25 years, seven companies had reassembled into two, and one was wearing the broken-up parent’s name like a costume.⁴ The remedy had a half-life of roughly a generation. The moment the legislature stopped holding the pieces apart, they drifted back together, because that’s what pieces of a natural monopoly do.

That, more than Standard Oil, is the shape I keep seeing. Not “concentration rises and gets fixed.” More like: concentration is the resting state, dispersal is the thing you have to actively hold in place, and the instant you stop holding, it flows back.

The state won the case and lost the remedy

Sometimes the state doesn’t even get that far. Microsoft was ruled an illegal monopoly, full stop. The judge found it had monopoly power and abused it. In June 2000 a court ordered it split in two. A year later, an appeals court threw out the breakup, partly because the trial judge had been giving press interviews trashing the company, and a new administration’s Justice Department settled for a wrist-slap: share some technical interfaces, no breakup, no real changes.⁵ One critic’s line stuck with me: after that, the only way Microsoft could die was suicide. It’s a three-trillion-dollar company now.

The lesson is brutal and simple. Proving the monopoly and breaking the monopoly are two different fights, and the powerful only have to win the second one. The state can be completely right about the diagnosis and still lose the remedy to politics and procedure.

Never broken at all, and the law made it worse

Then there’s the category where the correction just… doesn’t happen, or runs in reverse.

2008. The banks that were “too big to fail” got bailed out and then re-regulated under Dodd-Frank, which is to say supervised more tightly but not made smaller. And in 2018 Congress rolled back a chunk of Dodd-Frank, raising the threshold for the strictest oversight from 50 billion dollars in assets to 250 billion.⁶ Nobody got broken up. “Too big to fail” turned out to mean, functionally, “too big to break,” because the concentration got rebranded as stability and you don’t dismantle the thing holding the system up.

Media is the darker version, where the law itself was the concentrating force. The Telecommunications Act of 1996, the same one that let the Baby Bells re-merge, removed caps on media ownership.⁷ This is the exact inverse of Standard Oil. There the state broke the monopoly; here the state built it, by vote. Same institution, opposite sign. Which tells you concentration versus dispersal was never a law of nature. It was a choice Congress made, and it made it both ways within a century.

Ended only by revolt, and even then just changed owners

Now the one that hits closest to home, and the cleanest historical rhyme for the whole series.

The East India Company. A trading company that became a sovereign: its own army of roughly 260,000 men, twice the size of Britain’s, the power to tax, to administer justice, to govern most of the subcontinent.⁸ This is the purest version of the thing I’ve been describing. One entity holding the rule-book and the cashflow and the actual guns over a fifth of humanity. If you want to know what “rule-book plus cashflow plus attention plus force, fused in one body” looks like at full scale, it already existed, and Indians lived under it.

How did it end? Not through a court. Not through a law passed by sober reformers. It ended after the Rebellion of 1857, a violent shock from below, after which the British Crown dissolved the Company and took India directly.⁸ And here’s the part that matters for the whole argument: that wasn’t dispersal. It was a transfer. The private monopoly became a Crown monopoly. The Indian under the Company got a new master, not freedom. When power fuses to that degree, the “resolution” tends to be absorption by something even bigger, paid for in blood, not a breaking-up into many hands.

Nobody had to break De Beers

And then, to keep myself honest, the cases where nobody corrected anything and the empire fell anyway, because the comforting “they always fall eventually” story is real. It’s just slower and dumber than people imagine.

De Beers ran something like 80 to 85% of the world’s diamonds for most of the twentieth century. No antitrust court ever broke it; it simply ran the cartel from London, out of reach. What hollowed it out wasn’t a regulator at all. It was lab-grown diamonds, a cheaper substitute, which dragged its share down toward 29% and left it sitting on billions in unsold stock.⁹ A substitute did what a century of law never managed.

Go back two thousand years and you find Crassus in Rome, richest man of his era, running a private fire brigade that extorted burning buildings before it would put out the fire, plus vast slave-worked estates. The original one man with the rule-book, the cashflow, and his own private force.¹⁰ Go forward eighteen centuries and you find the Rothschilds, the largest private fortune of their age, never broken by any law either; they were ground down by ordinary joint-stock banks competing them flat, and then by the First World War cutting the cross-border capital flows their whole model ran on.¹¹ I bring those two up for one reason. This is not a 2026 panic. People have been concentrating money and power and a private army into one pair of hands for as long as there’s been money to concentrate. What’s different now is the speed and the reach, not the underlying move.

But notice what “they always fall eventually” actually buys you. The fall is usually fragility: a lost war, a cheaper rival, a bad bet. Not reform. And “eventually” can mean generations and a lot of damage along the way. Which is the line I want to land hard: eventually is not a policy. “It’ll sort itself out in due course” is not a plan you can vote for.

The home-soil arc, both directions

India ran its own version of this whole cycle, and it cuts both ways, which is exactly why it’s useful. In 1969 the government nationalized the major banks, and through the License Raj the state clamped down hard on private industrial concentration, the MRTP regime, permits for everything.¹² That was concentration broken by state seizure. Then in 1991, facing a balance-of-payments crisis, India liberalized and dispersed that state grip back out toward private enterprise.¹³ Concentrate, seize, disperse, all within living memory, on home soil. Whatever you think this series is leading toward, India has already been around the loop once.

And there’s a cleaner, more uncomfortable version next door. The East Asian developmental states built their giants on purpose and could also break them. South Korea grew its chaebol with directed credit and then, in the 1997 crisis, watched 11 of the top 30 collapse, with Daewoo going under for, of all things, defying the government’s restructuring orders.¹⁴ The cashflow-king who ignored the rule-book-setter, and got destroyed for it. That’s a correction the American and Indian cases mostly lack, and it’s worth sitting with: a concentration the state built, the state could also un-build, by force.

What the ordering actually says

Put the endings side by side and the argument writes itself.

The empire	How it ended	What that proves
Standard Oil	Court breakup, 1911	Works, but the owner got richer
AT&T	Broken 1984, re-merged by 2007	Doesn’t stay broken
Microsoft	Breakup ordered, then reversed	Proof isn’t a remedy
2008 banks	Re-regulated, never shrunk	The fix got rolled back
US media	Law removed the caps	The state built the monopoly
East India Co.	Revolt, then nationalized	A transfer, not a dispersal
De Beers	Undercut by a substitute	No reformer needed, just a rival

Read down the right-hand column. In not one of these did a concentrated power voluntarily disperse itself. Every correction came from outside: a muckraker, a court, a legislature, a crash, a rebellion, a cheaper product. The thing being corrected never reformed itself, because why would it? Concentration is the resting state, and the resting state has no incentive to get up.

And the corrections that did land mostly didn’t stay landed. AT&T flowed back together in a generation. Dodd-Frank got gutted in eight years. Microsoft’s breakup died on appeal in one. The pieces drift back together unless something keeps actively holding them apart, and “something” means sustained outside force: a press that exposes, a public that’s angry, a state willing to use the law. Remove any one of those and there’s no correction at all. The whole Standard Oil win took about a quarter-century to build that machinery, from the 1890 law to the enforcement bodies of 1914.¹⁵

So when someone tells you not to worry, that these things always sort themselves out, the historical record technically agrees with them and completely fails to reassure. Yes, the empires fall. Eventually. After decades, often after a catastrophe, frequently into the hands of something larger, and only when an outside force pays the cost of breaking them and then keeps paying to hold the pieces apart. That is not the same as “it self-corrects.”

Which lands me on a question I genuinely don’t have settled. If this keeps coming back no matter how often we break it, if the default just reasserts itself the moment we stop pushing, then maybe I’ve been asking the wrong thing. Maybe the question isn’t how you stop concentration. Maybe it’s whether stopping it is even possible, or whether the most you can ever do is hold it off for a while and call that a victory.

That’s the next part.

References

Footnotes

1. Standard Oil (Wikipedia); on the 1911 dissolution into 34 successor firms, the ~90% peak share, and the post-split share rise that made Rockefeller richer (he held roughly a quarter of every successor). Several of those successors later re-merged, Exxon and Mobil in 1999. ↩ ↩² ↩³ ↩⁴

2. The History of the Standard Oil Company (Ida Tarbell, serialised in McClure’s, 1902–04). ↩

3. The revisionist case is associated with Dominick Armentano, Antitrust and Monopoly: Anatomy of a Policy Failure (1982); on Standard’s pre-1911 share decline see Standard Oil (Wikipedia). ↩

4. Breakup of the Bell System (Wikipedia) for the 1984 divestiture into seven regional companies; the Telecommunications Act of 1996 (Wikipedia) enabled the re-mergers that collapsed them back to roughly two carriers by the mid-2000s, with SBC acquiring and renaming itself AT&T in 2005. ↩ ↩² ↩³

5. United States v. Microsoft Corp. (Wikipedia); the breakup was ordered in June 2000 and reversed by the D.C. Circuit in June 2001, after which the case settled with no breakup. ↩

6. Significant Rollback of Dodd-Frank Signed Into Law (Yale Journal on Regulation, 2018); the 2018 law raised the “systemically important” oversight threshold from US$50bn to US$250bn in assets. ↩

7. Telecommunications Act of 1996 (Wikipedia); the Act loosened media-ownership caps. ↩

8. Armies of the East India Company (National Army Museum, UK) on the ~260,000-strong private army; East India Company (Wikipedia) and the Government of India Act 1858 (Wikipedia) on the post-Rebellion transfer of the Company’s territories to the Crown. ↩ ↩²

9. De Beers and the rise of lab-grown diamonds (Fortune, 2024); De Beers ran roughly 80–85% of the diamond market for most of the 20th century and has fallen to around 29% of supply, hollowed out by lab-grown stones rather than by any regulator. ↩

10. Marcus Licinius Crassus (Wikipedia); Crassus (c. 115–53 BCE) built his fortune partly through a private fire brigade that bought burning properties cheaply, plus large slave-worked estates. ↩

11. Rothschild family (Britannica); the family’s dominance faded as joint-stock banks competed it down and the First World War severed the cross-border capital flows its model depended on. ↩

12. Bank nationalisation in India (Wikipedia) on the 1969 nationalization of the major banks. ↩

13. Economic liberalisation in India (Wikipedia) on the MRTP Act / License Raj and the 1991 liberalization that reversed much of that state concentration. ↩

14. 1997 Asian financial crisis (Wikipedia); during the 1997–99 crisis, 11 of South Korea’s top 30 chaebol collapsed, with Daewoo (around US$50bn in debt, estimates vary) failing after defying government restructuring orders. ↩

15. History of United States antitrust law (Wikipedia); the correction machinery built up over roughly a quarter-century, the Sherman Antitrust Act of 1890 supplying the weapon and the Clayton Antitrust Act and Federal Trade Commission Act of 1914 supplying the enforcement architecture. ↩