In this article, we discuss and analyse poverty in Germany from a longitudinal perspective. Using data from the German Socio-Economic Panel Study (SOEP), we first show that the general poverty rate in Germany has constantly increased since the late 1990s. Shifting to a life-course perspective, we show that not only socio-structural characteristics have a strong impact on the poverty risk, but also critical life-events. While focusing on dynamics of poverty within individual life-courses, it appears that incidents like formation of a 'new household, birth of a child and separation from partner are associated with an immediate increase of the poverty risk. The event of becoming unemployed stands particularly out. Comparing longitudinal and fixed-effects approaches on the one hand with simple cross-sectional procedures on the other, our analyses finally emphasises that cross-sectional analyses are not sufficient to fully understand or to explain poverty. Therefore, our study can be interpreted as a claim to make stronger use of the benefits of longitudinal data in the context of poverty research.