Household Vulnerability to Climate Variability and Food Insecurity: Evidence from Aceh and West Nusa Tenggara, Indonesia

Technical Summary: Household Vulnerability to Climate Variability and Food Insecurity in Aceh and West Nusa Tenggara

Problem Statement
Climate variability increasingly threatens household welfare and food security, particularly in developing nations dependent on climate-sensitive sectors like agriculture and fisheries. While global literature acknowledges the link between environmental shocks (floods, droughts, irregular rainfall) and food insecurity, significant gaps remain in understanding these dynamics at the micro-level within Indonesia. Specifically, existing research often focuses on aggregate agricultural productivity rather than household-level food insecurity, lacks integration of nationally representative microdata with localized climate indicators, and insufficiently examines distributional heterogeneity across vulnerable demographic groups. Furthermore, the specific role of relative humidity as a climate driver and the moderating effects of education and adaptive capacity remain under-explored in the Indonesian context. This study addresses these gaps by investigating how climate variability, specifically relative humidity, influences household food insecurity in two contrasting provinces: Aceh (coastal, high rainfall) and West Nusa Tenggara (NTB, arid, rain-fed).

Methodology and Data
The study utilizes pooled cross-sectional microdata from the 2018 and 2024 rounds of Indonesia's National Socioeconomic Survey (SUSENAS), encompassing 294,919 households in Aceh and 13,424 in NTB. Household-level data is merged with district-level climate indicators, including annual total rainfall, average relative humidity, average temperature, and wind speed.

• Dependent Variables: Food insecurity is measured using the Food Insecurity Experience Scale (FIES). The study constructs both a binary indicator (households experiencing ≥3 of 8 food insufficiency conditions) and a continuous score (0–8) representing the severity of food insecurity.
• Empirical Strategy: The analysis employs a stepwise hierarchical modeling approach.
  • Logistic Regression: Used to model the binary probability of food insecurity, controlling for socio-demographic, economic, and infrastructure variables.
  • Ordinary Least Squares (OLS) and Clustered Robust Regression: Used to analyze the continuous food insecurity score, with clustered robust standard errors applied to account for intra-cluster correlation (e.g., within villages or districts).
  • Heterogeneity and Moderation: The study conducts subgroup analyses (urban vs. rural, male-headed vs. female-headed households) and introduces interaction terms to test if household head education moderates the impact of relative humidity on food security.

Key Results

1. Impact of Relative Humidity: Empirical results consistently demonstrate that higher relative humidity significantly increases the likelihood and severity of household food insecurity across both provinces. In the logistic regression, a one-unit increase in average humidity correlates with a statistically significant rise in the probability of food insecurity.
2. Heterogeneous Effects: The adverse effects of climate variability are not uniform.
   • Vulnerable Groups: The negative impact of humidity is substantially stronger for households already experiencing high levels of food insecurity. Subgroup analyses reveal that rural and male-headed households bear the brunt of these shocks.
   • Regional Differences: In NTB, higher temperatures and humidity significantly increase food insecurity, reflecting the province's arid climate and reliance on rain-fed agriculture where heat stress accelerates soil moisture depletion. Conversely, in Aceh, humidity acts differently due to its tropical rainforest climate, though it remains a significant factor in the broader model.
3. Protective Factors:
   • Economic Status: Household expenditure is the strongest predictor of food security. Households in the highest expenditure quintile face a 90% lower risk of food insecurity compared to the poorest quintile.
   • Education: Higher education levels of the household head are significantly associated with reduced food insecurity. Crucially, education acts as a moderator; for more educated household heads, the adverse effects of humidity are effectively neutralized, suggesting enhanced adaptive capacity.
   • Infrastructure: Access to basic infrastructure, particularly safe drinking water, improved sanitation, and electricity, serves as a crucial mitigating factor. In NTB, the protective effect of safe drinking water is notably stronger than in Aceh, highlighting a greater dependence on infrastructure to offset natural environmental deficits.
4. Non-Significant Factors: Total precipitation (rainfall) generally did not show a statistically significant impact on food insecurity in the full models, suggesting that in these specific contexts, humidity and temperature are more critical drivers than total rainfall volume.

Significance and Contributions
The paper claims to make three primary contributions to the literature on climate vulnerability and food security:

1. Micro-Level Evidence: It provides household-level evidence on climatic vulnerability in Indonesia using nationally representative microdata, moving beyond aggregate agricultural outcomes.
2. Distributional Heterogeneity: By employing subgroup and clustered robust regression analyses, the study moves beyond average-effect estimations to reveal how climate variability disproportionately affects specific vulnerable groups (rural, male-headed, and already food-insecure households).
3. Adaptive Capacity: It examines the role of education as an adaptive capacity, demonstrating that human capital can mitigate the adverse impacts of climate shocks.

Policy Implications
The findings underscore the urgent need for policymakers to design climate-resilient rural development policies that integrate targeted adaptation strategies with poverty reduction, educational access, and infrastructure improvements. The study suggests that:

• Targeted Interventions: Rural, male-headed households require robust agricultural extension services focused on climate-smart practices and crop diversification.
• Infrastructure Investment: Expanding access to electricity, clean water, and sanitation is essential for securing food storage and hygiene, particularly in arid regions like NTB.
• Early Warning Systems: Incorporating local humidity and climate metrics into early warning systems can help anticipate food security risks before environmental shocks escalate.
• National Relevance: While focused on Aceh and NTB, the findings are relevant to the broader Indonesian archipelago, which shares high relative humidity and climate sensitivity, suggesting a need for national strategies that account for micro-level climate drivers in poverty alleviation programs (e.g., PKH and BPNT).