Adoption of climate-resilient groundnut varieties increases agricultural production, consumption, and smallholder commercialization in West Africa

Smallholder commercialization is widely promoted as a pathway out of poverty in developing countries, but its success hinges on agricultural productivity growth, which remains relatively low in Sub-Saharan Africa. Building on Green Revolution lessons, climate-resilient, high-yielding, disease-resistant varieties are central to climate-smart agriculture (CSA), offering potential productivity, resilience, and emissions benefits. Groundnut is a key food and cash crop in West Africa with agronomic and nutritional benefits. This study examines whether adoption of climate-resilient groundnut varieties is associated with increases in production (yields and output value), household consumption, and smallholder commercialization (market participation, quantities sold, and sales value) in Ghana, Mali, and Nigeria. The research addresses how adoption and sustained adoption relate to these outcomes and explores heterogeneity across countries and producers.

The paper situates its inquiry within literature on agricultural commercialization and productivity growth as drivers of rural development, and on the role of improved and climate-resilient varieties in boosting yields and welfare under stress conditions. Prior work documents significant impacts of improved seeds on yields and market participation in various contexts, including drought-tolerant maize and improved groundnut varieties; and emphasizes the triple-win potential of CSA. The nonseparable agricultural household model provides a framework: production and consumption decisions are jointly determined under imperfect markets, implying that increases in production may translate into market participation only after household consumption needs are met. The study also references evidence from Malawi linking improved groundnut varieties to market participation, and broader findings on technology adoption constraints (information exposure, liquidity) and econometric approaches to address endogeneity in adoption-impact studies.

Data and sampling: The study uses a three-wave panel (2017–2019) of 2,868 farm households (balanced panels: Ghana 498, Mali 840, Nigeria 1,530), totaling 8,604 household-year observations, from regions targeted by a USAID-funded groundnut upscaling project in Ghana (Northern, Upper East, Upper West), Mali (Koulikoro, Sikasso; Mopti excluded in later waves due to insecurity), and Nigeria (Jigawa, Katsina, Kano, Kebbi, Sokoto). A multistage sampling selected districts/LGAs, then 4–6 villages per district/LGA, then ~30 households per village. Adopters and non-adopters were sampled under similar agro-ecological and institutional conditions. Attrition between 2018 and 2019 was modest (Ghana 8%, Mali 7%, Nigeria 4%); probit tests found no attrition bias. Measures: Outcomes include production (total quantity, production value, yield per hectare), household consumption of groundnut (kg), and commercialization (binary market participation; quantity sold; sales value based on prices received). Commercialization variables with zeros were transformed using the inverse hyperbolic sine (IHS) for regression. Adoption is measured both extensively (binary adopter vs non-adopter of climate-resilient varieties) and intensively (area under climate-resilient varieties, ha). Adoption identification focused on well-promoted climate-resilient varieties known locally (e.g., Ghana: Samnut 22, Yenyawoso, Nkatiesari; Mali: ICGV 86124, ICGV 86015, ICGV 86024, Fleur 11; Nigeria: Samnut 23–26), recognized by farmers and enumerators. Econometric strategy: The core estimation employs panel two-stage least squares (2SLS) with household fixed effects (FE) and correlated random effects (CRE, Mundlak–Chamberlain) to control time-invariant unobserved heterogeneity and allow correlation between covariates and individual effects. Time fixed effects are included. Endogeneity from reverse causality and measurement error in adoption is addressed via an instrumental variable (IV): willingness to adopt climate-resilient groundnut varieties (access and willingness), capturing preferences and information exposure. First-stage shows strong relevance (p < 0.001; F = 357.5), exceeding weak-instrument thresholds. Controls include demographics (age, education, gender of household head, household size, dependency ratio), social capital (cooperative membership, training), services (public/private extension), credit (cash and in-kind), market access (distance to nearest urban/village markets), agronomic practices (crop rotation, mixed cropping), inputs and costs (labor, input costs, area cultivated), market price, off-farm income, and soil type. Robustness: Multiple sensitivity checks include control-function/two-stage residual inclusion (2SRI), Hausman–Taylor IV estimator, and Lewbel’s heteroskedasticity-based internal instruments (with over-identification tests such as Hansen J and Hayashi C). Results are consistent across FE and CRE estimators and across identification strategies. Additional pooled FE-OLS models complement the main estimates. Figures and analyses were produced in STATA 17 and R 4.3.1. Heterogeneity and mechanisms: Cross-country models examine Ghana, Mali, and Nigeria separately. Quantile regressions assess heterogeneity in commercialization effects across the distribution. Simultaneous models relate production (yields) and consumption to commercialization to explore pathways. Sustained adoption is defined as adoption in all three survey years to compare with one- or two-year adoption.

- Adoption boosts production and consumption: Relative to non-adoption, adopting climate-resilient groundnut varieties increases yield by about 345 kg/ha and production value by approximately USD 476; household home consumption increases by about 213 kg. - Commercialization increases with adoption: Adoption is associated with 5–6 percentage points higher market participation, 54–59% higher quantity sold, and 53–57% higher sales value. Intensive margin (area under adoption) also increases market participation by 3–4 percentage points, quantity sold by 37–41%, and sales value by 35–39%. - Sustained adoption amplifies gains: Continuous adoption over all three years yields substantially larger impacts than one- or two-year adoption; for example, yields increase by about 1,242 kg/ha and quantity sold by roughly 442 kg, with significant positive effects on consumption as well. - Cross-country heterogeneity: Strongest yield effects are observed in Ghana and Nigeria; statistically significant commercialization effects are evident particularly in Nigeria, potentially reflecting larger, more diversified markets and favorable pricing conditions. - Mechanisms: Yields are positively associated with commercialization, supporting yield growth as a pathway to market engagement; household consumption is negatively associated with commercialization at the margin, consistent with nonseparable household production-consumption decisions where sales occur after meeting own-consumption needs. - Inclusivity: Quantile analyses indicate adoption benefits across the commercialization distribution, with the largest relative gains among smaller-scale farmers, suggesting inclusive impacts. - Descriptive adoption trends (2017–2019): Share of adopters (percent of households) evolved as follows—Ghana: 34.3%, 31.7%, 29.1%; Mali: 17.0%, 17.5%, 23.6%; Nigeria: 55.0%, 52.7%, 48.7%.

The findings address the central question of whether adopting climate-resilient groundnut varieties enhances smallholder production, consumption, and commercialization. The positive impacts on yields and output value directly support higher on-farm productivity, while increased household consumption underscores the food security and nutritional benefits of the crop. Importantly, higher yields translate into greater market participation and sales once household consumption needs are met, aligning with the nonseparable agricultural household framework under imperfect markets. The strongest commercialization responses in Nigeria likely reflect better market access and pricing opportunities, reinforcing the role of enabling market environments. The pronounced gains under sustained adoption highlight the importance of durability in technology use for realizing transformative effects. The inclusivity of benefits, with larger proportional gains among smaller-scale farmers, suggests that promoting climate-resilient varieties can contribute to equitable rural development and broader agricultural transformation. Collectively, results support scaling climate-resilient technologies within CSA to bolster resilience to climate shocks while improving welfare through production and commercialization gains.

Adoption of climate-resilient groundnut varieties in Ghana, Mali, and Nigeria is associated with significant increases in yields, production value, household consumption, and commercialization outcomes, with the largest impacts realized under sustained multi-year adoption. Yield growth appears to be a key mechanism driving market participation, conditional on satisfying home consumption. Effects are heterogeneous across countries and inclusive across the commercialization distribution, with smaller producers often experiencing the largest relative gains. Policy implications include strengthening seed systems and delivery, ensuring reliable access and information exposure, and reducing transaction costs to facilitate sustained adoption and market engagement. Future work could further examine long-term welfare and resilience trajectories under sustained CSA adoption, interactions with market development and pricing, and complementary interventions that enhance equitable access to improved seeds and services.

- Identification relies on an instrumental variable (willingness to adopt); while highly relevant, exclusion restrictions cannot be directly tested and the instrument may be imperfect. - Potential measurement error is acknowledged (e.g., adoption and outcomes), though mitigated by focusing on well-known, promoted varieties and supervised data collection. - Gender analysis is limited: surveys targeted household heads and did not allow individual-level gender disaggregation. - Panel attrition and sample reduction due to financial constraints and security issues (notably in Mali) occurred, though attrition rates were low and tests did not indicate bias. - External validity may be bounded to comparable contexts within West Africa and the specific period studied; the design is observational, so interpretations rest on econometric identification rather than randomized assignment.