Firm-to-Firm Relationships and Price Rigidity

Economists have long suspected that firm-to-firm relationships might increase price rigidity due to the use of explicit or implicit fixed-price contracts. Using confidential, transaction-level import data from the U.S. Census, I study the pass-through of exchange rate changes and show that prices are in fact substantially more responsive to cost shocks in older versus newly formed relationships. Based on additional stylized facts about a relationship’s life cycle and interviews I conducted with purchasing managers, I develop a model in which a buyer-seller pair subject to persistent, stochastic shocks to production costs shares profit risk under limited commitment. Relationships that experience good shocks have lower costs, trade more, and survive longer, which generates the life cycle. Furthermore, since partners in older relationships on average enjoy a greater relationship surplus, alternative matches are less attractive to them, which enables the firms to share profit risk more completely by setting prices that are more responsive to shocks. As qualitatively predicted by the model, pass-through in the data is correlated with trade growth within a relationship and with exporter risk aversion. Once structurally estimated, the model replicates the empirical correlation between relationship age and pass-through. My results suggest that changes to the average length of relationships in the economy – e.g., in a recession, when the share of young relationships declines – can influence price flexibility and thus the effectiveness of monetary policy. ∗I am extremely grateful to my advisors, Eduardo Engel, Pinelopi Goldberg, Giuseppe Moscarini, Peter Schott, and Aleh Tsyvinski for their advice and support throughout this project. I am also thankful to Costas Arkolakis, Michael Peters, and Tommaso Porzio for their comments. This material is based upon work supported by the National Science Foundation under Grant Number 1427027. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation or of the U.S. Census Bureau. All results have been reviewed to ensure that no confidential information is disclosed.