Very preliminary and incomplete. Comments welcome. NONPARAMETRIC ESTIMATION OF RETURNS TO SCALE

Abstract. Estimating the returns to scale is a classic problem of economics. In this paper we show how to efficiently estimate the returns to scale without making unnecessary parametric or exogeneity assumptions. Specifically, the production function and its degree of homogeneity are both assumed to be unknown and we show how to estimate the latter — which can now be interpreted as the returns to scale — as efficiently as possible when some or all of the factor inputs are endogenous. This is a nonstandard problem because the production function is ill-posed in this setting. Hence, care has to be taken while deriving the efficiency bound and constructing an efficient estimator for the returns to scale. 1