On Point Predictions and Reference Dependence in Behavior-Based Pricing Experiments

It has been shown that the comparative static results of two-period behavior-based pricing models hold in laboratory experiments, while point predictions do not. This study aims to check whether these findings replicate and to evaluate why observed prices deviate from point predictions. We report observed prices in conformity with point predictions through: (1.) a uniform pricing benchmark, (2.) a replication of a behavior-based pricing experiment, and (3.) a follow-up experiment in which we consider the second period disjointed from the first period. By disjoining the two periods, we show that reference dependence toward first-period prices shifts the second-period pricing behavior of participants upwards. In a post hoc analysis, we show that considering consumers' myopic instead of strategic explains a downward shift in first-period prices and rationalizes prior experimental findings. Volatile price levels affect price-based welfare measures – such as seller profits and total customer costs. We show that transport costs are a robust welfare measure that alleviates the impact of distorted prices. Ultimately, our findings are relevant for the design and assessment of multi-period pricing experiments.