Correlations: the trap of false diversification
Correlation measures whether two assets move together. If your 10 holdings rise and fall at the same rhythm, you are not diversified: you own the same risk under ten different names. The line between real protection and false safety comes down to a single number, sitting between -1 and +1.
True diversification for a FIRE portfolio
True diversification is not about multiplying holdings, but about combining assets with low correlation. A three-position portfolio of well-chosen assets (for example: world stocks + bonds + gold) can be more diversified than twenty tech stocks. Worth remembering: think in asset classes, not in number of holdings.
The correlation scale (-1 to +1)
+1 = perfectly correlated: both assets always move in the same direction. 0 = no correlation: movements are independent. -1 = perfectly inverse: when one goes up, the other goes down. In practice, perfect correlation never exists, but the trends are clear.
Real-world correlation examples
US stocks / European stocks: ~0.85 (highly correlated, no real geographic diversification). Stocks / Government bonds: ~-0.3 (slightly inverse, a good diversifier). Stocks / Gold: ~0.05 (nearly independent, an excellent diversifier). Bitcoin / Nasdaq: ~0.55 since 2021 (so much for the «crypto decoupling» story).
The crisis trap: correlations spike
In normal times, asset classes move to different rhythms. In a crisis, they all converge toward +1. In 2008, stocks, listed real estate, and commodities fell together. Only government bonds and cash held up. Practitioners call it «correlation breakdown»: it is precisely the moment when naive diversification stops protecting you.
Key Takeaways
- 1Correlation +1 = identical moves; 0 = independent; -1 = opposite (perfect diversification).
- 2During crashes (2008), correlations converge to 1: this is the «correlation breakdown».
- 3US stocks vs European stocks: ~0.85. Geographic diversification across developed markets is limited.
- 4Three truly uncorrelated holdings beat twenty tech stocks correlated at 0.95 with each other.
Keep going
Frequently asked questions
Correlation measures whether two assets rise and fall together. +1 = identical moves (two MSCI World ETFs); 0 = independent; -1 = opposite (very rare). The less correlated two assets are, the more combining them reduces overall portfolio volatility without sacrificing expected return.
'Correlation breakdown' phenomenon: in 2008, stocks, listed real estate, commodities and high-yield all fell together toward correlations near +1. Only short government bonds (US Treasuries, Bund) and cash kept their decorrelation. Diversification protects against the everyday, not the systemic.
Around +0.85 over a 20-year rolling window. Developed markets are tightly integrated: a US crash contaminates Europe within hours. Geographic diversification across developed markets is more marketing than reality. For real decorrelation, add emerging markets (~0.65 vs S&P) or gold (~0.1).
Compute the correlation matrix between monthly returns of your lines over 5+ years. Tools like Portfolio Visualizer or our simulator generate it automatically. Aim for an average correlation <0.7 between your main lines to benefit from real statistical diversification.