Climate change, food security, and diarrhoea prevalence nexus in Tanzania

The study addresses how climate change affects both food security and public health—particularly diarrhoea—within the context of Tanzania, a country among those projected to drive substantial global population increases by 2050. Rising temperatures, altered precipitation, and extreme weather events disrupt agricultural systems, compromise immunity, and influence infectious disease dynamics, elevating risks of diarrhoea. Global trends show increased food insecurity and disease burdens linked to climate variability, with Sub-Saharan Africa experiencing significant declines in staple crop yields and elevated food insecurity. In Tanzania, increasing climate-related extremes, dense population pressures, and economic challenges heighten vulnerability, especially in rural areas, through impacts on food and water quality, pathogen ecology, and disease distribution. Despite many studies on adaptation and resilience, there is a gap in empirical evidence on direct impacts of climate change on health and food security in Tanzania. This study aims to fill that gap by empirically examining causal links between climate change, food security, and diarrhoea prevalence using nationally representative agricultural survey waves and econometric strategies to handle endogeneity. The paper also highlights the complex pathways by which climate influences health outcomes and food systems, including bidirectional feedbacks and the role of health system capacity.

Prior research documents the negative effects of rising temperatures and changing precipitation on agriculture, including reduced yields, increased pest pressures, and more frequent extremes such as droughts and floods, all of which threaten food security. Climate variability is also linked to increased diarrhoeal diseases, especially among vulnerable groups such as children and older adults. While some studies underscore uncertainties and call for more empirical and mechanistic systems-based research, a general pattern emerges: higher temperatures and lower water availability are associated with higher diarrhoea rates. In Sub-Saharan Africa, projections indicate reduced crop production and higher malnutrition, disproportionately affecting poorer populations. In Tanzania, climatic variability has been associated with increased diarrhoeal incidence, and dry years elevate food insecurity risks. Households dependent on agriculture with fewer working-age members face compounded vulnerability due to concurrent food insecurity and diarrhoeal risks. Despite escalating concerns, empirical studies quantifying causal relationships remain constrained, motivating the use of instrumental variable approaches in this study.

Data: The study uses three rounds of Tanzania's National Agriculture Sample Census Survey (NASC): 2002/2003 (3rd round), 2007/2008 (4th round), and 2019/2020 (5th round). The surveys employ a stratified multistage design consistent across rounds. The analysis focuses on smallholder farmers (over 75% of the agricultural workforce), enhancing sample homogeneity; large-scale farmers are excluded. A total of 32,560 smallholder farmers are used across estimations; regression samples report n = 28,080. Key variables: Diarrhoea presence (binary): D = 0 if no cases in the household in the past 12 months; D > 0 if at least one member affected (treated as presence/absence). Food security (binary): Based on FAO access dimension; households classified as food secure if F ≥ threshold T, and food insecure if F < T. Climate change status (binary): Based on annual precipitation relative to a threshold T; y = 1 if annual precipitation P exceeds T, else 0. Instrument and construction: Average temperature anomalies are used as an instrument for climate change. For month j in year i, temperature anomaly Aij = Tij − Tj, where Tj is the long-term average for month j (computed from historical data). These anomalies proxy deviations from typical climate patterns and satisfy relevance and exogeneity conditions as argued in the literature and validated by tests. Econometric approach: To estimate climate change effects on diarrhoea prevalence, an Instrumental Variable (IV) Probit model is employed to address endogeneity between climate change and health outcomes. For food security, the Control Function Approach (CFA), including a two-stage residual inclusion (2SRI), is applied to control for endogeneity and unobserved heterogeneity, using predicted residuals from auxiliary regressions. The instrumented endogenous regressor is climate change; instrument is average temperature anomalies. Instrument validity and diagnostics (Table 2): Relevance test F = 37.58 (p = 0.0000) indicates relevance; Weak instrument test statistic = 99.48 exceeds max critical value 32.56; First-stage R^2 = 0.6632 (p = 0.0005) shows strong correlation with climate change; F-statistic for instrument validity = 35.78 (p = 0.0006); Exogeneity test p = 0.4281 indicates exogeneity; Heteroskedasticity and autocorrelation tests statistic = 47.23 (p = 0.5630) indicate no issues; Endogeneity test = 3571.43 (p = 0.0040) confirms endogeneity; Overidentification (Sargan) = 25.17 (p = 0.3613) supports model specification. Controls: Socio-demographic and economic covariates include sex, age (and age squared), household size, education, marital status, household income, financial services access, membership in agriculture societies, extension services, years of schooling, employment, fertilizer use, sanitation status, residence (rural/urban), and government support.

Descriptive patterns: Rural households constitute 56.41% of the sample; 52.35% are female-headed. Only 6% report access to extension services, and 28.05% belong to agricultural societies. Government support is reported by 35.48%; financial services access by 23.40%. Trend analyses across socio-economic groups show rising climate change prevalence over time alongside deteriorating food security and increasing diarrhoea, especially among rural and poor households. Bivariate relationships: Higher precipitation is associated with higher diarrhoea prevalence; higher average temperatures are associated with increased food insecurity. Increasing precipitation is also linked to increased food insecurity due to flooding, waterlogging, crop disease, and soil fertility loss. IV Probit (diarrhoea prevalence, n = 28,080): Climate change increases diarrhoea prevalence by 0.214602 (p < 0.01). Older age is associated with higher diarrhoea prevalence (0.09565, p < 0.01), and rural residence increases likelihood by 0.26001 (p < 0.01). Factors reducing diarrhoea prevalence include higher household income (−0.24064, p < 0.01), more years of schooling (−0.17721, p < 0.05), financial services access (−0.18098, p < 0.05), government support (−0.29210, p < 0.05), good sanitation (−0.37217, p < 0.05), and membership in agriculture societies (−0.18752, p < 0.05). Married status is positively associated with diarrhoea (0.07591, p < 0.05). Instrumented variable: climate change; instrument: average temperature anomalies. Control Function Approach (food security, n = 28,080): Climate change significantly reduces the probability of being food secure, with the largest effect under CFA: −0.331142 (p < 0.01). Age reduces food security (−0.10082, p < 0.01). Positive determinants of food security include fertilizer use (0.27218, p < 0.01), household income (0.210043, p < 0.01), years of schooling (0.13105, p < 0.01), being married (0.1103841, p < 0.01), employment (0.26004, p < 0.01), access to extension services (0.31691, p < 0.01), membership in agriculture societies (0.24621, p < 0.01), government support (0.417841, p < 0.01), and financial services (0.18003, p < 0.01). Rural residence lowers food security (−0.191022, p < 0.01). Climate change residuals are significant in 2SRI and CFA, supporting endogeneity correction. Instrument tests: Multiple diagnostics confirm instrument strength, relevance, and exogeneity (e.g., weak instrument statistic 99.48 > critical value 32.56; Sargan p = 0.3613).

The study provides causal evidence that climate change worsens both public health and food security among Tanzanian smallholder households. The positive association between climate change and diarrhoea prevalence aligns with mechanisms whereby extreme precipitation and flooding degrade water and sanitation infrastructure, create pathogen-conducive environments, and increase water-borne disease transmission. Concurrently, climate change reduces food security through stressors on crop production (temperature stress, precipitation variability, droughts/floods, pests and diseases) and market disruptions, thereby lowering access and stability. The pronounced vulnerabilities of rural and poorer households reflect limited access to clean water, sanitation, healthcare, extension services, financial resources, and institutional support. The significance of income, education, sanitation, financial inclusion, government support, and farmer organizations in mitigating adverse outcomes highlights actionable policy levers. Overall, the findings directly address the research question by quantifying the climate–health–food nexus and demonstrating that climate change simultaneously elevates diarrhoeal risk and undermines food security, necessitating integrated agricultural and public health strategies.

The study concludes that climate change exerts substantial adverse impacts on both food security and diarrhoea prevalence in Tanzania, with stronger effects observed on food security. It underscores the urgency of coordinated global and national actions—particularly by high-emitting nations—to finance and implement adaptation and mitigation measures such as tree-planting, water reservoirs for irrigation, watershed and catchment protection, and widespread community awareness. For developing countries with constrained budgets, integrating climate considerations into health programs and expanding locally grounded adaptation strategies are essential. Empirically, the paper contributes by applying instrumental-variable methods to quantify the climate–health–food nexus, identifying older individuals and rural populations as especially vulnerable, and highlighting the roles of income, education, sanitation, financial services, extension, government support, and farmer organizations in building resilience. Future research could extend these analyses, deepen mechanistic understanding, and evaluate the effectiveness of integrated interventions that jointly target climate adaptation, food system resilience, and public health.

The analysis focuses exclusively on smallholder farmers to enhance homogeneity, excluding large-scale farmers; consequently, findings may primarily generalize to smallholder agricultural households in Tanzania rather than the broader population.