LG Corp. (KRX:003550) reported total equity of ₩30,334,813,000,000 at the end of December 2025 and ₩32,722,423,000,000 at 30 June 2026. Book value grew ₩2,387.6bn in six months.
Net income across those six months was ₩912.8bn.
The reconciliation is almost exact. Total comprehensive income for the half was ₩2,770.5bn. Subtract the ₩382.9bn of dividends paid and you get ₩2,387.7bn — within a rounding error of the equity movement.
So the company's book value grew by two and a half times its earnings, and the difference came from other comprehensive income: gains recognised directly in equity that never appear on the income statement.
The pattern runs through the whole history.
FY2022: net income ₩2,115.8bn, comprehensive income ₩4,964.1bn — a gap of ₩2,848.3bn. FY2024: net ₩790.5bn, comprehensive ₩1,815.5bn. FY2025: net ₩1,000.1bn, comprehensive ₩2,353.9bn.
Across FY2022 to FY2025, LG reported cumulative net income of roughly ₩5.32tn and cumulative comprehensive income of roughly ₩10.62tn. Half of what accrued to shareholders never went through profit.
For most industrial companies other comprehensive income is a small item — pension remeasurements, a bit of currency translation. At LG it is the main event.
A holding company's balance sheet is other companies, and LG's is more so than most.
Total assets were ₩33,982,889m at the end of FY2025, of which property, plant and equipment was only ₩1,656.4bn. Non-current assets of ₩28,285.1bn are overwhelmingly investments. Note also that LG Chem on its own reported total assets of ₩112.2tn at 30 June — several times LG Corp's entire balance sheet — which tells you that the major listed affiliates are carried as investments rather than being line-by-line consolidated into these statements.
That structure is what produces the OCI. Under the equity method, an investor recognises its share of an associate's other comprehensive income directly in its own equity, so every currency translation gain or loss and every pension remeasurement at LG Electronics, LG Chem and the rest passes through to LG Corp's book value without touching its profit. Financial assets designated at fair value through OCI — holdings the company neither consolidates nor equity-accounts — get marked to market with the gain landing in equity too.
None of this is unusual accounting. It is what IFRS prescribes for exactly this kind of company. The consequence is that LG's reported earnings describe a small part of what happened to shareholders' capital in any given period.
The obvious objection is that I am describing a company whose book value goes up for free. It does not.
In the second quarter of 2025, LG reported net income of ₩244.2bn and total comprehensive income of negative ₩374,528m. Other comprehensive income was a loss of roughly ₩619bn in a single quarter, and total equity fell from ₩28,929.0bn to ₩28,554.4bn.
So the same mechanism that added ₩1.86tn in the first half of 2026 subtracted ₩619bn a year earlier. A currency move at the affiliates, or a repricing of a held stake, does both.
That volatility is why quarterly book value at this company should be read as a range rather than a level, and why a price-to-book multiple computed at one date can look meaningfully different a quarter later without anything happening at the operating businesses.
Two practical consequences.
The first is that earnings-based multiples are close to useless here. LG's FY2025 net income of ₩1,000.1bn on a market value near ₩16.8tn — derived from the 0.9% yield stated on the 27 August dividend filing and the 151.19m common shares identified in it — implies about 17 times earnings. That number tells you almost nothing, because the earnings figure omits half the economics and because holding company income is a residual of what affiliates choose to distribute.
The second is that book value, while more meaningful, is still not what most people think it is. Equity of ₩32.7tn is the accumulated carrying value of long-held stakes under the equity method — original cost plus decades of retained profits and OCI — not their market value. It can sit above or below what the stakes would fetch, and for holdings acquired generations ago it is usually below. So comparing a ₩16.8tn market capitalisation to ₩32.7tn of book is not a sum-of-the-parts calculation, and should not be presented as one.
Retained earnings are the one line that behaves conventionally: ₩21,396.5bn at the end of December, ₩21,675.0bn at 30 June. Up ₩278.5bn, which is roughly net income less dividends. That line has barely moved in three years — ₩21,301.1bn at the end of FY2023 — because LG pays out most of what it earns, with a payout ratio near 69% on average and 74.6% in FY2025.
Earnings out the door as dividends; book value up through OCI. That is the shape of this company.
The strongest counter is that other comprehensive income is not phantom. A currency translation gain at an overseas operation reflects a genuine increase in the won value of assets the group's affiliates own. A fair value gain on a held stake reflects what someone would pay for it. Both would convert to cash on a sale.
There is also a fairness point about the direction of travel. LG has been criticised for a decade for trading below the value of its parts. If OCI is inflating book value, then the discount to book that critics cite is being measured against a number that is itself rising for reasons management cannot claim credit for. The two arguments cannot both be made at once.
And for a holding company, book value is genuinely the better anchor than earnings. The business is owning things. What those things are worth is the question, and OCI is part of the answer.
The composition of other comprehensive income in the half-year report notes. Share of associates' OCI, currency translation and fair value gains on financial assets are disclosed separately. If the ₩1.86tn was mostly currency translation, it reverses when the won strengthens and should be discounted heavily. If it was fair value gains on stakes the company intends to hold, it is closer to a real increase in worth.
Second, the investment note listing each affiliate, the percentage held, the carrying value and — where the affiliate is listed — the market value. That comparison is the holding company discount properly measured, and everything else written about this stock depends on it.
Third, whether management's 8-10% return on equity target for 2027 is defined on net income or on comprehensive income. The two produce very different companies, and at a business where the gap between them is this large, the definition is not a technicality.
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.