Context
This is a follow-up to our narrow-market, foreign-flow and semiconductor relative-alpha work. Related hub: Exclusive Analysis Hub.
TL;DR
- Korea looks like a roaring bull market at the index level. As of June 19, 2026, KOSPI traded around the 9,050 level and was up roughly 115% YTD.
- Under the index, the market is extremely narrow. In our proprietary 2,730-stock Korea dataset, the median stock return was -7.6%, only 37% of names were up, and only about 30% traded above their 200-day moving average.
- Monte Carlo simulations show how hard the pure KOSPI benchmark was to beat. A random equal-weight 20-stock portfolio beat KOSPI only 1.0% of the time. A 30-stock version beat it only 0.5% of the time.
- Turnover-weighted portfolios did much better, with 10-stock and 30-stock versions beating KOSPI 60.1% and 70.8% of the time. But that was essentially a concentrated exposure to the most popular mega-cap AI-memory winners.
- The conclusion: in 2026, pure KOSPI is not “the average Korean equity market.” It is closer to a concentrated Samsung Electronics / SK Hynix AI-memory benchmark.
The Index And The Median Stock Diverged
The KOSPI level around 9,050 on June 19 is visible on external index screens such as Investing.com and Trading Economics. KOSDAQ, by contrast, was near 966.6 and up roughly 11% YTD on public screens.
Our proprietary breadth dataset shows the internal divergence.
| Metric | Value |
|---|---|
| KOSPI YTD | around +115% |
| KOSDAQ YTD | around +11% |
| Median Korean stock | -7.6% |
| Advancing-stock ratio | 37% |
| Cumulative A-D line | -20,391 |
| Stocks above MA200 | 30% |
| Top-decile return | +88% or more |
| Top-1% return | +537% or more |
The gap between KOSPI and the median Korean stock is roughly 122 percentage points. That is not a normal broad bull market. It is a narrow leadership regime.
Monte Carlo Result
We simulated random Korean equity portfolios from a 2,730-stock universe, using returns from the start of 2026 to June 19. Each case ran 20,000 trials.
| Portfolio type | Holdings | Beat KOSPI | Median portfolio return |
|---|---|---|---|
| Equal-weight random | 5 | 5.3% | +2.6% |
| Equal-weight random | 10 | 2.8% | +9.8% |
| Equal-weight random | 20 | 1.0% | +14.1% |
| Equal-weight random | 30 | 0.5% | +15.9% |
| Turnover-weighted | 10 | 60.1% | +127% |
| Turnover-weighted | 30 | 70.8% | +131% |
Only 7.9% of individual stocks beat the KOSPI return threshold.
The interpretation is straightforward. Diversification reduced benchmark-relative performance because the benchmark itself was heavily concentrated in a tiny winner cohort. Turnover weighting did better because it leaned into the same crowded winners that drove the index.
What It Means
An active account lagging pure KOSPI in 2026 was not necessarily a failed account. It may have simply avoided the extreme concentration embedded in the benchmark.
The right evaluation has three layers:
| Layer | Question |
|---|---|
| Absolute return | Did the account make money? |
| Median-stock comparison | Did it beat the typical Korean stock? |
| Pure KOSPI comparison | Did it beat the mega-cap AI-memory benchmark? |
Those are different questions. A diversified account could do very well in absolute terms and still lag pure KOSPI.
Final View
In 2026, beating the pure KOSPI benchmark required more than good stock picking. It required concentration in the narrow mega-cap AI-memory path, early entry and long holding periods. Diversification was not wrong, but it carried a relative-performance cost.
Coverage: proprietary 2,730-stock Korea dataset, price returns from the start of 2026 to June 19, 2026, and 20,000 Monte Carlo trials per portfolio structure. External index levels checked against Investing.com, Trading Economics, Yahoo Finance KOSDAQ and MarketWatch KOSDAQ.