Asset Pricing, Factor Models and Market Efficiency

Introduction and Context

Investment analysis requires more than estimating whether a security is attractive. It requires a disciplined explanation of the return that investors should demand for bearing risk, the sources of that risk, and the extent to which market prices already reflect available information. Asset-pricing models provide this framework. They connect expected returns with systematic risk, distinguish diversifiable from non-diversifiable risk, and support decisions concerning valuation, portfolio construction, performance evaluation and investment policy.

The Capital Asset Pricing Model (CAPM) is the central single-factor model. It proposes that the expected return on an asset is determined by its exposure to the market portfolio. Multifactor models extend this logic by allowing several systematic sources of risk, such as size, value, momentum, profitability, investment, interest rates or inflation. Arbitrage Pricing Theory (APT) provides a more general no-arbitrage foundation for factor pricing, although it does not identify the factors in advance.

These theories must be interpreted critically. Asset-pricing models are useful approximations rather than immutable laws. Their predictions depend on assumptions about investor behaviour, expectations, market integration, portfolio measurement and data quality. Market anomalies—including momentum, value, size and post-earnings-announcement drift—challenge simple versions of market efficiency, although some anomalies may reflect risk, transaction costs, data mining or behavioural biases.

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Explain the logic, assumptions and implications of CAPM.
2. Calculate and interpret beta, expected return and the Security Market Line.
3. Distinguish the Security Market Line from the Capital Market Line.
4. Explain multifactor models and the arbitrage pricing framework.
5. Evaluate major empirical anomalies and competing explanations.
6. Compare active and passive investment management.
7. Assess weak, semi-strong and strong forms of market efficiency.
8. Identify empirical, methodological and practical limitations of asset-pricing theories.
9. Apply asset-pricing concepts to portfolio and investment-policy decisions.

1. Risk, Return and the Pricing of Securities

An investor’s total return over a holding period commonly consists of income and price appreciation:

where R_i is the holding-period return on security i, P₀ is the initial price, P₁ is the ending price, and D₁ is cash income during the period. The quantities must be measured in the same currency and for the same investment horizon.

Expected return is a probability-weighted average of possible returns:

Here, p_s is the probability of state s, R_is is the return in that state, and the probabilities sum to one. Risk may be represented by variance or standard deviation:

Variance and standard deviation measure dispersion, not necessarily economically undesirable risk. Investors may care particularly about permanent loss, downside risk, liquidity risk, inflation risk or shortfall relative to a liability.

For a portfolio, expected return is the weighted average of asset expected returns:

where w_i is the portfolio weight and the weights sum to one for a fully invested portfolio. Portfolio variance depends on covariances as well as individual variances:

Diversification reduces risk when asset returns are not perfectly positively correlated. The crucial asset-pricing question is which part of risk receives compensation in equilibrium.

2. Systematic and Unsystematic Risk

Unsystematic risk is security-specific risk arising from events such as a product failure, litigation or an unexpected management change. It can be reduced through diversification. Systematic risk affects many securities and may arise from economic growth, monetary policy, market-wide liquidity, geopolitical events or changes in risk appetite. It cannot be eliminated by holding more securities within the same market.

CAPM argues that only systematic risk is priced. If investors can diversify security-specific risk at low cost, they should not receive an expected-return premium for bearing it. This does not mean unsystematic risk is irrelevant to every investor; concentrated owners, employees and poorly diversified funds may still be exposed to it.

Covariance measures how two returns move together. Correlation standardizes covariance:

Correlation lies between −1 and +1. Beta measures the sensitivity of a security to market movements:

where R_M is the market return. Equivalently, beta is the slope coefficient from a regression of security excess returns on market excess returns, subject to the chosen frequency, sample period and market proxy.

A beta of 1 indicates market-like systematic sensitivity; a beta above 1 indicates greater sensitivity, while a beta below 1 indicates lower sensitivity. A negative beta implies that the security tends to perform well when the market performs poorly, although estimates may be unstable.

3. The Capital Asset Pricing Model

The CAPM expected-return equation is:

where R_f is the risk-free return and E(R_M) − R_f is the expected market risk premium. The expression is applicable as an equilibrium expected-return relationship under CAPM assumptions; it is not a guaranteed realized-return formula.

Assumptions

The standard CAPM assumes, in simplified form:

• Investors are risk-averse and choose portfolios using expected return and variance.
• Investors have homogeneous expectations about expected returns, variances and covariances.
• Investors can borrow and lend at a common risk-free rate.
• Markets are frictionless: there are no taxes, transaction costs, restrictions or information costs.
• Assets are infinitely divisible and investors can trade freely.
• Investors have a common single-period horizon.
• All relevant assets are available and the market portfolio includes all investable assets.

These assumptions are deliberately strong. Their value is analytical: they clarify the mechanism through which equilibrium risk premiums arise. Their weakness is that real markets contain taxes, short-sale constraints, heterogeneous beliefs, funding constraints and multiple horizons.

Security Market Line

The Security Market Line (SML) plots expected return against beta. Its intercept is the risk-free rate and its slope is the market risk premium. A security above the SML offers a positive alpha relative to CAPM; a security below it offers negative alpha.

In empirical work, alpha is estimated rather than observed directly. A positive historical alpha may result from omitted risk factors, estimation error, luck or genuine abnormal performance.

The SML applies to individual securities and portfolios and uses beta on the horizontal axis. It differs from the Capital Market Line (CML), which plots expected return against total standard deviation for efficient portfolios that combine the risk-free asset with the tangency portfolio:

The CML is relevant only to efficient portfolios under the CAPM setting, whereas the SML can be used to assess any asset or portfolio through systematic risk.

Worked Example: CAPM and Alpha

Suppose the annual risk-free rate is 4%, the expected market return is 10%, and a stock has beta 1.25. CAPM implies:

If an analyst forecasts a return of 14%, the forecast alpha relative to CAPM is 2.5 percentage points. This conclusion is conditional on the beta estimate, the market-risk-premium estimate and the appropriateness of CAPM. It is not proof that the stock is mispriced.

4. Multifactor Models

A multifactor model explains returns using exposure to several systematic factors. A general factor specification is:

Here, F_k is the excess return or factor realization associated with factor k, β_ik is security i’s sensitivity to that factor, α_i is abnormal return after controlling for the factors, and ε_i is an idiosyncratic residual. A common assumption is that the residual has zero conditional mean and is not systematically related to the included factors.

A factor may be a traded portfolio return, a macroeconomic variable or a characteristic-based return spread. The distinction matters. A traded factor portfolio can often be interpreted as an investable risk premium; a macroeconomic factor may require a separate mapping between economic shocks and returns.

Fama–French and Related Models

The three-factor model adds size and value factors to the market factor:

SMB (“small minus big”) represents the return difference between portfolios of smaller and larger firms. HML (“high minus low”) traditionally represents the return difference between high and low book-to-market portfolios. The model was developed to explain cross-sectional return patterns that CAPM beta alone did not adequately explain.

The five-factor extension adds profitability and investment factors:

Here, MKT is the market excess return, RMW is a robust-minus-weak profitability factor and CMA is a conservative-minus-aggressive investment factor. Exact construction varies by dataset and implementation.

Momentum is frequently represented by a factor based on recent winners minus recent losers. Momentum does not fit neatly into the original CAPM and may create substantial turnover and crash risk. Factor definitions, portfolio breakpoints, rebalancing frequency, liquidity screens and geographic coverage materially affect results.

Interpreting Factor Exposures

A factor loading is not automatically a causal explanation. A high value loading may proxy for distress, leverage, earnings risk or limits to arbitrage. A momentum loading may reflect behavioural underreaction, compensation for crash risk or trading-related effects. Factor models are therefore both valuation tools and empirical research instruments.

A portfolio manager can estimate factor exposures using time-series regression, monitor unintended bets, and decompose performance into factor returns and residual alpha. However, multicollinearity, unstable exposures and factor crowding can make interpretation difficult.

5. Arbitrage Pricing Theory

APT begins with a factor-generating process rather than the CAPM’s market-portfolio equilibrium. The return-generating model is commonly written as:

Under APT, sufficiently diversified portfolios with the same factor exposures should have approximately the same expected return. Otherwise, investors could construct a zero-net-investment portfolio with positive expected profit and little or no factor risk. The APT pricing relation is often expressed as:

where λ_k is the risk premium associated with factor k. A factor’s premium is not necessarily positive; it depends on whether investors require compensation for exposure to that factor.

APT is less restrictive than CAPM because it does not require the market portfolio to be mean-variance efficient and does not specify the factors. Nevertheless, it relies on diversified portfolios, stable factor relationships, sufficiently competitive markets and the absence of persistent arbitrage opportunities. In practice, identifying the economically meaningful factors remains the central challenge.

6. Market Efficiency

The efficient market hypothesis (EMH) states that security prices reflect available information sufficiently rapidly that risk-adjusted excess returns are difficult to earn consistently after costs. Efficiency is a statement about the joint behaviour of prices, information and competition—not the claim that prices are always correct or that prices never change.

Three forms are commonly distinguished:

The joint-hypothesis problem is fundamental. Testing market efficiency requires a model of expected returns. If a trading strategy earns an abnormal return, the result may indicate market inefficiency or an inadequate asset-pricing model. Therefore, rejecting CAPM does not by itself reject EMH.

Evidence and Anomalies

Important reported patterns include:

• Size effect: smaller firms historically earned higher average returns than larger firms in some samples.
• Value effect: firms with high book-to-market or other value characteristics sometimes outperformed growth firms.
• Momentum: recent winners tended to outperform recent losers over intermediate horizons, followed by possible long-horizon reversals.
• Post-earnings-announcement drift: prices appeared to adjust gradually after earnings surprises.
• Short-term reversal: very recent losers sometimes outperformed very recent winners.
• Low-volatility effect: low-beta or low-volatility portfolios sometimes earned returns higher than CAPM would predict.

Possible explanations include omitted risk, behavioural biases, transaction costs, illiquidity, short-sale constraints, publication bias and multiple testing. A pattern that was profitable before publication may become less profitable after investors trade on it. A professional evaluation must use out-of-sample tests, realistic costs, appropriate risk adjustment and economic rationale.

7. Active and Passive Management

Passive management seeks to replicate a benchmark at low cost. Its advantages include diversification, transparency, low turnover and predictable exposure. Exchange-traded funds and index funds have made broad passive exposure accessible to many investors. Passive does not mean risk-free: the investor remains exposed to market, factor, liquidity, tracking and benchmark-construction risks.

Active management seeks returns or risk characteristics different from a benchmark through security selection, tactical allocation, factor timing or portfolio construction. Active management can add value when information is scarce, mandates are flexible, markets are segmented or the manager has a genuine process advantage. It also incurs research costs, trading costs, market impact, management fees and potential behavioural errors.

The arithmetic of active management implies that, before costs, the aggregate holdings of active investors equal the market portfolio. Relative to the market, active investors must collectively earn zero before costs; after costs, they earn less than the market in aggregate. This does not imply that no active manager can outperform. It implies that identifying persistent skill is difficult and that the average investor must consider fees and implementation.

Performance evaluation should use a suitable benchmark and risk-adjusted measures. Jensen’s alpha under CAPM is:

For realized returns, alpha is normally estimated over a sample rather than calculated from one observation. A factor benchmark may be more appropriate than a market-only benchmark when a portfolio has systematic size, value, momentum or other exposures.

8. Empirical Methods and Limitations

Estimation Issues

Beta and factor loadings depend on the measurement window, return frequency, benchmark, currency, corporate actions and treatment of missing observations. Daily data may contain nonsynchronous trading and microstructure noise; monthly data provide fewer observations. A beta estimated from historical data may not equal the future beta.

Expected returns and risk premiums are unobservable and must be estimated. Historical averages are noisy, especially for long-horizon premiums. Survivorship bias excludes failed funds or firms; look-ahead bias uses information unavailable at the decision date; data-snooping bias makes some apparently successful strategies products of extensive search.

Model Misspecification

CAPM’s market portfolio is not directly observable. Common equity indices omit private businesses, human capital, housing and other wealth. A poor proxy can make the CAPM appear to fail even if the theoretical market portfolio is relevant. Multifactor models improve explanatory power but may proliferate factors without improving economic understanding.

High explanatory power is not proof of a valid pricing model. A regression can fit historical returns while failing out of sample. Factor premiums may be nonstationary, crowded or conditional on economic regimes. Structural breaks can arise from regulation, technology, monetary policy or changes in investor participation.

Practical and Ethical Limitations

A model should not be used to disguise uncertainty. Forecasts should include sensitivity analysis, confidence intervals where appropriate and explicit assumptions. Investment professionals must disclose conflicts of interest, distinguish back-tested results from live performance, avoid misleading claims about alpha, and apply suitable suitability and fiduciary standards. Model outputs should support—not replace—judgment concerning liquidity, concentration, client objectives and liabilities.

9. Professional Application: A Factor-Aware Portfolio Decision

Background

The following is an illustrative case. Horizon Retirement Fund is considering whether to replace part of an actively managed domestic-equity mandate with a low-cost multifactor index. The fund has a long horizon but must meet regular pension payments. Its current manager has outperformed the market by 1.2% annually over five years before fees, but the portfolio has substantial value and small-cap exposures.

Problem

The investment committee must determine whether the apparent alpha represents skill or compensation for factor exposures. It must also assess whether a passive or factor-based solution improves expected net outcomes without creating unacceptable liquidity or tracking risks.

Analysis

The consultant estimates a market-only regression and then a multifactor regression. Under CAPM, the manager’s alpha is 1.2%. After adding size, value and momentum factors, alpha falls to 0.3% and is statistically imprecise. The manager’s annual fee is 0.9%, while the proposed index costs 0.2%. The active portfolio also has higher turnover and greater exposure to less liquid securities.

The analysis does not prove that the manager lacks skill. It indicates that much of the historical outperformance can be explained by systematic factor exposures. The committee tests the result using different periods, regional benchmarks, transaction-cost assumptions and crisis periods. It also examines whether the manager’s factor exposures are stable and whether the factor index can be traded at its stated costs.

Approach and Decision

The committee retains a smaller active allocation where the manager has a differentiated research capability, replaces the broad core mandate with a low-cost diversified index, and sets explicit limits on unintended factor, liquidity and concentration exposures. It evaluates performance against a factor-aware benchmark rather than the market index alone.

Outcome and Learning Points

The decision recognizes that passive exposure is not automatically superior and active management is not automatically value-adding. The relevant comparison is expected net return, risk, liquidity, governance burden and reliability of the investment process. Attribution to systematic factors makes the decision more transparent and reduces the risk of paying active fees for readily available exposures.

10. Contemporary Developments

Technology has increased the speed and scale of factor research. Machine learning can detect nonlinear relationships, interactions and regime patterns, but flexible models increase overfitting risk and reduce interpretability. A valid research process requires economic motivation, strict separation of training and testing data, economically realistic costs, robustness checks and documented model governance.

Environmental, social and governance information may enter asset pricing through preferences, regulation, cash-flow risk, transition risk or reputational effects. Whether ESG characteristics represent a priced risk factor, an investor preference, or a source of measurement noise remains an empirical question. Managers should define the objective clearly: risk reduction, values alignment, impact, or return enhancement are not interchangeable goals.

Factor investing has also become more accessible, increasing the possibility of crowding. When many investors hold similar exposures, expected premiums may decline and liquidity losses may become severe during stress. Portfolio construction should therefore consider factor diversification, turnover, capacity, drawdown behaviour and correlations that may rise in crises.

Best-Practice Framework

A professional asset-pricing analysis should:

1. Define the investment objective, horizon, constraints and liability context.
2. Select a theoretically and economically defensible benchmark.
3. State the model, factor definitions, data sources and estimation window.
4. Separate in-sample discovery from out-of-sample evaluation.
5. Include fees, spreads, market impact, taxes and rebalancing costs.
6. Test robustness across periods, markets and plausible specifications.
7. Distinguish factor exposure from manager skill.
8. Monitor changes in beta, factor loadings, liquidity and concentration.
9. Communicate uncertainty and avoid presenting historical alpha as a promise.
10. Review governance, conflicts, model risk and suitability continuously.

Chapter Summary

Asset-pricing theory explains why investors demand compensation for risk and provides tools for valuation and performance evaluation. CAPM prices systematic risk through beta and represents the expected-return relationship with the Security Market Line. The Capital Market Line is a different relationship involving efficient portfolios and total risk. Multifactor models extend CAPM by incorporating systematic exposures such as size, value, momentum, profitability and investment. APT derives a factor-pricing relationship from the absence of arbitrage but does not specify the factors.

Market efficiency has weak, semi-strong and strong forms, each making different claims about information. Reported anomalies challenge simple efficiency and CAPM, but their interpretation is complicated by omitted risk, behavioural effects, transaction costs, data mining and the joint-hypothesis problem. Active management may add value, but aggregate active performance is reduced by costs; passive and factor-based approaches offer transparent, diversified exposures but retain market and implementation risks. Sound professional practice combines theory with robust empirical testing, realistic costs, governance and explicit recognition of uncertainty.

Key Terms

Systematic risk: Risk affecting many securities that cannot be eliminated through ordinary diversification.

Unsystematic risk: Security-specific risk that can generally be reduced through diversification.

Beta: Sensitivity of a security’s return to the market return.

Market risk premium: Expected market return minus the risk-free return.

CAPM: A model linking expected return to beta and the market risk premium.

Security Market Line: The CAPM relationship between expected return and beta.

Capital Market Line: The relationship between expected return and total risk for efficient portfolios.

Alpha: Return unexplained by a specified asset-pricing model.

Covariance: A measure of the joint movement of two variables.

Multifactor model: A return model using several systematic explanatory factors.

Factor loading: Sensitivity of an asset or portfolio to a factor.

Arbitrage Pricing Theory: A no-arbitrage factor-pricing framework.

Factor premium: Compensation associated with exposure to a systematic factor.

Market efficiency: The degree to which prices incorporate relevant information.

Momentum: The tendency for recent relative winners to continue outperforming over a specified horizon.

Value effect: The historical tendency of certain value-oriented stocks to earn higher returns than growth-oriented stocks.

Tracking error: Volatility of the difference between a portfolio return and its benchmark return.

Survivorship bias: Bias caused by analysing only securities or funds that remain in the sample.

Joint-hypothesis problem: The difficulty of testing market efficiency without simultaneously specifying an expected-return model.

Factor crowding: Concentration of investors in similar factor exposures, potentially increasing stress losses.

Review Questions

Short Answer Questions

1. Distinguish systematic risk from unsystematic risk.
2. What does a beta of 1.4 indicate?
3. Why does the CAPM use beta rather than total standard deviation for individual securities?
4. State the difference between the Security Market Line and the Capital Market Line.
5. What is meant by semi-strong market efficiency?

Long Answer Questions

1. Explain the assumptions of CAPM and critically evaluate their realism.
2. Compare CAPM, multifactor models and APT in terms of assumptions, factors and practical use.
3. Explain major market anomalies and discuss alternative risk-based and behavioural interpretations.
4. Discuss the advantages and limitations of active and passive management.
5. Explain why the joint-hypothesis problem complicates empirical tests of market efficiency.

Analytical / Application Questions

1. The risk-free rate is 3%, the expected market return is 9%, and a portfolio beta is 0.8. Calculate the CAPM expected return. If the analyst forecasts 10%, calculate the forecast alpha and explain two reasons why it may not represent skill.
2. A fund has a market beta of 1.1, a size loading of 0.4 and a value loading of 0.7. Explain how a factor-aware benchmark could change the interpretation of its apparent market-index outperformance.
3. Design an empirical test of a momentum strategy that addresses transaction costs, look-ahead bias, survivorship bias, data snooping and out-of-sample performance.

Multiple Choice Questions with Answers

Question 1

Under CAPM, which risk is expected to receive a risk premium?

A. Firm-specific risk

B. Systematic market risk

C. Accounting risk only

D. Diversifiable risk

Correct Answer: B. Systematic market risk

Question 2

A security has a beta of 1.5, a risk-free rate of 4%, and an expected market return of 10%. Its CAPM expected return is:

A. 7.0%

B. 10.0%

C. 13.0%

D. 15.0%

Correct Answer: C. 13.0%

Question 3

The Security Market Line plots expected return against:

A. Total standard deviation

B. Portfolio size

C. Beta

D. Trading volume

Correct Answer: C. Beta

Question 4

Which statement best distinguishes the CML from the SML?

A. The CML uses beta for all securities

B. The SML applies only to risk-free assets

C. The CML relates return to total risk for efficient portfolios

D. The SML excludes portfolios

Correct Answer: C. The CML relates return to total risk for efficient portfolios

Question 5

In a multifactor model, a factor loading measures:

A. A security’s sensitivity to a factor

B. The security’s accounting profit

C. The number of shareholders

D. The risk-free interest rate

Correct Answer: A. A security’s sensitivity to a factor

Question 6

APT is primarily based on the principle that:

A. All investors have identical preferences

B. Persistent arbitrage opportunities should not exist

C. Only market beta matters

D. Prices never change

Correct Answer: B. Persistent arbitrage opportunities should not exist

Question 7

Which anomaly refers to continued relative performance of recent winners over an intermediate horizon?

A. Momentum

B. Short-term reversal

C. Liquidity neutrality

D. Pure diversification

Correct Answer: A. Momentum

Question 8

Semi-strong-form market efficiency implies that prices reflect:

A. Only private information

B. Historical prices only

C. All publicly available information

D. No information

Correct Answer: C. All publicly available information

Question 9

Before costs, aggregate active investors collectively hold:

A. The market portfolio

B. Only Treasury bills

C. A zero-risk portfolio

D. Only small-cap stocks

Correct Answer: A. The market portfolio

Question 10

If the risk-free rate is 5%, the market risk premium is 6%, and beta is 0.5, CAPM expected return equals:

A. 5.5%

B. 6.0%

C. 8.0%

D. 11.0%

Correct Answer: C. 8.0%