What foreign net buying means — and why outlets print different numbers
Foreign net buying is the amount foreign investors bought minus the amount they sold over a period, usually one session, and the published figure varies sharply between outlets because tallies cover different markets — main board only, or including the KOSDAQ, off-exchange and basket trades
The three lines
- Net buying = purchases minus sales, measured in value, not trade count, and normally per session
- One session in August 2026 was reported as anywhere from 2.106tn to 3.706tn won of foreign net buying — a coverage difference, not an error
- Foreign flow moves the index not because it is large but because it concentrates in the heaviest index constituents
Key questions
- What does 'foreign net buying' mean?
- It means foreign investors bought more value than they sold over a period, usually one trading session. Buy 10bn won and sell 6bn and net buying is 4bn; the reverse is net selling. The key point is that it measures direction, not activity. A session with enormous turnover can still show net buying near zero if purchases and sales roughly match.
- Why do outlets publish different numbers for the same day?
- Because they are counting different things. Whether the tally covers the main board only, includes the KOSDAQ, or includes off-exchange and basket trades changes the total. On August 13, 2026 Korean outlets reported foreign net buying for the same session as 2.106tn, 2.1187tn, 2.4tn and 3.7058tn won. The 1.6tn won spread is not measurement error — it is different coverage. So read the market basis before the number, and only compare figures from the same source and basis.
- Retail investors trade far more. Why does coverage focus on foreigners?
- Because of where the money goes, not how much of it there is. Foreign flow concentrates in the largest listed companies, and the KOSPI is market-cap weighted, so movement in the heaviest names is movement in the index. Retail flow spreads across a far larger number of mid- and small-cap names, contributing less to the index. August 13, 2026 is the textbook case: foreign buying pushed into electronics and the KOSPI rose 3.56%, while the retail-heavy KOSDAQ managed 0.29%.
- So if foreigners buy, does the market go up?
- There is a statistical tendency, not a guarantee. One analysis put the correlation with KOSPI moves at 0.54 for foreign investors, 0.35 for institutions and -0.7 for retail. But correlation does not establish direction. It cannot separate foreigners buying and thereby lifting prices, foreigners buying because prices were about to rise, and both reacting to the same news. Counter-examples are easy to find: on August 10, 2026 foreigners sold a net 1.4887tn won on the KOSDAQ and the index rose 6.97%.
Almost every Korean market close report contains the same sentence: "Foreign investors bought a net X trillion won."
This page writes it too. But on August 13, 2026, the value differed by outlet. 2.106tn, 2.1187tn, 2.4tn, 3.7058tn won. The spread between lowest and highest is 1.6tn won.
That is not an error. They counted different things.
1. The definition
| Term | Calculation | Meaning |
|---|---|---|
| Purchase value | — | total bought |
| Sale value | — | total sold |
| Net buying | purchases − sales > 0 | tilted toward buying |
| Net selling | purchases − sales < 0 | tilted toward selling |
The point is that it measures direction, not activity.
If foreigners bought 10tn won and sold 9.9tn won, net buying is 100bn. Turnover was heavy; the change in position was small. Conversely a quiet session with one-sided buying registers a large net figure.
Investor groups are conventionally reported in four categories.
| Group | Composition |
|---|---|
| Retail | individual investors |
| Foreign | holders of a foreign investor registration |
| Institutional | securities firms, asset managers, pension funds, insurers, banks, private funds |
| Other corporate | domestic corporates outside the above |
"Institutional" is not one thing. Pension funds buying while securities firms sell is routine. A line reading "institutions bought a net 681.5bn won" can contain constituent groups moving in opposite directions and netting out.
2. Why the numbers diverge — coverage
This is the practically useful part of the article.
| Coverage | Included? |
|---|---|
| Main board (KOSPI) | almost always |
| KOSDAQ | varies by outlet |
| Off-exchange trades | varies by outlet |
| Basket and block trades | varies by outlet |
A session reported as both 2tn-something and 3tn-something is most plausibly explained by that combination. Net buying of 2.106tn won on the main board can become a 3tn-something figure once the KOSDAQ and off-exchange trades are added.
Two rules for reading follow.
① Find the basis inside the article. A well-written close report says "on the main board."
② Compare like with like. To set today's figure against yesterday's, both must come from the same outlet on the same basis. Reading a main-board figure against an all-market figure and concluding that foreign buying grew is simply wrong.
The safest option is the Korea Exchange's own trading-by-investor-type data. Even there, the tables are split by market, so which table you are reading still matters.
3. Why foreign flow moves the index
Measured by turnover, retail investors often dominate. Yet index commentary leads with foreign flow. The reason is where the money lands.
| Group | Where the money goes | Index impact |
|---|---|---|
| Foreign | concentrated in the largest listed names | large |
| Institutional | large caps plus index tracking | medium |
| Retail | spread widely across mid- and small-caps | relatively small |
The KOSPI is market-capitalisation weighted. Larger companies move the index more. So 1tn won buying Samsung Electronics and SK Hynix and 1tn won spread across 200 small caps have entirely different index effects.
August 13, 2026 is the textbook case.
| Metric | August 13 |
|---|---|
| KOSPI | +3.56% |
| KOSDAQ | +0.29% |
| SK Hynix | +5.92% |
| Samsung Electronics | +4.89% |
Foreign buying concentrated in electronics; the KOSPI has names of that weight and the KOSDAQ does not. The two indexes diverged twelvefold on the same day. The session is covered in "KOSPI closes at 6,813.34 on August 13, 2026."
4. Net buying is not the same as gains
One private analysis put correlations with KOSPI moves as follows.
| Group | Correlation |
|---|---|
| Foreign | 0.54 |
| Institutional | 0.35 |
| Retail | -0.7 |
Do not read the retail figure as "when retail buys, prices fall." Most trades have a counterparty — on a day of heavy foreign buying, retail is structurally on the sell side. That is closer to accounting symmetry than to causation.
And correlation says nothing about direction. Whether foreigners bought and prices rose, or prices were about to rise and foreigners bought, or both responded to the same news, is not separable from this number.
Clear counter-examples exist.
| Date | What happened |
|---|---|
| August 10, 2026 | Foreigners sold a net 1.4887tn won on the KOSDAQ; the index rose 6.97% |
| August 14, 2026 | Institutions sold a net 1.03tn won; the KOSPI rose 2.42% |
The first was retail buying carrying the index; the second was a single foreign bid absorbing two selling groups. Both are covered in "KOSDAQ closes at 854.47 on August 10, 2026" and "KOSPI closes at 6,977.94 on August 14, 2026."
5. What to read alongside it
A single net-buying line supports a firmer reading when paired with these.
| Metric | Why |
|---|---|
| Consecutive sessions | persistence of direction says more than one day's size |
| Sector breakdown | index impact depends on whether the buying hit electronics |
| Won-dollar rate | foreign inflows create won demand; if the won weakens on heavy buying, other flows offset it |
| Futures positioning | cash and futures pointing opposite ways may indicate hedging |
| The KOSDAQ, same day | the fastest control group for whether a rally is broad or narrow |
On August 14, 2026 foreigners bought a net 3.0387tn won and the won strengthened 1.10 won — internally consistent. On August 13 they bought more than 2tn won and the won weakened 3.7 won, which means something else was offsetting.
6. What is still open
The sample period and method behind the correlation figures were not verified. They are reproduced from a single private analysis.
Outlets were not asked directly about their tally bases. Off-exchange and basket coverage is the leading inference offered across sources.
The legal definition of 'foreign' is not the same as the origin of the capital. Classification follows foreign investor registration, so domestic money routed through an offshore entity registers as foreign. This article does not quantify that share.
Other mechanisms that move the index are set out in "What triggers a sidecar" and "What shareholder returns mean."
Sources
- Korea Exchange — KRX Data Marketplace, trading by investor type
- Hankyung Markets — Foreign trading
- Daum Finance — Foreign and institutional trading
- Goodkyung — Why does the market rise when foreigners buy?
- KB Asset Management — Foreign vs retail vs institutional: whose buying moves prices?
- Newspim — Market close: KOSPI up a fifth day on 3tn won of foreign buying