Towards a healthier future for the achievement of SDGs: unveiling the effects of agricultural financing, energy poverty, human capital, and corruption on malnutrition

The paper addresses the persistently high burden of child and maternal malnutrition in Sub‑Saharan Africa, particularly in the western sub‑region, where malnutrition rates exceed global averages and contribute substantially to morbidity, mortality, and economic losses. Against a backdrop of rapid population growth, stagnating agricultural productivity, energy poverty, and governance challenges, the study asks how agricultural financing (credit, research spending, and foreign aid), energy poverty, human capital, and corruption affect child and maternal malnutrition. It motivates the inquiry by linking nutrition to human capital accumulation, productivity, and SDGs (notably SDGs 2, 3, 4, 7, 8, and 17). The study emphasizes the western SSA sub‑region due to its highest malnutrition burden and contributes by jointly examining children and mothers, integrating multiple financing channels with energy, human capital, and corruption, and applying MMQR to capture heterogeneous effects across the malnutrition distribution.

The literature review synthesizes evidence on four domains: (1) Agricultural financing and malnutrition: Studies generally find that access to agricultural credit, market access, and certain forms of aid improve dietary diversity and food security (Amao et al., 2023; Kihiu & Amuakwa‑Mensah, 2020; Annim & Frempong, 2018; Dhahri & Omri, 2020), though results can be mixed depending on credit source and outcomes (Iftikhar & Mahmood, 2017; Islam et al., 2016). Evidence on agricultural research spending is mixed or inconclusive with respect to nutrition outcomes (Adjaye‑Gbewonyo et al., 2019; McDermott et al., 2015; Harriss, 1987), and external development assistance can help but may be unevenly distributed (Fleuret & Fleuret, 1980; Berg & Muscat, 1973; Reutlinger & Selowsky, 1976). (2) Human capital and malnutrition: Nutrition knowledge and parental education often reduce child undernutrition, though findings vary by context (Fadare et al., 2019a,b; Amare et al., 2021; Hasan et al., 2015; Smith & Haddad, 2015; El Mouzan et al., 2010; Benson et al., 2018). (3) Energy poverty and malnutrition: Energy poverty is associated with worse health and higher undernutrition risk; electrification improves nutritional and infant health outcomes (Dake & Christian, 2023; Kose, 2019; Fujii et al., 2018; Lewis, 2018; Thomson et al., 2017). (4) Corruption and malnutrition/food security: Higher corruption worsens food and nutrition security, while governance quality and control of corruption improve outcomes (Onder, 2021; Qingshi et al., 2020; Anik et al., 2013; Cassimon et al., 2021, 2022, 2023; Ogunniyi et al., 2020).

Data cover nine Western Sub‑Saharan African countries from 1990–2019. The dependent variable is child and maternal malnutrition measured as disability‑adjusted life years (DALYs) from IHME/GBD. Explanatory variables include: agricultural credit and external agricultural aid (FAOSTAT), agricultural research spending (ASTI), energy poverty proxied by access to electricity (WDI), human capital index (Penn World Table), and corruption (Bayesian corruption index by Standaert, 2015). All series are log‑transformed. The baseline model: lnMALNUT_it = β0 + β1 lnAGRCREDIT_it + β2 lnRAGR_it + β3 lnFAID_it + β4 lnEPOV_it + β5 lnHCI_it + β6 lnCORRUPTION_it + μ_it. Estimation strategy: (i) Test cross‑sectional dependence using Breusch–Pagan LM, Pesaran scaled LM, bias‑corrected LM, and Pesaran CD; (ii) Apply second‑generation unit‑root testing via Pesaran’s CIPS to address cross‑sectional dependence; (iii) Test cointegration with Westerlund (2007); (iv) Estimate long‑run heterogeneous effects across the conditional distribution using Method of Moments Quantile Regression (MMQR; Machado & Silva, 2019), which handles unobserved heterogeneity, nonlinearity, and endogeneity; (v) Conduct robustness via bootstrap quantile regression (Koenker, 2005); and (vi) Explore directional relationships with panel Dumitrescu–Hurlin Granger causality tests. Variable sources and data links are provided (IHME, FAOSTAT, ASTI, WDI, PWT, BCI).

- Cross-sectional dependence is present across variables (multiple CD test statistics significant), motivating second‑generation methods. CIPS tests indicate stationarity after first differencing for several variables; cointegration tests support long‑run relationships. - MMQR results (selected ranges across quantiles q=0.1–0.9): • Agricultural credit (lnAGRCREDIT): negative and generally significant effects on malnutrition; coefficients roughly −0.09 at q10 to around −0.12 at q90. • Agricultural research spending (lnRAGR): positive and highly significant across quantiles; coefficients near 0.94–1.06, indicating higher research spending is associated with higher malnutrition. • External agricultural aid (lnFAID): mostly negative and significant, with effects becoming more negative at higher quantiles; approximate range from −0.07 to −0.23 at mid-to-upper quantiles (some variability at specific quantiles). • Human capital (lnHCI): negative and significant across quantiles; effects strengthen at higher quantiles, ranging roughly from −0.10 to −0.36. • Energy poverty (lnEPOV; higher access to electricity implies lower energy poverty): generally negative effects on malnutrition across most quantiles (e.g., about −0.23 to −0.01), with significance varying by quantile; results broadly imply improved electricity access reduces malnutrition. • Corruption (lnCORRUPTION): positive and significant; coefficients around 0.09–0.10 across quantiles. - Bootstrap quantile regressions largely corroborate MMQR: credit remains negative and significant across quantiles; research spending remains positive and significant (~0.82–1.01); aid mostly turns negative and significant at higher quantiles (e.g., q70 −0.105**, q90 −0.230***); human capital remains negative and increasingly large in magnitude (q10 −0.069*** to q90 −0.406***); energy poverty mostly negative up to mid quantiles (e.g., q10 −0.221*** to q60 −0.134**), turning insignificant/positive at upper quantiles. - Dumitrescu–Hurlin panel causality highlights: one‑way causality from human capital to malnutrition, malnutrition to corruption, human capital to agricultural credit, energy poverty to agricultural credit, human capital to foreign aid, human capital to energy poverty, and foreign aid to agricultural research spending. Bidirectional causality between human capital and corruption; research spending and corruption; research spending and human capital; human capital and foreign aid; and energy poverty and foreign aid.

Findings indicate that expanding agricultural credit and external aid to the agricultural sector reduce child and maternal malnutrition, consistent with pathways through improved production, incomes, and diet diversification. The strong positive association between agricultural research spending and malnutrition suggests misalignment between research priorities and nutrition outcomes—potentially reflecting a focus on short‑term productivity or profit rather than nutrient‑dense foods and equitable access—highlighting the need to reorient R&D towards nutrition‑sensitive agriculture. Energy access emerges as an enabling factor for better nutrition (storage, food preparation, health services), while higher human capital (education and knowledge) consistently lowers malnutrition, underscoring synergies between SDG 3 and SDG 4. The positive effect of corruption on malnutrition reflects governance failures that undermine service delivery, resource allocation, and program effectiveness; thus, anti‑corruption and governance reforms are integral to nutrition strategies. Heterogeneous effects across the malnutrition distribution (quantiles) suggest that policies may have stronger impacts where malnutrition burdens are higher, informing targeted interventions.

The study contributes by jointly examining multiple agricultural financing channels with energy poverty, human capital, and corruption for both child and maternal malnutrition in Western SSA using advanced panel quantile methods. Key conclusions: (1) Agricultural credit and external agricultural aid reduce malnutrition; (2) Access to electricity reduces malnutrition; (3) Human capital improvements lower malnutrition; (4) Corruption increases malnutrition; (5) Agricultural research spending, as currently allocated, is associated with higher malnutrition, suggesting a need to reorient R&D towards nutrition outcomes. Policy implications include expanding nutrition‑focused agricultural credit and external investments (aligning with SDG 17.2), re‑aligning agricultural R&D to nutrition‑sensitive and smallholder‑appropriate innovations, investing in renewable and reliable energy (SDG 7), strengthening governance and anti‑corruption measures, and empowering women through education (SDG 4). Future research should integrate additional determinants (e.g., climate change) and extend analysis to other SSA sub‑regions for comparative insights.

Data constraints limited inclusion of potentially important variables (e.g., climate change and other environmental shocks), and the analysis focused on nine Western SSA countries over 1990–2019. Measurement limitations (e.g., proxies for energy poverty and corruption) and potential data quality issues may affect estimates. Although MMQR and robustness checks mitigate endogeneity and heterogeneity concerns, causal interpretation remains bounded by observational data.