The main purpose of this study is to find the optimal automobile market structure for the semi-medium and medium-sized car markets in Korea using the three stage nested logit model with explanatory variables derived from the previous study. In addition, it is intended to increase the understanding of the car market structure by post analysis using marginal effect and elasticity. Gender and age were set as individual characteristics, and fuel efficiency divided by fuel price, size, and car price divided by household income were set as alternative attributes. In addition, six hypothetical market structures having a three-stage hierarchical structure were established according to criteria such as segment, vehicle type, and corporate brand. After estimating the model using the new car buyer's purchase data, the goodness of fit, statistical significance of variables, and inclusive values for each hypothetical market structure were evaluated. The author found that the semi-medium and medium-sized car markets in Korea were structured, and also found that the segment-corporate brand-vehicle type structure is the optimal market structure. Among the alternative attributes put into the utility function, fuel efficiency/fuel price, size, and price/income were all found to be significant. Among individual characteristics, gender affects most alternative choices, while age has a limited effect. In order to interpret the effects of alternative attributes and individual characteristics on each alternative, marginal effect analysis for discrete variables and elasticity analysis for continuous variables were conducted.