Samsung and SK Hynix Buybacks Beat MSCI Rebalancing Flows

Samsung Electronics and SK Hynix buybacks absorbed MSCI rebalancing outflows on Monday, turning a 2.58% opening plunge in the KOSPI into a 0.46% gain. The daily wrap-up already covers the scoreboard; this post takes apart the one mechanism that decided the session — a collision between forced, calendar-driven passive selling and two enormous, standing corporate bid programs, and what that collision does and does not tell you about where the market goes from here.

A Scheduled Seller Met a Standing Buyer

Monday, August 31 was the closing-price rebalancing day for the MSCI Korea Index August regular review. When MSCI changes its index — this round adds LG Innotek and removes HLB, LG Display, POSCO International, and Samsung Epis Holdings, effective September 1 — passive funds that track the index must trade the changes, and they overwhelmingly do so at the closing auction of the effective-date eve to minimize tracking error. That concentrates a large, price-insensitive order flow into a single print at the end of the day. The source data notes exactly this: passive in-and-out volume piled into the closing single-price auction and amplified late-session volatility.

Normally that mechanical selling pressure lands on a market with no comparably mechanical buyer on the other side. Monday was different. Samsung Electronics is running a 15 trillion won buyback program through around November 21, and SK Hynix a 40 trillion won program through around November 19. A buyback of that size operates like the mirror image of passive rebalancing: it is also largely price-insensitive, also executed on a schedule, and it shows up in the order book day after day regardless of the tape. The session became what one of the source reports called a closing-price war — and the buybacks won it.

The Size of the Breakwater

The scale matters, so here are the figures. As of August 28, the two companies had executed a combined 10,277.7 billion won of purchases — 2,549.4 billion won by Samsung Electronics and 7,728.3 billion won by SK Hynix — leaving roughly 44.7 trillion won of remaining buying capacity across the two programs. That residual firepower dwarfs anything a single index review can throw at the market on one closing auction.

The intraday shape of the session shows the mechanism working. The KOSPI opened down 2.58% at 6,613.58 and sank as low as 6,547.76 before the large-cap semiconductor names, lifted by the buyback effect, dragged the index all the way back to a 6,820.02 close, up 0.46%. Samsung Electronics finished up 1.17% and SK Hynix up 1.27% — on a day when foreign investors were net sellers specifically in semiconductors and electronics. The KOSDAQ, which has no equivalent standing buyer and where heavyweight battery names in the Ecopro group fell 2% to 4%, recovered from an early 2.00% drop to 821.67 but could not turn positive, closing at 834.29, down 0.49%. Same macro news, same opening shock, opposite closes — the presence or absence of the buyback bid is the cleanest available explanation for the divergence.

What the Flow Data Shows

The investor-flow breakdown (press tallies near the close, which can differ slightly from the exchange's final figures) is consistent with that read. Foreigners sold a net of roughly 250 billion won, concentrated in semiconductors and electronics. Institutions sold roughly 140 to 150 billion won, led by brokerage proprietary desks and private funds. Individuals bought roughly 330 billion won, treating the plunge as a dip-buying opportunity. Program trading tells the passive story directly: non-arbitrage programs — the channel through which index-tracking rebalance flows typically travel — sold a net 303.5 billion won, against a tiny 11.9 billion won of arbitrage buying, for an overall program sell imbalance of about 290 billion won. Selling of that shape, on that specific date, is the signature of a rebalance, not of a change in conviction.

Why This Usually Matters

The practical distinction is between flow-driven and fundamentals-driven price moves. Rebalance selling is date-bound: once the closing print settles, the obligation is discharged and the seller disappears. Buyback demand, by contrast, persists — in this case until mid-to-late November. When a date-bound seller pushes prices down and a persistent buyer pushes them back, the textbook implication is that the dip was a liquidity event, not a repricing, and that the persistent buyer remains in place for subsequent sessions while the seller does not.

But the caveat is real, because Monday's opening shock was not purely mechanical. Fed Chair Kevin Warsh's Jackson Hole keynote reaffirmed a preference for tightness — summer inflation prints were better than expected, he argued, but the core inflation trend has not meaningfully improved — reigniting upward pressure on long rates. On August 28, the previous trading day, the Philadelphia Semiconductor Index fell 3.47%, with Nvidia down 4.57%, Intel down 2.85%, and AMD down 2.33%. And Google's unveiling of its TurboQuant AI compression algorithm has raised questions about memory demand. Those are genuine fundamental worries about exactly the stocks the buybacks are supporting. The breakwater absorbed Monday's wave; it does not settle the argument about the tide.

What Would Confirm or Break the Read

  • September 1 residual flow. Most rebalance volume should have printed at Monday's close, but the source flags possible leftover volume early in the first September session. If the chip large-caps hold their gains without the closing-auction distortion, the flow-driven interpretation strengthens.
  • Persistence of foreign selling. If foreigners keep selling semiconductors well after the rebalance date, the selling was never mainly about MSCI — it was about Warsh, rates, and memory demand, and the buybacks are merely slowing a fundamental exit.
  • Behavior of the deleted names. HLB, LG Display, POSCO International, and Samsung Epis Holdings leave the index from September 1 with no buyback shield. They are the control group: if they stabilize quickly, the passive overhang is spent.
  • The remaining 44.7 trillion won. Watch whether the execution pace holds. The programs run to November 19 and 21; the market's working assumption that they act as a breakwater against foreign selling only holds while they are actually buying.

Monday's session was, in effect, a natural experiment: forced selling of known size met programmatic buying of known size, in public, at a single auction. The buyers won by 31.14 points. The more useful takeaway is not the score but the structure — for the next several weeks, anyone selling Samsung Electronics or SK Hynix is selling into a bid that does not flinch, and every macro scare will be tested against it.

This post takes one theme from the session in depth. For the index-by-index wrap of the same trading day, see Seoul Closing Bell.

Disclaimer: This post is for informational and educational purposes only. It is not investment advice, a recommendation, or an offer to buy or sell any security. Figures are drawn from same-day exchange and press data and may be provisional. Do your own research before making investment decisions.

댓글

댓글 쓰기

이 블로그의 인기 게시물

KOSPI's 2.9 Trillion Won Foreign Sell-Off and Who Absorbed It

Foreign investors sold a net 2.9 trillion won of KOSPI shares on Tuesday, and retail buyers absorbed most of it while the index lost just 0.27%. That flow split — the largest force in the session — is the theme worth taking apart, because the exact figures were 2.903 trillion won of foreign net selling against 1.1432 trillion won of individual net buying, and the gap between the size of the outflow and the modesty of the closing loss says a lot about where this market's support currently comes from. The split, in numbers The KOSPI closed at 6,870.81, down 18.93 points or 0.27%, its second consecutive decline after the previous session's drop of more than 2%. Behind that headline, the investor-type ledger was lopsided: foreigners net sold 2.903 trillion won, individuals net bought 1.1432 trillion won, and institutions added a comparatively small 121.1 billion won. The KOSDAQ ran the same pattern in miniature — and finished in the opposite direction. Individuals net bought 148.6 ...
자세한 내용 보기

Why KOSDAQ Stayed Green as KOSPI Dropped 1.79%

KOSDAQ closed up 0.09% at 838.41 while KOSPI fell 1.79% to 6,788.88 — a split that owes more to index structure than to sentiment. What makes Friday's session worth a closer look is that both markets saw the same investors doing the same things: individuals buying, foreigners and institutions selling. One index absorbed the pressure; the other took a 123.49-point hit. This post is about why an identical flow pattern produced such different outcomes, and what would tell us whether KOSDAQ's resilience is real or just arithmetic. One Flow Pattern, Two Results Start with the flows themselves. On KOSPI, morning-session readings showed individuals net buying roughly 65.3 to 94.7 billion won, foreigners net selling 23.6 to 80.9 billion won, and institutions net selling 81.9 to 114.8 billion won, with program trading contributing a further 68.6 billion won of net selling. On KOSDAQ, the structure was the same in miniature: individuals net buyers at around 68.8 billion won, foreigners n...
자세한 내용 보기

Foreigners Sold 3 Trillion Won of KOSPI but Spared the KOSDAQ

Foreign investors sold over 3 trillion won of KOSPI stock on Monday but left the KOSDAQ nearly untouched — a chip unwind, not an exit from Korea. That asymmetry is the single most informative detail in Monday's session, and it is worth examining on its own, because a 3.26% index decline driven by foreign selling can mean two very different things depending on where the selling lands. This post looks only at that flow pattern: what it was, why it took the shape it did, and what would confirm or break the benign reading of it. One Day, Two Very Different Order Books On the KOSPI, the flow picture near the close was stark: Foreign investors: net sellers of 3.0502 trillion won Institutions: net sellers of 1.0838 trillion won Retail investors: net buyers of 3.1315 trillion won Foreigners and institutions together unloaded supply in the 4 trillion won range, and retail absorbed almost all of the foreign leg by itself. Now compare the KOSDAQ, measured near the close with the ind...
자세한 내용 보기