Potencialidad de la modelizacion State-Space y el filtro de Kalman para el analisis regional. Una aplicacion para el indice de actividad industrial

The deficit existent in our country in order of the available of quantitative indicators to made a regional industrial activity conjuncture analysis have initiated a discussion in different forums about which is the best methodology for elaborating indicators of this characteristics. In this context, in this paper we summarise and extend the main conclusions obtained in previous studies (Clar, et al., 1997a, 1997b and 1998) about the possibility of extending the indirect methods that nowadays have been used in the Spanish regions to analyse the short term evolution of regional industrial production. This conclusions takes us to propose a different strategy for elaborating indirect quantitative indicators. In particular, (following Israilevich and Kuttner, 1993) a latent variable model for estimating the regional production activity is proposed. This kind of models can be specified in terms of state-space model and estimated by the Kalman filter. To validate the proposed methodology an indicators are estimated following this methodology for three of the four Spanish regions that have an Industrial Production Index (IPI) obtained by direct methods (Andalucia, Asturias and Pais Vasco) and we compare them with the published IPIs. The results obtained shows the good accuracy of the proposed strategy, opening a working line for surpassing the deficit existent nowadays in Spanish regions.