Despite the introduction of major social policy reforms, since the mid-1980s Germany has been experiencing a steady increase in poverty rates among general population. It gave a boost to a great number of studies trying to explain why it happened. By applying decomposition techniques they have been exploring what kind of changes in household structures and labor market conditions of households could have potentially contributed to the dramatic increase in poverty rates over the recent two decades. At the same time, the literature available in the field does not shed light on how poverty duration has been evolving over time as well as what kind of forces might be responsible for it.
Hence, this paper aims at exploring and explaining changes in the patterns of poverty duration in Germany over 1992-2010. To be more precise, the following set of questions will be addressed: (1) How the distribution of time spent by individuals in poverty evolved over 1992-2010? (2) Have the probabilities to exit (re-entry) poverty for spells with different durations changed over the period of interest and, if yes, how? (3) Which socio-economic and demographic characteristic of individuals and their households became more important determinants of poverty exits (re-entries) and which ones lost their significance with time? (4) What kind of social and economic policies introduced in 1992-2010 can be potentially responsible for the observed changes in the patterns of poverty duration?
The primary method applied for the empirical part of the paper is discrete-time survival analysis based on the joint modeling of poverty entries and exits, controlling for unobserved heterogeneity and initial condition problem. In order to capture changes in poverty durations over time and explore potential contributions of different forces to these changes a five-year moving window is created and used for the analysis. It permits to investigate when exactly a change occurred and what forces might be potentially responsible for it.
The data for the analysis is derived from the German Socio-Economic Panel. The waves covering 1990-2010 are used for the sample construction with the purpose to incorporate information on both Western and Eastern parts of Germany into analysis.