Estimability in the multinomial probit model

摘要: Random utility models often involve terms which represent alternative-specific errors, and the main attractive feature of multinomial probit (MNP) model is that it allows a rather general covariance structure for these errors. However, since observed choices only reveal information regarding differences, scale cannot be determined, not all parameters in an arbitrary MNP specification may identified. This paper examines identification restrictions arise linear-in-parameters framework, provides discussion recommendations estimation analysis normalizations.