I study the degree of returns to scale in aggregate U.S. data by estimating the standard RBC model with and without variable capacity utilization using Bayesian methods. The source of increasing returns is the presence of an externality in production. I find that, at best, returns are mildly increasing at a level statistically indistinguishable from the constant-returns case. A high enough degree of increasing returns is needed to support equilibrium indeterminacy and sunspot-driven business cycle fluctuations. In my estimation results, I find that I can conclusively rule out this possibility on account of a low estimated degree of increasing returns and a fairly inelastic labor supply.