Capital Asset Pricing and Arbitrage Pricing Theory

Capital Asset Pricing and Arbitrage Pricing Theory Notes

7.1 The Capital Asset Pricing Model

  • Assumptions:

    • Markets are competitive and equally profitable.

    • No investor is wealthy enough to individually affect prices.

    • All information is publicly available; all securities are public.

    • There are no taxes on returns and no transaction costs.

    • Unlimited borrowing/lending is available at the risk-free rate.

    • Investors are alike except for differences in initial wealth and risk aversion.

    • Investors plan for a single-period horizon; they are rational and mean-variance optimizers.

    • All investors use the same inputs and consider identical portfolio opportunity sets.

  • Equilibrium:

    • All investors choose to hold the market portfolio.

    • The market portfolio contains all securities, with the proportion of each security corresponding to its market value as a percentage of total market value.

    • The market portfolio lies on the efficient frontier, representing the optimal risky portfolio.

  • Expected Return on Individual Securities:

    • The risk premium on individual securities is a function of the security's contribution to the risk of the market portfolio.

    • The individual security’s risk premium is determined by the covariance of its returns with the assets making up the market portfolio.

  • Example (Dell):

    • Rearranging provides the CAPM's expected return-beta relationship.

  • The Security Market Line (SML):

    • Represents the expected return-beta relationship as described by the CAPM.

    • Graphs individual asset risk premiums against asset risk.

    • Alpha:

    • Defined as the abnormal rate of return on a security that exceeds the rate predicted by the equilibrium model (CAPM).

  • Applications of CAPM:

    • The SML serves as a benchmark for fair returns on risky assets.

    • The SML provides a “hurdle rate” for evaluating internal projects.

  • Portfolio Beta:

    • The beta of a portfolio is the weighted average of the betas of the individual stocks within the portfolio.

    • Example Calculation:

    • If the portfolio consists of stocks with weights:

      • 40% with a beta of 1.2

      • 60% with a beta of 1.5,

    • Then Portfolio Beta = 0.4 imes 1.2 + 0.6 imes 1.5 = 1.38.

7.2 CAPM and Index Models

  • Estimation of the Index Model:

    • Utilizes historical data from T-bills, S&P 500, and individual securities.

    • Involves regressing risk premiums for individual stocks against those for the S&P 500.

    • The slope of the regression (beta) represents the beta for the individual stock.

    • Index Model Mean-Beta Equation:

    • r{it} - r{ft} = \alphai + \betai (r{Mt} - r{ft}) + e_{it} where:

      • r_{it}: Holding period return for asset $i$ at time $t$.

      • r_{ft}: Risk-free return.

      • \alpha_i: Intercept of the security characteristic line.

      • \beta_i: Slope of the security characteristic line.

      • r_M: Index return.

      • e_{it}: Firm-specific effects.

  • Expected Return Equation:

    • E(r{it}) - r{ft} = \alphai + \betai [E(r{Mt}) - r{ft}].

  • Figure 7.4 Scatter Diagram for Google vs. S&P 500 (01/06-12/10):

    • Displays excess rate of return on Google arranged against the excess rate of return on S&P 500.

  • Table of Regression Statistics for Google (S&P 500), 01/06-12/10:

    • R = 0.5914

    • R-squared = 0.3497

    • Adjusted R-squared = 0.3385

    • Standard Error of regression = 8.4585

    • Total observations = 60

    • Regression equation:

    • Google (excess return) = 0.8751 + 1.2031 × S&P 500 (excess return).

    • ANOVA results:

    • Regression degrees of freedom = 1, Sum of Squares = 2231.50, Mean Square = 2231.50, F-statistic = 31.19, p-level = 0.0000.

    • Residual degrees of freedom = 58, Sum of Squares = 4149.65, Mean Square = 71.55, Total = 59, Sum of Squares = 6381.15.

    • Coefficients:

    • Intercept: 0.8751 ext{ (SE: 1.0920, t: 0.8013, p: 0.4262)}

    • S&P 500: 1.2031 ext{ (SE: 0.2154, t: 5.5848, p: 0.0000)}

7.3 CAPM and the Real World

  • The CAPM was introduced by Sharpe in 1964.

  • Numerous tests followed the theory, including:

    • Roll’s critique in 1977.

    • Fama and French study in 1992.

  • Validations and Critiques:

    • CAPM is considered false based on the validity of its original assumptions but remains a useful predictor of expected returns.

    • CAPM is regarded as untestable as a theory; nonetheless, its principles are still relevant:

    • Investors should diversify their portfolios.

    • Systematic risk remains the risk that matters most.

    • A well-diversified risky portfolio can serve a wide range of investors effectively.

7.4 Multifactor Models and CAPM

  • Limitations of CAPM:

    • The market portfolio is not directly observable in actual markets.

    • Research indicates other factors can influence returns beyond those accounted for in CAPM.

  • Fama-French Three-Factor Model:

    • Suggests returns are tied to factors other than merely market returns:

    • Size.

    • Book value relative to market value.

    • This three-factor model offers a better explanation of return variations compared to CAPM.

7.5 Arbitrage Pricing Theory (APT)

  • Arbitrage:

    • Defined as the practice of capitalizing on relative mispricing in order to generate riskless profits.

  • Arbitrage Pricing Theory (APT):

    • Describes risk-return relationships originating from no-arbitrage conditions in expansive capital markets.

    • APT asserts that for well-diversified portfolios (with no firm-specific risk), the no-arbitrage condition yields the same risk-return relationship as CAPM.

  • Comparisons Between APT and CAPM:

    • APT applies to well-diversified portfolios but not necessarily to individual stocks.

    • Some individual stocks can be mispriced and not align with the SML within the APT framework.

    • APT is more versatile, deriving expected return and beta relationships without relying on the existence of a market portfolio.

    • APT can be adapted to include multifactor models, extending its applicability to various risk factors.