SOEPpapers on Multidisciplinary Panel Data Research 857
In general, poverty measures are estimated by applying income information. However, only using income data for calculating relative poverty might lead to an incomplete view. For example, a household can be under a poverty threshold even if a household member owns real estate or equity. In this thesis, at risk of income poverty in Germany is estimated. In order to get a more complete picture of at risk of poverty, a multidimensional approach is applied. Not only at risk of income poverty, also at risk of wealth poverty is measured. Moreover, households that are both at risk of income and wealth poor are analyzed. Furthermore, several poverty groups are identified: twice-poor which are households that are, at risk of income and wealth poverty; protected-poor, households that are at risk of income poverty but not at risk of wealth poverty; vulnerable-poor, households which are at risk of wealth poverty but not at risk of income poverty; non-poor, households which are either at risk of income poverty nor at risk of wealth poverty. Poverty profiles in Germany and their changes over time are analyzed for the years 2002, 2007 and 2012. In fact, it is investigated to which degree at risk of poverty rates differ in socio-economic characteristics. A logit regression is applied for each dimension and each wave for estimation. For robustness checks, 95 percent bootstrap confidence intervals are calculated for all results. Findings suggest that young age, region East Germany, single, lone parent, unemployment and low education are factors that condition the at risk of poverty rates. The definition of a certain rate influences the percentage of households that are affected by at risk of poverty, however, has a limited effect on poverty profiles. Poverty profiles have not changed over time but some factors such as unemployment and low education have significantly increased.