It is common to analyse poverty data broken down by household or economic status. Implicitly, it is assumed that people change state (for example, single, married, children, no children) for exogenous reasons. If we bring economic behaviour into the problem, then such transitions become endogenous. The data are then insufficient to identify the claims made from them. The distribution of the characteristics of the individuals in the states will be endogenous, and the state average poverty rate will depend on the composition of the individuals in the state as well as on the economic impact of being in that state per se. In this paper, we set out a simple model with endogenous transitions to make our point, and apply this to Family Expenditure Survey data for Britain. We show that our argument has empirical content for Britain.