In this paper we use nonparametric statistical tools to quantify motion-picture profit. We quantify the unconditional distribution of profit, the distribution of profit conditional on stars and sequels, and we also model the conditional expectation of movie profits using a nonparametric data-driven regression model. The flexibility of the nonparametric approach accommodates the full range of possible relationships among the variables without prior specification of a functional form, thereby capturing nonlinearities and interactions without introducing possible specification bias. We find that marginal returns to budgets and opening screens vary over the domain of these variables. We also find that the conditional distribution of movie profit and the expected level of profit are related to the use of movie stars and sequels.
