Effects of the current land tenure on augmenting household farmland access in South East Ethiopia

The study examines whether Ethiopia’s current state-ownership land tenure system augments household access to farmland, amid longstanding controversies over its effectiveness. Ethiopia has undergone successive land tenure regimes (Imperial, Derg, and current post-1991 system), with the current policy permitting allocation, inheritance, gifts, and leasing but prohibiting sale, exchange, and mortgage. Despite legal access provisions, national evidence shows small average farm sizes, landlessness, and fragmentation. The study focuses on Arsi Zone, Oromia Region, to assess: (1) household farmland sizes, (2) the perceived legitimacy of the current land tenure, and (3) determinants of household farmland access, using a conceptual framework where institutional, demographic, socioeconomic, and agro-ecological factors influence household farmland access.

The paper reviews Ethiopia’s land tenure evolution and comparative evidence on smallholder farm sizes globally and in Africa. Under Imperial tenure (rist/gult), land markets existed for a minority; the Derg’s 1975 reform nationalized land and set ceilings but led to unintended outcomes (fragmentation, villagization). The current constitution maintains state ownership with limited transfer rights and conditional redistribution. Literature indicates Ethiopia’s small average farm sizes (0.8–1.15 ha), high landlessness, and fragmentation, with regional disparities. Studies identify age, oxen ownership, education, income, technology use, and institutional arrangements (leasing, sharecropping) as correlates of land access. Debates persist over whether state ownership prevents concentration or undermines tenure security and efficiency. Evidence suggests growing pressures (population, infrastructure, urban expansion) may spur illicit land markets.

Design and setting: Concurrent mixed-methods design combining quantitative household survey with qualitative focus group discussions (FGDs) and key informant (KI) interviews. Study area is Arsi Zone, Oromia Region, Southeast Ethiopia, covering diverse agro-ecological zones (Kolla, Woina-Dega, Dega) across three purposively selected districts (Dodota, Hitosa, Tiyo). Data collection occurred May–June 2019. Sampling: Multistage sampling. From 25 districts, three were purposively selected due to acute land access issues and diverse agro-zones. Two rural kebeles per district were randomly selected (total six). Sample size determined using Kothari’s formula with z=1.96, p=0.75, q=0.25, e=0.05, N=5213, yielding n=273; with 15% allowance, 314 households were sampled (6.02% of population) via systematic random sampling from kebele frames. Qualitative sample included six FGDs (6 participants each; total 36) comprising development agents, farmers, kebele admins, women and youth leaders, plus 8 KIs (district and zonal administrators/agricultural heads). Instruments and procedures: Pre-tested closed-ended questionnaire on demographics, socioeconomics, tenure and land access, administered by trained development agents. Of 314 sampled, 310 responded (98.7%). FGDs used semi-structured guides on land access trends and perceptions of tenure; KIs addressed land access realization, gender balance, and landlessness. Measures: Dependent variable—household farmland access/total farmland size. Explanatory variables—sex, age, education, household size, number of oxen, total crop production, annual income, credit access, evaluation of land tenure (coded per Table 1). Analysis: Quantitative analysis with SPSS v22 and Stata v15.1 using descriptive statistics, Pearson correlations, chi-square tests for association (gender by tenure arrangement), independent t-tests (on-farm vs non-farm), and one-way ANOVA with Tukey post-hoc (agro-zones differences). An econometric Heckman two-stage sample selection model was used to estimate determinants of total farmland size and control selection bias; model fit assessed via Wald chi-square and significance of lambda. Qualitative data were analyzed thematically (Braun & Clarke) and integrated with quantitative findings. Ethical approval obtained from Bahir Dar University; informed consent, confidentiality, and anonymity assured.

- Descriptive farm size: Mean household farmland size = 1.59 ha (SD 0.83; range 0.2–4.5 ha). Male-headed households averaged 1.62 ha (SD 0.88) vs female-headed 1.50 ha (SD 0.63). By age: 65–76 yrs mean 2.26 ha vs 25–34 yrs mean 1.12 ha. - Tenure arrangements (primary mode reported): Government allocation 41.9% (n=130), inheritance 25.8% (n=80), gift 22.5% (n=70), leased 6.5% (n=20), sharecropping 2.3% (n=7), purchased 1.0% (n=3). Average farm size by arrangement: allocation 1.95 ha; inheritance 1.53 ha; gift 1.24 ha; sharecropping 1.41 ha; leased 0.81 ha; purchased 1.17 ha. - Gender and access mode: Chi-square significant association between gender and farmland access mode (χ2=56.36, df=5, p<0.001); male heads more advantaged. - Agro-ecological differences: Mean farm size by zone: Kolla 1.78 ha, Dega 1.51 ha, Woina-Dega 1.46 ha. One-way ANOVA significant: F(2,307)=4.70, p=0.010. Tukey: Kolla significantly greater than Dega (mean diff 0.276, p=0.043) and Woina-Dega (mean diff 0.320, p=0.014); Dega vs Woina-Dega not significant. - Socioeconomic correlates: Number of oxen: 0 ox → 1.20 ha; 1–2 oxen → 1.55 ha; 3–4 oxen → 2.14 ha; 5–6 oxen → 1.83 ha. Total crop production (quintals/HH): min 0, max 258, mean 38.04; 75.8% produced 1–50 qt. Pearson correlations: farm size vs total crop production r=0.43 (p<0.001); farm size vs annual income r=0.25 (p<0.001). Education inversely related to farm size (e.g., cannot read/write 1.99 ha; primary 1.45 ha; secondary 1.03 ha); correlation negative (r=-0.37, p=0.001). On-farm vs non-farm: annual income higher among on-farm (M=37,781 Birr) vs non-farm (M=18,252 Birr), t(308)=6.28, p<0.001; farm size larger for on-farm (M=1.64 ha) vs non-farm (M=0.69 ha), t(308)=8.46, p<0.001. - Illicit transactions: Despite a legal ban on sales, clandestine farmland purchases occur, often formalized as gifts; reported land prices 200–1000 Birr/m², with kebele officials allegedly facilitating legalization. Land purchased as primary mode only 1% vs gifts 22.5%. - Perceptions of current tenure: 37% rated current land tenure as good; 63% perceived it as not good for augmenting farmland access. Dissatisfaction higher among female-headed households and younger heads (<45 years). - Regression (Heckman two-step): Model significant (Wald χ2(9)=160.19, p<0.001). Determinants of larger household farmland size included number of oxen (positive), total crop production (positive), annual income (positive), education (reported as negative in correlations and narrative), and access to credit (positive). Household size showed mixed evidence (limited significance in model; positive correlation earlier).

The findings show that Ethiopia’s current state-ownership land tenure has not effectively augmented equitable household access to farmland in the study area. Access remains imbalanced by gender (male advantage), age (older farmers holding more land), and agro-ecology (Kolla zone with larger holdings than Dega and Woina-Dega). Institutional arrangements intended to provide access—allocation, inheritance, gifts, leasing, sharecropping—are important but insufficient in aggregate, with government allocation constrained and young adults and women disadvantaged. The widespread perception (63%) that the current tenure is not good for augmenting access, coupled with evidence of clandestine land transactions masked as gifts, indicates gaps between legal frameworks and local practices, undermining tenure legitimacy and policy goals. Socioeconomic endowments drive access: oxen ownership, production capacity, income, and credit enhance households’ ability to secure more land through leases or sharecropping, while higher formal education correlates with smaller holdings, likely reflecting historical allocation patterns and livelihood diversification outside agriculture. These results suggest that without targeted measures to address gender and generational inequities and to curb illicit land markets, the current tenure framework will continue to fall short in expanding fair farmland access.

This study adds empirical evidence from Arsi Zone that household farmland access under Ethiopia’s current land tenure remains limited and uneven. Average holdings are small (1.59 ha), with significant disparities by gender, age, and agro-zone. Institutional access pathways are dominated by government allocation (41.9%), inheritance (25.8%), and gifts (22.5%), yet clandestine sales occur under the guise of gifts. Most households (63%) perceive the current land tenure as inadequate for augmenting farmland access. Determinants of larger farmland holdings include oxen ownership, crop production, income, education (negatively associated in descriptive relationships), and credit access. The study recommends: empowering female-headed households; supporting younger households to expand access; enforcing legal tenure regulations and eliminating illicit land markets; considering land redistribution/reallocation in consultation with communities to address imbalances; and controlling rapid conversion of farmland to urban uses. Future research should employ longitudinal, multi-source designs to capture dynamics of land access, tenure security, and market evolution over time.

The study relied on a single-round household survey due to national and local security issues, time, and financial constraints, limiting longitudinal insights. The sensitive, policy-related nature of land access and clandestine transactions also constrains generalizability and may affect reporting accuracy.