Estimation in ARCH Models with Missing Data: Applications to Latin-American Capital Markets

• Activity Number: 240
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Natalia Bahamonde*+
• Companies: Pontificia Universidad Católica de Valparaiso
• Keywords: Time series ; ARCH models ; missing data ; financial data

Abstract

In this work, we are interested in parametric estimation of stationary time series models in presence of missing data. This situation can be found in many scientific areas and arises for a variety of causes. The estimation of the parameters of a time series model in this kind of problems is more difficult than in the complete case. In this work we present a least squares method of estimation for the parameters of a stationary AutoRegressive Conditional Heterocedasticy (ARCH) model with missing observations. The ARCH models have been introduced by Engle (1982) and are extensively used in economics. The statistical properties of this estimator are explored and the asymptotic normality is derived. Applications to real data sets involving Latin-american capital markets are presented in which our procedure is compared to directly imputation methods.