This paper examines outsourcing to test whether productivity-enhancing specialization is facilitated in bigger cities. First, the paper provides a theoretical model which shows that greater local demand for a given input promotes the entry of suppliers into a city; the increased number of suppliers then results in lower outsourcing prices and a higher use of outsourcing by final producers, therefore reducing the final producers' production costs. I then test the predictions of the model by examining manufacturing plants' practices of outsourcing business services, by using plant-level data from the 1992 Annual Survey of Manufactures. The empirical results show that an exogenous increase in local demand promotes the entry of service suppliers and increases a firm's probability of outsourcing for white-collar services. In particular, I found that doubling the intensity of the use of a service in a U.S. county, which can be attributed to the industrial composition of the county, results in a 7% to 25% increase in the probability of outsourcing.