The research team, led by Joshua Muloi of the Center for Development Research at the University of Bonn, conceptualized regenerative agriculture adoption not as a single decision but as a continuum of engagement. Working across two contrasting Kenyan counties—semi-arid Makueni in the southeast and the humid, commercially intensive Kirinyaga on the slopes of Mount Kenya—the researchers combined an exploratory qualitative phase with a large household survey. The qualitative work involved 33 individual farmer interviews, 14 focus group discussions, 9 expert interviews, and a stakeholder workshop with 196 participants, conducted in Swahili, Kamba, and Kikuyu languages between September and December 2024. These insights then informed the design of a survey administered to 949 farmers between August and September 2025.

Regenerative agriculture is organized around five core principles rather than a fixed package of technologies: minimizing soil disturbance, maximizing soil cover, increasing agrobiodiversity, maintaining living roots in the soil, and integrating livestock into cropping systems. Because farmers engage with combinations of interdependent practices rather than a single technology, the researchers argue that conventional binary measures collapse substantively different forms of engagement into one misleading distinction. A farmer experimenting with zai pits on a small plot, a farmer who delayed uptake for years after first hearing about the approach, and a farmer who abandoned the practices altogether may all be lumped together as non-adopters—or, conversely, as adopters—despite occupying fundamentally different positions.

To capture this heterogeneity, the team developed a seven-state typology based on four interconnected dimensions: awareness, uptake timing, continuity of use, and post-adoption behavioral adjustment. The states distinguish farmers who are not aware of regenerative agriculture, aware non-adopters, disadopters who previously practiced but stopped, trailing adopters who experienced delayed uptake, new adopters, plateau adopters who maintain practices without intensifying, and intensifying adopters who expand their engagement over time. The classification draws on social practice theory, which conceptualizes change as the reconfiguration of everyday practices through the interaction of materials, competences, and meanings rather than as a sequence of isolated individual decisions.

The results expose how much the binary lens conceals. A conventional classification would identify 49 percent of surveyed farmers as current implementers and 51 percent as non-implementers. But the seven-state breakdown reveals that within the non-implementer group, 33.7 percent had never heard of regenerative agriculture, 13.8 percent were aware but had not implemented it, and 3.5 percent had previously practiced and discontinued. Among implementers, 11.1 percent were trailing adopters with delayed uptake, 20.7 percent were new adopters, 2 percent had plateaued, and 15.3 percent were intensifying their engagement. The researchers emphasize that these states represent cross-sectional positions consistent with potential adoption pathways rather than directly observed transitions over time.

Qualitative evidence reinforced the quantitative patterns. Farmers described testing practices on small portions of their land before expanding, with one reporting that she started on one-eighth of an acre and subsequently expanded to three-quarters of her property. Another explained that each farmer picked what worked for their own farm from the range of practices they had been trained on. Discontinuation took varied forms: some farmers stopped individual practices that produced limited benefits or did not fit their conditions, while others withdrew from groups implementing the approach because of heavy labor requirements or unmet expectations of financial support. In the coffee-growing context of Kirinyaga, farmers described converting gradually to avoid disrupting existing production, for instance by grafting a limited number of coffee stems at a time rather than changing the entire farm at once.

Using a multinomial logit model with the not-aware group as the base outcome, the researchers examined which socioeconomic and institutional factors were associated with membership in each state. Credit access emerged as the most consistent correlate across engagement states beyond awareness. Farmers with access to credit had a 12-percentage-point higher predicted probability of being new adopters, while their probabilities of being not aware and aware non-adopters were 14.3 and 6.7 percentage points lower, respectively. Qualitative accounts provided a plausible mechanism: farmers described needing financial resources to hire labor, obtain manure, and acquire other inputs required for implementation. In Kirinyaga, some farmers accessed loans or advances through coffee cooperatives and tea factories connected to the Rainforest Alliance’s Mount Kenya Sustainable Landscape and Livelihoods programme.

Other factors showed more selective and sometimes counterintuitive associations. Land tenure security was negatively associated with being an aware non-adopter, disadopter, or trailing adopter relative to not being aware, but positively—though weakly—associated with intensifying adoption. The qualitative findings help explain why formal ownership should not be equated with operational control: participants described households where land was registered to men while women managed the farming, or where members cultivated separate portions. Marriage was positively associated with trailing and intensifying states, but the researchers interpret this as a marker of heterogeneous household decision and support arrangements rather than a single mechanism. Men were less likely than women to be classified as disadopters or new adopters relative to the not-aware category, while other gender comparisons were not statistically significant. Farm income was not significantly associated with any engagement state, and off-farm income was only weakly and negatively associated with intensifying adoption, suggesting that resource endowments alone are insufficient to explain engagement patterns.

The study identifies two critical transition gaps. First, a third of all farmers were unaware of regenerative agriculture, indicating an awareness gap that reflects limited knowledge diffusion among smallholders. Second, aware non-adopters, farmers with delayed uptake, and disadopters highlight a conversion gap: awareness alone does not resolve the practical constraints of labor requirements, limited income, perceived risks, and the need to observe convincing results before committing resources. The predominance of new adopters points to a recent wave of initial uptake, potentially linked to growing visibility of the approach through governmental and non-governmental programmes. Meanwhile, plateau engagement emerges as a distinct and relatively underexplored state in which farmers maintain practices without further intensification, for reasons the researchers hypothesize may include standardized training framings that limit continued experimentation, biophysical and production constraints, deliberate maintenance when benefits of expansion do not outweigh costs, or social identity dynamics that confer recognition without incentivizing further change.

The findings carry direct implications for how regenerative agriculture is promoted and financed. The researchers argue that extension systems and public-private advisory services should shift from one-off dissemination toward long-term accompaniment that supports iterative learning and progression across engagement states. Subsidies, grants, concessional credit, or in-kind support could help reduce the financial constraints that appear to bind most tightly during initial implementation, when farmers face upfront costs and delayed or uncertain returns. Programme monitoring systems could incorporate indicators of awareness, adoption timing, implementation intensity, and discontinuation to identify farmers requiring different forms of support. The seven-state typology, the team suggests, offers a potentially transferable framework for diagnosing where farmers sit along the engagement continuum and for designing state-contingent interventions that address both entry barriers and progression constraints—moving the field beyond the binary categories that have long obscured the differentiated reality of agricultural transitions among smallholders.

Subject of Research: Heterogeneous states of regenerative agriculture adoption among smallholders in Kenya

Article Title: Heterogeneous states of regenerative agriculture adoption among smallholders in Kenya

Article References: Muloi, J., Kamau, H., Betancur-Corredor, B., Kadenge, E., Rwanda, C., Wario, D. K., Stellmacher, T., Borgemeister, C., & Beuchelt, T. D. (2026). Heterogeneous states of regenerative agriculture adoption among smallholders in Kenya. Discover Sustainability, 7(1), Article 1737. https://doi.org/10.1007/s43621-026-05058-9

Image Credits: AI Generated

DOI: 10.1007/s43621-026-05058-9

Keywords: regenerative agriculture, smallholder farmers, Kenya, adoption pathways, credit access, mixed methods, multinomial logit, soil health, extension services, disadoption, agroecology, sustainability transitions

The research team, led by Stephen Prah and including John N. Ng’ombe, Bright Owusu Asante, and Jonathan Masasi, framed their analysis around a deceptively tricky statistical problem. Farmers who adopt climate-smart technologies are not a random sample of the farming population. They tend to be better educated, better connected to extension agents, and more likely to have access to credit. Any naive comparison between adopters and non-adopters therefore risks conflating the effects of the technologies themselves with the pre-existing advantages of the people who choose them. To untangle this, the authors deployed a dose-response model with Heckman selection correction, a two-stage econometric approach that explicitly models both the decision to adopt and the intensity of adoption, while correcting for the selection bias that would otherwise distort the estimates.

The dose-response framework is what makes this study unusual. Most adoption studies in agricultural economics treat technology use as a binary variable: a farmer either adopts or does not. But in the field, adoption is a continuum. One household might plant a small portion of its land with drought-tolerant maize while continuing business as usual elsewhere; a neighbor might restructure the entire farm around a full suite of climate-smart practices. By measuring adoption intensity, the share of recommended technologies and practices a household actually implements, the researchers could trace out how welfare outcomes change across the entire spectrum of engagement, from minimal uptake to comprehensive adoption.

The headline finding is a U-shaped relationship between adoption intensity and maize yield. At low levels of adoption, yields actually decline relative to non-adopter baselines, a pattern the authors attribute to the transition costs and learning curves that come with new practices. Partial adoption can mean partial disruption: new seeds planted without the complementary fertilizer schedules, or mulching introduced without adjusted tillage, may initially perform worse than the familiar system they replace. But as adoption intensity climbs past a threshold, the curve turns upward, and heavily adopting households pull decisively ahead. On average across the sample, higher adoption intensity was associated with a 21 percent increase in maize yield compared with what farmers would have achieved without the technologies.

Productivity, however, was only half of the story. The researchers also examined household asset accumulation, a slower-moving but arguably more meaningful indicator of long-term welfare. Assets, livestock, equipment, housing quality, and similar holdings, capture whether productivity gains translate into durable improvements in living standards rather than being consumed away season by season. Here again the relationship was nonlinear and positive: greater adoption intensity corresponded to an average 19.5 percent increase in household assets. The parallel results suggest that climate-smart farming is not merely boosting one season’s harvest but is reshaping the economic trajectory of adopting households over time.

Perhaps the most policy-relevant insight concerns heterogeneity. The effects of adoption were not uniform across the sample. Households differed in how much they benefited depending on their resource endowments, education levels, and exposure to climate stress. This matters because development programs often roll out technology packages with the implicit assumption that benefits will diffuse evenly. The Ghanaian evidence suggests otherwise: the same technology bundle that transforms one farm’s fortunes may deliver only marginal gains to another, and the distributional consequences of promotion campaigns deserve as much attention as their average effects.

The study also identified the factors that push farmers toward adoption in the first place. Education emerged as a consistent driver of both the likelihood and the intensity of adoption, reinforcing a well-established pattern in which human capital determines the capacity to absorb and implement complex agronomic information. Access to credit mattered for obvious reasons: many climate-smart inputs, from certified drought-tolerant seed to soil amendments, require upfront cash that cash-constrained households cannot easily mobilize. Contact with agricultural extension services was another strong predictor, since extension agents remain the primary conduit through which technical knowledge reaches rural communities. Finally, farmers who perceived drought as a significant and growing threat were more likely to invest in adaptive technologies, suggesting that climate risk awareness is itself a lever for change.

These determinants point toward a practical agenda. If adoption intensity is what drives the welfare gains, and if the drivers of intensity are education, credit, extension, and risk perception, then policy should aim squarely at deepening adoption rather than simply counting adopters. Programs that celebrate the number of farmers reached with a new seed variety may be missing the point: a farmer who plants 10 percent of her land with drought-tolerant maize may see no benefit at all, while one who reorganizes her whole operation around climate-smart principles reaps gains on the order of a fifth of her yield and assets. The difference between shallow and deep adoption is, on this evidence, the difference between wasted subsidy and transformative investment.

For Ghana, where maize is a staple crop and rainfed agriculture dominates, the stakes are high. Climate projections for West Africa point to rising temperatures, more erratic rainfall, and increased frequency of drought stress, all of which threaten the productivity of smallholder systems that already operate close to the margin. Climate-smart agriculture has been promoted by international initiatives as a triple-win strategy, raising productivity, enhancing resilience, and reducing emissions where possible, but the evidence base for its welfare effects has remained thin, particularly regarding how benefits scale with the depth of adoption. This study helps fill that gap with rigorous causal identification and a design that respects the reality of partial, incremental, and bundled technology use on real farms.

The broader lesson may resonate well beyond Ghana’s maize belt. As governments and donors channel billions into climate adaptation for agriculture, the question of dosage, how much adoption is enough, is becoming central to program design everywhere from the Sahel to South Asia. The Ghanaian evidence offers a clear if demanding answer: partial measures may disappoint, and the transition period may even hurt, but sustained, intensive adoption of climate-smart technologies pays off in both harvests and household wealth. Getting farmers to that threshold will require the unglamorous work of building credit systems, strengthening extension networks, and convincing rural households that a changing climate makes the investment worthwhile. The science of the technologies, in other words, is only half the equation; the economics of getting farmers fully on board is the other half, and it is where the battle for climate-resilient agriculture will be won or lost.

Subject of Research: Effects of climate-smart agricultural technology adoption intensity on maize yield and household welfare in Ghana

Article Title: How much climate-smart agriculture is enough? A dose-response analysis of technology adoption effects on farmer welfare in Ghana

Article References: Prah, S., Ng’ombe, J. N., Asante, B. O., & Masasi, J. (2026). How much climate-smart agriculture is enough? A dose-response analysis of technology adoption effects on farmer welfare in Ghana. PLOS Climate, 5(10), e0000886. https://doi.org/10.1371/journal.pclm.0000886

Image Credits: AI Generated

DOI: 10.1371/journal.pclm.0000886

Keywords: climate-smart agriculture, Ghana, maize, adoption intensity, dose-response analysis, farmer welfare, crop yield, household assets, extension services, credit access, drought resilience, agricultural economics

The study focused on two regions that could hardly be more different in the hazards they present. The Sava region in northeastern Madagascar lies squarely in the path of tropical cyclones that sweep in from the Indian Ocean, battering vanilla farms and rice paddies with destructive winds and flooding rains. Atsimo Andrefana in the southwest, by contrast, endures chronic drought, with erratic and declining rainfall slowly eroding the productivity of semi-arid croplands. By comparing these two contrasting hazard environments, the researchers were able to ask whether the relationship between exposure, perception, and preparedness follows a consistent logic, or whether local economic and social structures override the simple geography of risk.

To answer that question, the team surveyed 245 households selected through a stratified sampling design, ensuring that the sample captured the diversity of livelihoods and exposure conditions within each region. Rather than relying on any single measure, the researchers constructed three distinct analytical instruments. A composite Exposure Index quantified the physical hazards each household faced, drawing on both environmental data and household reports of past shocks. Perceived risk was measured through questions about the likelihood and severity of future climate events, capturing the psychological dimension of hazard. And preparedness was assessed through a multidimensional Shock Readiness Index that combined productive assets, access to credit, and financial buffers—the concrete resources a family could mobilize when disaster strikes.

The results revealed a pattern the authors describe as a decoupling of the three dimensions. In cyclone-prone Sava, households reported both the highest exposure and the highest perceived risk, yet scored lowest on preparedness. The reasons were structural rather than psychological. Many Sava households depend on specialized, crop-based livelihoods—particularly vanilla, a high-value but fragile cash crop whose fortunes rise and fall with both weather and volatile global prices. This specialization concentrates risk rather than spreading it. Compounding the problem, families in Sava faced limited access to credit and severe liquidity constraints, meaning that even when they understood the dangers, they lacked the financial machinery to build buffers against them.

Atsimo Andrefana told the opposite story. Households in the drought-prone southwest perceived less risk than their cyclone-exposed counterparts in the north, yet demonstrated comparatively higher levels of preparedness. The key difference lay in the structure of their livelihoods. Families in Atsimo Andrefana maintained diversified income streams and, crucially, held wealth in the form of livestock—zebu cattle and goats that function as convertible assets, a walking savings account that can be liquidated quickly when crops fail. This finding carries a provocative implication: preparedness in smallholder systems may depend less on how frightened people are and more on whether their economic lives are organized in ways that allow them to act.

Perhaps the most policy-relevant result concerns early warning systems. Across both regions, access to early warning information showed only a limited association with actual preparedness. In other words, knowing that a cyclone is coming or that a drought is likely does relatively little good when a household has no savings to draw on, no credit to access, and no way to protect or relocate its assets. The authors interpret this as evidence that information alone has limited influence where response options are constrained by structural poverty. This finding resonates with a growing body of international research on the risk perception paradox, which has repeatedly found that heightened awareness of hazards does not automatically produce protective behavior, particularly among populations with the fewest resources.

The methodological contribution of the study is as significant as its substantive findings. By integrating a physical exposure measure, a psychological perception measure, and a structural readiness measure within a single household-level framework, the researchers created what they call an exposure-readiness typology. This diagnostic tool classifies households according to their combined risk profiles, allowing practitioners and policymakers to distinguish, for example, between households that are highly exposed but unprepared and those that are moderately exposed but resilient. Such differentiation matters because the interventions appropriate for each group differ profoundly: the former may need emergency financial instruments and asset protection, while the latter may need little more than maintenance of existing diversification strategies.

Madagascar provides an especially compelling setting for this research. The island is routinely ranked among the countries most vulnerable to climate change, with its smallholder farmers facing extreme exposure to agricultural risks ranging from cyclones and floods to droughts and locust outbreaks. Previous studies have documented social-ecological traps in southwestern Madagascar that hinder rural development, and recent work has explored how climate perceptions among northeastern farmers shape adaptive behavior. The new study builds on this literature but departs from it by refusing to treat perception as the central variable. Instead, it places structural capacity—the assets, credit, and liquidity that determine what a household can actually do—at the center of the analysis, and treats perception as one factor among several that may or may not align with it.

The implications extend well beyond Madagascar’s borders. Across tropical smallholder regions, from the Sahel to South Asia, hundreds of millions of farming families face intensifying climate variability with thin margins for error. If preparedness does not systematically track exposure in these systems, then adaptation strategies built primarily on risk communication and awareness campaigns may systematically miss their targets. The study’s authors argue that strengthening structural capacity through improved financial access, livelihood diversification, and institutional support will be critical to narrowing preparedness gaps—not only in Madagascar but in climate-exposed rural systems worldwide. In practice, this could mean expanding rural credit markets, developing insurance products suited to smallholder agriculture, and supporting the livestock and off-farm income streams that give households options when harvests fail.

The research also carries a caution for the rapidly expanding global investment in early warning infrastructure. The United Nations has championed multi-hazard early warning systems as a cornerstone of disaster risk reduction, and such systems undeniably save lives. But the Madagascar findings suggest that warnings are only one link in a chain that must also include the financial and material means to respond. A family that receives a cyclone warning but has no savings, no insurance, and no safe place to store its assets can do little more than brace itself. As climate hazards intensify across the developing world, the study’s central lesson is likely to grow only more urgent: the households that need resilience most are often those least equipped to build it, and closing that gap will require changing economic structures, not just changing minds.

Subject of Research: The relationship between climate hazard exposure, risk perception, and household preparedness in smallholder farming systems in rural Madagascar

Article Title: Decoupling exposure, risk perception, and preparedness in smallholder systems: evidence from rural Madagascar

Article References: Joseph, J. E., Umutoni, C., Raharison, J., Whitbread, A. M., & Wane, A. (2026). Decoupling exposure, risk perception, and preparedness in smallholder systems: evidence from rural Madagascar. Regional Environmental Change, 26(4), Article 214. https://doi.org/10.1007/s10113-026-02707-7

Image Credits: AI Generated

DOI: 10.1007/s10113-026-02707-7

Keywords: climate risk, Madagascar, smallholder farmers, risk perception, preparedness, early warning systems, livelihood diversification, Shock Readiness Index, cyclones, drought, credit access, climate adaptation

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

Economists Ayodeji Damilola Kehinde of Obafemi Awolowo University and Abiodun Akintunde Ogundeji of the University of the Free State surveyed 240 smallholder crop farmers across four local government areas in Osun State, using a multistage sampling design drawn from registered farmer databases maintained by the Agricultural Development Programme and the Nigerian Agricultural Insurance Corporation. Their analytical weapon of choice was Cragg’s double hurdle model, an econometric framework first proposed in 1971 that deliberately breaks a single economic decision into two sequential stages. The first hurdle asks whether a farmer is willing to adopt insurance at all; the second, estimated only among those who pass the first, asks how much money that farmer is actually prepared to pay. This distinction matters because conventional single-equation models, such as the Tobit model, assume that the same underlying mechanism governs both the decision to participate and the intensity of payment, an assumption the new evidence decisively contradicts.

The portrait of the average respondent is striking. The typical farmer in the sample was about 50 years old, had accumulated 32 years of farming experience, and cultivated just 1.65 hectares of land. Nearly 91 percent were married, with an average household of five people, and three-quarters had received formal education. But the institutional connections that typically lubricate access to financial products were conspicuously thin: only 31 percent belonged to cooperative societies, a mere 17.5 percent had any contact with agricultural extension agents, and just 12.5 percent reported access to formal credit. Most tellingly, only 34 percent of respondents were even aware that agricultural insurance existed as an option.

That awareness figure may be the study’s most consequential finding. From the perspective of innovation diffusion theory, awareness is the first and most critical stage in any adoption process; farmers who do not know a product exists cannot evaluate its benefits or incorporate it into their farm management decisions. The authors attribute the information gap to weak extension services, inadequate communication channels, and limited interaction between farmers and insurance providers. Compounding the problem, most farmers in the sample exhibited low risk perception, with roughly three-quarters falling into the low-risk-awareness category. The researchers invoke behavioral risk theory to explain this pattern: people tend to underestimate low-frequency, high-impact events, particularly when they have not recently experienced severe losses. Years of successfully managing shocks through informal coping strategies, from family support networks to community risk-sharing arrangements, may breed an optimism bias that reduces the perceived need for formal protection.

When the researchers ran the first hurdle of their double hurdle model, a probit regression estimating the latent propensity to adopt, a distinctive set of factors emerged. Age and farming experience both carried negative and statistically significant coefficients, meaning older and more seasoned farmers were less likely to express willingness to adopt insurance. The authors suggest this may reflect deep familiarity with traditional risk-management practices rather than lower exposure to risk. Marital status, farm size, and access to credit all pushed in the opposite direction, increasing the probability of willingness to adopt. Married farmers face greater household financial responsibilities and therefore stronger incentives to protect farm income, while larger farms place more land, labor, and inputs at risk, raising the potential financial consequences of a single bad season. Access to credit relaxes the liquidity constraints that often make upfront premium payments impossible.

One result stands out as genuinely counterintuitive: years of formal education had a negative and significant effect on the willingness to adopt. This finding runs against a substantial body of evidence from Rwanda, Ethiopia, and elsewhere in Nigeria showing education positively associated with insurance uptake. The authors offer a subtle interpretation. More educated farmers may be better equipped to scrutinize premiums, coverage conditions, exclusions, and claim procedures, and where available products are perceived as costly, opaque, or ineffective, that critical capacity may translate into greater selectivity rather than greater demand. Crucially, the picture flips in the second hurdle: among farmers who had already expressed willingness to participate, education significantly increased the amount they were prepared to pay. Education, in other words, appears to operate differently across the two decision margins, making farmers more skeptical at the entry gate but more generous once inside.

The second hurdle, estimated with a truncated regression on the subset of willing farmers, identified five significant drivers of payment intensity: education, farm size, cooperative membership, risk perception, and insurance awareness. The cooperative effect is particularly interesting from a policy standpoint. Membership in farmer associations improves access to information through meetings, peer interactions, and collective learning, and it allows farmers to learn from the experiences of others. Evidence from China has shown that social networks can substantially increase demand for weather insurance by facilitating the transmission of insurance-related information, and the Osun State results are consistent with that mechanism. Risk perception and insurance awareness both increased the intensity of willingness to pay, underscoring that subjective assessments of danger and basic product knowledge, not just objective exposure, shape how much financial commitment farmers will make.

Diagnostic tests lend the findings credibility. Variance inflation factors averaged 2.41 and 2.38 for the first and second hurdles respectively, well below the conventional threshold of 10, and Breusch-Pagan tests found no evidence of heteroskedasticity in either equation. Both hurdle equations were jointly significant, and the authors were careful to frame the estimated coefficients as conditional associations rather than definitive causal effects, noting that variables such as credit access and insurance awareness could plausibly be correlated with unobserved farmer characteristics like intrinsic risk orientation. The study’s scope is also bounded: with only four of Osun State’s thirty local government areas sampled, the authors caution that findings should be interpreted within the study area rather than generalized to all Nigerian farming households.

The policy implications, however, travel well beyond southwestern Nigeria. Because the drivers of market entry differ from the drivers of payment intensity, insurance programs that treat demand as a single decision risk designing products that miss their target. The authors recommend bundling insurance with subsidized input packages or formal credit facilities so that premiums are covered upfront through loan disbursements, eliminating immediate liquidity constraints. They also point to mobile technology as a game changer: parametric weather-index insurance delivered through USSD and SMS platforms, with satellite remote-sensing data triggering automated payouts, could eliminate costly field inspections, accelerate claim settlement, and build the institutional trust that currently discourages participation. Cooperatives, they argue, should serve as primary distribution channels, aggregating smallholders to lower transaction costs and enable group-discounted premiums.

For a region where climate change is intensifying production risks year after year, the study’s central message is both sobering and actionable. The latent demand for agricultural insurance among Nigerian smallholders is real, but it is gated by two distinct barriers: structural constraints like credit access and farm scale that determine who enters the market, and informational and behavioral factors like risk perception and policy literacy that determine how much those entrants will invest. Unlocking the first barrier requires financial engineering; unlocking the second requires education, extension, and trust. Only by clearing both hurdles, the researchers conclude, can agricultural insurance fulfill its promise of transforming rural vulnerability into resilience.

Subject of Research: Determinants of smallholder farmers' willingness to adopt and pay for agricultural insurance in Osun State, Nigeria

Article Title: Investigating smallholder farmers’ willingness to adopt and pay for agricultural insurance policies in Osun State, Nigeria: a cragg’s double hurdle approach

Article References: Kehinde, A. D., & Ogundeji, A. A. (2026). Investigating smallholder farmers’ willingness to adopt and pay for agricultural insurance policies in Osun State, Nigeria: a cragg’s double hurdle approach. SN Social Sciences, 6(10), Article 489. https://doi.org/10.1007/s43545-026-01779-w

Image Credits: AI Generated

DOI: 10.1007/s43545-026-01779-w

Keywords: agricultural insurance, smallholder farmers, Nigeria, Cragg's double hurdle model, willingness to pay, risk perception, climate risk, cooperative membership, credit access, Osun State, crop insurance, rural finance