Abstract:
This thesis presents applied microeconomic analyses pertaining to production risk management on small farms, with a focus on distributional aspects of farmer decision-making and farm productivity. In the first paper, I study the heterogeneous impact of agricultural credit on crop yield quantiles for households operating millet-based cropping systems. Equilibrium credit rationing under asymmetric information is the economic channel underlying heterogeneous credit-yield linkage. Therefore, I exploit within sample, quantile- based variation in credit-yield linkage to detect credit rationing incidence among sample households. Specifically, I propose a new metric for household-specific ex-ante odds of credit rationing, beyond identifying whether the average-household was rationed. I interpret these odds within a state contingent technology framework, which reveals how credit rationing influences farm-level adaptation under uncertainty. In the second paper, I employ a beta regression model in conjunction with a non linear instrumental variable approach, i.e., Beta- IV framework to estimate input-conditioned crop yield densities from plot-level observational data for wheat, rice, cotton and pigeonpea. The key inputs considered are nitrogen, phosphorus and irrigation levels. My work reveals novel evidence on how crop inputs influence production risk across major crops. I use these density estimates to evaluate the role of mixed-cropping to manage production risk among risk-averse, smallholder farmers in semi-arid India (Di Falco and Chavas 2006). My third paper is a data contribution to algorithmically reconcile identifier inconsistencies in land quality information in the Village Dynamics Studies in South Asia (VDSA)-plot-level panel dataset comprising 1,689 farm households across multiple survey waves during 1975-2014. My algorithm uses VDSA's plot nomenclature and phonetic/vernacular similarities to reduce data loss extent from 52% to <2%. In the fourth paper, I examine farmers’ subjective soil quality perceptions and assess how these perceptions diverge from data-based evidence. I construct and model a plot-level measure of soil quality misperception as the gap between survey-based soil quality metrics and corresponding data-based measures from high-resolution gridded maps. Further, I examine previously-unexplored two-way causal relations between soil misperceptions and input application decisions by using appropriate IVs for each directional relationship. Overall, my essays advance academic and policy research on smallholder agriculture in the global South.
