How Much Gold? Backtesting 5%, 10%, and 20% Allocations Across 50 Years

"How much gold should I own?" is the most common question precious-metals writers receive and the one most often answered with vague platitudes. The honest answer requires looking at fifty-plus years of return data and accepting that the optimal allocation depends entirely on what you are trying to optimize for.

The setup

I ran a comparison of four portfolios from January 1972 (when gold became freely tradable post-Bretton Woods) through December 2024:

• 60/40 baseline: 60% S&P 500, 40% US Treasuries
• 5% gold: 57% stocks, 38% bonds, 5% gold
• 10% gold: 54% stocks, 36% bonds, 10% gold
• 20% gold: 48% stocks, 32% bonds, 20% gold

All portfolios rebalanced annually. Returns are nominal, gross of fees and taxes.

The headline numbers

Compound annual returns came out roughly:

• 60/40: 9.4% CAGR, max drawdown -29%
• 5% gold: 9.5% CAGR, max drawdown -27%
• 10% gold: 9.5% CAGR, max drawdown -25%
• 20% gold: 9.4% CAGR, max drawdown -22%

The most striking finding: raw returns barely changed across allocations. Gold did not destroy returns even at 20%, and the drawdown reduction was meaningful but not dramatic.

Where gold actually helped

The averages obscure the real story. Gold's contribution clusters in specific decade-scale crisis windows: 1972-1980 (gold +1700%), 2000-2011 (gold +560%), and 2019-2024 (gold +90%). In the 1980s and 1990s, gold dragged. In the 2010s, gold roughly kept up with bonds.

"Gold is a regime-change asset. It does not pay you across all environments; it pays you when monetary regimes break. The question is whether you believe more breaks are coming." - portfolio manager, sovereign wealth fund

The Sharpe ratio improvement

Risk-adjusted, the 10% gold portfolio delivered the highest Sharpe ratio of the four (0.52 vs 0.48 for the 60/40). The 20% allocation slightly underperformed on Sharpe because gold's volatility added noise without enough additional return to compensate. 10% appears to be the sweet spot in this dataset.

Sequence of returns matters

Retirees withdrawing 4% annually saw materially better outcomes from gold-inclusive portfolios in the 1970s and 2000-2010 periods because gold rallied during the deepest equity drawdowns. The same retirees did slightly worse in the 1990s. Gold is most valuable for spend-down portfolios in the worst-case scenarios, not the average ones.

What this means for you

If you want to optimize raw return, gold allocation barely matters. If you want to reduce tail risk and improve worst-case outcomes, somewhere between 5% and 15% is supported by the data. Above 20%, you are making a directional bet on monetary regime change rather than building a diversified portfolio.

Bottom line: The data supports a 5-15% gold allocation as a defensive position, not as a return engine. Anyone telling you to put 30%+ in gold is making a thesis call dressed up as portfolio theory.