SK Square Co., Ltd. (KRX:402340) rose 6.12% on Friday, September 4. It closed at ₩1,041,000. It was the first close above ₩1 million. The company's market capitalisation was ₩137.3 trillion at Friday's close. That's roughly $98bn at an approximate ₩1,400 per dollar. SK hynix is the memory maker that makes up almost all of SK Square's value. It rose about 4% on the same day from a prior close of ₩1,596,000.
A 6% day on a stock this size gets attention. The press framed it as the holding-company discount narrowing. Korean coverage on Friday put SK Square's net asset value at about ₩240.5tn. Of that, ₩235.7tn is the SK hynix stake. Against a ₩137.3tn market cap, that is a discount of roughly 43%.
Here is the arithmetic that the headline skips. If SK Square rose 6.12% and its net asset value rose about 4% because SK hynix did, the ratio improved. Price to asset value improved by about two percent. On a discount of 44% before Friday, that gets you to about 43% after. One point. The discount was 46.4% in late March, when the company's own target was to get below 50%. It was 43.8% in early July. It is 43% now. My view is that Friday moved the price and left the discount roughly where it was, and that the discount will not close on days like this. It closes on policy. The policy that matters belongs to SK hynix, not SK Square.
The mechanics are simple once written down. SK Square's value is a stake in SK hynix. When SK hynix rises, the stake rises. If SK Square's price rises by the same percentage, the discount is unchanged. The discount only narrows when SK Square outruns its own asset. On Friday it did so by about two points of relative performance.
That's why the discount has been so sticky through a year in which SK hynix has roughly tripled. SK Square has tripled too, from a 52-week low of ₩149,400 to over ₩1 million. The discount has drifted from the high 40s to the low 40s. The stock has been a leveraged way to own SK hynix. It has not been a way to capture the gap between the two.
Management set a new target in March: discount below 30% by 2028. It also targets a price-to-book ratio above one. The tools disclosed so far include a ₩310bn shareholder-return programme for completion by early 2027. They also include a cancellation on July 31 of 34,388 shares. That's 0.026% of the count. I wrote on August 28 that the buyback is two orders of magnitude too small for the gap, which was about ₩112tn at the time. That hasn't changed.
The reason buybacks can't close it is that SK Square doesn't have the cash. Its cash balance was ₩1.32tn at the end of June, against a ₩103tn gap to asset value on Friday's numbers. First-half operating cash flow was ₩260.8bn. First-half net income was ₩27.05tn. Almost all of it is equity-method accounting for SK hynix's profit that never arrives as money.
What does arrive is SK hynix's dividend. That line sets how much SK Square can ever return to its own holders. By extension, it sets what its shares are worth as a stream of cash rather than a claim on a stake that will never be sold. SK hynix resolved on August 19 to buy back and cancel ₩40tn of its own stock. That raises SK Square's percentage ownership but sends it no cash. A larger SK hynix dividend would send cash.
That is where Friday's policy backdrop matters. The expansion of separate taxation on dividend income (배당소득 분리과세) lowers the tax on dividends from high-payout companies. It gives SK hynix's own shareholders, SK Square among them, a reason to push for a higher payout ratio. If SK hynix moves from a buyback-heavy return policy toward a dividend-heavy one, SK Square's cash income rises. Then the discount has a reason to close. If it stays buyback-heavy, SK Square keeps booking paper profits and the discount keeps drifting.
On reported numbers, SK Square trades at about 2.5 times book equity of ₩55.0tn and about 4 times trailing net income. Neither figure means much. The book value carries SK hynix at equity-method cost, far below market. The earnings are the equity-method share of a memory maker at the top of its cycle.
The only valuation that matters is the discount to the market value of the stake. That is 43%. For comparison, the market prices SK Inc., one layer further up, at roughly the value of its SK Square stake alone, which I covered separately. The discounts compound going up the chain.
The obvious risk is that SK hynix's stock falls. SK Square would fall with it, probably by more. A wide discount tends to widen further in a downturn as hopes of it closing fade. Memory prices are at a cyclical high. Samsung's own capital allocation this year favours banking cash rather than building fabs. That suggests the industry's leader doesn't expect the high to last.
The second risk is that the 2028 target is a target. A discount below 30% requires either a much larger buyback, funded by something SK Square doesn't have. Or it requires a much larger dividend from SK hynix, which SK Square doesn't control. Korea's corporate value-up program (기업가치 제고 계획) rewards companies that publish targets. It doesn't penalise them for missing.
The third is dilution at the asset level. SK hynix's July Nasdaq ADR offering cut SK Square's stake from 20.5% to about 20%, which I covered last week. Further raises at SK hynix would do the same. The buyback at SK hynix pushes the other way. Where the stake settles by year-end is a number worth checking in the third-quarter report.
SK hynix's next dividend decision is the real event for SK Square. The memory maker's board sets its return policy annually and adjusts quarterly. If the mix shifts toward cash dividends under the new tax regime, SK Square's income statement gains a line that is actually cash. Then the discount has a mechanism to close. The filing to watch is a cash dividend decision (현금ㆍ현물배당결정) from SK hynix, not anything SK Square files itself.
The second is SK Square's own third-quarter report in mid-November. It will show the stake percentage and the cash balance. If cash rises beyond the ₩310bn already committed, the buyback can grow. If it doesn't, the 2028 target rests entirely on the parent company's cousin one level down.
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.