Golden Butterfly Portfolio: does it still work? Replicated through 2026
We re-ran it on S&P 500, Gold from 2010-01-04 to 2026-11-10 (16.8 years, before costs): +7.5% a year, Sharpe 0.97, worst drawdown -14.6%. The benchmark, S&P 500 buy-and-hold, did +13.8% a year (Sharpe 0.85, worst drawdown -33.7%).
Verdict: Smoother ride than its benchmark (higher Sharpe), but a lower return. Historical, hypothetical results before costs, taxes and fees. Not investment advice.
This is our adapted version, not the real portfolio or fund. Read what is lost before comparing it with anything published about the original.
What is lost
- Both Treasury sleeves (20% long-term, 20% short-term) are cash earning 0%.
- The 20% small-cap value sleeve is folded into SPY, so the value/size tilt is gone.
40% of the original allocation is government bonds. We have no bond data, so that 40% is held as cash earning 0%.
What this means: no coupon income, no price gains when yields fall or stocks crash, and none of the losses bonds had when yields rose (2022). Read the results as the non-bond part plus idle cash. The real portfolio would have done differently, better in some years and worse in others.
Where it comes from
A popular Permanent Portfolio variant with a stock tilt that became a staple of the 'lazy portfolio' community.
Rules, weights and thresholds are given as commonly cited. We have not re-checked them against the primary text.
The original rules
20% each in total US stock market, small-cap value stocks, long-term Treasuries, short-term Treasuries and gold, rebalanced annually.
How we replicate it
What we changed
Small-cap value is folded into SPY (40% SPY total) and both bond sleeves (40%) become cash; gold stays 20%.
The exact rule we run
Buy 40% SPY (20% total market plus 20% standing in for small-cap value) and 20% GLD and keep 40% in cash (0%). Rebalance at the last close of each year.
Target weights: 40.0% S&P 500, 20.0% Gold, 40.0% Cash (0%).
Conventions: Position decided at a close earns the next day's return; no look-ahead; before costs; cash earns 0%; starts 2010-01-04, when all its assets have data; weights drift between rebalances (annually); rebalanced at the last close of the period.
Growth of 1 vs the benchmark
Benchmark: S&P 500 buy-and-hold, same period. Bought on the first day and held.
Drawdown from the previous peak
Key numbers
Sample: 2010-01-04 to 2026-11-10, 16.8 years, 4240 daily observations. Annualised from the number of observations per year of the strategy's own calendar. Before costs.
Recent and full-period returns
| Period | Replica | Benchmark |
|---|---|---|
| Last month | -2.3% | -3.7% |
| Last year | +4.4% | +11.2% |
| Last 5 years | +50.4% | +72.9% |
| Whole period (cumulative) | +236.2% | +788.9% |
Calendar-year returns
| Year | Replica | Benchmark | Replica minus benchmark |
|---|---|---|---|
| 2026* | +3.0% | +10.8% | -7.9% |
| 2025 | +19.8% | +17.7% | +2.1% |
| 2024 | +15.3% | +24.9% | -9.6% |
| 2023 | +13.0% | +26.2% | -13.2% |
| 2022 | -7.4% | -18.2% | +10.8% |
| 2021 | +10.7% | +28.7% | -18.1% |
| 2020 | +12.3% | +18.3% | -6.0% |
| 2019 | +16.1% | +31.2% | -15.2% |
| 2018 | -2.2% | -4.6% | +2.4% |
| 2017 | +11.2% | +21.7% | -10.5% |
| 2016 | +6.4% | +12.0% | -5.6% |
| 2015 | -1.6% | +1.2% | -2.9% |
| 2014 | +4.9% | +13.5% | -8.5% |
| 2013 | +7.3% | +32.3% | -25.1% |
| 2012 | +7.7% | +16.0% | -8.3% |
| 2011 | +2.7% | +1.9% | +0.8% |
| 2010 | +10.5% | +13.1% | -2.6% |
* Partial year. The first year starts on the first day of the replica; the current year is year to date.
Position as of 2026-11-10
| Holding | Weight |
|---|---|
| S&P 500 SPY | 43.1% |
| Gold GLD | 18.1% |
| Cash (0%) | 38.8% |
Weights today, after drifting since the last rebalance on 2025-12-31. Target weights are in the rule above. Drawdown from its peak today: -3.6%.
What it did in 2020 and 2022
| Episode | Dates | Replica | Benchmark |
|---|---|---|---|
| COVID crash | 2020-02-19 → 2020-03-23 | -14.5% | -33.7% |
| 2022 bear market | 2022-01-03 → 2022-10-12 | -11.3% | -24.5% |
Close to close between the two dates. Two episodes only: not enough to judge a strategy.
How much to trust it
- A fixed allocation has no signals, so it is one path of history (16.8 years), not a test.
- This is one backtest on one path of history, before costs and with hindsight on the choice of strategy. It does not show the strategy will work in future.
Honest caveat
Loses the small-cap value factor and bonds, and the publication year is approximate, so it only resembles the original.
Related
Other famous strategies
Historical, hypothetical results before costs, taxes and fees. Not investment advice.