Abstract
The thesis "Concerned with everything: an objective index for local well-being" tackles the issue of the estimating the well-being at municipal level and how it relates to other micro-level indicators. In this field of research, the literature is usually bipartite, i.e. based on subjective or objective data; the first one is based on psychological surveys results, where it is essentially the answer to questions about the life satisfaction. It has been shown that these measures show a high degree of reliability. The second approach is based on data derived from official and univocally measurable statistics, such as, for example the data arising from national accounts. It can be further divided into two subareas of research, one known as "batches of indicators" and the other known as "objective well-being index", which are synthetic measures of well-being. All 3 strands have their own limitations: given the very nature of their sampling techniques, subjective well-being (SWB) has very high costs; the batch of indicators approach, although very comprehensive and with a lot of information, requires high statistical-technical expertise in order to be consulted, and are not efficient. Finally, an objective index of well-being (OWB), responds to the needs of the groups of indicators synthesis, but, of course, are likely to omit a certain quantity of information. This thesis addresses the issue of building a batch of 160 indicators at the municipal level for the Lombardia region, from 2001 to 2015, an effort which led to the statistic platform 100% Lombardy, in collaboration with Éupolis and Regione Lombardia; this is a first element of innovation, since the literature presents only batteries which "stop" at the regional level as a minimum level of analysis (or very few exceptions for big cities and notable provinces). This allows to evaluate and consider phenomena that occur only at micro-regional level. From 52 indicators chosen in the battery, a cluster analysis was carried out, aimed at identifying 11 uniform groups in Lombardia, starting from the 1531 cities; then, the 52 indicators have been condensed into one single indicator of objective well-being, called WIT (acronym for Well-being Index of Towns), explaining at least 40% of the variance of the original data in each year, from 2001 to 2015. The WIT, together with income per taxpayer, average quotation of residential buildings, 3 year migration rate, tourism index and the birth rate, have been included in a Panel Vector AutoRegression model, along with an exogenous variable, therefore forming a PVARX model. The exogenous variable is a dummy active only for the data of 2011, given the numerical difference due to the different collection of national census data. The final model has a temporal depth of 57 quarters, for 6 variables (plus the exogenous one), with a spatial component of 11 clusters. The analysis of the WIT movements showed specific local patterns, related to the presence of infrastructures (more or less developed) and congestion (in terms of persons) of a territory, also due to the presence of economic activity. The impulse response function of the PVARX model showed that the other 5 variables respond positively to WIT shock, thus making such results very useful for policy making.