Drivers of rural–urban food insecurity disparities in war-affected settings of Tigray, Ethiopia: a decomposition approach to policy and intervention insights

The importance of food security in promoting sustainable development is paramount. Policymakers and researchers worldwide have been grappling with the challenges of ensuring a stable and sufficient food supply for growing populations, particularly in the face of climate change, economic disparities, and geopolitical tensions. Global initiatives such as the United Nations Sustainable Development Goals (SDGs) and continental frameworks such as Africa Agenda 2063 have been implemented to achieve comprehensive food security (African Union, 2015; United Nations, 2015). SDG Goal 2 endeavors to “end hunger, achieve food security, improved nutrition, and promote sustainable agriculture,” underscoring the global commitment to this cause (United Nations, 2015). The Africa Agenda 2063 set forth objectives to achieve food security by 2063, such as eradicating hunger, enhancing intra-African agricultural trade, modernizing agriculture, and empowering women and youth in agriculture (African Union, 2015).

These frameworks emphasize sustainable and inclusive solutions for hunger and malnutrition, resulting in substantial reductions in global food insecurity. Nevertheless, current data suggest a reversal of this trend. The Global Report on Food Crises (GRFC) 2024 reported that acute food insecurity will affect 282 million people in 59 countries by 2023, increasing from 105 million in 2016. The numbers have risen annually: 112.7 million in 2018, 137.4 million in 2019, 155.3 million in 2020, 192.8 million in 2021, and 257.8 million in 2022 (FSIN, 2024). Armed-conflicts disrupt food systems and livelihoods, exacerbating food insecurity.

The Tigray conflict in Ethiopia exemplifies how prolonged hostility disrupts food systems and affects both rural and urban populations. In 2020, rural areas suffered direct agricultural damage, whereas urban centers faced supply chain and market access issues. By August 2023, 26.5% of internally displaced individuals and 61% of pregnant and breastfeeding women in Tigray had been severely malnourished (FSIN, 2024). Crop destruction, livestock theft, and market disruptions worsen food insecurity, which is further aggravated by limited humanitarian aid due to blockades and violence (Geremedhn and Gebrihet, 2024). These conditions highlight the urgent need for tailored recovery strategies in conflict-affected regions.

Armed-conflicts significantly affect food insecurity by disrupting food production and access. Several studies have emphasized how conflicts damage food systems and have explored coping strategies in conflict-affected regions (Gebregziabher et al., 2023; Trogisch, 2023; Yohannes et al., 2023; Béné et al., 2024; Salihu et al., 2024; Eshetu et al., 2025; Gebrihet et al., 2025). These studies highlight the broad implications of conflicts in food security. However, they failed to explore the differences in the prevalence of household food insecurity between rural and urban settings in conflict-affected regions. Addressing food insecurity in rural and urban conflict zones requires tailored strategies that reflect each community’s unique challenges and that incorporate targeted interventions to mitigate conflict impacts and foster long-term resilience.

This study examines the rural-urban food insecurity disparities in Tigray. The Oaxaca–Blinder decomposition technique was also used to identify the key factors to the rural-urban disparities in household food insecurity. The findings aim to inform policy interventions that address the unique challenges faced by rural and urban populations during post-conflict recovery. This study aligns with SDGs 2030 and the Africa Agenda 2063, offering a data-driven examination of the difficulties faced by rural and urban communities in conflict zones. It also suggests policies to reinforce resilience and to incorporate food security into wider development strategies.

Food insecurity has become a significant global issue, closely linked to sociopolitical instability and economic limitations, especially in conflict-affected regions. This multifaceted issue encompasses four key dimensions: availability, access, utilization, and stability. Goal 2 of the SDGs seeks to eradicate hunger and ensure the availability, accessibility, utilization, and stability of food by 2030 (United Nations, 2015). In Africa, Agenda 2063 aligns with these objectives, aiming to modernize agriculture, empower marginalized groups, and eliminate hunger through enhanced intra-African trade (African Union, 2015). However, conflict often undermines these goals, exacerbating structural vulnerabilities and intensifying rural-urban disparities.

Theoretical perspectives on food insecurity in conflict settings are effectively framed by Amartya Sen’s Entitlement Theory (Sen, 1983), Stephen Devereux’s refinements (Devereux, 2001), and the Sustainable Livelihood Framework (SLF) (DFID, 2008), later adapted by Nithya Natarajan and colleagues (Natarajan et al., 2022). Sen’s Entitlement Theory argues that food insecurity arises not only from food shortages, but also from failures in individuals’ ability to access food due to diminished entitlements, such as income, land, or labor (Sen, 1983). Devereux critiques this framework for neglecting the political and social dynamics shaping food access, emphasizing that state actions such as aid restrictions and food distribution control play a crucial role in famine outcomes (Devereux, 2001). In conflict zones, disrupted livelihoods, market collapse, and displacement further erode entitlements, shifting focus from food production deficits to access barriers. The SLF complements these insights by emphasizing how livelihood sustainability depends on access to diverse assets within social, political, and environmental contexts (DFID, 2008). Natarajan et al. (2022) argued that the framework must evolve to address contemporary global challenges by incorporating power dynamics, ecological limits, historical legacies, and diverse knowledge systems.

Both frameworks highlight the fact that food insecurity is shaped by economic, political, social, and ecological factors. Devereux’s refinements provide a critical lens for examining how state power exacerbates rural-urban food insecurity disparities in conflict settings. Similarly, Natarajan’s updated SLF emphasizes the importance of addressing power asymmetries and environmental vulnerabilities. In the war-affected Tigray, these theoretical advancements clarify how disrupted entitlements, weakened institutions, and ecological pressures intensify food insecurity. This decomposition-based approach identifies the structural drivers of these disparities and offers policy insights into conflict-affected regions.

Empirical studies in conflict-affected settings have highlighted the pervasive impact of war on food systems and the diversity of its manifestations. In Tigray, Ethiopia, the 2020–2023 conflict resulted in 77% of households experiencing food insecurity, driven by destroyed agricultural infrastructure, livestock losses, and restricted market access (Araya and Lee, 2024; Gebrihet et al., 2025). Comparable patterns have emerged worldwide.

In Yemen, prolonged conflict has triggered a humanitarian crisis, 17 million people are food insecure, 4.7 million of them at crisis levels by 2023, as rural households faced a 40% reduction in agricultural output while urban areas grappled with import disruptions (Hashim et al., 2021; UNICEF, 2025). South Sudan’s civil war similarly devastated rural livelihoods, with 7.7 million people acutely food insecure in 2024, driven by a 30% decline in cereal production and urban inflation exceeding 50% (FAO and UNICEF, 2024). In Syria, conflict weaponization of food—through sieges and agricultural destruction—left 12.1 million people food insecure by 2023, with rural areas losing 60% of arable land and urban centers reliant on erratic aid (WFP, 2023). The ongoing crisis in Syria has greatly disrupted agriculture in a country that was previously food producing and self-sufficient (Ibrahim et al., 2024). In Burkina Faso, terrorism has reduced crop yields and household income, with a 2023 study reporting a 35% drop in rural food access linked to displacement (Kafando and Sakurai, 2024). Meanwhile, in Gaza, despite humanitarian aid, conflict undermines household resilience, with urban food insecurity rising by 20% in 2024, owing to blockade-induced price hikes (Brück and d’Errico, 2019). These cases underscore the common drivers of food insecurity in conflict zones, including disrupted agricultural production, economic instability, and gendered vulnerability.

Prior studies that explicitly compare rural and urban food insecurity in conflict settings further justify this study’s focus on these disparities. In South Sudan, there are significant disparities in food security, with approximately 75% of rural residents experiencing moderate to severe food insecurity compared to 52% of urban residents (World Bank, 2024). A study in Somalia found that rural food insecurity was exacerbated by livestock losses, reducing dietary diversity by 30%, whereas urban households mitigated shortages through remittances, albeit at a higher cost (Maxwell and Fitzpatrick, 2012). An analysis of Syria’s conflict revealed that rural households were twice as likely to face acute malnutrition compared to their urban counterparts because of destroyed irrigation systems, whereas urban food access hinged on volatile aid deliveries (Doocy and Lyles, 2017). More recently, in northern Ethiopia, rural households in conflict-affected areas are more vulnerable to food insecurity than urban households (Eshetu et al., 2025). These studies highlight distinct rural-urban vulnerabilities—production losses versus market dependency—reinforcing the need to decompose these dynamics in Tigray’s post-conflict context.

Coping strategies vary by context and reflect entitlement and livelihood adaptation. In Mali, food aid supported household expenditure and nutrient availability during conflict, although rural-urban differences persisted (Tranchant et al., 2019). In Nigeria, strategies vary among internally displaced individuals and host communities shaped by demographic and contextual factors (Salihu et al., 2024). In the Democratic Republic of Congo (DRC), women in conflict zones adopted adaptive strategies such as seed-sharing networks, yet 65% remained food insecure in 2024 due to ongoing violence (Trogisch, 2023). In Ethiopia’s Offa District, households reduce their consumption and sell assets (Masha et al., 2023), whereas Tigrayan households skip meals or seek community support (Gebregziabher et al., 2023). Studies emphasize the need for interventions tailored to local contexts, integrating traditional coping mechanisms with modern frameworks, such as the SLF, to rebuild resilience (Yohannes et al., 2023). For instance, WFP interventions in Uganda reduced household food spending by 35% and improved meal frequency, although gender disparities persisted (Han et al., 2020). These findings align with Sen’s argument that addressing entitlement failure through income generation or asset restoration is critical to overcoming food insecurity in conflict settings.

This study builds on these insights to examine rural-urban disparities in food insecurity prevalence and coping strategies in post-conflict Tigray. By grounding the analysis in entitlement theory and the sustainable livelihood framework, it seeks to decompose the drivers of these disparities, offering policy insights to enhance food system resilience amid the lingering effects of war.

Tigray, bordered by Sudan, Eritrea, Afar, and Amhara, is divided into seven zones, 93 Weredas, and 814 “Tabia” (Araya and Lee, 2024). Seasonal labor and smallholder agriculture, focusing on livestock and cereals, are vital for rural food security in the region, which has one main rainy season from June to September (Clark, 2021). Tigray’s population of 5,685,598, with an average household size of 4.2, comprises 49% males, 51% females, 50% under 18 years old, and 30% urban residents (Araya and Lee, 2024). Figure 1 shows a map of the study area.

This study used a cross-sectional household survey conducted in war-affected Tigray from May to June 2024 to examine the drivers of rural-urban disparities in household food insecurity. A multistage sampling approach was employed. First, the Central and Northwestern Tigray zones were purposively selected due to resource constraints and their exposure to the 2020–2022 conflicts with the Ethiopian and Eritrean forces. Second, five rural Weredas (Ahferom, Adwa, Tahtay Machew, Tahtay Adiyabo, and Tahtay Koraro) and five towns (Enticho, Adwa, Axum, Shiraro, and Shire) were chosen randomly using the lottery method. Three tabias (the smallest administrative units) were randomly selected from each Wereda and town to ensure unbiased rural and urban representation. Finally, 1,800 households were randomly selected using systematic sampling.

The sample size was computed using Yamane’s (1967) formula as follows:

where $n$ is the required sample household size, $N$ is the target finite household population, $e$ is the marginal error (0.05) and 1 is a constant. This yielded 1,800 households (1,060 rural and 740 urban), which were subsequently selected randomly through systematic sampling from the household list.

A structured questionnaire based on the Food Insecurity Access Scale (FIAS) and Food Insecurity Experience Scale (FIES) investigated food insecurity drivers, covering household demographics, food access, and conflict impacts. Refined with local expert input, it was translated into Tigrinya and pre-tested with 30 households in Adwa for clarity and relevance. The data collectors received training on data collection methods. Face-to-face interviews using the structured questionnaire were conducted with household heads or adult representatives for 30–45 min. Key informant interviews (KIIs) focused on identifying coping strategies using a semi-structured guide with 15 purposively selected informants—town administrators and community leaders (three per Wereda)—each lasting about 60 min. A team of 15 trained enumerators fluent in Tigrinya conducted both processes following training in questionnaire administration, KII techniques, and ethical standards.

When analyzing food insecurity, it is essential to recognize that the assessment of food security relies on various scales developed by analysts, using diverse indicators and benchmarks. Prominent techniques include the Household Food Insecurity Access Scale (FIAS) and Food Insecurity Experience Scale (FIES), both of which provide valuable frameworks for classifying and understanding food security levels.

The FIAS assesses household food insecurity over 30 days before the survey and provides household-level data on food insecurity (Coates et al., 2007). It includes nine “occurrence” and “frequency-of-occurrence” questions to differentiate between food-secure and food-insecure households. Households were asked whether these conditions had occurred over the past 30 days (yes or no). If “yes,” a follow-up question determined if the condition occurred rarely (once or twice), sometimes (three to ten times), or often (more than ten times). Each household’s FIAS score (0–27) was calculated by summing the “frequency-of-occurrence” codes, indicating food insecure access (Coates et al., 2007). The FIAS assesses food access, availability, and stability by multiplying “occurrence” with “frequency-of-occurrence” questions (Tables A1 and 1). Table A1, which includes the FIAS questions, is provided in the  Appendix. The total frequency of occurrence of the nine food insecurity conditions over the past 30 d was calculated as follows:

Where: $Q1,a_Q9,a$ refer to the “frequency-of-occurrence” question (once or twice, sometimes, or often) and $xi$ refers to the $ith$ sample household.

The FIES measures food insecurity experienced at the household level using a series of eight dichotomous yes or no questions over the past 30 days. It evaluates individuals’ food security based on their direct responses (FAO, IFAD, UNICEF, WFP. and WHO., 2023). The scores ranged from 0 to 8, with higher scores indicating a greater food insecurity (Tables A1 and 1).

The sum of the “frequency of occurrence” of the eight dichotomous food insecurity conditions over the past 30 days was computed as

Households were categorized into four food security statuses: food secure, mild, moderate, and severe. By combining mildly, moderately, and severely food insecure households into a single category, the dependent variable had two categories: food security and food insecurity. Owing to this dichotomous nature, a logistic regression model was used to analyze the data and predict outcomes within 0 and 1 (Long and Freese, 2014). A binary variable was created, assigning one to food-secure households and zero to food-insecure households (Table 1). Binary logistic regression was applied when the dependent variable was dichotomous and the independent variables were continuous or categorical (Long and Freese, 2014). A logistic regression analysis examines a binary outcome $(Yi)$ and independent variables or covariates $(X1,X2,X3…Xi)$ to infer event probabilities in a population. Each household in the sample had a probability “p” of experiencing food security or insecurity. In a sample of n” households, $Yi$ represents the likelihood of being food-secure or -insecure for the $ith$ household, where $Yi$ = 1 (food-secure) or $Yi$ = 0 (food insecure).

The logistic regression model with natural log odds can be written as:

Where: $X$ is the independent variable; ′$ln′$ is the natural logarithm; ′$p′$ is the probability of the interested outcome that $Yi$ for food insecurity equals 1, $p(Yi=1$⁠); ′$1−p′$ is the probability that $Yi$ for food insecurity equals 0, $1−p$ (⁠$Yi=1$⁠); $p1−p$ is the odds; $ln(p1−p)$ is the “logit” or the “log odds.” The logistic regression parameters were $α0$ and $αk$⁠. This study employs a simple logistic regression model. By taking the antilog of Equation (5), the probability of household food insecurity status can be expressed as:

Extending the natural logarithm of Equation (6) with multiple covariates can be rewritten as

The households were classified into rural and urban areas. The dependent variable was a dummy variable indicating food security (1) or insecurity (0) based on the FIAS and FIES scores. Table 2 presents the study variables based on the literature and theory.

The Oaxaca-Blinder decomposition (OB), developed by Ronald Oaxaca and Alan Blinder in the 1970s, is a statistical method used to analyze outcome differences between two groups, such as the wage gap between men and women (Blinder, 1973; Oaxaca, 1973). While commonly applied to wage disparities, it can also explain differences in binary outcomes such as food security status between groups (Fairlie, 2005). Unlike logistic regression, which identifies associations, or structural equation modeling, which explores latent relationships, OB decomposition offers a distinct advantage by partitioning the gap into explained and unexplained components, thus providing actionable insights for policy. This study applies OB decomposition to examine rural-urban disparities in household food insecurity as it quantifies the contributions of observable characteristics to binary outcomes. OB has been used effectively in similar contexts, such as in Gebresilassie et al. (2021), who explored rural-urban differentials in under-five child mortality in Ethiopia, and Samuel et al. (2021). Assessed maternal health care utilization. Allen et al. (2022) used OB to analyze health inequalities in Wales, demonstrating its versatility in addressing socioeconomic disparities in conflict-affected areas; for example, Tigray and Gebre et al. (2021) applied OB to explore gender gaps in food insecurity among maize-producing households in Ethiopia.

For binary variables, such as the distinction between food-secure and food-insecure households, traditional OB decomposition is insufficient. Instead, an extended OB decomposition technique is required to attribute differences in household food insecurity levels to various determinants (Fairlie, 2005). The factors contributing to rural-urban disparities in household food insecurity, analyzed using the extended OB decomposition estimation technique proposed by Fairlie (2005), are described as follows:

where $X̅i$ is a row vector of the average values of the covariates and $σ̅i$ denotes the vector of the estimated coefficient for rural and urban areas.

An extension of this OB decomposition to a nonlinear equation, $FIN=(Xσ̅i)$ was computed as follows:

We define $FIN̅i$ as the average probability of the binary outcome of interest group $i$ (household food insecurity), and $F$ as the cumulative distribution function from the logistic distribution. ′$N′$ represents the sample size of urban and rural areas. The first term on the left-hand side of Equations (8) and (9) provides an estimate of the contribution of the entire set of covariates to rural-urban disparities at the household level of food insecurity. Thus, following Fairlie (2005), based on the estimated coefficients from the logit regression model for the pooled sample (Equation 7), the contribution of individual covariates to disparities in the household level of food insecurity between rural and urban areas was decomposed as follows:

The contribution of the individual covariate to household food insecurity in rural and urban areas equals the change in the average predicted probability when urban distribution is replaced with rural distribution, keeping all other covariates constant.

Quantitative analysis was performed using Stata (version 17), employing logistic regression to model food insecurity odds and OB decomposition to dissect rural-urban disparities. Robustness checks included testing for multicollinearity (Variance Inflation Factor <5), assessing model fit (Hosmer-Lemeshow test, p > 0.05), and conducting sensitivity analyses by re-estimating models with subsamples (e.g. excluding displaced households), confirming result stability. These procedures ensured reliability and validity of the findings. The results are presented in the Tables and Figures.

Enumerators obtained verbal consent, ensured confidentiality, and explained the study’s aim of assessing food security in conflict-affected areas. Ethical Approval was granted by the authorized institution, as detailed in the ethical approval section. Data quality was ensured through daily supervisor checks for accuracy and resolution of inconsistencies with the enumerators. The responses were recorded on paper and later entered into Excel by a data entry team.

Table 3 presents the prevalence of food insecurity across the sampled households. This provides a comprehensive overview of food insecurity levels and their rural-urban distribution, establishing the critical disparity that underpins this study’s investigation into conflict-driven food security challenges in Tigray. Of the sampled households, 42% were female and 30.67% had food insecurity. Among 58% of male-headed households, 42.22% were food insecure. Overall, 73% were food insecure, 33.44% experienced mild, 28% moderate, and 11.44% severe food insecurity. Rural households had higher food insecurity (73%) than did urban households (27%). The significant differences in food insecurity between rural and urban households were due to disparities in education, marital status, humanitarian aid, employment, and conflict-related damages.

Examining the drivers of rural-urban household food insecurity disparities requires a separate assessment of the impact of these factors on household food insecurity in urban, rural, and combined areas. Therefore, a logistic regression analysis was conducted on the factors influencing household food insecurity in war-affected settings for urban, rural, and pooled data to determine if the factors differed between these areas.

This section provides an overview of the determinants of household food insecurity based on logistic regression analysis. While this analysis is important for understanding the factors influencing food insecurity, the main focus of this study is on the drivers of rural-urban disparities in food insecurity, as discussed in Section 4.2.2. The detailed determinants and regression results are presented in Table A2 in the  appendix to maintain a clear focus on the key drivers of rural-urban disparities. This approach emphasizes the broader patterns and insights in Section 4.2.2, while informing readers of the factors contributing to food insecurity.

Using FIAS and FIES estimates, households with a high school education had a lower likelihood of food insecurity than those with primary or lower education. Female-headed households had significantly higher odds of food insecurity than male-headed households, in both rural and urban settings. Widowed households, both rural and urban, were more likely to experience food insecurity than married and divorced households were, holding other factors constant. In addition, households with a high school education level had lower odds of food insecurity. Families with a higher dependency ratio are more likely to experience food insecurity, while those with access to credit services are less likely to face such challenges, assuming that all other factors remain constant. Humanitarian aid recipients had lower odds of food insecurity than did non-recipients. Households affected by armed-conflict in Tigray from 2020 to 2022 have higher odds of food insecurity.

Table 4 details the decomposition of household food insecurity disparities between rural and urban households. This table detailed the breakdown of the rural-urban food insecurity gap using OB decomposition, identifying the proportional contributions of key drivers and offering a foundation for understanding structural disparities and informing targeted policy responses.

A positive covariate contribution indicates widening disparities, whereas a negative contribution indicates that a covariate helps to reduce these disparities. The explained components accounted for 67 and 56% of rural-urban disparities in household food insecurity prevalence using FIAS and FIES estimates, respectively. The remaining differences were due to the unexplained covariates.

Despite the varying magnitudes of the contribution between the FIES and FIAS scores, the direction of the covariate contributions remains consistent. A detailed decomposition analysis showed that female-headed households contributed 7.25 and 6.43% to food insecurity disparities between rural and urban households using the FIAS and FIES, respectively, ceteris paribus. Proportional differences in the dependency ratio accounted for 4.32 and 5.98% of these disparities, using the FIAS and FIES scores, respectively.

The difference in household size explained 4.32 and 4.98% of the disparities using the FIAS and FIES, respectively. In addition, the marital status of household heads played a crucial role, and the proportional differences in households led by widowed heads significantly contributed to the prevalence of food insecurity in both rural and urban areas, explaining 12.58 and 10.16% of the disparities, respectively, according to FIAS and FIES estimates. The proportional differences among households affected by prolonged armed-conflict accounted for rural-urban disparities in food insecurity prevalence of 58.39 and 49.58%, using FIAS and FIES estimates, respectively. This indicates that armed-conflict damage had a strong, statistically significant effect on rural-urban gaps in household food insecurity prevalence. In contrast, the proportional differences among households receiving humanitarian aid significantly widened, contributing 17.38 and 19.76% to rural-urban disparities in household food insecurity according to FIAS and FIES estimates, respectively.

Based on the FIAS and FIES estimates, differences in household employment status and income widened these gaps by 9.37 and 6.98%, respectively. The age of the household head accounted for 3.84 and 2.1% of the rural-urban disparities in food insecurity using the FIAS and FIES scores, respectively. Differences in the educational status of household heads (high school education) reduced rural-urban food insecurity disparities by 11.28 and 7.82% using FIAS and FIES, respectively. Employment differences among households explained the rural-urban food insecurity gaps by 9.37 and 6.98% using FIAS and FIES estimates, respectively.

Figure 2 illustrates the coping strategies for war-induced food insecurity among rural and urban households.

Rural households use various coping mechanisms. The most common stress-level strategies were consuming less-preferred food (56%), reducing meal size (45%), off-farm income and petty trading (34%), and food loans (31%). More than half adopted crisis-level strategies such as reducing healthcare expenses (50%), eating reserved seeds (51%), and withdrawing children from school for food and wages (56%). The least-adopted crisis strategy was to sell productive assets (31%). Additionally, 23% engaged in begging and 22% rented their farmlands. These findings suggest that food-insecure households in war-affected Tigray use multiple strategies to address severe food insecurity.

Urban strategies included selling assets (41%), reducing meal size, consuming less preferred food (35%), borrowing money (27%), and saving (23%). Despite being uncommon, 21% of urban households turned to daily labor in Tigray. The least-used strategy was to move the children to less expensive schools (12%). Urban households primarily employed stress-level strategies such as selling productive assets (22%), reducing healthcare spending (17%), and withdrawing children from school (16%). Emergency strategies included migration (13%), selling or mortgaging houses (12%), hazardous work (11%), and begging (8%).

This section is organized into three subsections, focusing on rural-urban disparities, coping strategies, and unexplained factors, with their broader implications. The remaining descriptive statistics are not the focus, as the primary concern is to analyze these three dimensions, where deeper theoretical and contextual insights are most relevant.

The analysis revealed explicit rural-urban disparities in Tigray’s food insecurity landscape, with rural households experiencing greater vulnerability than urban households do. A disparity decomposition analysis attributed this to factors such as gender, employment status, education, armed-conflict damage, dependency ratio, humanitarian aid, and marital status (see Table 4, Section 4.2.2). The decomposition presented in Table 4, accounting for 67% (FIAS) and 56% (FIES) of the observed gap, is crucial for identifying the conflict damage (58.39% FIAS) as the predominant factor. This finding highlights the extent to which disruptions in rural agriculture exacerbate disparities and offers critical insights for directing recovery efforts. The employment status of the household head was found to be significant in determining rural-urban gaps in the prevalence of food insecurity. This factor indicates that the probability of being food secure increases with an increased probability of being employed. A possible explanation is that prior to the conflict, unemployment in Tigray stood at a modest 4.49%; however, post-war, nearly half of the households reported job losses, severely curtailing income generation (Araya and Lee, 2024). This result is consistent with the findings of Araya and Lee (2024), and Gebrihet et al. (2025).

Rural areas that are heavily reliant on small-scale subsistence agriculture have been disproportionately affected by the destruction of crops and livestock, a vulnerability compounded by limited infrastructure and market access (FAO et al., 2017; IFAD, 2021). By contrast, urban households benefit from proximity to markets and aid networks, although they remain susceptible to poverty and income instability (Ngema et al., 2018; Lutomia et al., 2019). Araya and Lee (2024) corroborate this disparity, noting higher insecurity rates in rural Tigray than in urban areas, which is consistent with our results.

The educational attainment of the household head contributes to rural-urban disparities in food insecurity prevalence. Education equips individuals with the knowledge to develop their livelihood strategies. This finding highlights the importance of education in improving household food security, as educated households are more likely to engage in family planning, limit household size, and enhance their capacity to manage food demands. Educated households often engage in off-farm income-generating activities. Educated farmers are more likely to adopt drought-resistant agricultural technologies, resulting in higher yields at harvest, and contributing to food security. Education can enhance household resilience through improved income opportunities and nutritional knowledge (UNESCO, 2021; Srivastava and Muhammad, 2022), its scarcity in rural settings exacerbates this gap. Additional burdens, such as widowhood and high dependency ratios, further disadvantage rural households (Weldegiargis et al., 2023), underscore structural inequities that shape divergent coping responses, as detailed below.

The gender of the household head is a significant demographic factor explaining rural-urban disparities in household food insecurity prevalence, and the higher proportion of female-headed households is a notable factor in explaining rural-urban gaps. This can be attributed to the fact that male-headed households possess greater resource endowments than their female-headed counterparts in rural Ethiopia. Furthermore, even when female- -and male-headed households have equivalent levels of resource endowments, unobservable characteristics also play a role in the disparities in their food security prevalence.

The damage resulting from armed conflict plays a significant role in widening the rural-urban disparities in the prevalence of household food insecurity in Tigray. The protracted armed-conflict affects household food insecurity in several ways, including the reduction of crop production, the induction of forced migration, and the reduction of productive labor forces, all of which mitigate household food insecurity (Araya and Lee, 2024).

The dependency ratio constitutes a significant socioeconomic factor contributing to the exacerbation of rural-urban household food insecurity disparities in the war-affected region of Tigray. A plausible explanation for this gap is that a higher dependency ratio in rural settings imposes a substantial burden on the working-age groups of the population within the household, thereby reducing the household income available for food consumption that exacerbates food scarcity. Similarly, urban households characterized by high dependency ratios experience a reduction in financial income, which in turn reduces their capacity to purchase food items. This finding aligns with those of previous studies (Araya and Lee, 2024; Gebrihet et al., 2025).

Rural–urban food insecurity disparities in Tigray reflect deeper structural vulnerabilities that extend beyond the immediate impacts of conflict. While war exacerbates the crisis, the pre-existing fragility of rural economies, which depend on small-scale subsistence agriculture, amplifies these disparities. Conflict damage accounts for much of the disparity, revealing that food security is not just about access or availability, but is deeply embedded in economic structures, social stratification, and historical inequalities. Rural areas, unlike urban centers, lack the capacity to absorb economic shocks, leading to higher unemployment and fewer income-generating opportunities. These disparities are not just a symptom of war, but a reflection of the long-standing structural divides that conflict exacerbates.

In response to food insecurity, Tigray’s households deploy a range of coping strategies that reflect both their immediate circumstances and underlying vulnerabilities. Rural households facing the profound effects of conflict often resort to extreme measures, such as consuming seed reserves, withdrawing children from educational institutions, and engaging in begging. These responses are particularly prevalent among female-headed households, which are constrained by traditional gender roles (Weldegiargis et al., 2023; Yohannes et al., 2023).

Conversely, urban households adopt stress-level strategies, including asset sales, meal reduction, borrowing, capitalizing on market access, and humanitarian support (Sani and Kemaw, 2019; Srivastava and Muhammad, 2022). Education mitigates the severity of these responses, with more educated households being less likely to engage in detrimental long-term trade-offs (Masha et al., 2023), This is evident in the lower incidence of school withdrawal in urban areas than in rural ones. These findings resonate with broader patterns in conflict-affected regions, where rural coping often compromises future stability, whereas urban strategies provide temporary relief (Bahiru et al., 2023).

The sustainable livelihood approach (SLA) (Natarajan et al., 2022) frames Tigray’s coping strategies as asset depletion: natural capital erodes in rural areas, whereas financial resources strain urban households. Prolonged conflict dismantles livelihoods, forcing reliance on fragile traditional practices (Sani and Kemaw, 2019) and raising questions about long-term resilience. Rural households adopt irreversible measures—depleting seed reserves and withdrawing children from school— whereas urban households use market access for short-term relief. Female-headed households bear the greatest burden, reflecting the entrenched structural inequities. Through the SLA lens, these responses signal not only asset depletion but also the collapse of resilience. When survival strategies intensify vulnerability, the core challenge shifts from recovery to breaking the crisis cycle.

Despite the robustness of our decomposition analysis, a notable portion of the rural-urban disparity–33% (FIAS) and 44% (FIES)–remains unexplained, suggesting the influence of factors beyond the scope of our measured variables. Potential contributors include gender norms that may restrict rural women’s access to aid or markets (Weldegiargis et al., 2023), as well as fluctuations in urban market conditions that escape capture within our dataset (IFAD, 2021). Policy shortcomings, such as inconsistent aid distribution or degraded infrastructure, likely elaborate on rural disadvantages (OCHA, 2021), while urban social networks may confer unmeasured resilience absent in rural isolation (Yohannes et al., 2023). These possibilities align with George et al. (2021), who highlight the role of latent social and economic dynamics in conflict settings.

From a theoretical perspective, Sen’s Entitlement Theory interprets the impact of conflict as a fundamental loss of access to resources, with gender disparities underscoring systemic inequities (Devereux, 2001). The sustainable livelihood approach (SLA) situates these findings within a framework of asset shocks—employment losses and aid dependency eroding household capital. However, questions arise about the feasibility of recovery in such contexts (Natarajan et al., 2022). Rural coping strategies threaten long-term viability, while urban reliance on external support suggests fragile equilibrium, indicating the need for structural interventions. Building on Kafando and Sakurai (2024), these implications underscore the complexity of food insecurity in Tigray and frame subsequent recommendations for addressing this multifaceted crisis.

This study reveals that food insecurity in war-affected Tigray is not merely a consequence of conflict, but a structural crisis that deepens pre-existing rural-urban inequalities. Using the Oaxaca-Blinder decomposition, the analysis attributes 67% (FIAS) and 56% (FIES) of the rural-urban food insecurity gap to observable factors, including conflict-induced agricultural damage, gender inequalities, employment disruptions, education levels, and unequal distribution of humanitarian aid. Rural households, reliant on subsistence farming, experienced a food insecurity rate of 73%—nearly triple that of urban households (27%)— because of crop and livestock losses, weak infrastructure, and poor market access. Although urban households face economic instability, their proximity to markets and aid networks provides a degree of structural resilience in rural areas. Conflict damage alone accounts for up to 58.39% of the disparity, while the erosion of entitlements, particularly through lower educational attainment in rural areas, further entrench vulnerability. This study underscores education as a critical protective factor and highlights how gender disparities and household dependency burden perpetuate cycles of hunger.

Coping strategies reflect rural-urban disputes. Rural households resorted to extreme, irreversible measures, such as consuming seed reserves and withdrawing children from school, which compromised long-term stability. In contrast, urban households rely on fragile, market-based solutions, such as asset sales, which provide only temporary relief. These findings emphasize that sustainable post-conflict recovery in Tigray requires more than immediate humanitarian aid; it demands comprehensive efforts to rebuild food systems and address the socioeconomic inequalities exacerbated by war.

To facilitate sustainable recovery, we propose three targeted policy interventions:

1. NGOs should prioritize female-headed rural households by providing seeds, fertilizers, and credit access to address gender disparity along with long-term land tenure support to rebuild livelihoods.

2. Governments should implement youth-focused vocational training and job creation programs in urban areas, mitigate unemployment, and foster economic recovery.

3. Humanitarian agencies should deploy mobile aid and cash transfers for displaced communities, addressing the conflict impact, while supporting rural cooperatives to enhance market access. In the long term, rebuilding the educational infrastructure is essential to empower youth and women, while fostering market linkages for traders can reduce aid dependency and promote commercial resilience.

These strategies address food insecurity while advancing SDG 2 (Zero Hunger) and SDG 1 (No Poverty) through improved food systems and economic opportunities. Supporting education and gender equality aligns with SDG 4 (Quality Education) and SDG 5 (Gender Equality), while improving market infrastructure supports SDG 8 (Decent Work) and SDG 10 (Reduced Inequalities). They also align with Africa’s Agenda 2063: Aspiration 1 (inclusive growth and sustainable development), and Aspiration 3 (good governance, democracy, and human rights).

This study explored the drivers of rural-urban disparities in food insecurity and coping strategies in post-war Tigray, Ethiopia. However, this study has some limitations. Self-reported quantitative data may have introduced recall bias, affecting the accuracy of past food consumption and coping strategies. Additionally, the household-level focus may overlook community dynamics and the influence of external factors, such as government policies and humanitarian aid. Including community perspectives can offer a more comprehensive understanding of war-related food insecurity. Given the gendered impact of food insecurity, future research should investigate intrahousehold gender disparities resulting from armed conflicts in Tigray, Ethiopia.

The authors express their gratitude to the rural and urban households for their participation in the survey and acknowledge Aksum University for providing the research grant for this study.