Weak Separability in Applied Welfare Analysis

摘要: The focus of this paper is the impact on welfare measurement a common assumption in applied demand analysis, weak separability. A standard empirical practice to estimate conditional models for set goods which are assumed be weakly separable from all other with expenditure as right-hand-side variable, tacitly assuming that predetermined (Alston and Chalfant 1987, 1991; Alston, Carter, Green, Pick; Blackorby, Boyce, Russell; Blisard Blaylock; Brown Heien; Cashin; Choi Sosin; Clements E. Selvanathan; S. Selvana-than; Deaton 1975; Goddard Amuah; Gould, Cox, Perali 1990, Heien Pompelli 1988, 1989; Murray; Safyurtlu, John-son, Hassan; Theil 1976, 1980; Van Ko-oten; Yen Chern). Once estimated, such often used cal-culate exact measures, or changes true cost living index, economic effects consumers due prices under study (Heien; Pom-pelli Kooten; Blaylock). This shown here produce biased results. There two sources bias both can traced inappropriate treatment predetermined. In an empiri-cal application U.S. dairy program, comparison estimates incomplete system specification versus suggests negative around 17% compensating variation eliminating price discrimination federal milk marketing orders 1990. Results Monte Carlo simulation suggest downward closer 40%. I first present analysis separability, then application, summary conclusions. (This abstract was borrowed another version item.)