Permanent
Four equal parts, one for each economic climate Harry Browne named: stocks for prosperity, long-term Treasuries for deflation, gold for inflation and Treasury bills for recession.
Designed by Harry Browne (1987; Fail-Safe Investing, 1999).
Before the numbers. A hypothetical backtest, not a forecast: Jun 30, 2000 to Oct 9, 2026 (26.2 years), the window every published portfolio shares. Rebalanced once a year at 10 basis points a trade; Harry Browne's own rule was different: checked once a year, rebalanced when any holding is below 15% or above 35%.. The bar is the three-fund at the same stock share, 25%.
Funds younger than the window are extended by older series, each junction validated (how): BIL by TBILL.IRX before 2007-05-30; GLD by GC=F before 2004-11-18, then CEF before 2000-08-30; VGLT by VUSTX before 2010-01-04; VOO by VFINX before 2010-09-09. Gold before August 2000 is the Central Fund of Canada (CEF), which held silver as well as gold.
What it holds
| Asset class, and the fund this backtest holds it through | Weight | Fund fee |
|---|---|---|
| US large-cap blend VOO · Vanguard S&P 500 |
25% | 0.03% |
| Long Treasuries VGLT · Vanguard Long-Term Treasury |
25% | 0.04% |
| Treasury bills (cash) BIL · SPDR 1-3 Month T-Bill |
25% | 0.14% |
| Gold GLD · SPDR Gold Shares |
25% | 0.40% |
| Together | 100% | 0.15% |
The funds are the ones this backtest uses for each asset class, not a list to buy: any fund tracking the same asset class holds the same thing.
Its record against the three-fund at 25% stocks
| Statistic | Permanent | Three-fund at 25% stocks |
|---|---|---|
| Value per $10,000 | $56,203 | $35,429 |
| Total return | 462.03% | 254.29% |
| Annual return (CAGR) | 6.79% | 4.93% |
| Standard deviation | 6.81% | 5.42% |
| Max drawdown | -16.55% | -19.24% |
| Worst calendar year | -11.73% | -14.37% |
| Longest time underwater | 546 days | 713 days |
| Ulcer Index | 3.64% | 4.01% |
| Sharpe ratio | 0.73 | 0.57 |
| Sortino ratio | 1.18 | 0.84 |
| Ulcer Performance Index | 1.35 | 0.77 |
The crises
| Deepest decline from a prior peak | Permanent | Three-fund at 25% stocks |
|---|---|---|
| 2000–02 | -6.6% | -4.5% |
| 2008 | -15.2% | -16.0% |
| 2020 | -10.3% | -11.8% |
| 2022 | -16.6% | -19.2% |
2000–02 is measured from June 2000, the window's start, after the March peak; 2008 from October 2007 to June 2009; 2020 from February to April; 2022 through 2023.
Calendar years
The first and last years are partial.
The trade-off
Browne’s argument was that no one can predict which climate comes next, so the portfolio holds something for each, in equal parts, and leaves it alone. Only a quarter is in stocks. He checked it once a year and rebalanced only when a part fell below 15% or rose above 35%; here it is rebalanced every year, like every row.
Over this window its worst decline was -16.6%, against -19.2% for the three-fund at 25% stocks. What it gives up is the long-run growth of a stock-heavy portfolio: half of it is in gold and cash, which earn little in long, calm bull markets.