Data Integration Techniques for the Identification of Poverty Profiles

Abstract

Economic connotation of poverty concerns mainly the level of household’s spending and income as privileged indicators. Sometimes, data sources have several lack of information, but such information is available in other data sources related to the same statistical units, or to other households. In order to identify records relating to the same or similar units belonging to two data archives, some methodologies (record linkage in the first case, statistical matching in the latter) are used for integrating data. In this study we illustrate a model of data integration between two surveys conducted by the Italian National Institute of Statistics (Statistics on Income and Living Conditions and Survey on Household Expenses). The construction of an integrated database starting from those two surveys is useful for studying consumer behaviours in relation to specific groups of goods, analysing decisions related to household savings, examining economic and social inequalities, such as studying the impact of public policies through simulations.