You are on page 1of 15

Preliminaries

Optimal portfolio allocation

Conclusion

Lecture 3: Risk and return


SAPM [Econ F412/FIN F313 ]

Ramana Sonti
BITS Pilani, Hyderabad Campus
Term II, 2014-15

1/15

Lecture 3: Risk and return

Ramana Sonti

Preliminaries

Optimal portfolio allocation

Conclusion

Agenda

1 Preliminaries

Introduction
2 Optimal portfolio allocation

The capital allocation line


Optimal portfolio
3 Conclusion

Differential lending and borrowing rates


Final thoughts

2/15

Lecture 3: Risk and return

Ramana Sonti

Preliminaries

Optimal portfolio allocation

Conclusion

Introduction

The basic problem


Investor has the following investment alternatives
1 risky asset with random returns: expected return E(r) and standard
deviation
1 risk-free asset with constant return

For a portfolio of these two assets (with weight w on the risky asset),
h
i
Portfolio expected return: E(rp ) = wE(r) + (1 w)rf = rf + w E(r) rf
Portfolio variance: p2 = w2 2
Portfolio standard deviation: p = w
Recall our earlier example: E(r) = 22%; rf = 5%; = 34.29%. A

60:40 portfolio of the two assets has


Expected return: E(rp ) = (0.6 22%) + (0.4 5%) = 15.2%
Standard deviation: (rp ) = 0.6 34.29% = 20.57%

Big Question
How should the investor optimally allocate his investment?

3/15

Lecture 3: Risk and return

Ramana Sonti

Preliminaries

Optimal portfolio allocation

Conclusion

Introduction

Investor with A = 3.0


Utility calculations
Asset
Risk free
Risky
60:40 portfolio

E(r)
0.05
0.22
0.152

(r)

Utility

0
0.3429
0.2057

0.05
0.22 (0.5)(3)(0.34292 ) = 0.0436
0.152 (0.5)(3)(0.20572 ) = 0.0885

Obviously, the 60:40 portfolio is better than either asset alone


But which is the optimal portfolio?

Algorithm
Step 1: Generate the feasible set of all possible combinations of the

risky and the risk-free assets


Step 2: Among all feasible combinations, choose the one with the

highest utility
4/15

Lecture 3: Risk and return

Ramana Sonti

Preliminaries

Optimal portfolio allocation

Conclusion

The capital allocation line

The feasible set of investment opportunities


Portfolio expected return:

h
i
E(rp ) = wE(r) + (1 w)rf = rf + w E(r) rf
Portfolio variance: p2 = w2 2
Putting together
we get
h these equations,
i
h E(r)r
i
f
E(rp ) = rf + w E(r) rf = rf +
p

i
h
0.17
For our example, E(rp ) = 0.05 + 0.3429 p
Thus, we see that the feasible set is represented by the equation of

a straight line
This feasible set is known as the capital allocation line (CAL)
i
h

The slope of the CAL is

E(r)rf

. This ratio is the expected excess

return per unit standard deviation


In other words, the slope represents the reward-to-risk ratio, and has a
special name: the Sharpe ratio

5/15

Lecture 3: Risk and return

Ramana Sonti

Preliminaries

Optimal portfolio allocation

Conclusion

The capital allocation line

Going up the CAL


0.4

0.35

CAL
w=147.06%

Portfolio expected return

0.3

0.25
w=100%
0.2
w=58.82%
0.15

0.1
w=0%
0.05

0
0

6/15

0.1

0.2

0.3
0.4
Portfolio standard deviation

Lecture 3: Risk and return

0.5

0.6

0.7

Ramana Sonti

Preliminaries

Optimal portfolio allocation

Conclusion

The capital allocation line

Lending and borrowing portfolios


Investor Lalita wants an expected return of 15%
Solve: w(22) + (1 w)5 = 15 w = 58.82%
Lalita starts with Rs. 100, invests Rs. 58.82 of her money in the

risky asset, and invests the remaining Rs. 41.18 in T-bills, i.e. lends
this money to the government
Portfolio standard deviation: w = 0.5882 34.29% = 20.17%

Investor Bharati wants an expected return of 30%


Solve: w(22) + (1 w)5 = 30 w = 147.06%
Bharati starts with Rs. 100, borrows Rs. 47.06 (at the risk-free rate),

and invests the sum Rs. 147.06 in the risky asset


Portfolio standard deviation: w = 1.4706 34.29% = 50.43%
7/15

Lecture 3: Risk and return

Ramana Sonti

Preliminaries

Optimal portfolio allocation

Conclusion

The capital allocation line

Lending and borrowing on the CAL


0.4

0.35

CAL
w=147.06%

Portfolio expected return

0.3

0.25

0.2
w=58.82%
0.15

0.1

0.05

0
0

8/15

0.1

0.2

0.3
0.4
Portfolio standard deviation

Lecture 3: Risk and return

0.5

0.6

0.7

Ramana Sonti

Preliminaries

Optimal portfolio allocation

Conclusion

Optimal portfolio

The optimal allocation


Mathematics
Since we can calculate the utility of a particular investor (fixed A) for different values of
w, an obvious way to get the optimal portfolio is to maximize the utility function over w
1
= E(rp ) Ap2
w
2
i
h
s.t. E(rp ) = rf + w E(r) rf and p2 = w2 2

The problem is: max U(w)

Standard calculus solution sets the first derivative of the utility with respect to w to
zero:

h
i 1
E(r) rf
= E(r) rf A(2w ) 2 = 0 w =
w
2
A 2

For our example investor (with A = 3), we have: w =

0.220.05
3(0.34292 )

= 0.4819

Intuition: Optimal allocation to risky asset increases as


expected return on risky increases
risk of risky asset decreases
investor risk aversion decreases
9/15

Lecture 3: Risk and return

Ramana Sonti

Preliminaries

Optimal portfolio allocation

Conclusion

Optimal portfolio

The highest utility allocation


0.2

0.1

Optimum w *=0.4819, Optimum U *=0.09096

Utility

-0.1

-0.2

-0.3
0

10/15

0.5

1
Risky asset weight

Lecture 3: Risk and return

1.5

Ramana Sonti

Preliminaries

Optimal portfolio allocation

Conclusion

Optimal portfolio

Analysis using indifference curves


0.4

0.35

Optimal U=0.09096

Portfolio expected return

0.3

U=0.12

CAL

U=0.05

0.25

0.2

0.15

0.1

0.05

0
0

11/15

0.1

0.2

0.3
0.4
Portfolio standard deviation

Lecture 3: Risk and return

0.5

0.6

0.7

Ramana Sonti

Preliminaries

Optimal portfolio allocation

Conclusion

Optimal portfolio

Lalita and Bharati again


0.4

0.35

CAL

0.3
Portfolio expected return

Bharati
0.25

0.2

0.15
Lalita
0.1

0.05

0
0

12/15

0.1

0.2

0.3
0.4
Portfolio standard deviation

Lecture 3: Risk and return

0.5

0.6

0.7

Ramana Sonti

Preliminaries

Optimal portfolio allocation

Conclusion

Differential lending and borrowing rates

Differential rates
Can we usually borrow at the risk-free rate?
What if the borrowing rate were some rb greater than rf ?
Note that hwith a higher
rate, the reward-to-risk ratio of the new CAL
i
E(r)r

b
becomes
, lower than before

This leads to a kinked CAL, instead of the usual straight one


b
The optimal allocation becomes w = E(r)r
, less aggressive than
A 2
before

Allocation algorithm
Step 1: Calculate the formula w =

E(r)rf
A 2

Step 2: If answer from Step 1 is greater than 1, recalculate using the

revised formula w =

13/15

E(r)rb
A 2

Lecture 3: Risk and return

Ramana Sonti

Preliminaries

Optimal portfolio allocation

Conclusion

Differential lending and borrowing rates

With a kinked CAL, allocation becomes less aggressive


0.4

0.35

Portfolio expected return

0.3

New CAL

0.25

Bharati

0.2

0.15
Lalita
0.1

0.05

0
0

14/15

0.1

0.2

0.3
0.4
Portfolio standard deviation

Lecture 3: Risk and return

0.5

0.6

0.7

Ramana Sonti

Preliminaries

Optimal portfolio allocation

Conclusion

Final thoughts

Final thoughts

So far, we have assumed only one risky asset: how general is this?
Actually quite so; remember that the sole risky asset we looked at

could be a portfolio of assets


In fact, we will derive the optimal risky portfolio in the next lecture

15/15

Lecture 3: Risk and return

Ramana Sonti

You might also like