| The dissertation comprises three self-contained essays in the general field of industrial organization, and focuses on two sources of market imperfections: Consumer search costs in the automobile insurance industry and consumer myopia in the banking industry.;In Chapter 1, "Estimating the Effects of Non-informative Advertising when Consumers Search Sequentially," I use the auto insurance industry as a case study to establish that consumer search costs have fallen significantly, at least in some industries. I rigorously establish that search costs have been falling over the last two decades in the private passenger auto insurance industry of California, while I simultaneously estimate the sensitivity of the probability that a firm is visited to the advertising levels of that firm and of its competitors. This requires that I estimate the consumer search costs distribution function. The challenge is to undertake an estimation procedure that generally requires detailed consumer-level data, when we only have highly aggregated firm-level data. I make use of Hortacsu and Syverson's (2004) model, which addresses this problem. Consistent with a priori expectations, I find that overall consumer search costs have been falling over time, especially between 2005 and 2010. The results also confirm that consumers are more likely to visit a firm that advertised in the past, and that this advertising effect is increasing over time.;Chapter 2, "Dynamic Choice of Non-informative Advertising when Consumers Search Sequentially: A Numerical Approach," builds on the findings of chapter 1 by providing a theoretical model that shows how decreasing consumer search costs can lead to increased advertising activity, when firms produce a homogeneous product. I establish reasonably weak conditions under which firms actually choose optimally to increase advertising in response to a decrease in search costs. I do so by showing that the relationship between advertising and search costs is not monotone because the relationship between search costs and the firms' profitability is first not monotone. I use a dynamic model, where firms choose advertising to affect their market exposure, the state variable; and choose prices in a product market, where consumers with heterogeneous search costs search sequentially. In short, I embed a version of Carlson and McAfee's (1983) demand model into Doraszelski and Markovich's (2007) dynamic advertising framework. The results of chapter 2 have significant policy implications, as they suggest that though markets may ultimately become very competitive at very low levels of search costs, markets may first have to pass through stages of being highly concentrated as search costs fall.;Chapter 3, "Overdraft Pricing and Myopic Consumers," investigates whether the presence of myopic consumers impacts how banks choose the prices of bank overdraft services. Gabaix and Laibson (2006) propose a model where banks optimally choose to suppress information on the price of add-on services, and charge prices that are substantially higher than marginal cost. They argue that firms are able to do so because some consumers are myopic, in the sense that they do not consider some add-on services when choosing the base product. The authors also show that no firm has an incentive to "educate" the myopic consumers if the fraction of myopic consumers is sufficiently large. I test the applicability of Gabaix and Laibson's (2006) model to the pricing of overdraft services using a repeated cross-section of U.S. banks and credit unions. I use three demographic characteristics to serve as proxies for the fraction of myopic consumers in each market, and find that two of the three proxies are significant and have the predicted signs. These results, though not conclusive, suggest that structural estimation of Gabaix and Laibson's model may add substantially to our understanding of pricing decisions in aftermarkets. (Abstract shortened by ProQuest.). |
