Determinants of diversification by cocoa smallholders in Sulawesi

Purpose - The aim of this study was to evaluate determinants of four diversification practises by cocoa smallholders in West Sulawesi, Indonesia: (1) growing other crops, (2) keeping livestock, (3) off-farm work for wages (4) off-farm self-employment, and the impact of diversification on welfare of community members. Design/methodology/approach - Household interviews (n = 116) conducted in two subdistricts (Anreapi and Mapilli) of Polewali-Mandar District, West Sulawesi, provided quantitative data on household characteristics, crop and livestock production, income sources, expenditure and credit access. Two villages per subdistrict were included in the study, each producing cocoa as the main crop but differing in their proximity to a market town. Logistic regression was applied to identify determinants of diversification by households. Multiple linear regression (MLR) models evaluated the impact of diversification practices and other explanatory variables on two proxies of welfare (or household wealth): per capita value of durable assets (household assets other than land or livestock) and per capita expenditure for each household. Findings - Mean per capita cocoa production in the sample was low (51 kg dry beans/annum). The mean dependency ratio (proportion of household occupants age 64) was 35%, with an average of five occupants per household. Household heads were predominantly male (95%), averaging 46 yo and 7 years of formal education. Most households (72%) depended on loans, but only 24% accessed formal loans. Significant determinants of diversification practices were access to formal credit for self-employment and subdistrict for livestock, with Mapilli subdistrict households more likely to keep livestock. Household predictors in the MLR accounted for 28% variation of the dependent, per capita value of durable goods. Off-farm self-employment and raising livestock significantly improved welfare, but growing other crops or off-farm work for wages had little effect. Other household variables demonstrated to have significant positive effects on welfare were education of the household head, proximity to a market town and land area per household. Research limitations/implications - The study was restricted to a relatively small sample size (n = 116). Studies including panel data or larger numbers of households could enable the identification of further determinants of diversification. Practical implications - The study demonstrates that diversification has the potential to improve rural livelihoods, but that obstacles, especially formal credit access, may deter poorer households from diversifying their income sources. Social implications - Programs and policies that facilitate access to formal finance by smallholders could encourage diversification into small business and improve livelihoods in cocoa-dependent communities. Originality/value - In the light of the decline in cocoa farm productivity in West Sulawesi, the study demonstrates the potential benefits, as well as limitations, of income diversification by smallholders.