A Resilient Framework for Smallholder Postharvest Loss Reduction Modeling

Thesis Abstract

Smallholder farmers face persistent postharvest losses (PHL) that undermine food security and farmer livelihoods, driven by weak cold chains, inadequate storage, spoilage, and market access constraints. This study develops and tests a resilient framework for reducing PHL among smallholders by integrating socio-economic, technical, and institutional dimensions into a cohesive modeling approach that supports adaptive decision-making under uncertainty. The aim is to articulate a systems-oriented framework that can predict PHL under variable yields, climate shocks, and market fluctuations, and to identify leverage points for targeted interventions. Specific objectives are (i) to synthesize theory from resilience, supply chain risk management, and agri-food systems to construct a multi-layered PHL reduction model; (ii) to operationalize the model as a dynamic simulation with stochastic inputs representing climate variability, pest pressures, and price volatility; (iii) to calibrate and validate the framework using empirical data from 450 smallholder households across three agro-ecological zones; (iv) to quantify the relative effectiveness of interventions at community (storage capacity, drying facilities), value-chain (collective marketing, price guarantees), and policy levels (input subsidies, extension services); and (v) to formulate a decision-support tool for stakeholders. The methodology adopts a mixed-methods research design grounded in a comparative case study and system dynamics simulation. The population comprises smallholder maize and legume farmers in a mid-latitude tropical region with documented PHL ranging from 15% to 40%. A stratified random sample of 450 households is drawn, with 150 per zone, and data are collected through structured surveys (100-item questionnaire), in-depth interviews (34 key informants including extension agents and traders), and focus group discussions (three per zone). Instrument validity and reliability are established through pilot testing (n=50) and Cronbach’s alpha exceeding 0.78 for multi-item scales. Secondary data on seed varieties, storage technologies, and market prices are sourced from agricultural offices and trader associations. Analytical techniques include partial least squares structural equation modeling (PLS-SEM) to test the relationships among resilience factors, storage capacity, and PHL outcomes; system dynamics (SD) modeling to simulate dynamic interactions under alternative scenarios, and multi-criteria decision analysis (MCDA) to rank intervention packages. The SD model captures feedback loops among climate shocks, pest outbreaks, labor availability, and market access, while the PLS-SEM clarifies mediation effects of organizational and capability factors on PHL. The study anticipates findings that resilient capabilities—efficient drying and drying-to-storage sequencing, community-based storage cooperatives, diversified off-farm income, and transparent price information—significantly reduce PHL and improve income stability, with climate variability intensifying losses in the absence of adaptive capacity. The contribution to knowledge lies in (i) a theoretically grounded, integrative framework that translates resilience theory and supply chain risk management into a practical PHL reduction model; (ii) empirical estimation of the relative impact of technical, organizational, and policy interventions on PHL; (iii) a scalable, user-friendly decision-support tool that can be adopted by extension services, farmer cooperatives, and policymakers to simulate interventions under uncertainty; and (iv) methodological advancement by combining PLS-SEM with SD modeling to capture causal pathways and dynamic complexity in agri-food systems. The study concludes that a combination of scalable technical improvements, such as modular drying facilities and solar-powered cold storage, with cooperative governance and price risk management, yields the largest and most robust reductions in PHL under diverse climate scenarios. Recommendations include policy support for shared storage infrastructures, capacity-building programs to enhance organizational resilience, and investment in ICT-enabled market information systems to stabilize farmer incomes. Limitations acknowledge potential regional transferability constraints and data quality variability due to recall bias; future research should extend the framework to other crops and incorporate participatory scenario planning with farmer communities.