Cities exist because of the productivity gains that arise from clustering production and workers, a process called agglomeration. How important is agglomeration for aggregate growth? This paper constructs a dynamic stochastic general equilibrium model of cities and uses it to estimate the effect of local agglomeration on aggregate growth. We combine aggregate time‐series and city‐level panel data to estimate the model's parameters via generalized method of moments. The estimates imply a statistically and economically significant impact of local agglomeration on the growth rate of per capita consumption, raising it by about 10%.