L’impiego di algoritmi AMOEBA per lo studio delle variazioni temporali di un fenomeno economico: prime evidenze generalizzabili da un caso di studio

Every economic phenomenon presenting a spatial diffusion is strongly affected, even in its variations, by the relationships existing among neighbouring territorial units. The present work represents an example of the analyses of the changes in the different kinds of expenditure in the regional budget (as, for instance, the restoration, maintenance, and preservation of cultural heritage). In particular, the analysis focuses on the healthcare expenditures observed at the municipality level. This work aims at overcoming the traditional tendency of classical economic theory to neglect the influence of the territory on economic phenomena. In particular, new analytical tools capable of exploiting information on the territorial structure must be used in the practice in order to explain the temporal variations of these phenomena. A classical approach to the analysis of the temporal variations of an economic phenomenon is the so-called Shift-Share Analysis. For over half a century, this methodology has been considered by the statistical and the economic literatures as a tool for territorial analysis. The geographic literature pointed out that also the relationships among the territorial units must be considered in order to define a “territorial model”. Typically, the relational structure (neighbouring system) is assumed ex-ante by the researchers. The present contribution aims to show how this hypothesis can be removed. In particular, the use of a modified version of the AMOEBA algorithm (A Multidirectional Optimum Ecotope-Based Algorithm) in the Spatial Shift-Share Analysis is proposed to account for the spatial structure in accordance with the well-known Tobler's "First Law of Geography".