Description

We leverage a comprehensive dataset of electronic invoices from Chilean firms to document new facts on price dispersion across buyers of the same manufactured intermediate goods. Over half of firm-tofirm sales in manufacturing are accounted for by products that are purchased by more than one buyer in a given month, with prices ranging by 40 percentage points across buyers for the average product. Price dispersion is pervasive across all manufacturing sectors. Observable characteristics of products and of buyer-seller pairs (including distance, mode of payment, and size of the parties and of the transaction) explain only a small fraction of the variance of price gaps in the data. We use a workhorse model of production networks to quantify the productivity gains from eliminating markup dispersion across buyers of individual products, inferring initial differences in markups from observed price gaps. The increase in aggregate productivity relative to the sales share of treated multi-buyer firms ranges from 2 to 7 percent, depending on the calibration of elasticities of substitution. The gains from eliminating markup dispersion across buyers are as large as those of eliminating markup dispersion across products.