Part I

Portfolio Theory
1. Introduction

Before discussing the portfolio, it is important to make sure the following concepts
are understood:
E¢cient Portfolios: That is when investors seek to maximize the expected
return from their investment given some level of risk they are willing to accept.
Risk Aversion: Individuals according to those theories are assumed to be
risk averse: is one who, when faced with two investments with the same expected
return but two di¤erent risks, will prefer the one with the lower risk.
Risky assets: Those are the ones which the return that will be realized in
the future is uncertain. Corporate bonds are riskier than public bonds, because
of the possibility of default, in‡ation and so on.
Assets in which the return that will be realized in the future is known with
certainty today are referred to as risk-free assets or riskless assets.

1.1. Returns And Variances

In order to understand the relationship between risk and return, we must be able
to measure both risk and return for an investment.

1.1.1. Dollar returns

The return on an investment has two components: the income component; and
the capital gain (or capital loss) component. For a common stock, your return
comes in two components: dividends ( periodic income) and capital gains.
Suppose you buy 100 shares of ABC common stock for $25 per share. Over
one year, the ABC company pays a cash dividend of $2 per share and the value of
the stock rises to $30. At the end of the year you have received ($2 £100) = $200
in dividends. In addition you own 100 shares of stock worth $30 per share; The
capital gain is [($30-$25)£100] = $500:
Then:
Total Dollar Return = Dividend Income + Capital Gain (or loss) = $200
+ $500 = $700.
The total value of your position is $3200, total investment is $2500, and total
return is $700.
1.1.2. Percentage Return
Percentage returns are easy to work with. In the previous example, you received
$700 on a $2500 investment, so your total percentage return is:
[$200 + ($3000 - $2500)]/$2500 = .28 = 28%
That can be expressed in terms of per share basis, that is:[$2 + ($30 -
$25)]/$25 = .28 = 28%
The 28% total return is comprised of the dividend yield plus the percentage
capital gain (or capital gains yield). The dividend yield is given by:
Dt+1 =Pt = $2=$25 = 0:08 = 8%
Where Pt is the price of the stock at the beginning of the year, and Dt+1 is
the dividend paid during the year. The capital gains yield is given by:
(Pt+1¡ Pt)=Pt = ($30 ¡ $25)=$25 = :20 = 20%

1.1.3. Average Returns

The average return (or mean) is simply the sum of all returns divided by the
number of observations:
Mean = R = E(R) = [R1 + R2 + :::::::::::::: + RT ]=T
Where T is the number of time periods.

1.1.4. The variability of returns

Since riskiness a¤ects return, we must be able to measure the degree of risk
associated with an investment in order to understand the relationship between risk
and return. The most common measures are the Variance and the Standard
Deviation. In …nancial context, we use the variance to measure the variability
of returns from the average return; the greater the deviation from the average,
the more variable the rate of return and the higher the level of risk.
Take an example, say an investment returned 8%, 14%, -4%, and 6% over the
last 4 years. The average return is:
R = E(R) = [R1 +R2+R3 +R4T ]=4 = (0:08+0:14¡0:04+0:06)=4 = 0:06 = 6%
Rate of Return Average Return Deviation from Average(R-R) (R-R)2
0.08 0.06 0.02 0.004
0.14 0.06 0.08 0.0064
-0.04 0.06 -0.10 0.0100
0.06 0.06 0.00 0.000
The calculation of the variance is based on :
Var (R) = ¾2 = (1=(T ¡ 1)) [(R1-R)2+(R2-R)2+(R3-R)2 + :::::::::::::(RT -R)2 ]
The standard deviation is the square root of the variance.

2
2. Expected Portfolio Return

2.1. Expected Returns and Variances

The previous discussion about the calculation of average returns are based on
historical data. But usually the focus is on rates of return that might occur in
the future. A useful way to quantify the uncertainty about the portfolio return is
to specify the probability associated with each of the possible future returns. So
each outcome has a probability. The average return in this case will be referred to
as the expected return. An expected value in an average, or mean, of possible
future outcomes. To illustrate consider the following example:

State of the Economy Probability Possible Rates of Return

1 0.20 0.10
2 0.20 0.04
3 0.20 -0.09
4 0.20 0.20
5 0.20 0.05

The table provide forecasts of possible return of a stock of ABC during the
coming year. The forecasts are based on …ve equally-likely possible states of the
economy. A state of the economy might be a severe recession, or a depression,
or moderate growth in the economy. (Note it is by coincidence that Pri the
probability of state i is the same, but in general, di¤erent states will be attached
di¤erent probabilities. In case where we have equal weights for Pr’s, E (R) is
exactly calculated as in the case of the historical data).
Expected return is calculated as a weighted average of the possible returns.
Each possible return is multiplied (weighted) by its probability of occurrence.
Then the expected return will be:
E(R)= R = (Pr1 £ R1)+(Pr2 £ R2 ) + (Pr3 £ R3 ):::::::::::::::::::(PrT £ RT )
Where T is the number of possible states of the economy, Pr’s are the proba-
bilities of the respective states, and R’s are the possible rates of return. For the
above example, E(R) = 0.06
To calculate the variance of the data representing future possible rates of
return, we have to determine the squared deviations from the expected return.
Each possible squared deviation is multiplied by its probability of occurrence.
These are added; the sum is the variance. The standard variance is simply the
square root of the variance, or to say:
Var (R) = ¾2 = [Pr1 £(R1 ¡R)2]+[Pr2 £(R2 ¡R)2]+::::::::+[PrT £(RT ¡R)2]

3
 Figure 2.1: Return distribution

State Probability (Pr) Rate of Return (R ¡ R) (R ¡ R)2

1 0.20 0.10 0.04 0.016
2 0.20 0.04 -0.02 0.0004
3 0.20 -0.09 -0.15 0.0225
4 0.20 0.20 0.14 0.0196
5 0.20 0.05 -0.01 0.0001

So the Variance can be calculated, for the

p above example, we get Var(R) =
0.00884, and then the Standard deviation is 0:00884 = 0:09402:

2.2. Return and Return Volatility

To explain the relationship between expected return and return volatility, and
to show why it is important to consider more than just expected returns when
selecting assets for inclusion in a portfolio.
Expected Return The return expected on an investment (an asset or a port-
folio) based on a probability distribution, taking into account all possible return
scenarios.
Return Volatility Represents the variability or uncertainty of an assets return;
it is measured by a value called standard deviation.
Essentially, return volatility (or standard deviation) tells us how much an as-
sets actual return is likely to deviate above or below its expected return. Consider
the graph above, where each segment represents one standard deviation:
”Red” Zone = The assets actual return has approximately 68% probability of
falling within this zone (i.e. within 1 standard deviation of the assets expected
return).

4
 ”Red+Green” Zone = The assets actual return has approximately 95% prob-
ability of falling within this zone (i.e. within 2 standard deviations of the assets
expected return).
”Red+Green+Blue” Zone = The assets actual return has approximately 99%
probability of falling within this zone (i.e. within 3 standard deviations of the
assets expected return).
Asset A has an expected return of 22%, and a return volatility (standard
deviation) of 15%. With this information, we can infer the following:
Asset A has a 68% probability of achieving an actual return between 7% and
37% (i.e. one standard deviation below and above expected return On the graph,
this range is represented the red area).
Asset A has a 95% probability of achieving an actual return between -8%
and 52% (i.e. two standard deviations below and above expected return On the
graph, this range is represented the red+green area).
Asset A has a 99% probability of achieving an actual return between -23%
and 67% (i.e. three standard deviations below and above expected return On the
graph, this range is represented the red+green+blue area).
The important point here is that return volatility (standard deviation) can
have a tremendous impact on actual return. The oft-quoted cliché says, High risk,
high reward, but thats only half of the story. With an understanding of return
volatility, it’s clear that the cliché fails to mention that high risk also means the
potential for great loss.
When considering an asset for inclusion in a portfolio, it’s natural to look to
the upside associated with expected return. However, we also need to factor in
return volatility and decide whether or not we can live with the potential downside
associated it represents.

2.3. Portfolios
An investor’s portfolio of assets is the combination of assets the investor owns.
A portfolio consists of several assets. Therefore, an investor is more concerned
about the characteristics of the portfolio than the characteristics of the individual
assets which comprise the portfolio.
Portfolio Weights: the respective percentages of portfolio’s total value in-
vested in each of the assets in the portfolio are referred to as the portfolio
weights.
Example: An investor’s portfolio has a total value of $1000 and is com-
prised of three assets. The values of assets A, B and C are $450, $250 and $300
respectively. What are the portfolio weights for this portfolio.

5
 12000

10000

8000

6000

4000

2000
1988 1990 1992 1994 1996 1998 2000 2002

T S E 3 0 0

6
4

-2
-4

-6

-8

-10
1988 1990 1992 1994 1996 1998 2000 2002

R T S E 3 0 0

Figure 2.2: TSE 300 in levels and returns. Sample period is from 11/16/1987 to
11/14/2002

The percentage of the total value of the portfolio invested in asset A is

($450/$1000) =.45 = 45%. Similarly, the percentages for assets B and C are
($250/$1000) = 25% and 30%. Therefore, the weights are 45%, 25% and 30%.

2.3.1. Portfolio Expected Returns

Consider again the following example from before, except we have a new asset,
let us call the two assets, X and Y.

State Pr Rate of Return on X Rate of Return on Y

1 0.20 0.10 0.03
2 0.20 0.04 -0.08
3 0.20 -0.09 0.07
4 0.20 0.20 0.12
5 0.20 0.05 0.21

Suppose that an investor invests 60% of his money in Asset X and 40% in
Asset Y; that is the portfolio weights are 60% and 40%, respectively. For each
of the …ve possible states of he economy indicated in the example, the investor’s

6
return on his investment will be equal to a weighted average of the returns in
that state for the respective stocks. For example, in state 1, the portfolio return:
RP = (0:60 £0:10)+(0:40£0:03) = 0:072 = 7:2%: That says, if state 1 occurs,
then the rate of return for the portfolio will be 7.2%. The table below makes all
the calculation to derive at the expected return of the portfolio E(RP )

State Probability (Pr) RX RY RP (Pr £ RP )

1 0.20 0.10 0.03 0.072 0.0144
2 0.20 0.04 -0.08 -0.008 -0.0016
3 0.20 -0.09 0.07 -0.026 -0.0052
4 0.20 0.20 0.12 0.168 0.0336
5 0.20 0.05 0.21 0.114 0.0228
Total 1.00 0.30 0.35 0.320 0.0640

So the expected return for the portfolio is 0.064 = 6.4%.

A simpler approach to compute the expected return for a portfolio is based
on the fact that the expected return for the portfolio is a weighted average of the
expected returns of the securities which comprise the portfolio; the weights are
the weights of portfolio. The general formula would be:
E(RP ) = [x1 £ E(R1 )] + [x2 £E(R2)] + [x3 £ E(R3)] + ::::::::::: + [xn £ E(Rn )]
where n is the number of assets in the portfolio, x1; x2; x3; and xn are the port-
folio weights for the assets and E(R1); E(R2 ); E(R3) and E(Rn ) are the expected
returns for the assets. For the above example, we have:
E(RP ) = [x1 £ E(R1)] + [x2 £ E(R2 )] = (0:60 £ 0:06) + (0:40 £ 0:07) = 6:4%

2.3.2. Portfolio Variance

As opposed to the expected return for a portfolio, the variance and standard
deviation (of a portfolio) are not equal to weighted averages of the corresponding
characteristics of the individual securities. The actual standard deviation for
a portfolio is virtually always less than the weighted average of the standard
deviations or the securities in the portfolio. Let us …nd the variance and standard
deviation for the portfolio we have above:

7
 State Pr RX RY RP (RP ¡ E(R P ))2 [Pr £ (RP ¡ E(RP ))] 2
1 0.20 0.10 0.03 0.072 0.000064 0.0000128
2 0.20 0.04 -0.08 -0.008 0.005184 0.0010368
3 0.20 -0.09 0.07 -0.026 0.008100 0.0016200
4 0.20 0.20 0.12 0.168 0.010816 0.0021632
5 0.20 0.05 0.21 0.114 0.002500 0.0005000
Total 1.00 0.30 0.35 0.320 0.026664 0.0053328

So the variance of the portfolio is 0.0053328 and the standard deviation

would be 0.07303.

2.3.3. Portfolio standard deviation and Diversi…cation

That means, constructing a portfolio in such a way as to reduce portfolio risk
without sacri…cing return. Observations from the capital markets indicate that
deviation of returns for the stock alone can be greater than that of the portfolio,
and the stock average return is less than the portfolio return.
The answer to this question lies in the fact that not all risk is relevant. Much
of the total risk is diversi…able. That is, if that investment had been combined
with other securities, a portion of the variation in its returns could have been
smoothed or canceled by complementary variations in the other securities.
It is very important to note here, that in order to reduce risk, diversi…cation
of combining securities should involve less than perfectly correlated ones. Diver-
si…cation will not a¤ect the portfolio return, but it will reduce the variability of
return. The less the correlation among security returns, the greater the impact
of diversi…cation on reducing variability. Empirically speaking, securities are al-
ways positively correlated, they are a¤ected always by business cycles and interest
rates: elimination of risk is impossible in this case.
Systematic risk, which is related to market conditions, will always exist and
a¤ect the portfolio variability.
In general, combining securities into portfolios reduces the level of risk. The
actual standard deviation for a portfolio is virtually always less than the weighted
average of the standard deviations for the securities in the portfolio. This is be-
cause the correlation between the returns of the securities in the portfolio. Math-
ematically, it can be shown that the portfolio variance for a two stock portfolio,
stock A and stock B is:

¾2P = XA
2 2 2 2
¾A + XB ¾B + 2XAXB CorrA;B¾A¾B

8
 where CorrA;B = Correlation betwee the returns of A and B and CorrA;B ¾A¾B
= Covariance between the returns of A and B. As long as the correlation is less
than 1.0, the standard deviation of a portfolio of two securites is less than the
weighted average of the standard deviation of the individual securities.
Combining securities or assets whose Corr is less than +1.0 o¤ers the investor
the opportunity to reduce portfolio risk as measured by the portfolio standard
deviation. This process is commonly known as diversi…cation. As investors
diversify their portfolios by adding securities that have correlation coe¢cients less
than 1.0, and continue to adjust the respective weights in these portfolios, the
many possible combinations can be respresented in a feasible set of the e¢cient
set of portfolio combinations.
To illustrate through an example, take a portfolio of two assets A, and B.
Suppose that each asset makes up 50% of the total value of the portfolio. The
expected return and standard deviation for Asset A are 10% and 0.30, and for
Asset B, 25% and 0.60, respectively.
Then, the expected return and variance on the portfolio are:
E(RP ) = 0.5(RA) + 0:5(RB ) = 0:5(0:1) + 0:5(0:25) = 0:175:
The standard deviation is derived from:
¾P = [XA 2 ¾ 2 + X 2 ¾2 + 2X X Corr 1=2 = [0:52 (0:3)2 + 0:52 (0:6)2 +
A B B A B A;B¾ A¾ B ]
1=2
2(0:5)(0:5)CorrA;B (0:3)(0:6)] =
=[0.1125 + 0.09CorrA;B]1=2
Now taking di¤erent values for CorrA;B = +1; 0; ¡1 we get the risk of returns
below:

Risk of returns with

Case ½AB;BC Expected return AB Risk of Return AB
replacement asset C
1 +1.0 17.5% 45% 50%
2 0.0 17.5% 33% 36%
3 -1.0 17.5% 15% 10%

The table shows the “magic” of Tobin-Markowitz diversi…cation. The risk of

return of the portfolio is shown to decline as the degree of correlation between
the assets is reduced, and the expected return on the portfolio is not a¤ected.
There is another lesson we can learn from diversi…cation. One can increase
the expected return of a portfolio, and at the same time reduce the risk of the
expected return, by substituting a riskier asset for a less risky asset - if the
expected return of the replacement asset is correlated much less with the other
assets than was the case of the asset it replaced.

9
 Suppose that, in the above example, asset A is replaced by asset C with an
expected rate of return of 20% and a standard deviation of 0.40. The expected
return of the portfolio is now 22.5% rather than 17.5%, and the risk of returns,
(standard deviatioin ) becomes [0.13 + 0.12 CorrC;B ]1=2 : We can see from the
table that in case 3, risk of portfolio has been reduced.
When there are only two assets in a portfolio, we could get the variance of
this portfolio in the following manner:

V ar(RP ) = wi2V ar(Ri) + wj2V ar(Rj ) + 2wi wj std(Ri )std(Rj )cor(Ri ; Rj )

wi and wj are the percentage of he portfolio’s funds invested in asset i, and

asset j.
i, and j are the only assets in the portfolio. The above equation states that the
variance of the portfolio is the sum of the weighted variances of the two assets plus
the weighted correlation between the two assets. So the correlation between the
two assets a¤ect the variance of the portfolio. The lower the correlation between
the two assets, the lower the portfolio’s variance, and vice versa. (perfect positive
correlation, when it is 1, and -1 means negative perfect correlation). To expand
the above equation to n assets in portfolio, say we have three assets, i; j and k,
we get:
Var (RP ) = w2i Var (Ri) + w2j Var (Rj ) +w2k V ar(Rk )+ 2wiwj std(Ri ) std(Rj )
cor(Ri;Rj )+2wi wk std(Ri) std(Rk ) cor(Ri;Rk )+2wj wk std(Rj) std(Rk ) cor(Rj;Rk )
again, wk is the percentage of the portfolio’s funds invested in asset k.
To explain the concept of asset correlation, and to show why it is important
to understand the implications of correlation when considering assets for possible
inclusion in your portfolio.

2.3.4. Adding Securities to eliminate risk

In this part, we further illustrate the power of diversi…cation. First note that the
general formula for the variance of a portfolio is as follows:

N
X N X
X N
V ar(RP ) = w2i var(Ri ) + wi wj Cov(Ri; Rj ) for i 6= j (2.1)
i=1 i=1 j=1

The formula says that the variance of a portfolio is a weighted average of the
variances of the individual securities plus the covariance between each security
and every other security in the portfolio. If securities have zero correlation, and
hence zero covariance, the second term goes to zero, and the expression reduces
to:

10
 N
X
V ar(RP ) = wi2var(Ri ) (2.2)
i=1
Assume for purposes of illustration that only securities with zero covariance
are available, that each security has the same variance, and that equal amounts
are invested in each security,

N
X
V ar(RP ) = (1=N )2 var(Ri) = N(1=N)2var(Ri ) = 1=N var(Ri) (2.3)
i=1

and

p
var(Ri ) S tan dard deviationR i
s tan dard deviationRP = p = p (2.4)
N N
Using this formula and assuming that the individual securities have standard
deviations of (0.4 percent), the data in the following table show how the risk
(standard deviation) of the portfolio declines as zero-correlated securities with
identical standard deviations are added to the portfolio. Risk is reduced to less
than 10 percent of that of a single-stock portfolio when 128 securities are in the
portfolio and to less than 2 percent of the original risk when 510 are included. As
the number of securities added becomes larger and larger (approaches in…nity, in
technical terms), the standard deviation of the portfolio approaches zero.

Portfolio risk and number of securities (zero correlation)

Number of securities 1 2 8 16 32 128 510
Standard deviation of portfolio return 40.0 28.3 14.1 10.0 7.1 3.5 1.8

The principle here is that if there are su¢cient numbers of securities with zero
correlation (zero covariance), the portfolio analyst can make the portfolio risk
arbitrarily small. This is the basis for insurance, which explains why insurance
companies attempt to write many individual policies and spread their coverage so
as to minimize overall risk. It also has direct relevance in providing a benchmark
for assessing the extent to which diversi…cation can be e¤ective in reducing risk
for equity investors.

Example 2.1. Correlation

11
 Asset correlation measures the extent to which the returns on two assets
move together (i.e the extent to which those returns behave similarly in response
to market events or stimuli).How it works
Asset correlation ranges from a maximum of +1.00 to a minimum of -1.00. If
two assets have a perfect positive correlation (+1.00), their returns will tend to
move simultaneously in the same direction. With a perfect negative correlation
(-1.00), their returns will tend to move simultaneously in opposite directions. A
correlation of 0 indicates that there is no relationship at all between the price
movements of two assets.
Since few asset pairs will come anywhere close to perfect positive or negative
correlation, the following rules of thumb can be helpful:
High Correlation: Asset correlation greater than 0.75; implies that the two
assets respond very similarly to the market and that their prices will very often
move in the same direction.
Moderate Correlation: Asset correlation between 0.25 and 0.75; implies that
the two assets respond in somewhat similar ways to the market and that their
prices will move more or less in the same direction, depending on how strong the
correlation.
Low Correlation: Asset correlation between 0.00 and 0.25; implies that the
two assets respond fairly independently to the market and that their prices also
tend to move independently of one another.
Negative Correlation: Asset correlation below 0.00; implies that the two assets
respond fairly di¤erently to the market and that their prices will tend to move in
opposite directions.
An individual investor maintains a portfolio that includes shares of the fol-
lowing assets: Intel (INTC), AT&T (T), Walmart (WMT), Vanguard Wellington
Income Fund (VWELX), Janus Fund (JANSX), and Vanguard Total Interna-
tional Stock Index Fund (VGTSX). The matrix below shows how those assets
correlate to one another:

Correlation INTC T WMT VWELX JANSX VGTSX

INTC 1.00 0.53 0.21 0.35 0.42 0.25
T 0.53 1.00 0.29 0.27 0.42 0.12
WMT 0.21 0.29 1.00 0.24 0.31 0.12
VWELX 0.35 0.27 0.24 1.00 -0.08 0.84
JANSX 0.42 0.42 0.31 -0.08 1.00 0.72
VGTSX 0.25 0.12 0.12 0.84 0.72 1.00

In viewing this matrix, we can make the following observations:

12
 To illustrate the extent to which one asset can behave di¤erently from other
assets, the Vanguard Wellington Income Fund serves as a good example: This
fund has moderate or low/moderate correlation with the Intel, AT&T and Wal-
mart stocks, a negative correlation with the Janus Fund, and a high correlation
with the Vanguard Total International Stock Index Fund. While not precise, it is
generally true that the closer your median or average correlation is to +1.00, the
less diversi…ed your portfolio is likely to be. In this portfolio, VGTSX’s high cor-
relations with VWELX and JANSX suggest that this fund may be contributing
less to portfolio diversi…cation than the investor thinks.
Correlation is important because it serves as a checking mechanism on port-
folio diversi…cation. This is true because portfolio diversi…cation depends upon
both the number and weightings of portfolio assets, as well as their relative corre-
lations to one another. By including low-correlation asset pairs in your portfolio,
you can ”hedge” the risk of otherwise volatile assets, diversify, and possibly lower
your portfolio’s overall volatility.

2.4. Constructing Markowitz E¢cient Portfolios

To illustrate how the risk-return character of a portfolio changes as we vary the
weighting of securities in the portfolio, we use the data shown in the following
table for two hypothetical stocks. Panel shows the expected returns and standard
deviations for each of the stocks. Note that stock A has both a higher expected
return and standard deviation than stock B to illustrate a trade-o¤ between risk
and return. Also, for purposes of illustration, we assume that the correlation
between the two stocks is (1), +1.0 to represent perfect correlation; (2) 0 to
represent independence; or (3) +0.5 for a correlation that is approximately in
line with the average correlation between stocks in the United States. Using
this date in Panel B, we calculate the expected return and risk for a portfolio of
these two stocks over a range of portfolio weights for each of the three correlation
assumptions. In making these calculations, we assume that the portfolio is fully
invested and that the weight of stock A is WA, so the weight of stock B is (1¡WA )

Panel A
Expected Return Standard Deviation Correlation
Stock A 15% 24% +1.0, 0.0, +0.5
Stock B 125 18%

13
 Panel B
WA 0 20 40 60 80 100
WB = (1 ¡ WA) 100 80 60 40 20 0
Expected return 12.0 12.6 13.2 13.8 14.4 15.0
Standard deviation
Corr=1.0 18.0 19.2 20.4 21.6 22.8 24.0
Corr=0.5 18.0 17.3 17.7 19.0 22.0 24.0
Corr=0.0 18.0 15.2 14.4 16.1 19.5 24.0

Panel B shows the calculated risk and return data for these portfolios where
the weight of stock A ranges in increments of 0.20 from 0 percent to 100 percent
with the corresponding weight of stock B ranging down from 100% to 0 percent.
The third row in the panel shows the expected return for the portfolio and indi-
cates that the portfolio consisting of 100 percent of stock A has the highest return
and the portfolio consisting of 100 percent of stock B has the lowest expected re-
turn. The bottom three rows show the calculated portfolio standard deviation
for each of the three correlations. This illustrates how portfolio variance changes
as the security weighting vary.
For ease of comparison, we show a plot of the varying risk-return trade-o¤ of
the two stock portfolios at each of the three assumed correlation levels. Note that
there is a direct linear trade-o¤ of risk and return when the assumed correlation
is perfectly positive. On the other hand, there is a curvilinear relationship and
a lesser level of risk for the portfolios when the correlation is zero and at 0.5.
This illustrates the general principle that there is productive diversi…cation and
risk reduction when correlation is less than perfectly positive. Correspondingly,
there is a more favorable risk-return trade-o¤ at a correlation of zero than at a
0.5 correlation level, again illustrating the potential gains from diversi…cation as
correlation is lower.
In general, investors will calculate the risk and expected return for each port-
folio. For the portfolios with the same level of risk, there will be a large number
of portfolios, each with a di¤erent expected return. The investor will choose the
portfolio with the greatest expected return for a given level of risk. This portfolio
is the Markowitz e¢cient portfolio. Any portfolio that can be created is
called a feasible portfolio. The collection of all the feasible portfolios is called the
feasible set of portfolios. The Markowitz e¢cient portfolio is that one that gives
the highest expected return of all feasible portfolios with the same risk. (it is
called also the mean-variance e¢cient portfolio).

14
 15 Risk-Return Trade-off
0.05+Expected Return*E-2

10
5

1 2 3 4 5 6
0.04+Standard Deviation*E-2

Figure 2.3:

2.5. Announcements, surprises and expected returns

The actual return on a security (R) consists of two parts: the expected return,
which is predicted by participants in the …nancial markets; and, a surprise or
unexpected part. The expected return is based on a large number of factors
that may in‡uence a given company. The risk for a particular asset comes from
the unexpected part of the return. For a given stock, we can write:

R = E(R) + U (2.5)
The di¤erence between R and E(R) is attributed to surprises re‡ected in the
unexpected return, U.

2.5.1. Risk: Systematic and Unsystematic

The risk of owning an asset, such a common stock, emanates from the unexpected
events; if events always occurred exactly as predicted, there would be no risk. We
identify two types of unexpected events, which give rise to two categories of risk:
systematic and unsystematic risk.

15
 A systematic risk tends to a¤ect a large number of assets to a greater or
lesser degree; systematic risks are also called market risks. An unsystematic
risk a¤ects only a single asset or a small group of assets. Systematic risks arise
from uncertainty about economy-wide factors, such as in‡ation and interest rates,
whereas the possibility of a labor strike or lawsuit involving a single company is
an unsystematic or “idiosyncratic” risk.
We know that :

R = E(R) + U (2.6)
where U is the surprise component. The surprise component is in‡uenced by
two kinds of risk. So we write:

U = Systematic portion + Unsystematic portion (2.7)

or
U = m+² (2.8)
Then:
R = E(R) + m + ² (2.9)
The principle of diversi…cation indicates that spreading an investment across
many assets eliminates some of the risk. The risk which is eliminated by diver-
si…cation is called diversi…able risk. However, there is a minimum level of risk
in a portfolio which can not be eliminated by diversi…cation. This minimum risk
is called nondiversi…able risk.
Why is not all risk diversi…able:
The answer is found in the distinction between systematic and unsystematic
risk. In a large portfolio, the unsystematic risk associated with one stock typ-
ically has no impact on the unsystematic risk associated with any other stock.
Therefore, in a large portfolio, we would not expect to observe any particular
relationship between the unsystematic risks for the individual stocks in the port-
folio.
Systematic risk has an impact on all the stocks in a portfolio. For example,
a recession will have a negative impact on virtually all the securities in the port-
folio, and consequently on the investor’s return from the portfolio. Systematic
risk cannot be diversi…ed away; hence, the terms systematic risk and nondi-
versi…able risk are synonymous.
In this respect, investors will not accept risk without compensation in the
form of a higher expected return. Investors can eliminate unsystematic risk by
diversifying, and the market will not reward investors for taking unnecessary

16
risks. So, investors are not paid for bearing unsystematic risks. In other words,
the reward for bearing risk depends only on the systematic risk of an investment.

2.5.2. Measuring Systematic Risk

The systematic risk of an asset is measured by its beta coe¢cient, represented
by ¯: Beta measures the systematic risk for a particular asset, relative to the
systematic risk for the average asset. For example, the average TSE common
stock has a beta value of 1.0, so a stock with ¯ = 0:5 has half the systematic risk
of the average TSE stock. Or, a stock of ¯ = 1:25 has 25% more systematic risk
than the average stock.
It is important to remember that, since systematic risk is the relevant risk for
an investor, the expected return for an asset is dependent on ¯, a stock with a
high value of ¯ has a high expected return, and a stock with a lower value of ¯
has a lower expected return.

2.5.3. Portfolios Betas

The value of beta for a portfolio (¯ P ) is a weighted average of the betas of the
securities which comprise the portfolio, as indicated in the following:

¯ P = (x1¯ 1 ) + (x2¯ 2 ) + (x3¯ 3 ) + ::::::::: + (xn ¯ n) (2.10)

Where, x’s are the weights, n is the number of assets in portfolios, and ¯ 0s
are the beta values for the assets in the portfolio.

Example 2.2. Suppose, two assets, A and B, with : ¯ A = 1:10; ¯ B = 0:70; and
xA = 0:40 and xB = 0:60

then: ¯ P = 1:10 £ 0:40 + 0:70 £ 0:60 = 0:86:

2.6. The Security Market Line

Now we develop the relationship between systematic risk, measured by Beta, and
the expected return. The graphic representation is called the security market
line (SML) and the algebraic representation is the Capital Asset Pricing Model
(CAPM).

17
2.6.1. Beta and the Risk Premium
We take an example: Let asset A be an individual security with ¯ A =1.2 and
E(RA) = 16%: Assume that the risk-free rate (Rf ) = 5%: Note that beta for the
risk-free asset is zero. Since the risk-free asset does not have any risk, it clearly
does not have any systematic risk. If x is the percentage of the portfolio invested
in asset A, then (1-x) is invested in the risk-free asset. The expected return for a
portfolio is:
E(RP ) = [x £ E(RA )] + [(1 ¡ x) £ E(Rf )] = [x £ 16%] + [(1 ¡ x) £ 5%]
Beta for this portfolio is:
¯ P = [x £ ¯ A] + [(1 ¡ x) £ 0] = [x £ 1:2] + 0 = 1:2x
By selecting various values of x, and then computing the corresponding value
of E(RP ) and ¯ P we can plot expected returns and beta. This relationship be-
tween E(RP ) and ¯ P is a straight line through the point Rf and the point repre-
sented by ¯ A = 1:2 and E(RA )

x 0% 20 50 90 100 150 200

E(RP ) 5.0% 7.2 10.5 14.9 16.0 21.5 27.0
¯P 0.00 0.24 0.60 1.08 .120 1.80 2.40

The slope of the line, is called the reward-to-risk ratio. The slope of a straight
line can be computed between any two points on the line. Take ¯ A; E(RA);Rf to
…nd the slope:
Slope: = (E(RA) ¡ Rf ) = (¯ A ¡ 0) = 0:09167 = 9:167%
This can be regarded in the following manner: for each unit increase in sys-
tematic risk (an increase from beta =0 to beta = 1), an investor will increase her
expected return by 9.16%.
This reward-to-risk ratio can be explained as follow: assume an investor is
considering placing her entire investment portfolio in the risk-free asset. Prior
to making her decision, the investor might want to know what her compensation
would be for taking on risk. That is represented by the reward-to-risk ratio, for
each unit increase in the systematic risk, (Beta), she will increase her expected
rate of return by 9.16%.
Another example, say we have another asset B, with ¯ B = 0:6 and E(RB ) =
10%: Should an investor consider purchasing this asset?. In this case, we compare
asset B to a portfolio (P), comprised of asset A and the risk-free asset, with
¯ P = 0:6.
We know earlier that, ¯ P = 1:2x; so x must be 0.5, then:
E(RP ) = (0:5 £ 16%) + (0:5 £ 5%) = 10:5%

18
 For the same risk, portfolio P has a higher expected return than does asset
B, so that P is preferable to B.
B is inferior to A, and that can be demonstrated through the reward-to-risk
ration, that is:
Slope = (E(RB) ¡ Rf )=(¯ B ¡ 0) = 0:0833 = 8:33%:
8.33% is less than 9.167%, then, an investor would not consider buying asset
B.

25

20

15

10

5
-15 -10 -5 0 5 10 15
x

Security Market Line

2.6.2. The Security Market Line

The above calculation indicates that investors would prefer to hold asset A not
asset B. This would lead to a decrease in the price of asset B and thus an increase
in its expected return; similarly, the price of asset A would increase and its
expected return would decrease. Consequently, in equilibrium, asset B, and all
assets, must fall directly on a straight line, this line is called the Security Market
Line.
A portfolio comprised of all assets in the market is called the market portfolio.
The market portfolio itself is an asset, so its expected return E(RM ) and ¯ M must
be on the security market line.
what then is the value of ¯ M
The market portfolio is comprised of assets in the market and a portfolio beta
is a weighted average of the betas of the securities in the portfolio. Therefore, the
beta for a market portfolio must be equal to beta for the average security; that
is, ¯ M = 1:
The slope of SML would be:

19
 ³ ´ ³ ´
E(RM ) ¡ ¯ f =(¯ M ¡ 0) = E(RM ) ¡ ¯ f =(1 ¡ 0) = E(RM ) ¡ ¯ f
Since this slope is the excess return on the market over the risk-free rate, it
is often called market risk premium, and the equation for the SML becomes:
E(Ri) = Rf + [E(RM ) ¡ ¯ f ] £ ¯ i
where, E(Ri ) and ¯ i are the expected return and beta for any asset. This
equation is referred to as the Capital Asset Pricing Model (CAPM), and it
indicates the expected return for any asset for a given level of systematic risk.

2.7. The Capital Asset Pricing Model (CAPM)

The Capital Asset Pricing Model (CAPM) shows how the rate of return,
or alternatively the price of an asset, is determined in e¢cient markets.
Assumptions underlying the CAPM
1. All investors prefer the highest expected return for any given level of risk
2. Capital markets are e¢cient in that asset prices and yields re‡ect all
information.
3. All investors have the same expectations about each asset’s and return.
This condition is called homogeneous expectations.
4. Investors can both borrow and lend at a risk-free rate of interest.
Investors will calculate the risk and expected return for each portfolio. For the
portfolios with the same level of risk, there will be a large number of portfolios,
each with a di¤erent expected return. The investor will choose the portfolio with
the greatest expected return for a given level of risk (or alternatively, the lowest
risk for a given expected return.) This portfolio is the Markowitz e¢cient
portfolio.
Any portfolio that can be created is called a feasible portfolio. The collec-
tion of all the feasible portfolios is called the feasible set of portfolios. The
Markowitz e¢cient portfolio is that one that gives the highest expected return of
all feasible portfolios with the same risk. (it is called also the mean-variance
e¢cient portfolio).

2.7.1. Combining portfolios of Risk-Free and Risky Assets

Suppose investors want to hold a portfolio of assets that combines a risk-free asset
with an e¢cient portfolio of risky assets found on an e¢cient frontier.
The question would be: Which portfolio among those along the
e¢cient frontier should be included in the combined portfolio.
The question can be answered with the …gures. The curve AD is an e¢cient
frontier of portfolios of risky assets. We draw a straight line from RS the risk-free
rate of return for the risk-free asset, to intersect the e¢cient frontier at V. The

20
line RS V shows combinations of the risk-free asset with the V portfolio of risky
assets on the e¢cient frontier. At RS on the line only the risk-free asset is held,
while at V only the V portfolio of risky assets is held. In between, as we move
along the ling from RS to V, increasingly less of the investor’s given funds are
held in the risk-free asset and more are held in the V portfolio of risky assets.
An investor can go behind the broken portion of the line by borrowing at the
risk-free rate.
Combined portfolios along the straight line running from RS to V are inef-
…cient portfolios relative to those lying on a similar line drawn through RS but
intersecting the e¢cient frontier above V. This is because higher expected return
can be obtained at the same risk. Rotating the RS V line upward until it is just
tangent to the e¢cient frontier produces the optimum portfolio of risky assets
to be combined with the riskless asset. This optimum portfolio of risky asset is
shown at point M. The optimum portfolio M is called the market portfolio of
risky assets, or simply the market portfolio. The straight line RS MC is called
the capital market line (CML), which is the e¢cient frontier of combined
portfolios.
If all investors have the same expectations about asset returns and risk, each
will want to hold a combined portfolio lying along the capital market line. No
investor will want to hold any portfolio or risky assets other than the market
portfolio.

2.8. Arbitrage Pricing Theory (APT)

The CAPM is a normative theory, it does not necessarily describes what investors
do, but rather what they should do if their behavior re‡ects the underlying as-
sumptions of the market model. It has been subject to extensive empirical testing,
but there has been little evidence to support it. This is not surprising given its
assumption of investor behavior. It is unrealistic that all investors have the same
expectations concerning the future rate of return and risk of every asset in the
market, and even less likely that the lending rate equals the borrowing rate (which
equals the “risk-free” rate of interest).
An alternative to the CAPM, which attempts to avoid its shortcoming, is the
Arbitrage Pricing Theory (APT). It is not based on unrealistic assumptions,
however, like the CAPM, the APT states that the expected return, or alterna-
tively the price of an asset, depends on its systematic risk. Whereas the CAPM
de…nes only one source of systematic risk - market conditions - the APT assumes
multiple sources referred t as factors . That APT can be expressed by the
following:

21
 E(RA) = RF + ¯ 1F1 + ¯ 2F2 + ::::::::::::::::: + ¯ K FK (2.11)
where E(RA) is the expected return on an individual asset, RF is the risk-
free rate of return, F1; F2; ::::::FK are systematic risk factors (in‡ation, interest
rates, industrial production, ...), and ¯ 1; ¯ 2; ::::::::¯ K indicate the sensitivity of
the asset’s expected return to the respective factors.
Preliminary …ndings suggest that APT is better at explaining the historical
data than the CAPM, but that may be the result of the empirical methods and
tests employed by di¤erent researchers.

2.8.1. Derivation of the Capital Market Line

It was demonstrated where individuals would like to be regarding risky assets. We
constructed the Markowitz e¢cient frontier. Also we derived the security market
line, where individuals are able to borrow at the risk free rate. Individuals facing
the possibility of borrowing and being on the e¢cient frontier, would prefer to
have the highest expected return for a given level of risk. It was demonstrated
by Sharpe, Treynor and Mossin that the opportunity to borrow or lend at the
risk-free rate implies a capital market where risk-averse investors will prefer to
hold portfolios consisting of combinations of the risk-free and some portfolios M
on the Markowitz e¢cient frontier. Sharpe called this line (up to the tangent
point on the e¢cient frontier) the capital market line (CML).
How does an investor constructs portfolio M. Eugene Fama answered this
question by demonstrating that M must consist of all assets available to investors,
and each asset must be held in proportion to its market value relative to the total
market value of all assets. So, for example, if the total market value of some
asset is $500 million and the total market value of the all assets is $X, then the
percentage of the portfolio that should be allocated to that asset is $500 million
divided by $X.

2.8.2. How to derive the Capital Market Line

Suppose an investor creates two-fund portfolio: a portfolio consisting of wf in-
vested in the risk-free asset and wM in the market portfolio, where w represents
the corresponding percentage (weight) of the portfolio allocated to each asset.
Thus, wf +wM = 1
What is the expected return and risk of this portfolio: E(RP ) =wf Rf +wM E(RM )
since we know that wf = 1¡wM
then E(RP ) = (1¡wM )Rf +wM E(RM ) = R f +wM (E(RM ) ¡ Rf )

22
 What about the risk
Var(RP ) =w2f var(Rf )+w2M var(RM )+2wM wf std(Rf )std(RM )corr(Rf ; RM )
we know that var(Rf ) = 0, also corr(Rf ; RM ) = 0 this is because the risk-free
asset has no variability and therefore does not move at all with the return on the
market portfolio which is a risky asset. So the portfolio variance becomes:
Var(RP ) =w2M var(RM (RM )
or to say:
Std (RP ) =wM Std(RM )
and therefore: wM =Std (RP )=Std(RM )
Now let us use the equation for the expected return:
E(RP ) = Rf +wM (E(R M ) ¡ Rf ) = Rf + [Std (RP )=Std(RM )](E(RM ) ¡ Rf )
This is the equation for the Capital Market Line. The numerator is the
expected return on the market beyond the risk-free return. It is a measure of the
risk premium or the reward for holding the risky market portfolio rather than the
risk-free asset. The denominator is the risk of the market portfolio. Thus, the
slope measures the reward per unit of market risk. Since the CML represents the
return o¤ered to compensate for a perceived level of risk, each point on the line
is a balanced market condition, or equilibrium. The slope is also referred to as
the market price of risk.
According to capital market theory, the risk premium is equal to the market
price of risk times the quantity of risk for the portfolio (as measured by the
standard deviation of the portfolio), that is:
E(RP ) = Rf + Market price of risk £ Amount of portfolio risk

2.8.3. The Minimum Variance Portfolio

In a two-security portfolio, some particular combination of the two securities will
result in the least possible variance. Not surprisingly, this basked of securities is
called the minimum variance portfolio. To …nd the security proportions that
will produce the minimum variance portfolio, it is necessary to understand basic
calculas. To …nd the minimum risk portfolio, take the …rst derivative of the risk
requation, set is equal to zero, and solve for the value of the variable needed.
The variance of a two-security portfolio is as follows:

¾2P = w2a ¾2a + wb2¾2b + 2wawb½ab¾a ¾b (2.12)

We also know that wa + wb = 1, so wb = 1 ¡ wa. Substituting this expression
for wb gives the following version of the above equation:

¾2P = w2a ¾2a + (1 ¡ wa )2 ¾2b + 2wa (1 ¡ wa )½ab¾a ¾b (2.13)

23
 Taking the derivative of the equation for ¾2P with respect to wa and setting it
equal to zero, and solving for wa
we get:

¾2b ¡ ½ab ¾a¾b

wa =
¾2a + ¾2b ¡ 2½ab¾a ¾b

24