Economic base analysis is frequently used to describe employment profiles and to predict project-related impacts in small communities. Considerable evidence suggests, however, that economic base multipliers should be estimated from survey data and not from shortcut methods. In this paper two competing versions of the economic base model are developed and then these two models are estimated by use of the Arizona community data set. In both cases, marginal multiplier estimates, controlled for transfer payments, are generated for ten individual sectors in five different types of communities. Results from these two disaggregate economic base models are assessed and then compared with results provided earlier by more aggregate models. The better of these two new models closely resembles the popular input—output model.