Zero, Not Always a Special Price : 330,000 Consumer Drug Choices Give No Support for the Zero-price Effect

We use Swedish data on 330,000 consumer choices of medically equivalent drugs to study the zero-price effect first documented by Shampanier et al. (2007) in experimental settings. The Swedish benefit scheme implies that, during a given month, all consumers face the same price-differences between generic substitutes and that about a fifth of the consumers pay a zero price if they choose the cheapest substitute. Using both regression discontinuity designs and discrete choice models, we find no evidence for the zero-price effect in our study