My thesis consists of three chapters that study the introduction of differing priors in classic models. Aumann’s renowned result that agents who share a common prior cannot agree to disagree triggered a debate on the common prior assumption. For instance, Gul points out that “the prior view is an inherently dynamic story” and “it amounts to asserting that at some moment in time everyone must have identical beliefs”. In the long run, when agents learn or communicate this seems reasonable. In noncooperative one-shot games as analyzed in my thesis, however, it seems conceivable that agents have differing priors. I interpret this assumption as expert knowledge in Hotelling’s line of horizontal differentiation and as prejudice in a principal agent framework.

Chapter 1, Public Procurement with an Expert, studies two exemption clauses to the general obligation that public contracts must be awarded by means of tendering procedures. Typically, public procurement policies allow for direct awards to experts if the buyer is uncertain about his specific need or the contract’s value is low. In a model of horizontal differentiation with demand uncertainty, I derive the expert’s offer for a directly awarded contract and the offers of duopolists with differing priors in a tendering procedure. The expert’s price in the former case appropriates the buyer’s willingness to pay less a discount that—according to the expert’s prior—fully compensates the buyer for unsuitable project specifications. Consistently, the expert minimizes unsuitability by customizing the project’s specification to the buyer’s need. In the tendering procedure, however, prices equal implementation cost plus a premium that the competitors charge due to horizontal differentiation. Comparisons of the two cases in terms of welfare as well as the buyer’s total cost reveal that, concerning both criteria, direct awards are beneficial if expert knowledge is valid and the contract’s net value is low. These insights support common policies on public procurement.

Chapter 2, Profitability of a Self-Proclaimed Experts Knowledge, more thoroughly analyzes the specification-then-price game that provides the theoretical foundation of chapter 1. In the case of converging priors, the model is equivalent to Hotelling’s line with uniformly distributed demand and quadratic cost. Accordingly, firms offer maximally differentiated projects to mitigate price competition. In case of differing priors, however, I derive other equilibrium types. Their existence depends on the expert’s knowledge: if expertise is moderate and includes awareness of unsuitable project specifications that affect the expert’s special field, she customizes her offer and launches price competition. If expertise is deep, in contrast, the expert’s pricing strategy stifles competition, and her specification is contingent on the perceived unsuitability of her competitor’s project. Comparative statics determine conditions for the profitability and market power of expert knowledge.

Chapter 3, Prejudice, Quotas, and Wages, suggests prejudice as another interpretation of differing priors in a simple formulation of a static principal agent model with discrete effort choice. The prejudiced principal wants to hire two agents to carry out a common project. Without regulation he would hire two male team members because he wrongly believes that women bear higher cost of performing demanding work than men. However, even with a quota that forces the principal to offer jobs to candidates of both genders, he fails to hire a mixed team consisting of two hard working peers. This is not due to discrimination, but self-selection: the underestimated woman rejects her offer as a peer because her overestimated coworker shirks. In contrast, she accepts the offer for a trivial job even though her coworker shirks. This result constitutes a new rationale for the gender gap in workplace hierarchies. I propose wage equality as a remedy for the free-riding problem, and consequently, the underrepresentation of women in demanding positions.