This paper uses longitudinal data from the BHPS, waves 1-7, to document low-income dynamics for individuals living in Britain in 1990s. Poverty entry and exit hazard rates are estimated and used to calculate the distribution of time spent poor over a six-year period. The results underline the importance of accounting for individuals repeated spells of poverty when measuring poverty persistence. Using discrete-time proportional hazard rate models, the paper then seeks to explain and forecast the observed chances of exit/entering poverty and the distribution of time spent in poverty for individuals with selected characteristics. The socio-economic correlates of the observed poverty patterns are investigated, including the relative importance of both household and individual characteristics.