How does the number of products sold by a firm affect its decisions regarding price setting and information acquisition?
Using a firm-level survey from New Zealand, I show that firms that produce more goods have both better information
about aggregate inflation and more frequent but smaller price changes. To characterize the implications of these empirical
findings for the ability of monetary policy to stimulate the economy, I develop a new dynamic general equilibrium model with
rationally inattentive multi-product firms that pay a menu cost to reset their prices. I show that the interaction of the menu
cost and rational inattention frictions leads firms to adopt a wait-and-see policy and gives rise to a new selection effect:
firms have time-varying inaction bands widened by their subjective uncertainty about the economy such that price adjusters
choose to be better informed than non-adjusters. This selection effect endogenously generates a distribution of desired
price changes with a majority near zero and some very far from zero, which acts as a strong force to amplify monetary
non-neutrality. I calibrate the model to be consistent with the micro-evidence on both prices and inattention and find
two main quantitative results. First, the new selection effect, coupled with imperfect information of price setters,
leads to real effects of monetary policy shocks in the
one-good version of the model that are nearly as large as those in the Calvo model. Second, in the two-good version of the
model, as firms optimally choose to have better information about monetary shocks, the real effects of monetary policy shocks decline by 20%.