We model the dynamics of social assistance benefit receipt in Britain using data from the British Household Panel Survey, waves 1–15. First, we discuss definitions of social assistance benefit receipt, and present information about the trends between 1991 and 2005 in the receipt of social assistance benefits, and in annual rates of transition into and out of receipt. Second, we review potential multivariate modelling approaches especially the dynamic random effects probit models that are used in our empirical analysis and, third, discuss sample selection criteria and explanatory variables. Fourth, we present our regression estimation estimates and interpret them. The final section contains a summary of the substantive results, and highlights some lessons concerning application of the analysis for other countries and some methodological issues.