Risk, Return and Portfolio Theory

Introduction and Context

Investment and corporate finance decisions are made under uncertainty. A security may generate income and capital appreciation, but the realized outcome is unknown in advance. Finance therefore studies two linked questions: what return should an investor expect, and what risks must be accepted to obtain it?

Risk–return analysis supports asset selection, portfolio construction, capital budgeting, performance evaluation, cost-of-capital estimation, and strategic decisions such as acquisitions or international expansion. The central insight is that risk is not a single, indivisible concept. Some uncertainty can be reduced through diversification, whereas some is associated with economy-wide forces and cannot be eliminated by holding more securities.

This chapter develops the measurement of return and risk, modern portfolio theory, the Capital Asset Pricing Model (CAPM), beta, multifactor models, and performance measures. It also examines assumptions, empirical limitations, governance issues, and managerial applications.

Learning Objectives

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

1. Calculate holding-period return, expected return, variance, standard deviation, covariance, and correlation.
2. Explain how diversification affects portfolio risk.
3. Construct and interpret efficient portfolios using modern portfolio theory.
4. Distinguish systematic risk from unsystematic risk and explain beta.
5. Apply CAPM and multifactor models to investment and corporate decisions.
6. Evaluate portfolios using Sharpe, Treynor, Jensen’s alpha, and related measures.
7. Critically assess model assumptions, estimation problems, ethical risks, and practical limitations.

1. Return: Measurement and Interpretation

1.1 Holding-period return

For an investment purchased at price P₀, sold at P₁, and generating cash income D₁ during the period, the holding-period return is:

Here, R is the holding-period return, P₀ and P₁ are beginning and ending prices, and D₁ is dividend, interest, or other cash distribution. The formula assumes that cash flows are measured over the same period and that the beginning price is the capital committed.

The return has two components: capital gain yield, (P₁ − P₀) ÷ P₀, and income yield, D₁ ÷ P₀. For example, if a share is purchased for $50, sold for $54, and pays a $1 dividend, the return is (54 − 50 + 1) ÷ 50 = 0.10, or 10%.

For multiple periods, the arithmetic average is useful for estimating a one-period expected return, while the geometric average describes compounded historical growth:

The geometric mean is generally less than or equal to the arithmetic mean when returns vary. It is the more appropriate measure of realized compound growth; the arithmetic mean is commonly used as an unbiased estimator of a future single-period return under particular statistical assumptions.

1.2 Expected return

When there are possible states of the world, expected return is the probability-weighted average:

where p_s is the probability of state s, R_s is the return in that state, and ∑p_s = 1. Probabilities must be non-negative and collectively exhaustive. Expected return is not a guaranteed return; it is a probability-weighted forecast.

For a portfolio with asset weights w_i and expected returns E(R_i):

The weights normally sum to one for a fully invested portfolio. If short selling or borrowing is permitted, some weights may be negative or exceed one, but the economic and regulatory implications must be examined.

2. Risk: Variance, Volatility, Covariance and Correlation

2.1 Variance and standard deviation

For a discrete probability distribution, variance is:

Standard deviation is the square root of variance:

Returns are dimensionless percentages, so variance is expressed in squared percentage units and standard deviation in percentage units. Standard deviation measures total dispersion around the expected return. It treats positive and negative deviations symmetrically and is most informative when returns are reasonably represented by a stable distribution and investors have appropriate preferences toward risk.

For historical sample data, the sample standard deviation uses n − 1 rather than n in the denominator:

The distinction reflects estimation from a sample rather than observation of the complete probability distribution.

2.2 Covariance

Covariance measures how two returns move together:

Positive covariance indicates a tendency to move in the same direction; negative covariance indicates opposite movement. Covariance is scale-dependent and can be difficult to interpret alone.

Correlation standardizes covariance:

Correlation ranges from −1 to +1. A correlation of +1 means perfect co-movement, 0 means no linear co-movement, and −1 means perfect inverse co-movement. Zero correlation does not prove independence and does not rule out nonlinear dependence.

3. Diversification and Portfolio Risk

For a two-asset portfolio, expected return is:

with w_A + w_B = 1. Portfolio variance is:

Using correlation, the equivalent expression is:

The covariance term explains the diversification benefit. Portfolio risk depends not only on the risk of each asset but also on how the assets interact. Even if both assets are individually volatile, a correlation below +1 can reduce portfolio volatility.

For N assets, portfolio variance is:

In matrix notation:

where w is the vector of portfolio weights, Σ is the covariance matrix, and w′ is the transpose of the weight vector. This formulation is essential for practical portfolio optimization.

Worked example: two-asset diversification

Asset A has expected return 8% and standard deviation 10%. Asset B has expected return 14% and standard deviation 20%. An investor allocates 60% to A and 40% to B. Their correlation is 0.20.

Expected return:

Variance:

Therefore:

The portfolio’s expected return is 10.4%, while its risk is approximately 10.92%. The result is lower than a simple weighted average of individual standard deviations, 14%, because correlation is below +1.

Systematic and unsystematic risk

Unsystematic risk is security-specific risk arising from events such as product failure, managerial misconduct, litigation, or an isolated supply disruption. It can be substantially reduced through diversification.

Systematic risk arises from broad economic or market forces, including inflation, interest rates, recessions, exchange rates, and geopolitical shocks. It cannot be eliminated merely by holding many securities in the same market.

Total risk is often expressed conceptually as:

This is a risk decomposition rather than a universal additive equation for variances. In variance terms, the relationship depends on the model and on whether covariance terms are included. A diversified investor should focus particularly on risk that contributes to portfolio-wide fluctuations.

4. Modern Portfolio Theory

Harry Markowitz’s modern portfolio theory (MPT) formalized portfolio selection as a trade-off between expected return and risk. Under the mean–variance framework, investors compare portfolios using expected return and variance, assuming that either returns are approximately normally distributed or investors’ preferences can be represented adequately by mean and variance.

A portfolio is efficient if no alternative portfolio offers higher expected return for the same or lower risk, or lower risk for the same or higher expected return. The set of efficient portfolios forms the efficient frontier. Portfolios below the frontier are dominated.

A constrained optimization problem can be written as:

Here, μ is the vector of expected asset returns and μ_p is the target portfolio return. Additional constraints may prohibit short selling, impose liquidity limits, or require minimum and maximum allocations.

The risk-free asset and the capital market line

If a risk-free asset offering return R_f is available, combinations of the risk-free asset and a risky portfolio lie on a straight line in expected-return–standard-deviation space. The line with the greatest slope is the capital market line (CML), associated in the standard CAPM setting with the market portfolio:

The slope, (E(R_M) − R_f) ÷ σ_M, is the market portfolio’s Sharpe ratio. The CML applies to efficient portfolios formed from the risk-free asset and the market portfolio; it should not be confused with the security market line.

5. CAPM and Beta

The Capital Asset Pricing Model links an asset’s expected return to its non-diversifiable risk:

where E(R_i) is the required or expected return on asset i, R_f is the risk-free rate, E(R_M) is the expected market return, and β_i is the asset’s market beta. The quantity E(R_M) − R_f is the market risk premium.

Beta measures sensitivity to market returns:

A beta of 1.0 implies market-level sensitivity; a beta above 1.0 indicates greater sensitivity, and a beta below 1.0 indicates lower sensitivity. A negative beta is possible but uncommon for ordinary equities. Beta is not a complete measure of total risk: a low-beta security may still have substantial idiosyncratic volatility.

The security market line (SML) graphs CAPM-required return against beta. Unlike the CML, it applies conceptually to individual securities and portfolios, including inefficient portfolios, provided the CAPM assumptions hold.

CAPM assumptions and interpretation

The textbook CAPM assumes, among other conditions:

• investors are risk-averse and choose portfolios using mean and variance;
• investors have homogeneous expectations;
• investors can borrow and lend at a common risk-free rate;
• markets are frictionless, competitive, and liquid;
• assets are divisible and investors can trade over the same horizon;
• systematic risk is represented adequately by market beta.

These assumptions are simplifications rather than literal descriptions of markets. CAPM remains useful as a benchmark for estimating required returns and cost of equity, but beta and the market risk premium must be estimated with judgment.

Worked CAPM example

Suppose the risk-free rate is 4%, the expected market return is 10%, and a project or security has beta 1.3. The required return is:

A project expected to generate 14% would exceed this benchmark by 2.2 percentage points, although the decision should also consider estimation risk, liquidity, taxes, financing effects, and project-specific risks not captured by beta.

6. Multifactor Models and Extensions

Empirical research shows that market beta alone may not explain all systematic patterns in returns. A multifactor model expresses return as a function of several common risk factors:

Here, α_i is abnormal average performance after controlling for factors, F_k are factor premiums, β_ik are sensitivities, and ε_i is security-specific residual risk. The residual is assumed, in the model, to have an expected value of zero and to be uncorrelated with included factors; these assumptions require empirical testing.

Important approaches include:

• Arbitrage Pricing Theory (APT): expected returns reflect exposure to multiple pervasive factors, with arbitrage restrictions supporting the relationship but not requiring a single market portfolio.
• Fama–French models: commonly include market, size, and value factors; later versions add profitability and investment factors.
• Momentum models: include a factor based on recent relative performance.
• Macroeconomic models: use variables such as inflation, industrial production, interest rates, or credit spreads.

Multifactor models can improve risk attribution and portfolio construction, but factors may be correlated, unstable, crowded, or difficult to trade. A statistically significant factor is not automatically a reliable investment premium. Researchers must address data mining, transaction costs, survivorship bias, look-ahead bias, and factor-definition ambiguity.

7. Portfolio Performance Measurement

Performance measures must match the risk being evaluated and the investor’s opportunity set.

Sharpe ratio

The Sharpe ratio measures excess return per unit of total volatility. It is suitable for evaluating a complete portfolio when total risk matters. For realized performance, average excess return and sample standard deviation are used. It can be distorted by non-normal returns, illiquidity, smoothing, and unusual choice of measurement interval.

Treynor ratio

The Treynor ratio measures excess return per unit of systematic risk. It is most appropriate when the portfolio is already well diversified and beta is the relevant marginal risk.

Jensen’s alpha

A positive alpha indicates performance above the CAPM benchmark for the measured period, while a negative alpha indicates underperformance. Alpha is model-dependent: changing the benchmark or factors may change the conclusion.

Information ratio

Here, R_B is benchmark return and TE is tracking-error volatility, the standard deviation of active returns. The information ratio evaluates active management relative to a chosen benchmark.

No performance measure should be interpreted without considering fees, taxes, leverage, liquidity, benchmark appropriateness, time horizon, and statistical uncertainty.

8. Corporate Finance Applications

Risk–return analysis informs the cost of equity and weighted average cost of capital (WACC). A simplified WACC expression is:

where E and D are market values of equity and debt, R_e is the required return on equity, R_d is the pre-tax cost of debt, and T is the corporate tax rate. This formula assumes debt interest is tax-deductible and that the tax treatment is applicable.

CAPM may estimate R_e, but project risk should be assessed rather than automatically equated with the firm’s historical beta. A project in a different industry or country may have a different operating and financial risk profile. Practical methods include pure-play comparables, bottom-up beta estimation, scenario analysis, sensitivity analysis, simulation, and risk-adjusted discount rates.

Diversification also affects corporate strategy. A conglomerate may reduce exposure to one industry, but shareholders can often diversify more cheaply in capital markets. Therefore, corporate diversification creates value only when it supplies operating synergies, financing benefits, information advantages, or risk reduction that investors cannot obtain independently.

9. Illustrative Case: Meridian Energy Storage Ltd.

The following is a fictional illustrative case.

Background

Meridian Energy Storage is considering a $40 million investment in a battery-storage project. Management expects the project to produce cash flows that are sensitive to electricity prices and interest rates. The company’s current equity beta is 0.90, but comparable listed storage businesses have betas between 1.20 and 1.50. Meridian’s existing operations are more stable than the proposed project.

Problem

The chief financial officer must recommend a discount rate and decide whether to include the project in the firm’s strategic portfolio. Using the company-wide beta would potentially understate the project’s systematic risk.

Analysis

The CFO first identifies pure-play comparables and estimates an asset beta by removing the effects of their capital structures. A simplified unlevering relationship is:

where β_A is asset beta, β_E is equity beta, D ÷ E is the debt-to-equity ratio, and T is the tax rate. The formula assumes a conventional tax shield and is an approximation; more advanced analysis may address changing leverage and multiple debt classes.

The median asset beta of the comparables is then relevered to Meridian’s intended financing structure. Suppose the resulting project equity beta is 1.35. If the risk-free rate is 4.5% and the estimated market risk premium is 5.5%, CAPM gives:

Management also performs scenarios: a base case, a low electricity-price case, and a high-interest-rate case. It evaluates whether the project remains acceptable under adverse conditions rather than relying solely on a single expected cash-flow forecast.

Approach and outcome

Meridian approves the project only if its risk-adjusted net present value is positive under the base case and remains tolerable under downside scenarios. It establishes investment limits, monitors debt exposure, and reports assumptions to the board. The company does not treat diversification as a substitute for liquidity planning or operational controls.

Learning points

1. The relevant risk is the risk of the project, not necessarily the historical risk of the parent company.
2. Comparable-company analysis can improve beta estimation when direct history is limited.
3. CAPM is a benchmark, not a mechanical answer.
4. Scenario analysis complements, rather than replaces, portfolio and market-risk models.
5. Transparent assumptions are essential for governance and accountability.

10. Limitations, Governance and Contemporary Practice

Estimation and model risk

Expected returns, covariances, betas, and risk premiums are estimated from imperfect data. Covariance matrices become unstable as the number of assets grows relative to the number of observations. Historical relationships may break during crises. Optimization based on small differences in estimated inputs can generate extreme and fragile portfolios.

Best practices include shrinkage or robust covariance estimation, sensitivity analysis, resampling, economically justified constraints, out-of-sample testing, and explicit treatment of transaction costs and liquidity. Value at Risk (VaR) and expected shortfall may supplement mean–variance analysis, but they do not eliminate model risk. Stress testing is especially important for nonlinear exposures and extreme events.

Behavioral and institutional considerations

Investors may display loss aversion, overconfidence, home bias, narrow framing, and excessive trading. Institutions face mandates, benchmark pressure, liquidity needs, regulatory capital constraints, and agency conflicts. A theoretically efficient portfolio may therefore be unsuitable for a pension fund, bank, family office, or operating company with different liabilities.

Ethics and regulation

Risk models can create false confidence if assumptions are hidden or results are selectively reported. Professionals should disclose data limitations, model uncertainty, conflicts of interest, fees, leverage, and downside scenarios. Misrepresenting back-tested factor performance, omitting material risks, or using unsuitable benchmarks can harm clients and violate fiduciary or securities obligations. Governance should include independent model validation, documented approval thresholds, segregation of duties, audit trails, and periodic review.

Technology and AI

Modern portfolio systems use high-dimensional data, machine learning, alternative data, and automated rebalancing. These tools can improve forecasting or detect nonlinear patterns, but they also introduce overfitting, data leakage, unstable relationships, cybersecurity risks, and explainability problems. AI-generated forecasts should be benchmarked against simple models, tested out of sample, monitored for drift, and subject to human oversight. Automation does not transfer accountability away from investment decision-makers.

Chapter Summary

Risk and return are inseparable foundations of finance. Holding-period return combines income and price change, while expected return is a probability-weighted forecast. Variance and standard deviation measure dispersion; covariance and correlation explain how assets interact. Portfolio risk depends on weights, individual volatilities, and co-movement, so diversification can reduce unsystematic risk but cannot eliminate systematic risk.

Modern portfolio theory identifies efficient portfolios through the trade-off between expected return and variance. The CAPM states that required return depends on beta, or sensitivity to market risk, rather than total stand-alone volatility. Multifactor models extend this framework by incorporating additional systematic sources such as size, value, profitability, investment, momentum, and macroeconomic conditions.

Sharpe, Treynor, Jensen’s alpha, and information ratios provide different performance perspectives and must be matched to the relevant risk and benchmark. Corporate decisions require project-specific risk assessment, not uncritical use of company-wide averages. Finally, all models are estimates subject to data, behavioral, liquidity, governance, and ethical limitations. Effective practice combines quantitative analysis with judgment, transparency, scenario testing, and continuous monitoring.

Key Terms

Arbitrage Pricing Theory (APT): A multifactor asset-pricing framework based on the absence of persistent arbitrage opportunities.

Beta: Sensitivity of an asset’s return to market return, commonly measured as covariance with the market divided by market variance.

Capital market line (CML): The risk–return line for combinations of the risk-free asset and the efficient market portfolio.

Capital Asset Pricing Model (CAPM): Model linking required return to the risk-free rate and systematic market beta.

Correlation: Standardized measure of linear co-movement between two returns.

Covariance: Measure of joint movement between two random variables.

Diversification: Combining assets to reduce portfolio-specific risk.

Efficient frontier: Portfolios offering the highest expected return for each level of risk.

Expected return: Probability-weighted average of possible returns.

Information ratio: Active return divided by tracking-error volatility.

Jensen’s alpha: Return above or below the CAPM-predicted return.

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

Modern portfolio theory: Framework that evaluates portfolios through expected return, variance, and covariance.

Multifactor model: Return model using several systematic risk factors.

Security market line (SML): Relationship between expected return and beta under CAPM.

Sharpe ratio: Excess return per unit of total volatility.

Standard deviation: Square root of variance and a measure of return volatility.

Systematic risk: Market-wide risk that cannot be eliminated through diversification.

Treynor ratio: Excess return per unit of systematic beta risk.

Unsystematic risk: Security-specific risk that diversification can substantially reduce.

Review Questions

5 Short Answer Questions

1. Distinguish between arithmetic and geometric average return.
2. What does covariance reveal about two assets?
3. Why can diversification reduce risk?
4. Define beta and explain what a beta of 1.4 implies.
5. State one difference between the CML and the SML.

5 Long Answer Questions

1. Explain the assumptions, logic, and limitations of modern portfolio theory.
2. Derive and interpret the two-asset portfolio variance formula.
3. Discuss systematic and unsystematic risk and their relevance to corporate decision-making.
4. Compare CAPM with multifactor models, including their practical advantages and weaknesses.
5. Explain how Sharpe, Treynor, Jensen’s alpha, and information ratio measure performance differently.

3 Analytical / Application Questions

1. Asset X has an expected return of 9%, standard deviation of 12%, and weight of 70%; Asset Y has an expected return of 15%, standard deviation of 22%, and weight of 30%. If their correlation is 0.10, calculate portfolio expected return and standard deviation.
2. A firm is evaluating a project with beta 1.25. The risk-free rate is 3.5% and the expected market risk premium is 6%. Calculate the CAPM required return and explain two reasons why management should not rely on this number alone.
3. Design a governance process for an AI-supported portfolio optimizer. Identify at least four model, data, ethical, or operational risks and propose controls for each.

Multiple Choice Questions with Answers

Question 1

An investment is purchased for $40, sold for $43, and pays a $1 dividend. What is its holding-period return?

A. 2.5%

B. 7.5%

C. 10.0%

D. 12.5%

Correct Answer: B. 7.5%

Question 2

Which measure standardizes covariance into a range from −1 to +1?

A. Variance

B. Beta

C. Correlation

D. Alpha

Correct Answer: C. Correlation

Question 3

Two assets have a correlation of less than +1. Holding both assets can reduce portfolio risk primarily because:

A. Expected returns become risk-free

B. Their returns do not move perfectly together

C. Both assets necessarily have negative returns

D. Portfolio variance ignores asset weights

Correct Answer: B. Their returns do not move perfectly together

Question 4

An asset has beta 1.2, a risk-free rate of 4%, and a market risk premium of 6%. According to CAPM, its required return is:

A. 7.2%

B. 10.0%

C. 11.2%

D. 12.0%

Correct Answer: C. 11.2%

Question 5

Which risk is generally reducible through broad diversification?

A. Inflation risk

B. Economy-wide recession risk

C. Security-specific litigation risk

D. Market-wide interest-rate risk

Correct Answer: C. Security-specific litigation risk

Question 6

The Sharpe ratio uses which denominator?

A. Portfolio beta

B. Portfolio standard deviation

C. Tracking error only

D. Market variance

Correct Answer: B. Portfolio standard deviation

Question 7

A portfolio’s expected return is calculated as:

A. The highest asset return in the portfolio

B. The unweighted average of asset returns in all cases

C. The weighted sum of individual expected returns

D. The portfolio standard deviation multiplied by beta

Correct Answer: C. The weighted sum of individual expected returns

Question 8

Which statement correctly distinguishes the CML from the SML?

A. The CML uses beta, while the SML uses total volatility

B. The CML applies to combinations of the risk-free asset and market portfolio, while the SML relates return to beta

C. The CML measures alpha, while the SML measures covariance

D. The CML applies only to individual securities, while the SML applies only to firms

Correct Answer: B. The CML applies to combinations of the risk-free asset and market portfolio, while the SML relates return to beta

Question 9

A portfolio earns 12%, the risk-free rate is 3%, and its standard deviation is 15%. Its Sharpe ratio is:

A. 0.20

B. 0.60

C. 0.80

D. 1.25

Correct Answer: B. 0.60

Question 10

In a multifactor model, the residual term generally represents:

A. The risk-free rate

B. The market risk premium

C. Security-specific variation not explained by included factors

D. The portfolio’s expected return

Correct Answer: C. Security-specific variation not explained by included factors