Real-time regional GDP forecasting : statistical aspects and a forecasting model

Abstract

The monitoring of the regional economic situation takes on particular importance in highly decentralised countries, such as Spain. Against this background, this article summarises the key aspects of the BayFaR model (Bayesian Factor model for Regions), a new tool used by the Banco de España for the short-term forecasting of the GDP of the Spanish economy and of its biggest regions. The tool allows monthly indicators of activity to be combined with quarterly GDP, the benchmark variable for conjunctural analysis. In Spain’s case, however, it is necessary to use alternative measures of this variable. This is because the official statistics do not provide a homogeneous measure for all the regions, nor one that is consistent with INE’s national quarterly GDP