If you spend any time researching tactical asset allocation, you will encounter the term “dual momentum” repeatedly. It is one of the most well-researched and widely used concepts in quantitative investing, forming the backbone of several popular TAA strategies. But what exactly is dual momentum, and how do the main implementations differ?
What Is Momentum?
Momentum in investing is the empirically observed tendency for assets that have performed well recently to continue performing well, and for assets that have performed poorly to continue performing poorly. This is not a fringe idea—the momentum factor has been documented in academic research across equities, bonds, commodities, currencies, and real estate, in dozens of countries, over more than a century of data.
The simplest form of momentum is relative momentum (also called cross-sectional momentum): compare the returns of several assets over a lookback period and buy the ones with the strongest performance. If US stocks returned 12% and international stocks returned 5% over the past year, relative momentum favors US stocks.
Absolute momentum (also called time-series momentum) asks a different question: is an asset’s return positive compared to a benchmark, typically cash? If US stocks returned 12% and Treasury bills returned 4%, US stocks have positive absolute momentum (12% > 4%). If US stocks returned -5%, they have negative absolute momentum, signaling a potential downtrend.
What Is Dual Momentum?
Dual momentum combines both types. First, use relative momentum to select the strongest asset from a set of candidates. Then, apply an absolute momentum filter: if the selected asset has negative absolute momentum, move to a defensive position (typically bonds or cash) instead.
The dual approach addresses the main weakness of each type alone. Relative momentum picks winners but will keep you in a declining asset class if everything is falling. Absolute momentum catches downtrends but gives no guidance on which asset to hold when conditions are favorable. Together, they provide both offensive selection and defensive protection.
Gary Antonacci formalized this framework in his 2014 book Dual Momentum Investing and the associated Global Equities Momentum (GEM) strategy. Since then, several researchers have proposed modifications. Let us compare the three most prominent implementations.
GEM: Global Equities Momentum
How It Works
GEM is the original and simplest dual momentum strategy. Each month, it compares the 12-month total return of US stocks (SPY) against international stocks (VEU/EFA). The winner is selected via relative momentum. Then, if the winner’s 12-month return is below that of Treasury bills (BIL), the portfolio moves entirely to US aggregate bonds (AGG/BND).
At any given time, GEM holds exactly one ETF. There is nothing to optimize, no parameters to tweak (the 12-month lookback is fixed from the published research), and trading happens at most once per month.
Historical Performance (1974-2025)
| Metric | GEM | 60/40 Benchmark |
|---|---|---|
| CAGR | ~14.5% | ~9.6% |
| Max Drawdown | ~-19.6% | ~-32.5% |
| Sharpe Ratio | ~0.75 | ~0.54 |
| Annual Turnover | ~150% | ~5% |
| Avg. Positions | 1 | 2 |
Strengths and Weaknesses
GEM’s greatest strength is simplicity. One ETF at a time, one lookback period, no ambiguity. It is easy to understand, easy to verify, and easy to execute. The academic foundation is solid.
The weakness is whipsaw sensitivity. A single 12-month lookback means GEM can be slow to react to rapid market changes, and it can flip-flop between positions when momentum is borderline. It also concentrates the entire portfolio in one asset class, which means a bad signal leads to full exposure to the wrong asset.
ADM: Accelerating Dual Momentum
How It Works
Accelerating Dual Momentum, developed by Chris at EngineNo1, modifies GEM in a key way: instead of using a single 12-month lookback, ADM uses a weighted average of multiple lookback periods (typically 1, 3, 6, and 12 months), with shorter periods weighted more heavily. This means recent momentum has more influence on the signal than distant momentum.
The asset universe is also slightly broader than GEM. ADM typically considers US stocks, international developed stocks, and emerging market stocks for the offensive side, with aggregate bonds as the defensive position. Some variants extend to additional asset classes.
Historical Performance (1974-2025)
| Metric | ADM | GEM |
|---|---|---|
| CAGR | ~15.2% | ~14.5% |
| Max Drawdown | ~-17.8% | ~-19.6% |
| Sharpe Ratio | ~0.80 | ~0.75 |
| Annual Turnover | ~200% | ~150% |
| Avg. Positions | 1-2 | 1 |
Strengths and Weaknesses
ADM’s multi-period momentum scoring makes it more responsive to regime changes. When a crash starts, the shorter lookback periods drag the score down faster than a pure 12-month lookback would, triggering an earlier exit. Historically, this has translated to slightly better drawdown protection and faster recovery.
The trade-off is higher turnover and more parameter choices. With four lookback periods and their relative weights, there are more degrees of freedom than GEM. This raises the question of overfitting—are the parameter choices robust, or are they optimized to look good in backtests? Walk-forward analysis (testing on out-of-sample data) is essential for evaluating ADM variants.
CDM: Composite Dual Momentum
How It Works
Composite Dual Momentum takes a different approach to solving GEM’s concentration problem. Instead of holding one ETF, CDM runs dual momentum independently across multiple asset pairs and combines the results. For example, it might run separate dual momentum comparisons for US vs. international equities, US vs. emerging markets, and stocks vs. REITs, then allocate to the winners from each pair.
The defensive mechanism works at the pair level: if the winner of a pair has negative absolute momentum, that pair’s allocation goes to bonds. This means CDM can be partially offensive and partially defensive at the same time, unlike GEM and ADM which are all-or-nothing.
Historical Performance (1974-2025)
| Metric | CDM | GEM |
|---|---|---|
| CAGR | ~13.1% | ~14.5% |
| Max Drawdown | ~-14.2% | ~-19.6% |
| Sharpe Ratio | ~0.78 | ~0.75 |
| Annual Turnover | ~120% | ~150% |
| Avg. Positions | 3-5 | 1 |
Strengths and Weaknesses
CDM’s diversification across multiple pairs significantly reduces max drawdown compared to GEM. Because it can be partially in offensive assets and partially in bonds at the same time, it avoids the binary all-in or all-out problem. The lower turnover is also advantageous for tax efficiency.
The cost is lower absolute returns. By diversifying across pairs, CDM dilutes the strongest momentum signal. In strong trending markets, GEM’s concentrated bet on the single best asset outperforms. CDM is the more conservative dual momentum approach.
Side-by-Side Comparison
| Feature | GEM | ADM | CDM |
|---|---|---|---|
| Complexity | Low | Medium | Medium |
| Positions Held | 1 | 1-2 | 3-5 |
| Regime Responsiveness | Slow | Fast | Moderate |
| Drawdown Protection | Good | Better | Best |
| Return Potential | Highest | High | Moderate |
| Whipsaw Risk | Higher | Moderate | Lower |
| Tax Efficiency | Low | Low | Moderate |
| Backtest Overfitting Risk | Minimal | Moderate | Low |
Which Should You Choose?
Choose GEM if you want maximum simplicity and the highest return potential, and you can tolerate the occasional whipsaw and deeper drawdowns. GEM is the best starting point for someone new to dual momentum because there is nothing to configure wrong.
Choose ADM if you want faster reaction to market regime changes and are comfortable with slightly more parameters. ADM is a good choice for investors who have experience with GEM and want an evolution that addresses its lag problem.
Choose CDM if drawdown minimization is your top priority and you are willing to accept lower absolute returns for smoother ride. CDM is also better for larger portfolios where the diversification across multiple positions improves execution quality.
Or blend them. Running GEM, ADM, and CDM together in equal weights gives you exposure to three different interpretations of dual momentum with partially decorrelated signals. When GEM and ADM disagree on direction, CDM’s diversified approach provides stability. BestFolio tracks all three strategies (plus many more dual momentum variants) and shows their correlations to help you build effective blends.
Beyond Dual Momentum
Dual momentum is a powerful foundation, but it is not the only approach. Strategies like Vigilant Asset Allocation (VAA), Hybrid Asset Allocation (HAA), and Defensive Asset Allocation (DAA) extend the concept with breadth indicators, canary assets, and multi-factor scoring. If you find that dual momentum resonates with your investment philosophy, these advanced strategies are worth exploring as you grow more comfortable with tactical allocation.