Economic Impact of Climate Variability on Smallholder Farming Systems

Abstract

Climate variability is affecting smallholder agriculture systems, particularly in Sub-Saharan Africa where over 90% of croplands are rainfall dependent and 55-62% of people rely on agriculture for livelihood. A problem-driven, mixed method approach was used to assess the economic impacts of climate variability on smallholder farmers in three agroecological zones. Data of 1,236 households from a survey was integrated with 30 years of climate data, and tested using fixed effects panel regression, logistic regression, propensity score matching, quantile regression, spatial econometrics, structural equation modelling and nine machine learning models. The study demonstrates that a unit increase in temperature anomaly reduces the net farm income by 0.567-0.823 units, and severe drought lowers the yield of maize by 47.3% and food security index by 0.36% (p<0.01). Climate shocks account for 72.5% of household losses of income (mean USD 487 per year) due to crop failure (54.2%), livestock mortality (18.3%), post-harvest losses (12.4%), losses in labour productivity (8.7%) and price volatility (6.4%). Spatial spillovers reveal local vulnerability is compounded by climate shocks in adjacent regions, and impulse response functions reveal that economic shocks last up to eight seasons. Adaptation capacity provides significant protection, declining at index values >0.7. Machine learning shows the XGBoost predictor is the highest performing, and structural equation modelling confirms the climate shock-food insecurity link. Climate variability creates large, persistent and contagious economic effects on smallholder systems, demanding quantile-sensitive, gender-responsive and location-specific climate adaptation policies which address both direct and indirect (market) effects.