This dissertation contributes to three distinct literature in Applied Microeconomics and Applied Microeconometrics. Chapter 1 empirically investigates the determinants of sellers' optimal pricing formats using data from the used car retail market. The focus is on the two popular pricing formats of posted price and haggling. A structural demand model is developed to estimate of dealers' profits under posted price and haggling without having to observe the actual transaction prices. A few counterfactual experiments based on the estimated structural model are conducted to investigate the conditions leading to the heterogeneous choices of pricing formats that we observe. The experiment results show that posted price tends to be the superior pricing format when a seller has significant market power.

Chapter 2 proposes a new hypothesis for explaining the voluntary exchange of proprietary information between insurance companies. It is well-known that information asymmetry in the insurance market induces the so-called adverse selection problem, and which in turn could lead to very inefficient market outcomes and large social welfare loss. However what has been often ignored is that insurance companies obtain considerable information about their customers through transactions with them. Therefore over time the informational asymmetry may largely unravel if the insurers were to voluntarily share such privately gained information with each other. Such voluntary information exchange institutions actually have been established in the US and many other countries for a few decades, though why such institutions can be sustained in the market is still an unsettled question. This chapter develops a game-theoretic model to shed light on a new perspective for understanding such voluntary information exchange existing in the insurance market and many other markets. The key insight is that sharing the private customer information commits insurers to low premium for low risk customers in the future, and withdrawing from such an institution would necessarily attract disproportionately less low risk customers and more high risk customers. It is shown in this chapter that this type of consumer adverse selection can provide enough discipline to support information sharing in the equilibrium.

Chapter 3 provides a new method for estimating production functions with robustness against errors in the proxy variables. The estimate of production functions is a key input in the empirical analysis of productivity. The difficulty in obtaining consistent estimate of production functions has been how to control for the unobserved productivity that varies across firms. One important approach originally proposed by 011ey and Pakes (1986) is to use proxy variables to control for the latent productivity, under the restrictive assumption that the latent productivity is the only unobserved factor affecting the proxy variables. This chapter develops a new method for estimating production functions relaxing the assumption of scalar unobservable determinant for the proxy variables.

Keywords: Pricing Format, Information Sharing, Insurance Market Production Functions, Measurement Error