Impact of third-party enforcement of contracts in agricultural markets—A field experiment in Vietnam

Chapter 11 (Saenger et al.) provides a perfect example of a randomized controlled impact evaluation. The authors conducted a randomized controlled trial and field experiment with dairy farmers and a milk-processing company in Vietnam. Their approach, designed ex ante, is a theoretically ideal approach to constructing a valid counterfactual and to ensuring that there is no selection bias, given that the farmers are randomly assigned to treatment (beneficiaries) and control groups. This randomization ensures that all farmers have the same chance of participating in the program and that the distribution of the two groups’ characteristics (both observed and unobserved) are statistically indistinguishable. The authors tested whether the quality-control procedures used by the processing company were leading farmers to underinvest. The risk on the farmers’ part came from the possibility that the company would manipulate the process and say that the milk delivered was of low quality and therefore deserved a lower price. By introducing vouchers for third-party quality measurement, the program improved the company’s credibility with the farmers. With this increased trust, the farmers then had more incentive to invest in techniques to improve milk quality and increase revenue. This chapter is unique in that it focuses on the mechanisms and incentives for different value-chain actors to contract with one another. The authors’ proposed contract-farming designs make both parties better off, rather than trying to cut out the intermediary or encourage smallholders to take over other capacities in the value chain.