Defensive Asset Allocation (DAA)
Developed by Keller & Keuning · Canary Universe · Med Risk
Defensive Asset Allocation was published by Wouter Keller and Jan Willem Keuning in their 2018 SSRN paper (#3212862), introducing what has become one of the most influential architectural innovations in tactical allocation research: the canary universe. Rather than using the same assets for both crash detection and investment, DAA separates these functions entirely — dedicating specific sentinel assets to the sole task of monitoring market health while maintaining a completely independent set of assets for portfolio construction.
This separation was a direct response to a conceptual weakness Keller identified in earlier strategies, including his own VAA. When the same assets serve as both danger sensors and investment targets, a circularity problem emerges: an asset's declining momentum triggers defensive action, but that same decline is partly what the investor would need to detect and avoid. By isolating the detection function in purpose-selected canary assets, DAA breaks this circular dependency and creates a cleaner signal architecture that has influenced numerous subsequent strategies in the tactical allocation literature.
The two canary assets — emerging market equities (VWO) and US aggregate bonds (BND) — were selected because they tend to exhibit weakness ahead of US equity declines through different transmission channels. Emerging markets function as a gauge of global risk appetite: when international capital begins flowing out of higher-risk emerging economies, it typically signals a broader de-risking cycle that will eventually reach developed market equities. Aggregate bonds capture interest rate and credit stress: when bond prices decline due to rising rates or widening credit spreads, it often precedes equity weakness as tightening financial conditions feed through to corporate earnings and valuations.
Unlike VAA's binary all-or-nothing trigger, DAA implements a graduated response based on how many canaries are signaling stress. When both canaries are healthy, the portfolio is fully invested. When one turns negative, half shifts to defense. When both are negative, full defensive mode activates. This graduated approach reflects a more nuanced risk assessment — acknowledging that stress in a single market segment represents a warning that may or may not escalate, while simultaneous stress across both segments constitutes stronger evidence of broad environmental deterioration.
How It Works
The Canary Universe
Two canary assets — emerging markets (VWO) and aggregate bonds (BND) — are scored monthly using the 13612W weighted momentum composite, the same multi-timeframe formula used in VAA. These canary assets serve exclusively as sensors — they never receive portfolio allocation. This separation ensures the canary signal remains uncontaminated by position management concerns and prevents the circularity problem where an asset's declining price simultaneously triggers a sell signal and represents a loss the investor has already incurred.
The choice of two canaries rather than one provides coverage across different risk channels. VWO detects risk appetite deterioration — capital flight from emerging markets, currency stress, and trade flow disruptions. BND captures rate and credit conditions — rising yields, widening spreads, and liquidity tightening. Together they provide advance warning across the two primary transmission mechanisms through which market stress typically propagates before reaching US large-cap equities.
Graduated Defensive Response
The canary count determines the portfolio's defensive posture on a three-step scale. When both canaries show positive momentum, zero percent of the portfolio is defensive. When one canary turns negative, fifty percent shifts to defense. When both are negative, one hundred percent goes defensive. This graduated response avoids the all-or-nothing transitions that produce whipsaw losses in binary systems.
The fifty-percent partial defense during single-canary alerts is a distinctive design choice. It acknowledges uncertainty — one negative canary suggests the environment may be deteriorating but doesn't confirm it. By reducing risk exposure proportionally rather than eliminating it entirely, the strategy maintains some upside participation during periods of ambiguous signals. If the second canary subsequently turns negative, confirming the stress, the portfolio completes its shift to full defense. If the first canary recovers, the portfolio returns to full offense without having missed the entire intervening period of returns.
Asset Selection
During risk-on periods, four assets — US stocks (SPY), international developed stocks (EFA), emerging market stocks (EEM), and aggregate bonds (AGG) — are ranked by their 13612W momentum scores. The top-scoring asset receives the risk-on allocation. For the defensive portion, the strategy selects the strongest performer among three safe-haven assets: short-term Treasuries (SHY), intermediate-term Treasuries (IEF), and investment-grade corporate bonds (LQD).
The four-asset offensive universe is deliberately compact, covering the major global equity regions plus bonds as a potential risk-on holding during equity downturns. The inclusion of bonds in the offensive universe is noteworthy — it allows the strategy to hold bonds as a momentum-driven position during deflationary environments or rate-cutting cycles, rather than relegating them exclusively to the defensive role. This dual use of fixed income increases the strategy's adaptability across different macroeconomic regimes.
Explore Defensive Asset Allocation (DAA)
See the full backtest across 18 years of market data, or run your own what-if scenarios by adjusting all research parameters.