# Korea Market Briefing | October 29, 2026: KOSPI Loses 6,000 as Chip Stocks and Capital Flows Diverge

The KOSPI had barely reclaimed 6,000 before losing it again. Yet Thursday's close was more complicated than a simple retreat from Korean stocks. Samsung Electronics fell while SK hynix rose. Foreign investors were net buyers on the KOSPI and net sellers on the KOSDAQ. The won and government bonds strengthened even as equities declined.

October 29 was a session for separating things that often get bundled together: corporate earnings and shareholder payouts, large companies and smaller growth stocks, a day's rebound and a month's recovery. Those distinctions make the closing numbers much easier to understand.

*Equities use the October 29 regular-session close; USD/KRW uses the Seoul 15:30 observation. Research cutoff: October 29, 2026, 21:46 KST. The U.S. inflation release and Korea's bond issuance announcement arrived after the Korean equity close and are treated separately.*

## The day, the week and the longer decline

| Measure | October 29 | Daily change | Five sessions | Twenty sessions |
| --- | ---: | ---: | ---: | ---: |
| KOSPI | 5,937.66 | −74.78 points / −1.24% | +7.30% | −13.58% |
| KOSDAQ | 753.23 | −18.05 points / −2.34% | −1.42% | −11.36% |
| USD/KRW | KRW 1,336.80 per USD | −KRW 0.90 | +KRW 12.60 | −KRW 19.90 |
| Korea government yield, 3 years | 3.766% | −3.3 basis points | — | — |
| Korea government yield, 10 years | 4.081% | −2.7 basis points | — | — |

Five-session comparisons start on October 22; twenty-session comparisons start on September 29. Equity returns exclude dividends. Currency observations use the same Seoul 15:30 convention. A higher USD/KRW rate means a weaker won. Yields are KOFIA final quoted yields; one basis point equals 0.01 percentage point. Dashes indicate comparisons not calculated. Sources: [KRX][krx], [FX closing report][fx] and [bond closing report][bonds].

![Twenty-session KOSPI and KOSDAQ trends from 21 actual closes, September 29 to October 29, starting at 100: −13.58% and −11.36%. Below: daily changes of −1.24%, −2.34%, −4.13% and +2.62% for KOSPI, KOSDAQ, Samsung Electronics and SK hynix.](/briefing-assets/kr-market-briefing-2026-10-29.en.4bcc245495fa.png)

*The upper panel uses 21 actual closing observations, with September 29 set to 100 and trading sessions equally spaced. Holidays are omitted. The lower panel compares this session alone. Space calculated the history from [KOSPI][history-kospi] and [KOSDAQ][history-kosdaq] data and cross-checked starting observations against reports for [October 22][five] and [September 29][twenty].*

## A strong profit report can still meet a falling share price

The KOSPI's five-session gain of 7.30% sits beside a 1.42% loss for the KOSDAQ. That gap matters: a rebound in the headline index does not establish a recovery across the whole market. Over twenty sessions, both indexes remain below their starting points. A short recovery and an unfinished longer decline can exist at the same time.

Samsung Electronics fell 4.13%, while SK hynix gained 2.62%. [YTN's closing report][close] identified disappointment over Samsung's shareholder returns as a contributing explanation. That is an attributed market interpretation, not proof that every seller had the same motive or that one factor explains the whole decline.

Samsung's own [release][samsung] reported third-quarter revenue of KRW 195.3 trillion and operating profit of KRW 107.4 trillion. It announced a quarterly cash dividend of KRW 4,601 per share, totaling KRW 30 trillion. Final details and the amount of remaining shareholder returns are due to be determined by the board in late January 2027.

Investors begin putting their expected earnings into a share price long before a company reports. Excellent absolute results can therefore accompany a falling stock if holders wanted a larger change or faster execution. What matters is the difference between what was delivered and what had already been priced in. This briefing does not use an unverified consensus estimate to label the earnings a beat or a miss.

A dividend and a proposed buyback also mean different things. A dividend pays eligible shareholders under specified conditions; a buyback means the company purchases its own shares. Announcing the latter does not mean those purchases have been completed. Companies in the same semiconductor industry can receive different valuations because their earnings mix and distribution plans differ. For an ETF reader, that is a reason to look at constituent weights rather than relying on the sector name alone.

## “Foreign buying” needs a market and an asset attached

The table below consistently uses KRX's 15:45 snapshot. Later revisions or different collection times can produce slightly different amounts in news reports.

| Equity net purchases, KRW 100 million | Foreign investors | Institutions | Individuals |
| --- | ---: | ---: | ---: |
| KOSPI | +681 | −4,465 | +334 |
| KOSDAQ | −1,306 | −754 | +2,166 |

Positive figures mean purchases exceeded sales. Other investor categories are omitted, so the three columns need not sum to zero. Source: [KRX investor data][krx].

Foreign investors did buy more KOSPI shares than they sold. They also sold more KOSDAQ shares than they bought, alongside institutional selling there. Those aggregate figures cannot identify a particular global fund's intentions. Nor do they establish that capital was leaving Korea as a whole: cash equities, bonds and currency positions are separate exposures.

Sentiment is clearer when more than one signal is considered. KRX's large-, mid- and small-cap KOSPI groups fell 1.3%, 0.9% and 1.0%, respectively. Institutional selling and weakness across size groups together suggest a broad measure of caution. Foreign KOSPI purchases and the rise in SK hynix are counterevidence to a claim of indiscriminate flight.

Turnover is not a net inflow either. Every secondary-market purchase has a seller on the other side. More trading tells us that ownership changed hands more actively; investor-class balances tell us something different about who accumulated shares. Keeping those concepts apart helps prevent a busy market from being mistaken for a universally confident one.

## Why the won and bonds did not follow equities down

USD/KRW fell KRW 0.90 to 1,336.80, a modest strengthening of the won. It was nevertheless KRW 12.60 above its five-session starting point and KRW 19.90 below its twenty-session starting point. The currency story changes with the measurement window, just as the equity story does.

The [FX report][fx] described competing exporter dollar sales and settlement demand at lower exchange rates. USD/KRW reached 1,331.30 in the morning but did not stay there. Exporters converting receipts supply dollars; companies paying dollar bills demand them. Such commercial transactions and hedging needs help explain why stocks and exchange rates do not obey one fixed daily relationship.

The bilateral won rate should also be separated from broader dollar performance. KRX's 15:45 reference snapshot showed the dollar index at 102.1, with a displayed change of −0.2%. That is a timestamped reference, not a completed New York daily close. A basket of major currencies and the won alone answer different questions.

Korean government yields declined to 3.766% at three years and 4.081% at ten years. Bond prices generally rise when yields fall: an existing stream of promised payments becomes more attractive relative to newly available yields. The [bond report][bonds] also recorded foreign net purchases of 1,809 three-year and 2,940 ten-year government-bond futures contracts. Those are contract counts, not cash inflows into government bonds.

Overnight, the [Federal Reserve][fed] kept its policy range at 3.75% to 4.00% in a 10–2 vote. The two dissenters wanted a quarter-point increase. A hold is not a rate cut. The decision was available before Korea opened, but it did not settle how investors should respond to subsequent U.S. data.

## Two announcements after the Korean close

At 17:00 KST, Korea's finance ministry announced a KRW 5 trillion reduction in November government-bond issuance relative to the original plan. It plans KRW 12 trillion in competitive auctions and KRW 2 trillion in early repurchases. Final gross issuance depends partly on noncompetitive purchase options. The comparison is with the original plan, not a claim that issuance fell KRW 5 trillion from October. [Official announcement][issuance]

Less supply can support bond prices, but overseas yields and investor demand may pull the other way. The release followed both the equity close and the yield assessment used here. It is information for the next session, not an explanation to attach retrospectively to today's completed decline in yields.

At 21:30 KST, the [BEA's September release][pce] showed monthly PCE inflation of 0.5%, or 0.3% excluding food and energy. Annual rates were 3.6% and 3.1%, respectively. Nominal consumption rose 0.5%, while real consumption increased by less than 0.1%.

Spending more money does not necessarily mean buying proportionately more goods and services. Adjusting for prices changes the picture. The figures leave room for concern about persistent inflation, but without a verified prerelease forecast this briefing does not call them an upside surprise. They also arrived after Korea's regular session. It is too early in this research window to assign a settled direction to U.S. stocks or the dollar on the strength of an immediate reaction.

For Korea's October 30 session, three developments will be useful to watch: whether trading after Samsung's announcement spreads to other industries; whether foreign and institutional selling on the KOSDAQ eases; and how currencies and bond yields absorb the new inflation and issuance information. A broader participation in recovery would tell us more than the return of one round number on an index screen.

*This is an educational market briefing, not a recommendation to trade an individual security or ETF.*

## Sources and calculation notes

The official KRX calendar, latest notices and dated closing brief confirmed the session. Historical index changes were calculated from Yahoo Finance daily closes; currency comparisons use reports with matching Seoul daytime observations. Both editions and charts use the same evidence.

[krx]: https://open.krx.co.kr/contents/MKD/01/0101/01010000/MKD01010000.jsp#view=2026102921
[close]: https://www.ytn.co.kr/_ln/0102_202610291641201378
[samsung]: https://news.samsung.com/kr/삼성전자-2026년-3분기-실적-발표
[fx]: https://news.einfomax.co.kr/news/articleView.html?idxno=4442019
[bonds]: https://news.einfomax.co.kr/news/articleView.html?idxno=4442020
[fed]: https://www.federalreserve.gov/newsevents/pressreleases/monetary20261028a.htm
[pce]: https://www.bea.gov/news/2026/personal-income-and-outlays-september-2026
[issuance]: https://www.mofe.go.kr/nw/nes/detailNesDtaView.do?menuNo=4010100&searchBbsId1=MOSFBBS_000000000028&searchNttId1=MOSF_000000000079801
[history-kospi]: https://query1.finance.yahoo.com/v8/finance/chart/%5EKS11?range=3mo&interval=1d
[history-kosdaq]: https://query1.finance.yahoo.com/v8/finance/chart/%5EKQ11?range=3mo&interval=1d
[five]: https://www.newsis.com/view/NISI20261022_0021500548
[twenty]: https://www.hankyung.com/article/2026092923736

## Sources
- https://open.krx.co.kr/contents/MKD/01/0110/01100305/MKD01100305.jsp
- https://open.krx.co.kr/contents/MKD/01/0101/01010000/MKD01010000.jsp#view=2026102921
- https://www.ytn.co.kr/_ln/0102_202610291641201378
- https://news.samsung.com/kr/삼성전자-2026년-3분기-실적-발표
- https://news.einfomax.co.kr/news/articleView.html?idxno=4442019
- https://news.einfomax.co.kr/news/articleView.html?idxno=4442020
- https://www.federalreserve.gov/newsevents/pressreleases/monetary20261028a.htm
- https://www.bea.gov/news/2026/personal-income-and-outlays-september-2026
- https://www.mofe.go.kr/nw/nes/detailNesDtaView.do?menuNo=4010100&searchBbsId1=MOSFBBS_000000000028&searchNttId1=MOSF_000000000079801
- https://query1.finance.yahoo.com/v8/finance/chart/%5EKS11?range=3mo&interval=1d
- https://query1.finance.yahoo.com/v8/finance/chart/%5EKQ11?range=3mo&interval=1d
- https://www.newsis.com/view/NISI20261022_0021500548
- https://www.hankyung.com/article/2026092923736