Conservation Agriculture Evaluation Project in Northern Ghana: a formative evaluation using a framed field experiment

3ie evidence programme:  Agricultural Insurance Evidence Programme
Author(s): Kate Ambler, Alan de Brauw, Nicole Gargano, Mike Murphy, Usamatu Salifu
Institutional affiliation(s): International Food Policy Research Institute, International Food Policy Research Institute, Innovations for Poverty Action, International Food Policy Research Institute, Innovations for Poverty Action
Grant-holding institution: International Fund for Agricultural Development (IFAD)
Main implementing agency: Ghana Agricultural Sector Investment Programme (GASIP)
Sex disaggregation: Yes
Gender analysis: No
Equity focus: Yes, Heterogeneity Analysis
Study type: Formative evaluation

Context

While conservation agriculture (CA) provides public benefits, farmers adopting CA do not usually reap immediate private benefits. Since it takes time for organic matter to build up in the soil, productivity gains are not immediately realised. Farmers adopting CA, therefore, typically face an immediate cost for a potentially uncertain future benefit; hence, adoption rates in Africa have been low.

Given that there is a public good component to CA and potential private benefits to adopters in the medium-term, one method of reducing farmers’ perceived risk of CA techniques is to provide incentives for adopting CA. Within this context,  Innovations for Poverty Action (IPA) and  International Food Policy Research Institute (IFPRI) collaborated with The Ghana Agricultural Sector Investment Programme (GASIP), a programme within Ghana’s Ministry of Food and Agriculture (MoFA), to design and implement a framed field experiment (FFE) that incentivised farmers to take up CA practices.

Intervention design

The goal is to promote CA by educating farmers on how to best implement CA practices and encouraging the uptake of CA techniques. This study used an FFE to test whether incentives or information about peers affects the adoption of minimal soil disturbance (MSD).

In the FFE, farmers made decisions about adopting MSD over ten rounds representing agricultural seasons. Each ‘season’ was framed in three stages: a first stage in which farmers decided whether to adopt conventional practices or MSD for that season (with MSD requiring an associated cost for weeding); a second stage in which they received a random draw that determined whether the rainfall for that season was ‘normal’ or ‘poor’ and a third stage in which they received a payout for their ‘harvest’ for that season. The rainfall probabilities were designed to reflect conditions in the study areas, and the payouts were structured to reflect that adopting CA has short-term costs but long-term payoffs if adopted for multiple seasons consecutively.

For each round, farmers were also given information about the adoption decisions of a hypothetical peer in that ‘season’, and about that peer’s yields for that season.

Theory of change

The theory of change behind this incentive-based intervention to catalyse CA adoption followed two main mechanisms: (i) incentives could act as a nudge to access available information, which will improve knowledge of CA techniques and reduce uncertainty surrounding adoption and (ii) incentives may provide the financial compensation that will allow the financial benefits of transitioning to CA agriculture to outweigh its cost.

The peer-observation component posits that farmers will be more likely to adopt CA if others in their peer group are doing so, especially if farmers can observe positive results enjoyed by peers who adopt CA.

Incentives, peer observation, or their combination may lead to greater rates of CA adoption.

Evaluation design and methodology

A small randomised experiment was conducted in late 2019 and early 2020 to assess the effects of incentives and peer information on CA adoption. The experiment is ‘framed’ in such a way that participants choose between different options explicitly labelled as agricultural practices, rather than abstract choices.

The first stage of the research was a qualitative study to understand the feasibility of providing incentives for adopting CA techniques and to inform the design of a baseline survey instrument for an eventual impact evaluation of the GASIP programme once it is implemented. The research was conducted in districts  GASIP was targeting and in villages where farmers had some familiarity with CA from previous programmes.

The study sample for the baseline survey and the FFE included FBO members in the 66 communities who were direct beneficiaries of the GASIP programme in 2018 or targeted to become GASIP beneficiaries in 2019. The sample design anticipated 20 beneficiary farmers in each community. Ultimately, 1,324 individuals from the sample frame of 1,328 completed the experiment.

Primary evaluation questions

The focus of the study was to evaluate the best means of promoting adoption of the CA technique. The study also looked into factors helping or hindering adoption in the long term. Formative evaluations are crucial for proper design of large impact evaluations for future.

There are two main research hypotheses. First, if farmers receive incentives for adopting CA techniques promoted through extension agents, their adoption of CA techniques will increase. Second, if farmers observe farmers in their community or surrounding area practicing CA techniques, they will become more likely to also adopt CA.

Primary findings

1. Farmers who were exposed to the incentives in the FFE were 7.6 per cent more likely to choose MSD and around 8 per cent more likely to achieve production gains built into the experiment.
2. There was a continued higher take up of CA even after incentives were withdrawn. This suggests that targeting incentives over a fixed period could help farmers in adopting CA practices, as they cross hurdles of costs and uncertainty associated with adoption. MSD adoption was particularly notable and was sustained.
3. A positive effect was found among treated individuals who were told that a peer had successfully adopted CA over a long time period, suggesting that, for example, a demonstration plot might be effective if maintained for a long time period.
4. The focus groups also suggested that incentives could improve CA adoption. When specifically asked about incentives in focus groups, farmers generally agreed that incentives would increase the likelihood of them giving CA a chance. Farmers cited fertiliser and herbicide as potentially strong incentives. Durable goods like hand seeders or wheelbarrows were also commonly cited as incentives that would be effective, although the authors note that it is unclear how a one-time transfer would incentivise maintaining adoption over a sustained period.