Poverty risk estimation on the basis of Sen’s capability approach with the application of the MIMIC model

Abstract

Poverty is a problem which concerns not only the affected individual, but also a whole society. Poverty tends to hinder or even prevent the economic growth and social development of a country. Therefore, undertaking action which could reduce poverty is of great importance, as is the provision of ongoing help to impoverished individuals (households). The main aim of the research presented in the paper is a multidimensional analysis of poverty risk in the framework of capability approach developed by Amartya Sen. The article also aims at defining the determinants of poverty which increase or decrease the risk of its occurrence, and comparing the poverty risk estimations obtained through the adoption of the multidimensional and unidimensional approach. The research concerning Poland (e.g. by regions) used data from the European Union Statistics on Income and Living Conditions (EU-SILC) survey, carried out by Statistics Poland in 2018. The measurement was operationalised by means of a special variant of structural equations modelling (SEM) – the multiple indicators and multiple causes model (MIMIC). The MIMIC model used deprivation symptoms that reflected the degree to which respondents were unable to satisfy their needs. The model also took into account the factors that increase or reduce the risk of poverty. The results of the estimations demonstrate that the level of education has the most significant impact on reducing the risk of poverty, while a low activity on the job market, i.e. being a pensioner or unemployed, increases the odds of becoming poor to the greatest extent. The regions most prone to poverty in multidimensional terms are: the Eastern macro-region, rural areas, pensioners’ households and non-family multi-person households. The aforementioned results have been compared with the measurement of poverty according to the unidimensional approach. The study shows the differences between the rankings of poverty risk by macro-region, size of the place of residence, source of household income and by household type from obtained through multidimensional and unidimensional approaches. This comparison demonstrates that the unidimensional approach is insufficient for the identification of the poor.