Badanie dynamiki ubóstwa w Polsce z wykorzystaniem modeli analizy historii zdarzeń o czasie dyskretnym

Studies on poverty are based predominantly on cross-sectional analysis. Including a time dimension in the analysis allows us to better understanding the dynamics of poverty. In the long run, a unit (individual, household, family) can enter and exit poverty several times. An analysis of determinants of these events allows us to identify groups of households which are particularly likely to enter into poverty, and those with a high chance of exiting poverty. The main aim of this article is to identify determinants of transitions into and out of poverty in Poland in 2000–2011. To achieve this, I use logit regression models for discrete-time event history analysis. I present two specifications of the model, one with number of years spent outside poverty and a second with additional selected socio-economic characteristics of household and household head. The results of estimation suggest that the amount of time spent out of poverty or in poverty does not have a significantly effect, with the exception of one case, on the probability of a change in the status (in poverty/out of poverty). In the case of entries into poverty, including additional variables improves the model. I find that a change in probability of entry into poverty is significantly affected by age and education of household head, place of residence and labour force status of household. However, expanded model of poverty exits is worse than the base model. None of the included variables are statistically significant.