LG Chose Its Heir in 2004. The Lawsuit Came Fourteen Years Later.

Take the 11.28% stake in LG Corp that Koo Bon-moo left behind when he died in 2018, and look at how his family actually divided it. His adopted son and successor, Koo Kwang-mo, got 8.76 percentage points of it, 77.7% of the block. His widow and two daughters split the remaining 2.52 points, 22.3%, between the three of them.¹³

Now look at what Korea's default inheritance formula would have given them. A spouse takes one and a half shares to each child's one share. With one spouse and three children, including the adopted son, that works out to roughly 22.2% for Koo Kwang-mo alone and roughly 77.8% for the other three combined.¹⁵

Put those two numbers next to each other. What Koo Kwang-mo actually received, 77.7%, is almost identical to what the law would have given the three women combined, 77.8%. What the three women actually received combined, 22.3%, is almost identical to what the law would have given Koo Kwang-mo alone, 22.2%. The family didn't deviate from the statutory formula. They flipped it.

That is not a coincidence anyone has to prove, and this piece isn't going to try. It is close enough that a reader can sit with it for a minute, because it explains why the resulting court fight was never really about a bad deal. It was about whether a family that had already decided who was going to run the company in 2004 could write an agreement in 2018 that made the numbers match the decision, and whether anyone who signed that agreement could later argue they hadn't meant to.

LG had settled who would run it more than a decade before the estate was there to divide. Koo Bon-moo's only biological son died in an accident in 1994. Korean primogeniture custom in the chaebol families runs to the eldest son, and LG's own succession pattern followed it strictly enough that a chairman with two living daughters and no son was, by the family's own rules, a chairman without an heir.⁹¹¹

In 2004, Koo Bon-moo adopted his brother Koo Bon-neung's eldest son, Koo Kwang-mo, then 26. The adoption made Kwang-mo his legal child, with the same inheritance standing as any biological child under Korean law. It also made his father's wife, Kim Young-sik, his legal mother.⁹¹⁶

Notice what that adoption actually accomplished, separate from anything sentimental about it. It didn't just supply a son. It fixed, fourteen years in advance, which of the four eventual heirs the family had already agreed would receive the company. Everything that happened between 2018 and 2026, the settlement, the lawsuit, the ruling, was downstream of a decision the family had already made and simply needed the estate to catch up to. A will announces an intention after the fact. An adoption enacts one in advance and lets everyone spend a decade and a half getting used to it.

Here is the part that makes this case worth a founder's attention rather than a chaebol-watcher's. Korean inheritance law does not, on paper, prefer sons over daughters. Article 1009 of the Civil Code gives every child, regardless of sex or birth order, an identical share, with the surviving spouse getting one and a half times a child's share on top. Koo Yeon-kyung and Koo Yeon-soo were legally entitled to exactly what Koo Kwang-mo was entitled to, as siblings under the same article.¹⁵

What the law does not do is force anyone to take that default. Article 1013 lets co-heirs divide an estate by whatever agreement they can reach, at any time, and that agreement can look nothing like the statutory ratio.¹⁶ The default exists for heirs who can't agree. It is not a floor for heirs who can.

So the 66-year-old promise, sons and daughters split the estate identically, and the release valve that has always sat next to it, unless everyone agrees to something else, are the same statute. One clause hands daughters an equal claim. The next one lets a family sign that claim away, as long as everyone puts their name on the same page. Korea rewired its family law for gender equality in the 2000s and left the exit door exactly where it was.²⁰

The three women who eventually sued weren't arguing that Article 1009 had been violated. They knew they'd signed something else. Their argument was that the signature shouldn't count.

In February 2023, Kim Young-sik and her daughters filed suit at the Seoul Western District Court to unwind the 2018 settlement and force a redistribution along the statutory formula, the one that would have handed each daughter roughly what her brother got.⁵⁴ If they'd won outright, Koo Kwang-mo's total LG Corp stake would have dropped from roughly 16% to roughly 9.7%, a large enough cut to put his control of the group's holding company in genuine question.²³

Their theory was mistake, not merely regret. They said they had agreed to the 2018 split believing a legally valid will already dictated it, so that signing was a formality rather than a negotiation. No such will existed. Korea requires a holographic will to be entirely handwritten, dated, and signed by the testator's own hand to count at all; an informal note doesn't meet that bar no matter how sincerely it states an intention.⁶ If you believe you are ratifying a legal fact and you are actually creating one, Article 109 says you can undo your signature.¹⁷

The court didn't buy it. It found the 2018 negotiation had produced a memorandum, and separately a note signed by Kim Young-sik herself, both pointing the same direction: that Koo Bon-moo had wanted his adopted son to inherit the shares that ran the company.⁴ More damaging to the mistake theory, the record showed the plaintiffs had been briefed at length, had counsel in the room, and had gone back to the table once already to get the agreement revised so the daughters received something rather than nothing.¹⁶

That last fact does most of the legal work, and it's worth sitting with because it inverts the intuition a layperson brings to this kind of case. Successfully renegotiating your own allocation should be evidence you understood exactly what was on the table. In this court, it became evidence you understood the whole table, which is precisely what a claim of mistake needs you not to have understood. The daughters' one demonstrated win, getting Koo Kwang-mo to give up 2.52 points instead of zero, became Exhibit A for why they couldn't later claim they never knew there was anything to negotiate.

Kwang-mo's side also raised a second defense that the court didn't need to reach a verdict on to feel comfortable: a settlement four years old sits close to, if not past, the window Korean law gives someone to rescind for mistake once they've had the chance to discover it.⁶ A claim that arrives that late has to work twice as hard to explain why it took so long to notice the mistake.

One more asymmetry in the 2018 settlement is easy to miss if you only look at percentages. Kim Young-sik, the widow, had the single largest statutory entitlement of any heir, one and a half units against everyone else's one. She received none of the LG Corp stock. Her share of the family's compensation came entirely out of the roughly ₩500 billion in cash, real estate, and art that she split with her daughters.¹⁷

Every one of those substitute assets can be spent, displayed, or willed to someone else on its own timeline. None of them carries a vote at an LG Corp shareholder meeting. The family didn't just give the operating asset to the person chosen to run the company; they specifically kept it away from the person who, on paper, had the strongest claim to the largest piece of it. That isn't proof of anything sinister. It's a fairly clean illustration of what "inheritance" means once you stop assuming an estate is one asset and start noticing it's several, only one of which is a steering wheel.

LG Corp has since cancelled treasury shares in tranches, most recently in May 2026, which mechanically lifts every remaining shareholder's percentage without anyone buying or selling a thing. Koo Kwang-mo's stake rose from 15.95% to 16.60% on the back of one such cancellation; Kim Young-sik's rose from 4.20% to 4.37%, Koo Yeon-kyung's from 2.92% to 3.03%, Koo Yeon-soo's from 0.72% to 0.75%.²² Everyone got slightly richer in percentage terms and nobody's relative position changed at all. The gap between the chairman and the three women who sued him widened in absolute points and stayed exactly where the 2018 agreement had put it in relative terms.

That flatness is the tell. A cap table that only ever grows all its holders in the same proportion is a cap table where the last contested decision was the only decision that mattered. Everything downstream of it, buybacks, dividends, years passing, just rescales the same ratio.

For a period, the same four people were on the same side of a different courtroom. In late 2022, Koo Kwang-mo, his mother, and both sisters jointly sued Korea's National Tax Service, arguing the roughly ₩990 billion combined inheritance tax bill on the family, including Koo Kwang-mo's own ₩720 billion share, had been calculated using an inflated valuation of a 1.12% unlisted LG CNS stake. They wanted about ₩1 billion back.¹² In April 2024, the Seoul Administrative Court dismissed that claim too, without elaborating much on why.¹³¹⁴

So for a stretch of 2022 and 2023, the family's official position in one courtroom was that the state had overvalued the estate, while its unofficial position was hardening, in preparation for a different courtroom, that the estate had been undervalued for the three women specifically. Both claims can be true at once without contradicting each other on paper. A stake can be worth too much to the tax authority and still have been divided too unevenly among the heirs. But it's the kind of overlap that would make a skeptical judge, and a mildly amused reader, want to see the family's internal memos from that period rather than just its filings.

The Seoul Western District Court's 11th Civil Division dismissed the women's suit on February 12, 2026, finding the 2018 partition agreement lawfully executed and free of the fraud or actionable mistake the plaintiffs needed to prove.²³ The plaintiffs called the outcome reliant on one-sided testimony from LG officials and said they would appeal immediately.⁴ As of this writing that appeal hadn't yet produced a result.

What the ruling settled has a narrower shape than the headlines about it. It didn't decide whether the 2018 split was fair, generous, or gender-blind. Korean courts reviewing a partition agreement under Articles 109 and 110 aren't weighing fairness; they're checking whether a specific, narrow kind of deception or confusion occurred at the moment of signing. An agreement can be lopsided by any outside measure and still be completely enforceable, provided everyone understood what they were agreeing to. That is close to the whole holding here: not that 77.7% to 22.3% was the right split, but that it was a split everyone signed with open eyes, which is a different and much lower bar.

That is also, not incidentally, the same bar a founder clears or fails to clear every time a co-founder, an early employee, or an investor signs away something valuable in a term sheet, a vesting schedule, or a buy-sell agreement. Courts don't rescue a bad bargain. They rescue a bargain nobody actually understood they were making.

If you are ever the one drafting an agreement that concentrates control in one person's hands while everyone else takes a smaller, non-voting consolation, the LG case is a usable checklist for making that agreement survive contact with a later regret.

1. Give every signer their own lawyer, not the family's lawyer. The strongest evidence against the women's mistake claim was that they'd been advised and had negotiated a change. A single law firm representing "the family" produces exactly the ambiguity a mistake claim needs.

2. Write down, at signing, why each party is agreeing. Not just what they're getting, but their stated understanding of the alternative. If a mistake claim ever surfaces, the first thing a court asks is what the signer believed at the time. Make that belief a matter of record before it becomes a matter of memory.

3. Never let a real document sit next to an informal one and call them equivalent. Koo Bon-moo's memorandum functioned as strong evidence of intent precisely because nobody had to pretend it was a will. If an informal note is doing real work in your plan, say so explicitly to everyone bound by it, rather than letting people assume a document has a legal weight it doesn't.

4. Separate the control-bearing asset from the compensating one, on purpose, and say so. LG's settlement quietly kept every voting share with the person chosen to run the company and paid everyone else in things that don't vote. That can be a legitimate design. It stops being one if nobody named it during the negotiation and someone finds out about it afterward.

5. Know your jurisdiction's window to challenge, and don't assume silence is agreement forever. A rescission theory that shows up years after the fact has to explain the delay as convincingly as it explains the original mistake. If you are the party who might later want to challenge a division, the clock does not wait for your certainty to arrive.

Yes, and that refusal would have mattered. Under Article 1013, a partition agreement requires every co-heir's consent; without it, the estate falls back to the statutory formula in Article 1009, which would have given each daughter roughly what their brother received.¹⁵¹⁶ The women's later argument wasn't that they lacked the power to refuse. It was that they hadn't understood they had it.

No. Adoption under Korean law gives the adopted child identical statutory inheritance standing to a biological child, nothing more. It doesn't override Article 1009's equal treatment of siblings, and it doesn't bind other heirs to any particular division. What actually delivered control to Koo Kwang-mo was the 2018 partition agreement his mother and sisters signed, not the 2004 adoption on its own. The adoption decided who the company wanted. The agreement is what made it legal.

A will has to satisfy Korea's formal execution requirements to bind anyone, and testators, like everyone else, sometimes die without one that meets the bar, or without one at all.⁶ A partition agreement doesn't need a will behind it. Co-heirs can reach any division they can agree to, with or without a document expressing what the deceased wanted. The tension in this case was that the daughters treated an informal memorandum as if it carried a will's binding force, when its only real function was to explain, not command, what the family later agreed to on its own.

Every fact in this case that sounds like a smoking gun, the near-perfect inversion of the statutory shares, the widow left holding art instead of votes, the family suing the tax office together before suing each other, points toward the same unglamorous conclusion. None of it mattered to the court as much as whether four adults had understood, on one specific day in 2018, what they were putting their names to. They had. The company their signatures assigned to one of them was decided in 2004. The lawsuit just needed until 2026 to find out that showing up late to a decision isn't the same as being deceived about it.