Hi.
I am running nlogit, with two levels. All dissimilarity parameters I obtain (they are 8) are bigger than 1 (implying that the model is inconsistent with RUM).
According to the Cameron Trivedi ("Microeconometrics using Stata") this is not an unusual finding for NL models. I read in the article "Specification and Estimation of the Nested Logit Model: Alternative Normalisations" of Hensher and Greene (2000) that this may be due to the normalization process used.
I am using Stata12 and the normalized parameterization (the default option). Is there any way to fix the problem?
Thank's,
Simona
I am running nlogit, with two levels. All dissimilarity parameters I obtain (they are 8) are bigger than 1 (implying that the model is inconsistent with RUM).
According to the Cameron Trivedi ("Microeconometrics using Stata") this is not an unusual finding for NL models. I read in the article "Specification and Estimation of the Nested Logit Model: Alternative Normalisations" of Hensher and Greene (2000) that this may be due to the normalization process used.
I am using Stata12 and the normalized parameterization (the default option). Is there any way to fix the problem?
Thank's,
Simona
Comment