Agent-based simulation modeling for analysis and support of rural producer organizations in agriculture

摘要

Development of smallholder agriculture is widely recognized as an important pathway to poverty reduction in rural areas, particularly in Sub-Saharan Africa. Many researchers propose collective action of smallholder farmers by means of rural producer organizations (RPO) as a promising opportunity to improve commercialization and market access of small farms, which in turn willresult in improvement of rural livelihoods. However, little is known about the determinants of RPO success. Currently, there is a broad demand for detailed analyses of RPO performance and for ex-ante assessments of the developmental interventions and policies of RPO support. This thesis focuses on the provision of high-resolution quantitative data for the design of such interventions using a case study of coffee producers in the lake-shore Uganda and their RPO. This work demonstrates the effective ways to increase farmers’ welfare through the network of RPO and analyzes the associated risks and opportunities.This work applies agent-based computer simulation to analyze the RPO. Designed virtual simulation experiments assess the broad portfolio of development interventions and economic scenarios that are challenging to investigate by means of real-world empirical research. The agent-based nature of the model allows for a holistic integration of several modeling concepts in the developed model application. This leads to the inclusion into the model of a number of important aspects of the bio-economic system of coffee production in Uganda. The first aspect is the heterogeneity among farming households, reflected by differences in natural conditions, resource endowments, production and market constraints, time and consumption preferences. The second aspect is the inseparability of decisions that are taken on the farm (i.e. investment, production, consumption and marketing) from one another. The third aspect is human-environment interaction cycles and the dynamics of the bio-economic system, including interactions across levels of hierarchy (here: individual farmers and RPO).The constructed model is parameterized, calibrated and validated using the empirical data from project and country-level surveys. The set-up of the model and the results of simulation experiments are further complemented by (i) a detailed review of relevant literature, (ii) community-based participatory research with members of RPO and (iii) interviews with key informants.Results of simulation experiments indicate that RPO activities can cause significant increases in members’ sales revenues and consequently can improve their household incomes. The posvitive impacts of RPO can be amplified through external assistance. Recommended RPO-level interventions include (i) on-the-spot payments for RPO members’ transactions and (ii) support for group certification. Both are expected to have high cost efficiency and a low risk of failure. In addition, results of this thesis suggest that improvement of agricultural productivity through the provision of quality planting material and the promotion of good agricultural practices is likely to be highly beneficial for the rural households. In order to stream the related development policies to smallholder farmers it is recommended to use RPO networks.Findings of the participatory research in Ugandan RPO indicate that the establishment of transparent rules of reception of RPO services and allocation of earned benefits, together with frequent and formal reporting of RPO administration might increase members’ cooperation within an RPO.This thesis also shows the vulnerability of coffee producing households and their RPO to the risks imposed by the volatility of agricultural prices. The role of development policy is, therefore, to provide price risk insurance for smallholder farmers and to facilitate the formation of accurate price expectations. However, viable and sustainable models of smallholder risk insurance are yet to be found.