Scribd
Upload
3 views2 pages

Exercise 1

This document outlines the coursework for ECON 5412, which includes three questions related to stochastic processes and their properties. The first question focuses on deriving a weak law of large numbers (WLLN) for an AR(1) process, the second question examines a linear process and its covariance, and the third question provides an example of a sequence that is integrable but not uniformly integrable. The deadline for submission is set for February 7, 2025.

Uploaded by

qygkwjj6rt
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
3 views2 pages

Exercise 1

This document outlines the coursework for ECON 5412, which includes three questions related to stochastic processes and their properties. The first question focuses on deriving a weak law of large numbers (WLLN) for an AR(1) process, the second question examines a linear process and its covariance, and the third question provides an example of a sequence that is integrable but not uniformly integrable. The deadline for submission is set for February 7, 2025.

Uploaded by

qygkwjj6rt
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

ECON 5412 Exercise 1

2024/25

This coursework will be marked and will counts towards your …nal grade.

[Deadline: 16:00 on 7 February 2025]

Question 1
Let X1 ; :::::; Xn be random variables generated by an AR(1) process

Xt = Xt 1 + "t ; 2 ( 1; 1) t 2 f1; :::; ng (1)

where "1 ; :::; "n are i.i.d. (0; 2 ) and the initial condition X0 is a constant or a random
variable satisfying EX02 < 1. The following argument takes advantage of the simple
recursive propertyP (…rst order stochastic di¤erence equation) to derive a bare hands
WLLN for n 1 nt=1 Xt2 1 .

(a) By squaring both sides of (1) and summing over t 2 f1; :::; ng, prove the identity

X
n X
n X
n
2
1 Xt2 1 = Xn2 + X02 +2 Xt 1 "t + "2t : (2)
t=1 t=1 t=1

Why would you need to multiply (2) by (at least) n 1 ?

(b) Show that n 1 EXn2 ! 0 and that n 1 EX02 ! 0.

(c) Explain why fXt 1 "t : t 1g is a martingale di¤erence sequence clearly stating
which …ltration Ft you are using. Use the above property to prove that
!2
1 Xn
E Xt 1 "t ! 0:
n t=1

1
Pn
(d) Put the above results together to obtain a WLLN for n t=1 Xt2 1 .

Question 2
Consider a linear process
X
1 X
1
Xt = c j ut j ; jcj j < 1 (3)
j=0 j=0

1
where (ut ; Ft ) is a martingale di¤erence sequence with E (u2t j Ft 1 ) = 2
for all t. Show
that E (Xt ) = 0 and compute the covariance

(k) := cov (Xt ; Xt k ) = E (Xt Xt k )

when k 0 and when k < 0. Show that (k) is independent of t and that

( k) = (k)

for all k 0.

Question 3
This provides a simple example of a sequence (Xn )n2N that is integrable but not uniformly
integrable. Consider the random experiment of choosing a number ! at random from
[0; 1]. The sample space of all possible outcomes is = [0; 1]. We may assign a probability

P [a; b] = P (f! 2 [a; b]g) = b a

to each subinterval [a; b] of [0; 1] and de…ne the (discrete) random variables

n; ! 2 (0; 1=n)
Xn (!) =
0; ! 6= (0; 1=n) :

(i) Show that E (Xn ) = 1 for all n and conclude that supn2N E jXn j = 1.

(ii) Show that for all n >


E [jXn j 1 fjXn j > g] = 1
so (Xn )n2N is not uniformly integrable.

(iii) Show that Xn !p 0 but that (Xn ) does not converge to 0 in L1 . Is this a surprise?

You might also like