Actively Learning by Pricing: A Model of an Experimenting Seller

This paper presents a model of a rational seller who is actively learning the slope of his demand curve via his pricing strategy.  Consequently, this seller optimally experiments with his price.  Resulting price patterns show a lot of discreteness (as observed in the data), which has proved to be a major challenge to most price setting models.  This model's learning dynamics are able to reconcile individual price flexibility with aggregate price sluggishness, while the experimentation motive can explain the presence of many idiosyncratic price changes in the data as well as the observation that prices are more volatile than costs.