Forest income and livelihoods on Pemba: A quantitative ethnography

This paper offers a systematic approach to quantifying the socio-economic role of forests for 'forest-dependent' communities. Focusing on the island of Pemba (Zanzibar, Tanzania), we investigate how forest income contributes to livelihood portfolios, local inequality, and households' insurance against shocks. We also examine how forest income is affected by local institutions and household socio-demographics. We use a series of non-parametric measures in conjunction with multi-level Bayesian models supported by directed acyclic graphs to address these questions. On average, we find that 27% of household income comes from forests, with 83% of that value deriving from fuel products, and that 62% of the total value of forest products are harvested from the agroforestry scrub matrix. At the same time, forest income scales positively with income, forest-dependency scales negatively. Top income earners control similar to 4 times more forest income than low earners. However, when we consider forestry against other economic sectors, forest income reduces overall income inequality on the island. Despite forests being critical for the poor, we find it offers little insurance against shocks, especially for the vulnerable. In fact, in contrast to expectations, we find that the well-insured are the most likely to increase forest use in response to shocks. Regarding institutions, most forest products come from either government land or land owned by other private individuals, indicating weak tenure institutions on the island. Finally, young, poorly educated male-headed households, which are not integrated into markets, are the most likely to have high forest income. However, female-headed households are generally more dependent due to a lack of alternative income sources. Our results are encouraging as the use of tools from formal causal inference and detailed Bayesian modelling, in conjunction with a quantitative ethnography, build upon previous findings while improving our understanding of local socio-ecological systems. (C) 2022 Elsevier Ltd. All rights reserved.