Rural women’s adaptive agency: sustaining food security and agricultural production efficiency under climate stress in Murehwa District, Zimbabwe

Abstract

The smallholder farming system in sub-Saharan Africa is in crisis due to climate variability‚ soil degradation‚ and the increasing cost of agricultural inputs․ In the Murehwa District of Zimbabwe‚ these shocks manifest in transitory household food insecurity․ Finally‚ socio-technical changes identify enabling adaptive capacity‚ with SAPs being framed as de-risking instruments for climate‚ but with uneven uptake owing to gender asymmetries in SAP design and access․ Smallholder units are typically described as a homogeneous aggregation in the climate-impact literature‚ and therefore the structural resource restriction and human labour burden of independent female heads and married female plot managers under customary male control are blurred․ The overall objective was to analyse rural women's adaptive agency‚ technology adoption and production efficiencies under climate stress conditions in Murehwa District‚ Zimbabwe‚ to provide evidence to inform gender-responsive climate adaptation policies․ Specific objectives included: (1) estimating the causal effect of women's SAP adoption decisions on household food security; (2) identifying socio-economic and institutional constraints to women's SAP adoption; and (3) examining differences by gender in the effect of climate stressors on the reallocation of labour‚ production costs and production technical efficiency․ We adopted a quantitative‚ cross-sectional survey research design and collected data based on the agricultural household modelling framework․ A multi-stage stratified random and quota sampling of 12 villages was used to gather micro-level data from a verified and audited representative sample of 101 smallholder households․ First‚ we used an Inverse Probability Weighted Regression Adjustment (IPWRA) estimator to estimate the average treatment effect of adoption on HFIAS‚ correcting for self-selection bias․ Second‚ we applied a maximum likelihood binary probit model and a non-linear Blinder-Oaxaca (Fairlie's) decomposition to decompose the gendered technology adoption gap into asset endowments and structural institutional return effects․ Third‚ we estimated technical‚ allocative‚ and economic efficiency frontiers under environmental stress using a non-parametric Data Envelopment Analysis (DEA) and a farm-budgeting gross margin framework․ Qualitative data collected through KIIs and FGDs also helped triangulate the econometric estimates from both project and comparison areas․ The double-strong IPWRA models found a statistically important positive causal effect of women's SAP uptake on household food security mediated through a large and negative effect on baseline HFIAS score․ Cumulative years of farming experience and the size of the FTE labour pool were statistically important positive predictors of adoption in the Probit models (γ = 0․284‚ p < 0․05)․ Distance to commercial input markets was a major non-negotiable barrier: a 1 km increase in distance was associated with a 1․8 percentage point lower baseline probability of technology adoption․ The Blinder-Oaxaca non-linear decomposition showed this barrier accounted for the 16․7 percentage point technology adoption gap․ Decomposing this gap empirically‚ this study isolated its structural-return and endowment effect (p < 0․01)‚ revealing that: 70․7% of the gap in women's outcomes was due to structural discrimination; women's returns to formal agricultural extension services and credit conduits were considerably lower than men's at the same pre-existing endowment levels․ The asymmetric climate vulnerability in agricultural production was demonstrated by the independent female managers' considerably higher average labour shock score (μ₂ = 3․030) during climatic shocks than those of men․ This indicates that most households could not hire external labour when faced with adverse climate shocks when liquidity restrictions did not allow for hiring external labour․ 5․ The adoption of SAPs (which are a source of liquidity) led to a reduction of total variable production inputs by USD 98․070 per hectare (p < 0․10) and considerably improved technical efficiency (μ₀= 0․612 and μ₁= 0․748‚ p < 0․01) and allocative efficiency (μ₀= 0․485 and μ₁= 0․693‚ p < 0․01) through successful substitution of commercial synthetic chemicals with biological organic structures․ Scaling up sustainable intensification without uniform technology transfer will require efforts to induce long-term resilience through the MLAWRD that: (a) invests in inclusive‚ decentralized village-level input supply hubs to reduce spatial transaction costs; (b) guarantees gender-symmetrical representation and provides support for female-led peer-to-peer extension training networks to counteract gendered information asymmetries; and (c) invests in subsidized‚ scale-appropriate mechanization (e.g. two-wheel direct-seeders‚ ripping implements) to soften the impact of rising labour intensity on independent female smallholders․