This dissertation contains three chapters that are concerned with the fields of behavioral economics and industrial organization.

In the first chapter, I extend Regret Theory to a dynamic framework and identify conditions such that a choice overload effect and a status quo bias can arise in a sequential search model. Furthermore, I show that the optimal search behavior of the agent is non-stationary.

In the second chapter, I revisit and extend a classical model of vertical differentiation by assuming that firms are uncertain about their rival's fixed costs of quality improvements, measured by a cost parameter. This assumption gives rise to a competitive equilibrium, which is in stark contrast to the well-known result of maximum differentiation. In the competitive equilibrium, firms' quality levels are close to each other, whenever they have similar cost parameters. As a consequence, the firms' profits decrease, while consumer rent increases. Finally, I show that within the concepts of stochastic orders, the monotone likelihood ratio order is a very useful tool to study comparative statics in models of vertical differentiation when considering fairly general assumptions concerning the primitives. Among many results, I showed that an increase in the probability that consumers have a high willingness to pay for quality could increase the consumer rent for all types of consumers in the population.

The third and final chapter of this thesis is joint work with Kangkan Choudhury. This chapter is connected to the second chapter in that we embedded my second chapter’s model in a dynamic framework. We are concerned with how firms choose to innovate over time when they have uncertainty concerning their rival’s fixed costs of quality improvements in general and about their own fixed costs of quality improvement in the future. We find that a Markov Perfect equilibrium exists where, surprisingly, firms will never choose intermediate quality levels but will either stay at the lowest possible quality level or choose the highest possible quality level.