Missing price and coupon availability data in scanner panels: Correcting for the self-selection bias in choice model parameters

Discrete choice models have been widely estimated on scanner panel data to study consumer choice. One challenge in scanner panel research is that only the prices of the items bought are recorded. The ad hoc models used to fill in the missing prices of non-purchased brands may create a self-selection bias in estimating consumer price sensitivities. This type of bias is also present in existing studies of coupon effects. To obtain consistent estimates of price elasticities in the presence of missing price and coupon values, we estimate a brand choice model jointly with models for the price and coupon processes.