Risk, Return

,
and the Capital Asset Pricing Model
(CAPM)

Topics in Chapter
Basic return concepts
Basic risk concepts
Stand-alone risk / Total risk
Portfolio (market) risk
Risk and return:
CAPM/SML

2

What are investment returns?
Investment returns measure the financial results of an
investment.
Returns may be historical or prospective
(anticipated).
Returns can be expressed in:
Dollar terms.
Percentage terms.

3

Expected (holding period) returns
Expected rate of return on an investment asset can
be calculated as follows:

Expected Return  Expected ending value 
Cost
Example: if \$1,000 is invested and \$1,100 is returned after
one year, the rate ofCost  for this investment is:
return
(\$1,100 – \$1,000)/\$1,000 = 10%.

4

What is investment risk?
Typically, investment returns are not known
with certainty (as they depend on an
expected ending value; e.g., price, dividends.)
The greater the chance that the expected
returns will not be realized, the riskier
the investment.
An asset’s risk can be analyzed in two
ways
On a stand-alone basis (stand-alone risk)
On a portfolio basis (portfolio risk)

5

Probability Firm X Firm Y Rate of -70 0 15 100 Return (%) Expected Rate of Return 6 .Probability Distributions: Which stock is riskier? Why? A listing of all possible outcomes. Can be shown graphically (or in tabular form). and the probability of each occurrence.

2 5.0% -3.1 5.0% -21.0% 25.0% 13.0% -11. T-Bill HT Coll USR Market Recession 0.0% Risk free? 7 .0% 38.0% Above avg 0.0% 41.0% Below avg 0.2 5.0% -17.0% -14.0% 0.0% 10.5% 30.0% 6.0% 27.5% 15.4 5.5% -7.0% Boom 0.0% 26.Investment Alternatives Economy Prob.0% Average 0.1 5.0% 3.5% 45.5% -27.

very little unexpected inflation is likely to occur over such a short period of time.Why is the T-bill return independent of the economy? Do T-bills promise a completely risk-free return? T-bills will return the promised 5. regardless of the economy and hence your nominal return is truly risk free. 8 . as they are still exposed to inflation. T-bills are also risky in terms of reinvestment rate risk.5%. T-bills are risk-free in the default sense of the word. But. Although. T-bills do not provide a completely risk-free return.

and has a negative correlation. behave in relation to the market? HT – Moves with the economy. and has a positive correlation. Coll. – Is countercyclical with the economy. This is unusual. This is typical. 9 .How do the returns of HT and Coll.

4) riPi + (30%)(0.Calculating the Expected Return ˆr = Expected rate of return N ˆr =  ˆrH = (-27%)(0.2) + (45%)(0.1) + (-7%)(0.4% 10 .2) + (15%)(0.1) T i=1 = 12.

4% Market 10. but is it really? Have we failed to account for risk? 11 . and appears to be the best investment alternative.Summary of Expected Returns Expected return HT 12.5% USR 9.0% HT has the highest expected return.5% Coll. 1.8% T-bill 5.

Calculating Standard Deviation σi = Standard deviation σi = Variance = σ 2i N σi =  (r - i=1 rˆ)2P i i 12 .

5) (0.5)2(0.5) (0.8% σM = 15.Standard Deviation for Each Investment N σi =  (r -i i i=1 ˆr)2P 2   T-bills 2  + (5.5)2(0.2% σHT = 20% σ = (5.5-5.5-5.5-5.4) + (5.5-5.5-5.1) + (5.2% 13 .5)2(0.1)    1/2 σT-bills=0.0% (5.2)  σColl = 13.2) σUSR = 18.

r i )2 σi = N-1  An example: Page 228 of the Text.Expected Return and Risk Using Historical Data Expected Return = Average (arithmetic) of historical returns. t . N  (r t=1 i. N r i. 14 . ri = t t=1 Standard Dev = Square N root of sum of the squared deviations of historical individual returns from the average/expected return.

0 L-T corporate bonds 6.3 20.2 8.2 U. Bills. 15 . 2008).4 L-T government bonds 5.1 Source: Based on Stocks.S.8 3. and Inflation: (Valuation Edition) 2008 Yearbook (Chicago: Morningstar.8 9.1926-2007 Average Standard Return Deviation Small-company stocks 17. Treasury bills 3.1% 32. Inc..Selected Realized Returns.6% Large-company stocks 12. Bonds.

T-bill USR HT 0 5.4 Rate of Return (%) 16 .5 9.8 12.Comparing Standard Deviations Prob .

17 .alone. The larger σi is. risk. or stand. Larger σi is associated with a wider probability distribution of returns. the lower the probability that actual returns will be closer to expected returns.Comments on Standard Deviation as a Measure of Risk Standard deviation (σi) measures total.

8 Market 10. 18 .0 Coll* 1.0 13.2 USR* 9.5 15. ˆr Risk.0% HT 12.2 *Seems out of place.4 20.5% 0.Comparing Risk and Return Security Expected Return.8 18.  T-bills 5.

that shows the risk per unit of return.Coefficient of Variation (CV) A standardized measure of dispersion about the expected value. Standard deviation σi CV i = Expected return = ˆ i r 19 .

2 USR 1. has a relatively average CV.6 Coll. despite having the highest standard deviation of returns.0 HT 1.Risk Rankings by Coefficient of Variation CV T-bill 0. 13.4 Collections has the highest degree of risk per unit of return. HT. 20 .9 Market 1.

A B 0 Rate of Return (%) σA = σB . but A is riskier because of a larger probability of losses. the same amount of risk (as measured by σ) for smaller returns. 21 .Illustrating the CV as a Measure of Relative Risk Prob. In other words.

which serves as compensation for investors to hold riskier securities. Diversification: Shed part of the total (stand- alone) risk by investing in a portfolio of assets. Risk premium: Difference between the return on a risky asset and a riskless asset.Investor Attitude towards Risk Risk aversion: Investors are assumed to dislike risk and require higher rates of return to encourage them to hold riskier securities. 22 .

Portfolio Construction: Risk and Return Assume a two-stock portfolio is created with \$50. A portfolio’s expected return is a weighted average of the returns of the portfolio’s component assets. 23 . Standard deviation is a little more tricky and requires that a new probability distribution for the portfolio returns be devised.000 invested in both HT and Collections.

Calculating Portfolio Expected Return rp is a weighted average: ˆ N ^ r =  w i ri ˆp i=1 rp = 0.7% ˆ 24 .5 (12.0%) = 6.5 (1.4%) + 0.

4 15.0% 0.0%) + 0.10 (12.0% -21.40 (7.2 -7.0% 3.1 -27.0% Average 0.An Alternative Method for Determining Portfolio Expected Return Economy Prob.0% Below avg 0.20 (9. Recession 0.20 (3.0% 7.0% 12.0% ˆ = 0.2 30.0% 27. HT Coll Port.0% 13.0% 0.5% Above avg 0.0%) = 6.10 (0.0%) + 0.5%) rp + 0.0% -11.5% Boom 0.5%) + 0.7% 25 .0% 9.1 45.

6.6.Calculating Portfolio Standard Deviation and CV 1  2  0.4% σ p = + 0.7) 2  3.20 (3.5 .51 26 .40 (7.10 (12.7)2   + 0.5 .10 (0.4% CVp = 6.0 .20 (9.7) 2   2  + 0.7)2   + 0.7) = 3.7% = 0.0 .6.0 .6.6.

σColl.0%.2%). Why? Negative correlation between stocks.4% is much lower than the σi of either stock (σHT = 20. σp = 3. = 13.’s σ (16. but lower than the average risk. Therefore.6%). 27 .4% is lower than the weighted average of HT and Coll.Comments on Portfolio Risk Measures σp = 3. the portfolio provides the average return of component stocks.

Most stocks are positively (though not perfectly) correlated with the market (i. 28 .. Combining stocks in a portfolio generally lowers risk. ρ between 0 and 1).General Comments about Risk σ  35% for an average stock.e.

0) 29 .Returns Distribution for Two Perfectly Negatively Correlated Stocks (ρ = -1.

Returns Distribution for Two Perfectly Positively Correlated Stocks (ρ = 1.0) Stock M Stock M’ Portfolio MM’ 25 25 25 15 15 15 0 0 0 -10 -10 -10 30 .

Partial Correlation.35 31 . ρ = +0.

Eventually the diversification benefits of adding more stocks dissipates (after about 10 stocks). σp tends to converge to  20%. 32 . Expected return of the portfolio would remain relatively constant.Creating a Portfolio: Beginning with One Stock and Adding Randomly Selected Stocks to Portfolio σ decreases as stocks added. because they p would not be perfectly correlated with the existing portfolio. and for large stock portfolios.

Illustrating Diversification Effects of a Stock Portfolio 33 .

rM)/Var(rM) Diversifiable risk – portion of a security’s stand- alone risk that can be eliminated through proper diversification.Breaking Down Sources of Risk Stand-alone risk = Market risk + Diversifiable risk Market risk – portion of a security’s stand- alone risk that cannot be eliminated through diversification. Measured by beta. beta = Cov(ri. 34 .

diversifiable risk. which is based upon market risk. There can be only one price (the market return) for a given security. No compensation should be earned for holding unnecessary.Failure to Diversify If an investor chooses to hold a one-stock portfolio (doesn’t diversify). risk-averse investors are concerned with σp. would the investor be compensated for the extra risk they bear? NO! Stand-alone risk is not important to a well-diversified investor. 35 . Rational.

diversified (market) portfolio. 36 .Capital Asset Pricing Model (CAPM) Model linking risk and required returns. CAPM suggests that there is a Security Market Line (SML) that states that a stock’s required return equals the risk-free return plus a risk premium that reflects the stock’s risk after diversification. which is measured by the stock’s beta. ri = rRF + (rM – rRF)bi Primary conclusion: The relevant riskiness of a stock is its contribution to the riskiness of a well.

Beta Measures a stock’s market risk. 37 . and shows a stock’s volatility relative to the market. Indicates how risky a stock is if the stock is held in a well-diversified portfolio.

X) N-1 2 38 .Beta beta = Cov(ri.Y) Variance: Var(X) = (X .rM)/Var(rM) Covariance: Cov(XY) =  (X -NX)(Y -1 .

39 .0. If beta < 1.5.5 to 1.0. the security is riskier than average.0. the security is just as risky as the average stock. Most stocks have betas in the range of 0. the security is less risky than average.Comments on Beta If beta = 1. If beta > 1.

a negative beta is highly unlikely.e.Can the beta of a security be negative? Yes. However. ρi..m < 0). if the correlation between Stock i and the market is negative (i. If the correlation is negative. the regression line would slope downward. and the beta would be negative. 40 .

The slope of the regression line is defined as the beta coefficient for the security. A typical approach to estimate beta is to run a regression of the security’s past returns against the past returns of the market. analysts are forced to rely on historical data. 41 . Without a crystal ball to predict the future.Calculating Betas Well-diversified investors are primarily concerned with how a stock is expected to move relative to the market in the future.

. -5 ^ r = -2. 20 Year rM ri 1 15 18 15 2 % % 10 3 -5 - 12 10 5 16 -5 0 5 10 15 20 rM Regression line: .59 + 1.44 r^ i M -10 42 .Illustrating the Calculation of Beta ri .

X)(Y .X)2 The formula for the intercept is: a = Y .Y)  (X .Calculation of Beta The formula for the coefficient or slope in simple linear regression is: b=  (X .bX 43 .

87 -20 44 .32 40 20 T-bills: b = 0 -20 0 20 40 rM Coll: b = -0. Coll. and T-Bills ri HT: b = 1.Beta Coefficients for HT.

Analysts typically use four or five years’ of monthly returns to establish the regression line. 45 .Calculating Beta in Practice Many analysts use the S&P 500 to find the market return. Some analysts use 52 weeks of weekly returns.

5 1. so the rank order is OK. 46 .4% 1.32 Market 10.Comparing Expected Returns and Beta Coefficients Security Expected Return Beta HT 12.87 Riskier securities have higher returns.8 0.00 USR 9. 1.00 Coll.5 0.88 T-Bills 5.0 -0.

0%.5% and RPM = 5. 47 .The Security Market Line (SML): Calculating Required Rates of Return SML: ri = rRF + (rM – rRF)bi = rRF + (RPM)bi Assume that rRF = 5.

What is the market risk premium? Additional return over the risk-free rate needed to compensate investors for assuming an average amount of risk. 48 . but most estimates suggest that it ranges between 4% and 8% per year. Its size depends on the perceived risk of the stock market and investors’ degree of risk aversion. Varies from year to year.

00) = 5.5% + (5.88) = 9.0%)(1.50% rM = 5.5% + (5.00) = 10.0%)(1.50% rT-bill = 5.0)(0.15% rColl 49 .0%)(0.Calculating Required Rates of Return = 5.5% + (5.5% + 6.5% +(5.90% rUSR = 5.6% = 12.5% + (5.87) = 1.32) rHT = 5.10% = 5.0%)(-0.

5 5.1% Undervalued ( r >r) ˆ Market 10.5 10.4% 12.8 9.5 Fairly valued ( ˆ= r r) Coll.Expected vs.5 Fairly valued ( ˆr =r) USR 9. 1. Required Returns r ˆr HT 12.0 1.9 Overvalued ( ˆ< r r) T-bills 5.15 Overvalued ( ˆ< r r) 50 .

rRF = 5. bi USR T-bills 51 . . Risk.0%)bi ri (%) SML rM = 10.Illustrating the Security Market Line SML: ri = 5.5 .5 H T . .5% + (5.

bP = wHTbHT + wCollbColl bP = 0.5(1.32) + 0.225 52 .An Example: Equally-Weighted Two-Stock Portfolio Create a portfolio with 50% invested in HT and 50% invested in Collections.87) bP = 0. The beta of a portfolio is the weighted average of each of the stock’s betas.5(-0.

625% 53 . CAPM can be used to solve for expected return.5(12. using the portfolio’s beta.0%)(0.Calculating Portfolio Required Returns The required return of a portfolio is the weighted average of each of the stock’s required returns.625% Or. rP = rRF + (RPM)bP rP = 5.225) rP = 6.5% + (5.5(1.10%) + 0. rP = wHTrHT + wCollrColl rP = 0.15%) rP = 6.

Factors That Change the SML What if investors raise inflation expectations by 3%. bi 0 0.5 54 .5 10.0 1. what would happen to the SML? ri (%) ΔI = 3% SML2 13.5 5.5 1.5 Risk.5 SML1 8.

5 1. what would happen to the SML? r (%) i SML2 ΔRPM = 3% 13.5 55 .0 1. causing the market risk premium to increase by 3%.5 SML1 5. Factors That Change the SML What if investors’ risk aversion increased. bi 0 0.5 Risk.5 10.

56 . Statistical tests have problems that make verification almost impossible. other than the market risk premium.Verifying the CAPM Empirically The CAPM has not been verified completely. Some argue that there are additional risk factors. that must be considered.

the SML may not produce a correct estimate of ri. ri = rRF + (rM – rRF)bi + ??? CAPM/SML concepts are based upon expectations. 57 . Therefore.More Thoughts on the CAPM Investors seem to be concerned with both market risk and total risk. A company’s historical data may not reflect investors’ expectations about future riskiness. but betas are calculated using historical data.