CFA Level 3 Topic Breakdown

CFA Level 3 Asset Allocation: The Framework Behind Every Portfolio Decision

Asset allocation is the most important decision in portfolio management. At Level 3 it carries major exam weight and underpins everything in the individual and institutional portfolio management readings. This is the complete framework.

Asset allocation is, by most research estimates, the single most important determinant of long-run portfolio performance — more important than security selection, market timing, or fee management. At CFA Level 3, it occupies a corresponding place of prominence in the curriculum: asset allocation concepts appear in the morning essay section, the afternoon item sets, and the individual and institutional portfolio management readings that together dominate the exam.

This guide covers the full Level 3 asset allocation framework: the distinction between strategic and tactical allocation, the approaches to determining the strategic asset allocation, the Black-Litterman model, and the rebalancing and implementation considerations that the exam regularly tests.

35–40%
Portfolio management weight at L3
#1
Determinant of long-run returns
Both
AM and PM sections tested

The Asset Allocation Decision: A Framework

The asset allocation process starts with the investor's Investment Policy Statement (IPS). The IPS defines the investor's risk tolerance, return objectives, and constraints — and these parameters directly determine the appropriate strategic allocation. There is no optimal asset allocation independent of the investor's specific situation; what is optimal for a 30-year-old growth investor is inappropriate for a pension fund with near-term liabilities.

The key concepts:

Strategic Asset Allocation is the long-run target allocation aligned with the IPS. Tactical Asset Allocation is the short-term deviation from the SAA based on near-term views. Rebalancing returns the actual allocation to the SAA target when drift occurs.

Approaches to Determining the SAA

The CFA Level 3 curriculum covers four main approaches to determining the strategic asset allocation:

SAA Approaches
Mean-Variance Optimisation (MVO)

The classical Markowitz framework: maximise expected return for a given level of risk (or minimise risk for a given return). Produces the efficient frontier. Limitations: highly sensitive to input assumptions (small changes in expected returns produce large changes in the optimal portfolio), tends to produce concentrated allocations, and relies on historical data that may not be representative of the future.

Black-Litterman Model

Addresses MVO's sensitivity problem by starting from market equilibrium returns (implied by the market-cap-weighted portfolio) and adjusting these returns based on the investor's specific views, weighted by confidence. Produces more diversified, stable portfolios than unconstrained MVO.

Factor-Based Asset Allocation

Allocates across risk factors (equity risk, interest rate risk, credit risk, inflation, liquidity) rather than traditional asset classes. Asset classes that load on the same factors provide less true diversification than they appear to in traditional allocation frameworks. Particularly useful for identifying hidden concentrations.

Liability-Relative Asset Allocation

Used by entities with defined liabilities (pension funds, insurance companies). The asset allocation is evaluated relative to the liability profile — the goal is to maximise the probability of meeting liabilities, not to maximise absolute returns. The hedging portfolio (assets that offset liability risk) is separated from the return-seeking portfolio.

The Black-Litterman Model in Depth

The Black-Litterman model is the most technically demanding asset allocation concept at Level 3 and requires genuine understanding of both the mechanics and the logic. The model combines two sources of information:

  1. Equilibrium returns (the prior): The expected returns implied by assuming the current market-cap-weighted portfolio is optimal. These are derived by reverse-optimising the market portfolio — working backwards from the allocation to the implied expected returns. The equilibrium represents a neutral starting point.
  2. Investor views (the likelihood): The investor's specific, quantified views on expected returns for particular assets or asset classes, along with a confidence level for each view. Views can be absolute ("I expect US equities to return 8%") or relative ("I expect EM equities to outperform DM equities by 2%").

The model combines these using Bayesian updating: views with high confidence pull the combined estimate toward the view; views with low confidence leave the estimate close to the equilibrium. The result is a set of expected returns that are more stable and produce more diversified portfolios than pure MVO.

Exam Application
The Level 3 exam typically tests Black-Litterman by presenting a scenario where a manager has specific views and asking: (1) how would BL adjust equilibrium returns to incorporate the view, (2) what direction would the BL portfolio tilt relative to the equilibrium portfolio, and (3) why does BL produce a more stable portfolio than unconstrained MVO? Know the logic well enough to answer these questions without memorising the full matrix algebra.

Tactical Asset Allocation

Tactical Asset Allocation involves short-term deviations from the SAA to exploit perceived mispricings or to respond to near-term economic developments. There are two primary TAA approaches:

  • Discretionary TAA: Based on the portfolio manager's qualitative assessment of macroeconomic conditions, valuation levels, and market sentiment. Provides flexibility but introduces manager subjectivity and behavioural biases.
  • Systematic TAA: Rules-based signals (momentum, valuation metrics, macro indicators) that generate allocation shifts. More disciplined and back-testable but may be slow to adapt to structural changes.

The evidence on TAA is mixed. Research suggests that valuation-based signals (overweight cheap asset classes, underweight expensive ones) have modest but persistent predictive power over long horizons. Momentum-based signals have shorter-horizon predictive power but higher transaction costs. Most institutional investors use TAA conservatively — with limits on how far the actual allocation can deviate from the SAA — because the cost of large TAA bets that go wrong is significant.

Rebalancing Strategies

Rebalancing Approaches Compared
Calendar Rebalancing
Threshold (Percentage Band)
Tactical Rebalancing
Rebalance on a fixed schedule (monthly, quarterly, annually). Simple and disciplined. May trade when unnecessary or miss timely rebalancing opportunities.
Rebalance when any asset class drifts beyond a set band (e.g., ±5% of target). Trade-off: tighter bands = lower tracking error but higher turnover and costs.
Combine threshold rules with discretionary views — allow drift when TAA signal supports the drifted allocation; rebalance when signal reverses or band is breached.

The optimal rebalancing strategy depends on the investor's tax situation, transaction costs, and tolerance for tracking error. In tax-advantaged accounts with low transaction costs, tighter bands and more frequent rebalancing is justified. In taxable accounts with significant capital gains, less frequent rebalancing that avoids triggering realised gains may be preferable despite the higher tracking error.

For a deeper exploration of how asset allocation connects to individual investor IPS construction, see our guide on the Level 3 morning essay section, where IPS and asset allocation questions are the most frequently tested topics. For the fixed income side of the Level 3 allocation puzzle, our Level 2 Fixed Income guide covers the LDI and duration management concepts that underpin liability-relative approaches.

Related Reading

Found this useful?