Fifty Years of Data Crown No Single Winner, but a Home-Heavy Mix Wins on Risk

Which portfolio would have served investors best over the last half century? A new analysis of BullionVault annual return data from 1972 to 2025 offers an answer, and a warning about trusting it.

The study compared nine asset classes in real, inflation-adjusted terms: U.S. stocks, international stocks, REITs, U.S. corporate bonds, 10-year Treasuries, 3-month T-bills, gold, commodities and U.S. homes.

On raw growth, nothing came close to the S&P 500. One dollar invested in 1972 would have grown to $35.90 by 2025. Gold turned $1 into $12.89, REITs into $12.26 and international stocks into $11.27. Commodities lost ground, shrinking $1 to $0.42 despite annual volatility of 14%. T-bills, the least volatile option, managed just $1.14.

Gold was the most erratic performer. It topped the table in 13 of 54 years, roughly one year in four, but it also tied commodities for the most last-place finishes, with 11 each. Its best year was a 106% gain in 1979 and its worst a 37.6% drop in 1981, giving it the highest standard deviation of the group at 24.7%.

Inflation told a mixed story. In years when CPI ran above 4%, commodities held up best. Stocks, international equities, REITs, corporate bonds and Treasuries all did worse, with returns about 8 percentage points lower or more. Gold's average return rose by 5.6 points in high-inflation years, but its median fell by 7.7 points. That suggests its inflation-hedge reputation rests on a handful of outsized years such as 1973, 1974 and 1979.

When the author searched for the mix with the best Sharpe ratio, the result was 36% U.S. homes, 27% S&P 500, 18% gold, 16% 10-year Treasuries and 3% REITs. Homes earn a big slice because their prices are marked to market infrequently, so they look steady while still returning well.

That portfolio averaged a 4.7% real annual return with 6.6% volatility and a maximum drawdown of about 13%, in 2022. The S&P 500 averaged 8.3% with 16.8% volatility and a 48% drawdown in 1974. Using geometric returns, the gap narrows to 6.9% for stocks versus 4.5% for the optimal mix, because volatility costs the S&P 500 about 1.4 points a year against 0.2 for the blend.

Leveraging the optimal mix to match stock returns is hard in practice, since it requires cheap borrowing and the ability to survive margin calls. Housing is the exception, as a fixed-rate mortgage cannot be margin called. The author estimates the median homeowner already holds 51% of net worth in home equity.

Removing housing, the optimal mix becomes 43% S&P 500, 26% gold, 22% Treasuries and 9% REITs. That looks like an aggressive version of the Permanent Portfolio, which splits 25% each among stocks, gold, bonds and cash.

The key caveat is stability. Run the same exercise on 1972 to 1998 only, and the answer is 68% S&P 500, 24% gold, 6% international stocks and 2% corporate bonds, with no homes, REITs or Treasuries at all. The author's conclusion is that the best portfolio is always a solution to yesterday's world. The practical optimum, they argue, is whichever allocation an investor can hold through whatever comes next.