Segregation and Mobility in Economic Activity: Deriving Features and Drivers from Income and Spending Patterns

Milan van den Heuvel (Ghent University)
Tarik Roukny (MIT)
Ken Bastiaensen (Ghent University)
Benjamin Vandermarliere (Ghent University)
Koen Schoors (Ghent University)
Jan Ryckebusch (Ghent University)

Abstract: In this paper, we study the dynamics of income and spending behaviors using panel data spreading 10 years (2006-2016) and containing 3,5 million clients from the belgian market of a large european bank. For this representative sample of the belgian population, we collect every transactions statement at the individual level, as well as a wide range of individual characteristics (e.g. loan debt, age, postal code, marital status, etc.). First, we identify distinct clusters based on income and spending behavior using a selected set of individual characteristics data. We then study the evolution of each cluster over time. Furthermore, we characterise transfer flows of individuals from one cluster to another. When an individual moves from one cluster to another, we study other changes that subsequently take place in aspects of the economic behavior of the individual that is not part of the clustering method. Doing so, we are able to study economic and social mobility at a large and representative scale of the country population showing clear cases of intra-generational mobility upwards and downwards as well as the effects of mobility on other dimensions of individual’s economic life. This work offers a map of a country’s population relative economic dynamics, which factors come into play, and what second order effects may in turn occur. The results are relevant to discussions of government policy intended to identify and promote upwards trends for the welfare of its population.