If habitats contract or get fragmented, these reduced habitats cannot continue to support the same biodiversity as they did before, so they must lose species. While some species will be lost immediately, others can persist for a long time before going extinct. Indeed, areas recently protected may continue to lose species long after habitat loss has stopped. So, just because we don’t observe extinctions right away does not mean that everything is going to be OK. This phenomenon, often called “extinction debt”, also has implications for how ecosystems respond to climate change and to ecological invasions. A major scientific challenge is to understand the dynamics of this effect: how long before the extinctions are complete and which parameters drive it? Answering these depends to a large extent not only on having good empirical results but also on having a sound theoretical model of the community. I will be reviewing our current state of knowledge in this area and what case studies and larger meta-analyses have revealed. I will also discuss research on various models of the process and how our theoretical results can be used to design improved protected areas.