|
Abstract:
|
In marketing, much recent research has been concerned with the resultant biased estimates from demand models when ancillary information is ignored (e.g., endogenously chosen prices on the part of the retailer). In this research, we focus on a problem that, when ignored, also may lead to wrong inferences: the occurrence of unobserved out of stocks. In fact, every time there is a stockout, the set of available alternatives to a consumer changes. Consequently, the assumption that each consumer may consider all the alternatives in her decision process is no longer valid. The estimation approach is based on simulating the unobserved evolution of the inventory on a retailer's shelf via MCMC and data augmentation methods. The proposed methodology is tested using simulated data and will be illustrated using a dataset from a retailer in Spain.
|
