Economics 122

F I N AN CIA L ECO N O MICS (I N D E X M O D E LS 2)

M . D E B UQ UE - GO N ZALES
1ST S E M ESTE R , 2 0 1 4 - 2 0 1 5

SOURCE: BODIE ET AL. (2009)

The Single-index Model and
Portfolio Construction
The use of the index model to construct optimal risky
portfolios was originally developed by Jack Treynor and
Fischer Black in 1973.
The index model offers several advantages not only in
terms of parameter estimation, but also for the analytic
simplification and organizational decentralization it
presents.
Perhaps the most important advantage of the single-
index model is the framework it provides for
macroeconomic and security analysis in the
preparation of the input list that is so critical to the
efficiency of the optimal portfolio.
Alpha and Beta Security
Analysis
The Markowitz model requires estimates of risk
premiums for each security, where the estimate of
expected return depends on both macroeconomic and
individual-firm forecasts.
◦ If many analysts perform security analysis for, say a
mutual fund company, a likely result is inconsistency in
the macro forecasts that underlie expectations of returns
across securities.
◦ Moreover, the underlying assumptions for market-index
risk and return often are not explicit in the analysis of
individual securities.
Apha and Beta Security
Analysis
One can lay down a hierarchy of the preparation of
the input list using the framework of the single
index model:
◦ First, macroeconomic analysis is used to estimate
the risk premium and risk of the market index.
◦ Second, statistical analysis is used to estimate the
beta coefficients of all securities and their
residual variances, 𝜎 𝑒 .
Apha and Beta Security
Analysis
◦ Third, the portfolio manager uses the estimates for
market-index risk premium and the beta coefficient
of a security to establish the expected return of that
security absent any contribution from security
analysis.
The market-driven expected return is conditional on
information common to all securities, not on
information gleaned from security analysis of particular
firms.
This market-driven expected return can be used as a
benchmark.
Apha and Beta Security
Analysis
◦ Finally, security-specific expected return forecasts
(specifically security alphas) are derived using various
security-valuation models.
Thus, the alpha value distils the incremental risk
premium attributable to private information developed
from security analysis.
Apha and Beta Security
Analysis
Using the index model to disentangle premiums due to
market and nonmarket factors, a portfolio manager can
be confident that macro analysts compiling estimates
of the market-index risk premium and security analysts
compiling alpha values are using consistent estimates
for the overall market.
◦ Statistical methods of estimating beta coefficients are
widely known and standardized (this portion of the input
list is not expected to differ greatly across portfolio
managers).
◦ In contrast, macro and security analysis are far less of an
exact science.
“Seeking  Alpha”
From the expected return equation, 𝐸 𝑅 = 𝛼 +
𝛽 𝐸 𝑅 , the risk premium on a security that is not
subject to security analysis would be 𝛽 𝐸 𝑅 . That is,
it  would  derive  solely  from  the  security’s  tendency  to  
follow the market index.
Any expected return beyond this benchmark risk
premium (i.e. the security alpha, 𝛼 ) would be due to
some nonmarket factor that would be uncovered
through security analysis. The end result of security
analysis is the list of alpha values.
“Seeking  Alpha”
In the context of portfolio construction, alpha is more
than just one of the components of expected return.
◦ It is the key variable that tells us whether a security is a
good buy or a bad buy.
◦ A positive alpha security provides a premium over and
above the premium it derives from its tendency to track
the market index.
◦ A security with a positive alpha is hence a bargain and
should be overweighted in the overall portfolio compared
to the passive alternative of using the market index
portfolio as the risky vehicle.
“Seeking  Alpha”
◦ A negative alpha security is overpriced and, other things
equal, its portfolio weight should be reduced.
◦ In more extreme cases, the desired portfolio weight
might even be negative. That is, a short position, if
permitted, may be desirable.
The Index Portfolio as an
Investment Asset
The process of charting the efficient frontier using the
single-index model can be pursued much like the
procedure we used previously, where we used the
Markowitz model to fund the optimal risky portfolio.
Here, however, we can benefit from the simplification
the index model offers for deriving the input list.
The single-index model offers another advantage in
terms of a simple and intuitively revealing
representation of the optimal risky portfolio.
The Index Portfolio as an
Investment Asset
Before getting into the mechanics of optimization in
this setting, we consider first the role of the index
portfolio in the optimal portfolio.
◦ A simple way to avoid inadequate diversification is to
include the market index portfolio as one of the assets of
the portfolio
This will have a beta of 1.0 (its sensitivity to itself) and
an alpha of zero (there is no nonmarket component in
its expected return).
We can designate the market index as the (𝑛 + 1)th
asset to distinguish it from the 𝑛 securities covered by
the firm.
The Index Portfolio as an
Investment Asset
We can treat the market index as a passive portfolio that
the manager would select in the absence of security
analysis (i.e., gives broad exposure without the need for
expensive security analysis).
However, if the manager is willing to engage in research,
he or she can devise an active portfolio that can be mixed
with the market index to provide an even better risk-
return tradeoff.
The Single-index Model Input
List
If the portfolio manager plans to compile a portfolio from a
list of 𝑛 actively researched firms and a passive market
index portfolio, the input list will include the following
estimates:
◦ Risk premium on the market portfolio
◦ Estimate of the standard deviation of the market portfolio
◦ 𝑛 sets of estimates of
Beta coefficients,
Stock residual variances, and
Alpha values
The alpha values for each security, together with the risk
premium of the market index and the beta of each security
will allow for the determination of the expected return on
each security.
The Optimal Risky Portfolio of
the Single-index Model
The single-index model allows us to solve for the optimal risky
portfolio directly and to gain insight into the nature of the
solution. We can easily set up the optimization process to chart
the efficient frontier in this framework along the lines of the
Markowitz model. Recall that:
◦ With the estimates of the beta and alpha coefficients plus the
risk premium of the index portfolio, we can generate the 𝑛 + 1
expected returns (risk premiums) using the equation:
𝐸 𝑅 =𝛼 +𝛽𝐸 𝑅
◦ With the estimates of the beta coefficients and residual
variances, together with the variance of the index portfolio, we
can construct the covariance matrix using:
𝐶𝑜𝑣 𝑟 , 𝑟 = 𝛽 𝛽 𝜎
𝜎 =𝛽 𝜎 +𝜎 𝑒
The Optimal Risky Portfolio of
the Single-index Model
Given a column of risk premiums and the
covariance matrix, we can already conduct the
optimization program.
As in the discussion on diversification, let us
view the alpha, beta, and residual variance of a
portfolio as the weighted averages of those
parameters across component securities.
The Optimal Risky Portfolio of
the Single-index Model
Specifically, let the weighted averages be:
𝛼 =∑ 𝑤𝛼
For the index, 𝛼 =𝛼 =0

𝛽 =∑ 𝑤𝛽
For the index, 𝛽 =𝛽 =1

𝜎 𝑒 =∑ 𝑤 𝜎 𝑒
For the index, 𝜎 𝑒 =𝜎 𝑒 =0
The objective is to maximize the Sharpe ratio of the
portfolio by using the weights 𝑤 , … , 𝑤 .
The Optimal Risky Portfolio of
the Single-index Model
With this set of portfolio weights, the expected return,
standard deviation, and Sharpe ratio of the portfolio are:
𝐸 𝑅 =𝛼 +𝐸 𝑅 𝛽 = 𝑤𝛼 +𝐸 𝑅 𝑤𝛽
/

𝜎 = 𝛽 𝜎 +𝜎 𝑒 / = 𝜎 𝑤𝛽 + 𝑤 𝜎 𝑒

𝐸 𝑅
𝑆 =
𝜎
At this point, as in the standard Markowitz procedure, we
could use an optimization program to maximize the Sharpe
ratio subject to the adding-up constraint that the portfolio
weights sum to 1.
The Optimal Risky Portfolio of
the Single-index Model
Basic tradeoff of the model: The optimal risky portfolio
(ORP) trades off the search for alpha against the cost which
is the departure from efficient diversification.
◦ Note that if you were interested only in diversification,
you would just hold the market portfolio.
◦ Security analysis, however, gives you the chance to
uncover securities with nonzero alpha and take a
differential position in those securities.
◦ But the cost of that differential position is a departure
from efficient diversification (i.e. the assumption of
unnecessary firm-specific risk).
The Optimal Risky Portfolio of
the Single-index Model
It turns out that the optimal portfolio can be derived
explicitly from the index model (making the
optimization program unnecessary).
The optimal risky portfolio turns out to be a
combination of two component portfolios:
◦ An active portfolio, denoted by A, comprised of the 𝑛
analyzed securities (this term follows from the active
security analysis involved).
◦ The passive portfolio, denoted by M, which is the
market-index portfolio, the (𝑛 + 1)th asset that we
include to aid in diversification.
The Optimal Risky Portfolio of
the Single-index Model
Assuming first that the active portfolio has a beta of 1, the
optimal weight in the active portfolio would be proportional
to the ratio .
This ratio balances the contribution of the active portfolio
(its alpha) against its contribution to the portfolio variance
(residual variance).
The analogous ratio for the index portfolio is .
Hence, the initial position in the active portfolio (if its beta
were 1) is:
𝛼
𝜎
𝑤 =
𝐸 𝑅
𝜎
The Optimal Risky Portfolio of
the Single-index Model
We now amend this position to account for the actual
beta of the active portfolio.

Recall that, 𝐶𝑜𝑟𝑟 𝑟 , 𝑟 = =

Hence, for any level of 𝜎 , the correlation between the
active and passive portfolios is greater when the beta
of the active portfolio is higher.
◦ This implies less diversification benefit from the passive
portfolio and a lower position on it when beta is higher.
◦ Correspondingly, the position in the active portfolio
increases.
The Optimal Risky Portfolio of
the Single-index Model
We now amend the initial position in the active
portfolio, 𝑤 , to account for the actual beta of the
active portfolio.
The precise modification for the position in the
active portfolio is:
∗
𝑤
𝑤 =
1+ 1−𝛽 𝑤

Notice that when 𝛽 = 1, 𝑤 ∗ = 𝑤 .

The Optimal Risky Portfolio of
the Single-index Model
Hence, two equations yield the optimal position in the
active portfolio once we know its alpha, beta, and
residual variance:
𝛼
𝜎
𝑤 =
𝐸 𝑅
𝜎

∗
𝑤
𝑤 =
1+ 1−𝛽 𝑤
The Optimal Risky Portfolio of
the Single-index Model
With 𝑤 ∗ invested in the active portfolio and
1 − 𝑤 ∗ invested in the index portfolio, we can
compute the following for the optimal risky
portfolio:
◦ Expected return
◦ Standard deviation
◦ Sharpe ratio
Sharpe Ratio
Note that the Sharpe ratio of an optimally
constructed risky portfolio will exceed that of the
index portfolio (the passive strategy), the exact
relationship being:
𝛼
𝑆 =𝑆 +
𝜎 𝑒
This last equation shows that the contribution of
the active portfolio (when held in its optimal
weight, 𝑤 ∗ ) to the Sharpe ratio of the overall risky
portfolio is determined by the ratio of its alpha
𝛼 to it residual standard deviation 𝜎 𝑒 .
Information Ratio
This important ratio is called the information ratio.
This ratio measures the extra return we can obtain
from security analysis compared to the firm-
specific risk we incur when overweighting or
underweighting securities relative to the passive
market index.
To maximize the Sharpe ratio, we must maximize
the information ratio of the active portfolio!
Information Ratio
It turns out that the information ratio of the active portfolio will
be maximized if we invest in each security in proportion to its
ratio of !
Scaling this ratio so that the total position in the active portfolio
adds up to 𝑤 ∗ , the weight in each security is
𝛼
𝜎 𝑒
𝑤∗ = 𝑤∗ 𝛼
∑
𝜎 𝑒
The contribution of each security to the information ratio of the
active portfolio depends on its own information ratio. That is,
𝛼 𝛼
=
𝜎 𝑒 𝜎 𝑒
Index Model
The model thus reveals the central role of the
information ratio in efficiently taking advantage of
security analysis.
◦ The positive contribution of a security to the
portfolio is made by its addition to the
nonmarket risk premium (its alpha).
◦ Its negative impact is to increase the portfolio
through its firm-specific risk (residual variance).
Beta of a Security
In contrast to alpha, note that the market (or
systematic) component of the risk premium,
𝛽 𝐸 𝑅 ,  is  offset  by  the  security’s  nondiversifiable
(market) risk, 𝛽 𝜎 , and both are driven by the
same beta.
◦ Any security with the same beta makes the same
balanced contribution to both risk and return.
◦ Put differently, the beta of a security is neither vice
nor virtue (i.e., it is a property that simultaneously
affects risk and risk premium).
◦ Hence, we are concerned only with the aggregate
beta of the active portfolio, rather than the beta of
each individual security.
Alpha of a Security
If  a  security’s  alpha  is  negative,  the  security  will  assume  
a short position in the optimal risky portfolio.
If short positions are prohibited, however, a negative-
alpha security would simply be taken out of the
optimization program and assigned a portfolio weight
of zero! *IN THE PROBSET

As the number of securities with nonzero alpha values

(or the number with positive alphas if short positions
are prohibited) increases, the active portfolio will itself
be better diversified and its weight in the overall risky
portfolio will increase at the expense of the passive
index portfolio.
Alpha of a Security
The (market) index portfolio is an efficient portfolio only if
all alpha values are zero.
◦ Unless security analysis reveals that a security has a nonzero
alpha, including it in the active portfolio would make the
portfolio less attractive.
◦ In  addition  to  the  security’s  systematic  risk  (which  is  
compensated for by the market risk premium, through beta),
the firm would add its firm-specific risk to portfolio variance.
◦ Hence, if all securities have zero alphas, the optimal weight in
the active portfolio will be zero and the weight in the index
portfolio will be one.
But when security analysis uncovers securities with
nonmarket risk premiums (nonzero alphas), the index
portfolio is no longer efficient.
Summary of Optimization
Procedure
Once security analysis is complete and the index-model estimates of
security and market index parameters are established, the optimal risky
portfolio can be formed using the following steps:
1. Compute the initial position of each security in the active portfolio:
𝛼
𝑤 =
𝜎 𝑒
2. Scale those initial positions to force portfolio weights to sum to 1 by
dividing their sum:
𝑤
𝑤 =
∑ 𝑤
3. Compute the alpha of the active portfolio:

𝛼 = 𝑤 𝛼
Summary of Optimization
Procedure
4. Compute the residual variance of the active portfolio:

𝜎 𝑒 = 𝑤 𝜎 𝑒

5. Compute the initial position in the active portfolio:

𝛼
𝜎
𝑤 =
𝐸 𝑅
𝜎
6. Compute the beta of the active portfolio: 𝛽 = ∑ 𝑤 𝛽.
7. Adjust the initial position in the active portfolio:
𝑤
𝑤∗ =
1+ 1−𝛽 𝑤
*The optimal risky portfolio now has weights: 𝑤 ∗ = 1 − 𝑤 ∗ ; 𝑤 ∗ = 𝑤 ∗ 𝑤 .
Summary of Optimization
Procedure
9. Calculate the risk premium of the optimal risky portfolio from
the risk premium of the index portfolio and the alpha of the
active portfolio:
𝐸 𝑅 = 𝑤 ∗𝛼 + 𝑤 ∗ + 𝑤 ∗𝛽 𝐸 𝑅
(Note that 𝛽 = 1.)
10. Compute the variance of the optimal risky portfolio from the
variance of the index portfolio and the residual variance of the
active portfolio:
𝜎 = 𝑤 ∗ + 𝑤 ∗𝛽 𝜎 + 𝑤 ∗𝜎 𝑒
(Note that 𝜎 𝑒 = 0.)
Portfolio Management with
the Index Model
The index model seems to be the preferred one for
practical portfolio management.
But, is it really inferior to the Markowitz (full-
covariance) model?
Important to answer this if making a switch.
Is the Index Model Inferior to
the Markowitz Model?
This question is related to the more general question of
the value of parsimonious models.
The Markowitz model allows far more flexibility in
modelling the asset covariance structure compared to
the single-index model.
However,  the  advantage  may  be  illusory  if  we  can’t  
estimate those covariances with any degree of
confidence.
Using the full-covariance matrix invokes estimation risk
of thousands of terms.
Is the Index Model Inferior to
the Markowitz Model?
Hence, even if the full Markowitz model would be
better in principle or potentially, it is very possible
that cumulative effect of so many estimation errors
will result in a portfolio that is actually inferior to
that derived from a single-index model.
Thus, the single-index model has a practical
advantage, plus its other advantage of helping
decentralize macro and security analysis.
Industry Version of the Index
Model
To the extent that it is approximately valid, the
index model provides a convenient benchmark for
security analysis.
Thus, it has attracted the attention of practitioners.
A portfolio manager with no special information or
insight  about  a  security  will  take  the  security’s  
alpha to be zero and will forecast a risk premium
for the security equal to 𝛽 𝑅 .
Industry Version of the Index
Model
In terms of total returns, one would expect:
𝐸 𝑟 =𝑟 +𝛽 𝐸 𝑅 −𝑟

A portfolio manager who has a forecast for the market

index, 𝐸 𝑅 , and observes the risk-free rate, 𝑟 , can
use the model to determine the benchmark expected
return for any stock, 𝐸 𝑟 .
The beta coefficient, 𝛽 , the market risk, 𝜎 , and the
firm-specific risk, 𝜎 (𝑒),meanwhile, can be estimated
from historical SCLs (i.e., from regressions of security
excess returns on market index excess returns).
Industry Version of the Index
Model
 The industry variant (e.g., Merrill Lynch) of the index model is:
𝑟 = 𝑎 + 𝑏𝑟 + 𝑒 ∗
This is instead of:
𝑟−𝑟 =𝛼+𝛽 𝑟 −𝑟 +𝑒
𝑟 = 𝑟 + 𝛼 + 𝛽𝑟 − 𝛽𝑟 + 𝑒 = 𝛼 + 1 − 𝛽 𝑟 + 𝛽𝑟 + 𝑒
o Note that the intercept that what the industry calls alpha is
really an estimate 𝛼 + 1 − 𝛽 𝑟 . Note, in addition, that
the intercepts are equal when 𝛽 = 1.
o If 𝑟 is constant over the sample period, both equations have
the same independent variable, 𝑟 , and residual, 𝑒, and
therefore, the coefficient will be the same in the two
regressions (𝑏 = 𝛽).
Industry Version of the Index
Model
o The apparent justification for this procedure is that, on a
monthly basis, 1 − 𝛽 𝑟 is small and apt to swamped by the
volatility of actual stock returns.
 Another way the industry procedure departs from the
index model is its use of percentage changes in price
instead of total rates of return (i.e., it ignores the dividend
component of stock returns).
 Beta is also adjusted towards 1, the motivation being that
beta coefficients move toward 1 over time on average:
𝑀𝑒𝑟𝑟𝑖𝑙𝑙 𝐿𝑦𝑛𝑐ℎ 𝑎𝑑𝑗𝑢𝑠𝑡𝑒𝑑 𝑏𝑒𝑡𝑎 = 2/3 𝑠𝑎𝑚𝑝𝑙𝑒 𝑏𝑒𝑡𝑎 + (1/3)1.
Industry Version of the Index
Model
Meanwhile, note that the alpha estimates are ex-
post (after-the-fact) measures.
◦ They do not mean that anyone could have forecast these
alpha values ex ante (before the fact).
◦ The name of the game of security analysis is to forecast
alpha values ahead of time.
◦ A well-constructed portfolio that includes long positions
in future positive-alpha stocks and short positions in
negative-alpha stocks will outperform the market index.
◦ “Well-constructed”  here  means  that  the  portfolio  has  to  
balance concentration on high-alpha stocks with the need
for risk-reducing diversification.
Industry Version of the Index
Model
Since betas estimated from past data may not be
the best estimates of future betas, one might want
a forecasting model for beta.
One simple approach would be to collect data on
beta in different periods and then estimate a
regression equation
𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑏𝑒𝑡𝑎 = 𝑎 + 𝑏(𝑃𝑎𝑠𝑡 𝑏𝑒𝑡𝑎)
Given estimates of 𝑎 and 𝑏, one could then forecast
future betas using the rule
𝐹𝑜𝑟𝑒𝑐𝑎𝑠𝑡 𝑏𝑒𝑡𝑎 = 𝑎 + 𝑏(𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑏𝑒𝑡𝑎)
Industry Version of the Index
Model
An expanded version might be
𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑏𝑒𝑡𝑎
= 𝑎 + 𝑏 (𝑃𝑎𝑠𝑡 𝑏𝑒𝑡𝑎) + 𝑏 (𝐹𝑖𝑟𝑚 𝑠𝑖𝑧𝑒) + 𝑏 (𝐷𝑒𝑏𝑡 𝑟𝑎𝑡𝑖𝑜)
Such an approach was followed by Rosenberg and Guy
(1976) who found the following variables helpful in
predicting beta:
◦ Variance of earnings
◦ Variance of cash flow
◦ Growth in earnings per share
◦ Market capitalization (firm size)
◦ Dividend yield
◦ Debt-to-asset ratio
◦ Industry group
Industry Version of the Index
Model
Suppose a portfolio manager believes he or she has
identified an underpriced portfolio, from
estimating the index model for thus portfolio in
excess return form and getting the following
estimates:
𝑅 = .04 + 1.4𝑅 + 𝑒

𝑃 has an alpha of 4% and a beta of 1.4.

Portfolio seems underpriced on a relative basis.
Concept Check
The data below describe a 3-stock financial market that satisfies the
single-index model.
Stock Capitalization Beta Mean excess Standard
return deviation
A $3,000 1.0 10% 40%
B $1940 0.2 2% 30%
C $1,360 1.7 17% 50%
The standard deviation of the market index portfolio is 25%.
◦ What is the mean excess return of the index portfolio?
◦ What is the covariance between stock A and stock B?
◦ What is the covariance between stock B and the index?
◦ Break down the variance of stock B into its systematic and firm-
specific components.
Concept Check
Suppose that the index model for the excess returns of stocks A
and B is estimated with the following results:
𝑅 = 1.0% + .9𝑅 + 𝑒
𝑅 = −2.0% + 1.1𝑅 + 𝑒
𝜎 = 20%
𝜎(𝑒 ) = 30%
𝜎(𝑒 ) = 10%
Find the standard deviation of each stock and the covariance
between them.
Now, suppose we form an equally weighted portfolio of stocks A
and B. What will be the nonsystematic standard deviation of that
portfolio?