Scribd
Upload
69 views11 pages

Unit-5 Index Tracking Complete PDF

This document discusses portfolio optimization through index tracking models. It begins by setting the market context, including return vectors for benchmarks and assets. It then discusses minimizing tracking error by finding the optimal portfolio weights that minimize variance between benchmark and portfolio returns. Finally, it formulates tracking error minimization as linear programming problems for different error metrics, such as mean absolute deviation, MinMax, mean absolute downside deviation, and downside MinMax. The linear programming formulations allow minimizing these error metrics by introducing variables for deviations between returns and constraints linking them to optimal portfolio weights.

Uploaded by

M Madan Gopal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
69 views11 pages

Unit-5 Index Tracking Complete PDF

This document discusses portfolio optimization through index tracking models. It begins by setting the market context, including return vectors for benchmarks and assets. It then discusses minimizing tracking error by finding the optimal portfolio weights that minimize variance between benchmark and portfolio returns. Finally, it formulates tracking error minimization as linear programming problems for different error metrics, such as mean absolute deviation, MinMax, mean absolute downside deviation, and downside MinMax. The linear programming formulations allow minimizing these error metrics by introducing variables for deviations between returns and constraints linking them to optimal portfolio weights.

Uploaded by

M Madan Gopal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 11

Portfolio Optimization

Unit 5: Index Tracking Model


Duan LI & Xiangyu Cui
India Institute of Technology Kharagpur
May 26 - 30, 2014

Market Setting

Outline

Market Setting
Tracking Error Minimization
Linear Programming Formulation

1 / 10

Market Setting

Market Setting

Let T be the number of observations and n be the number of


assets.

Y is a vector of continuously compounded benchmark returns,


which is a T dimensional vector.

X is the matrix of continuously compounded returns on n


assets, which is a T n matrix.

is the portfolio weights to be determined, which is an n


dimensional vector.

2 / 10

Tracking Error Minimization

Outline

Market Setting
Tracking Error Minimization
Linear Programming Formulation

3 / 10

Tracking Error Minimization

Minimization of Tracking Error Variance

The investor chooses the portfolio such that tacking error variance
is minimized,
min (Y X )0 (Y X ).
The optimal portfolio is :

= (X 0 X )

X 0Y .

4 / 10

Tracking Error Minimization

Minimization of Various Tracking Error


Functions
I

Mean absolute deviation (MAD):


min 10 (|X

MinMax:

min
I

max |Xt
t

Yt | .

Mean absolute downside deviation (MADD):


min 10 (min(X

Y |).

Downside MinMax (DMinMax):

min max (min(Xt


t

Y , 0)).

Yt , 0)) .

5 / 10

Linear Programming Formulation

Outline

Market Setting
Tracking Error Minimization
Linear Programming Formulation

6 / 10

Linear Programming Formulation

MAD
Let zt+ 0 be a positive deviation and zt
0 the absolute value
of a negative deviation between the portfolio and benchmark
returns. Then
Yt > 0 , Xt

Xt

zt = Yt ,

Yt < 0 , Xt + zt+ = Yt ,

Xt
which implies
min

zt+ ,

zt ,

s.t.

T
X

(zt+ + zt )

t=1

Xt + zt+
0

zt+

zt = Yt , t = 1, . . . , T ,

= 1,
0, zt

0, t = 1, . . . , T .
7 / 10

Linear Programming Formulation

MinMax
Let z

0 be an upper boundary of the absolute deviations:


Yt |, t = 1, . . . , T .

|Xt

z
We have
min z
z,

s.t. Xt

z Yt , t = 1, . . . , T ,

Xt + z
0

= 1,

0.

Yt , t = 1, . . . , T ,

8 / 10

Linear Programming Formulation

MADD
If investors are concerned about negative deviations between the
portfolio and the benchmark, the zt+ are dropped from the
optimization MAD, which results in
min

zt ,

T
X

zt

t=1

s.t. Xt + zt
0

zt

Yt , t = 1, . . . , T ,

= 1,
0, t = 1, . . . , T .

9 / 10

Linear Programming Formulation

DMinMax
Only the restrictions Xt Yt are relevant. Thus, we only care
about Xt + z Yt .
min z
z,

s.t. Xt + z
0

= 1,

0.

Yt , t = 1, . . . , T ,

10 / 10

You might also like