Non-Convex Technologies and Cost Functions: Definitions, Duality and Nonparametric Tests of Convexity

This contribution is the first systematic attempt to develop a series of nonparametric, deterministic technologies and cost functions without maintaining convexity. Specifically, we introduce returns to scale assumptions into an existing non-convex technology and, dual to these various technologies, define non-convex cost functions. These non-convex cost functions are never lower than their convex counterparts. Both non-convex technologies and cost functions (total, ray-average and marginal) are characterised by simple, closed form expressions. Furthermore, a local duality result is established between a local cost function and the input distance function. Finally, nonparametric goodness-of-fit tests for the convexity axiom are developed. This is a first step towards making convexity a statistically testable hypothesis.