Supply chain disruptions hamper the gains from globalization and require costly investments in resilience. I study how input sourcing through specialized intermediaries helps firms to mitigate disruptions in risky markets. Combining customs and tax records from Chile, I document that the share of intermediated imports rises with supply chain risk, as intermediaries maintain more diversified and robust supply networks. These facts motivate a model of global sourcing with costly supplier matching, insecure supply relationships, and access to intermediaries. Heterogeneous producers balance input prices and disruption probabilities across locations to minimize expected production costs. More productive firms match with multiple suppliers per location, while less productive firms contract with intermediaries, paying higher markups for a more resilient network than they could build directly. Despite double marginalization, intermediation relaxes the efficiency-risk trade-off due to greater supply network operability. Model quantification reveals sizable profit losses from disruptions, which intermediaries halve for mid-size producers that lack the scale to diversify. Intermediation is thus instrumental for supply chain resilience, suggesting a role for policies that make these services more accessible.

This paper examines intermediation in production networks to unpack the firm attributes and matching costs that govern firm-to-firm networks and the gains from trade. Exploiting rich customs data for Chile, we show that exporters of all sizes use intermediaries, mix trade modes across buyers, and set lower prices on intermediated flows. We rationalize these facts in a model of network formation with suppliers of heterogeneous productivity and matchability, buyers of heterogeneous productivity, and intermediaries that reduce matching costs for a brokerage fee. Empirical evidence on trade activity across firms and countries corroborates the model, and informs how geographic distance, logistics and customs efficiency, formal institutions, and cultural-linguistic similarity shape network costs. Model estimation reveals that sellers’ attributes are negatively correlated, such that intermediaries enable highly productive sellers with low matchability to reach smaller buyers. This amplifies the welfare gains from intermediation due to wider and deeper network connectivity.

We study the role of firm heterogeneity and imperfect competition for global production networks and the gains from trade. We develop a quantifiable trade model with two-sided firm heterogeneity, matching frictions, and oligopolistic competition upstream. More productive buyers endogenously match with more suppliers, thereby inducing tougher competition among them to enjoy lower input costs and superior performance. Transaction-level customs data confirms that downstream French and Chilean firms import higher values and quantities at lower prices as upstream Chinese markets become more competitive over time, with stronger responses by larger firms. Moreover, suppliers charge more diversified buyers lower mark-ups. Counterfactual analysis indicates that entry upstream benefits high-productivity buyers, while lower matching or trade costs benefit all buyers, with the biggest boost to mid-productivity buyers. All three shocks generate sizeable welfare gains, especially under package reforms. Global production networks thus mediate bigger effects and cross-border spillovers from industrial and trade policies.

This paper studies the role of customer portfolio management in exporters' growth. Using panel data on Chilean exporters and their foreign buyers (2002–2019), we document new facts on export dynamics. Growing exporters replace old customers with more profitable matches, accounting for nearly 40% of firms' export growth, while increasing both their average price and price dispersion across customers. We build a dynamic model of exporting with endogenous network formation, managerial costs of customer relationships, and bilateral bargaining that can rationalize these patterns. Each period, firms decide how much to invest in customer search, which customers to retain, and what prices to charge. Additional customers enhance exporters' sales and negotiating positions, but they also increase managerial costs. Using the model, we characterize firms' dynamic export strategies and quantify the role of managerial costs in export growth. We then evaluate policies that reduce search and managerial costs, considering both individual and combined reforms, and their interaction with policies that increase competition through market entry.