SNOWSHAGAL
INVESTMENT INSIGHT
NO.32 ⁠· 2026.11.12
BY AIR MAILPAR AVION
2026.1~9₩214tnForeign net selling

WHO SOLD
₩214 TRILLION
IN KOREA?

U.S.-linked accounts trimmed as stocks rose;
U.K.-linked accounts sold into the downturn

SENT FROMU.K.-linked accounts
ULTIMATE OWNERGlobal hedge funds?
Monthly net flows: U.K. vs. U.S.₩tn
Jun–Sep0JanFebMar⁠·Apr⁠·MayJunJulAugSep−32.2−28.9+13.4−20.2
U.K. accountsU.S. accountsFSS
Foreign flows · Leverage · AIsnowshagal.com
Investment Insight ⁠· NO.32

Foreign investors net sold ₩214 trillion in Korean equities this year. But that does not mean they abandoned Korea. U.S.-linked accounts trimmed their positions while the market climbed; U.K.-linked accounts sold aggressively after it turned. We trace who may have sold, why, and what would count as a genuine return.

34.8%Foreign ownership at end-September, up from 30.8% at end-2025 despite ₩214tn in net sales
61%Share of May's ₩47tn in foreign net selling attributed to U.S.-linked accounts
76%Share of the ₩100tn sold in June–September attributed to U.K.-linked accounts

Published Nov. 12, 2026 · Data cut-off: after the Korean market close on Nov. 12 (KST) · Approx. 7-minute read (15 minutes with the deep dive)
Sources: Financial Supervisory Service; Daishin, Korea Investment & Securities, Hana, Hanwha and Samsung Securities; U.K. Financial Conduct Authority (FCA); Goldman Sachs prime-brokerage note (via Reuters); Bloomberg and Korean press reports

Contents

The story in five points

  1. Foreign investors did not abandon Korea. Net sales of ₩214tn amounted to 16% of their holdings at the start of the year. The value of what they kept rose, lifting their share of the market from 30.8% to 34.8%.
  2. There were two very different kinds of selling. U.S.-linked accounts pared holdings in the rising markets of February and May; U.K.-linked accounts led selling during March's shock and the June–September slump.
  3. Single-stock concentration limits helped drive sales even as prices rose. As Samsung Electronics and SK hynix surged past some funds' 10% single-name limits, managers mechanically sold down their excess exposure.
  4. Hedge funds trading through London are the leading explanation for U.K.-linked selling. The postmark says London, but the sender may be based somewhere else entirely.
  5. It is too early to call this a return. Foreigners bought KOSPI stocks for three straight sessions on Nov. 10–12, but Nov. 11 buying was concentrated in two chipmakers and Nov. 12 KOSDAQ selling nearly offset KOSPI purchases. The test is whether buying spreads to other sectors.

Hover over underlined terms for explanations, or tap on mobile. Red dotted underlines mark key concepts.

Introduction

On Nov. 10, foreign investors became net buyers of KOSPI stocks for the first time in eight trading sessions. They remained net buyers through Nov. 12—a three-day streak.1, 2, 3 The KOSPI gained more than 7% across Nov. 10–11,4 then fell 0.6% on Nov. 12 as domestic institutions sold.

Before that, foreigners had net sold ₩214tn of Korean equities in January–September alone.5 That is why headlines spoke of foreigners fleeing Korea. But the picture changes when we examine who was selling each month.

  • “If they sold ₩214tn, doesn't that mean they left?”
  • “Who actually sold—and why?”
  • “Does three straight days of buying mean they're back?”
01 · Did they leave?

They sold ₩214tn—
yet foreign ownership rose

The ₩214tn sold was equivalent to 16% of foreign investors' holdings at the start of the year. Far more remained invested than was sold, and the value of those remaining shares climbed enough to increase foreigners' share of the Korean stock market.

−₩214tnForeign net sales, Jan–Sep: 16% of holdings at the start of the year
+4.0%pForeign ownership: 30.8% at end-2025 → 34.8% at end-September
+73%Market value of foreign holdings: ₩1,327tn → ₩2,296tn
  • Semiconductor shares, which foreigners hold heavily, outperformed the broader market.5, 6 Imagine holding ₩1 million of stock, selling ₩200,000 worth, and then watching the remaining shares double: the investment is now worth ₩1.6 million.
  • It is therefore misleading to equate ₩214tn in net selling with foreigners abandoning Korea. They reduced part of their exposure, and some sales were simply a response to rising portfolio weights.
02 · Who sold?

U.S.-linked accounts trimmed as stocks rose;
U.K.-linked accounts sold into the downturn

The leading sellers changed from month to month. Three distinct episodes account for roughly 98% of the ₩214tn total.

Three episodes explain ₩210tn of the ₩214tn₩ trillion · Foreign net equity sales, Jan–Sep 2026

Sources: FSS, monthly Foreign Securities Investment reports; Daishin Securities. January and April (₩4tn combined) are omitted because their net flows were comparatively small.

Why sell into a rally?
The 10% single-stock limit

Many overseas funds cannot invest more than 10% of assets in a single security. As Samsung Electronics and SK hynix rallied, funds whose holdings exceeded that limit had to sell the excess mechanically. Goldman Sachs estimated such selling at roughly US$69bn since late October 2025.7 Daishin Securities likewise interpreted May's sales as rebalancing after Korea's weight in portfolios had risen.8

Selling generated by rising prices is not the same as capital fleeing Korea.

  • The intensity of selling was different. Relative to their year-start holdings, January–September net sales amounted to 12% for U.S.-linked accounts (excluding April) but 68% for U.K.-linked accounts.9 U.S.-linked accounts trimmed at the margin; U.K.-linked accounts cut much more deeply.
  • Indeed, U.S.-linked accounts bought ₩17tn of Korean stocks during the July–August slump.10
03 · Who is behind the U.K. accounts?

Postmarked London—
but the sender may be elsewhere

A “U.K.-linked account” does not necessarily hold British investors' money. Trades placed by hedge funds around the world through London investment banks can be attributed to the United Kingdom. Those funds are the leading candidates behind the selling.

When you order something from overseas, the shipping label may show a distribution hub rather than the sender's address. Securities transactions can work much the same way. If a hedge fund in the United States or Hong Kong borrows through a London investment bank (a prime broker) to buy Korean stocks, the bank's U.K. entity may be the name that appears on the books.10 Daishin Securities argues that the FSS attributes trades of this kind to the United Kingdom. A single account in the name of an overseas broker can also aggregate orders from numerous foreign investors.11

Korean equities reported by residents of each country ÷ Korean equities attributed to accounts in that countryEnd-2025; 1.0× means the two measures are equal
1.0×
Luxembourg1.04×
United States0.95×
Cayman Islands0.94×
Singapore0.56×
the United Kingdom0.53×
Netherlands0.49×
Hong Kong0.47×

the United Kingdom Account-attributed holdings: ₩144.1tn · Resident-reported holdings: about ₩76.4tn

Sources: IMF Coordinated Portfolio Investment Survey (CPIS), FSS and Daishin Securities

Korean equities reported by U.K. residents amount to just over half the value attributed to U.K.-linked accounts. For the United States, the two measures are much closer.

Three reasons to suspect hedge funds

  • Banks tightened financing. On June 12, Citi, JPMorgan and Goldman Sachs raised financing charges and restricted transaction sizes for leveraged positions in the two chipmakers.12 The KOSPI peaked ten days later.
  • Hedge funds did reduce positions. Goldman Sachs client hedge funds unwound positions at a record pace in July, with Korea leading semiconductor selling across Asia.13
  • Past market shocks show a similar pattern. U.K.-linked accounts accounted for 79% of foreign selling during the August 2024 slump and 65% in April 2025.14, 15 By contrast, they were the largest buyers in the rising market of October 2025.16 These flows have been highly sensitive to risk.

Why did the selling last more than four months?

  • June–July: the cost of leverage. Financing was tightened just as the KOSPI lost 22% in July, forcing leveraged investors to reduce positions.17
  • July–September: volatility. Annualized KOSPI volatility this year was 1.5 times its 2008 level. Funds that scale equity exposure to volatility sold more as the market became more turbulent.10
  • Autumn: mounting doubts about AI. Questions grew over whether the pace of AI investment could be sustained. Taiwan also experienced record monthly foreign selling in October.18
Another theory: U.K. pension withdrawals

Korea Investment & Securities pointed to U.K. inheritance-tax changes affecting pensions. From April 2027, unused pension savings left at death will be brought into the inheritance-tax net, potentially encouraging earlier withdrawals.19 U.K. pension withdrawals did, in fact, rise 21.7% over one year.20

But the amounts do not line up well. At last year's withdrawal pace, four months of withdrawals across the entire U.K. pension system would total about ₩55tn (the available statistics run only through March). Even if withdrawals accelerated this year, virtually all the withdrawn money would have needed to come from selling Korean stocks to approach the ₩76tn sold through U.K.-linked accounts in June–September. Pension-related selling may have contributed, but it is unlikely to explain most of that ₩76tn.

What this report cannot establish

Public data do not identify who sold how much within U.K.-linked accounts. Hedge funds' exact share cannot be verified, and country-by-country figures from October onward are not yet available.

04 · Are they coming back?

Three straight days of buying—but still too narrow
The test is whether it broadens

U.S.-linked accounts sold the equivalent of 12% of their year-start holdings and then bought again in July and August. The money that needs to return is the capital sold during the crash. Actual purchases will tell us more quickly than country-of-account statistics whether it is returning.

What to watchLatest readingWhere to check
① Is foreign buying sustained—and broadening?KOSPI net purchases: approximately ₩360bn on Nov. 10, ₩658.7bn on Nov. 11, and ₩370bn on Nov. 12 (provisional). On Nov. 11, purchases of Samsung Electronics and SK hynix exceeded total foreign net buying, implying net sales across the other stocks combined. On Nov. 12, foreigners sold about ₩350bn on KOSDAQ, leaving just ₩18bn in combined net buying across the two markets.2, 3Check “Trading by Investor Type” in a brokerage app. Watch for KOSPI foreign net purchases to exceed the combined purchases of Samsung Electronics and SK hynix—and for buying on KOSDAQ as well.
② Is volatility subsiding?VKOSPI 96.9 (June 29) → 52.76 (Nov. 11), still nearly three times its normal level of 184, 10Look up “VKOSPI” in a brokerage app or market-data portal. Is it still falling?
③ Have U.K.-linked accounts turned buyers?Net sales of ₩6.9tn in September, far less than June's ₩32.2tn5FSS monthly release. November's figures come out in December; this is a lagging confirmation.
  • What has improved. Concerns about AI investment eased,21 and VKOSPI fell more than 9% on Nov. 11 alone. Foreigners bought KOSPI shares even on Nov. 12, when the index fell as domestic institutions sold around ₩690bn.3
  • What remains unconvincing. Buying has yet to broaden. It was concentrated in two stocks on Nov. 11, while Nov. 12 KOSPI buying was almost entirely offset by KOSDAQ selling. Foreigners were also net sellers of futures on both Nov. 11 and 12.3, 22 November 12 was an options-expiration day, so expiry-related transactions are embedded in the figures.
  • Treat the Nov. 30 numbers with caution. MSCI's Nov. 12 index review made no constituent changes to its flagship Korea index, but SK hynix's weight will increase to reflect a higher share count. Samsung Securities estimates that at the Nov. 30 close, around passive index-tracking capital ₩580bn of buying will flow into SK hynix.23 Those are real purchases, but they are mechanically driven by an index rebalance. If foreign net buying surges that day, strip out this component before judging whether discretionary risk-taking has returned.
  • Hana Securities thinks a sustained return of foreign investors could take at least three months.24
Conclusion

They trimmed on the way up and sold hard on the way down—
actual buying will show whether they are back

The ₩214tn in foreign net sales reflects two distinct behaviors. U.S.-linked accounts reduced holdings during the rally; U.K.-linked accounts sold heavily during the crash. Hedge funds operating through London are the strongest explanation for the latter. Foreign investors reduced exposure by the equivalent of 16% of their starting holdings, while the value of what remained increased and their ownership share rose. That is not wholesale abandonment of Korea.

A slowdown in selling is not the same as a return to buying. Three tests matter now.

  1. Does foreign buying spread beyond semiconductors?Watch whether foreign net buying on KOSPI exceeds purchases of the two chipmakers combined, and whether investors also buy on KOSDAQ. Daily brokerage data can show this. Adjust for MSCI-related purchases on Nov. 30.
  2. Does VKOSPI keep falling?Less volatility would make it easier for both leveraged and longer-horizon investors to rebuild exposure.
  3. Have U.K.-linked accounts turned net buyers?The FSS November data, due in December, will provide a lagging check.

This is not a recommendation to buy or sell securities. The evidence and calculations appear in the “Deep dive” below.

Further reading

Deep dive

The supporting numbers and evidence omitted from the main narrative. Expand only the sections you want to explore.

Who sold in which month—and from which country?

The leading sellers changed from month to month. U.S.-linked accounts dominated sales during rising markets, while U.K.-linked accounts led during sharp declines.

Foreign net purchases of Korean listed equities, by month, 2026₩ trillion · settlement-date basis · Only August showed net buying under this measure

Sources: FSS monthly Foreign Securities Investment reports; Daishin Securities

Breakdown by country-linked account
MonthTotalU.K.-linked accountsU.S.-linked accounts
Jan−0.1+1.8−7.8
Feb−19.6−4.7−8.7
Mar−43.5−16.3−9.5
Apr−4.0−2.5Not disclosed
May−47.0Not disclosed−28.9
Jun−49.3−32.2−23.2
Jul−31.7−20.2+3.5
Aug+0.3−16.9+13.4
Sep−19.1−6.9−3.9
Jan–Sep−214.0−97.5−65.1 Excluding April

Unit: ₩ trillion. The FSS publishes only the leading net-buying and net-selling countries each month, leaving some cells blank. The Jan–Sep totals use Daishin Securities' data (U.S. excludes April).5, 10

Breaking down the ₩214tn into three episodes
EpisodeAmountMarket backdropLeading account
① Selling into strength
February · May		₩66.6tnA rising market in which foreign holdings grew to ₩2,026tn by end-February and ₩2,852tn by end-May5United States ₩37.6tn (56%) from U.S.-linked accounts; in May alone, ₩28.9tn (61%). The U.K. did not appear among the leading sellers that month.
② War shock
Mar		₩43.5tnU.S.–Iran war; KOSPI fell 12.06% on March 45the United Kingdom ₩16.3tn (37%) from U.K.-linked accounts; ₩9.5tn from U.S.-linked accounts
③ Peak and plunge
Jun–Sep		₩99.8tnThe index peaked on June 22 and then tumbled; July's 22% KOSPI loss was the worst monthly decline since October 2008.17the United Kingdom ₩76.3tn (76%) from U.K.-linked accounts. U.S.-linked accounts bought ₩16.9tn in July–August.10

January (−₩0.1tn) and April (−₩4.0tn) are omitted because their flows were comparatively small. June was a transition month: the index rose through June 22 before reversing, and both U.K. (−₩32.2tn) and U.S. (−₩23.2tn) accounts sold. The three-episode grouping is this report's analytical framework.

Whose money sits behind U.K.-linked accounts? Weighing the evidence

Who was responsible for the U.K.-linked accounts that made up 76% of June–September selling? First, distinguish the entity in whose name a trade is recorded, the manager who makes the investment decision, and the ultimate owner of the capital. They need not be the same.

Two sell-side explanations

  • Daishin Securities: hedge-fund deleveraging. Daishin argues that hedge funds using leverage through prime brokers and instruments such as total return swaps (TRS) and contracts for difference (CFDs) reduced their borrowing and sold positions.10
  • Korea Investment & Securities: U.K. pension inheritance-tax reform. From April 6, 2027, unused pension funds remaining at death will be included in the U.K. inheritance-tax estate. The theory is not that current pension balances face an immediate tax, but that savers have an incentive to withdraw earlier. The firm points to the rise in selling from June, after implementation details were published, while acknowledging that the impact cannot be quantified precisely.9, 19

These explanations are not mutually exclusive: they could affect different investor groups at the same time. The table separates what each dataset does—and does not—establish.

What the evidence can and cannot tell us
EvidenceWhat it showsWhat it cannot establish
Holdings ratio: 0.53×
U.K. resident-reported holdings ÷ U.K.-linked accounts		U.K.-linked accounts hold substantially more Korean equities than U.K. residents report owning. The corresponding ratios for the United States, Cayman Islands and Luxembourg are near 1.0×.10The precise shares attributable to non-U.K. capital (non-residents), hedge funds or pensions. Because the two datasets use different bases, the gap cannot be read as “47% foreign money.”
U.K.-linked net sales of ₩76.3tn, Jun–SepEquivalent to 45% of all Korean equities held by U.K. residents, valued at approximately ₩171tn using May prices10Which investors sold; account-level breakdowns by investor and security are not public
Semiconductor concentrationThe two chipmakers made up 75–84% of total foreign net selling after June 22.10Which securities U.K.-linked accounts sold; the available figure covers all foreign investors, not U.K.-linked accounts specifically
Variation in monthly net flows
Relative to holdings		U.K. 5.38% versus U.S. 1.30% in Jan–Sep 2026—much more volatile flows through U.K.-linked accounts10Holding periods or investor types
Bank financing restrictions and prime-brokerage dataBanks tightened swap-financing terms for the two chipmakers in June,12 and Goldman Sachs client hedge funds cut positions at a record pace in July, with Korea leading Asian semiconductor selling.13How much of the U.K.-linked net selling resulted from this activity. The sample covers Goldman Sachs clients; unwinding positions is not the same measure as net equity sales.
U.K. pension withdrawals£91.2bn withdrawn between April 2025 and March 2026, up 21.7% year on year (U.K. FCA)20How much of that money came from selling Korean equities, or whether the entire increase was caused by the inheritance-tax changes
Historical U.K.-linked sellingU.K.-linked accounts accounted for 79% of foreign sales in August 2024, before the pension-tax policy was announced in October 2024.14, 25Whether the 2026 tax changes increased the selling further
A reality check on the amounts

The Korea Investment & Securities argument is that prospective inheritance taxes prompted early pension withdrawals. The size of those withdrawals is therefore the relevant benchmark. Total U.K. pension withdrawals of £91.2bn were worth about ₩166tn. That is a full year's withdrawals across investments in stocks, bonds and cash; four months at the same rate comes to roughly ₩55tn. Yet U.K.-linked accounts sold ₩76tn of Korean shares in June–September alone. Even assuming withdrawals accelerated this year, nearly all the withdrawn money would have had to be funded by Korean stock sales to approach that figure.

The conversion assumes about ₩1,820 per pound. There are also limits to the comparison: withdrawal data end in March 2026, not September, and do not include the trading of employer-managed defined-benefit pension schemes. The figures likewise cannot establish whether pension managers reduced Korean equities for reasons unrelated to withdrawals. The conclusion is therefore limited to: pension withdrawals are unlikely to have been the primary driver.

Assessment. Hedge-fund risk reduction is independently visible in banks' financing restrictions and prime-brokerage data. It is also consistent with the size, volatility and historical behavior of U.K.-linked flows, making it the most plausible explanation. The increase in U.K. pension withdrawals is real, but the scale makes it unlikely to have been the dominant factor. Public data nevertheless cannot establish hedge funds' share of U.K.-linked net sales.

Eight historical episodes of foreign selling

This pattern is not unique to 2026. Across eight episodes since 2024 for which figures are available, U.K.-linked accounts led during abrupt market shocks, while U.S.-linked accounts tended to lead during gradual, multi-month portfolio reductions.

Leading account's share of total foreign net selling% · Annotations show the market backdrop and other accounts

Sources: FSS (monthly, settlement-date basis), Korea Exchange (Nov. 2025, trade-date basis), Daishin Securities and press reports

  • U.K.-linked accounts also led the buying. In the rising market of October 2025, U.K.-linked accounts bought ₩2.4tn of the ₩4.2tn in total foreign net purchases—the largest share—while U.S.-linked accounts sold ₩1.0tn.16 The pattern is consistent with fast-moving, risk-sensitive capital: buying most aggressively on the way up and selling most heavily when shocks hit. This is the report's interpretation, not a verified identification of investor type.
  • The hedge-fund evidence lines up. Goldman Sachs prime-brokerage data show that hedge funds bought Asian equities at a record pace in the second week of February, concentrated in Korea, Taiwan and China;26 then made their biggest sales of Taiwanese, Korean and Indian stocks in a year during the mid-March war shock, primarily through short sales.27 In July they reduced positions at a record pace, with Korea driving Asian semiconductor selling.13 These are Goldman Sachs client figures; changes in positioning are not identical to net sales.
  • There are counterexamples. During the three-month September–November 2024 slump driven by semiconductor earnings concerns, U.S.-linked and Luxembourg-linked accounts were the two largest sellers, while U.K.-linked accounts were net buyers in November.28, 29, 30 U.K.-linked accounts do not always lead selling; the pattern is strongest during rapid, disorderly moves.

Shares are calculated against total foreign net sales for the relevant period. November 2025 uses nationality-based Korea Exchange trades for Nov. 1–24; the other periods use FSS settlement data. “Shock” versus “gradual reduction” is this report's classification.14, 15, 31

Why the selling continued for more than four months after the peak

The post-June selling was unusually large and prolonged. Tighter financing, volatility and AI-investment concerns overlapped rather than following neatly one after another; their relative importance shifted over time.

  1. 1June–July

    Banks tightened the cost of leverage

    Event
    On June 12, ten days before the market peak, Citi, JPMorgan and Goldman Sachs raised financing costs and capped the size of leveraged trades in SK hynix and Samsung Electronics.12 Swap financing spreads jumped from SOFR + 1–2 percentage points in early May to as much as roughly SOFR + 10 percentage points by mid-June.17, 32 The KOSPI fell 10.0% on June 23, the day after its peak.10
    Impact
    Maintaining or expanding leveraged positions became harder, and hedge-fund deleveraging was subsequently observed. Public figures do not show how much was forced liquidation versus voluntary reduction. Daishin Securities argues prime brokers cut positions in baskets heavily exposed to semiconductors. Domestic leveraged ETFsalso traded about ₩9.6tn around the June 23 close as they rebalanced, magnifying volatility.10
    Evidence
    Asian equity long/short hedge funds lost 18.6% in July through the 28th, approaching their worst monthly decline on record in Goldman Sachs data.13 U.K.-linked net sales in June: ₩32.2tn
    How it evolved
    By an Aug. 20 report, SK hynix swap-financing spreads had eased to SOFR + 1.5–3 percentage points, less than half their peak.17, 32 U.K.-linked net selling also slowed from ₩32.2tn in June to ₩6.9tn in September. Yet selling persisted into autumn after funding costs eased, so financing conditions alone cannot account for the entire episode.
  2. 2July–September

    Volatility pushed even longer-term capital out

    Event
    Annualized KOSPI volatility jumped from 21.8% in 2025 to 60.7%, or 1.5 times its 2008 level. The index moved more than 5% in a day on 40 occasions this year, against 13 such days in all 16 years from 2010 through 2025 combined.10
    Impact
    For investors whose allocations are scaled mechanically to volatility, maintaining the same risk budget would require a 64% reduction in equity exposure. Daishin Securities' interpretation is that turbulence amplified by short-term trading eventually forced longer-horizon investors to sell as well.10
  3. 3September–early November

    AI concerns deepened—and they were not confined to Korea

    Event
    Concerns about AI investment were already present during July's selloff, but intensified in autumn. The question around U.S. Big Tech capital spending shifted from “How much will they spend?” to “Can they keep spending?” Whenever AI cloud provider CoreWeave's CDScredit-default swap spread (CDS) widened, foreign investors sold the memory-chip-heavy KOSPI.33 After the Nov. 5 U.S. midterm elections, worries about tighter regulation of AI data centers added to the pressure.24
    Taiwan, too
    Foreigners net sold US$31.78bn (about ₩42.4tn) of Taiwanese equities in October, a monthly record.18 Korean net selling in the same month amounted to ₩19.5tn.10 Market sizes differ, so the amounts are not directly comparable. They nevertheless indicate a broader reduction in semiconductor- and AI-related risk beyond Korea.
    Evidence
    Foreign net sales in Korea totaled ₩30.1tn from Oct. 1 to Nov. 6, with the two chipmakers accounting for 75%. CoreWeave's CDS reached a peak of 1,264bp on Nov. 5.

The three forces overlapped rather than operating in discrete phases. When October's account-country figures are released, they can show whether U.K.-linked accounts also led the autumn selling.

Why the entire market fell when semiconductors bore the selling—and who bought

The two chipmakers made up 52% of the KOSPI's market capitalization as of July 15, so their losses dragged down the index;34 they also undermined the margin positions of other leveraged investors, prompting sales of non-semiconductor stocks as well (Daishin Securities).10

On heavy foreign-selling days: where investors sold, and where prices fellThe ten largest foreign net-selling days between June 22 and Nov. 6 (at least ₩3.64tn in daily sales)

Source: Daishin Securities (top 300 companies by market capitalization)

KOSPI net flows by investor type during two declines₩ trillion · First decline: Jun. 22–Aug. 31 · Second: Sep. 1–Nov. 6

Source: Daishin Securities

  • Retail investors absorbed the first wave, buying ₩32.4tn, but switched to net selling of ₩5.5tn in the second. Brokerage customer cash deposits fell 35%, from ₩139.7tn on June 4 to ₩90.3tn on Nov. 6.10
  • Corporations absorbed the second wave. Other corporationsThey bought a net ₩33.5tn of SK hynix and ₩11.6tn of Samsung Electronics, consistent with buyback activity by the two companies (Daishin Securities).10 Samsung Electronics completed a ₩15tn buyback on Oct. 6;35 SK hynix completed its first ₩40tn round on Oct. 21. Its second ₩20tn program (half earmarked for employee compensation)36is proceeding at about ₩0.35tn per day, roughly one-third the first round's pace.
  • Between Nov. 3 and 6, after Samsung's and SK hynix's first buybacks had ended and only the slower second round remained, foreigners sold ₩10.3tn over four sessions.10 The timing overlaps, but causation has not been established. For more on the buybacks, see NO.19⁠·NO.22 .
Six signs of a foreign-investor return—and a volatility scenario

U.S.-linked accounts trimmed the equivalent of 12% of their starting holdings, and bought again in July–August. What needs to come back is the money sold after the reversal. Account-nationality data arrive more than a month late, so daily confirmation should focus on whether actual buying persists and broadens beyond the two chipmakers.

What to watchLatest readingWhat it showsCheck
① Persistence and breadth of actual purchasesKOSPI net buying: about ₩360bn on Nov. 10, ₩658.7bn on Nov. 11, and ₩370bn on Nov. 12 (provisional). Purchases of the two chipmakers totaled ₩668.1bn on Nov. 11, implying ₩9.4bn in combined net selling of other stocks. On Nov. 12, KOSDAQ net selling of about ₩350bn left roughly ₩18bn in net buying across both markets.2, 3Is demand for Korean equities returning? It needs to persist for several days and spread to other sectors.Korea Exchange · Daily
② VolatilityVKOSPI: 96.9 (June 29) → 65.2 (Nov. 6) → 52.76 (Nov. 11), still 2.9 times the 2010–25 average of 18.04, 10Are conditions improving enough for investors to take more risk?Korea Exchange · Daily
③ AI credit conditionsCoreWeave CDS peaked at 1,264bp on Nov. 5 and subsequently retreated.33Are funding concerns around AI investment easing? It remains to be seen whether the decline continues.Credit markets · Daily
④ Semiconductor earnings2027 net income estimate: ₩689tn (June 19) → ₩866tn (Nov. 6)10 CoreWeave's Q3 revenue of US$3.66bn beat the US$3.5bn expectation, but net interest expense of US$846m exceeded operating income of US$207m, and the company posted a US$709m net loss.37, 38Is AI demand holding up? Strong operating results and improving credit quality are separate questions.Broker consensus estimates
⑤ Financing termsAn August report said SK hynix swap-financing spreads had eased to less than half their June peak.17, 32 Current terms as of November have not been verified.Can investors once again finance purchases with borrowed money?International financial press
⑥ U.K.-linked accountsNet sales of ₩6.9tn in September (versus ₩32.2tn in June). End-September holdings were ₩172tn, close to the ₩144tn at the start of the year.9October data (expected in mid- to late November) will show who led selling that month. November data (due in December) will show whether U.K.-linked accounts turned buyers. Both are backward-looking checks.FSS · Monthly

Hana Securities cited a correlation of −0.46 between U.S. high-yield corporate bond spreads and foreign equity flows since 2019, suggesting that a sustained return of foreign capital could take at least three months.9, 24 Overseas leveraged funds' net position in Nasdaq-100 futures (as reported in the CFTC CFTC Commitments of Traders data, measured as a share of open interest ) shifted from −32.0% on Aug. 11 to +2.0% on Nov. 3. But these are U.S. futures, not a direct measure of flows into Korean stocks.10

How far has the Nov. 10–12 rebound progressed?

  • Buying continued, but remained narrow. Foreigners were net buyers on all three days. On Nov. 11, however, purchases were concentrated in the two chipmakers, while the rest of the market saw net selling. On Nov. 12, KOSPI purchases were offset by KOSDAQ sales, leaving just about ₩18bn of combined net buying.3 That is not yet convincing evidence of a market-wide return.
  • Conditions are beginning to ease. The Nov. 10 rebound came after global investment banks argued that prospective AI data-center regulation would more likely delay projects than halt them, viewing AI fears as overdone. The U.S. semiconductor index had risen 6.9% the previous day.21 VKOSPI then fell 9.33% on Nov. 11.4 On Nov. 12, reports concerning Nvidia's CEO weighed on sentiment, while institutional selling erased the KOSPI's intraday gains and pushed it down 0.6%.3, 39
  • Some judgments should remain open. Foreigners net sold 3,805 KOSPI 200 futures contracts on Nov. 11 and continued selling futures on Nov. 12.3, 22 We cannot tell whether those futures sellers were the same investors buying cash equities. November 12 was an options-expiration day, so the figures include expiry-related activity.

Potential capacity to rebuild exposure if volatility falls (illustrative)

If volatility falls to10% of holdings subject to the ruleApproximately 25% (extreme assumption)
45%₩57tn₩144tn
35%₩121tn₩304tn
25%₩235tn₩591tn

These are scenarios from Daishin Securities that assume a specified share of approximately ₩1,645tn in foreign-held Korean equities follows volatility-based allocation rules—not forecasts of actual inflows. The 25% case is extreme: it assumes all ₩214tn in foreign sales resulted from these rules. The volatility input is realized annualized equity-market volatility (60.7% this year), not the options-implied VKOSPI. Historical data also showed no statistically clear tendency for foreigners to buy more after volatility falls.10 Valuation alone is not enough. The two chipmakers' 12-month forward P/Eis 2.4×, or 76% below the 2005–25 median of 10.1×. But that figure also embeds interest-rate assumptions and concerns that earnings may be peaking. If earnings expectations (signal ④) deteriorate, a low multiple may be read as “peak earnings” rather than “cheap stocks.”

Past returns following foreign selling, and upcoming dates

Similar combinations of heavy selling and depressed valuations have often preceded rebounds. Yet the samples are small, and earnings, rates and market structure differed from one episode to another. These are not probabilities of a rebound this time.

ConditionSubsequent KOSPI returnCaveat
Four prolonged foreign-selling episodes (2006, 2007–08, H2 2008 and 2020), measured from the day of heaviest selling+12.1% after 60 days · +31.7% after 250 days
Unconditional averages over the same horizons: +2.5% and +12.1%		Only four cases; the 2007–08 and H2 2008 episodes overlap. Daishin Securities.
The day foreign investors turned net buyers in those same episodes+22.2% after 250 days, but −2.0% in the first 60 daysBuying on the turn did not guarantee an immediate rally. Daishin Securities.
Seven indicators reaching extremes: P/B, dividend yield, foreign net-buying intensity, turnover, drawdown, P/E and volatility+3% after 1 month · +4% after 3 months · +15% after 6 months · +24% after 1 year
95% of the historical cases studied were higher after one year		Median returns, excluding dividends. Several indicators flagged the same periods, so the result must not be treated as an independent probability. Hana Securities.

Cumulative foreign net sales in the current episode equal 3.44% of market capitalization, nearing the 3.97% recorded in 2007–08. Daishin Securities sees a medium-term buying opportunity;10 Hana Securities favors starting to scale in over three months.40

Dates to watch

Date (Korea time)EventWhy it matters
Early Saturday, weeklyCFTC weekly futures positioning reportLeveraged funds' positioning in U.S. tech futures (context only)
Mid- to late NovemberFSS October Foreign Securities Investment reportDid U.K.-linked accounts lead October's selling too? Whether they returned in November will be clearer in the November report, published in December.
11.26Bank of Korea Monetary Policy Board decisionKRW exchange rate and the rate path (NO.31)
Nov. 30 market closeMSCI November index review takes effect. No additions or deletions to the flagship Korea index; SK hynix weight increases (announced Nov. 12).23, 41Approximately ₩580bn in passive index buyingof SK hynix expected (Samsung Securities estimate). Adjust headline foreign net-buying figures for that mechanical flow.
Early Dec. 10U.S. FOMC decision (Dec. 8–9 meeting)42U.S. rates and global risk appetite
Dec. 10 market closeQuarterly KOSPI 200 futures and options expirationExpiry-related trading could distort foreign-flow figures that day.

KNOWNEstablished facts

  • Foreign net selling of ₩214.0tn in Jan–Sep; ₩97.5tn attributed to U.K.-linked accounts (FSS, settlement-date basis)
  • Foreign ownership rose from 30.8% at end-2025 to 34.8% at end-September.
  • U.S.-linked accounts accounted for 61% of May selling; U.K.-linked accounts, 76% of June–September selling.
  • Banks tightened swap-financing terms in June; hedge funds cut positions in July (Goldman Sachs client data).
  • U.K. pension withdrawals totaled £91.2bn, up 21.7% (April 2025–March 2026).
  • Taiwan saw record monthly foreign net selling in October.
  • Foreigners bought KOSPI equities for three straight sessions on Nov. 10–12 (Nov. 12 provisional); VKOSPI stood at 52.76 on Nov. 11.
  • MSCI November review: no changes to Korea index constituents, but a higher SK hynix weight

ESTIMATEAnalysis and assumptions

  • Sales into the rally reflected concentration limits and portfolio rebalancing (Goldman Sachs estimate: roughly US$69bn).
  • Hedge funds trading through London are the leading explanation for U.K.-linked selling (Daishin Securities; this report's assessment).
  • U.K. pension inheritance-tax changes may have contributed (Korea Investment & Securities); this report judges them unlikely to be the main driver on the basis of scale.
  • If realized volatility falls to 35%, illustrative capacity to increase exposure is ₩121tn.

UNKNOWNStill unknown

  • The portions of U.K.-linked selling attributable to hedge funds and pensions
  • Which accounts led selling from October onward
  • Current November financing terms and the ultimate buyers behind November purchases
Calculation notes
  • Size of the ₩214tn: ₩214.0tn ÷ ₩1,326.8tn of foreign holdings at end-2025 = 16.1% (Daishin Securities)
  • U.K.-linked shares: 97.5 ÷ 214.0 = 45.6%; June–September: 76.3 ÷ 99.8 = 76.4%
  • The 0.53× ratio: US$52.9bn of Korean equities reported by U.K. residents to the IMF, converted at the year-end USD/KRW rate to about ₩76.4tn, divided by ₩144.1tn held in FSS U.K.-linked accounts at end-2025 (Daishin Securities)
  • Growth in market value of holdings: total foreign 2,296.0 ÷ 1,326.8 − 1 ≈ 73%; U.S.-linked 1,007 ÷ 546 − 1 ≈ 84%; U.K.-linked 172 ÷ 144 − 1 ≈ 19% (end-2025 to end-September 2026)
  • Size of pension withdrawals: £91.236bn × about ₩1,820/£ ≈ ₩166tn (one year); four months ≈ ₩55tn. Exchange rate follows a Maeil Ilbo article converting £1.41bn to roughly ₩2.57tn.
  • Other KOSPI stocks on Nov. 11: ₩658.7bn − (₩537.5bn + ₩130.6bn) = −₩9.4bn
  • Three episodes: Feb + May = 19.6 + 47.0 = ₩66.6tn (U.S.: 8.7 + 28.9 = ₩37.6tn, 56%); March ₩43.5tn (U.K.: 16.3 ÷ 43.5 ≈ 37%); Jun–Sep ₩99.8tn (U.K.: ₩76.3tn; U.S.: −23.2 + 3.5 + 13.4 − 3.9 = −₩10.2tn). Add ₩4.1tn for January and April to reach ₩214.0tn.
  • Selling intensity: U.K. 97.5 ÷ 144 ≈ 68%; U.S. 65.1 ÷ 546 ≈ 12% (Jan–Sep net selling divided by end-2025 holdings; April U.S. data are unavailable and excluded).
  • U.K. share in shocks: August 2024, 1.97 ÷ 2.51 ≈ 79%; April 2025, 8.9 ÷ 13.59 ≈ 65%; Nov. 1–24, 2025, 4.99 ÷ 13.53 ≈ 37%; March 2026, 37%; Jun–Sep 2026, 76%.
  • U.S. share in gradual reductions: Sep–Nov 2024, (2.8 + 1.3 + 1.4) ÷ (7.36 + 4.39 + 4.15) ≈ 35%; February 2026, 8.7 ÷ 19.56 ≈ 44%; May 2026, 28.9 ÷ 47.02 ≈ 61%.
  • The 64% reduction in stage two: Under inverse-volatility allocation, 1 − 21.8 ÷ 60.7 ≈ 64% of the equity weight must be reduced (Daishin Securities).
  • Two-day rally: 1.0439 × 1.0260 − 1 ≈ 7.1% (Nov. 10–11)
  • Re-exposure simulation: approximately ₩1,645tn in foreign equity holdings (Oct. 22 holdings repriced using the Nov. 6 KOSPI level) × share following volatility rules × (current realized volatility ÷ new volatility − 1). At a 10% allocation share and 35% volatility: 1,645 × 0.1 × (60.7 ÷ 35 − 1) ≈ ₩121tn.
  • Extreme-case share: Raising volatility from 21.8% to 60.7% forces an investor following this rule to cut exposure by 64%. If all ₩214tn in selling (16.1% of holdings) came from those investors, then 16.1% ÷ 64% ≈ 25.2%. The ₩144tn, ₩304tn and ₩591tn table cases use 25.2%; using exactly 25% would yield ₩143tn, ₩302tn and ₩587tn.
  • P/E gap: 1 − 2.4 ÷ 10.1 ≈ 76%. Daishin Securities described this as a “76% discount to earnings estimates,” but the P/E multiple also reflects rates, risk premia and growth expectations; it should not be interpreted directly as a discount rate applied to earnings.
  • Combined Nov. 12 markets: KOSPI +₩370.8bn plus KOSDAQ −₩352.5bn = +₩18.3bn (based on reports such as News Tomato).
Why some sources differ
  • Settlement date versus trade date: The Jan–Sep ₩214.0tn total and monthly table use FSS settlement-date data. October's ₩19.5tn and Nov. 2–6's ₩10.7tn are on a trade-date basis, so they are not connected into the same chart.
  • Foreign flows in August: FSS settlement basis (KOSPI and KOSDAQ listed equities): +₩344bn; Korea Exchange trade basis (KOSPI): about −₩10.2tn. Different coverage and timing produce opposite signs.5, 43
  • First week of November: Daishin quant data give ₩10.7tn for Nov. 2–6 and ₩10.3tn for the four sessions Nov. 3–6. No. 29 used ₩10.32tn (Daishin strategy and Hana Securities). The periods and measurement bases differ.
  • U.S.-linked accounts in January: Asia Economy wrote “U.S. net buying of ₩7.8tn,” but the same article's regional figure shows ₩8.1tn in net selling from the Americas. The table treats the U.S. figure as net selling.
  • Foreign flows on Nov. 10: ₩358.6bn (Polinews), ₩369.6bn (Daishin Securities), approximately ₩360bn (Eugene Investment & Securities)
  • Foreign flows on Nov. 11: ₩449.1bn (EBN, 15:39 KST), ₩655.2bn (Opinion News, 15:56), ₩658.7bn (Polinews and The Korea Economic Daily, after the close). Earlier figures were provisional.44, 45
  • Foreign flows on Nov. 12: −₩244bn (Gukje News, 15:50), +₩150.7bn (Aju News, 15:56), +₩158bn (Etoday, 15:50), +₩347.5bn (Block Media, 15:55), +₩370.8bn (News Tomato, 15:59; M Economy News, 16:00). Immediate post-close figures were provisional and sometimes differed even in direction.3, 39, 46
  • Foreign purchases of the two chipmakers on Nov. 11: Samsung Electronics ₩537.5bn and SK hynix ₩130.6bn (The Korea Economic Daily), versus about ₩535.8bn and ₩128.2bn (Green Post Korea). Counting methods vary slightly.
  • Samsung Electronics buyback: Daishin Securities: “halted at ₩14.7tn in early October”; Oct. 12 completion filing: 56,856,305 shares acquired between Aug. 24 and Oct. 6.
  • Market value of foreign holdings: FSS end-September: ₩2,295.97tn (34.8% of market cap), versus Daishin's “currently about ₩1,645tn” after the market decline, versus ₩1,326.8tn at end-2025.
  • Cumulative sales through end-July: News Tomato: “about ₩195tn” (the original article apparently misstated the currency as dollars).
  • U.S.-linked accounts in Jan–Sep: Total −₩65.1tn excluding April, when only the Americas region's −₩1.6tn was disclosed
  • U.K.-linked accounts in Jan–Sep: May is undisclosed; adding only published months produces −₩97.9tn, versus Daishin Securities' total of −₩97.5tn.
  • VKOSPI: 65.2 on Nov. 6 (Daishin, apparently a closing reading) versus an intraday 64.74 (SBS Biz, 10:25)47 ; Nov. 11 close of 52.76 (Polinews and The Herald Business)
  • Hana Securities dividend yield: Summary: 3.4%; table: 3.35% (rounding)
Glossary
Account-country classification (FSS foreign-investment statistics)
The country assigned to holdings and flows in FSS foreign-investor data. If an overseas broker's account aggregates orders from multiple investors, positions may be attributed to the broker's country rather than the ultimate investors' residence. Daishin Securities interprets the classification using the location of the settlement account.
prime broker
A division of a large investment bank that finances hedge funds, lends securities and handles trading and settlement. Many prime-brokerage desks operate out of London.
total return swaps (TRS) and contracts for difference (CFDs)
Derivatives that provide economic exposure to a stock's price changes without necessarily owning the shares. They can support large positions with relatively little margin. When the dealer hedges its exposure by trading shares, those trades may appear under the dealer's account.
Leveraged ETF rebalancing
Trading near the close by 2× or 3× leveraged ETFs to restore their target daily leverage. These funds generally buy more after up days and sell more after down days; their required trades can be estimated from daily returns.
Volatility-targeted allocation
A strategy that cuts equity exposure as volatility rises and adds exposure when volatility falls. Funds and pension investors use such rules to stay within a predetermined risk budget.
settlement-date basis
Figures recorded on the date cash and securities settle, rather than the trade date (typically two business days later). FSS monthly statistics use this basis.
Trade-date basis
Figures recorded on the day an order is executed. Korea Exchange daily investor-flow data use this basis, so they may differ from FSS monthly statistics.
IMF Coordinated Portfolio Investment Survey (CPIS)
The IMF survey in which countries report their residents' holdings of foreign portfolio equities and bonds (CPIS). The focus is the residence of the reporting investor.
non-residents
An investor who does not reside in the country in question—for example, a U.S. or Hong Kong hedge fund using a U.K. account.
leveraged
Taking a position larger than invested capital by borrowing or using derivatives. When prices fall, the investor may need to post additional margin or reduce the position.
VKOSPI
The 30-day forward-looking volatility implied by KOSPI 200 option prices. A higher figure signals expectations of larger market swings.
Other corporations
A company outside the financial sector. When a corporation repurchases its own shares, those transactions can fall under this category.
12-month forward P/E
Share price divided by expected earnings per share over the coming 12 months. A low multiple can suggest cheap shares, but can also reflect fears that earnings are near a cyclical peak.
U.S. Commodity Futures Trading Commission (CFTC)
The U.S. regulator of futures markets. It publishes weekly positioning reports by trader category.
open interest
The number of outstanding futures contracts that have not yet been closed or settled.
Passive index-tracking capital
Funds that seek to replicate an index such as MSCI's. When constituent weights change, they trade mechanically—often at the scheduled closing auction—to match the new weights.
CDS
A credit-default swap (CDS) spread, the market price of protection against a borrower's default. Higher spreads generally signal greater perceived credit risk.
Sources

Data cut-off: after the Korean market close on Nov. 12, 2026 (KST). Most sell-side flow and valuation figures are as of Nov. 6.

  1. Polinews, “KOSPI and KOSDAQ Rebound More Than 4% on Bargain Hunting; Foreigners Buy for First Time in Eight Sessions,” Nov. 10, 2026 · https://www.polinews.co.kr/news/articleView.html?idxno=748187
  2. The Korea Economic Daily, “After the Selling Wave, Foreigners Pile ₩660bn into These Stocks,” Nov. 11, 2026 · https://www.hankyung.com/article/2026111106306
  3. News Tomato, “Two-Day Rebound Ends as KOSPI Slips to 5,340 and KOSDAQ Falls Below 700,” Nov. 12, 2026 · https://www.newstomato.com/ReadNews.aspx?no=1320250
  4. Polinews, “Closing Market: Chip Stocks Weather Rate Shock; KOSPI +2.6%, KOSDAQ +5.6%,” Nov. 11, 2026 · https://www.polinews.co.kr/news/articleView.html?idxno=748264
  5. Financial Supervisory Service, “Foreign Securities Investment Trends,” monthly reports Jan–Sep 2026. Media coverage: Asia Economy, Feb. 13 · The Herald Business, Mar. 26 · The Herald Business, Apr. 15 · The Herald Business, May 28 · The Herald Business, Jun. 25 · Newsis, Jul. 30 · YTN 8.21 · News Tomato, Aug. 21 · SBS Biz 9.18 · SBS Biz 10.30
  6. SBS Biz, “Foreigners Net Buy ₩1.5tn of Korean Stocks in December; Hold Around 30% of Market Cap,” Jan. 9, 2026 · https://biz.sbs.co.kr/amp/article/20000283669
  7. Tiger Brokers(Deep News), 「Samsung and SK Hynix Rally Forces Funds to Sell as Holdings Hit Limits, Sparking Record Foreign Outflows」, 2026.05.29 · https://www.itiger.com/news/1153238900
  8. Money Today, “Foreigners Sell ₩14.5tn This Month—A Historic Exit? Why Some Say Not to Worry,” May 13, 2026 · https://www.mt.co.kr/stock/2026/05/13/2026051218520456568
  9. Maeil Ilbo, “U.K.-Linked Accounts Sold ₩97tn of Korean Stocks as Funds Also Leave Domestic U.K. Investments,” Nov. 10, 2026 · https://www.m-i.kr/news/articleView.html?idxno=2016716
  10. Daishin Securities, “Review of Foreign Selling and Re-entry Scenarios under Volatility Constraints” (Quant), Nov. 10, 2026
  11. Financial Services Commission, “Guidelines to Allow Foreigners to Invest in Korean Equities through Overseas Brokers' Domestic Accounts Without Opening Separate Local Accounts” (press release), Nov. 27, 2025 · https://www.fsc.go.kr/no010101/85748
  12. News1, “Wall Street Curbs Leveraged Bets on Korea's Surging AI Chipmakers Samsung and SK hynix” (citing Bloomberg), Jun. 12, 2026 · https://www.news1.kr/amp/world/international-economy/6195916
  13. Reuters(Yahoo Finance), 「Asia stock-picking hedge funds set for record monthly loss, Goldman says」, 2026.07.30 · https://finance.yahoo.com/markets/stocks/articles/asia-stock-picking-hedge-funds-082032839.html
  14. Metro Newspaper, “Foreign Investors Net Sell ₩2.5tn of Korean Stocks after Ten Months,” Sep. 13, 2024 · https://www.metroseoul.co.kr/article/20240913500133
  15. SBS Biz, “Foreigners Dump ₩14tn of Korean Stocks in Biggest Monthly Outflow on Record,” May 16, 2025 · https://biz.sbs.co.kr/amp/article/20000234930
  16. Asia Economy, “Foreigners Buy ₩4tn of Korean Stocks in October, Extending Net-Buying Streak to Six Months,” Nov. 14, 2025 · https://core.asiae.co.kr/article/2025111315115362356
  17. Tiger Brokers(Stock News), 「AI deleveraging shock fades? Costs for leveraged SK hynix trades drop sharply, global banks reopen doors」, 2026.08.20 · https://www-web.itiger.com/news/1147028139
  18. Edaily, “Foreigners Sell a Record ₩42tn Amid AI Fears—A Major Exodus from This Market,” Nov. 2, 2026 · https://edaily.co.kr/News/Read?mediaCodeNo=257&newsId=02824086645641784
  19. Hankyung Business, “Was the U.K. Behind the KOSPI Crash?” (citing Korea Investment & Securities' Nov. 5 note, “Foreign Flows and U.K.-Linked Selling: The Butterfly Effect of U.K. Inheritance-Tax Reform”), Nov. 7, 2026 · https://magazine.hankyung.com/business/article/202611060202b
  20. Financial Conduct Authority, 「Retirement income market data 2025/26」 · https://www.fca.org.uk/data/retirement-income-market-data-2025-26
  21. Daishin Securities, “Quentin Strategy: Easing AI Concerns and Macro Pressures; KOSPI and KOSDAQ Buy-Side Circuit Breakers,” Nov. 10, 2026
  22. Green Post Korea, “KOSPI Gains 2.6% Even as Foreigners Sell 3,805 Futures Contracts—Focus on Nov. 12 Options Expiry,” Nov. 11, 2026 · https://www.greenpostkorea.co.kr/news/articleView.html?idxno=308012
  23. Saturday Economy, “No New Korean Stocks in MSCI Review; SK hynix's ₩580bn Flow Becomes the Main Variable,” Nov. 12, 2026 · https://sateconomy.co.kr/news/view/1065588902832841
  24. Etoday, “Foreign Selling Doubles in Speed, with ₩20tn Dumped in Samsung and SK hynix,” Nov. 6, 2026 · https://www.etoday.co.kr/news/view/2644716
  25. Norton Rose Fulbright, 「UK Pensions Briefing: Budget October 2024」 · https://www.nortonrosefulbright.com/en/knowledge/publications/2ccb90a5/uk-pensions-briefing-budget-october-2024
  26. Reuters(IFC Review), 「Hedge funds post largest purchases of Asian stocks in 10 years」, 2026.02.16 · https://www.ifcreview.com/news/2026/february/hedge-funds-hedge-funds-post-largest-purchases-of-asian-stocks-in-10-years/
  27. Reuters(The Star), 「Hedge funds' selling of emerging Asian market stocks last week was most in a year, says Goldman」, 2026.03.23 · https://www.thestar.com.my/business/business-news/2026/03/23/hedge-funds039-selling-of-emerging-asian-market-stocks-last-week-was-most-in-a-year-says-goldman
  28. Asia Economy, “Foreigners Sell Korean Stocks for Second Straight Month but Keep Buying Bonds,” Oct. 11, 2024 · https://core.asiae.co.kr/article/2024101014165741466
  29. Asia Economy, “Foreigners Sell Korean Stocks for Third Month, Including More Than ₩4tn in KOSPI Shares,” Nov. 8, 2024 · https://core.asiae.co.kr/article/2024110714334781138
  30. Metro Newspaper, “Fourth Consecutive Month of Foreign Selling: ₩4.1tn in Listed Equities,” Dec. 10, 2024 · https://www.metroseoul.co.kr/article/20241210500136
  31. Yonhap News Agency, “Who Led the KOSPI Correction? U.K. First, Japan Second in Foreign Net Sales,” Nov. 25, 2025 · https://v.daum.net/v/20251125104448553
  32. investingLive, 「SK Hynix swap costs, which spiked to 1000bp in June, now said to halve」, 2026.08.20 · https://investinglive.com/stock-market-update/sk-hynix-swap-costs-which-spiked-to-1000bp-in-june-now-said-to-halve/
  33. Hanwha Investment & Securities, “Orange Strategy: Visibility into Resolving, Rather Than Worsening, Bottlenecks,” Nov. 11, 2026
  34. The Fact, “New Curbs on Concentrated Samsung and SK hynix Leveraged Bets: Minimum Deposits Raised from ₩10m to ₩30m,” Jul. 16, 2026 · https://news.tf.co.kr/read/economy/2344296.htm
  35. DealSite, “Filing Roundup: Samsung Electronics Acquires ₩15tn of Treasury Shares,” Oct. 12, 2026 · https://dealsite.co.kr/articles/170496
  36. SBS Biz, “SK hynix Announces an Additional ₩60tn in Shareholder Returns via Dividends and Buybacks,” Oct. 28, 2026 · https://biz.sbs.co.kr/amp/article/20000342561
  37. Kiwoom Securities, “No Consensus Yet on a 50bp December FOMC Rate Hike” (Kiwoom Daily), Nov. 11, 2026
  38. CoreWeave, Form 8-K (Q3 earnings release, mirrored by StockTitan), Nov. 10, 2026 · https://www.stocktitan.net/sec-filings/CRWV/8-k-core-weave-inc-reports-material-event-d9367aabcf7e.html
  39. M Economy News, “KOSPI Falls for First Time in Three Days on Institutional Selling, Closes Near 5,342,” Nov. 12, 2026 · https://www.m-economynews.com/news/article.html?no=72941
  40. Hana Securities, “A Strong Signal to Scale into Positions—Is Now the Time?” (Practical Quant), Nov. 11, 2026
  41. MSCI, “MSCI November 2026 Index Review Announcement” (press release) · https://www.msci.com/downloads/documents/press-releases/media-room/MSCI_Nov_2026_IR_Pre-Ann_PR_Doc.pdf
  42. U.S. Federal Reserve, “Meeting Calendars and Information” · https://www.federalreserve.gov/monetarypolicy/fomccalendars.htm
  43. SBS Biz, “Foreigners Keep Selling in August; Year-to-Date Net Sales Reach ₩169tn,” Sep. 1, 2026 · https://biz.sbs.co.kr/amp/article/20000332046
  44. EBN, “KOSPI Gains 2.60% on Foreign Buying as Samsung and SK hynix Both Rally,” Nov. 11, 2026 · https://www.ebn.co.kr/news/articleView.html?idxno=1731281
  45. Opinion News, “Foreign Buying Lifts KOSPI Back Above 5,370; KOSDAQ Surges 5.62%,” Nov. 11, 2026 · http://www.opinionnews.co.kr/news/articleView.html?idxno=146957
  46. Block Media, “KOSPI Retreats to 5,340 on Institutional Selling; KOSDAQ Breaks Below 700,” Nov. 12, 2026 · https://www.blockmedia.co.kr/archives/1164941
  47. SBS Biz, “VKOSPI Jumps as KOSPI Falls Below 5,000, Reaching a Three-Month High,” Nov. 6, 2026 · https://biz.sbs.co.kr/amp/article/20000344701