Recent proposals to limit the participation of Verizon Wireless and AT&T in the upcoming broadcast TV spectrum incentive auction have rekindled the debate on the interactions between auction design and competition policy. Existing literature analyzes the efficiency of spectrum auctions by focusing on auction outcomes, noticing that there is little immediate post-auction resale in the secondary market for spectrum. However, the effects of specific pro-competitive auction policies on the market structure can only be fully understood by considering long-run market outcomes.

This dissertation provides a historical overview of the market for broadband PCS spectrum and then documents the construction of a unique and first comprehensive dataset of firms' spectrum licenses holdings. I then use this dataset to investigate the long-term effects of auction restrictions similar to the ones recently proposed.

I document how eligibility restrictions for entrepreneurs in the early years of the market for broadband PCS spectrum initially create two separate markets, the gradual opening of the restricted market, and its eventual convergence with the unrestricted one. Second, and contrary to the existing literature, the detailed nature of my data shows there was a significant level of market activity immediately following the auctions, a portion of which can be attributed to specific preferential treatment policies used. To investigate the actual transition of restricted spectrum from entrepreneurs to large companies, my identification strategy uses the institutional design of eligibility restrictions, resulting in considerable differences in the observed build-out behavior of entrepreneurs and large companies. While the latter have no incentive to meet the construction requirements ahead of time, doing so allows entrepreneurs to potentially sell their licenses to large companies. Results show a consistent pattern of entrepreneurs building out their licenses early only to immediately sell them to large companies.

By first constructing a unique spectrum license dataset and then exploring the effects of preferential treatment provisions used in spectrum auctions for PCS licenses, this dissertation provides evidence that these provisions did not achieve their long-run goal in terms of the desired market structure. Instead, there is strong evidence of opportunistic behavior of entrepreneurs, who get pulled in because of eligibility restrictions and then quickly resell their licenses to large companies. In general, it appears the presence of secondary markets undoes the outcomes generated in auctions (and desired by the policymaker). Because of this, FCC should consider regulating secondary markets in line with their existing regulatory practices rather than imposing auction restrictions, i.e. regulating auctions.

This dissertation consists of two chapters. In the first chapter, which is co-authored with Marcus Studart, we study the interaction of information and competition in incentivizing firms to produce high quality. We estimate a discrete quality choice game, using restaurant hygiene inspection data in Los Angeles County, from 1995 through 1998. Our results show that information is sufficient for competition to have an effect on quality provision. We also find that after the mandatory disclosure of information to consumers, a restaurant's equilibrium quality increases in the number of competitors up to a certain threshold. Beyond this threshold, an additional firm has a negative effect on quality choice. This result may indicate that too much competition reduces the returns of quality provision, which we interpret as a result of firms having more difficulty retaining customers.

In the second chapter, I investigate the notion of "hype" in the U.S. motion picture industry, which occurs whenever an upcoming film is heavily advertised irrespective of its underlying quality. I address the questions of 1) whether hype is a measurable phenomenon in data from the movie industry, 2) why the practice of hype is feasible, and 3) how the feasibility of hype impacts the movie industry. I first search for evidence of hype using a data set of weekly advertising expenditures, box office revenues, critical review outcomes, and movie characteristics of wide release theatrical films from 2003 to 2012. Using critical review outcomes as a measure of underlying quality, I observe that advertising levels are similar across film quality levels. I find that advertising levels fall dramatically between the pre-release and post-release phases for all movies, but they decay much more sharply for low quality films.

Building on this finding, I exploit weekly variation in the data by using regression analysis and a matching difference-in-differences estimator to quantify the causal effect of release and high critical reviews on advertising behavior. I find that advertising prior to a film's release is not statistically related to underlying film quality, but distributors with movies revealed to be high quality spend 62% more on advertising after release than other films. I then build a theoretical model of advertising on the movie industry based on Butters (1977) which incorporates consumer verification of film quality as a key feature to support this result. I provide conditions that ensure the existence and uniqueness of an advertising equilibrium in the model. I compute numerical examples of advertising equilibria and show that under certain parameter values, the distributors of bad films have no incentivize to advertise if consumers are sufficiently informed of quality. These findings suggest that consumer access to critical reviews could explain the feasibility of hype in the movie industry.

Advances in technology have made interactions between previously isolated markets, increasingly prevalent. Even so, much of the research on firm competition in economics has focused on independent markets and how firms compete within these distinct markets. This dissertation extends the literature by studying competition in settings with interrelated markets.

Chapter 1 studies how firm expansion into multiple geographic markets has affected local market competition. As a case study I examine the banking industry, where deregulation in the early 1990s encouraged banks to expand their branch networks into multiple markets. I estimate a model of branch entry that explicitly allows for spillovers across markets, which in banking include demand advantages in attracting more consumer deposits, cost advantages from economies of scale or density, or a diversification of risks. To do this I use a revealed preference approach that also deals with unobserved firm and market heterogeneity. I find that spillover benefits explain 20% of the branches built in the observed equilibrium. The additional observed branches increase consumer surplus, and in most markets lead to a larger share of deposits being collected by banks as opposed to non-bank alternatives. The exception is in the largest markets, where multi-market banks overemphasize competing with additional branches, as opposed to offering better service or higher deposit rates. Because of the lower deposit rates, this leads rich customers, whom are more price sensitive then the average consumer, to switch to outside options such as disintermediation or credit unions. This effect costs banks $115 billion in lost deposits.

In chapter 2, I study auction markets where bidders are competitors in some downstream market. I do this by extending the auction estimation literature to an auction model with externalities. In such a model, in addition to each bidder having a private value for the object, which they receive if they win the auction, some bidders, upon losing, will receive a negative externality that depends on which of their rivals has won the object in their stead. I identify and estimate the externality values by using structural auction estimation techniques to estimate bidder valuations as functions of the negative externalities, and then using variation in the sets of competitors to infer the externality values. I introduce three different estimators for the externalities and provide Monte Carlo results for each.

Two events in 2008-2009 had sweeping effects on the global economy and the US automotive sector in particular: the contraction of credit in 2008-2009 and the massive run-up in crude oil prices in late 2008. In 2008, as credit tightened, auto demand declined, and General Motors and Chrysler edged towards bankruptcy, begging the question: "What role does credit have in determining auto market outcomes?" In late 2008, crude oil prices rose sharply, leading many to ask "How will consumers respond to rising gasoline prices?" This dissertation addresses both of these questions.

In chapter one, I analyze car manufacturers and their lending subsidiaries, or "captive finance companies." I note that via their captive finance companies, car manufacturers compete against traditional banks and are majority lenders in the auto lending market. Government agencies like the Consumer Financial Protection Bureau and the Federal Deposit Insurance Corporation have recently expressed interest in captives and auto lending. They have asked for instance whether captives affect the market price for risk. This leads me to my research question: "How does the use of captive finance affect competition and risk in the auto lending and new vehicle markets?" I specify an equilibrium model of the new vehicle and auto lending markets. I estimate parameters from this model using the universe of loan originations and vehicle transactions from a 20% of all US vehicle dealerships. I find that consumers are more responsive to loan principal than to changes in interest rates. Using the estimated model to conduct a counterfactual in which captives cease to exist, I also find that competition between captives and traditional banks leads banks to under-price risk.

In chapter two, I ask "How do consumers' vehicle choices respond to changes in their expectations for future gasoline prices?" The 2008 spike in gasoline prices resurfaced questions asked about consumer gasoline consumption and vehicle choice which were originally posed in the 1970's. Since that time, economists have modeled consumers' dynamic responses to gasoline prices. A necessary component of such models is the specification of consumers' expectations over future values of relevant variables like gasoline prices. As data on gasoline price expectations were unavailable, researchers had been forced to make assumptions on how consumers form expectations over future gasoline prices. To address this, I introduce entirely novel data on consumers' gasoline price expectations and stated vehicle preferences. Using these data, I find that consumers' vehicle choices do respond to their expectations for future gasoline prices. Consumers vehicle preferences are persistent, however, which attenuates the effects of gasoline prices on vehicle choices.

This collection of essays considers three issues regarding the performance of the structural merger simulation. The first chapter addresses the underprediction problem of merger simulation by considering more flexible demand models. To be as flexible as possible in modeling the demand, I estimate the full variance--covariance structure of brand preferences and let the brand intercepts to be different across cities to exploit the multi--market structure of data. Compared to a simple logit demand model, a flexible setting allows the predicted price changes to be closer to the actual price changes. In Chapter 2, because the underprediction problem still exists with the flexible structural demand model, I consider the Almost Ideal Demand System as an alternative demand model to determine how the simulation performance differs from that of Chapter 1. I also address the firm-side modeling issues in Chapter 2. I examine, whether incorporating the retailer side in the pricing game would produce improved merger simulation results. In Chapter 3, I separately estimate the effect of several factors compounded in the actual price changes in the post--merger period. I explain that part of the reasons the merger simulation generally fails to result in accurate predictions is that both demand--side factors and supply--side factors considerably affect the post--merger prices.