ARCH modeling in the presence of missing data

Author

Bondon, Pascal

Author_Institution

Lab. des Signaux et Syst., Supelec, Gif-sur-Yvette, France

fYear

2013

fDate

3-6 Nov. 2013

Firstpage

39

Lastpage

43

Abstract

The problem of estimating an autoregressive conditionally heteroscedastic (ARCH) model in the presence of missing data is investigated. A two-stage least squares estimator which is easy to calculate is proposed and its strong consistency and asymptotic normality are established. The behaviour of the estimator for finite samples is analyzed via Monte Carlo simulations, and is compared to a Yule-Walker estimator and to some estimators based on a complete data set obtained after filling the missing observations by imputation procedures. An application to real data is also reported.

Keywords

Monte Carlo methods; autoregressive processes; estimation theory; least squares approximations; ARCH modeling; Monte Carlo simulations; Yule-Walker estimator; asymptotic normality; autoregressive conditionally heteroscedastic model; imputation procedures; missing data; two-stage least squares estimator; Data models; Estimation; Least squares approximations; Monte Carlo methods; Predictive models; Time series analysis; Yttrium; ARCH models; conditional heteroscedasticity; least squares estimation; missing observations;

fLanguage

English

Publisher

ieee

Conference_Titel

Signals, Systems and Computers, 2013 Asilomar Conference on

Conference_Location

Pacific Grove, CA

Print_ISBN

978-1-4799-2388-5

Type

conf

DOI

10.1109/ACSSC.2013.6810225

Filename

6810225