This post originally appeared on the IFPRI.org blog.
B Andrew Reid Bell, Patrick Ward, Lawrence Mapemba, Tim Benton, Klaus Droppelmann, Jennifer Zavaleta Cheek, Frazer Mataya, and Oliver Pierson
Conservation agriculture (CA) consists of three main farming practices: Keeping soil covered, reducing tillage, and enhancing soil nitrogen through crop rotation or intercropping. Broad adoption of these practices not only contributes to healthy and productive soils, it also reduces erosion that can compromise rainfall runoff and, downstream, waterways that generate hydropower. These benefits reach individual farming households that depend on healthy soils, as well as regions and even nations that rely on productive food and power systems.
In Malawi, CA has become a priority for the government and civil society, and could particularly benefit hydropower producers in the Shire River Basin. The country’s Agricultural Sector Wide Approach (ASWAp) incorporates the three CA practices, and several financial incentives for encouraging small-scale farmers to adopt them are under design and evaluation, including those launched by the United Nations Development Programme and the Millennium Challenge Corporation.
We recently completed a three-year study in partnership with the Malawi Department of Land Resources Conservation (DLRC) and the National Smallholder Farmers’ Association of Malawi (NASFAM) to evaluate how financial incentives affect the adoption of CA practices among smallholder farmers. We found evidence to suggest that incentives could trigger a tipping point in the number of farmers adopting the practices, leading to much broader use in the long run.
Decision trees toward adoption
We evaluated the impact of several financial inventive structures. While this was helpful for the specific context of Malawi, the broader set of questions can help us to understand just how farmers decided to adopt the three CA practices—and whether the incentives themselves are the strongest motivating factor.
We interviewed or surveyed small-scale farmers from a random sample of 63 villages in the Shire River Basin, including Balaka and Machinga districts and a portion of Zomba district. One set of randomly selected households that participated in the program received a standard subsidy payment equivalent to about $30 for up to one acre for adopting all three CA practices, while another set of randomly selected households received a smaller standard subsidy (about $15), with a bonus payment equivalent to about $5 for participation by neighboring farmers (up to 4 contiguous neighbors).
To explore how these farmers decided to adopt each individual practice, both with and without the presence of the subsidy, we combined ethnographic interviews with machine learning to help structure and classify the order of factors that went into their decision making. Formulating decision trees based off of ethnographic interviews has been used extensively to understand the dynamics of similar agricultural decisions such as farm planning and weed management, but combining results from machine learning algorithms with insights from ethnographic interviews is a novel contribution.
Following the lead of neighbors
At the start of the project, adoption rates were 60 percent for intercropping, 43 percent for mulching, but only 7.6 percent for zero tillage. By the end, adoption rates for all three practices increased: In a final survey of 1,923 farmers, 87 percent reported intercropping, 50 percent reported crop residue mulching and 39 percent reported zero tillage. Adopting all three practices was reported by 28 percent.
Drawing on our machine learning tools, we found that the most important factor shaping the decision to adopt any of the three practices was whether neighbors had adopted them. This factor transcended both the availability and structure of a financial incentive. In short, farmers who observed the practices benefiting neighbors were more likely to adopt those practices themselves.
We did find some variation in the factors influencing farmers’ decisions to adopt individual practices. The availability of a subsidy influenced decisions to adopt intercropping more than decisions to adopt crop residues and zero tillage, for instance. These variations support recent research showing that adopting CA is not a single, all-in commitment for farmers, but appears to emerge from two decisions—one related to intercropping and a second related to mulching crop residues over untilled soil.
The decision to use crop residues was most influenced by whether the farmer had ever been exposed to the practice. Farmers who mulched and were not offered subsidies were most influenced by whether their neighbors did it or if they had access to extension services. The main deciding factor for those offered subsidies, meanwhile, was the monetary incentive.
The decision to adopt zero tillage was also driven primarily by whether or not people had exposure to it. Farmers not offered subsidies adopted zero tillage when they had access to extension services or if their neighbors did it. To a lesser extent, these farmers were influenced by whether there was sufficient rain in the previous year. Those offered a subsidy were most influenced by whether their neighbors adopted the practice and, to a much lesser extent, by the monetary incentive and their access to residues.
Supporting early adopters
As the analysis shows that neighbors and incentives are the key factors shaping adoption, in places where adoption rates are low and there are few neighbors acting as role models, incentives are likely necessary to encourage CA practices. But our results also suggest that, thanks to the “neighbor effect,” there may be a tipping point beyond which further adoption is self-reinforcing, reducing the need for financial incentives.
Though these results are not necessarily universal, in Malawi’s Shire Basin they have clear implications for an array of development programs aimed at improving livelihoods, landscapes, and water systems. In the sustainable land management context, CA practices have the potential to reduce sedimentation in waterways, and thus could provide a clear rationale for payments from hydropower producers to encourage small-scale farmers to adopt CA.
This kind of program may have high upfront costs, but those are likely to fall over time as benefits accrue to farmers adopting CA practices. Eventually, such payments might even be phased out. The initial investment could be a relatively small price to pay for reaching the tipping point that addresses the broader challenges of land degradation.
Andrew Reid Bell is an Assistant Professor of Environmental Studies at New York University; Patrick Ward is a Research Fellow in IFPRI's Environment and Production Technology Division; Lawrence Mapemba is a Senior Lecturer at Lilongwe University of Agriculture & Natural Resources; Tim Bentonis a Professor of Population Ecology at the University of Leeds; Klaus Droppelmann is a Process Facilitator with PICOTEAM; Jennifer Zavaleta Cheek is a PhD student at the University of Michigan; Frazer Mataya is with the National Smallholder Farmers Association of Malawi; and Oliver Pierson is the Malawi Resident Country Director with the Millennium Challenge Corporation.