Gold Cross-Asset Momentum (Quantpedia gold and Treasury rule)
Cyril Dujava's gold timing rule from Quantpedia. At each month end it holds gold only while both gold and 7–10 year Treasuries have a positive 12-month total return; if either is flat or negative it holds Treasury bills until the next month end.
Designed by Cyril Dujava (Quantpedia), 2026. Implemented and tracked by Tactfolio.
| May 2007 – Sep 2026 | Strategy | SPY |
|---|---|---|
| Annual return (CAGR) | 7.3% | 10.7% |
| Worst drawdown | -33.7% | -55.2% |
| Sharpe ratio | 0.55 | 0.62 |
| Volatility | 14.8% | 19.7% |
| Annual return since publication (Feb 2026) | — | — |
| May 1969 – Sep 2026 | Strategy | SPY |
|---|---|---|
| Annual return (CAGR) | 10.7% | 10.7% |
| Worst drawdown | -34.6% | -55.2% |
| Sharpe ratio | 0.75 | 0.68 |
| Volatility | 15.0% | 17.2% |
| Annual return since publication (Feb 2026) | — | — |
Gold Cross-Asset Momentum is a gold timing rule that Cyril Dujava of Quantpedia published in January 2026 in his article Cross-Asset Price-Based Regimes for Gold. At each month end it holds gold only if both gold and 7–10 year Treasuries are up over the past twelve months; otherwise it waits in Treasury bills.
The idea#
Gold and Treasuries often respond to the same force: falling real interest rates. When bonds are rising, holding gold costs less in forgone interest, and gold has tended to do well. When bonds are falling, rates are usually climbing, and gold has had a harder time even when its own trend looked fine.
Dujava sorted decades of monthly data into four groups by the direction of gold's and bonds' 12-month returns. Gold did best when both were positive. Using bond momentum as a second key removes some of the months gold's own trend would have held, and most of the removed months were weak ones.
How it works#
At the close of the last trading day of each month:
- Measure the 12-month total return of gold (GLD) and of 7–10 year Treasuries (IEF).
- If both are above zero, hold 100% GLD for the next month.
- If either is zero or negative, hold 100% Treasury bills (BIL).
What the backtest shows#
The ETF-era test starts in mid-2007, when the Treasury bill fund begins. Since then the strategy returned about 7% a year, against about 11% for the S&P 500 and 8% for a 60/40 stock and bond mix. Its worst drawdown, about a third, was shallower than the S&P 500's but slightly deeper than the 60/40's, and its risk-adjusted return trailed both.
It held up in 2008, losing 7% while the S&P 500 lost 37%, and in 2022 it made a small gain while stocks fell 18%. Its best year was 2025, up 54%. The cost of holding one volatile asset showed from 2012 through 2016, when it lost money five years in a row. That stretch produced its worst drawdown, from August 2011 to December 2016, and it did not regain the 2011 high until July 2020.
It has been live for less than a year since publication in early 2026, and in that time it has fallen as much as a quarter from its peak. That is too short a record to judge.
With simulated history the test starts in 1969 and looks stronger: about 11% a year, level with the S&P 500, with a worst drawdown near 35% against 55% and a better risk-adjusted return. Most of that came from the gold booms of the 1970s, with gains of 94% in 1973 and 68% in 1979. In the 1990s it mostly moved a few percent a year either way.
When it struggles#
- Long gold bear markets. When gold drifts lower for years, as it did after 2011, the rule steps in and out and gives back money on each attempt.
- Sharp reversals. A 12-month signal checked monthly holds gold through sudden drops, such as its fall of about a quarter in 2026.
- As a whole portfolio. It is either all gold or all cash, so it can lag stocks for a decade. It makes more sense as a small sleeve.
Using it on Tactfolio#
The live strategy above runs these rules with the source's own funds; the only change is that cash earns the Treasury bill fund's return instead of the federal funds rate. Copy it to try a shorter lookback, or place it as the gold sleeve inside a larger portfolio.
Year by year
| Year | Strategy | SPY |
|---|---|---|
| 2026* | -0.7% | 14.0% |
| 2025 | 53.7% | 17.7% |
| 2024 | 16.7% | 24.9% |
| 2023 | 4.8% | 26.2% |
| 2022 | 1.4% | -18.2% |
| 2021 | -10.5% | 28.7% |
| 2020 | 24.8% | 18.3% |
| 2019 | 8.4% | 31.2% |
| 2018 | 2.5% | -4.6% |
| 2017 | 8.0% | 21.7% |
| 2016 | -7.7% | 12.0% |
| 2015 | -6.2% | 1.2% |
| 2014 | -3.9% | 13.5% |
| 2013 | -0.7% | 32.3% |
| 2012 | -0.5% | 16.0% |
| 2011 | 9.6% | 1.9% |
| 2010 | 19.3% | 15.1% |
| 2009 | 21.8% | 26.4% |
| 2008 | -7.0% | -36.8% |
| 2007* | 25.8% | -3.4% |
| Year | Strategy | SPY |
|---|---|---|
| 2026* | -0.7% | 14.0% |
| 2025 | 53.7% | 17.7% |
| 2024 | 16.7% | 24.9% |
| 2023 | 4.8% | 26.2% |
| 2022 | 1.4% | -18.2% |
| 2021 | -10.5% | 28.7% |
| 2020 | 24.8% | 18.3% |
| 2019 | 8.4% | 31.2% |
| 2018 | 2.5% | -4.6% |
| 2017 | 8.0% | 21.7% |
| 2016 | -7.7% | 12.0% |
| 2015 | -6.2% | 1.2% |
| 2014 | -3.9% | 13.5% |
| 2013 | -0.7% | 32.3% |
| 2012 | -0.5% | 16.0% |
| 2011 | 9.6% | 1.9% |
| 2010 | 19.3% | 15.1% |
| 2009 | 21.8% | 26.4% |
| 2008 | -7.0% | -36.8% |
| 2007 | 30.4% | 5.1% |
| 2006 | 36.5% | 15.8% |
| 2005 | 16.3% | 4.8% |
| 2004 | 4.0% | 10.7% |
| 2003 | 19.9% | 28.2% |
| 2002 | 25.6% | -21.6% |
| 2001 | -3.0% | -11.8% |
| 2000 | 4.1% | -9.7% |
| 1999 | 4.6% | 20.4% |
| 1998 | 4.7% | 28.7% |
| 1997 | 5.0% | 33.5% |
| 1996 | -1.6% | 22.5% |
| 1995 | -0.1% | 38.0% |
| 1994 | -1.0% | 0.4% |
| 1993 | 11.5% | 9.7% |
| 1992 | -2.3% | 7.6% |
| 1991 | 3.1% | 30.3% |
| 1990 | -8.6% | -3.2% |
| 1989 | 8.3% | 31.5% |
| 1988 | -2.0% | 16.4% |
| 1987 | 18.8% | 5.1% |
| 1986 | 19.0% | 18.6% |
| 1985 | 7.5% | 31.6% |
| 1984 | 9.9% | 6.2% |
| 1983 | -15.3% | 22.4% |
| 1982 | 15.5% | 21.6% |
| 1981 | -9.7% | -4.9% |
| 1980 | 62.6% | 32.4% |
| 1979 | 67.6% | 18.3% |
| 1978 | 25.6% | 6.5% |
| 1977 | 25.0% | -7.2% |
| 1976 | 5.0% | 23.8% |
| 1975 | -23.1% | 37.1% |
| 1974 | 51.1% | -26.5% |
| 1973 | 94.2% | -14.7% |
| 1972 | 48.8% | 18.9% |
| 1971 | 16.7% | 14.1% |
| 1970 | 5.5% | 3.9% |
| 1969* | -1.4% | -10.4% |
* Partial year.
The rules as implemented
This is the exact tree Tactfolio runs, rebalanced monthly with signals and trades at the close. Open it to inspect or copy it.
- StrategyDujava Gold Cross-Asset Momentum
- WeightEqual
- IfAll of 2 conditions
- 252d cumulative return of GLD is above 0
- 252d cumulative return of IEF is above 0
Then- WeightEqual
- TickerGLD
Otherwise- WeightEqual
- TickerBIL
- IfAll of 2 conditions
- WeightEqual
Sources and caveats
- Uses the source's own funds, GLD and IEF, and their 12-month total returns, measured over 252 sessions.
- BIL stands in for the source's cash, which earns the effective federal funds rate.
- Signals and trades use the close of the last trading day of each month, as in the source.
Common questions#
What is Gold Cross-Asset Momentum?#
It is Cyril Dujava's 2026 rule for timing gold with two 12-month returns: gold's own and that of intermediate Treasuries. It holds GLD only when both are positive and cash otherwise.
Why does a bond signal help time gold?#
Bond prices rise when interest rates fall, and lower rates reduce the cost of holding gold, which pays no income. Requiring positive bond momentum filters out many months when gold's own trend was up but rates were working against it.
What ETFs does it use?#
GLD for gold, IEF for 7–10 year Treasuries as the signal, and BIL for cash.
How is it different from HAA or US Cross-Asset Momentum?#
All three use bond momentum to decide when to own a riskier asset. HAA Simple watches TIPS to time the S&P 500, and US Cross-Asset Momentum uses Treasuries to time stocks. This rule applies the same idea to gold.