Beruflich Dokumente
Kultur Dokumente
2014851018
Returns for Securities
States
Probabilities
Security A Security B
Expected Returns
state 1
0,50
100%
0%
state 1
0,50X((*2)+((100-)*0))
State 2
0,49
95%
0%
State 2
0,49X((*1,95)+((100-)*0))
State 3
0,01
95%
5000%
State 3
0,01X((*1,95)+((100-)*5000))
Our Utility function is Log(w). Our goal is to maximize our utility. Here assume that we have
initial wealth of $100 all of which we can be invested. So utility of our current wealth is
Log(w0) which is Log(100) = 2. If we can manage to increase our utility over 2 through this
investment we go through with it f we cannot then we do not invest.
To find the optimal allocation we need to determine the value of for which
f ( )
reaches
maximum. To do this we take the derivative set it equal to zero and find , which is what we
invest in security A and the rest (100- ) is to security B.
f ( ) = 0,50*log((*2)+((100-)*0)) +0,49*log((*1,95)+((100-)*0))
+0,01*log((*1,95)+((100-)*5000)) =0
50 log 2
100
100
2
(1,95 )
+ 49 log
95,55
+ 1,95
f ' ( ) =
1
x
100
4998,05
+ 5000004998,05 )
4998,05
100
=
5000004998,05
2
95,55
+ 1,95
( )= 100.0003.
we do the math and becomes practically equal to 1 which means in order to maximize
(2)+((100 )0)
(1,95)+ ( ( 100 )0 )
A
A
A
0,50 e
f ' ( )=2 A0,50 eA ( 2 )+ 1,95 A0,49 e A (1,95 )=4998,050,01 eA (4998,05 )
For the sake of example assume that A=1 then equation becomes
2)
m=
U ' ( ct+1)
U ' (c 0)
Equation 1
Where stands for measure of patience, and U(c) represent marginal utility of consumption
ct +1
U ' stand for the marginal utility of consumption at future. It is called
at time 0,
marginal rate of substitution because what it does is it takes the ratio of marginal utility of a
next period consumption to todays marginal utility of consumption. It formulates that we
can make a trade off by consuming less today and investing a little bit more so we can
consume more in future with the return of previous investment made in time 0 by
consuming less, assuming marginal utility of future consumption is positive. So we compare
the marginal utilities to determine how much more we should save today and consume in the
future.
1
E (m)
Equation
2
With lognormal consumption growth and power utility. Rf =
Ct + 1
(
Ct
) when
people are less willing to save and consume more today interest rate is high because people
want more return to postpone the consumption. This relation works reverse, too. When
interest rates are high people save more today and spend more in future causing a growth in
consumption. So the Rf is really depend on marginal rate of substitution of future
consumption to todays consumption.
Xt +1
Rf
Equation 3
would become;
P=m=
U '(ct +1)
U ' (c 0)
Equation 4
But there is risk involved and to account for that risk we use m, which is uncertain (because
of unknown future consumption) where Xt+1 is the random payoff .
Our main purpose is to maximize utility function not the monetary wealth because same
amount of return measured by monetary value may provide different level of happiness at
different times. In difficult times 5% percent return for instance may be more valued than 6%
return of good times. Additionally the return or the asset that will give you that return is
more valuable when you feel poor and vice versa so the covariance of consumption with
return ( Xi) of an asset is what determines its price.
Pt =
E (Xi)
Rf
+ Cov (m, x)
Equation 5
First part of the equation is simple discount and second part is adjustment for risk. The asset
whose payoff covaries positively with consumption is more valuable today assume because it
means that you obtain that return in a time when you need more than any other times. You
get the payoff when it is valuable to you so that is positive covariance.
So the basic consumption based asset pricing model is then
P=E (mx)
Where P is the price and m is stochastic discount factor and x is the payoff. Further
decomposition is
Pi=
Equation 6
Risk premium:
If we divide both side of the equation with Pi ( price of the security ) we get
1=E( m) E(Ri)
Equation
7
Ri is the return for asset i. So 1=E(m) E( Ri)+Cov(m, Ri)
Equation 8
1=E(m) E [ Ri ] +
both sides by
Cov ( m, Ri )
E (m)
E(m) we get
Rf =( E [ Ri ] +
Cov ( m , Ri )
)
E (m)
Equation 9
Rearranging the Equation 9 we find
Ct , Rt + 1
U ' ()
Cov ( m, Ri )
E [ Ri ] =Rf +
= Cov
E (m)
Rf
Equation 10.
What this equation suggests is that when the covariance is negative between return of a
stock and consumption the investor can expect more. When consumption and returns are
both high marginal utility from consuming one more unit is not that satisfactory so the
E [ Ri ] is low but when the conditions are opposite for instance assume your house is
burned to the ground and you need money desperately so the marginal utility of
consumption coming from that investment is much more valuable. If in such timing the
returns are high there will be negative correlation and
investor should expect positive risk premium over Rf so that investor keep holding the stock
for higher return instead of cashing out and consuming. On the other hand when
Cov ( m , Ri ) is positive meaning when utility of marginal consumption is low and the
return is also low so investor are willing to hold theasset without requiring higher return
because marginal utility of consuming one more unit is not that pleasurable and
satisfactory at time t+1so it makes the risk premium lower.
Optimal Portfolios
18%
16%
14.00%
14%
12%
10%
E(r)
8%
7.00%
8.00%
11.00%
10.10%
10.00%
9.87%
6%
4%
2%
0%
0%
5%
10%
15%
20%
25%
30%
35%
This graph shows all the available portfolios Pi ( i= 1, 2, 7) to the investor laid out on a
E(r) and
p plane. Clearly some portfolios are inefficient due to lower return for the level
risk attached to them but on the other hand some portfolios provide highest possible return
for a lower level of risk we called them efficient portfolios and they are located on line
between 4 and 3 which is referred to as efficient frontier. Naturally portfolios P 4, P5, P6, and
P7 are not efficient because the return they offer can be achieved by taking lower risk. Our
objection is given as min E(R2P) minimize the portfolio variance while achieving below
constraints.
Subject to 1= and =p so the variance-minimized portfolios are located on the
efficient part of the frontier which is between P 2 and P3. and the rest are not applicable or
efficient because variance is not minimize with those portfolios. The portfolios on that curve
are called mean variance efficient portfolios because they have minimum variance for a given
level of return. Efficient portfolios form a parabola so the global minimum variance portfolio
can be found by
d 2
d
the portfolio that has the highest value offers the highest level of return for each unit of risk
being taken.
( 0C 0' p ) Rf +x
4.
( 0C 0)+ '(xRfp)
'
1 2
h( 0C 0+ ( xRfp ) )+ ah ' x
2
'
'
1 2
h ( 0C 0 + ( x Rfp ) )+ ah ' x
2
h ( xRfp ) +ah 2 xe
Since e cannot be zero
h ( xRfp ) +ah
=0
( xRfp )
h x
Additionally because the expected utility function monotonic in its exponent it can be rewritten as
1
'
max ( xRfp ) + h ' x
2
First order condition is
( xRfp ) +h x =0
( xRfp )
h x
solving for =
16%
14%
13.00%
13.00%
11.00%
11.30%
10.00%
9.50%
10.00%
9.00%
8.00%
12%
10%
E(r)
8%
7.00%
6%
4%
2%
0%
0%
5%
10%
15%
20%
25%
30%
35%
6. Mean variance analysis suggest that investor should construct a portfolio which provide
highest possible return for a minimum level of risk. Which is achieved by maximizing Sharpe
ratio
prf
p
rf
p is
Allocation Line) touches the efficient frontier Sharpe ratio is maximized. As shown by the
graph minimum variance portfolios that are a combination of market portfolio and risk free
asset provide better return for the same level of risk meaning they have higher Sharpe Ratio.
For example Portfolio A, which is combination of some risky asset and risk free asset provide
better return than P2 which has the same level of risk as B. The intuition behind is that it
measures the return or risk premium for every unit of risk an investor takes.
Deriving the Capital Market Line.
W m=
p
m
If we combine this with portfolio return formula above we get the formula for CML
(Rp)= Rf +
E ( R m )R f
m
E ( R m )R f
m
reward for investing beyond risk free asset and taking additional risk attached to the market
portfolio. Thus the slope of the tangent line measures the reward per unit of market risk,
which is maximized where the CML touches the efficient frontier, in above graph it is shown
by P1. So the investor by holding some combination of risk free asset and P 1 will maximize her
reward for taking the market risk and for every unit of market risk she bears she will hope to
get
E ( R m )R f
p
m
an individual asset to the market portfolio. So for example 1% of change in the reward for that
market risk we should expect to see % change in the reward of an individual asset because
is constant and it represents a stocks riskiness level relative to the market portfolio.