Rasch, Frisch, and Two Fishers: A Social History of the Econometric Origins of Some Widely Used Psychometric Models
William P. Fisher Jr.
University of California, Berkeley - BEAR Center
October 27, 2010
In 1979, the Danish mathematician Georg Rasch recounted a 1959 visit with his former teacher, and later economics Nobel Prize winner, Ragnar Frisch. At this time, Frisch prompted Rasch to formalize his work in a separability theorem. Previously unnoted is that Frisch’s close colleague, Irving Fisher, had decades before formulated an econometric "separation theorem," for which Frisch defined necessary and sufficient conditions of its satisfaction. Rasch attributes important features of his model to Ronald Fisher’s concept of statistical sufficiency, but he appears to have been prepared to find special value in sufficiency by two concepts circulating in Norway during his (Rasch's) time with Frisch. The first stems from Frisch's stress on structural invariance (referred to by him as autonomy), and the second from various economists' efforts to found econometric models in analogies from mathematical models in physics. Rasch’s models may have more significance for econometrics than has yet been accorded.
Number of Pages in PDF File: 39
Keywords: econometric models, history, biography, parameter separation, minimally sufficient statistics, structural invariance, Rasch models, Ragnar Frisch, Tjalling Koopmans, Ronald Fisher, Irving Fisher, Georg Rasch, University of Chicago, Cowles Commission
JEL Classification: B41, C10, C13, C20, C42, D70, E60, H54, I11, I21, I31, P11
Date posted: October 29, 2010 ; Last revised: February 8, 2011