Strategic and tactical asset allocation and the total portfolio approach

In depth12 min read · 16 sources

Strategic asset allocation (SAA) determines how wealth is divided over the long term between equities, bonds and other asset classes. Tactical asset allocation (TAA) deviates from it temporarily and within limits, in order to exploit opportunities or avoid risks. Large sovereign wealth and pension funds now often take a third route, the total portfolio approach: a single overall portfolio, measured against a simple reference portfolio, instead of fixed buckets for each asset class. Here you will learn all three approaches, work out the contribution of a tactical deviation and see why its added value is so hard to prove.

Three levels of an investment decision

Every portfolio results from three decisions. The first is the long-term split between asset classes, for example 60% equities and 40% bonds. That is the strategic asset allocation. The second consists of temporary deviations from it, for example a few percentage points more equities because valuations appear favorable. They form the tactical asset allocation. The third is selection within the asset classes: which fund, which stock, which bond.

In a pension fund, the board sets the SAA for several years, the investment team manages the TAA within predefined ranges, and fund managers select the securities. Private investors hold all three roles at once. That is precisely why a written SAA helps: later on, it shows whether a deviation was deliberate tactics or a reaction to headlines.

Strategic asset allocation: the policy portfolio

The SAA is laid down as a policy portfolio: target weights for each asset class, plus an index for each class as a yardstick. Together they form the benchmark against which everything else has to be measured.

It is based on the goals, the risk capacity and assumptions about the return, volatility and correlation of the asset classes. Optimizers such as the efficient frontier help with the translation, but do not replace judgment. The SAA also includes fixed rules:

• Risk budget: how much may the portfolio fluctuate overall, and how much deviation from the benchmark is allowed for TAA and selection? The measure for this is the tracking error.
• Ranges: a lower and upper limit for each asset class, for example 60% ± 5 percentage points. Within the band, the weight may drift or be managed tactically.
• Rebalancing: if a weight leaves its band, it is reset (threshold rebalancing); alternatively, on fixed dates. The trade-off between costs and deviation is described by Masters (2003) and Donohue and Yip (2003); more in the article rebalancing.

Tactical asset allocation: limited deviations

Tactical asset allocation makes deliberate use of the ranges. It shifts weights for months up to a few years when a signal makes an asset class look more attractive or riskier. Typical signals are:

• Valuation: historically, a high Shiller P/E went hand in hand with lower equity returns in the following years. It is a slow signal.
• Trend and momentum: if the price is above its moving average or the class has recently done better than others, it is overweighted, otherwise underweighted (trend following, momentum strategies).
• Macro and interest rates: economic indicators, the steepness of the yield curve or credit spreads.
• Carry: higher running yields of a class or currency, such as interest rate differentials in the carry trade.

The key is to measure deviations in terms of risk, not just weight. Overweighting equities by Δw at the expense of bonds creates a tracking error relative to the SAA that depends on the volatility of the return difference between the two classes:

This lets you translate a risk budget for the TAA into a permitted deviation. If an investment committee grants the TAA 1% tracking error and the return difference between equities and bonds has a volatility of 16.28%, the equity weight may deviate from the target by at most 6.1 percentage points. Such an active risk budget prevents tactics from quietly turning into a new strategy (see also market timing).

Worked example: what a deviation earns and costs

The SAA is 60% equities and 40% bonds. For one year, the TAA overweights equities by 5 percentage points to 65%. For the risk, we assume 16% volatility for equities, 5% for bonds and a correlation of 0.1. These are the same round values as in the article on the efficient frontier.

Measuring the added value of TAA

It is measured against the SAA benchmark. The Brinson-Fachler allocation effect shows what the weight deviations would have earned if each class had only achieved its index return; how to link effects across periods is explained in the article performance attribution, and for many asset classes, currencies and overlays see multi-asset attribution.

A good measure of quality is the information ratio: active return divided by tracking error. The fundamental law of active management by Grinold (1989) explains why it usually stays small for TAA: IR ≈ IC × √N. The information coefficient IC measures skill, N the number of independent decisions per year. An equity-bond bet has only a few of them.

The t-statistic of the active return grows roughly with IR × √years. Sharpe (1975) showed early on that market timing has to be right very often to pay off compared with buy and hold. The reason is simple: if you are out of equities at the wrong time, you forgo their risk premium.

Total portfolio approach: one portfolio instead of silos

The total portfolio approach (TPA) turns the logic of the SAA on its head. Instead of target weights for each asset class, the board sets a goal, a risk appetite and a simple reference portfolio, for example 80% equities and 20% bonds in low-cost index investments. The investment team may deviate within a risk framework if an investment improves the overall portfolio; every idea competes with every other for capital. CAIA Association (2024) describes the approach as a common measure of risk and return for the whole portfolio, implemented differently depending on the organization.

Well-known examples are large sovereign wealth and pension funds:

• New Zealand: the New Zealand Superannuation Fund measures itself against a notional, passive reference portfolio of 80% growth assets (75% global and 5% New Zealand equities) and 20% global bonds, with currencies fully hedged. It serves as the yardstick for added value after all costs and as a hurdle for active investments (NZ Super Fund 2023).
• Canada: CPP Investments itself calls its decision-making process a total portfolio approach; the reference portfolio of the base plan consists of 85% global equities and 15% Canadian government bonds (CPP Investments 2023).
• Singapore: on April 1, 2013, GIC introduced a reference portfolio of 65% global equities and 35% global bonds. It reflects the risk the government is willing to bear; below it sit a policy portfolio and an active portfolio (GIC 2013).
• Australia: the Future Fund aims to achieve at least inflation plus 4 to 5% a year over the long term, with an acceptable but not excessive level of risk (Future Fund Investment Mandate 2024); according to Urwin (2025), it has no classic SAA.
• US: in November 2025, CalPERS decided to switch from the SAA model to TPA from July 1, 2026, with a single reference portfolio of 75% equities and 25% bonds instead of eleven benchmarks for individual asset classes (CalPERS 2025).

The table summarizes the differences described by Urwin (2025) and Filbeck and Bok (2025). The boundary is fluid; many funds combine both approaches. Roger Urwin of the Thinking Ahead Institute puts it roughly like this in the report by Filbeck and Bok (2025): SAA is convenient for governance but cumbersome for the portfolio; TPA is efficient for the portfolio but makes greater demands on governance.

What the research shows

The importance of the SAA is documented above all by Brinson et al. (1986). In 91 large US pension funds from 1974 to 1983, investment policy explained on average 93.6% of the variation in total return over time. The funds earned an average of 9.01% a year, their policy benchmark 10.11%: market timing and security selection therefore cost 1.10% a year, even though individual funds gained up to 3.69% a year.

Ibbotson and Kaplan (2000) show that the often-quoted figure differs depending on the question: investment policy explains around 90% of a fund's variation over time, but only around 40% of the differences between funds, and on average about 100% of the level of return. Xiong et al. (2010) go further: once the movement of the overall market is stripped out, allocation and active management are equally important for the differences in return within a peer group. The SAA therefore determines the level, but not on its own the gap to the competition.

So far, the evidence on TPA mainly consists of reports from practice. Filbeck and Bok (2025) analyze conversations with more than a dozen investors and find that success depends more on culture and governance than on organizational structure. A comparative study by the Thinking Ahead Institute among 26 very large investors reports a return advantage of 1.3% a year over ten years for TPA funds (Urwin 2025). That is an indication, not proof: the sample is small, and differences in size or in the share of private markets could equally explain the advantage.

For advanced readers: risk budgets and competition for capital

Risk budgets do not add up linearly. If fund selection with 1.5% is added to the TAA with 0.81% tracking error and the two are uncorrelated, the total tracking error is √(0.81² + 1.5²) = 1.70%. Uncorrelated bets are therefore cheap, correlated ones expensive; in the framework of Grinold and Kahn (2000), you allocate the budget to sources with a high information ratio.

Competition for capital can be expressed formally. A new investment improves a portfolio's Sharpe ratio at the margin if and only if its own Sharpe ratio is greater than its correlation with the portfolio times the portfolio's Sharpe ratio:

With a portfolio Sharpe ratio of 0.40, an investment with a correlation of 0.5 only has to exceed a Sharpe ratio of 0.20, whereas one with a correlation of 0.95 has to exceed 0.38. This hurdle is the core of TPA: an investment counts not by its return alone, but by its risk contribution to the whole. In the SAA approach, by contrast, each class receives its budget even if its marginal benefit for the overall portfolio is small.

The reference portfolio is also a statement about risk. It should have as much risk as the actual portfolio; otherwise the difference merely measures additional market risk. With illiquid investments this is difficult: their valuations are smoothed, their measured risk is too low, and hidden leverage can easily look like added value.

Limits and common mistakes

• Assumptions instead of knowledge: the SAA is only as good as the capital market assumptions. An optimization using historical returns often produces extreme weights (see resampling optimization).
• Tactics driven by fear: deviations after losses are often procyclical, not tactics. A fixed signal and a risk budget help you see the difference.
• Costs and taxes: every tactical reallocation costs spread and fees, and for private investors in taxable accounts often also tax on realized gains. Added value of 0.16% a year is quickly eaten up.
• TPA without governance: without a clear goal, risk limits and accountability, competition for capital becomes arbitrary.
• Transferability: the TPA of large funds relies on teams, data and private markets. What remains for private investors is the mindset: look at your entire wealth (see total wealth).

How Portfolio Lab calculates it

In the Backtest, a portfolio's target weights are its SAA. Rebalancing takes place monthly, quarterly (default), semi-annually, annually, never or with a tolerance band: as soon as a weight deviates from its target by more than the band (default ± 5 percentage points, selectable from 1 to 50), the app resets all weights to target. With monthly data, it checks the band at each month-end; with daily prices, every day. You set up a simple reference portfolio, such as 80% global equities and 20% bonds, as a second portfolio and compare it over the same period.

You model tactical deviations in the Backtest with time-dependent weights: a table from a given date, a glide path, a value threshold or a file with your own signals. Every change triggers a rebalancing. The Tactical page calculates rule-based models on a monthly basis without look-ahead: moving average, dual momentum, relative strength, target volatility, adaptive allocation, valuation based on the Shiller CAPE, the rotation models PAA, VAA, DAA and accelerating dual momentum, and a regime model (hidden Markov model). The comparison is always with buy and hold of the base weights, rebalanced monthly, i.e. with the SAA.

The Risk page decomposes the added value relative to the SAA using Brinson-Fachler: another portfolio from your Grid serves as the benchmark, the segments are the asset classes by default, and the allocation effect corresponds to the contribution of the tactical weights. The app does not model a total portfolio approach with a board and competition for capital, but a reference portfolio as a yardstick can be recreated this way. Details are in the tactical allocation methodology and the backtest methodology.

This article explains a method; it is not investment or tax advice. Worked examples are simplified and rounded. All terms in the glossary.