Estimated study time: 45 minutes
Content:
Factor models decompose asset returns into systematic components (factors) and idiosyncratic residuals. In asset allocation, factor-based frameworks offer three advantages over traditional asset class approaches: better risk decomposition (two "asset classes" may share the same underlying risk factors), improved diversification (true factor diversification vs. correlation-based diversification), and more stable inputs for optimization (factor loadings are more stable than direct asset class correlations).
The foundational factor model is the Capital Asset Pricing Model (CAPM), which attributes all systematic risk to a single market factor (beta). Multifactor models extend this. Macroeconomic factor models use observable economic variables as factors — economic growth, inflation, interest rates, credit spreads, and liquidity are common choices. Asset returns are regressed on these macro variables to estimate factor sensitivities (betas). Statistical factor models use principal component analysis (PCA) or factor analysis to extract latent factors from historical return data — the factors have no predetermined economic interpretation but explain the covariance structure.
Fundamental factor models use firm characteristics — such as size (market cap), value (book-to-market), momentum, profitability, and investment — as proxies for systematic risk. The Fama-French three-factor model adds size (SMB: small minus big) and value (HML: high minus low book-to-market) to the market factor. Carhart extended this to four factors by adding momentum (UMD: up minus down). The Fama-French five-factor model further adds profitability (RMW) and investment (CMA).
At the portfolio level, factor models are used to construct risk-factor-based portfolios (sometimes called "smart beta" or "factor-tilted" portfolios). A long-only equity manager can tilt a portfolio toward value, quality, or momentum factors while controlling overall factor exposures. Factor-based asset allocation seeks to replace traditional asset class allocations with explicit allocations to risk factors — for example, allocating to the "equity risk" factor, "credit risk" factor, "duration" factor, "real assets" factor, and "liquidity" factor rather than to equities, bonds, real estate, and commodities separately.
Risk factor budgeting assigns portions of total portfolio risk (often measured as risk contribution, not weight) to each factor. Equal risk contribution (ERC) portfolios weight assets so each contributes equally to total portfolio variance — this tends to produce more balanced portfolios than equal-weight or market-cap-weight approaches, especially across asset classes with different volatilities.
Key Terms:
Quiz Questions:
Q1. An asset allocator notices that her equity portfolio and her corporate bond portfolio both have high loading on the "credit risk" factor. From a factor-based diversification perspective, she should:
A) Maintain both positions because they are in different asset classes B) Reduce total exposure to the credit risk factor since she is taking on concentrated factor risk C) Hedge the equity portion to reduce interest rate exposure D) Add more equity to offset the negative carry in the corporate bond portfolio
Answer: B — Factor-based thinking reveals that owning both equities and credit bonds can result in concentrated exposure to the credit/economic risk factor, even though they are categorized as different asset classes. The relevant risk is factor exposure, not asset class label.
---
Q2. A portfolio manager is using a Fama-French five-factor model to explain returns. Her portfolio shows positive alpha after controlling for market, size, value, profitability, and investment factors. This alpha:
A) Must reflect genuine skill since all major systematic factors are controlled B) May still reflect exposure to omitted factors (e.g., momentum, quality, low volatility) rather than pure skill C) Cannot exist in an efficient market after controlling for five factors D) Represents compensation for illiquidity, which five-factor models capture well
Answer: B — Five-factor models control for five known systematic sources of return, but many other factors (momentum, low volatility, quality variants) are not included. Apparent alpha may reflect loading on these omitted factors. A robust attribution requires including all relevant factors.
---
Q3. In a macroeconomic factor model, which factor would most likely have a positive loading for a long-duration government bond portfolio?
A) Economic growth surprise (positive) B) Inflation surprise (positive) C) Interest rate decline (positive — i.e., negative loading on rate increases) D) Credit spread widening (positive)
Answer: C — Long-duration government bonds benefit from falling interest rates (prices rise when rates fall) and suffer from rising rates. In a macro factor model, the portfolio would show a large negative loading on interest rate levels (or equivalently, a large positive loading on rate declines).
---
Q4. An equal risk contribution (ERC) portfolio is constructed across four asset classes with very different volatilities. Compared to an equal-weight portfolio, the ERC portfolio will:
A) Overweight the highest-volatility asset classes to equalize dollar risk B) Underweight the highest-volatility asset classes so each contributes equally to total portfolio variance C) Produce the same allocation as market-cap weighting D) Require leverage to achieve the same return as an equal-weight portfolio
Answer: B — ERC targets equal marginal risk contribution. High-volatility assets naturally contribute more risk per dollar invested, so they receive lower weights to equalize contributions. This typically leads to overweighting of lower-volatility assets (e.g., bonds) relative to equal-weight.
---
Q5. A factor-based asset allocation framework replaces "equity" and "corporate bond" allocations with allocations to "equity risk premium" and "credit risk premium" factors. The primary advantage of this approach is:
A) It reduces the number of instruments the portfolio needs to hold B) It reveals true risk diversification — assets from different classes may share the same factor exposures, and factor diversification is more meaningful than asset class diversification C) Factor-based allocations always produce higher Sharpe ratios than asset class allocations D) It eliminates the need for rebalancing since factor exposures are stable over time
Answer: B — The core insight of factor-based asset allocation is that the true unit of diversification is the risk factor, not the asset class. Equities and corporate bonds both have significant exposure to economic/credit factors, meaning a traditional 60/40 portfolio may be much less diversified in factor space than it appears.
---