LG Corp. (KRX:003550) no longer owns any of its own shares. It held about ₩500bn of treasury stock, cancelled half of it — 3.029m shares — in September 2025, and retired the remainder in May 2026 in a tranche worth around ₩350bn.
Under Korea's amended Commercial Act, effective March 2026, cancelling repurchased treasury stock is now mandatory rather than optional. LG got there before the deadline and did it completely.
Then on 27 August the board declared an interim dividend of ₩1,000 per share on both common and preferred, totalling ₩154,203,203,000, with a record date of 11 September and payment on 23 September. Four outside directors attended and none were absent.
The stated yield on the common is 0.9%. Work backwards and the shares sat near ₩111,000 in late August, implying a market value around ₩16.8tn on the 151.19m common shares the dividend filing identifies.
Total equity at 30 June was ₩32,722,423,000,000.
Half of book, on the face of it. But the comparison needs unpacking, and the way it is usually done is wrong.
LG Corp's consolidated total assets were ₩33,982,889m at the end of FY2025. LG Chem alone reported total assets of ₩112,209,782m at 30 June 2026. If LG Corp consolidated its major listed affiliates, its balance sheet would be several times larger than it is.
It does not. Property, plant and equipment is only ₩1,656.4bn of the ₩34.0tn. Non-current assets are ₩28,285.1bn, and the overwhelming majority of that is investments — the carrying value of stakes in LG Electronics, LG Chem, LG Uplus, LG H&H and LG Innotek, held at cost plus accumulated share of profits rather than at market price.
So the ₩32.7tn of equity is not a sum-of-the-parts valuation. It is an accounting carrying value for stakes the company has held for decades, and its relationship to what those stakes would fetch is not fixed. Anyone quoting "0.5 times book" for this company is comparing market value to a number that is neither market value nor liquidation value.
The right analysis compares market capitalisation to the market value of the listed stakes, less net debt and holding company costs. That calculation requires the shareholding percentages and current prices of five separate companies, which is why it is done less often than the lazy version.
What is true regardless is that Korean holding companies trade at a persistent discount to the market value of what they own. The discount exists because a holding company's shareholders receive dividends from the affiliates after tax and after the parent's own costs, and because control has historically been exercised for family purposes rather than minority ones.
Both of those are what the current reform wave is aimed at.
Look at what LG has actually done rather than what it has said.
Dividends paid: ₩526.6bn in FY2023, ₩550.1bn in FY2024, ₩745.6bn in FY2025. The first half of 2026 already shows ₩382.9bn. Against FY2025 net income of ₩1,000.1bn, the FY2025 figure is a 74.6% payout, and the company reports an average payout near 69% across 2021 to 2025.
For a holding company whose own income is largely dividends received from affiliates plus brand royalties, that is close to a full pass-through. LG is functioning as a conduit rather than as a capital allocator hoarding cash.
It has also adopted semi-annual dividends and changed its articles at the March 2024 annual meeting so that the dividend amount is fixed before the record date is set — the Korean disclosure reform that ended the practice of investors buying without knowing what they would receive. The August filing notes this explicitly.
Add the treasury cancellation and this is one of the more complete shareholder-return packages among Korean holding companies. It has not closed the discount.
Management targets consolidated return on equity of 8% to 10% by 2027.
FY2025 net income of ₩1,000.1bn against total equity of ₩30,334.8bn is about 3.3%. The gap to 8-10% is large.
There are only three ways a holding company gets there. Earn more at the affiliates, which LG Corp does not control day to day. Shrink the equity base through buybacks, which it has now largely exhausted by cancelling everything it held. Or change the mix of what it owns.
The third is what the company says it is doing: redeploying resources into artificial intelligence, biotechnology and clean technology — the businesses it groups under "ABC". That is a multi-year reallocation, and for now it means spending rather than earning.
Meanwhile the affiliates have their own problems. LG Chem reported a net loss of ₩977.1bn in FY2025 as battery assets were impaired, and LG Corp's share of that flows through its own income statement.
One pattern anyone modelling this company should know about before extrapolating a good half.
Fourth quarter 2024: operating loss of ₩241.9bn, net loss of ₩311.2bn, with gross profit of negative ₩120.1bn. Fourth quarter 2025: operating loss of ₩421.7bn, net loss of ₩225.9bn, gross profit negative ₩278.9bn — and a tax charge of ₩187.9bn against a pretax loss of ₩38.0bn.
Two consecutive year-end quarters in the red, at a company that earned a 24.0% operating margin across the first half of 2026.
Holding companies take their impairment reviews on investment carrying values at year end, and LG also consolidates operating businesses that settle costs annually. Whatever the specific driver, the pattern is consistent enough that a strong first half tells you less about the year than it would at most companies.
The sceptical view is that no amount of dividend policy fixes the structure.
An investor who wants LG Electronics can buy LG Electronics. Buying it through LG Corp means accepting a layer of holding company costs, a share of businesses you may not want, and the risk that capital gets allocated toward group priorities rather than yours. The discount is the market pricing that, and it has been remarkably stable for two decades across multiple reform waves.
The counter is that this reform wave has teeth the previous ones lacked. Directors now owe a duty of loyalty to shareholders as well as to the company, treasury cancellation is compulsory rather than optional, and the government has tied policy attention to the discount explicitly. LG cancelling its entire treasury holding ahead of the deadline is evidence a large group is responding rather than waiting.
The investment note in the FY2026 annual report, which lists each affiliate, the percentage held and the carrying value, alongside the market value where the affiliate is listed. That table converts a vague discount into a number, and it is the single most useful page in the document.
Second, the ROE trajectory. Management has committed to 8-10% by 2027, which is one more year. Anything above 6% in FY2026 would make it credible. Continued low single digits would make it a slogan.
Third, what happens now that the treasury is empty. LG has used up its cheapest tool. The next return decision has to be either a higher dividend — from a base already near 70% of earnings — or open-market repurchases funded with cash. Which one the board chooses will say whether the value-up programme has a second act.
kstock reads DART every morning and writes up what moved — the contract, the buyback, the number that does not add up. The daily post and a Saturday roundup, by email.