Copula-based multiple-discrete/continuous choice models

Abstract

Explaining consumer choices with mathematical model has been an objective not only for economists but also for marketing scientists. Starting from microeconomic utility theory, the consumer choice model has evolved in terms of what kind of choice outcomes the model can handle. Earlier consumer choice model can deal with multinomial choice which allows only one choice among choice set (pick only-one choice). Nowadays, the model can handle multiple choices which allow any possible combinations among choice set (pick any-alternative choice). In addition, the marketing scientists start incorporating the quantity information into the consumer choice model. We call this more general type of choice outcome as multiple-discrete/continuous choice. Because the multiple-discrete/continuous choice outcome contains both alternative choice and quantity choice, it is a compound of various factors. The factors can be observed marketing driven such as price and promotions. On the other hand, the factors can be unobservable such as budget, economic reasons, size of car, size of storage, characteristics of store, etc. Also, these unobservable factors might be correlated each other. Therefore, we must incorporate these correlations when it comes to explaining multiple-discrete/continuous choice outcomes. Therefore, this study aims to develop a new approach for multiple-discrete/continuous choice outcomes with allowing various correlations between the unobservable factors. The dissertation consists of three studies. First study and second study suggests new copula-based choice models for multiple-discrete/continuous choice outcomes. Third study is an application of the first study to investigate marketing problem, especially the reference price effect. First study is about copula-based multiple-discrete choice model. It not only handles multiple-discrete choice outcomes but also allows the various correlations between unobserved factors. Three types of correlations are introduce...