Contracts can take on a large variety of forms . They can help SHFs overcome market failures and government failures . For example, a contract in an out grower scheme with a sugar plantation in Kenya provides index insurance to individual farmers with the premium deducted from harvest-time payment for product delivery , providing a rare case of large take-up of index insurance without subsidies. Quality recognition can be an integral part of contracts, for instance inducing higher quality of cocoa production in Sierra Leone . Resource-providing contracts are however typically complex to put into place, implement, and enforce. Contract terms can be difficult for SHFs to respect. This applies for example to supermarkets that demand quality standards approximating international norms. Monopsony power of contracting agents can lead to hold-up practices with defaults or delays in payments and lower prices on deliveries. Side-selling by SHFs when the contract price is inferior to the local market price at harvest time can also undermine the enforcement and credibility of contracts. And principal-agent relationships facilitate appropriation by the lead agent, typically the commercial partner in a contract with farmers, rack system of the net social gains created by the contract thus limiting its potential for poverty reduction if there has been no improvement in farmers’ bargaining power over surplus.
Despite these difficulties, empirical studies tend to show positive benefits of these contracts for SHFs under the form of technology adoption, agricultural transformation, and income and employment gains . Most available impact studies are however recognized to suffer from selection biases and external validity issues .Other institutions are frequently necessary to complement or strengthen a simple contract. For example hold-up risk by the buyer can be resolved through a jointly accepted external certification agent. Saenger et al. show that providing dairy farmers contracted by a large company in Vietnam an independent milk quality assessment induced a 12% increase in input use, and a significant corresponding increase in output , ultimately translating into higher welfare for farmers. What is interesting in this case is that, while the company was playing fair with farmers, some of them believed that it was behaving opportunistically in deciding to reject some milk deliveries. Independent certification can thus benefit both the farmers and the company. More generally, quality certification can have a major role in value chain development where phytosanitary standards and other qualitative aspects of produce are essential. Important is for certification to happen low in the value chain, before aggregation of produce makes it impossible to reward individual quality contributions, which is particularly challenging when production originates in thousands of very small farmers . Where quality can be certified early in the value chain, incentives can be created for farmers to invest in quality enhancement, in particular through the adoption of technological change.
Beyond the direct contract between an agro-industry and farmers, arrangements may involve a separate input provider, or a financial actor that will provide credit to the farmers, possibly guaranteed by the agro-industry or the input provider, or a transporter, etc. These arrangements result in value chain models that are quite complex, entailing the need for coordination between multiple stakeholders, sometimes referred to as innovation platforms. As an example, the Bubaare Innovation Platform in Uganda developed since 2008 the sorghum value chain for the production of baby food and a traditional beverage in a context where most small farmers previously only produced sorghum for their own consumption. The platform includes Huntex Ltd, a large processor selling to supermarkets that invested in storage and packages and commercializes the products, credit partners that provide financing to farmers and the processor, the public agricultural research system that developed new high yielding varieties, extensions services that taught farmers improved crop management techniques, local parish governments that set standards for the commercialization of sorghum, and the University of Makerere that helped with technology development. This successful platform registered as a cooperative in 2013 and operates without any further donor intervention, solely relying on membership fees . A different story emerges from Peru, where a value chain was created to market native potatoes produced by SHF to high-income consumers in Lima.
Coordination is assumed by a strong intermediary which links farmers to a few supermarket groups. On the supply side, CAPAC coordinates services provided to producers by different NGOs, which include contract management, quality control, and delivery to the supermarkets. On the demand side, CAPAC participates in national advocacy, the promotion of events, and the creation of labels. The supermarkets themselves have developed their own promotion, including a cooking school and books. And researchers at the International Potato Center developed improved storage methods . There have been multiple initiatives by lead private enterprises, coalitions of private interests, and public-private partnerships to promote the development of similar vertically coordinated value chains. Over the last 15 years, the World Bank Group has spent heavily in value chain development in West African countries with investments in infrastructure , financing of private enterprises, support to producer organizations , development of supporting services , and public sector capacity . This has focused on value chains such as mangos, onions, meat, and poultry in Burkina Faso, and onions and rice in Senegal . Rigorous evaluation of these investments is still not available. These innovation platforms are to help actors in a value chain communicate and coordinate actions to address bottlenecks to value chain development. Swinnen emphasizes the role of identifying appropriate entry points that can consist in financing the lead firm in a value chain so it has resources to in turn finance farmers in interlinked contracts, and directly targeting constraints to value chain development such as farmer training, PO development, and presence of service providers. As revealed by the FARM Foundation’s review of contracting in value chains in Sub-Saharan Africa, lead private sector enterprises have been important in acting as coordinating agents for value chain development. Coordination can thus be achieved at the cost of competition, creating an interesting trade-off whereby monopsony power in value chains can help facilitate vertical coordination while enhancing value extraction to the benefit of the lead agent. Value chains for low-value domestic staple foods are particularly important for SHFs, but more difficult to develop as discipline is harder to achieve due to the large number of producers and availability of local buyers facilitating side-selling . Yet, success with value chain development for domestic producers is essential if they are to remain competitive with imports, and also potentially help the country make headways in substituting for rapidly rising food imports. Value chain development does not necessarily come top-down from commercial partners. It can also come bottom-up at the initiative of producer organizations. Collion thus contrasts top-down “aggregation schemes” in Morocco where an agroindustry contracts with producers to secure the provision of produce with quality specifications, mobile shelving to bottom-up “productive alliances” in Latin America where a producer organization develops a business plan that involves contracting with a commercial partner in resource-providing contracts.
Capacity of the producer organization to do this typically comes with technical assistance and subsidies provided by the public sector and with the support of international development organizations . Hence, the inclusive value chain development approach to modernization and transformation can come from upstream as well as from downstream agents in the value chain, even if the latter tends to dominate occurrences.Assessing the impact of a vertically coordinated value chain on farmers production or welfare is particularly challenging. All impact evaluations proceed from well-defined units that are treated and corresponding counterfactual that did not receive the treatment. In the case of value chains, the unit of intervention is most often either a farmer organization or a community, rather than an individual farmer. Even if not all farmers from a community participate in the value chain, this would be by choice, and hence non-participating farmers cannot be good counterfactuals to participating farmers. Any analysis will thus compare communities in which the value chain has intervened to communities not affected. This raises some statistical concerns. The number of communities any such intervention can incorporate is limited, simply because of financial and human resource constraints. The second challenge is to find counterfactual communities. This would consist on defining a priori a set of communities that would equally qualify for participating to the value chain and only randomly invite half of them to participate. Getting an institution or a private sector firm to select in this way clients with whom it will do business is difficult. But getting a large number of private and public sector partners to all agree around a scheme that will impose on them restrictions on who should and should not be incorporated in the process of value chain construction is even more challenging. In addition, the construction of a value chain is by essence gradual, and it is in part adjusted to the needs and with the contributions of the communities that are incorporated. It is rarely a blue print that is offered to communities. It is consequently not surprising that we find almost no rigorous impact evaluations of value chains. Those that exist proceed with an ex-post evaluation whereby researchers take stock of the communities that are included in the value chain and look for similar communities that are not included. Cavatassi et al. provide an interesting example with evaluation of a value chain linking smallholder potato farmers in the mountain regions of Ecuador to high-value markets. The authors proceed with a careful matching of communities, including propensity score matching using census data, verification by key local organizations and agronomists to determine whether they were comparable, and finally verifying that the treated and control communities did not differ regarding other interventions. Their impact evaluation is thus restricted to less than half of the treatment communities for which they have a good counterfactual. The authors then use different methods of weighting participants, nonparticipants, and control farmers. All of them show similar results showing that the value chain has brought a large increase in yield and in gross value per hectare, and that these benefits were obtained through both selling a higher percentage and quantity of potato harvest and selling at a 30 per cent higher price. Production costs also increased but increased revenues outweigh added costs, suggesting welfare benefits for the farmers. Impact analysis of contract farming presents somewhat different challenges, but is generally easier to do. Similar to what has been done to evaluate the impact of microfinance institutions , one can imagine a large processor who is extending its scale of operation willing to proceed with a randomized scheme, by which it will preselect more units than it wants to incorporate, or use a phased-in design whereby selected units are randomly assigned to be incorporated in two phases. Yet, the large literature attempting to measure the impact of contract farming on farmer welfare uses observational methods . Most studies use propensity score matching to find counter factuals to individual participating farmers, often within limited samples of non-participating farmers. The validity of the method in that context is questionable as identification of treatment effects is based on the assumption that selection into participation is fully captured by observable characteristics. As of yet, the closest to a randomized evaluation of getting smallholders access to high value markets through contract farming was done by Ashraf et al. . The authors worked with an NGO that serves as an intermediary linking SHF to commercial banks, retail providers of farm inputs, transportation services, and exporters. 36 farmer groups selected to participate in the experiment were randomized into treatment and control. The study shows that during the first year of operation, the contract increased production of export crops and lowered marketing costs, leading to a 32% income gain for new adopters. The experience however fell apart one year into the project, with defaulting on their commitments by buyers and farmers. It consequently has little external validity for what contract farming by a solid private sector partner could do . In view of the difficulty of organizing experiments on value chains, using natural experiments or statistical instrumentation can to be a fruitful approach. Macchiavello and Morjaria thus use an exogenous negative supply shock of Kenyan rose exports on the Dutch market due to post-electoral violence to study the importance of relational contracting when contract enforcement is weak.