How to Read Chaebol Governance as a Foreign Investor

Key Takeaways

• Chaebol (a large, family-controlled South Korean conglomerate spanning multiple industries—Samsung, SK, Hyundai Motor Group, and LG are the best-known) governance is not one thing to check off a list. It is a control-versus-ownership gap, and Korea’s Fair Trade Commission (FTC) measures it every year. The FTC designates and monitors these conglomerates. In its most recent count, the families that control Korea’s largest business groups held an average direct stake of just 3.5%–3.7% of the companies they run. That figure has not moved in five years, even as the groups’ own affiliated companies kept building up cross-holdings in each other.
• The tool that used to close that gap—circular shareholding, where three or more affiliates hold stakes in a closed loop (A owns B, B owns C, and C owns A)—has been banned for new formations since 2014. That ban didn’t unwind the loops that already existed. 2026 has already produced one live FTC enforcement case and one proposed rule closing a fresh workaround.
• Holding companies are the FTC’s preferred fix: collapse the web into one legally defined parent that owns clean, disclosed stakes in each subsidiary. Korea had 173 of them at last count. But only 47 of the 102 largest disclosure-target groups have converted their entire structure to one. LG did it in 2001–2003. SK did it in 2007. Samsung, Korea’s largest conglomerate, has not.
• None of this is news you can trade on a headline. It is the plumbing behind the Korea discount — the mechanical layer underneath that story, not a restatement of it.
• This is an explainer of structure, not a stock call. It doesn’t tell you which chaebol-affiliated company to buy or avoid. It tells you what questions the ownership chart should answer before you decide.

Contents

1. What “chaebol governance” actually means
2. The control-ownership gap, in the regulator’s own numbers
3. Circular shareholding: banned since 2014, not gone since 2014
4. Holding companies: the fix half of Korea hasn’t adopted
5. Board independence: what it changes, and what it doesn’t
6. Four traps investors keep falling into
7. FAQ

1. What “chaebol governance” actually means

Every explainer about Korean stocks eventually says some version of the same sentence: founding families control chaebol with a sliver of the equity. Fine. Now the actual question—the one that determines whether that sentence should change how you value a stock: control through what, exactly?

Three separate mechanisms get flattened into that one sentence. They behave differently, carry different legal exposure, and are moving on different timelines in 2026. Circular shareholding is a structure—a closed loop of cross-holdings that amplifies one family’s voting power far past its cash invested. A holding company is the opposite: a single, legally defined parent that owns clean, auditable stakes in its subsidiaries. Korean regulators have spent two decades trying to get more conglomerates to adopt one. Board independence is a separate lever entirely—who sits on the board, rather than who owns what beneath it.

Most English-language coverage of Korea’s 2026 governance reforms treats these as line items on a compliance calendar: dates, thresholds, board-reconstitution deadlines. We checked two of the closest English pieces on this exact subject in full while researching this one (see the FAQ for what we found and did not find). That kind of checklist is useful if you are the general counsel of a Korean subsidiary. It is close to useless if you are deciding whether a specific holding in your portfolio actually has the control structure the checklist implies—because none of those pieces name a single group or trace a single ownership chain.

This piece does both, using only regulator-published numbers, and stops exactly where the numbers stop.

2. The control-ownership gap, in the regulator’s own numbers

Start with the FTC’s own homework. Each year it designates Korea’s largest business groups for enhanced disclosure. For the biggest of those, it also restricts how they can invest in each other. As of May 1, 2026, that list ran to 102 groups holding at least ₩5 trillion (roughly $3.62 billion) in assets — up from 92 the year before — spanning 3,538 affiliated companies. A smaller, higher tier faces the tighter bans described in the next section: 47 groups with at least ₩12 trillion (about $8.68 billion) in assets, a threshold pegged to move with the size of the economy.

The FTC’s most recent full ownership analysis, covering the 92 groups designated the previous year, put a number on the gap directly. It examined 81 of those groups with an identifiable controlling individual and 3,090 affiliated companies. The average combined internal ownership ratio ran to 62.4%—the sum of the controlling family’s own stake, affiliated-company cross-holdings, nonprofit foundation holdings, executive holdings, and treasury shares. Strip that down to just the controlling family’s own direct stake, and the number is 3.5%–3.7%. It has sat in that narrow band for five straight years, even as the affiliated-company share of the total kept climbing.

Read those two numbers side by side, and the mechanism is obvious. The families who set strategy and appoint executives hold a direct stake that rounds to a twentieth of the group’s total internal ownership. Until 2025’s Commercial Act amendments, they also owed a formal duty only to “the company,” rather than to “the company and its shareholders”—the mechanics of that change are covered in detail in the pillar piece on the Korea discount. The other nineteen-twentieths of that 62.4% is companies holding stakes in each other. That is the governance story, not a footnote to it—and the next section is how it gets built.

3. Circular shareholding: banned since 2014, not gone since 2014

Here is the mechanism in its purest form. Company A holds shares in Company B. Company B holds shares in Company C. Company C holds shares back in Company A. No single link in that chain looks unusual on its own — cross-holdings between related companies are not inherently exotic anywhere in the world. What is unusual is the loop. A family that starts with a real stake in only one corner of that triangle can, through the loop, effectively control votes across all three. That voting power is built on capital the group already owns of itself.

Korea moved against new versions of this structure in a 2013-12-31 amendment to the Fair Trade Act, effective 2014-07-25. The mutual-investment-restricted tier of large groups — the ₩12-trillion-plus tier from Section 2 — is barred from forming new circular-shareholding links. The penalty for forming one anyway is real: up to a 20% surcharge on the value of the shareholding involved, plus criminal exposure running to three years in prison or a fine of up to ₩200 million (about $144,700).

Read that ban carefully, because what it doesn’t say matters as much as what it does. It stops new loops. It gave groups with loops already in place a grace period — six months to three years, depending on the circumstances — to unwind them; that isn’t an instant reset. More than a decade later, unwinding old structures and inventing new workarounds have both turned out to be live, ongoing activities, not settled history.

Two things happening in 2026 make that concrete.

First, the FTC is mid-review of a case alleging that a chain—Korea Zinc, routed through two Australian subsidiaries, into Young Poong, and back into Korea Zinc—functioned as exactly this kind of loop. It was deployed during a management-control fight, not an ordinary financing decision. The commission received a formal complaint in January 2025, sent Korea Zinc a review report in April 2026, and a decision was expected in August 2026. As of August 22, 2026, no ruling had been published, and this piece doesn’t predict one. What it demonstrates regardless of outcome: twelve years after the formation ban, this is still a structure regulators actively investigate, not one confined to pre-2014 textbooks.

Second, the FTC itself proposed a rule change on 2026-07-23, still a draft, open for comment through September 1. It targets a specific workaround: a holding-company-affiliated venture-capital fund investing in an outside fund that then reinvests in the holding company’s own affiliates. That achieves the substance of a circular link without technically forming one. The regulator finding new plumbing to close, twelve years into the ban, tells you something the ban’s passage date alone does not: the incentive behind circular structures did not disappear when the formation rule did.

4. Holding companies: the fix-half of Korea hasn’t adopted

If circular shareholding is the disease, the holding-company structure is the regulator’s preferred medicine. It is worth understanding as a real legal category, not just a synonym for “reformed.”

Under the Fair Trade Act, a company becomes a regulated holding company once two tests are met: its subsidiary shares are worth at least half its own total assets, and its total assets pass ₩500 billion (about $362 million)—a lower ₩30 billion (about $21.7 million) bar applies to venture holding companies. Once it qualifies, the rules get stricter, not looser. It must own at least 30% of a listed subsidiary, or 50% if unlisted. It cannot carry debt beyond twice its own equity. It generally cannot own more than 5% of any non-affiliated company and generally cannot own a financial-sector subsidiary at all. The entire design goal is to force a group’s ownership into one clean, auditable parent-subsidiary tree instead of a web only insiders can trace.

Here is the number that matters for reading any specific Korean holding. At the most recent full count (figures as of end-2025, released June 2026), Korea had 173 holding companies nationwide — down slightly from 177 the year before. Of the 102 large disclosure-target groups analyzed that cycle, 51 possessed at least one holding company — but only 47 had converted their entire group structure to run through one. Across all 173 holding companies, those structures controlled 2,357 subsidiary, grandchild, and great-grandchild affiliates, an average of 13.9 per holding company.

Put plainly: fewer than half of Korea’s largest conglomerates run their whole structure through the FTC’s preferred, transparent format. The other half either haven’t converted at all or have converted only part of the group. That split is where the checklist articles stop. It’s also where an actual investor question starts: which side of that line does the company you’re looking at sit on?

Two real examples make the mechanism easier to see than aggregate percentages do. LG Group began converting in April 2001 and had consolidated into a single holding company, LG Corp., by March 2003 — the first of Korea’s largest conglomerates to run a full holding-company structure. SK Group converted in July 2007, splitting its main operating entity into an investment holding company and a separate operating business. In both cases, an investor can look at one filing and see cleanly what the parent owns in each subsidiary.

Samsung, Korea’s largest conglomerate by assets, has not converted. Control instead runs through a chain of cross-affiliate stakes. The Lee family’s largest single holding sits in Samsung C&T, which holds a stake in Samsung Life Insurance, which holds a substantial stake in Samsung Electronics. We looked for a clean, single-source set of current ownership percentages for that chain and did not find one—different outlets report different figures for the same stakes. That gap is its own small lesson: even a subject this closely watched does not always resolve to numbers you can safely repeat. We are naming the structure here, not the disputed percentages. The mechanism is the same type as circular shareholding in spirit, controlling compounding through a chain of affiliate holdings, but it does not technically form the closed loop the 2014 ban targets, because the chain does not close back on itself.

Neither structure is automatically “safe” or “risky” as a blanket rule. A holding-company conversion buys transparency. It doesn’t buy good management, and it doesn’t by itself change the 3.5%–3.7% direct-ownership math from Section 2. What it buys a foreign investor, concretely, is a filing you can actually read, instead of a diagram you have to reconstruct.

5. Board independence: what it changes, and what it doesn’t

This site has already covered the board-independence reform itself in detail: the Commercial Act amendment that passed the National Assembly on August 25, 2025, adding cumulative voting and raising the minimum share of independent directors from one-quarter to one-third. If you haven’t read that part of the Korea discount piece, that’s where the legal mechanics and the compliance dates live. This section adds one structural point that piece doesn’t make, without re-deriving a single date from it.

A higher quota of independent directors changes who occupies seats on the board. It does not, by itself, change who nominates them. Nomination in Korea, as in most markets, still typically runs through the existing board. In a company still built on the ownership structure described in Sections 2 through 4, that means a board a controlling family already has outsized influence over. Cumulative voting is the one mechanism in the 2025 reform that changes this directly — it makes it mathematically easier for a minority shareholder bloc to elect at least one director without the controlling shareholder’s cooperation. But it requires that bloc to actually organize and vote. That’s a behavioral question, not a legal one.

None of that makes the 2025 reform meaningless. A board that’s legally required to be one-third independent is a different board than one that isn’t, and the reform is real. It just answers a different question than “who owns what.” Know which question you’re asking before you treat a board seat as proof that the ownership math from Section 2 has changed.

6. Four traps investors keep falling into

Trap 1: Assuming “holding company” means clean governance and “no holding company” means bad governance. A conversion buys a transparent, legally defined ownership tree. It says nothing about whether the people running that tree make good capital-allocation decisions. Samsung’s unconverted structure and its business performance are two separate questions; conflating them either way is a mistake.

Trap 2: Treating circular shareholding as a solved, historical problem because new formation has been banned since 2014. The ban stops new loops. It didn’t retroactively unwind old ones on a fixed clock. 2026 alone has produced a live enforcement case and a proposed rule closing a fresh workaround, both covered in Section 3. “Banned” and “gone” are different words for a reason.

Trap 3: Reading a higher independent-director quota as proof that board capture has ended. Section 5 covers why a seat quota changes composition, not nomination. Read the related-party-transaction disclosures alongside the board roster, not as a substitute for it.

Trap 4: Extrapolating one company’s structure onto “Korean chaebol” as a monolithic category. LG and SK run clean, converted holding-company structures. Samsung doesn’t. The FTC’s own count shows fewer than half of the 102 largest groups have fully converted. The sector-wide average is a starting point for research. The specific company’s own ownership chart is the answer.

7. FAQ

What is circular shareholding, in one sentence?

Three or more affiliated companies hold stakes in a closed loop—A owns B, B owns C, and C owns A—so a stake in one corner effectively carries voting influence over the whole loop.

Is circular shareholding illegal in Korea?

Forming a new one has been, at the largest tier of business groups (₩12 trillion or more in assets), since a 2014 rule change. Loops that existed before that date weren’t automatically illegal. Enforcement against structures alleged to evade the rule — like the case described in Section 3 — is still active in 2026.

What makes a company a “holding company” under Korean law, and does that guarantee good governance?

It’s a defined legal category under the Fair Trade Act: broadly, a company whose main asset is subsidiary shares, past a ₩500 billion asset threshold, subject to minimum ownership stakes in its subsidiaries and limits on debt and outside holdings. It guarantees a transparent, auditable ownership structure. It doesn’t guarantee good management or capital allocation. Those are separate questions the structure alone can’t answer.

How can I check a specific company’s ownership structure myself?

The FTC publishes its own ownership-chart tool for designated business groups, showing the affiliate-by-affiliate holding structure the agency itself uses for enforcement: the egroup. go.kr ownership-chart service (interface is Korean language). It’s the same category of primary source this piece draws its own figures from.

Does this connect back to the “Korea discount”?

Yes. This is the mechanical layer underneath one of the three structural causes that piece names for the discount. That piece covers the reform timeline: the Value-Up Program, the Commercial Act amendments, the dividend tax cut, and mandatory treasury-share cancellation. This piece covers how the underlying ownership structure actually works, mechanically, independent of any single reform’s timeline.

---

Last updated: August 23, 2026

This article is for informational purposes only and does not constitute investment, tax, or legal advice. It describes general corporate structure and regulatory mechanisms in Korea and is not a statement about any individual company’s current governance quality, risk, or investment merit. Ownership figures, legal thresholds, and the status of pending regulatory matters — including the enforcement case and the proposed rule discussed in Section 3 — can change. Verify current status before making any investment decision.