Fighting Over-Indebtedness: An Artificial Intelligence Approach: An Abstract

This research focuses attention on over-indebtedness (i.e., recurrent incapability to repaying credits) and its risk factors, among Portuguese households in the context of the recent European sovereign debt crisis. Different theoretical accounts of consumers decision behavior and risk of becoming over-indebted vary (among other aspects) on the emphasis they put on situational (socio-economic) versus individual (psychological) factors (Angel et al. 2009; Berthoud and Kempson 1992; Kamleitner and Kirchler 2007; van Staveren 2002). Although several of the identified factors have been shown to be associated with over-indebtedness, actual cases of over-indebted households are likely to be multifactorial. Remarkably, how these different risk factors combine in producing concrete situations of over-indebtedness is a highly important issue to avoid poverty that has received less research attention.

This paper examines how artificial intelligence may contribute to better understanding and overcome over-indebtedness in contexts of severe economic austerity. We analyze a field database of over-indebted households with a high risk of poverty. Artificial intelligence algorithms are used to identify distinguishable over-indebtedness clusters and to predict over-indebtedness risk factors within each cluster. First, unsupervised machine learning generated three over-indebtedness clusters of families affected by abrupt economic crisis. Second, supervised machine learning with exhaustive grid search hyperparameters suggest algorithms that best predict families’ over-indebtedness risk factors. These findings extend previous research by proposing a multifaced and yet organized bottom-up approach to over-indebtedness and poverty risk.