3210AFE

Advanced Corporate Finance

Jason West
Semester 1 2011
Week 2. Portfolio Theory and Evidence

3210AFE Advanced Corporate Finance

Jason West
Topic Objectives
After completing this topic, all students should be able to:
 understand the basic model that underlies modern portfolio theory;

 list the assumptions underlying portfolio theory;

 compute the covariance and correlation between two returns given historical data.

 identify a mean-standard deviation diagram and be familiar with its basic elements.

 use means and covariances for individual asset returns to calculate the mean and
variance of the returns on a portfolio of assets.
 Revision
 Last week we examined the simple investment-consumption model
that underlies much of finance theory and, in turn, financial
decision-making.
 This model is based on certainty; that is, the firm’s shareholders
could accurately forecast the firm’s cash flows and the returns from
investment (and the capital market) were set.
 Under these conditions, the utility of shareholders was based solely
on their own subjective tradeoff between current and future
consumption.
 As long as the firm objectively optimized the selection of projects,
the shareholders were always in an optimal position in the presence
of a capital market.
 However, in reality we are faced with uncertainty. This includes
uncertainty regarding future investment returns (risk) and our own-
subjective requirements for consumption.
 We then need a model of optimal decision-making in the presence of
uncertainty: this is where portfolio theory can help.
 C1 Slope = MRS
C0C1

Increasing utility

An aside on utility theory U1

U0
U2

C0
 What we call utility theory underlies our understanding of the theory of
investor choice.
 The notions of utility we develop underlie the asset pricing models we will
examine later along with the principles of portfolio theory.
 They are also important because the utility theory that underlies modern
finance has been relatively recently called into question.
 If true, then much of the investor and market behavior that we take for
granted is not the result of rational utility-maximizing behavior rather
irrational and subjective decision-making.
 In this instance, the asset pricing, portfolio and other models upon which
we rely are invalid.
 How then would we make and understand financial decisions?
 We need to examine more closely the axioms that underlie utility theory
when individuals are faced with the task of choosing risk alternatives
combined with the assumption of non-satiation (i.e. greed).
Axioms of choice under uncertainty
Axiom 1 Comparability (or completeness)
 This axiom says that for entire set of uncertain alternatives, S, the individual
can say that they prefer outcome x to y, y to x, or are indifferent between x
and y.

Axiom 2 Transitivity (or consistency)

 This axiom says if outcome x is preferred to y and y is preferred to z then x
must be preferred to z.

Axiom 3 Strong independence.

 This axiom says that if there is a probability of some outcome x and a
residual probability for outcome z, then if the individual is indifferent
between x and y then they will also be indifferent between a gamble with x
or one with y for z.
Axioms of choice under uncertainty cont…

Axiom 4 Measurability
 This axiom says that if outcome y is preferred less than x but more than
z, then there is some unique probability and residual probability that
will make the individual indifferent between y and a gamble between x
and z respectively.

Axiom 5 Ranking
 If alternatives y and u lie between x and z and we can construct a
gamble between y and x and z and a second between u and x and z (as
above), then if the probability of the first gamble is greater than the
second, then the y is preferred to u.
What do these axioms imply?
 These axioms, along with the assumption of non-satiation of wealth, combine
together to mean that individuals seek to maximise the expected utility of wealth.
 In this, all individuals always make completely rational decisions.
 They make these decisions from amongst an infinite number of alternatives.
 Importantly, there are any number of utility functions that can satisfy these
conditions and still have positive marginal utility.
 The trouble is there are many examples around us in the real world that could be
regarded as inconsistent with these concepts.

 Consider a lotto game where the odds of winning are 1 in 6,991,908 with a
jackpot of $1.6 mil.
 The expected return from a gamble on this jackpot is just $0.22.
 And yet most of us would be willing to pay say $0.50 to gamble even
through we are paying more than the expected return.
 Is it because we irrational or do we overestimate the probability of
success? Either way we usually violate the axioms of utility theory…
A fair game
 However, finance theory relies on the concept that individuals are risk averse.
What does this mean?
 Consider an investor with a legitimate square root utility function such that
U(w) = w0.5, i.e. if w = $1, U = 1; if w = $4, U = 2; if w = $9, U = 3 and so on.
 We assume that the investor’s current wealth is $5.
 To make things simple, we assume that there is only one investment available.
 In this investment a fair coin is flipped. If it comes up heads, the investor wins
$4, increasing his wealth to $9.
 If it comes up tails, the investor loses $4, decreasing his wealth to $1.
 Note that the expected gain is 0.5 × ($4) + 0.5 × (−$4) = $0.

This is called a “fair game.”

 But this game may seem more like a “bet” than an “investment.”
 Fair game cont…
 To see why it’s an investment, consider an investment which costs $5 (all of
our investor’s current wealth) and which has two possible future values: $1 in
the bad case and $9 in the good case.
 This investment is clearly exactly the same as the coin-flipping game.
 Note also that we have chosen a very volatile investment for our example.
 In the bad case, the rate of return is -80%.
 In the good case, the rate of return is +80%.
 Now assume that our investor has only two choices (the set of feasible
investment alternatives has only two elements).
 The investor can either play the game or not play the game (do nothing).
 Which alternative does the investor choose if he follows the principle of
expected utility maximization?
 The next figure shows our investor’s current wealth and utility, the wealth and
utility of the two possible outcomes in the fair game, and the expected outcome
and the expected utility of the outcome in the fair game.
 Fair game cont…
 If the investor refuses to play the game and keeps his $5, he ends up with the
same $5, for an expected utility of 50.5 = 2.24.
 If he plays the game, the expected outcome is the same $5 (0.5 × $1 + 0.5 ×
$9), but the expected utility of the outcome is only 0.5 × 1 + 0.5 × 3 = 2.
 Because he acts to maximize expected utility, and because 2.24 is greater
than 2, he refuses to play the game.
 In general, a risk-averse investor will always refuse to play a fair game
where the expected return is 0%.

Because the utility of expected wealth is greater than the expected utility of
wealth the individual is risk-averse.

 If the expected return is greater than 0%, the investor may or may not choose
to play the game, depending on his utility function and initial wealth.
 For example, if the probability of the good outcome in our example was 75%
instead of 50%, the expected outcome would be $7 (0.25 × $1 + 0.75 × $9).
 The expected gain would then be $2 ($7 - $5), the expected return would be
40%, and the expected utility would be 2.5 (0.25 × 1 + 0.75 × 3).
 Fair game cont…
 Because 2.5 is greater than 2.24, the investor would be willing to make the
investment.
 The expected return of 40% is a ‘risk premium’ which compensates him for
undertaking the risk of the investment.
 Another way of looking at this property of risk aversion is that investors attach
greater weight to losses than they do to gains of equal magnitude (we call this loss
aversion).
 In the example, the loss of $4 is a decrease in utility of 1.24, while the gain of $4 is
an increase in utility of only 0.76.

Now, at what level of wealth would our investor be indifferent between the
investment and doing nothing?
 In our example, the expected utility of the outcome is 2.
 The wealth value which has the same utility is $4 (2 2).
 This value of $4 is called the certainty equivalent (CE).
 As the investor is willing to accept a lower level of wealth but with the same utility,
he is risk-averse.
 If the investor has current wealth less than the CE ($4), he will consider the
investment attractive (although some other investment may be even more
attractive).
 Fair game cont…
 If his current wealth is greater than CE ($4), he will consider the investment
unattractive, because doing nothing has greater expected utility than the
investment.
 If his current wealth is exactly CE ($4), he will be indifferent between
undertaking the investment and doing nothing.
Given this information, what would our investor be willing to give up to
avoid the investment, that is, what is the risk premium?
 Well, the risk premium is defined as the difference between the expected
wealth ($5) and the certainty equivalent ($4).
 This $1 is the risk premium or compensation the investor requires for assuming
risk over certainty.
 As an alternative view, what would the investor be willing to pay to undertake
the risky investment.
 This is defined as the difference between current wealth ($5) and the certainty
equivalent wealth ($4).
 Because this is positive, the investor would need to be paid $1 to undertake the
investment.
 Note that the risk premium is always + (for risk averse investors) but the cost of
the gamble can be + or – depending on the risk and wealth impact of the
investment.
Fair game cont…
 Risk averse investors are always willing to have a lower level of certain
income over a risky investment with the same level of wealth.
 For example, in the CAPM I am only willing to invest in the market portfolio
of risky assets (equity, etc.) if I receive a premium over the risk-free return
(T-bonds).
 Risk averse investors are not risk avoiders, they just need a premium to take
on risk.
 But individuals do vary. Some of us are very risk averse, some less so.
 The cost of the gamble for our investment would have been negative (willing
to pay) if our CE was higher.
 This would occur if we had a greater likelihood of higher, positive outcomes.
 Risk neutral investors are entirely indifferent between the certainty equivalent
and the fair game investment since they have the same utility.
 Risk seekers will always prefer the risky investment over a certainty
equivalent with the same level of wealth…if you offered them $100 they
would say “I’ll flip you for it, double or nothing!”. Their expected wealth is
the same but in terms of utility they would favor the gamble.
Definition of Risk
Aversion
 A risk averse individual will refuse to accept a fair gamble versus the sure thing.
 The utility function for a risk averse individual is must be concave (from below)
such that the chord lies below the utility function.
 The chord or linear locus represents a risk neutral investor that is indifferent
between the fair gamble and the sure thing.
 For the risk neutral investor the expected utility of the gamble equals the
expected value of the gamble.
 A risk loving individual is represented by a utility function which is convex
(from below) such that the chord lies above the utility function.
 A risk loving individual will prefer the fair gamble over the sure thing.

We heavily rely on the assumption of risk-aversion in finance theory

 Combined with the assumption of a normal distribution of returns this says we
maximise utility by simply choosing the best combination of mean and variance.
 This has been critiqued in two ways. First, that the distribution of returns is not
normal. And second, that individuals do not behave in the same way as
suggested by the axioms.
 The Markowitz Model
 When we start to combine individual securities (or any asset for that
matter) together we are talking about a portfolio.
 The concept of portfolio theory encompasses the management of
portfolios.
 The most famous insight in the history of modern investment appeared
in a short paper titled “Portfolio Selection” in the Journal of Finance in
1952.
 The author, Harry Markowitz was then a 25-year old graduate
economics student from the University of Chicago and the idea for the
article came from Markowitz’s studies in linear programming. He
decided to use these techniques with his key insight: that risk is central
to the process of portfolios.
The Markowitz Model cont...
 Despite the acclaim this article would eventually receive, it was more
than 10 years before it attracted much attention.
 Before Markowitz's seminal portfolio theory study in 1952, decisions on
portfolios were made in a rather loose manner.
 The concepts of return and risk were understood, and portfolio holders
intuitively followed the heuristic "don't put all your eggs in one basket".
 However, this informal approach to portfolio behaviour lacked a
theoretical grounding.
 How could we answer questions like "Why does a diversified portfolio
reduce risk?" and "How can I construct such a portfolio?"
 Markowitz reorganised the existing qualitative thought into providing
quantitative measures of risk and return for a portfolio.
Why is Portfolio Theory Important?
 Portfolio theory is central to our understanding of finance for two reasons.
 First, portfolio theory provides a straightforward mathematical approach
whereby portfolios can be constructed whose risk is less than the weighted
sum of its components assets or securities.
 This means that we are creating synthetic risk-return combinations that are
not necessarily reflective of risk-return opportunities available in the
market.
 The widening of risk-return combinations is generally associated with an
increase in utility.
 Second, portfolio theory underlies our asset pricing models.
 For example, CAPM assumes that investors hold fully-diversified portfolios
such that only systematic risk remains.
 As such, expected return only reflects the risk-free rate, the market risk
premium and beta.
 There is no return compensation for unsystematic, non-market or non-
diversifiable risk of individual assets or securities.
Markowitz Portfolio Theory

Assumptions :
 Investors consider each investment alternative as being
represented by a probability distribution of expected returns
over some holding period.
 Investors maximise one-period expected utility, and their utility
curves demonstrate diminishing marginal utility of wealth.
 Investors estimate the risk of the portfolio on the basis of the
variability of expected returns.
 Investors base decisions solely on expected return and risk, so
their utility curves are a function of expected return and the
expected variance (or standard deviation) of returns only.
 Assumptions cont…
 For a given risk level, investors prefer higher returns to lower
returns.
 Similarly, for a given level of expected return, investors prefer less
risk to more risk.
 Under these assumptions, the Markowitz model aims to establish
an efficient portfolio.
 This is either a portfolio that maximises return for a given level of
risk, or one which minimises risk for a given level of return.
 The notion of ‘efficiency’ also came from Markowitz’s studies in
linear programming: it was all about maximising outputs from a
given set of inputs.
Inputs into Portfolio Theory
 One important feature of portfolio theory is that it is extremely parsimonious.
 That is, only a small number of statistics are used to cover a wide range of
outcomes.
 If we use Markowitz Portfolio Theory so as to construct an efficient portfolio,
what information do we need?

1. Portfolio weights
 This tells us how much we have invested in the assets/securities in our portfolio.

2. The expected return for every security considered

 This is a measure of location in the same manner as median and mode. The
concept of an average is standard in our culture, i.e. average grades, average
crime rates, average income.
Portfolio Theory inputs cont…
 Unfortunately, the average (or expected) return is only a good measure of location
in the absence of skewness, i.e. probability distributions that have long tails
extending to lower (left-skewed) or higher (right-skewed) values.
 The expected return, however, has two useful properties: (i) the expected value of
the sum of two returns is equal to the sum of the expected value of each return; (ii)
the expected value of a constant times a return is the constant times the expected
return.

3. The standard deviation of returns, as a measure of the risk of each security

 This is a measure of dispersion in the same manner as range.

4. The covariance (correlation) between the asset/security rates of return

 This is the critical insight into portfolio theory.
 In a portfolio context, it is not just the risk-return of each asset/security that
determines portfolio risk-return, but also how returns co-move or co-vary.
Portfolio weights
The portfolio weight for stock j, denoted xj, is the fraction of a portfolio’s
wealth held in stock j; that is:

Money held in stock j

xj 
Monetary value of the portfolio

By definition, portfolio weights must sum to 1.

Short sales: to sell short common stocks or bonds, the

investor must borrow the securities from someone who owns
them. These then have a negative weight in the portfolio

Long position: opposite to a short position, which was

achieved by buying an investment and hence has a positive
portfolio weight.
The Two-Stock Portfolio
Example 4.1 Computing Portfolio Weights for a
Two-Stock Portfolio

A portfolio consists of $1 million in Westpac stock (WBC) and

$3 million in ANZ stock (ANZ). What are the portfolio weights of
the two stocks?

Answer:
The portfolio has a total value of $4 million. The weight on
WBC is:
$1,000,000/$4,000,000 = .25 or 25 percent,

and the weight on ANZ is

$3,000,000/$4,000,000 = .75 or 75 percent.

Feasible Portfolio Weights
Suppose the world’s financial markets contain only two stocks, WBC and ANZ.
Describe the feasible portfolios.

Answer: In a two-stock world, the feasible portfolios consist of any two

numbers, xWBC and xANZ, for which xWBC = 1 xANZ. Examples of feasible portfolios
include
1. xWBC = .5 xANZ = .5
2. xWBC = 1 xANZ = 0
3. xWBC  2.5 xANZ = -1.5
2 1 2
4. xWBC xANZ 
5. xWBC  1/3 xANZ = 4/3

An infinite number of such feasible portfolios exist because an infinite number

of pairs of portfolio weights solve xWBC  xANZ 1.
Computing Portfolio Weights for a Portfolio of
Many Stocks

Describe the weights of a €40,000 portfolio invested in four

stocks. The amounts invested in each stock are as follows:

Stock 1 2 3 4
Amount €20,000 -€5,000 €0 €25,000

Answer: Dividing each of these investment amounts by the total

investment amount, €40,000, gives the weights

x1 0.5 x2 -0.125 x3 0 x4 0.625

Notation
 Portfolio Return
 When we combine securities into a
portfolio the return (or expected
return) we receive (or expect to
receive) is simply the weighted
average of the individual securities'
expected returns.
 The weight for each security is the
proportion of total investable funds
directed to it within the portfolio.
 Therefore increasing the weight of a
security with a return higher than the
portfolio, will increase portfolio return
and vice versa.
Portfolio Returns
Computing Portfolio Returns for a Two-Stock
Portfolio

A $4,000,000 portfolio consists of $1,000,000 of WBC and

$3,000,000 of ANZ.

If WBC stock has a return of 10 percent and ANZ stock has a

return of 5 percent, determine the portfolio return using

(1)the ratio method, and

(2)the portfolio-weighted average method.
Computing Portfolio Returns for a Two-Stock
Portfolio
• The ratio method divides the value of the portfolio at the end of the
period (plus distributed cash, such as dividends) by the portfolio value
at the beginning of the period and then subtracts 1 from the ratio.
• This means WBC has an end-of-period value plus dividends of
$1,100,000 (10% return). Similarly the end-of-period value of ANZ plus
dividends must amount to $3,150,000 to yield a 5% return.

1,100,000  3,150,000
 1  0.0625  6.25%
1,000,000  3,000,000
--------------------------------------------------------------------------------

•The portfolio-weighted average method weights the returns of the

investments in the portfolio by their respective portfolio weights and
then sums the weighted returns.
• The portfolio weight is .25 for WBC and .75 for ANZ. Thus, the
portfolio-weighted average return is .25(.1) + .75(.05) = .0625 =
6.25%.
Expected Portfolio Returns
The Palisades Quant Fund has a portfolio weight of x in the
Eurostoxx 50 index, which has an expected return of 11 percent. The
fund’s investment in Treasury bills, with a portfolio weight of 1 x,
earns 5 percent. What is the expected return of the portfolio?

Answer: Using equation (4.4a), the expected return is

 
E ( R)  0.11x  0.05(1  x)  0.05  0.06 x
Properties of expected return
1. The expected value of a constant times a return is the constant (x) times the
expected return (r), that is E ( xr )  xE (r )
 
2. The expected value of the sum or difference of two returns is the sum or
difference between the expected returns themselves; that is E (r1  r2 )  E (r1 )  E (r2 )
 
Combining these two equations implies: E ( x1r1  x2 r2 )  x1 E (r1 )  x2 E (r2 )

3. The expected return of a portfolio is the portfolio-weighted average of the

expected returns; that is, for a portfolio of two stocks n
R p   xi ri
i 1

4. The expected portfolio return is the portfolio-weighted average of the expected

returns of the individual stocks in the portfolio
 Portfolio Risk
 Unlike portfolio return, the risk of a portfolio is not just the weighted
average of the individual security's' risk.
 Portfolio risk also depends on the covariances among the returns on
securities in the portfolio.
 We can use the standard deviation of the portfolio to describe portfolio risk,
as well as variance.
 We can see that an important component of portfolio risk is how the returns
of securities within the portfolio move, do they move in the same direction,
opposite directions, strongly or weakly?
 Since covariance is an absolute measure, we cannot make valid
comparisons.
Variances and Standard Deviations
To compute return variances

Assuming that there are a finite number of possible future return

outcomes, that we know what these outcomes are, and that we know
what the probability of each outcome is. Given these possible return
outcomes:
1.Compute demeaned returns. (Demeaned returns simply subtract the
mean return from each of the possible return outcomes.)
2.Square the demeaned returns.
3.Take the probability-weighted average of these squared numbers.
 The variance of a return is the expected value of the squared demeaned
return outcomes, that is
Var (r )  E[(r  r ) 2 ]
where r is the return of the investment, is a random variable, and r is
the expected return of the investment
Computing Variances
Compute the variance of the return of a €400,000 investment in the
hypothetical company SINTEL. Assume that over the next period SINTEL
earns 20 percent 8/10 of the time, loses 10 percent 1/10 of the time, and
loses 40 percent 1/10 of the time.

Answer:
The mean return is 0.8 x 20%  0.1 x(10%)  0.1 x(40%)  11 %
 
Subtract the mean return from each of the return outcomes to obtain the
demeaned returns.
  20%  11%  9%  0.09
  10%  11%  21%  0.21
40%  11%  51%  0.51
The variance is the probability-weighted average of the square of these
three numbers

var  0.8 * (0.09) 2  0.1* (0.21) 2  0.1* (0.51) 2  0.0369

 
Estimating Variances: Statistical Issues
Estimating Variances with Historical Data

Estimate the variance of the return of the FT 100. Recall from Example 4.5 that
the annual returns of the FT 100 from 2003 to 2006 were 12.49 percent, 15.92
percent, 7.47 percent, and 10.71 percent, respectively, and that the average of
these four numbers was 11.65 percent.

Answer:
Subtracting the average return of 11.65 percent from each of these four
returns results in demeaned returns of -0.94 percent for 2003, 4.27 percent for
2004, -4.18 percent for 2005, and 0.84 percent for 2006. Thus, the average
squared demeaned return is:
(0.0094) 2  (0.0427) 2  (0.0418) 2  (0.0084) 2
 0.0009
4
 Covariance
The covariance between two securities i and j can be:
 Positive, indicating that the returns on the two securities tend to move in the
same direction at the same time; when one increases (decreases), the other
tends to do the same.
 Negative, indicating that the returns on the two securities tend to move
inversely; when one increases (decreases), the other tends to decrease
(increase).
 Zero, indicating that the returns on two securities are independent and have no
tendency to move in the same or opposite directions together.
Remember that covariance (like variance) is an absolute or unit specific
measure.
 The size of the covariance will depend on the variances of the security itself,
and its relationship with the other security.
Computing the Covariance from a Joint
Distribution

Determine the covariance between the returns of stocks

A and B given the following joint distribution.

Event Probability Return A Return B

  1 1/6 = .166 .1 0
2 1/6 = .166 .1 .5
3 ½ = .5 .2 0
4 0 .2 .5
5 1/12=.083 .3 0
6 1/12=.083 .3 .5
Covariance cont…

Answer:
The covariance calculation sums the probability-weighted
product of the demeaned outcomes. The mean return for
stock A is .183, and the mean return for stock B is .125,
implying

cov  .166 0.1  .183 0  .125  .166 0.1  .183 0.5  .125  .5 0.2  .183 0  .125
 0  .083 0.3  .083 0  .125  .083 0.3  .183 0.5  .125
 .002083
Computing the Covariance from Historical
Returns

Determine the covariance between the returns of the FT 100

and a portfolio of corporate bonds given the following
annual returns.
FT 100 Corporate FT 100 Corporate
Bond Bond
Return Return Demeaned Demeaned
 
Return Return
Year (Percent)a (Percent)b (Percent) (Percent)
2003 12.49% 7.49% 0.84% 1.49%
2004 7.47 7.36 -4.18 1.36
2005 15.92 8.06 4.27 2.06
2006 10.71 1.08 -0.94 -4.92
Answer: Recall from the earlier example that the average FT 100 return
from 2003 to 2006 was 11.65%. The average corporate bond return was
6.00%. Hence, the covariance estimate is
 
0.0084(0.0149)  0.0418(0.0136)  0.0427(0.0206)  0.0094(0.0492)
0.0002 
4
Correlation
 Correlation or the correlation coefficient corrects some of the ‘problems’
associated with covariance.
 The correlation coefficient is a standardised measure, it will always have a
range between +1 and -1, with
 ρ = +1.0 = perfect positive correlation
 ρ = -1.0 = perfect negative correlation
 ρ = 0 = zero correlation

An example to illustrate these concepts.

 Compare two bank securities, such as WBC and ANZ. Now compare the
returns on a US stock such as IBM to WBC or ANZ.
 Hint: Think of the factors that determine returns, how many would be
different and how many would be the same?
Covariances and Correlations
Variance Is a Special Case of the Covariance. It is useful to
remember that the variance of a return is merely a special case of
a covariance. The variance measures the covariance of a return
with itself; that is
cov(r1r2 )  var(r1 )

Translating Covariances into Correlations. The correlation

between two returns, denoted , is the covariance between the
two returns divided by the product of their standard deviations;
that is cov(r r )
 (r1r2 )  1 2
 1 2

Translating Correlations into Covariances. A formula for

translating correlations into covariances can be obtained by
rearranging the above equation as follows:
cov(r1r2 )   (r1r2 ) 1 2
Computing the Standard Deviation of Portfolios
of Two Stocks

Stock A has a standard deviation of 30 percent per year.

Stock B has a standard deviation of 10 percent per year.
The annualized covariance between the returns of the two
stocks is .0002.

Compute the standard deviation of portfolios with the

following sets of portfolio weights:
 
1. xA .75 xB .25 3. xA1.5 xB .5
2. xA .25 xB .75 4. xA  x xB 1 x

 
   x   x   2 x1 x 2 12
2
p
2
1
2
1
2
2
2
2

Answer: The stocks’ variances are the squares of their standard

deviations. The variance of stock A’s return is therefore .09, stock B’s is .
01. Applying equation (4.9a) above gives the variances for the four cases.
 
1. 0.752(0.09) 0.252(0.01) 2(0.75)(0.25)(0.0002) = 0.051325
2. 0.252(0.09) 0.752(0.01) 2(0.75)(0.25)(0.0002) = 0.011325
3. 1.52(0.09) (0.5)2(0.01) 2(1.5)(0.5)(0.0002) = 0.2047
4. 0.09x2 .01(1 x)2 2(0.0002)x(1 x)
 
Because these are variances, the standard deviations are the square roots
of these results. They are approximately
 
1. 22.65% 3. 45.24%
2. 10.64% 4. .09 x 2  .01(1  x) 2  2(.0002) x(1  x)
Understanding Portfolio Risk
 Having considered the correlation coefficient and the covariance we
are now in a position to understand portfolio risk.
 We have seen that to calculate portfolio risk, we must account for two
factors.
 Weighted individual security risks (ie. the variance of each individual
security, weighted by the percentage of investable funds placed in
each individual security).
 Weighted relationships between securities (i.e. the covariation
between the securities' returns, again weighted by the percentage of
investable funds placed in each security).
 As the number of securities in a portfolio increases, the importance of
an individual security's variance (risk) will fall, whilst the importance
of the covariance relationships will increase.
Understanding Portfolio Risk

 For example, in a portfolio of one security the only component of risk will be
the individual securities own risk.
 As we increase the number of securities the weighting given to this individual
risk will fall, whilst the risk attributable to the covariance between our security
and all other securities will increase.
 Eventually as n (the number of securities) becomes large, the portfolio risk
reduces to the sum of the weighted covariances.
 Such a portfolio will consist of N variance terms (one for each security) and N 2-
N covariance terms.
 For example, to calculate the portfolio risk of a set of just ten securities will
require ten variances and ninety covariances!
 In practice, there are many software packages available that assist us in the
rather mundane act of calculating the minimum variance set and its components.
 The linear programming tool in Excel can also help for a small number of
assets.
The efficient set with two risky assets
(and no risk-free asset)
 Let us assume that there is no-risk free asset (i.e. zero variance) in which we can
invest, just two risky assets.
 By altering the percentage of invested assets in each we eventually attain some
minimum variance set.
 The minimum variance set is the locus of risk and return combinations offered by
portfolios of risky assets that yields the minimum variance for a given rate of
return.
 As long as the assets included are less than perfectly correlated, this minimum
variance set must be convex.
 Obviously, we are not interested in interior solutions: these will always be
dominated by those on the frontier (i.e. less risk for the same return, more return
for the same risk).
 Note that in a two asset portfolio there are never any inefficient (or interior)
combinations.
Two risky assets
How does the investor decide in which portfolio combination to optimally
invest?
The answer is given by adding the utility curves discussed last week.

 Clearly, the investor will wish to lie on their highest attainable utility curve, but
these must be attainable.
 Eventually, we will pick a point of tangency between a utility curve and the
minimum variance set. This will be point C where the subjective MRS is
equated with the objective MRT.
 At point A the MRS is steeper that the MRT indicating that we would be
willing to give up a lot of return just to reduce risk a little.
 By moving along the frontier (to the left) this can be achieved by substituting
to a less-risky portfolio combination.
 This is in the direction of portfolio E which offers identical utility, but with
lower risk and return.
 At the optimal point, my subjective rate of tradeoff is equated with the
portfolio's capacity to tradeoff and I can be no better off.
 Note that a rational investor will, in fact, never prefer a point below the
minimum variance point, as all portfolios below this are dominated by those
above (i.e. same risk, more return).
The efficient set with one risky asset
and one risk-free asset
 The most important characteristics of the risk-free asset is that it has zero
variance (of course) and that its covariance with the risk asset is zero.
 Accordingly, portfolio variance is the variance of the risk asset and the
opportunity set is linear.
 That is, the slope of the minimum variance set is independent of the
proportion of investment in the risk-free asset.
 In order to attain portfolio combinations along the interval XV it is
necessary to borrow (sell-short) the risk-free asset (leverage) and invest
in the risky asset.
 Between Y and X we are investing some proportion in the risk-free asset.
The decrease in risk and return is constant as this increases.
 If we decide to invest more than 100% in the risk-free asset we must sell
short the risky asset. Assuming no restrictions on short sales the segment
YZ represents the various payoffs.
 In general, the availability of short-selling shifts the minimum variance
set to the left.
A worked example via Excel
Microsoft Office
Excel 97-2003 Worksheet

Portfolio analysis
Two-asset portfolio
Month Firm A Firm B w1 1-w1 St. Dev Mean

1 0.1126 -0.0324 1.0 0.0 0.0504 0.0188

2 0.0452 0.0051 0.9 0.1 0.0448 0.0179
2.00%
3 -0.0249 -0.0094 0.8 0.2 0.0393 0.0170 1.80%
4 -0.0403 0.0016 0.7 0.3 0.0339 0.0161
5 -0.0014 0.0243 0.6 0.4 0.0288 0.0152 1.60%
6 -0.0519 0.0200 0.5 0.5 0.0240 0.0144
7 0.0370 0.0398 0.4 0.6 0.0198 0.0135 1.40%
8 -0.0228 0.0067 0.3 0.7 0.0167 0.0126
1.20%
Mean return

9 0.0131 0.0185 0.2 0.8 0.0153 0.0117

10 0.0259 0.0198 0.1 0.9 0.0162 0.0108 1.00%
11 0.0885 0.0010 0.0 1.0 0.0189 0.0100
12 0.0441 0.0246 0.80%
Mean 0.0188 0.0100 0.60%
St. Dev. 0.0504 0.0189 0.40%
Corr. -0.3162 0.20%
0.00%
0.00% 1.00% 2.00% 3.00% 4.00% 5.00% 6.00%
Standard deviation
 Where to now?
 The principles of portfolio risk-return are easily extended to the N risky
asset case.
 We can see here that ultimately the risk of our portfolio collapses to the
sum of its (many) covariances and that individual risk becomes increasingly
less important.
 In such a situation, the portfolio of risky assets is comprised solely of non-
diversifiable (or market) risk.
 This puts us in the starting position for understanding asset pricing models.

 But to move to this point, we need some general conditions that imply that
all investors will hold identical portfolios of risky assets, otherwise the
optimal portfolio will vary subjectively.
 This is the essence of two-fund separation.
 Theory in Practice…

Week 2. Portfolio Theory and Evidence

 Mutual Funds
 We want to know if the time, talent and effort involved in picking securities pays
off – after transaction costs and other trading realities are taken into account.
 Previous studies indicate that return on actively managed funds, on average,
underperforms the market index (includes survivor bias).
 Any research must distinguish skill from luck – the only way to do this is to group
funds based on some observable characteristic and then examine the average
performance of the group.
 Skilled funds should have done better in the past and also in the future, on average
– if there is skill, we should notice persistence in performance.
 All studies indicate that there is NO PERSISTENCE AT ALL.
 Funds that did well in the past were no more likely to do well in the future, and
active management does not provide excess returns.
 Mutual Funds
 Funds that did well took on more market risk
(a bit of measure error).
 Wide dispersion in average returns is
surprising – looks like single stocks.
 This implies that funds are not well diversified
– they are loading up on specific bets.
 The average fund looks bad – but are the good
funds any good? NO MORE THAN THE
MARKET.
 This is a surprising fact – professionals in
every field should do better than amateurs.
 Like sports, we would expect a few stars to
shine (even via randomness) but this does not
occur.
Economic Interpretation of Strategies
 Prices reveal slow-moving market expectations of subsequent returns because potential
offsetting events seem sluggish…
 You have to buy stocks or long-term bonds at the bottom of a recession or the peak of a
financial panic.
 You have to buy stocks that everyone else thinks are dogs.
 You have to invest in value or small-cap companies with years of poor past returns, poor
sales or on the edge of bankruptcy.
 You have to sell the stocks or long-term bonds during the good times, when stock prices are
high, or the yield curve is inverted.
 You have to sell now the stocks that everyone else is buying, solid blue-chip companies with
good sales and earnings growth.

 If this feels uncomfortable, what you are feeling is risk.

Market Behaviour: Irrational or Not?
 Royal Dutch and Shell split their joint cashflows on a 60/40 basis. Therefore
we would rationally expect Royal Dutch stock to be worth 1.5 times the value
of Shell’s stock. What does the market think?
 LTCM bet a lot of money on this ‘arbitrage’ and lost.
Investors Do Not Outperform Simple Indexes

 While some professional managers do better than the market in any given year, some do
worse, and the outcomes look very much like luck.
 Funds that do well in one year are not more likely to do better than average the next year.
 The average actively managed fund performs about 1 percent worse than the market
index.
 The more actively a fund trades the lower the returns to investors.

63
For next week
 Read Chapter 4 and take appropriate notes.

 Complete Exercises 4.3, 4.5, 4.6, 4.22 and using

Excel 4.23(a-d) in HGT. Our first tutorial in Week
3 will cover the assigned work for Week 1 and
Week 2.
 Read Chapter 5 for next week if you have time.