Forecasting Spanish Inflation Using the Maximum Disaggregation Level by Sectors and Geographical Areas

This article evaluates different strategies for forecasting Spanish inflation using information from price series for fifty-seven products and eighteen regions in Spain. We consider vector equilibrium correction (VeqCM) models that include cointegration relationships between Spanish prices and prices in the regions of Valencia, Andalusia, Madrid, Catalonia, and the Basque Country. This approach is consistent with economic intuition and is shown to be of tangible importance after suitable econometric evaluation. It is found that sectoral disaggregate models are useful for forecasting inflation in the five largest Spanish regions. Moreover, aggregate inflation forecasts in Spain can be significantly improved by aggregating projections from different sectors and geographical areas and by considering cointegration relationships between regional and national prices. However, in spite of the existence of long-run relationships between sectoral and national prices, they include deterministic components that are not consistent with the law of one price.