Conservation policy requires reliable estimates of extinction rates that consider the interactions between population size (N) and habitat area. Current approaches to estimating extinction from the endemics-area relationship (EAR) estimate only the minimum number of species that can become extinct due to habitat loss (instantaneous extinction). EARs will therefore underestimate extinction if small populations and/or habitat area (SPHA) commit species to future extinction. We demonstrate this mathematically, by assuming species require a minimum population size of two individuals, and by randomly sampling habitat loss within stem-mapped forest plots. We then develop a general framework for incorporating SPHA effects into EARs that builds upon recent advances introducing N into estimates of extinction. By accounting for effects that modify N, our framework explains extinction debt and reduces the uncertainty associated with future estimates of extinction through carefully qualifying the spatial and temporal context of predictions.