Development and estimation of a semi-compensatory model with a flexible error structure

In decisions involving many alternatives, such as residential choice, individuals conduct a two-stage decision process, consisting of eliminating non-viable alternatives and choice from the retained choice set. In light of the potential of semi-compensatory discrete choice models to mathematically represent such decisions, research is inching ahead with the aim of alleviating their high computational complexity and their severe restrictive assumptions. To date, still a major barrier for the implementation of semi-compensatory models is their underlying assumption of independently and identically distributed error terms across alternatives at the choice stage. This study relaxes the assumption by introducing nested substitution patterns and alternatively random taste heterogeneity at the choice stage, thus equating the structural flexibility of semi-compensatory models to their compensatory counterparts. The proposed model is applied to off-campus rental apartment choice by students. Results show the feasibility and importance of introducing a flexible error structure into semi-compensatory models.