Skip to content
TEN Brief Ten verified stories a day 2026.09.03 KO

이 기사는 한국어로도 읽을 수 있습니다 →

Economy · 2 min read · Reference

What institutional net selling means in Korea — seven actors behind one number

When Korean market coverage says institutions sold two trillion won, the single number hides at least seven different actors: securities firms trading their own capital, asset management companies, private equity funds, insurers, banks, other financial firms and pension funds. They buy and sell for unrelated reasons over unrelated time horizons. Selling from a securities firm's own account is frequently the offsetting leg of a derivatives hedge or an ETF creation and redemption rather than any view on the market, while pension funds such as the National Pension Service hold positions for years and are often net buyers on sharp down days because a falling market pushes equity below its target weight. Company share buybacks, meanwhile, are not recorded under institutions at all but under other corporations — which is why that line showed 1.65 trillion won of net buying on September 2, 2026, as Samsung Electronics and SK hynix ran large buyback programmes

A sunlit reading room with a long wooden table, stacks of documents and an open notebook, green trees beyond tall windows

The three lines

  • Composition — institutions bundles securities desks, funds, insurers, banks and pensions
  • Reading — a securities desk's selling can be hedging or ETF flow, not a market view
  • Distinction — company buybacks appear under other corporations, not institutions

Key questions

What does institutional net selling mean?
**It means that for a given day, institutional investors' sales exceeded their purchases.** The Korea Exchange publishes daily flows split into four headline categories: **individuals, foreign investors, institutions and other corporations**. The third is a **container, not an actor**. Inside it sit securities firms trading their own capital, asset management companies running public funds, private equity funds, insurers, banks, other financial firms and pension funds. So 'institutions sold two trillion won' means **the sum across seven kinds of participant was negative two trillion**, not that any decision-maker chose to sell that amount. On September 2, 2026, institutions net sold 2.04 trillion won of KOSPI stock — and that number alone cannot tell you whether pension funds were buying or selling inside it.
If institutions sell, do prices fall?
**They often fall the same day, but reading it as cause is a frequent mistake.** The reason is the largest sub-category, **securities firms' own accounts**, whose trading is regularly unrelated to any market view. Three common cases: **derivatives hedging**, where a firm buys or sells cash equities to offset structured products or futures positions; **ETF creation and redemption**, where investor demand for an ETF forces the underlying basket to be bought and investor exits force it to be sold; and **arbitrage** between futures and cash, which is a spread trade with no directional opinion. **Pension funds** run the opposite way. They hold for years against fixed target weights, so when equity prices fall sharply their equity allocation drops below target and they have to **buy** to restore it — which is why pensions frequently print as net buyers on the worst days. The practical conclusion: institutional net selling is **an accounting total, not a signal**.
Where do company buybacks show up?
**Under other corporations, not institutions.** That category covers non-financial companies, and a company purchasing its own shares lands there. Missing this distinction inverts the reading of a session. September 2, 2026 is a clean example: foreigners −1.92 trillion won, institutions −2.04 trillion, individuals +2.30 trillion and **other corporations +1.65 trillion**. Behind that last line sat the buyback programmes of Samsung Electronics (15 trillion won through November 21) and SK hynix (40 trillion won through November 19). Without knowing the classification, the natural conclusion would be that retail investors absorbed the entire four trillion won of selling. In fact the issuers themselves absorbed a large share. One caution: other corporations also contains ordinary corporate stake transactions, so **the whole line cannot be attributed to buybacks**.

"Institutions sold two trillion won."

It appears in Korean market coverage every day. But "institutions" is not one investor. It is at least seven.

1. The four headline categories

CategoryWho
IndividualsRetail investors
Foreign investorsRegistered foreign participants
InstitutionsA bundle of domestic financial institutions and pensions
Other corporationsNon-financial companies — where buybacks land

The problem lives in row three. It is the name of a basket, not of a participant.

2. Opening the basket

Sub-categoryCharacterWhy it trades
Securities firms' own accountsProprietary capitalHedging · ETF creation/redemption · arbitrage
Asset managementPublic fundsFund inflows and outflows
Private equity fundsPrivate vehiclesStrategy-specific
InsurersLife and non-lifeLong-horizon asset-liability management
BanksBank accountsTreasury operations
Other financialSpecialty financePortfolio operations
Pension fundsNational Pension Service and peersMaintaining target weights

The first and last rows carry most of the interpretive weight, and they behave in opposite ways.

3. Securities desks: this selling is not an opinion

SituationWhat appears in the cash market
Hedging structured products or futuresDirection-neutral buying or selling
ETF creation (investors buy the ETF)Underlying basket bought
ETF redemption (investors sell the ETF)Underlying basket sold
Futures-cash arbitrageTwo-sided trading on the spread

So "securities firms sold" rarely means "securities firms are bearish." Money leaving an ETF produces selling even when nobody has changed their mind about anything.

4. Pension funds: buyers on the worst days

FeatureConsequence
Multi-year horizonDaily moves are not a trigger
Fixed target weights per asset classA price fall pushes equity below target
Weights must be restoredFrequently prints as buying into sharp declines

This is why the pension line can run opposite to the rest of its own basket. The aggregate hides it completely.

5. Reading a real session: September 2, 2026

The day the KOSPI fell 3.99 percent:

ParticipantNet
Foreign investors−1.92tn won
Institutions−2.04tn won
Individuals+2.30tn won
Other corporations+1.65tn won

The last row is the one most readers skip. Behind it were the buyback programmes of Samsung Electronics (15 trillion won through November 21) and SK hynix (40 trillion won through November 19).

Buybacks are other corporations, not institutions. Without that, the obvious story is "institutions dumped and retail caught it." The fuller story is that the issuers themselves absorbed a large part of the four trillion won.

Even so, the line also contains ordinary corporate stake transactions, so it cannot be read as buybacks in full.

6. Three working rules

RuleReason
① Do not treat institutional net selling as a directional signalIt is a sum across seven kinds of participant
② Look at sub-categories where they are publishedPensions and securities desks can move in opposite directions
③ Always read the other-corporations line alongside itDuring buyback programmes it is the main flow in the market

7. What is still open

  • Disclosure — sub-category flows are published unevenly by stock and date.
  • Buyback share — the portion of other corporations attributable to buybacks is not disclosed.
  • Labels — category names differ slightly between exchange statistics and brokerage screens. This page follows the exchange's classification.

Sources

  1. KRX — Korea Exchange Data Marketplace (investor trading trends)
  2. Opinion News — Explaining the two forces that move the market: institutions and foreigners
  3. Samsung Securities — Investor category screen guide
  4. Kiwoom Securities — Intraday investor trading screen guide
  5. Money Today — Foreigners and institutions dump 4 trillion won, KOSPI slides 4% (September 2, 2026 flows)

Verification

Published
Last modified
Cross-check
Checked against 5 independent sources.
Unverified
  • Disclosure of sub-category flows varies by stock and by date and is not established here
  • The share of the other-corporations line attributable to buybacks is not disclosed
  • Sub-category labels differ slightly between exchange statistics and brokerage screens
Authoring
Reviewed by a person before publication. The full process is described in the Editorial.

Ten stories, once each morning

We send the three-line summaries only; the full pieces stay on the site. One-click unsubscribe, any time.

Related