Structural and reduced-form approaches of ARIMA model based seasonal adjustment methods / Estela Bee Dagum. : CS11-614/87-7E-PDF

"Two strategies have been followed for the development of model-based seasonal adjustment methods. One, where each of the unobserved components, trend-cycle, seasonal and irregular is assumed to follow a normal stochastic process of the ARIMA class and the other, where the observed data are assumed to follow an ARIMA process and from it similar kind of models are derived for the components. The first approach is known as "structural" and the second, as the "reduced-form" given their similarities to the problems of identification of structures from the data in econometrics. This paper discusses the major properties and operational limitations of these two approaches. It also analyses the salient characteristics of the empirical comparisons made between model-based seasonal adjustment methods and the X-11-ARIMA which is used by the majority of government statistical agencies"--Abstract.

Lien permanent pour cette publication :
publications.gc.ca/pub?id=9.838105&sl=1

• Format MARC XML
• Format MARC HTML

Demander des formats alternatifs