文摘
This dissertation presents three topics in dynamic discrete choices. Chapter 1 sets up a dynamic model of demand to identify consumers preferences for "newness" of products in a new durable goods market,namely golf drivers market. Forward-looking heterogeneous consumers with preferences for newness of products decide when and what to purchase. The model also accounts for the fact that the market is highly subject to seasonal fluctuations. Using the aggregated data from the US golf drivers market the model succeeds at identifying consumers preference for newness of products when the seasonality and quality differences are controlled for. Experiments with different assumptions are performed to confirm the robustness of the model. Chapter 2 is a joint essay with Yoonseok Lee. This chapter finds that the state legislation decision on the mandatory motorcycle-helmet-use law is affected by the neighboring states choices. It turns out that such a social interaction is one of the key factors in this decision making,whereas the fatality rate from motorcycle-related accidents is not so. Using the US state level panel data,the analysis is conducted by developing a mixed proportional hazard model with grouped data,which allows for possible cross sectional duration dependence. It explains a behavioral aspect of the legislative decision making procedure i.e. social interactions) and empirically shows how the proximity between agents affects the decision making. Chapter 3 deals with a dynamic panel data model where the dependent variable is latent while only its ranking among individuals is observable at each time period. It sets up a dynamic panel data model where the latent dependent variable is characterized by its ranking in the previous period and current exogenous variables along with individual heterogeneity. To overcome a small sample problem when the number of individuals is large and the number of time periods is small,it uses the explosion property of logit models with ranked data. As an application,it applies the econometric model to the panel ranking data of the best states for business announced by Forbes magazine. It finds a significant relation between lagged ranking,along with selective covariates,and current business environment.