Experimental Studies on Public Good Allocation with Agents

Inefficiencies in private giving are a common occurrence in public good games. In this dissertation, we ask four questions: do subjects recognize the inefficiency due to the lack of coordination on group giving, is there a simple way to overcome this coordination problem, is there a mechanism for giving that improves group welfare that is preferred by subjects, and do subjects choose a predatory allocation strategy when given opportunities to discriminate. To investigate these questions, we design a public goods experiment where the contributions of each individual may be determined by a member of the group named the "agent." In the game with common wealth and preferences among the group, the dominant strategy for the agent is to choose the Pareto-efficient allocation. Thus, giving through an agent in this environment eliminates the group coordination problem seen in private giving. In the game with diversified wealth among the group, the dominant giving behavior for the agent becomes a predatory allocation strategy, where members of their own wealth group can free-ride off the contributions of the other wealth group.Using public good games with both boundary and interior Nash equilibria, results from the agent treatment are contrasted with results from a no-agent treatment. In addition, efficiency and allocation decisions are compared between the equal-endowment experiment and the diversified wealth experiment. Subjects do recognize the inefficiency of individual giving to a group, and therefore, higher contributions to the public good are observed under the agent treatment which improve social welfare.In addition, a third chapter is presented modeling where an industry with two differentiated firms producing a homogenous product priced by contracts. Each firm faces difficulty in pricing their product since they are competing in a market with two types of customers, "old" and "new." Two types of switching costs are considered, one explicit in the contract and another which is implicit. We examine how prices and fees in contracts are affected by the parameter of the model including the spread and expected value of this implicit switching cost.