On September 2, 2026, the KOSPI plunged 3.99% from the 6,560 level while foreigners recorded their largest net selling in eight months. Data revealed that entities classified as "other corporations" accumulated net buying of over one trillion won for ten consecutive trading days. These entities were identified as Samsung Electronics and SK Hynix. The combined share buyback amount approved by their boards of directors is 55 trillion won. The market calls this volume a "safety net," yet simple calculations of the purchase speed indicate the disclosed volume will be exhausted by early to mid-October. The question arises regarding what will support the KOSPI floor once the safety net is consumed.
This issue extends beyond simple supply and demand calculations because the Korean stock market in September 2026 relies on defense mechanisms after losing structural buying entities. Since the share buyback system was activated twenty years ago, it has mostly functioned as a short-term defense tool during periods of poor performance. The pattern of the market failing to find its own equilibrium point after the buyback volume is depleted has repeated. September 2026 represents the latest version of this cycle, and the October depletion point is viewed as the dividing point where this repetition is either proven or disproven.
What Happened
The starting point was the SK Hynix board's resolution on August 19 to cancel the entire 40.43 trillion won of share buybacks. Samsung Electronics also disclosed a plan to buy back over 15 trillion won of shares in the same month, bringing the combined total to 55 trillion won. The promise by the second and third largest market capitalization companies to purchase and cancel or hold shares long-term is typically a card played only when the board can increase per-share value at a lower cost than increasing dividends. Celltrion joined in on-exchange purchases in September, broadening the axis of supply and demand defense to the top three companies. However, the substantial weight remains concentrated over 95% in the two semiconductor firms.
The direct catalyst that shook the market was the sharp drop on September 2. Following foreign net selling exceeding two trillion won the previous day, institutional investors showed net selling in the trillion-won range, and the KOSPI fell 3.99% in one cycle. This trend is interpreted as the result of overlapping factors including the repricing of the US Federal Reserve interest rate path since late August, the weak won, and concerns over slowing export demand. On the same day, the net buying balance of other corporations recorded over one trillion won for nine consecutive trading days, interpreted as the two companies' buyback volume entering the exchange.
Simple calculations of purchase speed indicate the depletion point converges between early and mid-October. Dividing SK Hynix's 40 trillion won by the daily average purchase size based on trading volume and converting Samsung Electronics' 15 trillion won to the same axis yields a scenario where the remaining volume is consumed within 20 trading days. This figure assumes that actual on-exchange execution speed follows the disclosed amount directly after the board resolution, though the possibility remains that purchase speed will be adjusted based on intraday price fluctuations.
Structurally, share buybacks are executed through on-exchange closing price purchases or off-exchange block trades, with a daily execution limit restricted to within 5% of total issued shares. If both companies raise this limit to the maximum for execution, the depletion point is pulled forward. If they adjust speed to about 70% of the limit, the depletion could be pushed to late October. Based on currently public information, early to mid-October is read as the most likely depletion window.
Why This Structure Emerged
The motivation for companies to buy back shares is typically to enhance shareholder value, defend management control, and set a price floor during periods of uncertain visibility into performance. In the Korean market of August 2026, all three factors operated simultaneously. As the semiconductor industry entered its second correction phase after the fourth-quarter 2024 peak, allocation of institutional funds slowed due to prolonged guidance gaps for both Samsung Electronics and SK Hynix. Share cancellation is a minimum resistance means that increases per-share value without worsening cash flow, unlike dividends, making it a card the board had little choice but to choose.
The broader background is the structural outflow of foreign funds. A trend of decreasing allocation to Korean semiconductors in global asset allocation portfolios was confirmed from late 2025. In August 2026, this overlapped with dollar strength against the won, causing nominal won returns to lag behind competitor country stocks. Institutions also increased selling volume by reducing IT sector weight by about 2 percentage points during late-August rebalancing. While domestic funds playing the role of the last support pillar for supply and demand is unusual, in this instance, share buybacks are replacing that support.
Celltrion's September on-exchange buying participation is seen as a separate matter but belongs to the same axis in supply and demand logic. Since the bio sector has low correlation with semiconductors, Celltrion's buying helps diversify sector concentration risk and broadens the signal that "large enterprises are buying shares." However, Celltrion's buyback size is only one-tenth of the two major firms, so the substantial axis of supply and demand defense still relies on Samsung Electronics and SK Hynix.
Macro indicators show third-quarter 2026 export data slowed by about 4% compared to the second quarter, with fundamentals in non-semiconductor industries also weakening. While share buybacks appear to defend the KOSPI floor, they are actually concentrated on raising the prices of the two top market capitalization stocks. The KOSDAQ and small- to mid-cap stocks receive little direct benefit from this defense. This asymmetry must not be overlooked as a seed for volatility expansion after October.
How the Market Reacted
The KOSPI index traded in a 6,500-6,580 box range from September 3 to 5 after the sharp drop on September 2. Daily fluctuations in this period converged within ±1%, and the volatility index K225 VIX fell about 20% from 28 to 22. It is possible to read that buyback volume absorbed selling pressure and created downside rigidity. Conversely, the cumulative foreign net selling balance reached the four-trillion-won range during the same period, indicating the selling pressure itself had not slowed but was being absorbed by buying volume.
By stock, the daily average trading value of Samsung Electronics and SK Hynix increased by 35% compared to August, while small- and mid-cap semiconductor materials and equipment stocks in the same sector fell an additional 5-8%. The structure shows buyback benefits concentrating in the two top market cap stocks and intra-sector supply and demand being redistributed according to the two giants' buying volume. The fluctuation gap between the KOSPI 200 index and the KOSDAQ widened from an August average of 1.2 percentage points to 2.4 percentage points in September, with top market cap defense and small-cap weakness proceeding simultaneously.
By sector, the finance and retail sectors, excluding IT and semiconductors, showed relative strength with a 1-2% slight rise in the first week of September. Financial stocks are not directly related to share buybacks, but the sentiment that "large enterprises are defending the market" transferred a safe asset premium to bank and securities stocks. Conversely, sectors with strong thematic characteristics such as secondary batteries and robots showed intensified capital outflows as they did not receive direct benefits from share buybacks.
The reactions in the exchange rate and bond markets are also notable. The won-dollar exchange rate slipped from the 1,400 won range to the 1,415 won range simultaneously with the KOSPI sharp drop on September 2, and the 3-year treasury yield rose from 2.85% to 2.92%. The pattern of risk aversion proceeding simultaneously in both stocks and bonds shows the market does not interpret the situation solely through the single variable of share buybacks. If the correlation between these two assets weakens after the October depletion point, the supply and demand conditions for the stock market become even more unfavorable.
Looking Into the Mirror of the Past
The most direct comparison is the March 2020 pandemic shock. At that time, Samsung Electronics and SK Hynix announced share buybacks of over 10 trillion won each, and the KOSPI rebounded 14% from the 1,800 level to the 2,100 level. The common point is that domestic institutions and other corporations maintained buying pressure amid foreign selling. The difference is that in 2020, subsequent liquidity expansion and performance recovery filled the scenario after the buyback volume was depleted. In September 2026, performance guidance remains uncertain and macro downside risks have not been removed, making the buyback volume depletion structurally different.
The second comparison is the Bank of Korea rate cut transition period in April 2023. At that time, share buybacks also supported the KOSPI floor, and as buyback volume decreased after mid-April, the index retreated from the 2,400 level to the 2,300 level. Market participants felt the pattern of "returning to the original position when buying ends," and this time the early to mid-October depletion point throws the same question again. However, while individual stock buying was distributed across 3-4 companies in 2023, 2026 sees high concentration with the two top market cap companies accounting for over 95%, meaning the impact of the depletion point could be larger.
The third perspective is the persistence of the accounting and psychological effects of share cancellation. In 2020, the canceled volume was reflected as a one-time increase in per-share value, after which the stock price moved according to performance. The 40 trillion won cancellation resolution by SK Hynix has a definite short-term effect of raising EPS by 8-12%, but the subsequent stock price direction will be determined by the AI server demand cycle and the HBM4 shipment schedule. It must be viewed separately that cancellation is a short-term per-share value adjustment, not a permanent safety net.
The fourth comparison is the Japanese Nikkei share buyback cycle of 2023-2024. Japanese companies expanded share buybacks for six consecutive months from April 2023 to March 2024, and the Nikkei rose 15% during this period. However, after the buyback volume was depleted in April 2024, the Nikkei experienced an 8% pullback in three weeks. As the Japanese case shows, the gains created by share buybacks are largely returned when the buying volume disappears, and the October 2026 KOSPI is highly likely to apply the same dynamics.
After Depletion, What Remains
After the early to mid-October depletion of the share buyback volume, the KOSPI downside support line effectively disappears without a shift to net buying by foreigners and institutions. If the ten consecutive trading days of net buying by other corporations since September 2 represents the process of consuming this volume, a reorganization of buying entities is required after the second week of October. The core variables for this reorganization depend on the US Federal Reserve's October FOMC minutes and the Bank of Korea's October base interest rate path. If rate cut expectations do not materialize, the probability of foreign net selling continuing in October is high.
The second point of focus is the third-quarter earnings announcement timing for Samsung Electronics and SK Hynix. These typically concentrate from late October to early November. If there is no earnings surprise in a state where buyback volume is already depleted, the probability of selling pressure resuming is high. Conversely, if HBM4 order backlogs are confirmed enough to back guidance upgrades, the arguments of selling forces weaken, leaving room for supply and demand balance to be re-established. Volatility expansion one week before and after earnings disclosure is likely to appear similarly to the April 2023 pattern.
The third factor is the won-dollar exchange rate and the possibility of foreign won reinvestment. The weak won slipping from the 1,380 won range in late August 2026 to the 1,410 won range in September is a factor strengthening foreign selling pressure. If the won recovers to the 1,350 won range after buyback depletion, the psychological barrier for shifting to foreign net buying lowers, but there is a risk of structured selling pressure if it breaks the 1,450 won range. Share buybacks cannot replace the exchange rate variable, and the direction of supply and demand after October will depend more heavily on exchange rates and interest rate spreads.
The fourth factor is whether additional share buyback disclosures will emerge besides Celltrion after mid-October. In March 2020, additional disclosures from LG Chem and Hyundai Motor followed the two initial buybacks in early March and April, creating a second wave of buying volume that cushioned the sharp decline. Whether the same pattern repeats in October 2026 depends on each company's board schedule, but the prevailing assessment is that room for additional disclosures is limited given that 55 trillion won has already been disclosed as the maximum size.
Supply and Demand Dynamics, The Next Move
The largest legacy left by this 55 trillion won share buyback is the reconfirmation of the fact that "structural buying entities are absent in the Korean stock market." Since the share buyback system was activated in 2005, it has functioned only as a defense tool during performance downturns for nearly twenty years. The pattern of the market failing to find its own equilibrium point after the buyback volume is depleted has repeated. The situation in September 2026 is the latest version of this repetition, and the October depletion point becomes the dividing point where this repetition is proven or disproven.
Reflectively, for share buybacks to function as a safety net, structural drivers must exist to bring new buying volume in after the buyback period. In 2020, liquidity expansion and performance recovery played that role, but in 2026, that driver has not yet been confirmed. While SK Hynix's 40 trillion won cancellation raises EPS and Samsung Electronics' 15 trillion won buying supports the stock price, the market's real question is who will buy next. Unless an answer to this question appears before the early to mid-October depletion point, the term safety net will become synonymous with a consumable defense layer.
Ultimately, this 55 trillion won could be a bandage covering the structural vulnerability of the Korean stock market. Just as removing a bandage reveals the wound, the KOSPI price discovery process after the October depletion point will proceed purely on fundamentals and macro variables without the external supply and demand of share buybacks. Regardless of the direction of that price discovery, the 55 trillion won of September 2026 will remain as a single piece of experimental data regarding the supply and demand structure of the Korean capital market.
Ending the reflective thought, one point is added. Share buybacks are autonomous corporate decisions, with the board judging the scale and timing. However, the moment that decision assumes the responsibility of supporting the entire market floor, it becomes part of the market structure beyond corporate behavior. Who and how will fill that structural vacuum after the October depletion point is the largest variable determining the second half of 2026 Korean stock market, and the figure of 55 trillion won itself shows the size of that question.
Comments
Post a Comment