Standard poverty measures provide a snapshot of poverty at a single point in time but they cannot capture the distribution of poverty across time. This distribution matters at the individual level where it distinguishes between chronic and transitory poverty and at the aggregate level as it captures economic mobility across the poverty line. This paper proposes three new classes of poverty measures to account for how poverty is distributed across households over time. Each class of measures represents an intertemporal extensions of the standard Foster-Greer-Thorbecke poverty indices. The paper shows that each class of measures is an improvement on the standard practice of applying FGTs directly to intertemporal welfare data. Each class of measures requires choosing a parameter representing society’s preferences towards the intertemporal distribution of poverty. Choosing among them depends on the monitoring and evaluation problem of interest. The applications to panel data from rural Pakistan show that the different methods of accounting for the intertemporal distribution of poverty across households substantially increase estimates of poverty indices suggesting that the standard practice of using snapshot poverty measures underestimates the real level of aggregate poverty.