This report's gauges (§07) — live widgets update with new observations after publication · Trigger 2 Watch 3 Wait 3 · Each needle has its own date
On Friday, August 21, Samsung Electronics disclosed a KRW 90–110 trillion shareholder-return plan. On the next trading day, August 24, its shares fell −8.70%. An announcement to raise the company's value sent its price down. This year, foreigners have net sold KRW 163.4 trillion in net selling of Korean stocks (as of 8/28), yet the won has+11.7% appreciated, and just two stocks lifted the index. Value-up (Korea's corporate value-up program) was an instruction to raise corporate value. What is actually unfolding is a fight over who sets a company's value.
01
The Puzzle of August 24
Announcement: fair disclosure on 8/21 (Fri) · Market response: 8/24 (Mon). Read the configuration, not just the announcement
Remaining resources: 50% of 3-year FCF minus KRW 29.3 trillion already spent (dividends 20.9 + buybacks 8.4). Depending on earnings and investment, this can change.
"The same day's KRW 15 trillion buyback expands returns"
The acquisition's purpose is"employee share compensation" (53,285,968 shares, market purchases 8/24–11/21). The filing says share-compensation spending is deducted from FCF, and the KRW 90–110 trillion is separate. Adding them double counts.
"Shareholder-return disclosure on 8/24"
The shareholder-return disclosure was on 8/21. 8/24 was the KRW 15 trillion buyback's acquisition start date.
The answer had not emerged by week's end. Samsung Electronics stayed in the lower-to-mid KRW 260,000s for the next four days, ending August's last trading week without recovering (8/28 intraday: KRW 263,500, −0.94%, reported). The gap from the upper KRW 270,000s–280,000s before the announcement remained.
This Report's Rules — The Same as Before
Several individual companies appear, but this report does not recommend buying or selling any security. It does not establish any company's intent. The strongest claim we can make is that the configuration fits. Figures state their confirmation dates; ★ marks an original source awaiting recheck.
02
After Two Boards — Cancellation's Boomerang
Canceling treasury shares raises remaining ownership percentages. That arithmetic reaches the financial-industry law's 10% limit
The answer lay outside the announcement. Canceling treasury shares reduces shares outstanding, so every remaining shareholder's ownership percentage rises automatically. That includes holders of Samsung Electronics shares, Samsung Life and Samsung Fire, which hold Samsung Electronics shares. But the Act on the Structural Improvement of the Financial Industry bars financial companies from holding more than 10% of a nonfinancial company. The larger the cancellation, the closer Life and Fire get to the limit. At the limit, they must sell. (The same day's KRW 15 trillion buyback is outside this calculation because it is for compensation and will not be canceled.)
1
What the market calculated
A record cancellation plan can also mean a record warning of selling by Life and Fire. On 8/24, Samsung Life recorded a deeper fall than Electronics (−8.70%), at−10.81%, (★Correction 2026-09-02: observations from two Naver daily-price routes show−13.09% and a close of KRW 285,500. The basis of the reported −10.81% remains unconfirmed. Its fall was still deeper than Electronics'.) Brokerages pointed to the same place that day: "Life and Fire's combined Samsung Electronics stake is approaching the 10% level" (DS Investment & Securities, quoted in reports; translated). This suggests the market read this arithmetic rather than the announcement.
2
There is already a precedent
In March this year, Life and Fire sold Samsung Electronics shares in trillions of won through block deals. Cancellation → higher ownership → limit → selling is not a new circuit.
The limit of what we can claim
There is no basis to say returns were designed around the financial-industry law's limit. What we can say is that mandatory cancellation (3rd Commercial Act amendment) and the holding limit (financial-industry law) squeeze the same company from opposite sides. That configuration exists, and 8/24 prices fit it. The composition of returns is still unknown. The original states only that Q3's KRW 30 trillion is entirely cash dividends. The cancellation amount will be decided by the end-October and end-January boards.
03
Three Doors — Cancel, Explain or Leave
The 3rd Commercial Act amendment's question to every listed company. Five responses, three doors
The Question First — Main Points of the 3rd Commercial Act Amendment
Based on National Assembly passage on July 3★ · Detailed implementation dates await an original-source recheck
Provision
Details
New treasury shares
Must be canceled within 1 year of acquisition
Existing treasury shares
Must be dealt with within 6 months + 1 year = 18 months after implementation
Exception
Annual shareholder-meeting approval can extend retention
Sectors with foreign-ownership caps
3-year grace period
95% squeeze-out
A controlling stake of at least 95% permits compulsory purchase of remaining shares, widening the voluntary-delisting exit
The Cancellation Rush Is Observable — Monthly Decision Disclosures (All DART Stocks)
2024-01 to 2026-08 · Dashed = each year's monthly average · 2026 monthly average 52.0 decisions = 3.1x 2024 · August ongoing (tallied 8/30)
Decision disclosures lag execution (acquisition usually takes 3 months). Amounts require opening each original filing, so we count disclosures only. Cancellation, acquisition and trust disclosures must not be added together. As with SK hynix's KRW 40 trillion in August, cancellation and acquisition decisions can appear as a pair on the same day. Adding them doubles the count.
Treasury shares have long been an all-purpose tool in Korean governance: defending control, transferring shares to friendly holders, or simply holding them. Now companies have been told to cancel them within 1 year. In January–August alone, cancellation decisions reached 416 disclosures (tallied 8/30), a monthly average three times 2024's. Companies are lining up at the widest, first door. But it is not the only door.
1
Door ① Cancel — Comply
This is the widest door, the one intended by Value-up. Samsung's KRW 90–110 trillion formally belongs here too. But as §02 showed, inside it lies the financial-industry law's threshold.
2
Door ② Explain — Transform, Evade or Hold Out
Companies can turn treasury shares into exchangeable bonds (EBs) to avoid cancellation (the FSS responded with stronger disclosure★), transform through holding-company conversions or restructuring that changes the regulatory landscape, or hold out using annual shareholder approval. All explain why shares are not canceled.
3
Door ③ Leave — Voluntary Delisting
If staying listed costs more, a company can give up its listing. Inside this door are Three paths: buy up shares, exchange them, or dilute them. This year produced actual cases on all three routes. Two companies completed the journey:§04.
04
Three Exit Routes — Two Completions, One Brake
Share accumulation, exchange or dilution — actual voluntary-delisting routes after the Commercial Act amendment
Inside §03's third door are three more routes: share accumulation (tender offer), a share exchange (comprehensive share exchange), or dilution until holders give up (a large rights offering, checked by authorities). All three routes produced actual cases this year: two completed (Shinsung Tongsang and Shinsegae Food), one checked (SK D&D). All involve naming a price, but minority-shareholder outcomes differ by type, so we do not combine them.
Route ① Buy Up Shares — Tender-Offer Accumulation (Original Source: Tender-Offer Results)
★ = original source awaiting recheck · No recommendation to trade any security
A Korean route: neither 90% nor 95% is needed. What is needed is a shareholders' special resolution: two-thirds
Stage
Details
1. Tender offer
E-Mart, through a tender offer last December, acquired 66.45%, putting the special-resolution quorum within reach
2. Share-exchange resolution
Comprehensive share exchange (Commercial Act, Article 360, subarticle 2): remaining holders' Shinsegae Food shares are compulsorily exchanged for E-Mart shares (0.5031313 shares per 1 share). The exchange price is the KRW 48,729 reference price plus a 3% premium, giving KRW 50,191
3. Regulatory gate
The FSS required an amended filing. E-Mart held a shareholder meeting and responded by raising the planned appraisal price 30%, from KRW 48,876 to KRW 63,348. The filing cleared after that
A legal route around the squeeze-out (95%) and KOSDAQ practice (90%) thresholds, using two-thirds. Minority holders are not simply expelled; they become shareholders of the parent, while their ownership in this company ends. Appraisal claims: 585,898 shares · 1,132 people · KRW 36.7 billion (paid 2026-07-15). Of these, 22 people with 5,787 shares sought a price adjustment
Route ③ Dilute — Then Authorities Applied the Brake (SK D&D, Ongoing)
After two rejected tender offers, a reversal: lower the value of others' stakes instead of raising the offer
Stage
Details
1. Failed tender offers
Largest shareholder Hahn & Company (PEF, stake 47.42%): last October, an offer of KRW 12.75 thousand per share and another attempt; both missed their targets
2. Change of direction
7/28 2.4x existing shares (44.68 million new shares) approved in a rights offering, at an issue price of KRW 3.06 thousand · KRW 136.7 billion, all for debt repayment
3. The price's arithmetic
KRW 3,060 is ⑴ this share of the tender price: 24% ⑵ 0.1x P/B (reported), roughly one-tenth of book value per share. ⑶ Proceeds of KRW 136.7 billion equal this multiple of market capitalization at the correction request (about KRW 78 billion; share price KRW 4,180): 1.8x. A nonsubscribing holder transfers that share to the new subscribers
4. Market interpretation
The next day: lower price limit. With 0% subscription, the largest shareholder's stake could reach 92.62%, raising reported suspicions of a delisting shortcut. The debts are real, however: bonds maturing in September, KRW 107 billion; October, KRW 62 billion; and KRW 74.7 billion in joint guarantees
5. The brake
8/5: a shareholder coalition with 4.7% (875,164 shares) petitioned. The FSS requested corrections the same day. The filing's effectiveness was suspended; no correction within 3 months means withdrawal. → 8/6: upper price limit → 8/24: minorities with 3.63% demanded an extraordinary meeting. Outcome unknown
Comparing Two Routes — The Same Door, Different Prices
Both controlling holders named an exit price; both received regulatory correction requests. Their responses diverged
Failed tender offer → new shares at 2.4x existing stock
Minority options
Receive E-Mart shares (still listed and liquid) or exercise appraisal rights
Put in more cash (subscribe) or accept dilution: the value of staying is forcibly lowered
Pricing basis
Statutory formula (reference price +3% premium); the appraisal price could rise
No regulatory issue-price formula blocks even a price at one-tenth of net assets★
Response to the correction request
Raised the price(appraisal price +30%) → cleared → completed
Has not yet answered → effectiveness suspended; outcome unknown
Again, we do not establish intent. The company said debt repayment, the market read ownership arithmetic, and authorities returned the filing. We record only how these facts fit together. §05 examines the institutional structure.
Completed
Shinsegae Food delisted 8/11 (share exchange) · Shinsung Tongsang delisted 2025-07 (tender offer) Completed examples of two routes
~11/5
SK D&D — 3-month deadline for an amended filing (request date confirmed: 8/5) No correction means withdrawal · minorities' extraordinary-meeting demand (8/24) ongoing
end-October
Samsung Electronics board — details of Q3's ~KRW 30 trillion cash dividend Final decision on the rest at end-2027-01
The FSS can require explanations, not regulate prices. Yet that tool raised Shinsegae Food's price 30%
Authorities used the same tool in both cases: a correction request for the securities registration statement. It lets them return a filing if material statements appear false or missing. Effectiveness is suspended; no correction within 3 months means withdrawal. What matters is what this tool cannot do. Authorities cannot order a change to the exchange ratio or a higher issue price. As with tender offers, no formula blocks a rights-offering issue price. Authorities can require only this:"Make the filing explain why this price and why this method" (translated).
1
Shinsegae Food — An explanation request moved the price
After the FSS's correction request, E-Mart answered with a price rather than paperwork. After meeting shareholders, it raised the planned appraisal price to KRW 48,876 → KRW 63,348 (+30%), and the filing cleared. The authority's demand for explanation, without power to order a price, acted as a price-negotiation table. A deal with a statutory formula (reference price plus premium) could complete its explanation by raising the price slightly.
2
SK D&D — The explanation has not yet held
A filing to repay debt by raising 1.8x market capitalization at one-tenth of net assets★ must still answer,"Why this method and this price?" even when debt maturities are real, compared with alternatives such as asset sales or refinancing. A higher issue price disrupts the financing structure; a paperwork answer faces stark arithmetic. This is why the correction request has acted as a de facto veto. (This interpretation is ours; authorities have not disclosed their reasons.)
In that gap, negotiations concern not price but the denominator and the direction of the quote.
Telcoware Tendered Shares: 964,876 — Divided by Planned Purchases
41.4%
The tender-offer report's denominator (2,332,438 planned shares): "4 in ten tendered"
The Same Number — Divided by Total Shares Issued
10.4%
The reported denominator (9,241,038 issued shares): "only 10% tendered"
SK D&D — The Price Named to Leave (Tender Offer, Last October)
KRW 12.75 thousand
Hahn & Company's tender price. Both attempts missed the target
Then the price named to enter (rights-offering issue price)
KRW 3.06 thousand
The same company's shares were priced at one-quarter as much. The FSS's correction request applied the brake
Thresholds and Their Complements — Where Minority Bargaining Power Lives
At Viol, minority holders fought over the threshold rather than the price. Through ACT, they once assembled 7.26%, targeting the 10% complement of KOSDAQ's 90% voluntary-delisting practice. As the coalition fell to 5.97%, its leverage weakened. At SK D&D, a 4.7% petition coincided with the regulatory correction request. Commercial Act squeeze-out: 95% (complement 5%); KOSDAQ practice: 90% (complement 10%); special resolution: two-thirds (complement one-third). Every threshold has a complement, and that complement is minority bargaining power. But as Shinsegae Food showed, one tender offer can cross the two-thirds threshold, and as SK D&D showed, a rights offering can dissolve the complement itself. The authority's demand for explanation is currently holding that move in check. The Value-up paradox: rules intended to raise corporate value created a table for bargaining over it.
06
Two Faces of the Same Market
Foreigners sold KRW 163 trillion; the won rose 11.7%. A summary of the 8/26 assessment
Annual Foreign KOSPI Net Buying (2016–2026, KRW tn)
2026: January–August cumulative (as of 8/28, ongoing). This chart shows record net selling's scale itself of record net selling. Aggregated from the original daily series; monthly and annual totals reconcile (rounding ±0.1, as of 8/28)
The combined 2 pandemic years, 2020–2021, saw −KRW 50.6 trillion. Foreigners sold three times that in 8 months. The 5 worst daily net-selling records all occurred in 2026. This is the scale of a structural shift, rather than a correction.
2026 Monthly Net Buying (KRW tn)
Total −KRW 163.4 trillion (as of 8/28) · ★Won exchange-rate overlay will follow after ECOS becomes the canonical series · Over the same period, the won's currency-specific appreciation was+11.7%
In May–June alone:−KRW 88.4 trillion. Yet the won kept appreciating during that period. This paradox began the assessment.
The Assessment (Finalized 2026-08-26) — Three Layers
Canonical source: notes/2026-08-26-krw-verdict.md. This report's won discussion follows those rules
Layer
Details
Confirmed
Won-specific appreciation:+11.7% · DRAM unit prices → current account: r=+0.819 (causality upgraded through mechanism + statistics) · daily foreign flows → exchange rate: r=−0.273, R²=7.5% (more than 92% of won movements lie outside trading flows) · net selling −KRW 162.7 trillion
Judgment
The main driver of appreciation is the memory-driven current account surplus: dollar inflows from DRAM unit prices+240% YoY (elimination + quantity assessment, not formal statistical proof)
Prohibited
Claiming statistical proof that the current-account surplus lifted the exchange rate; turning trading flows' dominance over exchange rates into a law; causal explanations of a single day's move
Falsification conditions are monitored automatically: reject the assessment if the won keeps strengthening after unit prices turn down. Today's reading: joint strength (unit prices: 2026-07 MoM +6.4%, accelerating; won: −5.48% appreciation over 20 business days). This fits the assessment. The gauge does not predict the exchange rate; it monitors whether the assessment survives.
1
One face — A selling market
Foreigners: −KRW 163.4 trillion (8/28). As §04 showed, negotiations to end listings are unfolding within the market. These forces name a low company value.
2
The other face — A market pulled higher
Yet two semiconductor stocks lifted the index, and the dollars they earned pushed the won up +11.7%. The force naming a high company value is effectively memory alone.
07
gauge panel
Eight needles — trigger 2 · watch 3 · wait 3
This table is the gauge panel above the title. Like #13–#15, it is a live widget that updates with new observations after publication. Publication values remain in this table. The first four slots are Korea: cancellation decisions, the K-exit registry, cumulative foreign flows, and won–memory co-movement. The last four come directly from #15. Treasury buyback, TIC and pipeline needles keep updating live on #15's gauge panel.
Versus the base year's monthly average:≥3x → trigger; ≥2x → watch. Decision disclosures lag execution (acquisition usually takes 3 months). A 'decision this month' does not mean 'buying this month'. Counts only, not amounts.
≥3 ongoing cases → trigger; ≥1 → watch. Minority-shareholder outcomes differ by type. Tender offers (a premium), market-price mergers (0 premium) and appraisal claims on business transfers (statutory formula) must not be combined. tier 'Secondary sources' are excluded from the tally.
Current-year cumulative flows below the historical low (since 2016) → trigger; below half that level → watch. KRW execution basis cannot be directly compared with the balance of payments (dollars, net investment). Different numbers for'"foreign equity"' are normal.
④
Won–memory co-movement
Joint strength · unit price +6.4%MoM / won −3.2%20d [Wait · 08.28]
4 quadrants: co-movement (strength/weakness) fits the assessment → wait; the 2 divergence cases → watch; assessment falsification condition 1 ('continued strength despite falling unit prices') → trigger. Mixed frequency: monthly unit prices × daily exchange rates. Outside a unit-price update month, only the exchange rate has moved.
⑤
Real−BEI divergence
Real 30y -4bp / BEI -3bp (5 days) [Wait · 08.28]
BEI 5-day change |10bp| or more → trigger (intervention reaches expectations); divergence |5bp| or more → watch. While only real yields move, this is the fingerprint of supply-and-demand intervention.
⑥
Bill share
22.2% (band +2.2pp) · flow 81% [Watch · 07.31]
Bills outstanding as a share of marketable debt. Above the band's 20% upper bound → watch; ≥25% → trigger. The denominator is marketable debt, not total debt. The flow share (bills' portion of monthly net issuance) is an earlier-moving measure, shown alongside rather than on the same axis.
21-day change ≤−30bp → trigger; negative ERP → watch (speed becomes valid after enough observations). ERP is daily; the crack signal is weekly. The chip date refers to ERP; the crack date appears in the value. Do not read them as same-date values. While 13-week/4-week changes are None, accumulation is insufficient: do not infer a direction.
08
Korea Is Bargaining
#14 The US: Those Who Came to Set the Price · #15 Compute: A New Product's First Quote · #16 Korea —
In #14, we saw the US Treasury manage the price of money (interest rates). In #15, we watched the value of compute arrive on the auction block for the first time. Korea is doing neither. Around the value of a company, boards and minority holders, controlling shareholders and foreigners, rules and markets sit across from one another with different scales in a negotiation.
Value-up was an instruction to raise value. What it actually created was a table. Mandatory cancellation collides with the financial-industry law (§02). Three doors open (§03), with three routes through the exit (§04). Authorities cannot name prices, only demand explanations; yet those demands can move prices (§05). Meanwhile, memory alone pulls the value of the whole market (§06). Prices are not declared; they are named. For the first time this year, Korea's market is learning this across its entire institutional framework.
Two things to watch next: SK D&D's next move— whether it refiles or lets the 3-month deadline pass—will decide if the dilution route opens or closes. Samsung's end-October board will set the first reading on the largest scale.
Boboo from Sky
This report is for information only and does not recommend buying or selling any security or product. Individual companies are cases illustrating institutional change. Statements about their intentions go no further than assessing whether public configurations fit. Figures are snapshots as of their stated dates. ★ marks an original source awaiting recheck. Investment decisions and responsibility rest with the investor.
