We present an experimental study of dynamic pricing in which two retailers compete to sell perishable goods over a finite horizon. Consumers arrive at the market one at a time and remain there for a single period. Each consumer compares the two simultaneously posted prices, one by each retailer, and then decides probabilistically whether to purchase the good from Retailer 1, from Retailer 2, or not purchase it at all. The competing retailers are assigned to one of three between-subject experimental conditions. In one condition they start each session with equal inventories and in two other conditions with unequal inventories. Following a short learning period, equilibrium best-response solutions, which serve as benchmarks for our analysis, account well for the mean posted prices; however, small but systematic deviations from equilibrium play significantly diminish the retailers’ profits. These deviations decrease considerably in size with experience.

我们提出了一项动态定价的实验研究，其中两个零售商在有限的范围内竞争出售易腐商品。消费者一次到达一个市场，并停留一个时间。每个消费者比较两个同时发布的价格，每个零售商比较一个，然后根据概率决定是否从零售商1购买商品，还是从零售商2购买商品。竞争零售商被分配到三个受试者间实验条件之一。在一种情况下，他们以相同的库存开始每个会话，在另外两种情况下以不相等的库存开始。经过短暂的学习，均衡的最佳响应解决方案（作为我们分析的基准）很好地说明了平均发布价格；但是，均衡的微小但系统的偏差会大大削弱零售商的利润。这些偏差随着经验的减少而大大减小。