Spatial Discrete Choice and Spatial Limited Dependent Variable Models: a review with an emphasis on the use in Regional Health Economics

A.G. Billé,G. Arbia
DOI: https://doi.org/10.1111/joes.12333
2013-02-10
Abstract:Despite spatial econometrics is now considered a consolidated discipline, only in recent years we have experienced an increasing attention to the possibility of applying it to the field of discrete choices (e.g. Smirnov, 2010 for a recent review) and limited dependent variable models. In particular, only a small number of papers introduced the above-mentioned models in Health Economics. The main purpose of the present paper is to review the different methodological solutions in spatial discrete choice models as they appeared in several applied fields by placing an emphasis on the health economics applications.
Applications
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is to apply the methods of spatial econometrics to discrete - choice models and limited - dependent - variable models, especially in the field of health economics. Although spatial econometrics is already a relatively mature discipline, its application in discrete - choice models and limited - dependent - variable models is relatively scarce. By reviewing different methodological solutions, especially the application of these models in health economics, the author aims to fill this gap. Specifically, the paper focuses on the following aspects: 1. **Spatial discrete - choice models**: The paper elaborates on various methodological solutions for spatial discrete - choice models, including binary - choice models (such as spatial binary Probit/Logit models), ordered - choice models (such as ordered Probit/Logit models and their variants), multinomial - choice models (such as multinomial Probit/Logit models and their variants), and count - data models (such as spatial Poisson/negative binomial models and their variants). 2. **Bayesian methods**: The paper also discusses the application of Bayesian methods in spatial discrete - choice models, including Bayesian spatial discrete - choice models. 3. **Limited - dependent - variable models**: The paper introduces different limited - dependent - variable models and explores the application of these models in health economics. 4. **Estimation problems**: The paper pays special attention to the estimation problems of these models, including how to avoid heteroscedasticity, how to utilize the off - diagonal information of the covariance matrix, and how to avoid calculating the inverse matrix. For example, for the heteroscedasticity problem, the paper mentions the methods of Case (1992) and Pinkse and Slade (1998); for the multi - dimensional integration problem, the paper discusses techniques such as simulated maximum likelihood estimation (SML) and recursive importance sampling (RIS); for the inverse matrix problem, the paper introduces the linearized GMM method proposed by Klier and McMillen (2008). 5. **Practical applications**: The paper emphasizes the practical applications of these models in health economics, such as describing the unequal distribution of medical resources within a region and studying the relationship between the geographical distribution of specific diseases and environmental factors. In short, the main purpose of this paper is to promote the application and development of spatial discrete - choice models and limited - dependent - variable models in health economics by reviewing and analyzing existing methodological solutions.