Fields of Interest: Industrial Organization, Microeconomic Theory
A new draft of “Opacity and Search Markets” will be uploaded soon. See my Thesis (link) for my results related to opacity.
Abstract:Consider models where consumer choices are represented with either classic discrete choice or ordered search. If consumers correctly anticipate the choices of firms, the set of classic discrete-choice models and the set of ordered-search models generate the same set of demand and payoff functions. Set equality extends to the supply side of these models if consumers observe the choices of firms prior to search. Thus, in certain contexts, classic discrete choice and ordered search are interchangeable models where results in each literature apply to the other literature. This relationship is established with the Geometric Search Model (GSM) which can be used to match any classic discrete-choice model.
Opacity and Search Markets
Abstract: In search markets where the choices of firms (such as prices) may or may not be observed by consumers prior to search, the set of classic discrete-choice models and the set of ordered-search models generate the same set of equilibrium demand structures. This set equality extends to the supply side of these models if consumers know all prices prior to search (full transparency). However, in search markets with strategic complementarities, if any firm faces increased opacity, all equilibrium prices increase. Thus, if an ordered-search market is incorrectly modeled with the equivalent classic discrete-choice model, empirical estimates of firm profit margins and theoretical predictions of market prices are at the lower bound corresponding to full transparency.
Consumer Journeys, Generalized Weitzman Search, and Discrete Choice Demands with Complementarities (with Simon Anderson and Maxim Engers)
Abstract: To represent consumer journeys, we model ordered search using a tree to allow for shared search costs and search depth. We prove that Weitzman's optimal search algorithm extends to this more general search model: we define a score for each node such that the consumer is best off always selecting or searching the highest-scored node that hasn't been chosen yet. Search models in which products have no shared search costs generate the same set of demand functions as classic discrete-choice models. However, the possibility of overlapping paths in the general search model allows products to be gross complements for a range of prices, unlike classic discrete-choice models. Since products can be both close in value and close in search, the general model has different implications for mergers analysis, optimal pricing and advertising.