Abstract: How does competition affect market outcomes when formal contracts are not enforceable and parties’ resort to relational contracts? Difficulties with measuring relational contracts and dealing with the endogeneity of competition have frustrated attempts to answer this question. We make progress by studying relational contracts between upstream farmers and downstream mills in Rwanda’s coffee industry. First, we identify salient dimensions of their relational contracts and measure them through an original survey of mills and farmers. Second, we take advantage of an engineering model for the optimal placement of mills to construct an instrument that isolates geographically determined variation in competition. Conditional on the suitability for mills’ placement within the catchment area, we find that mills surrounded by more suitable areas: (i) face more competition from other mills; (ii) use fewer relational contracts with farmers; and (iii) exhibit worse performance. An additional competing mill also (iv) reduces the aggregate quantity of coffee supplied to mills by farmers and (v) makes farmers worse off. Competition hampers relational contracts directly by increasing farmers’ temptation to default on the relational contract and indirectly by reducing mill’s profits.

Since January 2021, we have been conducting research activities in collaboration with RwaCof, a coffee  sourcing company in Rwanda. RwaCof sources from approximately 30,000 smallholder farmers in Rwanda. These farmers tend to have remarkably low yields and income, and are highly fragmented. The main goal of the collaboration is to conduct research that supports RwaCof in fostering and strengthening sustainable sourcing practices in the supply chain. RwaCof and our research team jointly design interventions in the supply chain, which are then rolled out as pilots and rigorously tested through a combination of randomized control trial (RCTs) and original data collection. These activities aim to produce a fully traceable supply chain and support smallholder farmers’ incomes and investments.

Abstract: Well-functioning markets allocate assets to owners that improve firms’ management and performance. We study the effects of ownership changes on coffee mills in Rwanda – an industry in which managing relationships with farmers and seasonal workers is important and that has seen many ownership changes in recent years. We combine administrative data, a survey panel of mills and an original survey of acquirers that allows us to construct acquirer-specific and target-specific control groups. A difference-in-differences design reveals that ownership changes do not improve performance unless the mill is acquired by a foreign firm. Our preferred interpretation – supported by detailed survey evidence that considers alternative hypotheses – is that foreign firms successfully implement management changes in key operational areas. Upon acquisition, both domestic and foreign owned mills attempt to implement similar changes, but domestic firms face resistance from workers and farmers. Domestic owners have relationships with their local communities, which can create opportunities to establish new mills and acquire existing ones. However, these same relationships create pressure to maintain status-quo relational arrangements, which makes it harder to implement managerial changes

An in-progress collaboration with Davide Del Prete (Napoli Parthenope).

Abstract: We compare transactions within integrated firms and in long-term relationships between firms in the Costa Rica coffee chain. We unveil a qualitative difference between these two organizational forms: the size of forward sales supported by long-term relationships is limited because better opportunities arising after the forward transaction is agreed might tempt the seller into default. Integration shifts ownership of coffee away from the seller and removes such temptations. We model this difference and exploit weather-induced supply shocks to test for constraints to forward trade. We show that forward trade in long-term relationships is constrained, while forward trade within integrated firms and spot trade are not. The evidence supports models in which firm’s boundaries alter parties’ ability to enforce past promises to trade.

Abstract: Widespread market imperfections in agricultural value chains raise the possibility that regulatory interventions may enhance efficiency and farmers’ welfare. We develop a structural model of agricultural value chains and estimate it using rich data from Costa Rica’s coffee sector to evaluate common regulations. Farmers supply differentiated mills that strategically decide which rural markets to source from and bilaterally bargain prices with downstream exporters. Through counterfactuals, our analysis highlights the nuanced, and potentially counter-productive, effects of commonly observed pro-competitive regulations on farmers’ welfare. Tightening revenuesharing rules to increase farm-gate prices, increases farmers’ welfare on average but makes many worse off. Similarly, banning vertical integration raises farm-gate prices but harms most farmers by lowering valuable services provided by integrated mills 

Abstract: Do the returns to quality upgrading pass through supply chains to primary producers? We explore this question in the context of Colombia’s coffee sector, in which market outcomes depend on interactions between farmers, exporters (which operate mills), and international buyers, and contracts are for the most part not legally enforceable. We formalize the hypothesis that quality upgrading is subject to a key hold-up problem: producing high-quality beans requires long-term investments by farmers, but there is no guarantee that an exporter will pay a quality premium when the beans arrive at its mills. An international buyer with sufficient demand for high-quality coffee can solve this problem by imposing a vertical restraint on the exporter, requiring the exporter to pay a quality premium to farmers. Combining internal records from two exporters, comprehensive administrative data, and the staggered rollout of a buyer-driven quality-upgrading program, we find empirical support for the key theoretical predictions, both the lack of pass-through of quality premia under normal circumstances and the possibility of a buyer-driven solution through a vertical restraint. Calibration of the model suggests that one-third to two-thirds of the (substantial) gains from the program accrue to farmers, with the vertical restraint playing a critical role. The results are consistent with the hypotheses that quality upgrading can provide a path to higher incomes for farmers, but also that it is unlikely to be viable under standard market conditions in the sector.

An in-progress collaboration with Davide Del Prete (Napoli Parthenope), Pablo Fajgelbaum (UCLA) and Amit Khandelwal (Yale).

An in-progress collaboration with Davide Del Prete (Napoli Parthenope) and Arthur Blouin (Toronto).