S-Oil Corporation (KRX:010950) paid ₩617.1bn of dividends in FY2022. In FY2025 it paid ₩97m. Not ₩97bn. Ninety-seven million won, roughly $69,000, from a company with ₩8.9tn of equity at the time.
That is one of the more complete dividend suspensions a large listed Korean industrial has run without anyone calling it one, and it happened for a specific reason: every won the company could spare was going into a petrochemical complex at Onsan.
Put the two lines side by side across four years.
Dividends paid: ₩617.1bn in FY2022, ₩372.7bn in FY2023, ₩189.3bn in FY2024, ₩97m in FY2025.
Purchases of property, plant and equipment: ₩598.8bn in FY2022, ₩2,185.6bn in FY2023, ₩3,041.9bn in FY2024, ₩3,906.0bn in FY2025.
One line went to zero. The other went up more than sixfold. The Shaheen project, budgeted at ₩9,258bn when the board approved it in November 2022, is the reason for both.
A caveat on how to read these. Korean cash flow statements record dividends when cash leaves, so a given year's figure typically mixes the prior year's final dividend with the current year's interim. The FY2025 figure of ₩97m largely reflects a decision taken about FY2024, and FY2024 was a loss year — the company reported a net loss of ₩193.0bn. Not paying out after a loss is ordinary. What is not ordinary is the four-year trajectory, and that trajectory maps onto the capital spending line too closely to be coincidence.
On 10 August 2026 the board declared an interim cash dividend of ₩800 per share, on both common and preferred, with a record date of 25 August and payment scheduled for 15 September. Five outside directors attended, none were absent, and the company's audit committee is composed entirely of outside directors. The total comes to ₩93,136,511,200, of which ₩3,070,277,600 goes to preferred holders. Treasury shares excluded from the payout: 184,080 preferred.
Working back from those numbers gives roughly 112.6m common shares and 3.8m preferred shares receiving the dividend.
₩93.1bn is a real payment and a small one. Against first-half net income of ₩1,235.6bn it is a 7.5% interim payout. Against the ₩617.1bn S-Oil paid across FY2022 it is 15%. The board is signalling that cash returns have resumed. It is not yet signalling that they have returned to where they were.
That reticence is defensible with Shaheen still in commissioning. Operating cash flow in the first half of 2026 was negative ₩1,090.8bn, absorbed by a ₩2,447.0bn inventory build, and total liabilities of ₩20,156.5bn now sit against ₩10,085.6bn of equity. A company in that position paying out ₩93bn rather than ₩400bn is behaving sensibly.
Worth knowing what is producing the money that funds this. S-Oil reported ₩965.0bn of operating income in the second quarter. On the company's own presentation materials, the lubricants business contributed ₩477.4bn of that — a record quarterly figure for the segment, on base oil spreads the company described as record high.
Roughly half the operating income of a company with a 669,000 barrel-a-day refinery came from its smallest division.
That is not a criticism. Lube base oil is a genuinely differentiated product where S-Oil has invested for years and holds a strong position in Group II and III grades. A high-margin specialty business carrying a cyclical commodity business through a soft patch is the diversification working as intended.
It does mean the refining recovery is thinner than the headline suggests. Gross margin ran 16.2% in the first quarter and 10.4% in the second, and the second-quarter decline came despite revenue rising 26.8%. Strip out lubricants and the fuels business had an ordinary quarter, not a great one. Base oil spreads at record levels are, almost by definition, not a level to extrapolate.
A detail that will look strange to a US reader. The same ₩800 dividend was declared on the common and on S-Oil's preferred shares. The filing states the yield as 0.6% on the common and 1.3% on the preferred, computed on the same one-week average price convention before the board date.
Identical cash, roughly double the yield. That means the preferred trades at something close to half the common price. Working from the stated yields, the common sits near ₩133,000 and the preferred near ₩62,000.
Korean preferred shares — non-voting, usually with a small dividend preference — have traded at wide discounts for decades, and the gap has narrowed at some companies as governance reform advanced. A 50%-odd discount on a stock where the dividend is identical is at the wide end. For an income-focused holder indifferent to voting rights, that is the more efficient way to own the same cash flow, with the caveat that the preferred line is small and thinly traded.
The obvious objection to what I've written is that there is nothing to criticize here. S-Oil committed to a project worth more than its book value, funded it substantially from internal cash rather than diluting shareholders, and suspended distributions for the years the spending peaked. That is textbook capital allocation. The alternative — borrowing ₩9tn to protect a dividend, or issuing equity into a refining downturn — would have been worse on any measure.
Majority-owned by Saudi Aramco, S-Oil also has a controlling shareholder whose interests are aligned with a long project rather than an annual payout, which is unusual and in this case probably helpful. A widely held refiner would have faced far more pressure to keep paying through the build.
The counter to that counter is only about disclosure. A dividend that falls from ₩617bn to ₩97m over three years is a policy change, and it was never framed as one. Shareholders learned it from the cash flow statement.
The FY2026 final dividend, declared alongside full-year results early next year. That is the first payout decision the board makes with Shaheen's construction phase behind it, and the number will show whether the interim was the new base or the first step back toward the old one. Anything that takes the FY2026 total past ₩300bn would say the funding squeeze is over.
The second thing is the FY2027 depreciation charge. Once ₩9tn of assets start depreciating and capitalized construction interest stops being capitalized, reported net income falls even if cash generation improves. A board that anchors the dividend to reported earnings will find itself with a smaller number to work from; one that anchors to cash flow will not. Which of those S-Oil does is the question the next two dividend announcements will answer.
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.