This dissertation thesis contains three chapters in the fields of empirical industrial organization and applied econometrics. In my first chapter, I combine complementarity in bundle choice and consideration set of products into a demand model for differentiated products. I explore the identification when the market shares of products, not bundles are observed, and use a novel estimation approach via combining the loglikelihood of consumer choices and market shares with moment conditions. I apply the model to the yogurt market with consumer-level and store-level data. I find a considerable demand synergy in the bundle of different products and a significant proportion of consumer inertia, defined here as choosing from the last purchases, in consumer demand for yogurt. Compared to the standard discrete choice model, my estimation results suggest that accounting for complementarity between products and consumers’ limited consideration set can substantially affect price competition analysis. The second chapter is a collaborated work on the application of spatial econometrics in banking industry. We examine the direct and indirect impacts of natural disasters on deposit rates of bank branches during the 2008 – 2017 period. We find that spatial spillover effects substantially explain the total impact for deposit rate-setting branches. Our analysis and findings contribute to the existing literature by showing that the responses of branches to natural disasters are not confined only to those branches in counties directly affected but to branches in neighboring counties through competitive effects. Our results also confirm that spillover effects occur among branches across counties via a social connection and geographical network. The third chapter is on the identification and inference of a discrete choice model with partially unobserved attributes. It is motivated by the observation that consumers often do not account for unplanned purchases in their store choice when planning grocery trips. I show that the model is partially identified, and the sharp identified set is characterized via both moment-based inequalities and likelihood-based criteria. Using household grocery shopping data, I show that point estimation from the standard multinomial choice model assuming no unplanned purchases is rejected.