The research team, led by Fenet Belay of Jimma University together with colleagues at Bayero University Kano and Jimma University, drew on two waves of the Ethiopia Socioeconomic Panel Survey, collected in 2018/19 and 2020/21 under the World Bank’s Living Standards Measurement Study–Integrated Surveys on Agriculture program. After merging the waves and cleaning the data, the final analytical sample comprised 1,400 agricultural households tracked over time. Rather than asking whether farmers used sustainable practices in a single year, the researchers classified each household into distinct trajectories: persistent adopters who used a practice in both waves, new adopters who began between the waves, and non-adopters who used it in neither. A fourth category, dis-adopters, proved too small for reliable statistical analysis, containing only about a dozen households.

The study focused on three practices with sharply different technical profiles. Crop rotation is primarily knowledge-intensive, demanding planning and agronomic management but little cash. Chemical fertilizer use is capital-intensive, requiring a lump-sum cash outlay at the start of each planting season along with functioning input markets. Soil and water conservation, which includes structures such as terraces and bunds, demands heavy labor investment and delivers its benefits gradually through improved soil fertility and moisture retention. This deliberate contrast allowed the researchers to test whether the forces driving adoption differ according to what a practice actually costs a farming household in money, muscle, or management.

The descriptive results reveal striking heterogeneity. More than half of households, 53.6 percent, were persistent adopters of crop rotation, a figure the authors attribute to the practice’s low financial cost and compatibility with traditional farming systems, though 27.17 percent never adopted it. Chemical fertilizer showed a different pattern: 45 percent of households used it in both waves, 21.7 percent were new adopters, and roughly a third, 33.2 percent, never used it at all, likely reflecting constraints on input cost, availability, and credit. Soil and water conservation presented the most transitional picture, with only 36.68 percent persistent adopters, but a remarkable 31.6 percent new adopters, a surge the authors link to public works programs such as food-for-work and watershed management campaigns.

To explain these trajectories, the team employed a pooled multinomial logit model, a statistical framework grounded in random utility maximization in which each household is assumed to select the adoption pathway offering the greatest perceived benefit. Because the dependent variable summarizes behavior across two periods rather than a single snapshot, the researchers used lagged explanatory variables from the 2018/19 baseline, ensuring that the factors predicting trajectories were measured before the outcomes occurred. Standard errors were clustered at the household level to account for correlation across survey waves. The authors are careful to note that the coefficients represent conditional associations rather than causal effects, an honest caveat in a field where unobserved traits such as farmer innovativeness and risk appetite can never be fully measured.

The econometric results deliver a clear message about money. Access to credit increased the probability of persistent adoption of crop rotation by 22.6 percentage points and of chemical fertilizer by 16.2 percentage points, while cutting the probability of non-adoption by 25.5 and 20.0 percentage points respectively. The most dramatic descriptive finding concerned new crop rotation adopters: 73 percent of them had credit access, compared with only 16 to 17 percent of persistent adopters, suggesting that liquidity is a primary catalyst for initiating rotational farming. This makes economic sense, since fertilizer in particular demands a large upfront payment precisely when cash is scarcest, and farmers who cannot borrow either skip the input or abandon it after a single season.

Extension services told an equally important but more puzzling story. Access to agricultural extension raised the probability of persistent adoption by 19.5 percentage points for crop rotation, 24.8 for fertilizer, and 16.0 for soil and water conservation. Yet in the raw data, non-adopters reported the highest rates of extension contact, reaching 80 percent for crop rotation and 75 percent for fertilizer. The authors suggest two possible explanations: extension agents may deliberately target non-users, or the content of existing advisory services may simply be failing to change behavior. Either way, the finding points to a potential misalignment between how extension is delivered and what farmers actually need to sustain new practices over multiple seasons.

Household resources shaped trajectories in ways that varied by practice. Livestock ownership, measured in tropical livestock units, was consistently higher among persistent adopters, averaging 4.09 units for crop rotation and 4.02 for fertilizer use, consistent with integrated crop-livestock systems easing input constraints. Labor availability emerged as decisive for soil and water conservation, where persistent adopters reported an average of 4.26 adult-equivalent workers, reflecting the heavy demands of building and maintaining terraces. Intriguingly, age worked in opposite directions: older farmers tended to persist with rotation and fertilizer, while younger farmers, averaging about 42 years versus 46 for non-adopters, were more likely to sustain conservation investments, plausibly because they have longer horizons over which to reap the gradual benefits. Wealth itself showed a non-linear pattern, with the poorest households persisting in conservation practices, likely because public employment programs targeted them, while wealthier households led fertilizer adoption.

Geography mattered too. Distance from roads significantly reduced persistent fertilizer adoption and increased non-adoption, underscoring how remoteness and transport costs strangle access to inputs. Non-adopters of crop rotation were located closer to major roads but nearer national borders, hinting that proximity alone does not guarantee adoption when markets, extension infrastructure, and services remain out of reach. The authors also found evidence of substitution rather than synergy in some contexts: manure use was negatively associated with persistent adoption of rotation and fertilizer, suggesting that resource-constrained farmers relying on organic inputs may feel little need for purchased ones.

The policy implications are direct. Because adoption is dynamic and practice-specific, the authors argue, programs must move beyond celebrating initial uptake and instead build the institutional scaffolding, reinforced extension services, accessible rural finance, and better market connectivity, that keeps practices in place year after year. Interventions should be tailored to the resource profile of each practice: credit for capital-intensive inputs, labor support and community programs for conservation works, and integrated packages where practices complement one another. In a country where agriculture anchors rural livelihoods and climate shocks are intensifying, the study’s central insight is quietly radical: resilience is not built in a single planting season, but in the patient accumulation of sustained, supported practice.

Subject of Research: Determinants of smallholder farmers' adoption trajectories of sustainable agricultural practices in Ethiopia

Article Title: Adoption trajectories of sustainable agricultural practices for climate resilience in Ethiopia

Article References: Belay, F., Barau, A. S., Gemechu, A., & Oljira, A. (2026). Adoption trajectories of sustainable agricultural practices for climate resilience in Ethiopia. Discover Agriculture, 4(1), Article 314. https://doi.org/10.1007/s44279-026-00790-2

Image Credits: AI Generated

DOI: 10.1007/s44279-026-00790-2

Keywords: sustainable agriculture, Ethiopia, smallholder farmers, adoption trajectories, climate resilience, crop rotation, chemical fertilizer, soil and water conservation, panel data, multinomial logit, credit access, extension services