Date of Award

May 2019

Document Type


Degree Name

Doctor of Philosophy (PhD)



Committee Member

Matthew S. Lewis

Committee Member

Babur De los Santos

Committee Member

Patrick L. Warren

Committee Member

Andrew Hanssen

Committee Member

Bernardo F. Quiroga


This dissertation consists of two chapters. The first paper estimates demand for gasoline in the presence of two types of imperfect price information: ex-ante uncertainty about each station's price and uncertainty about the distribution of all stations' prices. Volatile wholesale cost causes retail gasoline prices to fluctuate regularly, making it difficult for consumers to remain aware of the overall price level in the market or the stations offering the lowest price. In this article, I develop a model in which consumers formulate their prior belief of the current price distribution using the prices observed during past driving trips, and then Bayesian update their beliefs with each new price observed, before deciding whether to purchase gasoline or continue searching for a cheaper price. I estimate this model by utilizing a unique data set of station-level daily gasoline sales and prices, combined with data on the empirical distribution of various traffic flows. My empirical results suggest that consumers are able to learn about the overall price increases or decreases resulting from the wholesale cost movements relatively quickly. In addition, I find that price distribution uncertainty is the primary component of imperfect price information, and if it were eliminated, consumers could achieve 70 percent of the total savings that could be realized by having perfect price information. Furthermore, by incorporating travel patterns, the estimation suggests that cross-price elasticity between two stations depends largely on the amount of common traffic they share.

My second paper studies the effect of complexity in multi-dimensional bidding and competition in A+B (price + quality) auctions using a laboratory experiment. I examine whether the behaviors of human bidders are consistent with the predictions of two alternative models of auctions: the Bayes-Nash Equilibrium model and the Quantal Response Equilibrium (QRE) model. I extend the QRE framework to multi-dimensional A+B auctions. The results indicate that the QRE model, as a generalization of the rational models of behavior by allowing decision making errors, predicts bidder behaviors well across different treatments as the number of bidders and the dimensionality of the bid vary.



To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.