Files in this item



application/pdf9305751.pdf (3MB)Restricted to U of Illinois
(no description provided)PDF


Title:Specification tests for autoregressive conditional heteroskedastic models with applications to exchange rates of Asian countries
Author(s):Zuo, Xiao-Lei
Doctoral Committee Chair(s):Bera, Anil K.
Department / Program:Economics
Degree Granting Institution:University of Illinois at Urbana-Champaign
Economics, General
Economics, Finance
Abstract:The White information matrix (IM) test is applied to the linear regression model with autoregressive conditional heteroskedastic (ARCH) errors. ARCH models are used widely in analyzing economic and financial time series data. However, in practice, the models are not often thoroughly tested. We derived some tests for these models using the IM test principle. It is found that these tests turned out to be equivalent to checking whether kurtosis is changing over time, i.e to test for heterokurtosity.
In the second part of the thesis, we considered Neyman's $C(\alpha)$ test for ARCH(p) against ARCH(p + r) and GARCH(p,q) against GARCH(p + r,q). We used the ergodic theorem to complete the proof of $\sqrt{n}$-consistency of ordinary least squares (OLS) type estimates of mean equation and variance equation parameters. Also we compared the $C(\alpha)$ test with Lagrange multiplier (LM) test for both models.
The third part deals with exchange rates of Asian countries. The theoretical results are applied to study the validity of commonly used ARCH type models.
Issue Date:1992
Rights Information:Copyright 1992 Zuo, Xiao-Lei
Date Available in IDEALS:2011-05-07
Identifier in Online Catalog:AAI9305751
OCLC Identifier:(UMI)AAI9305751

This item appears in the following Collection(s)

Item Statistics