Mid-Term Evaluation of the Food Security through Improved Resilience of Small Scale Farmers in Ethiopia and Kenya (FOSEK)

Overview

Mid-term evaluation of FOSEK, a public-private partnership of 10 partners (including Nestlé, Solidaridad, coffee marketers, the Coffee Research Foundation/Institute and three Ethiopian cooperative unions). The project aims to improve food security, diversify incomes and strengthen cooperatives among smallholder coffee farmers in Kenya and Ethiopia. Carried out by Development Frontiers and reported in March 2020, the evaluation assessed progress against outcome and output indicators, identified gaps and drew lessons for the rest of the project and for possible scale-up. It covered cooperative member farmers and cooperatives in both countries, plus partners and government officers.

Key findings

• In Kenya, average income per farmer fell from KES 130,607 (USD 1,230) at baseline to KES 75,156 (USD 708) per year, with income sources more diversified: coffee 53.5%, food crops 14.4%, dairy 10.6%, cash crops 9.6%.
• In Ethiopia, average income per farmer rose from Birr 20,394 (USD 621) at baseline to Birr 40,457 (USD 1,232); coffee provided 68.9% of income, food crops 13% and cash crops 6.2%.
• At baseline fewer than 1% of farmers reported earning income from food; at mid-term 14% in Kenya and 13% in Ethiopia named food crops as a major income source. Almost 90% of farmers in both countries practise intercropping.
• Cherry/parchment deliveries in sampled Kenyan cooperatives fell from 992.2 tonnes in 2013/14 to 643.78 tonnes in 2019/20. In Ethiopia they rose from 1,656 tonnes (2013/14) to 3,875 tonnes (2015/16), then fell to 1,006 tonnes in 2019/20.
• 80.2% of farmers in Kenya and 87.8% in Ethiopia rated their cooperatives above average, but poor governance (missed elections, unaudited accounts, late payments) was reported in some cooperatives.
• Implementation stalled in 2017 and part of 2018 in both countries. In Kenya, partners disputed their cash contributions. In Ethiopia, a government cap cut the unions' pledged 50.2% contribution to 25%. Political instability and the devolution of agriculture to Kenyan counties also delayed work.
• The M&E system could not report on most outcome indicators: indicators had no definitions, numerators or denominators, no targets were set from the baseline, there was no database or data QA system, and Ethiopia had no M&E officer.

In figures

Recommendations

• Review the allocation of partner responsibilities based on each partner's core strengths and mandate.
• Expand partnerships with county ministries/departments responsible for agriculture (food production) and cooperative development so that farmers keep support after the project.
• Build on gains in intercropping and food crop production, balancing the use of trained farmers with strong support from assigned partners.
• Decide on the most effective implementation arrangement in Ethiopia, given the experience with the cooperative unions.
• Have cooperatives in Kenya and Ethiopia take more ownership of supporting Trainers of Trainers and Promoter Farmers so their services continue after the project.
• Strengthen M&E: link every activity in the work plan to an indicator, revise the Theory of Change, set up a digital database and a data collection matrix with named responsibilities.
• Commission a study of actual nutritional changes, especially among children and the elderly, and widen the risk management framework to cover political risks.

Lessons learned

• Finding common ground between project objectives and each partner's core strategies and interests helps advance outcomes.
• Partners need a shared view, from the start, of the material and other contributions expected and how these link to benefits for farmers.
• Robust stakeholder analysis and management plans help avoid blind spots in a public-private partnership approach.
• Thorough due diligence on partners' capacities and interests is needed before engaging them in long-term implementation.
• Food production, health and nutrition work is stronger when led by partners whose core mandate and expertise lie in those areas.

Methodology

Mixed methods: household survey of cooperative member farmers (408 interviews in Kenya from 14 cooperatives, 184 in Ethiopia from 17 cooperatives, against a target of 400 per country), 12 cooperative interviews in Kenya and 9 in Ethiopia, plus focus group discussions and in-depth interviews with partners and officials. Quantitative data were analysed descriptively and inferentially, and qualitative data thematically.

Limitations

• Only 184 of the targeted 400 households were surveyed in Ethiopia because of research assistant capacity problems and poor internet connectivity, which reduced the analytical power of the Ethiopia findings.
• Some planned qualitative interviews did not take place in Kenya (Embu FGDs and officer interview, Nyeri cooperative officer, one marketer), nor with two of the three Ethiopian unions, for security reasons.
• Cooperative officials' information on governance, finances and gender could not be verified, and no anthropometric survey was done, so nutrition analysis relied on county-level secondary data.
• Respondents were asked to recall conditions from 2014 to 2019, which introduces recall bias.
• The mid-term sample included more women than the baseline, so differences from the baseline may partly reflect sampling.