i PhD Dissertation Research Proposal Name of the candidate: Mohammed Ebrahim Mohammed Title of Dissertation1: Inclusive Agricultural Insurance for Sustainable Wheat Intensification as a Pathway to Smallholder Resilience in Ethiopia. Department: College of Agriculture and Environmental Science Thesis Supervisors: Thesis Director: Professor. Bruno Gerard Thesis Co-Director: Dr. Wuletawu Abera Thesis Co-Director: Dr. Lulseged Tamene Thesis Co-Director: Dr. Thomas Assefa Academic Year: 2025 Date: October 26-2025 ii Executive Summary This study addresses the pressing need for inclusive and scalable agricultural insurance solutions for smallholder wheat farmers in Ethiopia, who face persistent yield risks across diverse agroecologies and farming systems. Despite the proven benefits of sustainable intensification (SI), adoption remains low due to risk exposure, financial constraints, and limited access to insurance. Existing area-based index insurance models often fail to reflect localized realities, resulting in high basis risk and poor uptake. To bridge this gap, this research will develop a dynamic farm-level and area yield index insurance model integrating sustainable intensification (SI) practices and risk- based farm typologies. The model will combine remote sensing, geospatial, and ground-truth agronomic data through machine learning and simulation to enable accurate yield prediction and premium estimation. Once calibrated, it will function with minimal inputs like NDVI, weather data, and location ensuring cost-effective, scalable, and timely payouts. The research will also evaluate the risk-reducing effects of SI, estimate SI-sensitive premiums, and assess adoption drivers and farmers’ willingness to pay to ensure alignment with smallholders’ needs. Beyond compensating losses, the study envisions insurance as a driver of technology adoption and farm investment. Developing such holistic tools can enhance resilience and environmental sustainability. Keywords: Sustainable Intensification, yield index-insurance, Premium estimation, Remote Sensing, Machine Learning, Wheat Smallholders iii Table of Contents Chapter 1: Introduction ............................................................................................................................. 1 1.1. Background of the Study .......................................................................................................... 1 1.2. Problem statement ..................................................................................................................... 3 1.3. Research Objectives .................................................................................................................. 6 1.3.1. General Objective: ................................................................................................................ 6 1.3.2. Specific Objectives: ........................................................................................................... 6 1.4. Research Questions and Hypothesis ........................................................................................ 6 1.4.1. Research Questions ............................................................................................................... 6 1.4.2. Research Hypothesis and Testing Approaches.................................................................... 6 1.5. Scope and Limitations of the Study ......................................................................................... 7 1.6. Significance of the Study ........................................................................................................... 8 Chapter 2. Literature Review .................................................................................................................. 10 2.1. Empirical Evidence on Sustainable Intensification In sub-Saharan Africa. ............................ 10 2.1.1. Nutrient Use Efficiency and Precision ................................................................................... 11 2.1.2. Soil Health and Nutrient Management ................................................................................. 12 2.1.3. Diversification and Integration of Crops .............................................................................. 13 2.2. Crop Insurance as Catalyst for Improved Practices and Technology Adoption ...................... 14 2.3. Trends Toward Context-Specific Agricultural Insurance in Ethiopia....................................... 15 2.4. Constraints and Drivers of Crop Insurance Uptake ................................................................... 17 2.4.1. Economic and Institutional Challenges in Crop Insurance Adoption ................................ 17 2.4.2. Behavioral Economics in Smallholder Crop Insurance Uptake ......................................... 18 2.4.3. Systematic Review of Key Challenges in Agricultural Insurance ....................................... 22 2.5. Crop Insurance and Research Gaps ............................................................................................. 23 2.5.1. Methodological Gap in Agricultural Insurance and SI Research ....................................... 24 3. Conceptual Framework ........................................................................................................................ 26 3.1. Sustainable Intensification: Concept and Implementation Pathways ....................................... 26 3.2. The Linking Between Agricultural Insurance and Sustainable Intensification ....................... 27 3.2.1. Sustainable Intensification and Risk-Typology for Inclusive Crop Insurance ...................... 30 4. Research Methodology ......................................................................................................................... 34 4.1. Study Areas Description ................................................................................................................ 34 4.2. Research Design and Sampling ..................................................................................................... 36 4.3. Types and Source of Data. ............................................................................................................. 39 iv 4.4. Methods of Data Analysis .............................................................................................................. 40 4.4.1. Descriptive Statistics ............................................................................................................... 40 4.4.2. Crop Insurance Adoption and Farmers’ Willingness to Pay (WTP) Analysis ................... 40 4.4.3. Machine Learning Model for Farm Level Yield Index and Risk-Typology Sensitive Area based Yield Index insurance Design. ............................................................................................... 44 4.4.4. Analysis of the Risk Reduction Effects of SI Practices and Premium Estimation ............ 45 4.4.5. Analysis of Risk-Typology–Based Area Yield Index Insurance (TAYII) ............................ 48 4.5. Ethical Considerations (Data Privacy, Storage, and Participant Protection)........................... 53 5. Work Plan .......................................................................................................................................... 55 6. Budget Breakdown ............................................................................................................................ 56 7. Anticipated Results, Relevance, and Communication Plan .......................................................... 57 7.1. Expected Results ............................................................................................................................ 57 7.2. Contribution to Scientific Knowledge ......................................................................................... 58 7.3. Policy, Practice, and Research Implications ................................................................................ 58 7.4. Research Dissemination Strategy ................................................................................................. 58 Annex A. Figures from Text Frequency Analysis of 810 Papers ............................................................. a Annex B. Agricultural Insurance Systematic Review Approaches ......................................................... b Anex C. Supervisors’ Approval Sheet for PhD Dissertation Research Proposal ................................... d Annex D. Disclosing Use of AI.................................................................................................................... e v List of Figures Figure 1. Conceptual framework illustrating pathways to inclusivity and uptake of crop insurance …………………………31 Figure 2. Study Site Map .......................................................................................................................... 36 Figure 3. Frequency Of Emerging Thematic Mentions in Agricultural Insurance Abstracts ............. a Figure 4. The Top 20 Keywords Plus in Agricultural Insurance Research ............................................ a Figure 5. Word cloud of specific Agricultural Insurance Challenges ..................................................... b Figure 6. Flow Diagram of the Publication Screening and Selection Process ....................................... c 1 Chapter 1: Introduction 1.1. Background of the Study Agriculture faces a growing challenge from climate change and extreme weather, including unpredictable changes in temperature and rainfall, more frequent floods and droughts (Chandio et al., 2020). Global surface temperature has increased by 1.1°C between 2011 and 2020. It has contributed to desertification and exacerbated land degradation, slowing agricultural productivity over the past 50 years (IPCC, 2023). Climate change directly affects crops productivity and indirectly affects them by increasing the occurrence of pest and disease. Growing frequency of extreme weather and climate events has exposed millions of people to the prospect of tragic crop failure or livestock mortality (Mahul & Stutley, 2010) and causing a substantial impact on food security for many countries (Han et al., 2023a). These challenges are more pronounced in regions like Africa and more negative impact to small-scale food producers, low-income households, and local peoples around the world (IPCC, 2023). Globally, small farms of less than 20 hectare produce 75 % of the world food, while very small farms under 2 hectares are contributing around 30% of most food commodities (Fanzo, 2017). These smallholder farmers agriculture in developing countries have been threatening by land degradation and land garbing which increased vulnerability to environmental and disaster risks (UNCCD, 2017). In Sub-Saharan Africa, 330 million ha are affected by human-induced degradation, and soil erosion, nutrient depletion and salinity leading decline agricultural productivity (FAO, 2021). In the region, about 70-80% farms are smaller than 2 ha (Lowder et al., 2016). These lands continued further fragmentation following the land inheritance practice (Aslam & Fazal, 2025) and leads to economical unsustainable to support the family food security by increasing production costs and reducing efficiency (Ntihinyurwa & de Vries, 2021). Africa remains the region with the largest estimated proportion of the population facing hunger 20.4 percent. It affected almost 300 million people in Africa, in 2023 (FAO, 2024b). In Ethiopia, the number reached close to 15.8 million (FAO, 2024). This food insecurity problems driven by declining productivity of soils, lack of access to sufficient inputs (seed and fertilizer), improved agricultural technologies and climate information (Mapfumo et al., 2013). In drought prone region of sub–Saharan Africa (SSA) drought causes high risk of crop failure and low adoption of high- 2 yielding, input-responsive crop varieties (Amede et al., 2023) and worsen the food security situation. The severity and frequency of drought will worsen in the future, with more negative consequences for rural community (Abdela, 2024). Climate change and traditionally agricultural intensification leads to land degradation, vulnerability to soils to erosion and nutrient loss consequently, diminishing agricultural yields and makes the agricultural more challenging (Mapfumo et al., 2013; Jia et al., 2022; Yang et al., 2024). Initiative, particularly the green revolution has emphasized the intensive use of agrochemicals, energy, land and water (Ickowitz et al., 2019a). Such overuse of natural resources focusing on increasing yields per unit of land causes negative environmental effects (Dobbs & Pretty, 2004). The growing awareness on these environmental costs, leads to emphasizes the need for the most efficient agricultural system and alternative agricultural practice to reduce the negative externality, focusing on three pillars: producing more food per unit of land, enhancing resilience to shocks and stresses while preserving ecosystem services (Pretty et al., 2011a). In this context sustainable intensification (SI) is an effective agricultural pathway to achieve green growth through increasing yields per unit area of land without negatively affecting the environment (Ickowitz et al., 2019c). SI promotes adaptation to climate change through resilient agricultural models that can address environmental challenges (Ajibade et al., 2023). However, despite the SI received growing attention by governments and international institutions for its multi-functional (sustainability, intensification, adapted approaches to local context and enhancing farmers’ livelihoods) advantages (S. Cook et al., 2015), improved practices and technologies have been implemented in a poorly integrated and less synergized manner resulting in low level of adoption of SI (Giller et al., 2009). Effective SI implementation requires tailored practices to the local conditions and simultaneous adoption of improved practices and technologies (Giller et al., 2009). More specifically, proper implementation of SI requires, the use of integrated agricultural inputs such as organic and inorganic fertilizer, high-yielding varieties, mechanization, integrated disease management, soil and water conservation practice (Pretty et al., 2011a), site-specific agro-advisory and climate information (Amede et al., 2023). These intensification practice needs heavy investment which often inaccessible and challenging for smallholder farmers particularly in Ethiopian and in Africa at large. Because of fear of crop failure, smallholder farmers often maintain risk-averse 3 traditional practices and stick with low-input, low-yield and low-risk practices. Without efficient financial support and derisking mechanism, the success of SI implementation could be unrealistic (Simutowe et al., 2024). To this end, agricultural insurance plays a crucial role in protecting farmers against production risks and losses, while also building their confidence to adopt climate-smart technologies and improved practices that enhance productivity, strengthen resilience, and reduce negative environmental impacts (Badani et al., 2020). Agricultural Insurance also helps microfinance and Banks to gain more confidence in extending credit to insured farmers as it reduces the risk of loan default and assure the repayment (Meyer et al., 2017). The participation of farmers in the insurance scheme largely depends on the principle of "fairness" and "affordability", which aim to create a more balanced system with minimal trade-offs between insurer profit and premium levels (Biagini & Severini, 2021). A more sustainable insurance scheme will be realized when the insurance system promotes the use of SI practice and improved technology that enhance farmers’ productivity by reducing risks and in turn, lower the insurer’s indemnity burden. In general, agriculture insurance not only protecting smallholder farmers from objective risks, but also serve as a gateway to access bundled services and supports to enhance farm productivity, resilience and sustainability. 1.2. Problem statement In Ethiopia wheat one of the most important crops for the national food security and economy of the country. Due to its strategic importance, wheat has become a priority crop for government programs and its area coverage increasing over time; for instance, the area coverage expands from 1.6 million 2013 hectare in 2013 to 1.86 million hectare in 2023 (FAOSTAT). In recent years, wheat has been expanding from the highland to the vast lowland areas with the support of irrigation (Y. Kuma et al., 2025). Significant research efforts have been made, ranging from wheat varietal development to site and context specific agronomic solutions aiming to promote SI of agriculture (Abera et al., 2022; Liben et al., 2024). Despite this progress, wheat production still faces several challenges such as disease and pest, land degradation, rainfall variability and drought, limited access to inputs and market. The large-scale adoption of those context specific agronomic solutions also remains challenging mainly because of high input cost, lack of supportive policies and farmers economic limitation (Tamene et al., 2017). 4 Small holder farmers in Ethiopia are producing in a full of risky and uncertain environment. This risk limits their investment on production inputs, further hindering farm productivity (Tamene et al., 2017). In the face of such climate-induced vulnerabilities and risks, crop insurance emerges as one of the best solutions for smallholder farmers. By providing a financial safety net against unpredictable risks like extreme weather events, crop diseases, and pests, insurance can significantly enhance their resilience and safeguard their livelihoods (Ankrah et al., 2021; Madaki et al., 2023; Tamene et al., 2017). However, the different types of crop insurance such as index- based, yield-based, and multi-peril insurance products are primarily focused on protection of farmers from financial loss arises from the objective risk (Sushchenko & Schwarze, 2021). Insurance products focused solely on farm-loss compensation are inadequate for the complex challenges of smallholder farmers and cannot sustain a viable agricultural insurance market. Insurance works best when embedded in comprehensive risk management combining prevention, adaptation, mitigation, and risk transfer (SDC, 2014; Ceballos et al., 2025). Despite decades of promotion, agricultural insurance products have limited adoption and less impacts on Ethiopian small holder farmers livelihood resilience (Kramer et al., 2022). This is mainly because of unaffordable premium price, low awareness (Blackmore et al., 2025), and limited access to finance further prevents farmers from investing on improved inputs and, poor data quality and infrastructure that limit its effectiveness for smallholder farmers (Mahul & Stutley, 2010). The potential benefits of index-based insurance largely depend on a strong correlation between the index-triggered indemnities and the actual losses experienced by farmers (Miranda & Farrin, 2012). However, the basis risk (discrepancy between index-based insurance and actual loss) often remains a critical challenge by eroding smallholders’ farmers trust and limiting their uptake of insurance (Porter et al., 2023). Basis risk arises mainly from two sources: heterogeneity within insurance zones (zonal risk) and low predictive accuracy of the index (Stigler & Lobell, 2023). As a result, crop insurance uptake remains limited and is often accompanied by high dissatisfaction with claim settlements (Njue et al., 2018). The existing agricultural insurance product in Ethiopia often lack flexibility and customization to the local contexts. As a result, the one-size fit all approach fails to meet the diverse risk preference and needs of farmers (Blackmore et al., 2025). While insurance can facilitate access to credit and inputs that support SI practices, most programs do not explicitly link insurance with improved 5 practices, technologies, or other risk-reducing measures (Adegoke et al., 2017). Scholars recommend bundling agricultural insurance with SI to enhance farm investment and productivity. However, this approach remains largely conceptual, as the effects of SI interventions on loss reduction and their influence on premium pricing are not yet well explored (Simutowe et al., 2024; Badani et al., 2020 ; Meyer et al., 2017). This lack of integration is particularly evident in the Ethiopian context, where many adoption studies have examined improved agricultural practices and crop insurance in isolation (Wassie et al., 2024; Teklewold et al., 2013; Abegaz et al., 2024; Bekuma, 2024; Eze et al., 2020), revealing a notable evidence gap on their combined role and potential synergy in enhancing adoption and smallholder resilience. These challenges emphasize the critical need for lower premiums, advanced data systems, and the development of more effective, inclusive, and innovative insurance products that empower smallholder farmers while ensuring insurer profitability. In this context, high-resolution index insurance at the farm, community, or recommendation domain level enhances loss estimation accuracy and unlocks the use of site-specific agronomic solutions and SI practices. In addition to spatial refinement, the insurance products should also promote bundling with various technologies and agro-advisories that go beyond loss compensation (Hellin et al., 2017; Choruma et al., 2024). In pursuit of this goal, the risk-reduction effects of SI practice need to be examined to design dynamic and sustainable crop-insurance products based on the level of bundled integrated technologies and practices (Simutowe et al., 2024). This research aims to model farm- level yield index insurance premium pricing based on risk-reduction strategies, specifically bundles of improved agricultural practices (IAP) as a pathway to sustainable intensification (SI), and to design area yield index insurance differentiated by levels of risk. It also seeks to examine how SI adoption influences farmers’ willingness to pay for crop insurance an area where evidence remains limited. 6 1.3. Research Objectives 1.3.1. General Objective: To model and empirically evaluate how integrating sustainable intensification practices, remote sensing, geospatial data and farm typologies can improve the design and performance of crop- insurance systems for Ethiopia’s wheat smallholders. 1.3.2. Specific Objectives: 1. To determine how behavioral, economic, and biophysical factors influence crop-insurance adoption and smallholders’ willingness to pay (WTP) in Ethiopian wheat farming systems. 2. To evaluate the risk-reducing effect of sustainable-intensification practice on wheat yield and their role in minimizing basis risk and improving premium targeting. 3. To design and evaluate a dynamic, typology-sensitive area-based yield-index insurance model across diverse wheat farming systems in Ethiopia. 1.4. Research Questions and Hypothesis 1.4.1. Research Questions 1. What are the factors influencing adoption of crop insurance and willingness to pay among wheat-growing smallholder farmers in Ethiopia? 2. What types of risks do wheat smallholder farmers face across different farming systems? 3. To what extent do SI practices contribute to minimizing basis risk and improving premium targeting in farm level yield index insurance schemes? 4. How risk-typology based area yield index insurance improve premium affordability and fairness? 1.4.2. Research Hypothesis and Testing Approaches Hypothesis Testing Approaches Hypothesis 1: The level of sustainable intensification (SI) practice significantly influences both crop insurance adoption and the willingness to pay for it. Check the significance of the SI composite index in DBH model 7 Hypothesis 2: Farmers with higher perceived production risks are more likely to demand for crop insurance. Check the significance of risk perception variables in DBH model Hypothesis 3: Higher level of Sustainable Intensification (SI) adoption lowers basis risk and improves the affordability of farm-level yield-index crop insurance. Compare Basis risk and premium affordability across 3 level of SI index (higher, medium, lower) using ANOVA. Hypothesis 4: A risk-typology–based area-yield index insurance design significantly improves premium affordability and pricing fairness? ANOVA will be used to check mean Loss ratio difference and premium to net income ratio across typology groups? (permutation ANOVA) 1.5. Scope and Limitations of the Study This study will proceed through a comprehensive sequence. It will begin by diagnosing constraints to crop-insurance adoption and assessing farmers’ risk profiles, then identify bundles of improved agricultural practices and extract geospatial and remote-sensing indicators from GPS-referenced plot data, and ultimately develop a dynamic insurance-premium product tailored to varying levels of Sustainable Intensification (SI) integration and risk profiles. This integrated approach is designed to address key challenges in crop insurance that are critical for enhancing wheat productivity. Farmers’ risk profiles, along with the economic and behavioral constraints shaping crop-insurance adoption and preferences, will be systematically assessed. Data from household surveys, focus group discussions (FGDs), and key informant interviews will be collected in six contracting districts of Ethiopia to analyze adoption constraints and farmers’ willingness to pay, while also examining insurance implementation and the availability and use of extension services. In parallel, context-specific improved agricultural practices, advisory information, historical and in-season wheat yield data, and geospatial and remote-sensing covariates will be gathered from the same six districts to evaluate the risk-reduction effects of SI and to inform the development of a farm-level dynamic crop-insurance premium. 8 Agronomic practices, wheat grain yield data, and geospatial and remote-sensing indices will also be systematically derived from GPS-referenced crop-cut measurements and household survey data collected across the 13 districts included in the wheat-yield gap study. Finally, risk-based farm typologies will be constructed, and a dynamic, typology-based area-yield index insurance product will be developed. Because of the logistical issues and time, the new adoption study will be geographically limited to six districts of three wheat growing region of Ethiopia. As result, these selected sites may not fully capture the heterogeneity across agroecology and farming system. Despite the robust household sample size and broad geographic coverage of the GPS-referenced crop-cut data, relying on yield measurements from a single production season may be insufficient to fully capture the range of yield variability caused by climatic fluctuations and seasonal differences. Consequently, this limits the study’s ability to capture inter-annual yield variability and its applicability across different years and climatic conditions. 1.6. Significance of the Study The research study will help to develop tailored, dynamic insurance products based on the specific farming realities, risks, and preferences of smallholder farmers. This approach is not merely protecting farmers from unexpected loss, but also enabling them to implement bundles of technologies, advisories, and improved practices that enhance productivity while reducing insurers’ indemnity payouts. The study will also help researchers and agricultural extension workers better understand the diverse risk profiles and adoption behaviors across different farming systems and farm typologies. This evidence will support more effective advisory services and technology promotion strategies. Furthermore, the study will provide evidence-based guidance for scaling sustainable intensification (SI) intervention options across diverse farming contexts. This, in turn, will enable policymakers and development partners to design more context-specific interventions that reflect farmers’ real challenges and priorities. For insurers, the study will offer a dynamic premium-making model that incorporates varying levels of technology adoption, risk and risk reduction strategies, while integrating data-driven 9 approaches for accurate and fair premium setting. These innovative and inclusive insurance approaches linking SI, advisory services, and insurance will promote resilience and productivity beyond simple loss compensation. Ultimately, this research will foster collaboration among insurers and key partners by providing concrete evidence and recommendations on how integrated, innovative, and bundled crop insurance and SI practices that can enhance insurance uptake, reduce crop loss risks, and ensure affordable and sustainable insurance systems. 10 Chapter 2. Literature Review 2.1. Empirical Evidence on Sustainable Intensification In sub-Saharan Africa. Sustainable Intensification is an efficient agricultural system that minimize environmental harm, increase food production per land unit, and enhance resilience while protecting ecosystems (Pretty et al., 2011b). Even though there is no clear-cut pathway for SI, experts agreed on the need to integrate crop productivity, profitability and preserving and enhancing ecosystem services (Pretty et al., 2011b; Vanlauwe et al., 2014). The integration of crop and natural resources plays a significant role in SI. The application of crop residue in the soil improves fertility and water retention and contributes to sustainability (Brandt et al., 2018; Sajjad et al., 2019; Sarkar et al., 2020). Various authors highlighted that improved sustainable agricultural practices such as agroforestry, intercropping, and conservation agriculture, along with techniques, such as precision farming, crop intensification, and diversification, improve soil health with symbiotic relation with microorganism, lower greenhouse gas emission, and reduce the effect of climate change (Ameur et al., 2020; Ayantunde et al., 2020; Jhariya et al., 2021; Kuyah et al., 2021a; Lal, 2020; Lemaire et al., 2023; Ntamwira et al., 2023; Suman et al., 2022; Swastika et al., 2024). The main aim of SI is to addresses questions related to low-risk production for consumption, marketing, profitability, crop diversity, agronomic practice for soil cover and nutrient cycling, legume nitrogen fixation, amount of external input to be applied, crop health, water use and mechanization (Vanlauwe & Dobermann, 2020). According to (2011) sustainable agriculture employs a combination of practices for maximum synergies and reduced environmental impacts. These practices include utilizing high-productivity crop varieties and livestock breeds, avoiding unnecessary external inputs, the use of nutrient cycling, biological nitrogen fixation, pest control, minimizing harmful technologies and practices for environment and human health, quantifying and minimizing impacts on externalities like greenhouse gas emissions, water availability, carbon sequestration and biodiversity. 11 2.1.1. Nutrient Use Efficiency and Precision Enhancing input use efficiency is crucial for improving crop productivity while minimizing environmental impacts in Sub-Saharan Africa. Precision nutrient management (PNM) offers a promising solution by aligning nutrient supply with crop demand. For example, in Ethiopia, the application of variable rate application (VRA) led to a 15% increase in maize yields and a 25% reduction in fertilizer costs, demonstrating both economic and agronomic benefits (Fue KG, 2025). Similarly, in Nigeria, the use of geospatially guided soil management improved crop yields while reducing the quantity of fertilizers applied (Aliyu et al., 2020). These findings underscore PNM’s role in enhancing sustainable farming practices. Site-specific fertilizer recommendations further improve productivity and efficiency. In Ethiopia, tailored nutrient strategies increased wheat yields by 16-25% and enhanced nitrogen and water use efficiencies by 30% and up to 0.83 kg grain/ha/mm of water, respectively, compared to uniform fertilizer application methods (Liben et al., 2024) These improvements indicate the value of adapting input levels to local soil and climatic conditions. Efficient nitrogen management is also critical for reducing greenhouse gas emissions, particularly nitrous oxide (N₂O). In western Kenya, applying nitrogen in split doses up to 200 kg N/ha kept emissions below 0.1% of applied nitrogen, significantly lower than the IPCC default estimate (Hickman et al., 2014). However, emissions increased markedly when nitrogen applications exceeded 100 kg N/ha, especially after a second application (Hickman et al., 2015). In Tanzania, emission factors varied between 0.13% and 0.42% depending on soil properties and nitrogen levels, reinforcing the importance of site-specific management to balance productivity with environmental protection (Zheng et al., 2019). Water management through precision irrigation also contributes to resource efficiency. In Zimbabwe, the use of wireless soil moisture sensors improved water use efficiency by up to 25% without reducing winter wheat yields, aligning irrigation more closely with crop water requirements (Munyaradzi et al., 2022). This technology supports sustainable intensification by conserving water while maintaining productivity. 12 2.1.2. Soil Health and Nutrient Management Soil health and nutrient management are essential components of sustainable crop production systems in Sub-Saharan Africa. The use of low carbon-to-nitrogen (C:N) ratio residues, such as Tithonia diversifolia, has demonstrated substantial benefits—boosting maize yields by 92% and increasing soil organic carbon by 57% compared to maize stover (Sprunger et al., 2019). Similarly, household-level studies indicate that retaining crop residues can increase wheat yields by approximately 22% and reduce the need for chemical fertilizers by 40%. The application of organic fertilizers led to a 28% increase in yields, while top-dressing with urea improved both fertilizer efficiency and economic gains. More precise nutrient strategies, such as site-specific fertilizer recommendations, further enhanced wheat yields by over 30% compared to traditional practices (Mohammed et al., 2024). Conservation tillage techniques also contribute to improved soil health and crop productivity. In Central Benin, no-till systems combined with mulch reduced soil loss by more than 30% compared to conventional tillage (Akplo et al., 2022). In Ethiopia, reduced tillage led to reductions in surface runoff and soil erosion by 30–60% and 49–76%, respectively, and increased wheat yields by 246– 323 kg/ha (Kebede, 2023). Conservation agriculture practices like cover cropping have also shown promising results. In South Africa, legume-based cover crops enhanced wheat protein content and soil nitrogen regardless of whether they were grazed, harvested, or mulched (Smit et al., 2021) When combined with mulching and crop rotation, cover cropping has led to yield gains of up to 8.4% across the continent, particularly under dryland conditions (Corbeels et al., 2020). Integrating cover crops and agroforestry practices enhances soil organic carbon (SOC) storage, further supporting environmental sustainability. In Malawi, agroforestry systems were found to sequester up to 4.17 Mg C ha⁻¹ annually, with soil carbon contributing around 12% of the total sequestered carbon (Thangata & Hildebrand, 2012). This highlights the dual benefit of these practices for both carbon sequestration and long-term soil health. Legume-wheat rotations in Sub-Saharan Africa boost yield stability, increasing cereal yields by 41% or up to 0.49 t/ha through nitrogen fixation and improved soil health (Franke et al., 2018). Moreover, the combined application of 10 t/ha compost and 3.17 t/ha lime in Northwestern 13 Ethiopia increased wheat yields from 0.91 t/ha to 3.33 t/ha—a 266% increase demonstrating the effectiveness of organic amendments on acidic soils (Addisu et al., 2025). 2.1.3. Diversification and Integration of Crops Crop diversification and integration offer resilience and productivity benefits in Sub-Saharan wheat farming. Diversifying crops reduces vulnerability to climate variability and market fluctuations (World Bank, 2018). while incorporating underutilized crops enhances climate adaptability, especially in dry wheat-growing zones (van Zonneveld et al., 2023). Crop intensification strategies using improved wheat varieties have also shown significant promise. In Ethiopia, across regions, applying the full recommended rate of fertilizers increased wheat yields to 4.9 t/ha, making local production more economically viable (Shiferaw et al., 2011). The integration of green manure crops such as lupin and vetch into wheat systems in the Ethiopian highlands improved yields by 49% in 2017 and 32% in 2018, with further gains of 18–26% in 2019 under moderate fertilizer regimes (Amede et al., 2021). Diversified grain-legume systems have also enhanced nutrient use efficiency; for example, pigeon pea–maize intercropping and groundnut–maize rotations increased nitrogen use efficiency by 42% compared to maize monoculture (Droppelmann et al., 2017). Additionally, conservation agriculture practices like bed-and-furrow planting increased wheat yields from 1.6 t/ha under conventional tillage to 2.6 t/ha—a 62.5% increase over three years in the Ethiopian highlands (Weldeslassie et al., 2015). These findings emphasize the productivity and sustainability advantages of combining improved agronomy with ecological practices in wheat-based systems. In Ethiopia, bundling insurance with credit and inputs increased adoption of improved maize seeds by 18 percentage points and fertilizer use by 16 points, highlighting the strong impact of integrated services (Belissa et al., 2018b). A similar finding from China potentially applicable to the African context showed that adopting integrated farming practices such as intercropping, no-till, and mulching led to yield increases of 15.6% for wheat, 30.1% for maize, and up to 49.9% for potato, along with a 39.2% increase in net returns and a 17.3% reduction in environmental impact over 12 years period (Chai et al., 2021). For sustainable long-term implementation of SI, it is important to evaluate the interventions based on their potential to enhance farm productivity and generate 14 income. These collective actions for ecosystem conservation require a supportive policy framework and incentives, especially when broader society benefits are targeted (Nackoney et al., 2013). SI which aims to increase productivity while minimizing environmental impact, faces barriers including insufficient incentives and weak policy support (Kassie et al., 2015; Yami & Van Asten, 2017). To achieve large-scale impacts on land productivity and environmental health, the synergistic effects of combined interventions should be explored, and adoption factors such as financial and labor constraints should be resolved (Droppelmann et al., 2017; Kuyah et al., 2021b). In this context, crop insurance can serve as a catalyst by enabling smallholder farmers to access loans and invest more confidently in their farms. For example, Kshetri (2021), found that by mitigating farming risks, crop insurance can lead to significant welfare gains, increasing smallholders’ investments and income by 20–30%. This financial uplift also strengthens their capacity to obtain credit for future agricultural improvements. 2.2. Crop Insurance as Catalyst for Improved Practices and Technology Adoption A growing body of experimental and observational research demonstrates that agricultural insurance is a powerful catalyst for modernizing agriculture in Sub-Saharan Africa. Agricultural insurance encourages the adoption of improved farming practices by reducing risk and boosting financial stability. It increases investment in machinery and modern (Fu et al., 2024), raises adoption of eco-friendly practices like straw return and biological pesticides by over 45% (Su et al., 2025). Crop insurance can boost smallholder investments and income by 20–30% (Kshetri, 2021), highlighting its substantial influence of agricultural insurance on promoting the adoption of improved farming practices. In Ethiopia, access to weather index insurance led to a 29% increase in fertilizer use, as farmers grew more optimistic about the returns on investment (Hill & Viceisza, 2012). Similarly, in Ghana, the provision of rainfall insurance boosted the adoption of fertilizer and hired labor by 19–25 percentage points substantially exceeding the impact of equivalent cash grants (Karlan et al., 2014). In Malawi, insured farmers increased their use of drought-tolerant seeds by 6–10 percentage points (Emerick et al., 2016) while multi-country trials reviewed by Carter et al. found that bundling insurance with inputs such as improved seeds or credit led to a 10–30% rise in technology adoption (Carter et al., 2017) Beyond on-farm investment, insurance also improves access to finance. According to the (World Bank, 2014) insured farmers are more likely to secure credit, 15 enabling them to adopt high-yield inputs such as hybrid seeds and pesticides. Taken together, these findings affirm that agricultural insurance is not merely a tool for managing risk it is a strategic lever for unlocking sustained investment, accelerating the adoption of modern technologies, and driving agricultural transformation in some of the world’s most vulnerable farming regions. Multiple studies across Sub-Saharan Africa provide compelling, quantifiable evidence that agricultural insurance significantly boosts the adoption of climate-smart agricultural (CSA) technologies. According to (Mnukwa et al., 2025) the combination of agricultural insurance, credit access, and extension services leads to a 45% increase in CSA adoption, with extension services alone increasing uptake by 2.8 times. In a complementary analysis, (Kombat et al., 2021) demonstrate that when insurance is bundled with essential inputs like drought-tolerant seeds and fertilizers, technology adoption rises by 10–30%, underscoring the role of insurance in unlocking high-risk, high-reward innovations. Further reinforcing these findings, (Simutowe et al., 2025) show that CSA interventions linked to insurance deliver an average 20.9% increase in adoption, with site-specific gains in Zambia and Tanzania ranging from 18% to 32%. Collectively, these results confirm that agricultural insurance not only reduces farmers' exposure to climate and market risks but also empowers them to confidently invest in technologies that enhance resilience and productivity. 2.3. Trends Toward Context-Specific Agricultural Insurance in Ethiopia In Ethiopia, several types of agricultural insurance products have been piloted to manage climate- related risks, particularly drought. The most extensively tested has been weather-indexed insurance (WII), which bases payouts on measurable weather indicators such as rainfall rather than actual crop loss. Pilots like the horn of Africa risk transfer for adaptation (HARITA) project began as early as 2009 in Tigray, using rainfall data as a proxy for losses (Oxfam, 2009) . WII has been promoted due to its cost-effectiveness and ability to minimize moral hazard and administrative (Gebrehiwot, 2015). However, uptake has remained low, largely due to farmers’ limited understanding, mistrust of payout mechanisms, and lack of affordability without subsidies (Ahmed et al., 2020a). Bundled insurance products, which combine index insurance with agricultural inputs and credit, have also been tested. These packages have shown promising results in improving the adoption of 16 modern farming technologies and increasing productivity. Notably, the combination of insurance, credit, and inputs proved more effective than insurance alone (Belissa et al., 2018a). Despite these benefits, adoption is still constrained by liquidity limitations and the complexity of bundled schemes. To address liquidity barriers, delayed payment insurance models where farmers pay premiums after harvest were introduced. These models increased uptake significantly, especially when promoted through local informal institutions like iddirs. However, high default rates on post- harvest premiums have raised concerns about long-term sustainability(Belissa et al., 2019a). Labor-based insurance models, or work-for-insurance schemes, were also piloted to allow farmers to pay for insurance with labor rather than cash. These were well-received among resource-poor farmers in Ethiopia’s highlands, offering a viable alternative in contexts where cash liquidity is low (Tadesse et al., 2017). While most of Ethiopia's insurance initiatives have focused on weather indices, there is a growing research and pilot interest in yield-based and area yield index insurance. Yield-based insurance, which compensates based on actual crop losses, is more responsive to farmers’ real experiences but is challenged by data and verification costs. Studies show that area yield indexes based on remotely sensed crop yields can provide more localized and reliable compensation mechanisms, especially in drought-prone areas like Tigray (Eze et al., 2020). Additionally, modeling studies assessing yield insurance and income protection schemes found the latter to be more cost-effective but confirmed the potential of yield-based schemes to improve household welfare (Eaves et al., 2017). To sum up, while uptake of agricultural insurance in Ethiopia faces structural and economic limitations, pilot programs continue to adapt through innovations like delayed premiums, Bundled insurance products and area yield modeling. There is a clear trend toward refining more context- specific products, especially farm level-yield-based schemes, to better align with smallholder realities. 17 2.4. Constraints and Drivers of Crop Insurance Uptake Despite the crucial role of agricultural insurance in enhancing farmers’ investment capacity, reducing agricultural risk and serving as a gateway to bundled services, the uptake of crop insurance in Sub-Saharan Africa remains limited and, in some areas, is declining. Key barriers include high premium costs, inaccurate index designs, and limited farmer understanding (Adelesi et al., 2024). Structural issues such as weak regulatory frameworks, inadequate financial infrastructure, and poor-quality weather data also hinder adoption (Ntukamazina et al., 2017). Addressing these challenges calls for more inclusive, farmer-centered insurance models and stronger collaboration between public and private sectors to improve accessibility and trust. 2.4.1. Economic and Institutional Challenges in Crop Insurance Adoption Premium price and affordability are among the key obstacles to the success of crop insurance in Sub-Saharan Africa, contributing to its limited uptake among smallholder farmers. This challenge is largely driven by high premium costs, the complexity of insurance terms, and limited awareness. Many farmers perceive insurance products as either unaffordable or too complicated to navigate, which discourages participation (Ntukamazina et al., 2017;Nshakira-Rukundo et al., 2021). that higher premiums reduce enrollment, reflecting the heavy burden on resource-poor households. Affordability therefore remains a central barrier to the adoption of agricultural insurance. High premium rates, inefficiencies in claim settlement, and administrative complexities often deter small-scale farmers (Nshakira-Rukundo et al., 2021). Operating with limited resources and narrow margins, many cannot afford even subsidized premiums. Furthermore, insurance options are often poorly communicated or insufficiently tailored to farmers’ needs. Addressing these constraints requires strategies that reduce premium costs, improve transparency, and expand outreach so that farmers are both aware of and able to access the protection insurance offers (Adelesi et al., 2024); Ntukamazina et al., 2017; Nshakira-Rukundo et al., 2021). Making insurance more affordable and streamlining service delivery through targeted research is essential to broaden participation and ensure equitable access (van Asseldonk et al., 2020). 18 Government and public-private partnerships (PPPs) play a critical role in expanding agricultural insurance. Government support through subsidies, infrastructure development, and regulatory frameworks can help reduce premium costs and build trust among farmers. Meanwhile, PPPs can share costs and risks between sectors, enabling more efficient service delivery and broader coverage. Effective collaboration between public and private actors is essential for scaling up crop insurance in a sustainable and inclusive manner (Osumba & Kaudia, 2020). However, many African countries lack the strong institutional support required to foster a reliable crop insurance market. Inadequate regulatory frameworks, limited financial infrastructure, and poor oversight hinder the development and effective implementation of insurance programs (Osumba & Kaudia, 2020) . Without clear rules and trusted enforcement, farmers are less likely to invest in or rely on insurance products. Furthermore, access to accurate and timely weather data remains a persistent challenge. Many parts of Sub-Saharan Africa lack the necessary meteorological infrastructure, such as a dense network of weather stations, to generate precise indices for insurance schemes. This deficiency undermines the accuracy of payouts and increases basis risk, further weakening the credibility of these products (Ntukamazina et al., 2017; Sinha & Tripathi, 2016; Osumba & Kaudia, 2020). 2.4.2. Behavioral Economics in Smallholder Crop Insurance Uptake Behavioral economics challenges and extends traditional models which emphasize price and risk preferences as primary drivers of adoption by revealing non-price barriers such as liquidity constraints, timing of payments, trust, complexity, social dynamics, and perceived fairness, all of which systematically reduce demand (J-PAL, 2024; Cole et al., 2012). Behavioral and informational barriers remain significant obstacles to crop insurance adoption. Farmers’ decisions are influenced by their perceptions of risk and trust in institutions, and many have limited exposure to the concept of crop insurance. Misconceptions and lack of awareness contribute substantially to low uptake. Addressing these challenges through targeted education and community outreach can help build understanding and confidence in insurance mechanisms (Nshakira-Rukundo et al., 2021). 19 Liquidity timing and premium design are major behavioral barriers to crop insurance adoption. Premiums are typically due at planting, when farmers face the greatest liquidity constraints. Evidence from Ethiopia shows that allowing payment at harvest increased take-up from about 8% to 24%, and up to 43% when promoted via local mutual support groups (Iddirs). This demonstrates that easing liquidity constraints and leveraging trusted networks can substantially increase demand, even without altering actuarial pricing (Belissa et al., 2019b). Another evidence from Kenya and Ethiopia shows that shifting premiums to harvest raised insurance uptake from 5% to 72%, highlighting liquidity constraints and present bias as key barriers to upfront payment (Casaburi and Wills, 2017). In addition, Accurate and timely loss assessment is another persistent issue. Delays or inaccuracies in estimating yield losses can lead to disputes and erode trust among farmers (Shenoy & Korb, 2024). Integrating advanced tools like satellite imagery and remote sensing technologies can enhance the speed and reliability of damage assessments, thereby enhancing insurance effectiveness (Wahab et al., 2018). Trust (Giampietri et al., 2020) and complexity (J-PAL, 2024) of insurance contracts are major obstacles. Field experiments in India showed that uptake remained low even when policies were highly subsidized, as farmers doubted whether payouts would be delivered or struggled to understand contract terms (J-PAL, 2024; Cole et al., 2012). In Ethiopia, willingness-to-pay was higher among wealthier and more educated households, whereas many women and less-educated farmers preferred group-based delivery mechanisms such as iddirs, reflecting the importance of trusted social institutions (Hill et al., 2011). Overall, in crop insurance adoption, trust and experiential learning are critical. Clear communication, repeated exposure, and delivery through trusted providers or peer groups can reduce skepticism and complexity, making farmers more willing to adopt insurance (J-PAL, 2024). Another key behavioral dimension is learning from experience. Insurance adoption is influenced not only by individual experience but also by community-level interactions. Evidence from India shows that when insurance payouts occurred within a village, the likelihood of purchasing insurance in the following season increased by 25–50%, highlighting the role of social learning and spillover effects (Cole et al., 2014). Similarly, in Ethiopia, learning and social influence were found to raise the willingness to pay for insurance. However, the overall impact remained limited due to persistent barriers related to liquidity constraints and lack of trust (Ahmed 20 et al., 2020b). Moritz et al. (2025) further demonstrate that the adoption of crop index insurance tends to increase gradually from 56% in the first round to as high as 89% in subsequent rounds as farmers gain direct experience and learn from their peers. Basis risk and perceptions of fairness further constrain adoption. In Ethiopia, evidence shows that high downside basis risk reduces index insurance demand significantly, particularly at higher prices (Hill et al., 2011). Basis Risk is another critical issue, especially in index-based insurance models. A significant technical limitation in the insurance scheme is the mismatch between actual losses experienced by farmers and the payouts determined by weather-based indices rather than actual farm-level losses. Since these indices often fail to capture local variations in climate conditions, farmers may not receive compensation even after incurring substantial losses (Ntukamazina et al., 2017; Sinha & Tripathi, 2016) or may be paid without incurring loss (Du et al., 2017). This misalignment erodes trust in insurance providers and discourages long-term participation. Kramer et al. (2025) also revealed that farmers adopt insurance when payouts are perceived as fair and consistent; unfair outcomes from basis risk quickly erode trust and reduce uptake. This suggests, the need for research to focus on improving the precision of local data and designing farm level and context-specific indices is essential to closing this gap. Insurance adoption is also shaped by social norms and peer effects. Insurance adoption is not only an individual choice but also a social process. Studies across Asia show that farmers often imitate peers, and village-level norms influence decisions. Early adopters and visible payout events create reputational signals that increase community-wide uptake, raising purchase rates by 25– 50% even among non-recipients (Cole et al., 2014). However, as farmers gain experience, their reliance on imitation and peer effects declines. Alongside these dynamics, ambiguity aversion and risk preferences strongly influence adoption. Farmers are not only risk-averse but also ambiguity-averse, they dislike uncertainty about how insurance rules translate into payouts. Recent experiments with Kenyan smallholders show that ambiguity aversion significantly reduces willingness-to-pay, particularly among women (Cecchi et al., 2024). Similar study in Uganda found that 58% of coffee farmers were ambiguity-averse, which significantly reduced adoption of weather index insurance (Lwiza & Barkley, 2025). Another evidence from Northern Ghana indicates that negative perceptions on value of insurance further lower farmers’ willingness to pay 21 for maize drought-index insurance (Abugri et al., 2017). This indicates the need insurers to emphasize in demystifying index rules. Farmers mitigate production risks through various strategies, primarily guided by their risk attitudes and perceptions, which are in turn shaped by socioeconomic characteristics (Diyyala et al., 2025). For example, Ethiopian wheat farmers prioritize risk-reducing traits such as drought and pest resistance, yield stability, and frost tolerance, while in India, 70–80% of farmers were found to be risk-averse, perceiving floods, heavy rains, and pest/disease outbreaks as the most critical threats to production. Older farmers, those with lower off-farm income, and those relying more on formal information sources tend to be more risk-averse (Ullah et al., 2015). Therefore, aligning insurance products and policies with farmers’ preferences is essential for designing more effective risk management tools that can improve adoption, resilience, and sustainability (Sarker et al., 2025; Turvey et al., 2013). Another pathway to higher adoption lies in bundling and delivery models. Insurance products gain more traction when combined with other services such as credit, inputs, mobile payments, or social safety nets. Notable programs include the Agriculture and Climate Risk Enterprise (ACRE) in East Africa and the Rural Resilience Initiative (R4) in Ethiopia and Senegal, both of which successfully scaled by embedding insurance in broader risk management packages Greatrex et al., 2015). Finally, cognitive biases also shape adoption. Farmers often overestimate their yields and revenues, which reduces their perceived need for insurance. Studies show that overprediction of future yields is a key reason for lower uptake Biswal & Bahinipati, 2022; Turvey et al., 2013). Misestimation of probabilities likewise distorts risk perceptions, making insurance appear less valuable than it truly is. In sum, behavioral economics provides a powerful lens to understand the persistent under-adoption of crop insurance. Barriers such as liquidity timing, trust, complexity, social learning, basis risk, ambiguity aversion, and cognitive biases systematically depress demand, even when products are actuarially sound or subsidized. Addressing these behavioral frictions through redesigned payment schedules, improved indices, bundling, trusted delivery channels, and clearer communication offers a path toward more inclusive and sustainable insurance adoption among smallholder farmers. 22 2.4.3. Systematic Review of Key Challenges in Agricultural Insurance Alongside a manual literature review, a systematic analysis of 810 academic papers was conducted using Python-based text analysis to explore key research themes and challenges in agricultural insurance (see Annexes 1–3). The word cloud in Figure 3 highlights a wide range of challenges identified in the literature. Notable terms such as moral hazard, basis risk, climate risk, and sustainability stand out, reflecting the most discussed and pressing issues in the field of agricultural insurance research. Moral hazard and adverse selection are two persistent behavioral challenges that significantly reduce the efficiency of agricultural insurance, particularly for smallholder farmers. When insurance is in place, farmers may take greater risks (moral hazard), and those facing higher levels of risk are more inclined to seek coverage (adverse selection). These dynamics lead to increased costs and imbalance within insurance pools. Addressing this requires the development of more adaptive and responsive contract designs. One example is dynamic pricing, where premiums are adjusted based on real-time data, individual risk profiles, and the farmer’s use of risk-reducing practices. This approach can help promote fairness and mitigate adverse selection (Mao & Ostaszewski, 2023;Du et al., 2017). Basis risk is a critical issue in index-based insurance, arising from the mismatch between farmers’ actual losses and payouts determined by weather-based indices rather than farm-level losses. Because these indices often miss local climate variations, farmers may receive no compensation despite substantial losses (Ntukamazina et al., 2017; Sinha & Tripathi, 2016)) or may be paid without incurring loss (Du et al., 2017). This misalignment erodes trust and discourages long-term participation, highlighting the need for more precise local data and farm-level, context-specific indices. Diversification presents both hurdles and possibilities. While it mitigates risks for farmers by spreading income sources or crop types, it complicates the design of insurance products (Jithitikulchai, 2023). Creating policies that accommodate varied agricultural practices requires innovative approaches (Adejuwon, 2020). Research can help develop insurance frameworks that better align with diversified farming systems. 23 Climate and systemic risks are becoming increasingly difficult to manage as extreme weather events grow in frequency and scale. When disasters strike entire regions simultaneously, it becomes challenging for insurers to distribute risk effectively and maintain financial stability. Current insurance models often fail to fully account for these wide-scale disruptions (Herbstein et al., 2013). There is a clear need for more research into climate-resilient insurance solutions that not only protect farmers but also encourage the adoption of climate-smart technologies to reduce long-term vulnerabilities (Phelan, 2011). Sustainability and profitability are both crucial to the long-term success of agricultural insurance. The model must remain financially viable for insurers while being accessible and affordable for farmers. Striking this balance is particularly challenging in the inherently high-risk agricultural sector. A system where either party consistently incurs losses is unlikely to be viable in the long run (Nshakira-Rukundo et al., 2021; Fusco et al., 2021). Research plays a key role in developing models that address and reconcile these competing needs. In addition to the above bottlenecks, the word cloud also highlights persistent challenges around product design and farmer participation. For effective and sustainable insurance uptake, farmers should be a part of the design process and needs work more on the awareness raising (Fonta et al., 2018). Together, these challenges highlight that while agricultural insurance play great role and serve as a gateway to access the bundle of service, it still requires targeted innovations and understanding to meet the needs of farmers effectively. 2.5. Crop Insurance and Research Gaps Python-based text analysis reveals (Annex 1, Figure 1) reveals that technology dominates agricultural insurance research, appearing in over 80 studies, while forecasting and technology bundling are moderately covered (around 48 papers). However, critical themes such as gender, risk-based classification, participatory design, reinsurance, and social-environmental integration are notably underrepresented (Vyas et al., 2021). Current insurance products often overlook farm diversity and sustainable practices, limiting their effectiveness in addressing basis risk, moral hazard, and climate resilience. This highlights the need to develop effective and inclusive insurance systems, through innovative models leveraging real-time data and tailored premiums. 24 2.5.1. Methodological Gap in Agricultural Insurance and SI Research Based on review of recent literature on agricultural insurance, sustainable intensification (SI), and remote sensing in Sub-Saharan Africa (SSA), while these studies have contributed important insights, they also reveal notable methodological limitations that this research directly addresses. Below is a comparative analysis focused on gaps in model design, data integration, premium calculation, and sustainability alignment. Many existing studies employ area-based index insurance models that rely on aggregated data at regional or administrative levels, which fail to capture intra-regional variability. For instance, Hochrainer-Stigler et al. (2014) develop index-based insurance using remote sensing data in Ethiopia, but their model aggregates risk at a broad spatial scale and does not account for variations in farm typologies or management practices, thereby increasing basis risk. Similarly, Mwungu et al. (2024) evaluate a soil-moisture-based insurance product in Kenya, but their approach does not integrate on-farm agronomic data or tailor the model to local agroecological realities. These models often sacrifice accuracy for scalability, which limits their effectiveness for heterogeneous smallholder systems. Methodologically, most studies focus on either remote sensing or agronomic data, rarely combining them with socioeconomic or farm-level variables in a predictive framework. Vanlauwe et al. (2014) emphasize SI’s role in smallholder resilience but treat it mainly as a separate agronomic strategy, without linking it to tools like crop insurance. This study, by contrast, integrates SI adoption into insurance pricing, directly tying risk-reducing practices to premium adjustments. Another gap lies in the use of static actuarial assumptions and limited premium differentiation. For example, Mwungu et al. (2024) use satellite-based models to assess crop damage from flash flooding, yet their premium rates are based on historical averages and lack dynamic adjustment mechanisms. Their approach excludes on-farm agronomic data and ignores farming system diversity. By same insurance trigger (soil moisture threshold) across varied contexts, the model sacrifices accuracy for scalability, limiting its effectiveness for smallholder systems. This contrasts with the present research, which introduces a SI bundles-based yield estimation, enabling more accurate and equitable premium pricing at the farm level. Moreover, while Benami et al. (2021) 25 emphasizes the need for high-resolution data and propose integrating crop models with economics, their framework remains conceptual, with limited operational application at the smallholder level. In addition to these design limitations, ML remains significantly underused in SSA insurance literature. Although Ahmed et al. (2024) applies random forest algorithms to explore climate variability, they stop short of integrating ML into premium prediction or insurance product design. Another study by (Gelagay et al., 2025) in Ethiopia maps wheat areas and defines agro-ecological zones; however, it does not model yield or production risk, lacks farm risk typologies, and does not explore the role of SI practices in enhancing yield stability or reducing production risk. It also does not establish clear linkages between these factors and insurance premium design, which limits its applicability for SI-informed crop insurance. This research will fill a critical gap by developing a dynamic, risk-based, farm-level yield index insurance model tailored to Ethiopia’s diverse wheat systems. The model will integrate remote sensing, geospatial, and ground-truth agronomic data using machine learning to enable accurate yield prediction and premium estimation across risk based-farm typologies. Once calibrated, it will require only minimal annual inputs such as weather grids, NDVI, and spatial data allowing for scalable implementation, timely payouts, and improved accuracy. By addressing the limitations of conventional area-based schemes, the model will help minimize basis risk and improve insurance accessibility for smallholders. The study will further close the research gap by incorporating SI-sensitive features into the insurance model. It will identify specific improved agricultural practice bundles (SI) such as improved seeds, crop rotation, residue retention and soil conservation that reduce yield variability and downside risk. Using Random Forest and Monte Carlo simulations, the analysis will show how these bundles contribute to more stable yields, lower premiums, and reduced basis risk. This will offer a novel, evidence-based approach linking SI adoption with insurance design, while also supporting scalable, data-efficient solutions that enhance smallholder resilience In addition, the study will generate evidence on the drivers of crop insurance adoption and assess smallholders’ willingness to pay for insurance. This will help evaluate how well current insurance offerings align with farmers’ risk perceptions, financial capacity, and interest providing insights to inform the design of more demand-responsive and inclusive insurance solutions. 26 3. Conceptual Framework Farmers’ participation in the insurance system mainly depends on the principle of "fairness" and "affordability", which aim to balance between the insurer profitability and premium levels (Biagini & Severini, 2021). This fairness and affordability also directly linked with key farmers adoption behaviors like farmers trust and perception of insurance relevance (Kramer et al. (2025). This study seeks to enhance crop insurance affordability, fairness, participation and inclusivity through two main approaches: (i) linking crop insurance (protection of farmers extreme loss) and sustainable intensification (risk reduction strategies) to design affordable, sustainable and well calibrated farm level yield index insurance (for accuracy of loss assessment) and (ii) developing risk-typology based area-yield index insurance that groups farmers by their risk level and sets premiums according to their degree of risk exposure. This section clearly presents the study’s conceptual framework. 3.1. Sustainable Intensification: Concept and Implementation Pathways Several authors define sustainable intensification (SI) broadly as a strategy to reconcile agricultural intensification with long-term sustainability. In this context, Pretty et al. (2018) define SI as maintaining or increasing yields on existing land while maintaining and enhancing environmental outcomes. (Pretty et al., 2024) highlight system redesign as key to co-producing agricultural and ecological outcomes, while Xie et al. (2019) frames SI as boosting production with reduced environmental impacts and safeguarding social, environmental, human, and economic assets through efficient input use. Similarly, Jain et al. (2023) describe SI as enhancing production while limiting ecological harm. Pretty & Bharucha (2014) also describe SI as a set of practices aimed at raising agricultural yields without adverse environmental impacts or the expansion of farmland. However, given its multi-objective nature, SI has no fixed endpoint; no single package of practices can remain effective across all contexts. For instance, adopting high-yielding crop varieties may qualify as SI if it reduces pressure on land expansion while improving profitability and yield (Pretty et al.,2018). Moreover, although there is no clear-cut pathway for SI, experts agree on the need to integrate Agricultural crop productivity and profitability with the preservation and enhancement of ecosystem services (Pretty et al., 2011b; Vanlauwe et al., 2014). 27 Together, these perspectives present SI as a systemic approach that links yield gains with sustainability through context-specific practices and innovations. In this study, SI is therefore defined as a bundle of improved agricultural practices that enhance productivity while contributing to the environmental outcome. Applied to Ethiopian wheat systems, IAP to be bundled as SI pathways include high-yielding and stress-tolerant wheat varieties, conservation agriculture (reduced tillage, residue retention, wheat– legume rotations), biodiversity integration through rotations and agroforestry, integrated pest management, soil and water conservation practices, mechanization, climate information, and balanced nutrient management and site-specific precision input use. These pathways have been emphasized by several authors, notably Pretty et al. (2018, 2024); Xie et al. (2019); Çakmakçı et al. (2023). 3.2. The Linking Between Agricultural Insurance and Sustainable Intensification In Ethiopia, evidences shows that agricultural insurance and sustainable intensification (SI) have largely been examined in isolation: agricultural insurance research has primarily focused on index design, adoption barriers and impacts on technology uptake (Belissa et al., 2019; Wassie et al., 2024; Karlan et al., 2014; Dercon et al., 2014; Eze et al., 2020), while SI research has remained centered on its performance indicators such as productivity, environmental, economic and social outcome, and system redesign (Hammond et al., 2021; Pretty et al., 2018; Pretty & Bharucha, 2014; Xie et al., 2019; Pretty et al., 2024). Yet, a critical conceptual link emerges when these domains are brought together, as several scholars argue for pairing risk transfer instruments with risk-reducing practices rather than treating them separately (Hellin et al., 2017; (Adegoke et al., 2017; Carter et al., 2017; Greatrex et al., 2015; Ofam, 2010). Agricultural insurance stabilizes income and strengthens financial security by protecting households from shocks (Key et al., 2017; Wang & Zhong, 2025; Zhou et al., 2023). Insurance participation also facilitates access to credit by supporting lender confidence and guaranteeing repayment capacity and serving as a collateral substitute (Heil et al., 2024; Schultz, 2021). Agricultural insurance is increasingly recognized as a catalyst for increasing smallholder farm investment and agricultural transformation (Kshetri, 2021). It encourages farmers to adopt improved practices and technologies by reducing production risk (Karlan et al., 2014; Carter et al., 2017) and ultimately it 28 enhances resilience and contributes to broader food security and development goals (Greatrex et al., 2015. Adegoke et al. (2017), in their review study “Successful Index-Based Insurance for Scaling Up Climate-Smart Agriculture,” highlighted that linking insurance premiums to the adoption of climate-smart technologies ensure more resilient agricultural systems and livelihoods. Similar, WFP & Oxfam America (2015) indicated that risk reduction practices were monetized as insurance premiums or used as eligibility conditions for coverage. Within the Ethiopian wheat system, several improved agricultural practices (IAPs) demonstrate their potential to reduce risks, enhance productivity, and strengthen sustainability. To begin with, improved seed, early maturing wheat varieties and soil and water conservation methods as sustainable solutions to mitigate against yield loss because of drought and heat stress (Belete et al., 2022). In addition, adopting drought-tolerant crops, small-scale irrigation, and efficient fertilizer use significantly improves food security, thereby highlighting the combined effect of improved practices in mitigating climate change risks (Zeleke Tessera & Demeke Molla, 2025). Boucher et al. (2021) further emphasize the complementarity between technologies and risk transfer mechanisms, showing that drought-tolerant maize varieties manage mid-season drought risk cheaply, while index insurance addresses systemic risks beyond the control of improved varieties. Likewise, improved disease resistant varieties reduce yield loss when compared with rust-susceptible ones (Abro et al., 2017). Furthermore, consistent legume–cereal rotation contributes to long-term soil health by improving soil quality, improves water-use efficiency, enhancing nitrogen uptake, reducing the demand for synthetic fertilizer, and delivering significant yield gains (Mesfin et al., 2023; Tzemi et al., 2024 (Wamicwe et al., 2023). These rotations also disrupt disease cycles insects, pests, diseases, and weeds, thereby lowering yield losses linked to monocropping (Wamicwe et al., 2023; Muche et al., 2021). Collectively, these benefits sustain yields over time, reduce variability, lower production costs, and promote environmental sustainability. Similarly, soil and water conservation (SWC) measures both physical (bunds, terraces, water harvesting) and biological (agroforestry, tree planting) reduce soil erosion, enhance soil fertility and moisture, improve crop productivity, and stabilize yields (EEPA, 2024). Notably, soil bunds significantly increase soil moisture by reducing runoff and enhancing infiltration, which in turn 29 improves wheat yields (Erkossa et al., 2018). Thus, by improving soil fertility and conserving moisture, SWC practices strengthen sustainability and reduce the risk of crop loss under climate and environmental stresses. Equally important, crop residue incorporation adds organic matter to the soil and substantially boosts wheat productivity while reducing dependence on chemical fertilizers. Evidence from household surveys shows that retaining residues after harvest increases wheat yields by 22% and lowers fertilizer requirements by 40% compared to households that completely remove residue (Mohammed et al., 2024). Moreover, integrating residue management with reduced tillage also supports wheat grain yield (Singh et al., 2019). In parallel, improving precision in nutrient management enhances crop yields while lowering fertilizer costs, thereby offering agronomic, economic, and environmental benefits (Fue KG, 2025 ; Aliyu et al., 2020). For instance, in Ethiopia, site-specific fertilizer application increased wheat yields by 16–25%, while improving both nitrogen and water use efficiency by 30% and 0.83 kg grain/ha/mm, respectively, compared to blanket application (Liben et al., 2024). With this approach, fertilizer use can be optimized to increase yields while contributing to long-term agricultural sustainability. Moreover, climate information services such as forecasts and early warnings when combined with indigenous and improved technologies (including SWC, improved seeds, and diversification), enable Ethiopian farmers to anticipate risks, adapt practices, and reduce crop yield losses under increasing climate variability (Kajumba, 2018). In addition, farm mechanization plays a vital role in improving productivity. Mechanization for land preparation, harvesting and threshing not only enhances wheat productivity and efficiency (Gebiso et al., 2024), but also reduces the risk of yield losses from untimely rainfall during harvest and pests damage caused (FDRE & FAO, 2018). Overall, greater economic and environmental gains arise from integrating multiple climate-smart practices rather than relying on single interventions (CIAT & USAID, 2017). These improved agricultural practices not only enhance wheat productivity but also reduce the risk of crop loss and yield variability under climate and environmental stresses, while contributing to sustainability. 30 Locally adapted bundles are particularly effective in boosting productivity, lowering risks, and strengthening resilient food systems (Sova et al., 2018). Thus, a strategic and synergistic bundle of practices ensures higher productivity, minimizes yield loss and variability, and contributes to a more resilient Ethiopian wheat production system. 3.2.1. Sustainable Intensification and Risk-Typology for Inclusive Crop Insurance Agricultural technologies and improved practices strengthen smallholders’ resilience by reducing crop losses under moderate weather risks, while crop insurance covers losses from extreme events. Combining the two lower both the likelihood and scale of widespread losses and makes coverage more cost-effective (J-PAL, 2024). Smallholder farmers are diverse, yet index insurance works best when losses are homogeneous and tightly correlated with the index (Hellin et al., 2017); therefore, segmenting farmers into more homogeneous groups can improve correlation and fairness. This study therefore aims to enhance the inclusivity and uptake of crop insurance by integrating crop insurance with sustainable intensification and applying risk-typology to group households for fairer area yield index-insurance (AYII) design. 31 Figure 1. Conceptual framework illustrating pathways to inclusivity and uptake of crop insurance: (A) linking sustainable intensification with crop insurance to lower risk and premiums, and (B) implementing risk-based area yield index pricing to improve fairness and reduce basis risk (Figure A is adapted from “Five-Step Dual-Wave Circle Cycle – Slide Template,” Presentationgo.com template, and an open-access wheat/ambarella image sourced via Google Images). 32 ➢ Crop Insurance and SI for Reinforced Risk Management Linking insurance with sustainable intensification (SI) creates a reinforcing cycle: insurance enables SI adoption, and SI stabilizes yields and reduces both the probability and size of indemnity payouts, consistent with expected-loss pricing (Mahul & Stutley, 2010). Similarly, Asfaw & Lipper (2016) emphasize that effective agricultural risk management requires combining adaptation strategies, such as CSA practices that reduce the likelihood or severity of adverse events, with agricultural insurance as a risk transfer mechanism. Operationally, embedding SI-sensitive bundles into premium modeling offers a novel pathway to design more inclusive, affordable, and fair insurance products. Pilot programs by United States Department of Agriculture demonstrate this by lowering premiums ($5 per acre) as adoption incentives for recognized risk-reducing practices, such as cover crops. This provides farmers with tangible financial rewards for sustainable farming (Sawadgo, 2024). Also, a multi-site trials by Awondo et al. (2020) indicates bundling drought-tolerant maize with rainfall index insurance yields lower premiums and higher coverage level. Specifically, in dry lowlands, the premium rate for the drought-tolerant variety was 44% to 500% lower than for the non-drought-tolerant variety, while in dry mid-altitudes, the reduction ranged from 9% to 83%. Conceptually, this integration advances the role of insurance beyond ex-post loss recovery toward a proactive instrument for enabling resilient, productive, and environmentally sustainable farming systems (WFP & Oxfam America, 2019; Greatrex et al., 2015). ➢ Farm-Level Yield Index Insurance to Reduce Basis Risk and Improve Design Another challenge is that area index-based crop insurance relies on aggregated indices that often miss the local conditions, leading to payouts without losses or uncompensated losses. This mismatch creates basis risk, eroding farmers trust in index insurance (Hellin et al., 2017; Porter et al., 2023) and causing widespread dissatisfaction with claim settlements (Njue et al., 2018). To address this, technology-driven farm-level yield index insurance using finer-resolution yield estimation is proposed as a more accurate and reliable alternative (Stigler & Lobell, 2023). By leveraging remote sensing, geospatial analytics, and machine learning (Tsiboe & Tack, 2022; (Hobbs & Khairnar, 2022; Sarkar et al., 2025), this approach integrates farmers’ actual yield data and management practices into well-calibrated predictive models. It enables accurate, timely, and 33 low-cost loss assessment (Hobbs & Khairnar, 2022), generating personalized thresholds and plot- level estimates that enhance fairness, precision, and trust (Hellin et al., 2017), thereby strengthening uptake of crop insurance. ➢ Integrating Farm Risk Typologies for Fairer Area-Yield Index Insurance The other drawback of index insurance is that farmers within a geographically defined area face the same premium rates and estimated payouts Elabed; Tsay & Paulson, 2024). In reality, however, farmers face different levels of risk, with varying risk profiles that affect both productivity and potential crop loss (Stigler & Lobell, 2023). Grouping farmers by administrative or purely geographic boundaries often ignores intra-administrative heterogeneity (soils, microclimates, management), creating high intra-zone variation and index mismatch (Estefania- Salazar et al., 2025). Applying uniform premiums and payouts across farmers with heterogeneous risk profiles raises fairness concerns and often leads to dissatisfaction. Grouping farmers according to dominant production risks such as drought, heat, excess rain, or their combinations enables more precise index-insurance thresholds and premiums, reducing basis risk and aligning payouts with actual losses (Benso et al., 2023). Scholars (Kuivanen et al., 2016; Meron et al., 2024; Kaur et al., 2021) highlight farm typology as a key tool to address this heterogeneity by grouping farms with similar characteristics, thereby enabling more targeted and context-specific interventions aligned with farmers’ constraints and priorities. The choice of variables for typology construction is guided by the objectives of the study (Kaur et al., 2021). Accordingly, in this study, wheat production-related risk variables such as topographic and soil features (Tsiboe & Tack, 2022), climatic and socioeconomic factors will be analyzed to construct a risk-based farm typology that reflects variations in production vulnerability. Grouping farmers with similar risk profiles can improve the design and targeting of crop insurance (Rigo et al., 2022). Therefore, introducing production risk differentiated premiums reduce the basis risk inherent in area index insurance and ensures fairer coverage. This, in turn, strengthens farmers’ trust in crop insurance and enhances its uptake. 34 4. Research Methodology 4.1. Study Areas Description This research will be conducted in Ethiopia’s primary rainfed areas of wheat-producing regions: Oromia, Amhara, and Central Ethiopia. In 2025, wheat occupied about 2 million hectares, producing 6.4 million tons at an average yield of 3.2 t ha⁻¹. Oromia accounted for 58% of production, followed by Amhara (28%) and the Southern Nations, Nationalities, and Peoples’ Region (8%). Between 2020 and 2025, wheat area expanded by 6%, total output rose by 10%, and productivity increased by 4%. On average, wheat represented 18% of cereal area and contributed 21% of national cereal production, underscoring its central role in food security and agricultural development (USDA, 2025). The study covers 17 districts across Oromia, Amhara, and Central Ethiopia, representing diverse altitudinal, climatic, and agroecological conditions. Elevations range from 670 m in Wemberma to 4,377 m in Goba, capturing both lowland and highland environments. Mean annual temperatures vary from 10.6 °C in Goba to 20.6 °C in Wemberma, averaging 16.6 °C across districts. Highland sites such as Goba, Gedeb Asasa, and Hitosa remain relatively cool (<15 °C), mid-altitude areas (Agarfa, Basona Worena, Siya Debirna Wayu) average 15–17 °C, while warmer conditions (>18 °C) prevail in Arsi Negele, Meket, Baso Liben, Minjar Shenkora, and Wemberma. Rainfall ranges from below 900 mm in Arsi Negele, Minjar Shenkora, and Meket to above 1,500 mm in Wemberma. The remaining districts receive 900–1,400 mm on average, providing generally favorable crop conditions but with localized risks of runoff or waterlogging. The temperature and precipitation were obtained from WorldClim v2.1. Topography, derived from NASA’s 30 m SRTM (Shuttle Radar Topography Mission) Digital Elevation Model (DEM) and classified using FAO slope classes (Jahn et al., 2006), ranges from nearly flat to moderately sloping. Arsi Negele has the flattest terrain (<2°), favoring mechanization, while Gedeb Asasa, Gimbichu, Hitosa, Minjar Shenkora, and Siya Debirna Wayu are gently sloping (2–5°). Most remaining districts (e.g., Agarfa, Baso Liben, Basona Worena, Doyogena, East Este, Goba, Jama, Lemmo, Lude Hitosa, Meket, Wemberma) are characterized by sloping terrain (5–8°), where erosion risk is more pronounced. 35 Soils are dominated by Vertisols (Agarfa, Baso Liben, Basona Worena, Gimbichu, Jama, Lude Hitosa, Siya Debirna Wayu), which support cereal production but suffer from poor drainage. Luvisols (Doyogena, East Este, Hitosa, Lemmo) are relatively fertile and better drained, while Leptosols (Goba, Meket, Minjar Shenkora, Wemberma) are shallow and erosion prone. Soil data are obtained from the EthioSoilGrids (250 m) national database (Ali et al., 2024). Less common soils include Andosols in Arsi Negele, of volcanic origin and fertile, and Mollic Gleysols in Gedeb Asasa, productive when drainage is managed (Jahn et al., 2006). Land use is predominantly cropland across districts, though Goba remains forest-dominated and Wemberma has extensive shrub and bush cover, reflecting localized variation in land cover (based on the Ethiopia Sentinel-2 Land Use/Land Cover 2016 (LULC2016) dataset). Local research shows that wheat is the dominant crop in Baso Liben, Doyogena, East Este, Gedeb Asasa, Gimbichu, Goba, Lemmo, Meket, Siya Debirna Wayu, and Wemberma, typically followed by barley, teff, maize, or faba bean. In Arsi Negele, Hitosa, and Lude Hitosa, maize leads, with wheat ranking second alongside barley, teff, or vegetables. In Jama and Minjar Shenkora, teff dominates, though wheat remains important, ranking second in Jama and fourth in Minjar Shenkora (Bushi et al., 2024; Salo, 2022; Kidane et al., 2022; Gidelew et al., 2022; Ayele, 2021 ; Ng’ang’a et al., 2025 ; Balemie & Singh, 2012; Worku et al., 2022; Mekonnen et al., 2014; Leta & Abi, 2025; Sida et al., 2021; Kuma, 2021; Gidelew et al., 2025; Kebede & Kassa, 2011; Demissie, 2029; Bizuayehu & Bahir, 2012). Informed by this agroecological diversity, a household survey will be conducted in six purposively selected districts; Siyadebir and Basona Worena (Amhara), Hitosa and Lode Hitosa (Oromia), and Lemo and Doyogena (Central Ethiopia) to assess barriers to crop insurance. In addition to household survey data, the study will use 2023 wheat crop-cut yield data collected in Ethiopia by the Alliance of Bioversity International and CIAT under my supervision. The dataset, designed to support yield- gap analysis and site-specific fertilizer scaling, spans high, medium, and low-productivity districts across three regions covering 11 zones and 13 districts such as Minjar Shenkora, Gumbichu, Jama Degollo, Siyadebir, Basoliben, Wemberma, Este, Meket, Negele Arsi, Gedeb Asasa, Agarfa, Goba, and Lemmo. 36 Figure 2. Study Site Map 4.2. Research Design and Sampling This study will follow different research design procedure for the different objectives. These including a household survey, focus group discussions (FGDs), key informant interviews (KIIs), and crop-cut yield measurements with systematic data management. In the following section, the research design is discussed in detail: The household survey will aim to identify barriers to adopting crop insurance, while also documenting farmers’ adoption of Sustainable Intensification (SI) practices, access to advisory services, risk exposure, and insurance preferences. To ensure representativeness, a stratified multi- stage random sampling design will be applied across major agroecological zones and farming 37 systems. Three key wheat-producing regions Oromia, Amhara, and Central Ethiopia (formerly SNNP) will be purposively selected based on national production statistics and their significance to Ethiopia’s wheat sector, as these regions collectively account for over 90% of the country’s wheat production. Within each region, two representative districts will be selected in consultation with regional agricultural experts, considering variations in agroecology, and cropping system farming systems to capture regional diversity. Subsequently, all wheat-growing kebeles within each district will be stratified by the level of improved agricultural practice (IAP) integration and access to extension services, classified as high and low, using information from district agricultural offices. One kebele will then be randomly selected from each stratum to ensure both representativeness and contrast in SI adoption. Finaly, household lists will be obtained from kebele administrations, and samples will be drawn proportionally and randomly. The primary farm decision-maker in each selected household will be interviewed using structured questionnaires across six districts in the Amhara, Oromia, and Central Ethiopia regions. In both district and kebele selection accessibility for fieldwork will be also considered. The survey will be employed using structure-OKD based questionnaire. In addition, focus group discussion (FDG) will be conducted in each district for in-depth understanding and complementing the interview data. Key informant interview (KII) from insurance company, research institutes, NGO, cooperatives and district agricultural extension offices to gather insights on institutional perspectives, support systems, and challenges related to crop insurance, wheat production, crop insurance. To ensure robust and representative data collection, two standard approaches will be explored to determine the household survey sample size. Approach 1: (Yamane, 1967) sample size approach: 𝑛 = 𝑁 1+𝑁𝑒2 will be applied using a total population (𝑁) of 610,560, based on the 2007 population census, and a precision level (𝑒) of 0.05. This calculation is expected to yield a base sample of 400 households, which will be adjusted for a 15% non-response rate, resulting in a final sample of 471 households. The total sample will then be proportionally allocated across the six study districts. 38 Approach 2: Cochran (1977) sample size formula: The sample size will be determined using Cochran’s (1977) formula: 𝑛0 = 𝑧2.𝑃.(1−𝑃) 𝑒2 where: 𝑛0 = 𝑆𝑎𝑚𝑝𝑙𝑒 𝑠𝑖𝑧𝑒, 𝑍 =Z-score for confidence level (1.96 for 95%), 𝑃 = estimated proportion of attribute or adoption (it is unknown, we use 0.5 for maximum sample size), 𝑒 = 𝑚𝑎𝑟𝑔𝑖𝑛 𝑜𝑓 𝑒𝑟𝑟𝑜𝑟 (0.05 𝑓𝑜𝑟 5%). This approach produced an estimated base sample size of 384 households, which, after applying a 15% adjustment for potential non-response, will result in a sample size of 453 households. Between the two approaches, the larger sample size of 471 households, derived from the Yamane formula, will be used to ensure representativeness across the six study districts. For the risk-typology–sensitive, yield-based insurance premium modeling, the study will utilize crop-cut sample–based yield data collected through a stratified random sampling design. Districts were selected based on their wheat production potential and logistical feasibility for field implementation. The crop-cut exercise was conducted among 2,360 households, but after data cleaning, 1,464 household plot records were retained. From each participating household, three wheat plots were diagonally selected to ensure representative sampling. Each plot measured 2 meters by 2 meters (4 m²), a size chosen for ease of measurement while maintaining the ability to extrapolate yields to standard units such as kilograms per hectare. This within-household replication allowed the study to capture yield variability not only across households and regions but also within individual farms. Field staff were trained in standardized crop-cut protocols to ensure consistency across all study locations. On the same households where the crop cuts were conducted, a household survey was also carried out using a structured, ODK-based questionnaire. The main objective of the survey was to complement the crop cut data with socioeconomic information, input and plot-level management practices, as well as nutrition and gender-related data. The survey questionnaire was reviewed and approved by the Institute’s Research Ethics Committee. 39 4.3. Types and Source of Data. The primary data for this study will be obtained from a household survey designed to capture behavioral and socioeconomic characteristics, biophysical conditions, and other factors influencing the adoption of Sustainable Intensification (SI) practices, access to advisory services, and participation in crop insurance. The survey will also collect information on risk exp