Test of Choice Set Misspecification for Discrete Models of Consumer Choice

摘要: We develop and evaluate a test of choice set misspecification for multinomial logit model. This determines whether the designated by researcher mistakenly assigns relevant substitutes to numeraire good. this generalizing traditional McFadden-type conditional model is conditioned on an overly restrictive substitution possibilities. The has convenient feature: while it requires information potentially relevant, but omitted, substitute goods, does not require observe consumers choices among these omitted potential if they select good (which contains substitutes). A comparison our more general suggests that may produce biased parameters distort welfare estimates consequential extent.