Correcting for endogeneity using the EMIS method: a case study with revealed preference data

Endogeneity is an important issue that often arises in discrete choice models leading to biased estimates of the parameters. We propose the extended multiple indicator solution (EMIS) methodology to correct for it and exemplify it with a case study using revealed preference data about mode choice in Switzerland. The same data is used as one of the first applications of the multiple indicator solution (MIS) method. These two methodologies - EMIS and MIS - are then compared to an integrated choice and latent variable model (ICLV), and a model without any correction. In order to compare the different methodologies between them, the value of time and the time elasticity in public transportation estimates are computed and reported for each of the methods.