Improving Incomes of Coffee Farmers (IICoF) In Jimma, Ethiopia

The project has four main intervention components: Digital Extension; Input credit; Access to inputs; and Women self-help groups (WSHGs). The main features of each project component are provided below.

Digital Extension. Building on their previous work with maize, teff, and wheat farmers in Ethiopia, Digital Green will develop and use video-based extension (VBE) and interactive voice recording (IVR) approaches to promote best practices in coffee and maize farming in the project areas. Carefully tailored to match the information provided by the public extension service, the content in the VBE and IVR will focus on good agricultural practices, improvement of soil fertility, prevention of pests and diseases, and harvest and post-harvest management.

Input credit. Partnering with local financial service providers (e.g., banks, and micro-finance institutions), Digital Green will pilot input credit to households farming coffee and maize. The voucher-based input credit can be used to purchase key farm inputs such as seeds, seedlings, fertilizers, agrochemicals, and farm tools.

Access to inputs. In this project component, Digital Green will ensure the availability of inputs and tools by collaborating with local cooperatives and agro-dealers.

Women self-help groups (WSHGs). Working with zone and woreda level officials, Digital Green will aim to strengthen existing and form new WSHGs in the project kebeles. Between 3 and 4 videos will be developed to educate the WSHGs on group governance, conflict management, financial literacy, business development, and entrepreneurial skills. The groups will also receive financial support to set up revolving funds to strengthen group lending capacity to develop new and strengthen existing livelihood activities.

2024-01-15

2027-12-31

The primary outcomes include farmers’ knowledge of good agricultural practices; adoption of coffee stumping and other good agronomic practices; coffee and maize yields; and household consumption-expenditures.

The secondary outcome indicators include alternative welfare indicators (household per capita income, household per capita income from activities managed by women, food security) and access to inputs, technologies, and credit.

The experiment is based on a cluster randomized controlled trial (RCT), where cluster is defined at the development (farmer) group level. This is a longitudinal study following coffee farming households for a four-year period covering pre-treatment (baseline) period, period during the treatment (midline) and post-treatment period (endline). The study clusters are randomly allocated to different study arms, permitting us to identify the causal impact of the project. Specifically, we followed two-stage randomization. First, we randomized the 75 eligible/project kebeles into 55 ‘Treatment’ kebeles and 20 ‘Pure Control’ kebeles. Second, in the ‘Treatment’ kebeles we draw a random sample of 625 development groups. To understand potential synergies between the digital extension (DE), input credit (IC), and WSHG project components, these 625 development groups were randomized into the following 5 study arms:

Study arm T1: DE on coffee - exposed to Digital Green’s Video Extension services that focus on coffee farming.
Study arm T2: DE on coffee & maize - exposed to video extension services that focus on both coffee and maize farming.
Study arm T3: DE on coffee & maize + IC - receive the same intervention bundle as households in study arm T2 but in addition, are offered input credit
Study arm T4: DE on coffee & maize + IC + WSHG - exposed to the digital extension on coffee and maize, receive input credit, and are offered to join women’s self-help groups
Study arm TC: Control - a control group within "Treatment" kebeles.

Not available

The randomization was done in office by a computer.

Cluster randomization at the kebeles and development (farmer) group levels.

Yes

75 kebeles for the 1st stage randomization and 625 development (farmer) groups for the second stage randomization.

3,000 farm households.

500 households per treatment arm.

Stumping adoption (the key best practice for coffee): 3.7 percentage points. Coffee yield: 30.2 percent. Maize yield: 13.4 percent. Consumption expenditure: 15.5 percent.