Partial-moment functions are proposed as a flexible way to characterize and estimate asymmetric effects of inputs on output distributions. Methods for econometric estimation of partial-moment functions, and tests for input symmetry and location-scale distributions, are presented. A Monte Carlo study demonstrates properties of proposed tests. A study of Ecuadorian potato production illustrates the methods. Hypotheses of input symmetry and location scale are rejected. A risk-value model based on partial moments implies that fertilizer is risk increasing and fungicides and labor are risk reducing in potato production, whereas an expected utility model based on full moments has the opposite implications. Copyright 2010, Oxford University Press.