Models of health plan choice

This paper considers alternative approaches to accounting for consumer heterogeneity in econometric models of health plan choice. In particular, we compare and contrast approaches used by health economists and marketing researchers to explicitly allow for consumer heterogeneity in assessment of plan provider quality and characteristics. As a result, we are able to identify conditions under which approaches utilized by health economists are preferred over those utilized by marketing researchers, and vice versa. For example, we find that in the absence of data on prior year choice, which is often the case of models estimated in the health economics literature, it would be valuable to estimate random coefficient models, which have not been estimated in such works. And, in the presence of data on prior year choice, which is often the case of models estimated in the marketing literature, a model wherein consumer characteristics simply interact with plan specific constants and premium, and which has not been estimated in that literature, offers better fits than random coefficient, semi-parametric random coefficient, and latent class models which have been previously estimated in that literature. The model offers important managerial insights useful for targeting benefits of various plans to different customer segments.