Beyond 60/40: How Adaptive Asset Allocation Outperforms Traditional Portfolios

Summary

Adaptive Asset Allocation (AAA) offers a dynamic, rules-based portfolio strategy designed to deliver steady returns while minimizing downside risk.
AAA stands out for its breadth across nine asset classes, agility via momentum/volatility/correlation scoring, and risk-aware covariance optimization.
The AAA approach has historically outperformed traditional 60/40 and bond portfolios, with a 20-year annual return of 14% and a max drawdown of 19%.
AAA is ideal for investors seeking consistent, risk-adjusted performance and willing to rebalance monthly using tools like M1 Finance or Interactive Brokers

(This article was written by our Summer Finance Intern, Tate P.)

An elusive investing goal is a portfolio that delivers steady, upward returns while minimizing downside risk. Popular asset allocation approaches that pursue this goal include the Balanced 60/40 Portfolio, Ray Dalio's All-Weather Portfolio, and Harry Browne's Permanent Portfolio. One challenge with these static allocation strategies is that they are left vulnerable to swings since they can’t adjust to market cycles.

Tactical (dynamic) portfolios, on the other hand, aim to minimize downside with periodic adjustments to a portfolio's holdings. However, these tactical portfolios can bring some complexity, since designing a tactical approach involves answering several questions, such as:

What is my asset universe? (e.g., asset classes, types of funds, sectors)
What input factors will I use (e.g., fundamentals, momentum, performance, macro metrics, events)
What decision-making framework or algorithm will I use?
How often will I rebalance? (e.g., on a fixed schedule, or whenever a criterion is met)

The Adaptive Asset Allocation [AAA] portfolio is a flexible upgrade to static, set-it-and-forget-it portfolios. It's not as simple as a static portfolio, but it does use a straightforward portfolio-building framework. AAA analyzes risk and return for a range of global asset classes, looks at what’s working, how volatile each asset is, and how the assets move together. Then Adaptive Asset Allocation creates a portfolio allocation that shifts toward the stronger options while dialing back risk. The aim is steadier upside, with smaller drawdowns. If you want a rules-based way to pursue returns without taking wild, swing-for-the-fences risks, AAA is worth a look.

Adaptive Asset Allocation was described by four financial advisors (Adam Butler, Michael Philbrick, Rodrigo Gordillo, and David Varadi) in a 2012 whitepaper titled, “Adaptive Asset Allocation: A Primer.” These authors described a dynamic approach to portfolio construction based on momentum, volatility, and correlation.

AAA has features that make it worth a close look for those looking for a solid tactical portfolio.

Breadth. It's broad; AAA draws from nine different asset classes to have a complete scope of the market
Agility. It's nimble; AAA uses a multi-factor tactical approach. AAA applies a scoring system for momentum, volatility, and correlation (M-V-C), which is used to decide the monthly reallocation.
Risk-awareness. It's analytical. Using covariance optimization, AAA allocates intelligently to reduce overall portfolio risk.
Performance. It’s done well, one AAA portfolio has a 20-year historical total annual return of 14% with a max drawdown of around 19%.

Let's explore AAA a bit more by looking at each of these features individually.

Adaptive Asset Allocation's Breadth

In the 2012 whitepaper, the authors recommend building an "investable universe" that includes a wide range of global asset classes. The advantage of using a broad set of asset classes is that there's "somewhere to run" when conditions change. Well, there may not be shelter in every market storm, but using a range of asset classes allows the AAA portfolio to avoid being ruined by a disturbance in one asset class.

This investable universe can be defined and implemented using asset class ETFs. At RecipeInvesting, we have chosen the following nine ETFs for our Adaptive Asset Allocation universe:

DBC - Broad commodities (energy, metals, agriculture)
EFA - Developed international equities (ex-U.S. and Canada)
EEM - Emerging markets equities
GLD - Gold bullion
TLT - Long-term U.S. Treasuries
SPY - U.S. large-cap equities (S&P 500)
QQQ - U.S. large-cap growth, tech-heavy (Nasdaq-100)
IYR - U.S. real estate (REITs)
IWM - U.S. small-cap equities (Russell 2000)

Adaptive Asset Allocation's Agility

The Adaptive Asset Allocation portfolio uses a multi‑factor scoring system that includes momentum, volatility, and correlation, combined with minimum‑variance optimization.

This combination of momentum, volatility and correlation is central to AAA. Blending these signals and then weighting the resulting portfolio by minimum variance dramatically improves risk‑adjusted returns. To implement an Adaptive Asset Allocation, we'll need to follow these steps:

Choose the investable universe, as discussed above.
Choose the input parameters, including the length of the lookback period (typically several months), the number of ETFs to select from the universe (e.g., the top three or top five), and the rebalance frequency.
Calculate momentum scores to pick the top assets to include in the portfolio.
Run a minimum‑variance optimization to determine the weightings (allocations) to each of the selected assets. This is described in more depth below.

Then rebalance periodically to maximize the portfolio's agility. We should note that this periodic asset turnover can have tax consequences. Caveat venditor.

Adaptive Asset Allocation's Risk-awareness

Once the AAA process identifies the top-ranked assets based on their momentum scores, the next step is to determine the percentage allocation to each of those assets. This is done using a covariance matrix. The covariance matrix is a statistical tool that shows how asset returns move in relation to one another over a given period. This is key for optimizing allocation between multiple assets like those in AAA. Using covariance helps create a portfolio allocation that minimizes the chance of all assets in the portfolio plunging at the same time.

Generally speaking, for assets X and Y, the covariance is:

Covariance of (X,Y) = (Correlation of X,Y) * (Volatility of X) * (Volatility of Y)

or mathematically, this is

Cov(X,Y) = ρ(X,Y) * σ(X) * σ(Y)

To help visualize this concept of covariance, in Exhibit A we look at the growth of $10,000 invested in five selected ETFs over the past 100 trading days.

Exhibit A. Growth of $10,000 for Five ETFs

The lines in Exhibit A are based on total return and a starting investment of $10,000 in each of five funds selected. We can see that when GLD (the red line) moved down in May 2025, QQQ (the purple line) moved in the other direction.

But to implement the Adaptive Asset Allocation model, we can't just eyeball the zigs and zags. We need a quantitative measure of how the assets vary with each other, so we can allocate intelligently among the assets, and not overload on highly correlated assets. To do this, The AAA algorithm creates a covariance matrix, shown below in Exhibit B.

Exhibit B. Covariance Matrix for Five ETFs

(For readability, we've multiplied each covariance value by 10,000)

This matrix shows how each asset varies in relation to every other asset in the portfolio. But these numbers are not the same as correlation between asset pairs. The correlation of SPY with itself is 1.00, but Exhibit B (in the upper left cell) shows the covariance of SPY with itself as 2.76. The SPY vs. SPY covariance is greater than 1.0 because the covariance is calculated by multiplying the correlation of two assets by the volatility of the two assets. So even though SPY has a perfect correlation of 1.0 with itself, to get the SPY vs. SPY covariance, we need to multiply the volatility of SPY (.0166) by itself. This results in a non-zero covariance of 0.000276, or 2.76 as shown in Exhibit B.

Or we can think of it this way:

Volatility, as measured by standard deviation, shows how much one asset varies.
Correlation measures how two assets vary together, using a standard range of -1 to +1.
Covariance measures the direction and strength of how two assets vary together, with no limit on the maximum or minimum.

And we can also describe what the possible covariance values mean. If the covariance is…

a big positive number, then the asset pair moves together with a strong, combined positive effect.
a smaller positive number, then the asset pair varies less together.
zero, then the two assets have no linear relationship. Their movements don't show a consistent pattern together.
negative, the asset pair moves in opposite directions. We can see this looking back at Exhibit A, where we can see GLD (the red line) moving in the opposite direction of QQQ (the purple line) in May 2025. And sure enough, the covariance for this pair is -0.15 ( x 10^-4) in Exhibit B.

Applying the covariance matrix to Adaptive Asset Allocation

But enough of the math! How does covariance help us? How does the Adaptive Asset Allocation algorithm use the covariance matrix to get better, risk-adjusted results?

The Adaptive Asset Allocation algorithm uses the covariance matrix to choose the percentage of the portfolio to allocate to each asset. This can be done using a variety of methods including Inverse-Risk Weighting, Mean-Variance Optimization [MVO], or Risk Parity. The result is an allocation that aims to minimize total portfolio volatility. Whatever method is used, some assets get bigger weights than others, and some assets may get an allocation at or near 0%.

This risk-adjusted portfolio features a higher allocation to assets that are less correlated with the others, and lower (or zero) allocation to highly volatile or highly correlated assets. This provides higher portfolio diversification, even when using a smaller number of assets.

Performance of the Adaptive Asset Allocation approach

How well has the Adaptive Asset Allocation performed? The AAA approach has shown strong historical performance over the past 20 years.

The total return of t.aaaf (one of six implementations of AAA that we track at RecipeInvesting) outperforms a traditional 60/40 balanced portfolio, a U.S. Bond portfolio, and even SPY over the last 20 years (which includes the annus horribilis of 2008).

As shown in Exhibit C, AAA balances compounded annual growth rates [CAGRs] and maximum drawdown effectively, when compared to other allocations.

Exhibit C. Growth of $10,000 over the past 20 years for AAA portfolio (t.aaaf), SPY, static balanced 60/40 portfolio (s.6040) and BND

Exhibit D shows that the AAA portfolio beaten SPY and the other benchmarks so far this year (and over the past 20 years), but AAA has lagged SPY over the past 15 years. However, just looking at the total return numbers does not account for risk.

Exhibit D. Total Return Comparison (periods longer than one year show annualized return)

AAA provides a respectable risk-adjusted return. Over the past 20 years, the maximum drawdown (i.e., the highest peak-to-trough fall in total return, measured at month-end) for SPY is 50.8%. But AAA has only 18.9% max drawdown. This is shown in Exhibit E.

Over the past five years, AAA's Maximum Drawdown is within 2% of BND, the broad bond ETF. This is despite t.aaaf having a much larger annual return than BND.

Exhibit E. Maximum Drawdown Comparison

Conclusion

Adaptive Asset Allocation provides consistent, long-term performance with a dynamic, risk-adjusted approach. Based on its breadth, agility, risk-awareness, and performance, the Adaptive Asset Allocation portfolio is worth considering. This approach can appeal to investors who are willing to rebalance their portfolio periodically. Portfolio updates can be eased by using a percentage-based portfolio rebalancing tool, like those from M1 Finance or Interactive Brokers.

as published on

Disclosure: I am/we are long EFA, IWM, GLD, TLT, QQQ, SPY, EEM. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it. I have no business relationship with any company whose stock is mentioned in this article.