NONPARAMETRIC-TESTS OF REGULARITY, FARRELL EFFICIENCY, AND GOODNESS-OF-FIT

The goal of this paper is to reveal the links among several seemingly unrelated literatures: 1) the nonparametric regularity test literature (see Afriat, 1972; Diewert and Parkan, 1983; Hanoch and Rothschild, 1972; Varian, 1984), 2) the efficiency measurement literature (Farrell, 1957, is the classic reference, but also included is what has been dubbed Data Envelopment Analysis), and 3) the 'goodness-of-fit' approach associated with Varian (1990). The unifying threads are the Shephard-type distance function and the Mahler inequality (Mahler, 1939). Our synthesis yields regularity tests of cost minimization and technical change which are nonparametric, yet provide information as to goodness-of-fit based on Farrell-type efficiency measures.