Spatial projection of input-ouput tables for small areas

Studies on regional economy have achieved a huge expansion in the last decades. In particular, from an input-ouput optic many efforts have been devoted to carring out a suitable methodology, which enable us to cover the overall and exhaustive knowledge of the economic reality of one region. Given that a input-output table (IOT) gathers both intersectorial relationships and the final demand of the economy, it allows us to provide a reliable picture of one economy in a certain moment of time. Nonetheless, the elaboration of a IOT is a complex task, which needs many human and economic resources. Thus, most of the tables elaborated using direct methods have as a benchmark frame either a country or a region, although it is difficult to find matrices related to smaller geografic spaces like, for instance, counties. So, if we attempt to perform a deeper study of both spaces, it would be of great help if we could dispose of estimation methods, which enable us to make tables with less information, i.e, indirect (semidirect) methods of estimation. Let´s say that the economy of a region is determined by the relations among productive structures of their counties, therefore a previous knowledge of these productive structures can be interesting. The basic aim of this work consists of estimating a TIO for each one of the eight Asturian Counties in 1995, since this is the last period in which we possess published information relative to regional accounts. To this end, a technique focused on mathematical theory of the information: cross entropy, will be employed. Such a technique has lately been applied to the construction of regional tables, largely for two reasons: one, flexibility as regards the information it needs; the other, to produce some rather suitable empiric results. From the tables estimates by this method we will be able to know the characteristics of economic structures of the counties. To achieve this scope, tools related to the graph theory, have been applied. Their application in input-output analisis has a great potential to provide a simple vision of the relations between the different sectors, as well as being able to integrate matters as important as the relative positions of the sectors, their orientation or paths in which drive the economic influence inside the corresponding structure.