Chapter 13.
Risk & Return in
Asset Pricing Models
• Portfolio Theory
• Managing Risk
• Asset Pricing Models
I. Portfolio Theory
• how does investor decide among
group of assets?
• assume: investors are risk averse
• additional compensation for risk
• tradeoff between risk and expected
return
goal
• efficient or optimal portfolio
• for a given risk, maximize exp.
return
• OR
• for a given exp. return, minimize
the risk
tools
• measure risk, return
• quantify risk/return tradeoff
Measuring Return
change in asset value + income
return = R =
initial value
• R is ex post
• based on past data, and is known
• R is typically annualized
example 1
• Tbill, 1 month holding period
• buy for $9488, sell for $9528
• 1 month R:
9528 - 9488
= .0042 = .42%
9488
• annualized R:
(1.0042)12 - 1 = .052 = 5.2%
example 2
• 100 shares IBM, 9 months
• buy for $62, sell for $101.50
• $.80 dividends
• 9 month R:
101.50 - 62 + .80
= .65 =65%
62
• annualized R:
(1.65)12/9 - 1 = .95 = 95%
Expected Return
• measuring likely future return
• based on probability distribution
• random variable
E(R) = SUM(Ri x Prob(Ri))
example 1
R Prob(R)
10% .2
5% .4
-5% .4
E(R) = (.2)10% + (.4)5% + (.4)(-5%)
= 2%
example 2
R Prob(R)
1% .3
2% .4
3% .3
E(R) = (.3)1% + (.4)2% + (.3)(3%)
= 2%
examples 1 & 2
• same expected return
• but not same return structure
• returns in example 1 are more
variable
Risk
• measure likely fluctuation in return
• how much will R vary from E(R)
• how likely is actual R to vary from
E(R)
• measured by
• variance (
• standard deviation
= SUM[(Ri - E(R))2 x Prob(Ri)]
SQRT(
example 1
= (.2)(10%-2%)2
+ (.4)(5%-2%)2
+ (.4)(-5%-2%)2
= .0039
= 6.24%
example 2
= (.3)(1%-2%)2
+ (.4)(2%-2%)2
+ (.3)(3%-2%)2
= .00006
= .77%
• same expected return
• but example 2 has a lower risk
• preferred by risk averse investors
• variance works best with symmetric
distributions
prob(R) prob(R)
R R
E(R) E(R)
symmetric asymmetric
II. Managing risk
• Diversification
• holding a group of assets
• lower risk w/out lowering E(R)
• Why?
• individual assets do not have same
return pattern
• combining assets reduces overall
return variation
two types of risk
• unsystematic risk
• specific to a firm
• can be eliminated through
diversification
• examples:
-- Safeway and a strike
-- Microsoft and antitrust cases
• systematic risk
• market risk
• cannot be eliminated through
diversification
• due to factors affecting all assets
-- energy prices, interest rates,
inflation, business cycles
example
• choose stocks from NYSE listings
• go from 1 stock to 20 stocks
• reduce risk by 40-50%
unsystematic
risk
total
risk
systematic
risk
# assets
measuring relative risk
• if some risk is diversifiable,
• then is not the best measure of
risk
• σ is an absolute measure of risk
• need a measure just for the
systematic component
Beta,
• variation in asset/portfolio return
relative to return of market portfolio
• mkt. portfolio = mkt. index
-- S&P 500 or NYSE index
% change in asset return
=
% change in market return
interpreting
• if
• asset is risk free
• if
• asset return = market return
• if
• asset is riskier than market index
• asset is less risky than market index
Sample betas
Amazon 2.23
Anheuser Busch -.107
Microsoft 1.62
Ford 1.31
General Electric 1.10
Wal Mart .80
(monthly returns, 5 years back)
measuring
• estimated by regression
• data on returns of assets
• data on returns of market index
• estimate
R R m
problems
• what length for return interval?
• weekly? monthly? annually?
• choice of market index?
• NYSE, S&P 500
• survivor bias
• # of observations (how far back?)
• 5 years?
• 50 years?
• time period?
• 1970-1980?
• 1990-2000?
III. Asset Pricing Models
• CAPM
• Capital Asset Pricing Model
• 1964, Sharpe, Linter
• quantifies the risk/return tradeoff
assume
• investors choose risky and risk-free
asset
• no transactions costs, taxes
• same expectations, time horizon
• risk averse investors
implication
• expected return is a function of
• beta
• risk free return
• market return
E( R ) R f [ E( R m ) R f ]
or
E( R ) R f [ E( R m ) R f ]
where
E( R ) R f is the portfolio risk premium
E( R m ) R f is the market risk premium
so if
E( R ) R f > E( R m ) R f
E( R ) > E( R m )
• portfolio exp. return is larger than
exp. market return
• riskier portfolio has larger exp.
return
so if
E( R ) R f < E( R m ) R f
E( R ) < E( R m )
• portfolio exp. return is smaller than
exp. market return
• less risky portfolio has smaller exp.
return
so if
E( R ) R f = E( R m ) R f
E( R ) = E( R m )
• portfolio exp. return is same than
exp. market return
• equal risk portfolio means equal exp.
return
so if
E( R ) R f =0
E( R ) = Rf
• portfolio exp. return is equal to risk
free return
example
• Rm = 10%, Rf = 3%, = 2.5
E( R ) R f [ E( R m ) R f ]
E( R ) 3% 2.5[10% 3%]
E( R ) 3% 17.5%
E( R ) 20.5%
• CAPM tells us size of risk/return
tradeoff
• CAPM tells use the price of risk
Testing the CAPM
• CAPM overpredicts returns
• return under CAPM > actual return
• relationship between β and return?
• some studies it is positive
• some recent studies argue no
relationship (1992 Fama & French)
• other factors important in
determining returns
• January effect
• firm size effect
• day-of-the-week effect
• ratio of book value to market value
problems w/ testing CAPM
• Roll critique (1977)
• CAPM not testable
• do not observe E(R), only R
• do not observe true Rm
• do not observe true Rf
• results are sensitive to the sample
period
APT
• Arbitrage Pricing Theory
• 1976, Ross
• assume:
• several factors affect E(R)
• does not specify factors
• implications
• E(R) is a function of several
factors, F
each with its own
E( R ) R f 1F1 2 F2 3F3 .... N FN
APT vs. CAPM
• APT is more general
• many factors
• unspecified factors
• CAPM is a special case of the APT
• 1 factor
• factor is market risk premium
testing the APT
• how many factors?
• what are the factors?
• 1980 Chen, Roll, and Ross
• industrial production
• inflation
• yield curve slope
• other yield spreads
summary
• known risk/return tradeoff
• how to measure risk?
• how to price risk?
• neither CAPM or APT are perfect or
free of testing problems
• both have shown value in asset
pricing
