Umeå Economic Studies, Department of Economics, Umeå University
No 436:
Generalized Method of Moment and Indirect Estimation of the ARASMA Model
Kurt Brännäs ()
and Xavier de Luna ()
Abstract: Estimation in nonlinear time series models has mainly been
performed by least squares or maximum likelihood (ML) methods. The paper
suggests and studies the performance of generalized method of moments (GMM)
and indirect estimators for the autoregressive asymmetric moving average
model. Both approaches are easy to implement and perform well numerically.
In a Monte Carlo study it is found that the MSE properties of GMM are close
to those of ML. The indirect estimator performs poorly in this respect. On
the other hand, the three estimation techniques lead to fairly similar
power functions for a linearity test.
Keywords: Estimation; Nonlinearity Test; Small Sample Properties; Time Series.; (follow links to similar papers)
JEL-Codes: C13; C15; C22; (follow links to similar papers)
10 pages, December 15, 1997
Download Statistics
- This paper is published as:
-
Brännäs, Kurt and Xavier de Luna, (1998), 'Generalized Method of Moment and Indirect Estimation of the ARASMA Model', Computational Statistics, Vol. 13, pages 485-494
Questions (including download problems) about the papers in this series should be directed to Kjell-Göran Holmberg ()
Report other problems with accessing this service to Sune Karlsson ()
or Helena Lundin ().
Programing by
Design by Joachim Ekebom
