Does Following Foreign Investors' Net Purchases Yield Returns? — A 4-Year, 986-Trading Day Test in the Korean Market

"Buy the stocks that foreign investors are buying" is one of the most widely accepted pieces of advice in the Korean stock market. We tested this with actual data. To start with the conclusion — there was a return that looked like it came from following them, but that return did not come from the directional information that 'foreign investors bought'.

Test Methodology

• Data: KRX daily investor trading trends data from 2022-08 to 2026-08, about 2.5 million rows, 986 trading days.
• Signal: Top 20 stocks by foreign investors' net purchase amount over the recent 5 trading days (filtered for a daily average trading value of 10 billion KRW or more and a stock price of 1,000 KRW or more).
• Entry & Exit: Entry at the closing price of the next trading day after the signal occurs (to block look-ahead bias), and exit at the closing price after holding for 20 trading days.
• Evaluation: Excess return compared to the market average (equal-weighted) over the same period. Reflects a round-trip trading cost of 0.25%.
• Sample: 49 non-overlapping samples (separated so that holding periods do not overlap).

Results

The key is the second row. Even when buying stocks that foreign investors heavily sold, the return was +1.74%p — which shows no difference from buying by following them (+1.81%p). The long-short difference between buying and selling being +0.03%p means that the direction itself—"foreign investors bought/sold"—contained no information regarding subsequent returns.

What These Numbers Mean

The true nature of the +1.8%p that looked like an excess return is interpreted as a trade concentration effect. A large net purchase amount by foreign investors means that trading was concentrated on that stock, and the result that stocks with concentrated trading (ignoring direction, top |net purchases|) were the highest at +2.47%p supports this. It is simply that stocks with concentrated attention and volatility beat the market during this period (2022~2026), not that the foreign investors' 'directional choice' was correct.

Also, the excess return of stocks with concentrated trading is difficult to use as a basis for a long strategy — the return distribution of highly volatile stocks has thick tails on both sides, and this sample period was generally a bull market.

Limitations

• Cross-event market correlation can inflate the t-value. We adjusted for this using non-overlapping samples, but it is not perfect.
• The 4-year sample period (2022-08~2026-08) may be dependent on specific market conditions.
• The test unit is a 20-day holding period. The results may be different for other holding periods, and this was not tested.

Data Sources

• KRX daily investor trading trends data. The daily raw data for investor flows can be found in the investor flows report, and the cumulative investor flows by foreign investors and institutions for each stock can be viewed on each stock page.