Experimental Design
Our experiment is designed to compare the “quota rule” to the “unanimity rule” in the context of strata sales. Participants engage in a multilateral bargaining problem between a buyer and three sellers. Each seller starts the game holding a single unit, which has market value V_A/3. Their personal value for holding the unit is V/3. The buyer has value S for owning all three units and thus “completing the project”, but only receives the market value for each unit if not all three units are purchased. Unanimity requires that the buyer make individual deals with each seller, while the quota rule forces the last seller to sell at the market price if any two sellers make deals.
Subjects begin the experiment by reviewing instructions that describe the structure of the games they will participate in. Subjects complete a quiz and receive feedback about the correct answers for the quiz. They then move on to complete bargaining matches. Subjects continue playing new matches until either they complete 30 matches or the 95-minute session limit is reached, with the first match to finish after this cutoff serving as the final match of the experiment.
In the experiment, participants are randomly assigned the roles of “buyer” or “seller”, with roles being re-randomized each match. In odd rounds, buyers must make offers between 500 and 5000 to each of the sellers that remain active (i.e. those that have not already sold their unit). Sellers then observe all offers that were made and can accept or reject their offer. In even rounds, each active seller makes a request to the buyer that is between 500 and 5000. The buyer observes all requests and choose which, if any, to accept. All offers and requests are restricted to being multiples of 50. After each round, subjects are reminded of the current state of the game, including who has already made deals and for how much. Then, if the project has not already been completed, the match continues with probability delta. In all sessions, V_A = 1500, A = 7500, delta = 0.6, and the buyer’s endowment (from which they make offers) is 9000.
Participants are provided with several tools to help them make decisions. In all rounds, all participants see a table describing the current state of the game. This shows them current payoffs, current offers or requests, whether those offers or requests were accepted, and any previously made deals. Additionally, when sellers are making the decision of whether to accept or reject, they are reminded of the consequences of each of these options and how those consequences depend on other sellers’ choices.
We use a mixed between-subject and within-subject treatment design. The within-subject treatment is variation of the “quota rule” and “unanimity rule”. Each match, half of the groups are randomly selected to be assigned to each. The between-subject treatment varies the value of V: in the “High V” treatment, V = 6000 while in the “Low V” treatment, V = 3000.