Mathematical Statistics

• Random variable
– A variable that records, in numerical form, the
possible outcomes from some random event
• Probability density function (PDF)
– A function f(x) that shows the probabilities
associated with the possible outcomes from a
random variable

1
 Mathematical Statistics
• Expected value of a random variable
– The outcome of a random variable that will
occur ‘‘on average’’
– The expected value is denoted by E(x)
– If x is a discrete random variable
n
with n
outcomes, then E ( x)   xi f  xi 
i 1

– If x is a continuous random variable, then


E ( x)   x f  x  dx

2
 Mathematical Statistics
• Variance and standard deviation of a
random variable
– Measure the dispersion of a random variable
about its expected value n 2

Discrete: Var ( x)   x2    xi  E  x  f  xi 

i 1

2
  x  E  x  f  x  dx
2
Continuous: Var ( x)    x


Standard deviation =  x   x2
3
 Fair Gambles and The Expected
Utility Hypothesis

• A ‘‘fair’’ gamble
– Specified set of prizes and associated
probabilities that has an expected value of
zero
– Common observation: people would prefer
not to play fair games

4
 St. Petersburg Paradox
• A coin is flipped until a head appears
– If a head appears on the nth flip, the player is
paid $2n
x1 = $2, x2 = $4, x3 = $8,…,xn = $2n
– The probability of getting a head on the ith
trial is (½)i
1=½, 2= ¼,…, n= 1/2n

5
 St. Petersburg Paradox
• The expected value of the St. Petersburg
paradox game is infinite
i
i1

E ( x)    i xi   2    1  1  1  ...  1  ...  
i 1 i 1 2
• Because no player would pay a lot to
play this game, it is not worth its infinite
expected value

6
 Expected Utility
• Individuals do not care directly about the
dollar values of the prizes
– They care about the utility that the dollars
provide
• If we assume diminishing marginal utility of
wealth
– The St. Petersburg game may converge to a
finite expected utility value
• This would measure how much the game is worth
to the individual

7
 Expected Utility
• Expected utility
– Can be calculated in the same manner as
expected value n
E ( x)    iU ( xi )
i 1

• Because utility may rise less rapidly than

the dollar value of the prizes
• It is possible that expected utility will
be less than the monetary expected
value
8
 7.1 Bernoulli’s Solution to the Paradox
and Its Shortcomings
• Utility of each prize in the St. Petersburg
paradox is U(xi)=ln xi
– Diminishing marginal utility (U’ > 0 but U’’ < 0),
– The expected utility value of this game
converges to a finite number:
 
1
expected utility =   iU ( xi )   i ln(2 )  1.39
i

i 1 i 1 2

9
 7.1 Bernoulli’s Solution to the Paradox
and Its Shortcomings

• Bernoulli’s solution to the St. Petersburg

paradox
– Does not completely solve the problem
– As long as there is no upper bound to the utility
function
• The paradox can be regenerated by redefining the
gamble’s prizes

10
 The von Neumann-Morgenstern
Theorem
• Suppose that there are n possible prizes
that an individual might win (x1,…xn)
– Arranged in ascending order of desirability
– x1 = least preferred prize  U(x1) = 0
– xn = most preferred prize  U(xn) = 1
– Show that there is a reasonable way to assign
specific utility numbers to the other prizes
available

11
 The von Neumann-Morgenstern
Theorem
• The von Neumann-Morgenstern method
– The utility of xi - the expected utility of the gamble
that the individual considers equally desirable to
xi
U(xi) = i · U(xn) + (1 - i) · U(x1)
• Since U(xn) = 1 and U(x1) = 0
U(xi) = i · 1 + (1 - i) · 0 = i
• The utility number attached to any other prize is
simply the probability of winning it
• This choice of utility numbers is arbitrary
12
 Expected Utility Maximization
• Consider two gambles:
– First gamble offers x2 with probability q and x3
with probability (1-q)
expected utility (1) = q · U(x2) + (1-q) · U(x3)
– Second gamble offers x5 with probability t and
x6 with probability (1-t)

expected utility (2) = t · U(x5) + (1-t) · U(x6)

13
 Expected Utility Maximization
– Substituting the utility index numbers gives
expected utility (1) = q · 2 + (1-q) · 3
expected utility (2) = t · 5 + (1-t) · 6
• The individual will prefer gamble 1 to
gamble 2 if and only if
q · 2 + (1-q) · 3 > t · 5 + (1-t) · 6

14
 Expected Utility Maximization
• Expected utility maximization
– If individuals obey the von Neumann-
Morgenstern axioms of behavior in uncertain
situations
– They will act as if they choose the option that
maximizes the expected value of their von
Neumann-Morgenstern utility index

15
 Risk Aversion
• Two lotteries
– May have the same expected value but differ
in their riskiness
• Risk
– The variability of the outcomes of some
uncertain activity
• When faced with two gambles
– With the same expected value, individuals will
usually choose the one with lower risk

16
 Risk Aversion
• Assumption: marginal utility of wealth falls
as wealth gets larger
– A flip of a coin for $1,000
• Small gain in utility if you win
• Large loss in utility if you lose
– A flip of a coin for $1
• Inconsequential: gain in utility from a win is not
much different as the drop in utility from a loss

17
 7.1
•Utility of Wealth from Two Fair Bets of Differing Variability

Utility (U)

U(W)
U(W0)
EU(A)=U(CEA)
EU(B)

Wealth (W)
W0-2h W0-h W0 W0+h W0+2h
CEA
•If the utility-of-wealth function is concave (i.e., exhibits a diminishing marginal utility of
wealth), then this person will refuse fair bets. A 50–50 chance of winning or losing h dollars,
for example, yields less expected utility [EU(A)] than does refusing the bet. The reason for
this is that winning h dollars means less to this individual than does losing h dollars.
18
 Risk Aversion
• Gamble A:
– A 50-50 chance of winning or losing $h
• Expected utility:
• EU(A) = ½ U(W0+h) + ½ U(W0-h)
• Gamble B:
– A 50-50 chance of winning or losing $2h
• Expected utility:
• EU(B) = ½ U(W0+2h) + ½ U(W0-2h)

19
 Risk Aversion
• Gamble A:
– A 50-50 chance of winning or losing $h
• Expected utility:
• EU(A) = ½ U(W0+h) + ½ U(W0-h)
• Gamble B:
– A 50-50 chance of winning or losing $2h
• Expected utility:
• EU(B) = ½ U(W0+2h) + ½ U(W0-2h)
U(W0)>EU(A)>EU(B)

20
 Risk Aversion and Insurance
• Risk aversion
– An individual who always refuses fair bets is
said to be risk averse
– If individuals exhibit a diminishing marginal
utility of wealth, they will be risk averse
– As a consequence, they will be willing to pay
something to avoid taking fair bets

21
 7.2 Willingness to Pay for Insurance

• A person with a current wealth of $100,000

– Faces a 25% chance of losing his automobile worth $20,000
– Von Neumann-Morgenstern utility index is:
U(W) = ln (W)
– Expected utility without insurance
• EU(no insurance) = 0.75U(100,000) + 0.25U(80,000) = 0.75
ln100,000+0.25 ln80,000 = 11.45714
Query: What will be the fair insurance premium?
- (25% of 20000= 5000)
– Expected utility with insurance
• EU(fair insurance)=U(95,000)=ln 95,000 = 11.46163

22
 7.2 Willingness to Pay for Insurance

– EU(maximum-premium insurance)=
U(100,000 – x) = ln (100,000 – x) =11.45714
– So x=5,426
• This person
– Would be willing to pay up to $426 in
administrative costs to an insurance company
– In addition to the $5,000 premium to cover the
expected value of the loss
– Is as well off as he or she would be when
facing the world uninsured
23
 Measuring Risk Aversion
• The most commonly used risk aversion
measure was developed by Pratt
U "(W )
r (W )  
U '(W )
• For risk averse individuals, U”(W) < 0
– r(W) will be positive for risk averse
individuals
– r(W) is not affected by which von
Neumann-Morganstern ordering is used

24
 Measuring Risk Aversion
• The Pratt measure of risk aversion
– Is proportional to the amount an individual will
pay to avoid a fair gamble
– h - winnings from a fair bet, E(h) = 0
– p - size of the insurance premium
• That would make the individual exactly indifferent
between taking the fair bet h and paying p with
certainty to avoid the gamble
E[U(W + h)] = U(W - p)

25
 Measuring Risk Aversion
• Expand both sides of the equation using
Taylor’s series
– Because p is a fixed amount, we can use a
simple linear approximation to the right-hand
side
• U(W - p) = U(W) - pU’(W) + higher order terms
• E[U(W + h)] = E[U(W) - hU’(W) + h2/2 U” (W) +
higher order terms
• E[U(W + h)] = U(W) - E(h)U’(W) + E(h2)/2 U” (W)
+ higher order terms

26
 Measuring Risk Aversion
– E(h)=0
– Drop the higher order terms
– Substituting k for E(h2)/2, we get

U (W )  pU '(W )  U (W )  kU "(W )
kU "(W )
p   kr (W )
U '(W )

27
 Risk Aversion and Wealth
• It is not necessarily true that risk aversion
declines as wealth increases
– Diminishing marginal utility
• Would make potential losses less serious for high-
wealth individuals
• Makes the gains from winning gambles less
attractive
– The net result depends on the shape of the
utility function

28
 Risk Aversion and Wealth
• If utility is quadratic in wealth,
U(W) = a + bW + cW 2
– Where b > 0 and c < 0
– Pratt’s risk U
aversion
"(W ) measure
2c is
r (W )   
U (W ) b  2cW
• Risk aversion increases as wealth
increases

29
 Risk Aversion and Wealth
• If utility is logarithmic in wealth,
U(W) = ln (W )
– Where W > 0
– Pratt’s risk aversion measure is
U "(W ) 1
r (W )   
U (W ) W

• Risk aversion decreases as wealth

increases

30
 Risk Aversion and Wealth
• If utility is exponential,
U(W) = –e–AW = –exp (–AW)
– Where A is a positive constant
– Pratt’s risk aversion measure
2  AW is
U "(W ) A e
r (W )     AW
A
U (W ) Ae
• Risk aversion is constant as wealth
increases

31
 7.3 Constant Risk Aversion
• How much ( f ) would an individual pay to
avoid the risk?
– Initial wealth is W0
– Utility function exhibits constant absolute risk
aversion
– A 50–50 chance of winning or losing $1,000
– To find f, we set the utility of W0-f equal to the
expected utility from the gamble
- exp [-A(W0-f)] = -0.5 exp [-A(W0+1,000)] -0.5 exp
[-A(W0-1,000)]
exp(Af) = 0.5exp(-1,000A)+ 0.5exp(1,000A)

32
 7.3 Constant Risk Aversion

• A person faces a random shock to his or

her wealth that follows a normal
(W )  1/ 2  e  z /2 , where z  [(W   ) /  ]
fdistribution
2

– With mean μ and variance  AW σ2

utility function: U (W )  e
– The probability density function for wealth:
expected utility from risky wealth:

1 [(W   )/ 2 ]/2
E[U(W)]=  U (W ) f (W )dW   e
 AW
e dW
 2
A 2
E[U(W)]   -
2
33
 Relative Risk Aversion
• Unlikely
– Willingness to pay to avoid a gamble is
independent of wealth
• More appealing assumption
– Willingness to pay is inversely proportional to
wealth
• Relative risk aversion:
U "(W )
rr (W )  Wr (W )   W
U '(W )
34
 Relative Risk Aversion
• The power utility
W R function
/ R   for R  1, R  0
U(W, R)  
ln R if R  0
– Diminishing absolute relative risk aversion
U "(W ) ( R  1)W R  2 ( R  1)
r (W )    R 1

U '(W ) W W
– But constant relative risk aversion

rr (W )  Wr (W )  ( R  1)  1  R
35
 7.4 Constant Relative Risk Aversion

• Constant relative risk aversion utility function

– What fraction of initial wealth ( f )
• Willing to give up to avoid a fair gamble of, 10% of
initial wealth
– Assume R = 0
– Logarithmic utility function
ln[(1-f)W0] = 0.5 ln(1.1W0) + 0.5 ln(0.9W0)
ln(1-f) = 0.5 ln(1.1) + 0.5 ln(0.9) = ln(0.99)0.5
f=0.005
– Sacrifice up to 0.5 percent of wealth to avoid the
10 percent gamble

36
Methods for Reducing Uncertainty and Risk

• Four different methods that individuals can

take to mitigate the problem of risk and
uncertainty:
– Insurance
– Diversification
– Flexibility
– Information

37
 Insurance
• Risk-averse people
– Pay a premium to have the insurance
company cover the risk of loss
– Would always want to buy fair insurance to
cover any risk he or she faces
• Insurance company
– Cannot stay in business if it offered fair
insurance
• Premium exactly equals the expected payout for
claims

38
 Insurance
• An insurance customer
– Can always expect to pay more than an
actuarially fair premium
• If people are sufficiently risk averse
– They will even buy unfair insurance
– The more risk averse they are, the higher the
premium they would be willing to pay

39
 Insurance
• Factors that make insurance difficult or
impossible to provide
– Large-scale disasters
– Rare and unpredictable events
– Informational disadvantage the company may
have relative to the customer
• Adverse selection problem
– Moral hazard problem

40
 Diversification
• Diversification
– Suitably spreading risk around - to reduce the
variability of an outcome without lowering the
expected payoff
• A person has wealth W to invest
– Can be invested in two independent risky
assets, 1 and 2
• Equal expected values (μ1=μ2)
• Equal variances (σ21=σ22)

41
 Diversification
• Undiversified portfolio, UP
– Includes just one of the assets
– Expected return: μUP = μ1=μ2
– Variance: σ2UP = σ21=σ22

42
 Diversification
• Diversified portfolio, DP
• 1 – the fraction invested in the first asset
• (1- 1) – the fraction invested in the second
– Expected return:
μDP = 1 μ1+(1- 1)μ2= μ1=μ2
– Variance;
σ2DP = 21 σ21+(1- 1)2 σ22 = (1-21+ 221 ) σ21
– Minimize σ2DP
1 = ½ ; σ2DP= σ21/2
43
 Flexibility
• Flexibility
– Allows the person to adjust the initial decision
depending on how the future unfolds
– The more uncertain the future, the more
valuable this flexibility
– Keeps the decision-maker from being tied to
one course of action
• And instead provides a number of options

44
 Flexibility
• Financial option contract
– Offers the right, but not the obligation, to buy
or sell an asset
• During some future period
• At a certain price
• Real option
– An option arising in a setting outside of
financial markets

45
 Flexibility
• All options share three fundamental
attributes
– Specify the underlying transaction
– Specify a period over which the option may be
exercised
– Specifies a price

46
 Flexibility
• Model of real options
– Let x embody all the uncertainty in the economic
environment
– The individual has some number, I = 1,…,n, of
choices currently available
– Ai(x) - payoffs provided by choice I
• (x) allows each choice to provide a different pattern of
returns depending on how the future turns out

47
 Flexibility
• Model of real options
– No flexibility
• Choose the single alternative that is best on
average
• Expected utility from this choice: max{E[U(A1)] ,…,
E[U(An)]}
– Flexibility
• Choose the best alternative
• Expected utility: E{max[U(A1),…,U(An)]}

48
 7.2
•The Nature of a Real Option

•Panel (a) shows the payoffs and panel (b) shows the utilities provided by two alternatives
across states of the world (x). If the decision has to be made upfront, the individual chooses
the single curve having the highest expected utility. If the real option to make either decision
can be preserved until later, the individual can obtain the expected utility of the upper
envelope of the curves, shown in bold
49
 Flexibility
• More options are better (generally)
– Options give the holder the right—but not the
obligation—to choose them
• In a strategic setting
– Economic actors may benefit from having
some of their options cut off
• Allow a player to commit to a narrower course of
action
• This commitment may affect the actions of other
parties

50
 7.3
•More Options Cannot Make the Individual Decision-Maker Worse Off

•The addition of a third alternative to the two drawn in Figure 7.2 is valuable in (a) because it
shifts the upper envelope (shown in bold) of utilities up. The new alternative is worthless in (b)
because it does not shift the upper envelope, but the individual is not worse off for having it.
51
 Flexibility
• Computing option value
– Let F be the fee that has to be paid
• For the ability to choose the best alternative after x
has been realized instead of before
– The individual would be willing to pay the fee
as long as:
E{max[U(A1(x)-F),…,U(An(x)-F)]} ≥
max{E[U(A1(x)] ,…, E[U(An(x)]}

52
 7.5 Value of a Flexible-Fuel Car

• A1(x)= 1-x
– The payoff from a fossil-fuel–only car
• A2(x) = x
– The payoff from a biofuel-only car
• State of the world, x
– Reflects the relative importance of biofuels
compared with fossil fuels over the car’s lifespan
– Random variable, uniformly distributed between 0
and 1
53
 7.5 Value of a Flexible-Fuel Car
• Probability density function (PDF) is f (x) =
1
– When the uniform random variable ranges
between 0 and 1
• Suppose first that the car buyer is risk
neutral
1
– Utility level = payoff level1
E[ A2 ]   A2 ( x) f ( x)dx 
– Forced to choose a biofuel 2 car
0
– Expected utility:
with f ( x)  1
54
 7.5 Value of a Flexible-Fuel Car

• Risk neutrality
– Flexible-fuel car is available
– Buyer: either A1(x) or A2(x), whichever is
higher under the latter
1
circumstances
E[max( A1 , A2 )]   max(1  x, x) f ( x)dx 
– Expected utility:
0
1/2 1
  (1  x)dx   xdx  3 / 4
0 1/2

55
 7.5 Value Uof( x)aFlexible-Fuel
x Car

• Risk aversion,
1 1
E[U ( A )] 
2  2
– Expected0 utility from a biofuel

A ( x) f ( x)dx  x1/2 dx  2 / 3
0 car:
• Expected utility from a fossil-fuel car; E[U(A1)]=2/3
• Expected utility from a flexible-fuel car that costs F
more than a single-fuel car:

E{max[U ( A1 ( x)  F ), U ( A2 ( x)  F )]} 
1 1/2 1
  max( 1  x  F , x  F ) f ( x)dx   1  x  F dx   x  F dx
0 0 1/2

56
 7.4
•Graphical Method for Computing the Premium for a Flexible-Fuel Car

•To find the maximum

premium F that the risk-averse
buyer would be willing to pay
for the flexible-fuel car, we plot
the expected utility from a
single-fuel car from Equation
7.55 (E[U(A2)]) and from the
flexible-fuel car from Equation
7.56 (E{max[U(A1(x)-F),
U(A2(x)-F)]})and see the value
of F where the curves cross.

57
 Flexibility
• Option value of delay
– Delay preserves options
– If circumstances continue to be favorable or
become even more so, the action can still be
taken later
– But if the future changes and the action is
unsuitable, the decision-maker may have
saved a lot of trouble by not making it

58
 Flexibility
• The cost–benefit rule
– An action should be taken if anticipated costs
are less than benefits
– Correct course of action in simple settings
without uncertainty
– In settings involving uncertainty
• Account for risk preferences
• Account for the option value of delay

59
 Information
• Information
– Difficult to quantify
– Has some technical properties that make it an
unusual sort of good
• Durable and retains value after it has been used
– Pure public good: non-rival and non-exclusive

60
 Information
• Quantifying the value of information
– Individual is uncertain about what the state of
the world (x) will be in the future
• Needs to make one of n choices today
• Ai(x) represents the payoffs provided by choice I
• F - the fee charged to be told the exact value that x
will take on in the future; maximum F:
E{max[U(A1(x)-F),…,U(An(x)-F)]}=
max{E[U(A1(x)] ,…, E[U(An(x)]}

61
 The State-Preference Approach to
Choice Under Uncertainty

• Outcomes of any random event

– Can be categorized into a number of states of
the world
– “Good times” or “bad times”
• Contingent commodities
– Goods delivered only if a particular state of
the world occurs
– “$1 in good times” or “$1 in bad times”

62
 States of the World
• It is conceivable that an individual could
purchase a contingent commodity
– Buy a promise that someone will pay you $1 if
tomorrow turns out to be good times
– This good will probably sell for less than $1

63
 Utility Analysis
• Assume two contingent goods
– Wealth in good times (wg) and wealth in bad
times (wb)
– Individual believes the probability that good
times will occur is 
– Expected utility:

V(Wg,Wb) = U(Wg) + (1 - )U(Wb)

• This is the value that the individual wants to
maximize given his initial wealth (W)

64
 Prices of Contingent Commodities

• Assume
– The person can buy $1 of wealth in good times
for pg and $1 of wealth in bad times for pb

– His budget constraint: W = pgWg + pbWb

• The price ratio pg /pb
– Shows how this person can trade dollars of
wealth in good times for dollars in bad times

65
Fair Markets for Contingent Goods

• If markets for contingent wealth claims

– Are well-developed
– And there is general agreement about 

– Prices for these goods will be actuarially fair:

pg =  and pb = (1- )
– The price ratio will reflect the odds in favor of
good times pg 

pb 1  
66
 Risk Aversion
• If contingent claims markets are fair
– A utility-maximizing individual will opt for a
situation in which Wg = Wb
• Arrange matters so that the wealth obtained is the
same no matter what state occurs
– Maximization of utility subject to a budget
constraint:
V / Wg  U '(Wg ) pg
MRS   
V / Wb (1   )U '(Wb ) pb
67
 Risk Aversion
• If markets for contingent claims are fair
U '(Wg )
1 or Wg  Wb
U '(Wb )
• When faced with fair markets in contingent
claims on wealth
– Individual will be risk averse
• And will choose to ensure that he or she has the
same level of wealth regardless of which state
occurs
68
 7.5
•Risk Aversions in the State-Preference Model

Wb

certainty line

W*

U1
I I’
Wg
W*
•The line I represents the individual’s budget constraint for contingent wealth claims: W =
pgWg + pbWb. If the market for contingent claims is actuarially fair [p g /pb = π/(1- π)], then
utility maximization will occur on the certainty line where W g = Wb = W*. If prices are not
actuarially fair, the budget constraint may resemble I’, and utility maximization will occur at a
point where Wg > Wb.
69
 7.6 Insurance in the State-Preference
Model
• A person with wealth of $100,000
– Faces a 25% chance of losing his automobile
worth $20,000
– Wealth with no theft (Wg) = $100,000 and
probability of no theft = 0.75
– Wealth with a theft (Wb) = $80,000 and probability
of a theft = 0.25
– Assume logarithmic utility
E(U) = 0.75U(Wg)+0.25U(Wb)=0.75ln Wg + 0.25ln Wb
E(U) = 11.45714
70
 7.6 Insurance in the State-
Preference Model
• The budget constraint
– Written in terms of the prices of the contingent
commodities
pgWg* + pbWb* = pgWg + pbWb
– Assuming that these prices equal the
probabilities of these two states
0.75(100,000) + 0.25(80,000) = 95,000
– The expected value of wealth = $95,000

71
 7.6 Insurance in the State-
Preference Model
• The individual will move to the certainty line
and receive an expected utility of
E(U) = ln 95,000 = 11.46163
– To be able to do so, the individual must be able
to transfer $5,000 in extra wealth in good times
into $15,000 of extra wealth in bad times
• A fair insurance contract will allow this

• The wealth changes promised by insurance

(dWb/dWg) = 15,000/-5,000 = -3

72
 7.6 Insurance in the State-
Preference Model
• A policy with a deductible provision
– Insurance policy costs $4,900, but requires the
person to incur the first $1,000 of the loss

Wg = 100,000 - 4,900 = 95,100

Wb = 80,000 - 4,900 + 19,000 = 94,100
E(U) = 0.75 ln 95,100 + 0.25 ln 94,100
E(U) = 11.46004
– The policy still provides higher utility than doing
nothing
73
 Risk Aversion and Risk
Premiums
• Two people
– Each starts with an initial wealth of W*
– Each seeks to maximize an expected utility
function of the form
WgR WbR
V (Wg ,Wb )    (1   )
R R
• Constant relative risk aversion
• R determines the degree of risk aversion
• R determines the degree of curvature of
indifference curves implied by the function
• A very risk averse individual- a large negative R
74
 7.6
•Risk Aversion and Risk Premiums

Wb

certainty line

W*
W* - h U1

U2

Wg
W* W2 W1

•Indifference curve U1 represents the preferences of a risk-averse person, whereas the person with
preferences represented by U2 is willing to assume more risk. When faced with the risk of losing h in bad
times, person 2 will require compensation of W2 – W* in good times, whereas person 1 will require a larger
amount given by W1 - W*.
75
 Asymmetry of Information
• The level of information that a person buys
– Will depend on the price per unit
• Information costs may differ significantly
across individuals
– Specific skills for acquiring information
– Experience that is relevant
– Different former investments in information
services

76
 The Portfolio Problem
• Basic model with one risky asset
– Assume an individual has wealth (W0) to
invest in one of two assets
– One asset yields a certain return of rf
– One asset’s return is a random variable, r
– k - the amount invested in the risky asset

77
 The Portfolio Problem
• The person’s wealth at the end of one period
W = (W0 – k)(1 + rf) + k(1 + r)

W = W0(1 + rf) + k(r – rf)

– W is now a random variable: it depends on r
– k can be positive or negative: can buy or sell
short
– k can be greater than W0: the investor could
borrow at the risk-free rate

78
 The Portfolio Problem
• U(W) - the investor’s utility function
– The von Neumann-Morgenstern theorem:
he will choose k to maximize E[U(W)]
• The first-order condition:
 
E U W  E U W0 1  rf   k  r  rf  
 
k k
 E U '  r  rf   0

79
 The Portfolio Problem
• As long as E(r – rf) > 0
– An investor will choose positive amounts of
the risky asset
• As risk aversion increases
– The amount of the risky asset held will fall
– The shape of the U’ function will change

80
 CARA utility
• The investor’s utility function - the CARA form:
U(W) = – exp (– AW)
– Marginal utility function: U’(W) = A exp(–AW)
– End-of-period wealth:
U’(W) = A exp[– A(W0(1+rf) + k(r – rf))] =
= A exp[–A(W0(1+rf)] exp[–Ak(r – rf)]
– Optimality condition:
E[U’·(r – rf)] = A exp[– AW0(1+rf)]
E[exp(– Ak(r – rf))·(r – rf)]=0
81
 CARA utility
• CARA function
– Implies that the fraction of wealth that an
investor holds in risky assets should decrease
as wealth increases
• CRRA form
– All individuals with the same risk tolerance
• Will hold the same fraction of wealth in risky assets
• Regardless of their absolute levels of wealth

82
 Portfolios of many risky assets
• Return on each of n risky assets
– The random variable ri (i = 1,…, n)
– Expected values: E(ri)=μi
– Variances: Var(ri) = σ2i
– An investor who invests a portion of his or her
wealth in a portfolio of these assets will obtain
n
a random return: n
rp    i ri where  i  0,  i 1
i 1 i 1

83
 Portfolios
• Expected ofonmany
return risky
this portfolio assets
n
E (rp )   p    i i
i 1

• If the returns of each asset are

independent
– The variance of the portfolio’s return:
n
Var (rp )      
2
p i
2
i
2

i 1

84
 Optimal portfolios
• Solving the optimal portfolio problem
– The first step: consider portfolios of just the
risky assets
– The second step: add in the riskless one
• Optimal portfolio of just the risky assets
– Choose a general set of asset weightings (the
i)
• Minimize the variance (or standard deviation)
– “Efficiency frontier’’ for risky asset portfolios
85
 Optimal portfolios
• Add a risk-free asset
– With expected return μf
– And standard deviation σf = 0
• Optimal portfolios
– ‘‘Market portfolio’’ consisting of all capital assets
held in proportion to their market valuations
• Expected return μm
• Standard deviation σm

86
 Optimal portfolios
m   f
• Mixed
 p portfolio
  f  (line RP) p
M
• Permits individual investors to ‘‘purchase’’
returns in excess of the risk-free return (μM-μf)
by taking on proportionally more risk (σp/σM)
• Points to the left of the market point M:
σp/σM<1 and μf < μp < μM
• High-risk points to the right of M: σp/σM>1 and
μp > μM
87
 E 7.1
•Efficient Portfolios

•The frontier EE represents

optimal mixtures of risky
assets that minimize the
standard deviation of the

σP, for each

portfolio,

expected return, μP. A risk-

free asset with return μ

f
offers investors the opportunity
to hold mixed portfolios along
RP that mix this risk-free asset
with the market portfolio, M.

88
 Individual choices
• Individuals with low tolerance for risk (I )
– Opt for portfolios that are heavily weighted
toward the risk-free asset
• Investors willing to assume a modest
degree of risk (II )
– Opt for portfolios close to the market portfolio
• High-risk investors (III )
– Opt for leveraged portfolios

89
 E 7.2
•Investor Behavior and Risk Aversion

•Given the market options

RP, investors can choose
how much risk they wish to
assume. Very risk-averse
investors (UI) will hold
mainly risk-free assets,
whereas risk takers (UIII) will
opt for leveraged portfolios.

90
 Mutual funds
• Mutual funds
– Pool the funds of many individuals
– Able to achieve economies of scale in
transactions and management costs
• Fund owners to share in the fortunes of a much wider
variety of equities
– Managers have incentives of their own
– Portfolios they hold may not always be perfect
representations of the risk attitudes of their
clients
91
 Capital asset pricing model
• Portfolio
– Small amount () of an asset with a random return
x
– Market portfolio, random return M
– Return on the portfolio: z = x+(1- )M
– Expected return:
μz = μx + (1- )μM
– Variance:

σ2z = 2σ2x + (1- )2σ2M + 2(1- ) σx,M

92
 Capital asset pricing model
z
 z   f  (M   f ) 
M
 z  M   f  z
  x  M  
 M 
 x,M
 x   f  (M   f )  2
M

93