• 検索結果がありません。

Econometrics I: Solutions of the homework #12

N/A
N/A
Protected

Academic year: 2021

シェア "Econometrics I: Solutions of the homework #12"

Copied!
7
0
0

読み込み中.... (全文を見る)

全文

(1)

#12

LU ANG July 7, 2020

Contents

1 Question 1 1

1.1 (1) . . . . 1

1.2 (2) . . . . 4

1.3 (3) . . . . 4

1.4 (4) . . . . 5

1.5 (5) . . . . 5

1 Question 1

1.1 (1)

In order to Obtain the variance matrix of y = (y

1

, y

2

, · · · , y

t

) we first rewrite y

t

as:

If you have any errors in handouts and materials, please contact me via

[email protected]

(2)

y

t

= ρy

t−1

+

t

= ρ

2

y

t−2

+

t

+ ρ

t−1

.. .

= ρ

τ

y

t−τ

+

t

+ ρ

t−1

+ · · · + ρ

τ−1

t−τ+1

.. .

=

t

+ ρ

t−1

+ ρ

2

t−2

+ · · ·

where ρ

τ

y

t−τ

goes to 0 when |ρ| < 1 as τ goes to infinity Then we can obtain variance as:

V (y

t

) = V (

t

+ ρ

t−1

+ ρ

2

t−2

+ · · · )

= V (

t

) + ρ

2

V (

t−1

) + ρ

4

V (

t−2

) + · · ·

= σ

2

(1 + ρ

2

+ ρ

4

+ · · · )

= σ

2

1 − ρ

2

Next denote the covariance of y

t

and y

t−τ

as γ(τ ):

γ(τ ) = Cov(y

t

, y

t−τ

)

= E(y

t

y

t−τ

) = E (ρ

τ

y

t−τ

+

t

+ ρ

t−1

+ · · · + ρ

τ−1

t−τ+1

)y

t−τ

= ρ

τ

E(y

2t−τ

) + E(u

t

y

t−τ

) + ρE(u

t−1

y

t−τ

) + ρ

2

E(u

t−2

y

t−τ

) + · · · ρ

τ−1

E(u

t−τ+1

y

t−τ

)

= ρ

τ

E(y

2t−τ

)

= ρ

τ

γ(0) Notice V (y

t

) = γ(0)

Going back to the vector form

E(y) = E

 y

1

y

2

.. . y

T

=

 0 0 .. . 0

(3)

V (y) = E(yy

0

) = E(

 y

1

y

2

.. . y

T

y

1

y

2

· · · y

T

)

=

γ(0) γ(1) · · · γ(T − 1)

γ(1) γ(0) γ(1) · · · γ(T − 2)

γ(1) γ(0) . .. .. . .. . .. . . .. . .. γ(1)

γ(T − 1) γ(T − 2) · · · γ(1) γ(0)

=

γ (0) ργ (0) · · · ρ

T−1

γ(0)

ργ(0) γ(0) ργ(0) · · · ρ

T−2

γ(0) ργ (0) γ(0) . .. .. . .. . .. . . .. . .. ργ(0) ρ

T−1

γ(0) ρ

T−2

· · · ργ(0) γ(0)

= γ(0)

1 ρ · · · ρ

T−1

ρ 1 ρ · · · ρ

T−2

ρ 1 . .. .. . .. . .. . . .. ... ρ ρ

T−1

ρ

T−2

· · · ρ 1

= σ

2

1 − ρ

2

1 ρ · · · ρ

T−1

ρ 1 ρ · · · ρ

T−2

ρ 1 . .. .. . .. . .. . . .. ... ρ ρ

T−1

ρ

T−2

· · · ρ 1

= Ω

(4)

1.2 (2)

Based on (1) we know that y ∼ N (0, Ω). Thus, the joint distribution of y = (y

1

, y

2

, · · · , y

T

) whihch is also the likelihood function can be obtained as:

f(y) = (2π)

−T /2

|Ω|

−1/2

exp(− 1

2 y

0

−1

y)

1.3 (3)

Unconditional Mean:

E(y

t

) = E(

t

+ ρ

t−1

+ ρ

2

t−2

+ · · · )

= E(

t

) + ρE(

t−1

) + ρ

2

E(

t−2

) + · · ·

= 0 Unconditional Variance:

V (y

t

) = V (

t

+ ρ

t−1

+ ρ

2

t−2

+ · · · )

= V (

t

) + ρ

2

V (

t−1

) + ρ

4

V (

t−2

) + · · ·

= σ

2

(1 + ρ

2

+ ρ

4

+ · · · )

= σ

2

1 − ρ

2

Conditional mean:

E(y

t

|y

t−1

, · · · , y

1

) = ρy

t−1

+ E(

t

) = ρy

t−1

Conditional variance:

V (y

t

|y

t−1

, · · · , y

1

) = V (

t

) = σ

2

(5)

1.4 (4)

From (3) we can know that the unconditional distribution of y

t

is given by:

f(y

t

) = 1

p 2πσ

2

/(1 − ρ

2

) exp(− y

t2

2

/(1 − ρ

2

) )

the conditional distribution of y

t

is given by:

f (y

t

|y

t−1

, · · · , y

1

) = 1

2πσ

2

exp(− (y

t

− ρy

t−1

)

2

2

)

The innovation form of the likelihood function can be written as:

f (y

t

, y

t−1

, · · · , y

1

) = f (y

t

|y

t−1

, · · · , y

1

)f(y

t−1

, y

t−2

, · · · , y

1

)

= f (y

t

|y

t−1

, · · · , y

1

)f(y

t−1

|y

t−2

, · · · , y

1

)f(y

t−2

, y

t−3

, · · · , y

1

) .. .

= f (y

t

|y

t−1

, · · · , y

1

)f(y

t−1

|y

t−2

, · · · , y

1

)f(y

t−2

, y

t−3

, · · · , y

1

) · · · f (y

2

|y

1

)f(y

1

)

= f (y

1

)

T

Y

t=2

f (y

t

|y

y−1

, · · · , y

1

)

= 1

p 2πσ

2

/(1 − ρ

2

) exp(− y

12

2

/(1 − ρ

2

) ) ×

T

Y

t=2

√ 1

2πσ

2

exp(− (y

t

− ρy

t−1

)

2

2

)

1.5 (5)

Set P

−1

such that:

P

−1

ΩP

0−1

= I

T

i.e.

Ω = P P

0

(6)

We can construct P

−1

as:

P

−1

= 1 σ

p 1 − ρ

2

0 · · · · · · 0

−ρ 1 0 · · · 0

0 −ρ 1 . .. ...

.. . .. . . .. ... 0

0 · · · 0 −ρ 1

Thus:

P

−1

y = 1 σ

p 1 − ρ

2

0 · · · · · · 0

−ρ 1 0 · · · 0

0 −ρ 1 . .. ...

.. . .. . . .. ... 0

0 · · · 0 −ρ 1

 y

1

y

2

.. . y

T

= 1 σ

p 1 − ρ

2

y

1

y

2

− ρy

1

.. . y

T

− ρy

T−1

Then:

y

0

−1

y = y

0

P

0−1

P

−1

y

= 1 σ

2

p 1 − ρ

2

y

1

y

2

− ρy

1

.. . y

T

− ρy

T−1

T

p 1 − ρ

2

y

1

y

2

− ρy

1

.. . y

T

− ρy

T−1

= 1

σ

2

(1 − ρ

2

)y

12

+

T

X

t=2

(y

t

− ρy

t−1

)

2

!

(1)

(7)

On the hand by using the matrix determinant property det(A

−1

) = 1 det(A) and det(AB) = det(A)det(B) we can easily obtain:

|Ω| = 1

|Ω

−1

| = 1

|P

0−1

||P

−1

| = σ

2T

1 − ρ

2

(2)

Substitute expression (1) and (2) back into f (y) = (2π)

−T /2

|Ω|

−1/2

exp(−

12

y

0

−1

y) we can easily find that the likelihood function we obtained in question (2) and (4) are the same.

Finally we can calculate the inverse of P

−1

to obtain:

P = σ

p 1 − ρ

2

1 0 0 · · · 0

ρ p

1 − ρ

2

0 · · · 0

ρ

2

ρ p

1 − ρ

2

p

1 − ρ

2

0 .. .

.. . .. . . .. . .. 0

ρ

T−1

ρ

T−2

p

1 − ρ

2

· · · ρ p

1 − ρ

2

p 1 − ρ

2

参照

関連したドキュメント

W ang , Global bifurcation and exact multiplicity of positive solu- tions for a positone problem with cubic nonlinearity and their applications Trans.. H uang , Classification

[11] Karsai J., On the asymptotic behaviour of solution of second order linear differential equations with small damping, Acta Math. 61

Now it makes sense to ask if the curve x(s) has a tangent at the limit point x 0 ; this is exactly the formulation of the gradient conjecture in the Riemannian case.. By the

Then, the existence and uniform boundedness of global solutions and stability of the equilibrium points for the model of weakly coupled reaction- diffusion type are discussed..

In recent years, several methods have been developed to obtain traveling wave solutions for many NLEEs, such as the theta function method 1, the Jacobi elliptic function

7, Fan subequation method 8, projective Riccati equation method 9, differential transform method 10, direct algebraic method 11, first integral method 12, Hirota’s bilinear method

Sun, Optimal existence criteria for symmetric positive solutions to a singular three-point boundary value problem, Nonlinear Anal.. Webb, Positive solutions of some higher

It is also well-known that one can determine soliton solutions and algebro-geometric solutions for various other nonlinear evolution equations and corresponding hierarchies, e.g.,