Professional Documents
Culture Documents
Lecture Notes
Introduction:
Why do we need a course in financial
econometrics?
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Syllabus: Motivation
The past few decades have been characterized by an
extraordinary growth in the use of quantitative methods and
analysis in financial markets for various asset classes; be it
equities, fixed income instruments, commodities, or derivative
securities.
Financial market participants, both academics and practitioners,
have been routinely using advanced econometric techniques in a
host of applications including portfolio management, risk
management, volatility modeling, and interest rate modeling, for
understanding pivotal issues in corporate finance and asset
pricing, as well as regulatory purposes, and more.
3
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Syllabus: Objectives
This course attempts to provide a fairly deep understanding of
topical issues in asset pricing and deliver econometric methods.
More specifically, the purpose is twofold, both to develop
research agenda in financial economics and to comprehend
complex investment designs employed by practitioners.
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Syllabus: Prerequisite #1
I will assume prior exposure to matrix algebra, distribution
theory, Ordinary Least Squares, Maximum Likelihood
Estimation, Method of Moments, and the Delta Method.
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Syllabus: Prerequisite #2
I will also assume you have some skills in computer
programing beyond Excel.
MATLAB and R are the most recommended for this course. OCTAVE
could be used as well, as it is a free software, and is practically
identical to MATLAB when considering the scope of the course.
If you desire to use STATA, SAS, or other comparable tools, please
consult with the TA.
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
10
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
11
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
12
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Session #1 Overview
13
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Let us Start
This session is mostly an overview. Major contents:
Why do we need a course in financial econometrics?
14
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Financial Econometrics
In previous courses in finance and economics you had mastered
the concept of the efficient frontier.
A portfolio lying on the frontier is the highest expected return
portfolio for a given volatility target.
Or it is the lowest volatility portfolio for a given expected
return target.
15
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
16
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
More generally,
if there are N investable assets, you need:
N estimates for means,
N estimates for volatilities,
0.5N(N-1) estimates for correlations.
Overall: 2N+0.5N(N-1) estimates are required!
18
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
19
=1
Less Noise =
=1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Estimation Methods
One of the ideas here is to introduce robust methods in which
to estimate the comprehensive set of parameters.
We will discuss asset pricing models and the Black-Litterman
approach for estimating expected returns.
We will further introduce several methods for estimating the
large-scale covariance matrix of asset returns.
20
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
21
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
22
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
23
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Drawback in Mean-Variance
A
B
S1
S2
S3
S4
S5
5.00%
3.05%
8.25%
2.90%
15.00%
2.95%
10.50%
3.10%
20.05%
3.01%
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Note that =
1
2 2
, and specifically
if = 1, then =
1
2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
28
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
29
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Confidence Level
Normality suggests that deviation of 2 SD away from the mean
creates an 80% range (from -32% to 48%) for the realized return
with approx. 95% confidence level (assuming ~(8%, (20%)2 )
30
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
31
0.08
0.2
<
00.08
0.2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Higher Moments
Skewness the third moment - is zero.
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Skewness
The skewness can be negative (left tail) or positive (right tail).
33
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Kurtosis
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
35
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2 ,
, , 2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Conditional Distribution
Conditional distribution:
2
2
~ + 2 , 2
37
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Conditional Moments
If we have no information about , or if and are
uncorrelated, then the conditional and unconditional expected
return of are identical.
Otherwise, the expected return of y is x .
If = 0, then: =
38
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2 1
2 2
1 2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Multivariate Normal
1
1
2
= . ~ 1 ,
.
Define Z=AX+B.
40
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Multivariate Normal
Let us make some transformations to end up with (0, ):
= 1 + 1 1 + 1 ,
1 + 1 0,
41
1
2
1 1 0,
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Multivariate Normal
Consider an N-vector or returns which are normally
distributed:
1 ~ 1 ,
Then:
~ 0,
1
2
~ 0,
1
What does
1
2
1
2
2
42
1
2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
43
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
44
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The F Distribution
Gibbons, Ross & Shanken (GRS) designated a finite sample asset
pricing test that has the F - Distribution.
If
1 ~ 2
2 ~ 2
1 2
then =
45
~ ,
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The t Distribution
As noted earlier, 1 , 2 , , is a sample of length T of stock
returns which are normally distributed.
The sample mean and variance are
1 +
=
1
=
1
=1
The resulting t-value which has the t distribution with T-1 d.o.f
is
=
1
46
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The t Distribution
The pdf of student-t is given by:
+1
2
2
1
, =
1+
2,1 2
47
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
48
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The t Distribution
The t-distribution is the sampling distribution of the t-value
when the sample consist of independently and identically
distributed observations from a normally distributed
population.
It is obtained by dividing a normally distributed random
variable by a square root of a Chi-squared distributed random
variable when both random variables are independent.
Indeed, later we will show that when returns are normally
distributed the sample mean and variance are independent.
49
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Regression Analysis
Various applications in corporate finance and asset pricing
require the implementation of a regression analysis.
We will estimate regressions using matrix notation.
For instance, consider the time series predictive regression
= + 1 1,1 + 2 2,1 +
50
= 1,2, ,
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1
2
= .
.
1, 1,0 , 2,0
1, 1,1 , 2,1
.
=
31 = 1
.
2
1, 1,1 , 2,1
1
2
1 = .
.
yields
R= +
We will derive regression coefficients and their standard errors
using OLS, MLE, and Method of Moments.
51
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
11
12
= .
.
1
52
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Multiplication of Matrices
22
22 22
11, 12
=
21, 22
22
11, 12
=
21, 22
11 11 + 12 21, 11 12 + 12 22
=
21 11 + 22 21, 21 12 + 22 22
22
,
11, 21
=
12, 22
() =
()1 = 1 1
53
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
54
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1 1
>0
56
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1,2 2 = 1 , ,
=
Let
1
1
,1, 2
Trace of a matrix is the sum of diagonal elements, therefore:
= 12 + 22 + + 2
57
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Matrix Vectorization
58
1
2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
+1
1
2
59
1 1
22
=
2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Partitioned Matrices
A, a square nonsingular matrix, is partitioned as
11 , 12
1 1 1 2
=
21 , 22
2 1 2 2
11
12 22 21
11
12 22 21
12 22
1 =
1
1 , ( 1 )1
1
12 22 21
22
21 11 12
22 21 11
Check that 1 =
60
1 +2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Matrix Differentiation
Y is an M-vector. X is an N-vector. Then
1 1
1
,
,,
1 2
,
,,
1 2
And specifically, if
=
Then:
61
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Matrix Differentiation
Let
1 1
62
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Matrix Differentiation
Let
If C is symmetric, then
63
=
1 1
= +
= 2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Kronecker Product
It is given by
=
64
11 , 1
=
1 ,
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Kronecker Product
For A,B square matrices, it follows that
= 1 1
65
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
( + ) = () + ()
() = ( ) ()
( ) = ( )
= ( ) = ( ) = + ()
where
= (), = ( )( )
66
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
68
1
= 2
3
31
1
= 2
3
12 , 12 , 13
= 21 , 22 , 23
33
31 , 32 , 32
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
69
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Paul Samuelson shows the relation between the LIE and the
notion of market efficiency which loosely speaking asserts
that the change in asset prices cannot be predicted using
current information.
70
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
[+1 | ] = +1 [ ] [ ] = 0
The quantity does not depend on the information.
71
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
+ |
+1
1+
+2
2
1+
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Variance Decomposition
=
>
73
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
74
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
TA in Finance
Several major applications in finance require the use of Taylor
series approximation.
The following table describes three applications.
75
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
76
Expected Utility
Bond Pricing
Option Pricing
First Order
Mean
Duration
Delta
Second Order
Volatility
Convexity
Gamma
Third Order
Skewness
Fourth Order
Kurtosis
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Taylor Approximation
Taylor series is a representation of a function as an infinite sum of
terms that are calculated from the values of the function's
derivatives at a single point.
Taylor approximation is written as:
1
= 0 + 0 0 +
1!
1
1
2
+ 0 0 + 0 0
2!
3!
It can also be written with
() =
77
notation:
( )
0
0
!
=0
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1 + = 1 + +1 1 + +2 1 + +
Applying ln on both sides of the equation:
ln 1 + = ln 1 + +1 1 + +2 1 + + =
= ln 1 + +1 + ln 1 + +2 + + ln 1 + +
78
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
+ = + = exp ( ) = exp ( )
80
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
81
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
exp ( ) +
exp ( )( ) +
1
1
2
2
exp ( )( ) +
exp ( )( )3 3 +
2
6
1
exp ( )( )4 4 +
24
And the expected value of (+ ) is:
1
[(+ )]
exp ( )[1 + () 2 +
2
1
1
3
() + () 4 ]
6
24
82
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Expected Utility
Let us assume that log return is normally distributed:
~ , 2
= 0, = 3 4
then:
2 2 4 4
exp ( ) 1 + +
2
8
2 2
+ =
exp +
2
Are solutions compatible?
83
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
=1 1+
0
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1 0
1
1
0 1 0 + 0 1 0
1!
2!
0 + 12 0
85
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
0 0
0
2 0
86
Instead of:
0
0
Instead of:
0
0
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2
+
2
What if the yield to maturity falls?
87
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2
= +
2
According to duration - bond price should increase.
According to convexity - bond price should also increase.
Indeed, the bond price rises.
88
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2
= +
89
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
90
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Option Pricing
A call price is a function of the underlying asset price.
What is the change in the call price when the underlying asset
pricing changes?
1 2
2
( ) ( ) +
+
2 2
1
( ) + + 2
2
Focusing on the first order term this establishes the delta
neutral trading strategy.
91
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
92
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
93
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Different writing:
1, 11 , , 1
, = ,
1, 1 , ,
1
= .
= +
Or:
V=
Define a function
of the squares of the errors, that we want to minimize:
2 = = = + 2
=
=1
94
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= ( )1
=1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
96
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Reminder:
= ( )1
97
= ( )1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= ( )1 ( )1 2 = ( )1 2
= ( )1 2
1
where is the number of explanatory variables in the regression specification.
2
98
( )1
=
1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
defined as = 2 .
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Implementing MLE
Assume that
~ (, 2 )
Let us estimate and 2 using MLE; then derive the joint
distribution of these estimates.
Under normality, the probability distribution function (pdf) of
the rate of return takes the form
1
1 2
=
exp
2
2
2
2
100
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2 2 2
1
exp
2
=1
1
2
ln = ln 2 ln
2
2
2
101
=1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
=1
=0
2
1
= 2+
2
2
2
=1
1
=
=1
=0
1
=
=1
102
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2
=
2
103
=1
=1
= 2
2
= 2
2
=0
2
= 4
2
=1
= 4
2
2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= ()1
Or put differently, the asymptotic distribution of is:
0, ()1
[The proof is beyond the scope of this course]
In our context, the covariance matrix is derived from the information matrix as follows:
2
2 , 0 "1" 2 , 0
,0
2 4
0, 4
0, 4
0,
2
2
104
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
To Summarize
Multiply by and get:
0, ()1
105
=1
=1
1
(
)2
0 2, 0
,
0 0,2 4
=1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
106
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
108
2 2
= 0
=1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Method of Moments
Continue estimation:
1
= 0 = 1, ,
1
=
=1
2 = 0
=1
2
109
1
=
=1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
MOM: Stage 3
=
2 ,
2 2
+ 4
2,
3
=
3 , 4 4
110
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
MOM: stage 4
Memo:
=
2 2
111
1,
=
2 ,
0
1,
=
1
0,
0
1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= 1
112
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
,
3
0
1
2
=
=
4 ,
0, 2
3 , 4 4
3 = = 3
(sk is the skewness of the standardized return)
4 = = 4
(kr is the kurtosis of the standardized return)
Sample estimates of the Skewness and Kurtosis are
1
3 =
113
=1
1
4 =
=1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Under Normality
the MOM Covariance Matrix Boils Down to
2
114
2,
=
3 = 0,
3 = 0
=
4
4
4 = 2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
=
= 0
115
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
=1
= 0
=1
=
=1
2
116
1,
=
1,
=1
1
=
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1
=
Stage 4:
1
=
117
()
()
=1
=1
()
1
=
1
=
=1
( )2
=1
( ) =
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
( )2 ( )1
= ( )1
=1
118
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
119
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Overview
A short brief of the major contents for todays class:
Hypothesis testing
120
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Hypothesis Testing
Let us assume that a mutual fund invests in value stocks (e.g.,
stocks with high ratios of book-to-market).
Performance evaluation is mostly about running the regression
of excess fund returns on the market premium
= +
+
H1: Otherwise
121
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Hypothesis Testing
Errors emerge if we reject H0 while it is true, or when we do
not reject H0 when it is wrong:
Dont reject H0
Reject H0
H0
Good decision
Type 1 error
H1
Type 2 error
Good decision
122
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
124
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Test I Skewness
TEST 1 - Skewness (third moment)
The setup for testing normality of stock return:
H0: , 2
H1: otherwise
Sample Skewness is
1
=
125
=1
6
0,
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Test I Skewness
Multiplying by
,
6
we get
~
6
0,1
126
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
TEST 2 - Kurtosis
Kurtosis estimate is:
1
=
=1
24
3,
After transformation:
0
3 0,1
24
127
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2
= +
3
6
24
2 2
Why 2 (2) ?
128
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
12 + 22 2 2
and
24
129
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
0,
Then: =
130
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
There are a few tests based on time series and cross section
specifications. Let us start with the time series.
)
The CAPM says that: ( ) = (
= +
131
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
If: ,
0
and 0,
We get:
0
1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= 0 + 1 1 + 2 2 + , = 1,2, ,
134
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1, 1 , 2
1 = 3 31 + 1
135
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
, 1 2
= 1
Joint hypothesis testing:
0 :
1 :
136
0 = 1, 2 = 0
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1 =
=
0,0,1
2
0
2
1 :
137
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= 0
,
H0
Under H0: ~ 0,
Chi squared:
~ 2 2
138
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1
= . ,6
1, 11 , , 51
61 = 0 , 1 , 6
1
.
=
,
.
1 = .
1, 1 , , 5
= +
139
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1
1
1
0 = 1, 1 = , 2 = , 3 = , 5 = 7
2
3
4
0 :
1 :
Here is a receipt:
1. =
2. =
3.
4. ,
140
1 2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1
1
2
1
3
1
4
6. 0 : = 0
7. ,
0
8. 0,
9.
141
0 2 5
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
142
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Shrinkage Methods
One of the problems with the tests we have just displayed is
that the decision is binary: Reject or do not reject (Classical
econometrics) the null.
The possibility of partly rejecting/accepting the null, for
example, does not exist.
143
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
144
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
50% CAPM : ( ) = (
)
+ = +
50% Data: 1 = 1 + 1
1
1
1
=0
145
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
146
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
MLE
0,
148
MOM
0, 2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
149
0,
0,
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2
We will show that 0,
2
We use the Delta method to derive
150
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
+
12
21
11
=
12
21
151
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
152
=0
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
153
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2 0.5
1
=
1 2 0.5
2
=
=
2
2
2 3
154
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1
=
,
2 3
2,
0,
0
2 4
2 3
1 2
0, 1 +
2
155
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
156
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Hypothesis Testing
Does the S&P index outperform the Rf?
H0: SR=0
H1: Otherwise
Under the null there is no outperformance.
(0,1 + 122 )
Thus, under the null
1
2
1+ 2
157
(0,1)
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
158
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
159
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The cross sectional tests apply to both portfolio and nonportfolio based factors.
Consumption growth in the CCAPM is a good example of a
factor which is not a return spread.
160
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
161
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1
= 1 + 1
+ 1
= +
+
1 = 2 = =
The CAPM says: H0: 1 = 2 = =
H1: Otherwise
162
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
164
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1
= 1 + 11 1 + 12 2 + + 1 + 1
= + 1 1 + 2 2 + + +
H0: 1 = 2 = =
H1: Otherwise
In the multi-factor context, the hypothesis is that some
particular combination of the factors is the tangency portfolio.
165
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
=
= =
166
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2 , , 2
= =
, 2
167
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Creating a Portfolio
1
A portfolio is investing = is the N assets.
1
= 1 1 + 2 2 + + =
= 1 1 + 2 2 + + =
168
=1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Creating a Portfolio
The variance of the portfolio is:
2 = = 12 12 + 1 2 12 + 1 1
+1 2 12 + 22 22 + + 2 2
+
+
+1 1 + 2 2 + + 2 2
169
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Creating a Portfolio
Thus 2 =
=1 =1
where is the coefficient of correlation.
170
1 1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
We know: 2 = = 12 12 + 21 2 12 + 22 22
Let us check:
2
= 1 , 2
12 , 12
12 , 22
1
2 2
2 2
=
+
2
1 2 12
1 1
2 2
2
So it works!
171
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
172
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1
where is the Greek letter iota 1 = ,
1
and where is the expected return target set by the investor.
173
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= 1 2 = 0
= 1 1 + 2 1
174
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1
= 1 1
175
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
176
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1
2
= = = = 1
1 1
1
1
= = = 0
Indeed, = 1.
177
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Recall:
= +
= +
So we need to show = 0
= 1
178
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1 1
1
1
= = = 0
179
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Efficient Frontier
The optimization program also delivers the shape of the frontier.
180
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Efficient Frontier
Where point A stands for the Global Minimum Variance
Portfolio (GMVP).
The efficient frontier reflects the investment opportunities; this
is the supply side.
Points below A are inefficient since they are being dominated
by other more attractive portfolios.
181
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
182
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Inflation risk.
Interest rate risk and the horizon effect.
183
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= 1
= 1
The tangency portfolio is investing all the funds in risky assets.
184
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Fund Separation
Interestingly, all investors in the economy will mix the
tangency portfolio with a riskfree asset.
The mix depends on preferences.
But the proportion of risky assets will be equal across the
board.
One way to test the CAPM is indeed to examine whether all
investors hold the same proportions of risky assets.
Obviously they dont!
186
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Equilibrium
The efficient frontier reflects
the supply side.
What about the demand?
The demand side can be
represented by a set of
indifference curves.
187
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Equilibrium
What is the slope of indifference curve positive?
The equilibrium obtains when the indifference curve tangents
the efficient frontier
No riskfree asset:
188
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Equilibrium
With riskfree asset:
189
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
190
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
191
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Utility Maximization
= +
1
2
= = 0
1 1
= 1
where reflects the fraction of weights invested in risky assets.
The rest is invested in the riskfree asset.
192
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1 1
=
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= exp > 0
= 2 exp < 0
194
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
195
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
max
+1 |
. .
+1 = 1 + + +1
196
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Then +1 1 + + , 2
197
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
198
1 2 2
= exp +
2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1 2
exp +1 = exp +1 + +1
2
1 2 2
= exp 1 + + +
2
= exp 1 +
199
1
exp
2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
200
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1 1
.
= 1
201
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
202
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Quadratic Preferences
The quadratic utility function is of the form
=+
2
2
where > 0
= 1
2
= < 0
2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Quadratic Preferences
The utility function looks like
204
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Quadratic Preferences
It has a diminishing part which makes no sense because we
always prefer higher than lower wealth
Utility is thus restricted to the positive slope part
Notice that = =
. . +1 = 1 + + +1
205
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Quadratic Preferences
Avramov and Chordia (2006 JFE) show that the optimization
could be formulated as
1
1
max
+
=
1
1+
206
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Quadratic Preferences
The tangency portfolio is
= 1
207
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Notice that
is actually the squared Sharpe Ratio of
the tangency portfolio.
Let us prove it
1
= 1
= + 1 = = 1
1
= = 1 2
208
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
209
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
210
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
211
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
212
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
213
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
214
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
215
1
= 1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2 Asset Allocation
Under the CAPM, the vector of expected returns and the
covariance matrix are given by:
=
2
=
+
where is the covariance matrix of the residuals in the timeseries asset pricing regression.
We denoted by the residual covariance matrix in the case
wherein the off diagonal elements are zeroed out.
216
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2 Asset Allocation
The corresponding quantities under the FF model are
+
=
+
=
where
217
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
3 Discount Factors
Expected return is the discount factor, commonly denoted by k,
in present value formulas in general and firm evaluation in
particular:
=
=1
1+
218
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
3 Discount Factors
Indeed, thus far we have examined models with constant beta
and constant risk premiums
=
where is a K-vector of risk premiums.
When factors are return spreads the risk premium is the mean
of the factor.
Later we will consider models with time varying factor
loadings.
219
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
4 Benchmarks
Factors in asset pricing models serve as benchmarks for
evaluating performance of active investments.
In particular, performance is the intercept (alpha) in the time
series regression of excess fund returns on a set of benchmarks
(typically four benchmarks in mutual funds and more so in
hedge funds):
= + ,
+
+ + +
220
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
5 Corporate Finance
There is a plethora of studies in corporate finance that use
asset pricing models to risk adjust asset returns.
Here are several examples:
Examining the long run performance of IPO firm.
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
5 Corporate Finance
Analyzing mergers and acquisitions
Analyzing the impact of change in board of directors.
Studying the impact of corporate governance on the cross section of
average returns.
Studying the long run impact of stock/bond repurchase.
222
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1
= 1 + 1
+ 1
= +
+
The null hypothesis is:
0 : 1 = 2 = = = 0
225
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Likelihood Ratio.
GRS (Gibbons, Ross, and Shanken (1989)).
GMM.
226
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The Distribution of .
Recall, is asset mispricing.
=
+
+
1
1 11
~ 0,
for = 1,2,3, ,
Let = , ,
227
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The Distribution of .
Under normality, the likelihood function for is
1
1
2
1
| =
exp
2
where is the constant of integration (recall the integral of a
probability distribution function is unity).
228
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The Distribution of .
Moreover, the IID assumption suggests that
1 , 2 , , | =
1
exp
2
=1
ln ln
2
2 =1
229
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The Distribution of .
Asymptotically, we have 0, ()
where
2
230
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The Distribution of .
Let us estimate the parameters
=
=1
1
231
=1
1
1 +
2
=1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The Distribution of .
Solving for the first order conditions yields
232
=1
=1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The Distribution of .
Moreover,
1
=
233
=1
1
=
1
=
=1
=1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The Distribution of .
Recall our objective is to find the standard errors of .
234
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The Distribution of .
The information matrix is constructed as follows
2
2
=
2
235
2
,
2
,
2
,
2
,
2
,
2
,
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
236
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The Distribution of .
We get:
1
, 1 +
Moreover,
1 1
, 2
2,
237
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The Distribution of .
Notice that (, ) stands for the Wishart distribution with
= 2 degrees of freedom and a parameter matrix = .
238
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
then
1 2
Here we test
0 : = 0
0
where ~ 0,
1 : 0
The Wald statistic is 1 2
239
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2 1
1 =
2
1 +
240
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1, 1
221
1
1
2 =
where
1,
= +
221
1
= ,
1
1 = 1 + 1
241
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
and = 1 , ,
242
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
243
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
un: = +
+
0,
res: =
+
0,
244
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1
=
=1
2
=1
=1
1
1
,
2
2
+
1,
245
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
246
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The LR Test
= ln
ln = ln ln
2
2 = 2 = ln ln 2
Thus,
1
2 = ln
+1
247
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
GRS (1989)
Theorem: let
~ 0,
let A ~ , where
1
248
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
GRS (1989)
In our context:
= 1+
2 2
0,
= ,
where
=2
Then:
1
3 =
1+
2 1
1 , 1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
GMM
I will directly give the statistic without derivation:
4 =
1 1 1
2 ()
where
, 0
1,
=
,
250
2
2
+
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
GMM
Assume no serial correlation but heteroskedasticity:
1
=
where
=1
= 1,
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1 = 1 +
1 1
1
3 =
1 + 1
1 1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
, ,
,
,
,
,
=
= , ,
2
, , ,,
253
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
254
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
where 0
255
=1
= +
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
256
1
2
2
2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
257
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
258
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
259
1 2
1 2 2
= = + 2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= |
=
=
+
1 2
|
+
2
2 + 2
where
260
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2
2 + 2
2
2
=
=
261
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Example
Let
2 = 0.01
= 0.05
2 = 0.3
For what values of 0 it is sill preferred to use the CAPM?
262
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Example
263
2
2
=
+
<1
1
1
2
0.05 0.7 + 0.3 +
<1
1
0.01
60
1
2
<
0.01 0.665
60
< 0.01528 = 1.528%
2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
264
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
265
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1 1
3 =
2
1 +
266
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
267
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
268
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Consider an investment universe that consists of + 1 assets the test assets as well as the market portfolio.
The expected return vector of the + 1 assets is given by
= ,
+1 1
11 1
where
is the estimated expected excess return on the
market portfolio and is the estimated expected excess return
on the test assets.
269
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
+1 +1
2,
=
2,
where
2 is the estimated variance of the market factor.
270
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= 1
271
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1 =
+ 1 ,
1 ,
1
1
272
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2
2
=
+ 1
or
2
2
1 =
273
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
274
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The test statistic examines how close the two sample Sharpe
ratios are.
Under the CAPM, the extra N test assets do not add anything
to improving the risk return tradeoff.
The geometric description of 1 is given in the next slide.
275
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2
1
1 = 12 22
276
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2
1 =
~
2
1 +
2
2
1
3 =
~ , 1
2
1 +
277
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
278
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
279
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
280
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Conditional Models
Here is a conditional asset pricing specification:
= +
+
= 0 + 1 1
= + 1 +
(
|1 ) =
cov ( , ) = 0
281
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= + 0
+ 1
1 +
Interestingly, the one factor conditional CAPM becomes a two
factor unconditional model the first factor is the market
portfolio, while the second is the interaction of the market with
the lagged variable.
282
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
283
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Conditional Moments
Suppose you are at time what is the discount factor for time
+ 2?
+2
= + 0 +2
+ 1 +2
+1 + +2
= + 0 +2
+ 1 +2
+ + +1 + +2
= + 0 +2
+ 1 +2
+ 1 +2
+ 1 +2
+1 + +2
284
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Conditional Moments
|
+2
= + 0
+ 1
+ 1
+
= + 0 + 1
1
285
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Conditional Moments
Could you derive a general formula in particular you are at
time what is the expected return for time + as a function of
the model parameters as well as ?
286
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Conditional Moments
Next, the conditional covariance matrix the covariance at
time + 1 given is given by
|
2 +
+1
= 0 + 1 0 + 1
287
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Conditional Moments
Could you derive a general formula in particular you are at
time what is the conditional covariance matrix for time +
as a function of the model parameters as well as ?
Could you derive general expressions for the conditional
moments of cumulative return?
288
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
289
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
290
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= + 1 +
291
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
292
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= + ,1 +
293
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
295
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
296
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
GMVP
Of particular interest to academics and practitioners is the
Global Minimum Volatility Portfolio.
For two distinct reasons:
1. No need to estimate the notoriously difficult to estimate .
2. Low volatility stocks have been found to outperform high volatility
stocks.
297
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
GMVP Optimization
= 1
Solution: =
1
1
298
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
GMVP Optimization
Ex ante, the GMVP is the lowest volatility portfolio among all
efficient portfolios.
Ex ante, it is also the lowest mean portfolio, but ex post it
performs reasonably well in delivering high payoffs. That is
related to the volatility anomaly: low volatility stocks deliver
higher payoffs than high volatility stocks.
299
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
300
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2 = = 2 + 2 + 2
304
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
. .
305
= 0
=
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
where
306
1 =
1 =
1 =
2 / =
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Questions
Is the sum of the components equal to zero?
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
. .
= 0
=
( + ) ( + ) = 2
310
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
311
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
312
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
314
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
315
, =
, =
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
316
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
317
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
11
+1
+1
where = ,
= ,
319
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= ,
320
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
4. Let be a diagonal matrix with the (i,i)-th component being
equal to the (i,i)-th component of .
321
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1 1
322
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
323
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
A Shrinkage Approach
Based on a Paper by Ledoit and Wolf (LW)
There is a well perceived paper (among Wall Street quants) by
LW demonstrating an alternative approach to estimating the
covariance matrix.
It had been claimed to deliver superior performance in
reducing tracking errors relative to benchmarks as well as
producing higher Sharpe ratios.
Here are the formal details.
324
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
A Shrinkage Approach
Based on a Paper by Ledoit and Wolf (LW)
Let be the sample covariance matrix, let be the equal
correlation based covariance matrix, and let be the shrinkage
intensity. and were derived earlier.
The operational shrinkage estimator of the covariance matrix
is given by = + (1 )
325
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= max 0, min
,1
2
and where =
(
)
with being the (, )
=1 =1
component of and is the (, ) component of .
326
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
=1 =1
1
=
{( )(
=1
+
=1
1
=
1
=
=1 =1,
_
) }2
_ 2
)
)2
, +
,
_
) (
_
) (
=1
_
)
_
)
=1
and with and being the time return on asset and the time-series
average of return on asset , respectively.
327
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
328
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Backtesting
We have proposed several methods for estimating the
covariance matrix.
Which one dominates?
We can backtest all specifications.
That is, we can run a horse race across the various models
recursive schemes.
329
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
330
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
332
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
,+4 = +4
,+5 = +5
,+6 = +6
Then at time t+6 you re-derive the GMVPs.
333
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
That is, you first estimate the GMVPs based on the first 60
observations, then based on 66 observations, and so on, while
in the rolling scheme you always use the last 60 observations.
Pros: the recursive scheme uses more observations.
Cons: since the covariance matrix may be time varying perhaps
you better drop initial observations.
334
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
335
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
MKT
FF
Recursive Scheme
FF+
MOM
LW
MKT
FF
FF+
MOM
Mean
STD
SR
SP
(5%)
alpha
IR
336
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
LW
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2
2.
=
,
and
=1
=1
We have =
339
and
2 4
.
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
and =
340
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2
.
= =
2 2
2
.
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2 4
.
22 4
1 2 4
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
343
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
PCA
The aim is to extract common factors to summarize the
information of a panel of rank .
In particular, we have a panel of stock returns where is
the time dimension and (< ) is the number of firms of
course <<
= 1 , ,
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
PCA
You have returns on stocks for periods
1 , ,
1 = 1 1 =
let
= 1 , 2 , ,
1
=
345
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
PCA
Extract K-eigen vectors corresponding to the largest K-eigen
values.
Each of the eigen vector is an N by 1 vector.
The extraction mechanism is as follows.
346
1 1 = 1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
PCA
1 is an eigen vector since
1 = 1 1
Moreover
1 = 1 1
347
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
PCA
Extracting the second eigen vector
max
2
348
2 2
2 2 = 1
1 2 = 0
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
PCA
The optimization yields:
2 = 2 2
2 = 2 2 < 1
The second eigen value is smaller than the first due to the
presence of one extra constraint in the optimization the
orthogonality constraint.
349
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
PCA
The K-th eigen vector is derived as
max
.
= 1
1 = 0
2 = 0
1 = 0
350
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
PCA
The optimization yields:
=
= < 1 < < 2 < 1
351
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
PCA
Then - each of the K-eigen vectors delivers a unique asset
pricing factor.
Simply, multiply excess stock returns by the eigen vectors:
1 = 1
1
352
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
PCA
Recall, the basic idea here is to replace the original set of
variables with a lower dimensional set of K-factors ( << ).
The contribution of the -th eigen vector to explain the
covariance matrix of stock returns is
=1
353
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
PCA
Typically the first three eigen vectors explain over and above
95% of the covariance matrix.
What does it mean to explain the covariance matrix?
Coming up soon!
354
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= 1 , ,
355
1, , 0
0, ,
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
356
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
357
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1, , 0
0, ,
358
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Is close to ?
This is the same as asking: what does it mean to explain the
sample covariance matrix?
359
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Is close to ?
Let us represent returns as
1 = 1 +
= +
360
11 1
12 2 +
1
+ 1
1 1
2 2 +
= 1, ,
= 1, ,
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Is close to ?
where 1 are the principal component based factors and
1 are the exposures of firm i to those factors.
361
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Is close to ?
is closed enough to - if
1. The variances of the residuals cannot be dramatically reduced
by adding more factors.
2. The pairwise cross-section correlations of the residuals cannot
be considerably reduced by adding more factors.
362
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
common factors.
Let be an N-vector of portfolio weights:
= 1 , 2 , ,
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1 1
364
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
+
+ 1 1 + 2 2 + +
365
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1 1
2
1 1
1 1
366
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
367
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
368
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
=1 ( )
369
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
where
=
1 1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Other Applications
PCA can be implemented in a host of other applications.
371
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Other Applications
In a matrix form
11 , 12 , , 1
1 , 2 , ,
372
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Other Applications
If > compute the covariance matrix of then extract
principal components such that you summarize the
M-dimension of the predictors with a smaller subspace of order
<< .
373
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Other Applications
Then you construct K predictors:
1 =
374
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
What if > ?
If > then you extract eigen vectors from the TT
matrix.
In this case, the -predictors are the extracted eigen vectors.
Be careful of a look-ahead bias in real time prediction.
375
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
376
1 1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
377
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= + max
ln
378
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
379
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
381
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
382
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
383
Likelihood: , 2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
384
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
385
1
1
= 2+ 2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
In particular,
1 =
386
1
2
1
1
2+ 2
2 = 1 1 =
1
2
1
1
2+ 2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
387
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= 1 1 + 2 2 + +
where 1 , 2 are the sample estimates of factor loadings,
and 1 , 2 are the sample estimates of the factor mean
returns.
388
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
389
1
11
+ 1
1
11
+ 1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2
where and
are the
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
394
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
cov ,
cov ,
=
=
=
2
2
395
=
2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
then
396
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
397
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
In our case = 2.
Then is a 2 matrix.
The first row is all zero except for the fifth entry which is one.
Likewise, the second row is all zero except for the fifth entry
which is one.
399
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
403
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
406
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The Matrix
is a covariance matrix expressing uncertainty about
views.
It is typically assumed to be diagonal.
In the absolute views case described above 1,1 denotes
uncertainty about the first view while 2,2 denotes
uncertainty about the second view both are at the discretion
of the econometrician/investor.
407
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
The Matrix
In the relative views described above: 1,1 denotes
uncertainty about momentum. This could be the sample
variance of the momentum payoff.
2,2 denotes uncertainty about the value payoff. This is the
could be the sample variance of the value payoff.
408
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Deciding Upon
There are many debates among professionals about the right
value of .
From a conceptual perspective it could be 1/ where denotes
the sample size.
You can pick = 0.01
You can also use other values and examine how they perform
in real-time investment decisions.
409
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
max
410
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
=
where
411
+ ( /)
1 1
1 + ( /)1
= ()1 + 1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
412
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Downside Risk
Downside risk is the financial risk associated with losses.
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Downside Risk
Example of downside risk measures
Value at Risk (VaR)
Expected Shortfall
Semi-variance
Maximum drawdown
Downside Beta
Shortfall probability
414
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
5%
95%
415
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
where
1 , the critical value, is the inverse cumulative distribution
function of the standard normal evaluated at .
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1 = + 1 = + 1.64
Check:
If
Then
417
, 2
Pr 1 = Pr
1.64
= Pr
<
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
0.08, 0.202
= 0.08 1.64 0.20 = 0.25
That is to say that we are 95% sure that the future equity
premium wont decline more than 25%.
If we would like to be 97.5% sure the price is that the
threshold loss is higher.
To illustrate,
0.975 = 0.08 1.96 0.20 = 0.31
418
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
VaR of a Portfolio
Evidently, the VaR of a portfolio is not necessarily lower than
the combination of individual VaR-s which is apparently at
odds with the notion of diversification.
However, recall that VaR is a downside risk measure while
volatility which diminishes at the portfolio level is a symmetric
measure.
419
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
422
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
After the year passes you count the number of wrong analysts.
An analyst is wrong if he/she predicts up move when the
market is down, or predict down move when the market is up.
423
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
424
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1 :
Under the null hypothesis it follows that
=
0 1 0
425
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
426
0
0 1 0
0,1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
20 12.6
0.05 0.95 252
= 2.14
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
428
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
429
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
430
1
= +
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
431
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
, 2
| 1
since
432
0
1
=
=
0
1
0 =
= 1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= 20%
1.64
0.10
95% = 0.08 + 0.20
0.08 + 0.20
= 0.32
0.05
0.05
1.96
0.06
97.5% = 0.08 + 0.20
0.08 + 0.20
= 0.40
0.025
0.25
433
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
434
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
435
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
436
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
437
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
438
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2
=
+ 2
+1
where
and are the PDF and CDF of a 0,1 variable,
respectively
Of course if =
then =
439
2
2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
440
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
441
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Downside Beta
1
min
min
, 0
, 0
The numerator in the equation is referred to as the co-semivariance of returns and is the covariance of returns below on
the market portfolio with return in excess of on security .
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Downside Beta
(2)
min , 0
=
min , 0 2
443
min , 0 min , 0
=
min , 0 2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Shortfall Probability
We now turn to understand the notion of shortfall probability.
444
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Shortfall Probability
in Long Horizon Asset Management
Let us denote by the cumulative return on the investment
over several years (say years).
Rather than finding the distribution of R we analyze the
distribution of
= ln 1 +
which is the continuously compounded return over the
investment horizon.
445
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Shortfall Probability
in Long Horizon Asset Management
The investment value after years is
= 0 1 + 1 1 + 2 1 +
Dividing both sides of the equation by 0 we get
= 1 + 1 1 + 2 1 +
0
Thus
1 + = 1 + 1 1 + 2 1 +
446
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Shortfall Probability
in Long Horizon Asset Management
Taking natural log from both sides we get
= 1 + 2 + +
Assuming that
, 2
= 1, . . . ,
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= 0 1 +
= 0 exp
where is the continuously compounded risk free rate.
449
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
ln
< ln
0
0
450
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
452
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Example
Take =0.04, =0.08, and =0.2 per year. What is the Shortfall
Probability for investment horizons of 1, 2, 5, 10, and 20 years?
Use the excel norm.dist function.
If T=1 SP=0.42
If T=2 SP=0.39
If T=5 SP=0.33
If T=10 SP=0.26
If T=20 SP=0.19
453
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
454
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
455
If > exp
you get
If < exp
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
456
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2. = exp .
3. Riskfree rate given by
4. Volatility given by
5. Dividend yield given by = 0
457
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= exp 2 exp 1
Which, given the parameter outlined above, becomes
1
1
=
2
2
458
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
d1 =
(/)+(+2 2 )
and 2 = 1
while 1 = 1 1
2 = 1 2
459
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= 1 2
where
1 =
2 = 1
and is <
460
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
461
T (years)
0.08
0.18
10
0.25
20
0.35
30
0.42
50
0.52
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
463
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
464
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
465
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
, , =
2 2
>0
2
If is a log-normally distributed variable, its expected value
and variance are given by
=
466
1
2
+ 2
2
2
2+
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
467
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
c E(x|x > c) =
1
2
x 2
where 1/ c is a normalizing constant.
468
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
c E xx>c =
=
=
=
469
2
1
+0.5 2
1
2 2
1
2 2 ++
1
2 2 + +0.5 2
+ =
1
2
2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
+0.5 2
1
2
1
2 2
470
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
+0.5 2
F c ELN x x > c =
1
2
+0.5
=
471
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
c ELN (x|x c) =
=
=
472
+0.5 2
+0.5 2
0 2
1
1
2
2
2
2
1
2 ++
2
+0.5 2
1
2
1
2
2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Punch Lines
+ + 2
473
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1
+2 2
or
| =
474
1
+2 2
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
476
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
478
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
479
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= Pr 0 < 0 = Pr ln 0 < ln 0
= Pr
480
<
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= 0
= 1
1
+
481
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
+1
2
)2
1
(
|, , =
1+
/2,1/2
482
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Partial Expectation
Let =
-standardized
Than,
| =
, =
+1
2
2
1
1+
/2,1/2
1
2
2
1+
1
/2,1/2
+1
+1
2
2
1+
/2,1/2 1
483
|
=
+ 2
= ,
1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Partial Expectation
And thus
2
1 , + 1
=
1
1
Where = 1 is quantile of
484
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
.
,
2
2
485
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
486
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
487
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= 0.102
= 0.273
700
600
= 0.1
= 0.274
500
=3
400
300
200
100
0
-3
488
-2
-1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= 1.011
= 0.856
= 1
= 0.866
250
=3
200
150
100
50
0
-4
489
-3
-2
-1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
300
250
= 0.099
= 0.174
= 0.1
= 0.177
200
= 10
150
100
50
0
-0.6
490
-0.4
-0.2
0.2
0.4
0.6
0.8
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
300
250
= 0.994
= 0.566
= 1
= 0.559
200
= 10
150
100
50
0
-1
491
-0.5
0.5
1.5
2.5
3.5
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Sub-Periods
Divide the sample into non-overlapping subsamples (to
preserve assumption) of length :
1 |+1 2 | |+1
, 0
=
, =
493
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Sub-Periods
Using estimations of , and we can compute VaR
+
ln 1 1 , 0
=
+ ln ln 1
, 0
Here we used the assumption that returns are iid and then
single period return is related to return over sub-period n as
follows
< =
< = <
=1
494
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
(1)
496
+
1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
1 1+
, 0
=
1
, 0
For
0,
497
0, /
<0
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
estimate
, where n is the total number of observations
and is the number of exceedances above the threshold :
1/
=1
1+
498
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
=+
And
= + | >
is mean excess function for GP distribution which equals to
499
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
500
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
501
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Option Pricing
The B&S call Option price is given by
C(S, K, , r, T, ) = SeT N(d1 ) KerT N(d2 )
The put Option price is
P(S, K, , r, T, ) = KerT N(d2 ) SeT N(d1 )
1
Where d1 =
502
(/)+(+2 2 )
and d2 = d1 T
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Option Pricing
There are six inputs required:
503
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
504
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
507
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
508
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Strike
Price
509
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
510
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Strike
Price
511
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
512
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
513
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
514
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
515
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
516
0.90
0.95
1.00
1.05
1.10
1 month
14.2
13.0
12.0
13.1
14.5
3 months
14.0
13.0
12.0
13.1
14.2
6 months
14.1
13.3
12.5
13.4
14.2
1 year
14.7
14.0
13.5
14.0
14.8
2 years
15.0
14.4
14.0
14.5
15.1
5 years
14.8
14.6
14.4
14.7
15.0
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
517
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Volatility Models
We describe several volatility models commonly applied in
analyzing quantities of interest in finance and economics:
ARCH
GARCH
EGARCH
Stochastic Volatility
Realized and implied Volatility
518
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Volatility Models
All such models attempt to capture the empirical evidence that
volatility is time varying (rather than constant) as well as
persistent.
The EGARCH captures the asymmetric response of volatility to
advancing versus diminishing markets.
In particular, volatility tends to be higher (lower) during down
(up) markets.
519
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
ARCH(1)
= +
= where ~ 0,1
2
2 = + 1
1 2 = 1 2 2 = 2
so
2 is the conditional variance.
520
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
ARCH(1)
2
1
= 2 is the unconditional variance.
2
2 = + 1
2
= + 1
2
2
= + 1
1
2
= + 1
2
2
2 = 2 = 1
= 2
521
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
ARCH(1)
Fat tail?
4 =
=
4
2
2
1 4 4
=3
522
1 4
2
1 2 2
1 2 2
4
2
34
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
ARCH(1)
The last step follows because:
2 = 4 2
so
4
2
2
523
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
GARCH(1,1)
= +
= where ~ 0,1
2
2
2 = + 1
+ 1
2
2
2 = + 1
+ 1
524
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
GARCH(1,1)
2 = + 2 + 2
4 =
525
1
1
3 1++ 1
1232 2
>3
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
EGARCH
= + where ~ 0,1
=
ln
=+
1
1
1
1
1
2
2
+ ln 1
= 1
526
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
EGARCH
The second component is
= 1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
EGARCH
More formally if 1 < 0 then the log conditional variance rises
by + .
If 1 > 0 then the log conditional variance rises by
This produces the asymmetric volatility effect volatility is
higher during down market and lower during up market.
528
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Realized Volatility
The realized volatility (RV) is a very tractable way to measure
volatility.
It essentially requires no parametric modeling approach.
Suppose you observe daily observations within a trading month
on the market portfolio.
RV is the average of the squared daily returns within that
month.
530
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Realized Volatility
Of course, volatility varies on the monthly frequency but it is
assumed to be constant within the days of that particular
month.
If you observe intra-day returns (available for large US firms)
then daily RV is the sum of squared of five minute returns.
531
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Where d1 =
532
(/)+(+2 2 )
and d2 = d1 T
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
534
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Return Predictability
If log returns are IID there is no way you can deliver better
prediction for stock return than the current mean return.
535
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Return Predictability
Also note that the variance of a two-period return + +1 is
equal to
+ +1 + 2 cov , +1 = 2 2
536
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
,+1
+
+1
=1+
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
2 1 0,1
538
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
=1
+1
0 : = 1 no auto correlation
1 :
539
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Variance Ratios
Test statistic:
1 0,
+1
e.g
4 1
=2
2 1 0,1
=3
3 1
540
20
0,
9
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Predictive Variables
In the previous specification, we used lagged returns to
forecast future returns or future volatility.
You can use a bunch of other predictive variables, such as:
The term spread.
The default spread.
Inflation.
541
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Predictive Variables
The aggregate book-to-market ratio.
542
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Predictive Regressions
To examine whether stock returns are predictable, we can run
a predictive regression.
This is the regression of future excess log or gross return on
predictive variables.
It is formulated as:
+1 = + 1 1 + 2 2 + + + +1
543
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Predictive Regressions
To examine whether either of the macro variables can predict
future returns test whether either of the slope coefficients is
different from zero.
Use the t-statistic or F-statistic for the regression R-squared.
There is a small sample bias if (i) the predictive variables are
highly persistent, (ii) the contemporaneous correlation between
the predictive regression residual and the innovation of the
predictor is high, or (iii) the sample is small.
544
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
545
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Newey-West Correction
Rewriting the long horizon regression
+1,+ = + +1,+
= 1,
= ,
546
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Newey-West Correction
The estimation error of the regression coefficient is represented
by
= 1
where is the Newey-West given by serially correlated
adjusted estimator
=
=
547
=+1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
548
1
3
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
K=1:
= 1, = 0, = 1
=
=1
549
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
K=2:
= 2, = 1, = 0, = 1, = 2
1 1
=
2
=1
1
+1 +
1
=
550
2
=1
1
=2
2 +
=1
1 1
+
2
+1
=1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
=
=
551
=1
=1
+1
=1
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
552
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
=1 ,1
2
= 1
2
=1
Where ,1 is the return forecast assuming the presence of
predictability, and is the sample mean (no predictability).
Can compute the MSE (Mean Square Error) for both models.
553
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Model Selection
When variable are potential candidates for predicting stock
returns there are 2 linear combinations of predictive models.
In the extreme, the model that drops all predictors is the nopredictability or IID model.
The one that retains all predictors is the all inclusive model.
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
= ln 2 ln
2
=1 1
1
1
= 1
555
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Model Selection
Notice that all criteria are a combination of goodness of fit and
a penalty factor.
You choose only one model and disregard all others.
Model selection criteria have been shown to exhibit very poor
out of sample predictive power.
556
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics
Model Combination
The other approach is to combine models.
Prof. Doron Avramov, The Jerusalem School of Business Administration, The Hebrew University of Jerusalem, Financial Econometrics