Data through September 2026. Trading costs included.
Ten years of one idea
Starting in 2016, Keller and Keuning published a run of tactical strategies that all ask the same question: when should a momentum portfolio stop taking risk? Each one answered it a little differently, and most of what came later was built on what came before.
| Strategy | Published | What it added |
|---|
| PAA | 2016 | Protective momentum: the more assets are falling, the more goes to bonds |
| GPM | 2016 | Protective momentum generalized, with correlation in the ranking |
| VAA-A | 2017 | Breadth momentum: one failing asset is enough to go defensive |
| DAA | 2018 | The canary universe: a small set of assets that only decides when to take risk |
| LAA | 2019 | A mostly static portfolio with one sleeve switched by growth and trend |
| BAA-B and BAA-A | 2022 | Separate offense, defense and canary universes, with a slower momentum filter |
| HAA | 2023 | The simplest of the line: one canary asset, one momentum formula |
Each strategy has its own page with the full rules and its record against its benchmark. This page puts all eight side by side on the same months, with 60/40 and the S&P 500 for reference.
What the record shows
From Feb 1980 through Sep 2026, the eight strategies compounded at between 10.1% a year (LAA) and 15.7% (BAA-A) after trading costs, against 9.8% for 60/40.
Their worst declines ranged from -9.6% (HAA) to -21.6% (VAA-A), against -32.3% for 60/40. Every one of them fell less than 60/40 at its worst.
On the Ulcer Performance Index, which weighs return against the depth and length of drawdowns, HAA leads at 3.70. All eight beat 60/40, which scores 0.90.
Through the global financial crisis, Nov 2007 to Feb 2009, results ran from +9.0% (BAA-B) to -8.0% (LAA), while 60/40 returned -32.3%.
2022, when stocks and bonds fell together, was the harder test for strategies that hide in bonds: three of them made money, with results from +2.8% (HAA) to -14.0% (LAA).
Since 2000
The modern half of the record, with the dot-com bust, 2008, the long zero-rate decade and the 2022 rate shock all inside it.
| Jan 2000 to Sep 2026 | CAGR | Max DD | Ulcer | UPI | Sortino | MAR |
|---|
| HAA - Hybrid Asset Allocation | 11.5% | -9.6% | 2.69 | 3.58 | 2.35 | 1.19 |
|---|
| BAA - Bold Asset Allocation - Balanced | 9.9% | -11.8% | 3.89 | 2.07 | 2.14 | 0.84 |
|---|
| DAA - Defensive Asset Allocation | 8.4% | -20.2% | 6.18 | 1.05 | 1.71 | 0.41 |
|---|
| VAA - Vigilant Asset Allocation - Aggressive | 10.3% | -21.6% | 7.71 | 1.10 | 1.72 | 0.48 |
|---|
| PAA - Protective Asset Allocation | 8.0% | -11.0% | 3.17 | 1.93 | 2.02 | 0.73 |
|---|
| BAA - Bold Asset Allocation - Aggressive | 11.9% | -19.0% | 5.76 | 1.74 | 1.67 | 0.63 |
|---|
| GPM - Generalized Protective Momentum | 7.5% | -11.7% | 3.79 | 1.49 | 1.66 | 0.64 |
|---|
| LAA - Lethargic Asset Allocation | 8.8% | -18.8% | 5.06 | 1.37 | 1.67 | 0.47 |
|---|
| 60/40 | 6.6% | -32.3% | 7.62 | 0.62 | 1.11 | 0.20 |
|---|
| S&P 500 | 8.1% | -51.0% | 15.89 | 0.39 | 0.88 | 0.16 |
|---|
Return by decade
| Annual return | 1980s* | 1990s | 2000s | 2010s | 2020s* |
|---|
| HAA - Hybrid Asset Allocation | +22.5% | +13.5% | +14.5% | +7.7% | +12.8% |
|---|
| BAA - Bold Asset Allocation - Balanced | +18.7% | +11.1% | +14.5% | +7.5% | +7.0% |
|---|
| DAA - Defensive Asset Allocation | +17.5% | +15.2% | +11.5% | +7.1% | +5.8% |
|---|
| VAA - Vigilant Asset Allocation - Aggressive | +20.5% | +13.9% | +20.3% | +7.1% | +1.4% |
|---|
| PAA - Protective Asset Allocation | +18.6% | +11.4% | +10.5% | +6.0% | +7.3% |
|---|
| BAA - Bold Asset Allocation - Aggressive | +22.8% | +19.2% | +18.1% | +10.0% | +6.0% |
|---|
| GPM - Generalized Protective Momentum | +17.7% | +13.3% | +9.9% | +6.6% | +5.4% |
|---|
| LAA - Lethargic Asset Allocation | +9.8% | +14.0% | +6.8% | +9.5% | +10.9% |
|---|
| 60/40 | +14.6% | +14.0% | +2.1% | +9.4% | +9.4% |
|---|
| S&P 500 | +16.5% | +18.1% | -1.0% | +12.9% | +15.3% |
|---|
* Partial decade, annualized over the months available.
Through the hard stretches
Total return over each period, start of the first month to the end of the last.
| Period | HAA | BAA-B | DAA | VAA-A | PAA | BAA-A | GPM | LAA | 60/40 | S&P 500 |
|---|
| 1987 crashSep 1987 to Nov 1987 | +1.3% | +1.8% | -10.5% | +0.7% | -12.5% | +1.8% | -10.1% | -16.0% | -18.5% | -29.8% |
|---|
| Dot-com bear marketSep 2000 to Sep 2002 | +11.2% | +17.5% | +12.3% | +15.1% | +19.8% | +5.9% | +26.6% | -11.5% | -23.5% | -44.8% |
|---|
| Global financial crisisNov 2007 to Feb 2009 | -1.7% | +9.0% | -5.9% | -0.4% | -0.7% | +7.2% | +2.9% | -8.0% | -32.3% | -51.0% |
|---|
| COVID crashFeb 2020 to Mar 2020 | +2.7% | +4.7% | +2.1% | +4.4% | +1.9% | +4.7% | +6.1% | -7.7% | -11.9% | -19.6% |
|---|
| 2022 rate shockJan 2022 to Dec 2022 | +2.8% | +0.9% | -12.4% | -13.1% | -6.8% | +1.3% | -5.7% | -14.0% | -15.9% | -18.2% |
|---|
How it behaves month to month
| Since Feb 1980 | Worst 12 months | Positive months | Positive when S&P 500 fell | Correlation to S&P 500 |
|---|
| HAA - Hybrid Asset Allocation | -9.1% | 70% | 44% | 0.45 |
|---|
| BAA - Bold Asset Allocation - Balanced | -10.9% | 70% | 52% | 0.29 |
|---|
| DAA - Defensive Asset Allocation | -14.4% | 67% | 38% | 0.53 |
|---|
| VAA - Vigilant Asset Allocation - Aggressive | -15.9% | 65% | 46% | 0.31 |
|---|
| PAA - Protective Asset Allocation | -10.9% | 69% | 44% | 0.44 |
|---|
| BAA - Bold Asset Allocation - Aggressive | -14.3% | 68% | 52% | 0.31 |
|---|
| GPM - Generalized Protective Momentum | -8.7% | 67% | 52% | 0.26 |
|---|
| LAA - Lethargic Asset Allocation | -15.2% | 64% | 23% | 0.78 |
|---|
| 60/40 | -27.5% | 66% | 10% | 0.97 |
|---|
The strategies
- HAA - Hybrid Asset Allocation: Hybrid Asset Allocation, 2023. One TIPS canary decides between eight global assets and Treasuries.
- BAA - Bold Asset Allocation - Balanced: Bold Asset Allocation, Balanced, 2022. Wider offensive universe, slower momentum filter.
- DAA - Defensive Asset Allocation: Defensive Asset Allocation, 2018. Introduced the canary universe.
- VAA - Vigilant Asset Allocation - Aggressive: Vigilant Asset Allocation, Aggressive, 2017. Breadth momentum with a fast weighted signal.
- PAA - Protective Asset Allocation: Protective Asset Allocation, 2016. The first of the line.
- BAA - Bold Asset Allocation - Aggressive: Bold Asset Allocation, Aggressive, 2022. The concentrated version, with VAA's offensive assets.
- GPM - Generalized Protective Momentum: Generalized Protective Momentum, 2016. Protective momentum with correlation in the ranking.
- LAA - Lethargic Asset Allocation: Lethargic Asset Allocation, 2019 (Keller). Mostly static, with one sleeve switched by growth and trend.
- 60/40: 60% US stocks, 40% US bonds, rebalanced annually.
- S&P 500: US large-cap stocks, held without rebalancing.
Questions
Who are Keller and Keuning?
Wouter Keller and Jan Willem Keuning are researchers who have published a series of tactical asset allocation strategies since 2016, including PAA, VAA, DAA, BAA and HAA. Dual Momentum Systems is not affiliated with them.
What is a canary universe?
A small set of assets that does not get invested in but decides when the strategy takes risk. When enough canaries show negative momentum, the strategy moves to its defensive assets. Defensive Asset Allocation (DAA) introduced the idea in 2018, and the later strategies in the line use it.
Which Keller and Keuning strategy has the highest return?
In this backtest, from Feb 1980 through Sep 2026, BAA-A had the highest annual return, 15.7% after trading costs. The figures update every month.
Which Keller and Keuning strategy has the smallest drawdown?
HAA, whose worst decline was -9.6% over the same months.
Which has the best risk-adjusted return?
On the Ulcer Performance Index, HAA scores highest at 3.70.
Which is the newest Keller and Keuning strategy?
Hybrid Asset Allocation (HAA), published in 2023. It is also the simplest of the line, with a single canary asset and one momentum formula.
Why are these numbers different from the papers?
This backtest uses DMS's own data, fund conventions and trading-cost model, and runs every strategy over the same months. The papers use their own data, funds and periods.
About the data
All portfolios run on the same monthly return series with the same trading-cost model and are rebalanced as each one specifies. Returns before an ETF's launch come from a matching index, as explained in where the data comes from. Before the ETFs they hold launched, the strategies' histories rely on matching index returns. The authors' own published results differ because they use different data sources, funds and cost assumptions.
Backtests are hypothetical and are not a promise of future results. This page is educational and is not investment advice.