Summary: An “index” stops being diversified when a few stocks dominate it. In 2026 Samsung and SK Hynix grew to about half of the KOSPI (usually closer to a quarter for its top names) only because they were the stocks retail investors could leverage; when they fell, the whole index fell with them. The S&P 500 is also unusually concentrated: seven tech companies make up ~36% of it, above the ~25% of the dot-com peak. Yet concentration alone has never been a reliable bubble signal: Swiss, Taiwanese and Israeli indexes have long been dominated by one or two names, and patient investors still came out ahead. Know what is inside what you buy, and diversify across companies and sectors.
Key ideas
- Concentration is never good; the KOSPI effectively became a 2–3 stock index (lazy-investing—108-the-leverage-trap, tamir).
- Israel’s TA-35 caps any single stock at 7%, a rule introduced after Teva grew to dominate the index (lazy-investing—108-the-leverage-trap).
- Seven tech companies make up ~36% of the S&P 500 and ~32% of the whole US market, versus ~25% at the dot-com peak (lazy-investing—111-the-ai-bubble, tamir).
- Concentration alone has never been a factor for a bubble or for poor returns: Swiss and Taiwanese indexes are very concentrated, and when Teva crashed and dragged the TA-35 down, long-term investors still made money as the other companies (banks etc.) grew (lazy-investing—111-the-ai-bubble, tamir).
Disagreements & open questions
- How dangerous is concentration? In #108 Tamir called it “never good”, tying it to the KOSPI crash (lazy-investing—108-the-leverage-trap); in #111 he argues it has never by itself signaled a bubble or poor long-term returns (lazy-investing—111-the-ai-bubble). The difference may be leverage: KOSPI concentration came from leveraged retail flows.
Takeaways
- Check the top holdings and their weights in any index fund you buy (lazy-investing—108-the-leverage-trap)