i Unlocking Agricultural Efficiency: A Stochastic Frontier Analysis of Smallholder Farmers in Rwanda Gilberthe Uwera Benimana, James Warner, Arnold Kwesi Missiame STRATEGY SUPPORT PROGRAM | WORKING PAPER 17 MAY 2025 ii CONTENTS TABLES ................................................................................................................................................. iii FIGURES ............................................................................................................................................... iii Abstract ................................................................................................................................................. 4 1. Introduction ................................................................................................................................ 5 2. Methods and Data ...................................................................................................................... 6 2.1 The stochastic frontier Analysis ................................................................................................... 6 2.2 Tobit Model ................................................................................................................................. 8 2.3 Data ............................................................................................................................................ 8 3. Results and discussion ............................................................................................................... 9 3.1 Descriptive statistics of variables ................................................................................................. 9 3.2 Key drivers of crop output using stochastic frontier analysis ...................................................... 10 3.3 Technical efficiency in smallholder farmers: Insights from stochastic frontier analysis of crop output value in Rwanda ................................................................................................................... 12 3.4 Factors influencing crop output technical efficiency ................................................................... 14 4. Conclusion ............................................................................................................................... 18 5. RECOMMENDATIONS AND IMPLICATIONS .......................................................................... 19 6. References ............................................................................................................................... 22 7. Appendix .................................................................................................................................. 25 ABOUT THE AUTHORS ..................................................................................................................... 27 ACKNOWLEDGMENTS ...................................................................................................................... 27 iii TABLES Table 1: Key determinants of crop output value from maximum likelihood estimations ........................ 11 Table 2: Technical efficiency of smallholder farmers in Rwanda: Maximizing crop output value ........... 13 Table 3: Projected socio-economic effects on technical efficiency ....................................................... 15 Table 4: Summary of input-output data used in the stochastic production frontier ................................ 25 Table 5: Summary of principal variables of technical efficiency used in the Tobit model ...................... 25 Table 6: Determinants of technical efficiency in crop output across Rwanda's farm types: Tobit model insights ................................................................................................................................... 26 FIGURES Figure 1: Distribution of Technical Efficiency ........................................................................................ 13 4 ABSTRACT Agriculture is central to Rwanda's economy, supporting the livelihood of about 70% of the popula- tion and contributing significantly to GDP. Smallholder farmers face many production challenges such as limited use of modern inputs, low productivity, and vulnerability to climate change. Despite efforts like the Crop Intensification Program and the Smart Nkunganire System, which aim to en- hance access to resources, agricultural productivity remains suboptimal for Rwanda smallholder farmers. This study seeks to identify specific sources of technical inefficiencies among smallholder farmers, focusing on the total value of farmer’s crop output. By using stochastic frontier analysis, a robust quantitative method for separating inefficiencies and random shocks, the study assessed the overall technical efficiency of smallholder farmers in Rwanda and identified the key factors in- fluencing crop output value. The analysis reveals that fertilizer use, pesticide application, labor, seed use, and land size are key drivers of crop output value. This research further indicates that farmers operate at only 45% of their potential productivity, given the same level of input and tech- nology, highlighting substantial room for efficiency improvements to reach the optimal output value frontier. Furthermore, additional analysis emphasizes the critical role of socioeconomic factors in shaping technical efficiency. The findings highlight the need for targeted interventions to optimize resource utilization, streamline labor allocation and strengthen access to extension services and government initiatives aimed at boosting agricultural production value. These strategies can sub- stantially improve technical efficiency, enabling farmers to achieve optimal crop output values and advancing Rwanda's agricultural development objectives. 5 1. INTRODUCTION Agriculture is a pivotal sector in Rwanda, serving as the primary source of livelihood for approxi- mately 70% of the population, and contributing significantly to the country's GDP (World Bank, 2020). Furthermore, a recent government census highlights the pivotal role of agriculture in the country's economy, with 53.4% of the employed population engaged in agriculture activities, solidi- fying it as the dominant economic sector (NISR, 2022). Smallholder farmers, who form the back- bone of the sector, face numerous challenges, including limited access to modern agricultural in- puts, insufficient extension services and increasing vulnerability to climatic change. In response, the Rwandan government has implemented a series of strategic initiatives, such as the Crop Inten- sification Program (CIP) and the Smart Nkunganire System (SNS). These programs focus on in- creasing land use efficiency, improving access to seeds, fertilizers, and extension services, and strengthening food security. The goals of these interventions align with the broader objectives of the previous PSTA 4 framework, while PSTA 5 will build on its achievements to drive further pro- gress. However, despite these efforts, agricultural productivity remains suboptimal, with many smallholder farmers unable to fully capitalize on the available resources (MINECOFIN, 2020). Ad- dressing these persistent challenges requires a deeper understanding of the factors influencing ag- ricultural efficiency, particularly technical efficiency,1 which reflects how effectively farmers convert input into outputs and subsequently translate those outputs into monetary value. The Food and Agriculture Organization (FAO) (2020) emphasizes the importance of measuring ag- ricultural efficiency to inform policy development and foster sustainable agricultural practices. Pre- vious studies conducted in Rwanda have also demonstrated that improving technical efficiency can lead to significant gains in agricultural output without the need for additional inputs, which is crucial in resource-constrained environments like Rwanda (Ngango and Hong, 2021; Ubarijoro et al., 2016). For instance, Bizimana et al. (2018) assessed the technical efficiency of maize farmers and identified inefficiencies largely driven by inadequate access to inputs and extension services. Simi- larly, Kanyamibwa et al. (2021) highlighted the role of farm management practices and financial resource access in determining technical efficiency in coffee production. Other studies, such as Ubarijoro et al. (2016), have shown the positive impact of technologies like improved seeds and organic fertilizers on agricultural production. While previous studies in Rwanda have provided valuable insights into the technical efficiency of single-crop outputs, they often overlook the broader efficiency dynamics of smallholder farmers managing diverse crops. This shortcoming limits our understanding of how smallholders optimize resources to maximize economic returns across their entire crop portfolio. This study addresses this by analyzing technical efficiency among smallholder farmers, focusing on the conversion of key inputs—land, labor, seeds, fertilizers, and pesticides—into optimal monetary returns from all household crop outputs. In addition, this research examines the role of socioeconomic factors, in- cluding household and farm characteristics, environmental conditions, and participation in agricul- tural initiatives, in shaping efficiency outcomes. By offering a holistic value perspective, the findings deliver actionable insights to optimize resource allocation and maximize the economic value of crop production among smallholder farmers in Rwanda. 1 When using crop output value as the point of analysis, technical efficiency refers to the ability of a farmer to maximize the value of crop production from the resources (inputs) they use, given the best available technology or production frontier. In this context, it measures how well a smallholder farmer converts inputs—such as land, labor, seeds, fertilizers, and pesticides—into the highest possible monetary value of crops, compared to the most efficient producers in the sample. 6 2. METHODS AND DATA 2.1 The stochastic frontier Analysis Stochastic frontier analysis (SFA)2 is a widely adopted methodology for evaluating production effi- ciency, particularly in agricultural systems. This model decomposes the unexplained aspect of the production function into two components: one capturing inefficiency in production, and the other capturing the random noise or statistical error - often due to measurement errors or external shocks (Battese & Coelli 1995, Aigner et al. 1977). Originally introduced by Aigner et al. (1977) and later refined by Meeusen and van den Broeck (1997), SFA assumes firms often operate below their potential due to inherent inefficiencies that need to be separated from stochastic error. This study estimates the technical efficiency of Rwanda smallholder farmers by analyzing how well they convert inputs into monetized crop outputs under given resource and structural constraints. To en- sure comparability across diverse crops, physical quantities have been converted into a standard- ized monetary metric, enabling the aggregation of all household crop production into a single measure of total crop output value. Such an approach provides a unified framework for analyzing production efficiency and economic outcomes. In a scenario with multiple inputs contributing to a single monetary-valued output. The general form of the stochastic frontier production function is given by: 𝑌𝑖 = 𝑓(𝑋𝑖; 𝛽) + 𝜀𝑖, … (1) Where 𝑌𝑖 represents the total value of crop output for the ith household, calculated by summing the monetary values of all crops produced in the ith household. This is done by multiplying the quantity of each crop produced by the household with its price and then adding these values together. 𝑋𝑖 is the vector of input variables for the ith household; β is the vector of parameters to be esti- mated; 𝑓(𝑋𝑖; 𝛽) represents the production function for the ith household (e.g., Cobb-Douglas, translog); and 𝜀𝑖 is the error term, capturing random inefficiencies and unobserved factors. The er- ror term can also be further decomposed into: 𝜀𝑖 = 𝜈𝑖 − 𝑢𝑖 … (2) where, 𝜈𝑖 is the random error term assumed to be normally distributed with zero mean and con- stant variance, 𝑁(0, 𝜎𝜈 2) and 𝑢𝑖 represents a non-negative inefficiency term, representing the devi- ation from the frontier (inefficiency), which is assumed to follow a half-normal, exponential, or trun- cated normal distribution with variance 𝜎𝑢 2 (Coelli et al., 2005). This approach was initially devel- oped by Aigner et al. (1977), who provided a foundation for analyzing production efficiency by dif- ferentiating between random shocks and inefficiency effects. The study specified the stochastic frontier model using a translog production function with input variables. The translog function is preferred over Cobb-Douglas due to its flexibility in capturing complex, non-linear relationships between inputs, such as labor, fertilizer, and land, which may in- teract in variable elasticities and interaction terms, an essential feature in agricultural production. The translog model enables us to better reflect these complexities and variations in production pro- cesses, providing a more accurate analysis of technical efficiency (Kumbhakar and Lovell, 2000; Christensen, Jorgenson, and Lau, 1973). 2 Stochastic frontier analysis is an econometric method used to measure the efficiency of production. In this study, it measures efficiency by estimating the maximum achievable crop value output value from given inputs. 7 Additionally, we assume a half-normal distribution for the inefficiency term due to its simplicity, ease of estimation, and widespread use in the literature (e.g., Aigner et al. 1977, Meeusen and van den Broeck 1977, and Greene 2008). This choice ensures non-negative inefficiency, making it a robust and interpretable assumption for our analysis. The translog production function includes both linear and interaction terms to account for the rela- tionships between inputs (𝑋𝑖) and their joint influence on the output value (𝑌𝑖) . The translog model is specified as follows: 𝑙𝑛𝑌𝑖 = 𝛽0 + ∑ 𝛽𝑖 𝑙𝑛𝑋𝑖 𝑛 𝑖=1 + 1 2 ∑ ∑ 𝛾𝑖𝑗 𝑙𝑛 𝑋𝑖 𝑙𝑛 𝑋𝑖𝑗 𝑛 𝑗=1 𝑛 𝑖=1 + (𝑣𝑖 − 𝑢𝑖)… (3) Where 𝑌𝑖 is the value of total crop output of the individual farm household; 𝑋𝑖 represents input vari- ables, such as land size, labor cost, fertilizer and pesticide, and seed cost; 𝛽0 is the intercept; 𝛽𝑖 are the coefficients corresponding to the input variables; 𝛾𝑖𝑗 represents the parameters for the in- teraction terms, capturing how different input variables jointly influence the output value; and 𝑣𝑖 − 𝑢𝑖 is the inefficiency term. After estimating the model, technical efficiency (𝑇𝐸𝑖) is calculated for each observation by compu- ting the inefficiency term 𝑢𝑖. Therefore, the TE for each observation is calculated as: 𝑇𝐸𝑖 = 𝑌𝑖 𝑌𝑖 ∗ = 𝑌𝑖 𝑓(𝑋𝑖;𝛽) 𝑒𝑥𝑝 (𝜈𝑖) = 𝑒𝑥𝑝 (− 𝑢𝑖)… (4) 𝑌𝑖 is the actual output, valued in monetary terms, for each farmer, and 𝑌𝑖 ∗ is the maximum possible output value based on the estimated production value frontier, 𝑓(𝑋𝑖, 𝛽) is the predicted output value from the translog production function. 𝑢𝑖 represent the inefficiency for the ith farmer. If 𝑢𝑖 = 0,the farmer is perfectly efficient (i.e., they operate on the efficiency frontier), and if 𝑢𝑖 > 0 , they are technically inefficient. The estimated TE scores for each farmer represent how efficiently they use their resources (inputs) compared to the highest possible monetary outcome (frontier). The value of TE ranges from zero to one, where a TE equal to one indicates that the farmer exhibits perfect efficiency and is operat- ing on the production value frontier. A TE less than one indicates inefficiency, with smaller values reflecting less efficiency. A farmer with higher technical efficiency is closer to the production value frontier, while those with lower scores are more inefficient (Aigner et al., 1977, Battese & Coelli, 1995). Using Stochastic Frontier Analysis (SFA), with a translog production function, provides quantitative estimates of technical efficiency and input elasticity, as well as input use efficiency and as well as potential drivers to improve agricultural output value in Rwanda. 8 2.2 Tobit Model In this study a Tobit Model3 was utilized to analyze the effect of various socioeconomic factors on technical efficiency, because the efficiency scores are censored or truncated (Tobin, 1958). Since the TE scores are between zero and one (i.e., they cannot be negative nor exceed 100 percent ef- ficiency), the Tobit model is appropriate for handling the specific distribution of the dependent vari- able. This model estimates the factors of various socioeconomic factors on the TE of farm produc- tion. These factors include household size, age and gender of the household head, access to ex- tension services, and participation in agricultural programs. The Tobit model is specified as: 𝑇𝐸𝑖 ∗ = 𝛽′𝑋𝑖 + 𝜀𝑖 … (5) 𝑇𝐸𝑖 ∗ is the latent variable representing the potential technical efficiency, 𝑋𝑖 denotes the vector of explanatory variables (socioeconomic factors) for the ith observation, 𝛽′ represents the parameters to be estimated that is a vector of coefficients corresponding to the explanatory variables, and 𝜀𝑖 is the error term assumed to be normally distributed 𝑁 (0, 𝜎𝑣 2) . The observed technical efficiency 𝑇𝐸𝑖 is censored between zero and one, reflecting the minimum and maximum bounds of efficiency scores. The observed TE score is defined as: 𝑇𝐸𝑖 = { 𝑇𝐸𝑖 ∗ 𝑖𝑓 0 ≤ 𝑇𝐸𝑖 ∗ ≤ 1 1 𝑖𝑓 𝑇𝐸𝑖 ∗ ≥ 1 0 𝑖𝑓 𝑇𝐸𝑖 ∗ ≤ 0 ….. (6) Tobit model results intend to provide insights into the factors influencing farmers' technical effi- ciency, providing insights to guide policy recommendations for maximizing smallholder crop output value. 2.3 Data This study leveraged data from the Smallholder Commercialization Survey, building on the work of Warner et al. (2023) and other key studies that have examined crop commercialization (Warner et al., 2024), farm typologies (Benimana et al., 2024), and the costs and returns in Rwandan small- holder agricultural production (Mugabo et al., 2024). The stochastic frontier analysis (SFA) framework, used to measure technical efficiency, incorpo- rates total crop output valued in monetary terms. The approach standardizes and aggregates di- verse crops, addressing the challenge of comparing physical quantities (e.g., 1 kg of maize vs. 1 kg of Irish potatoes), which differ in characteristics and market value. By valuing output in monetary terms, we create a unified, comparable measure of production efficiency across various crops and regional prices. 3 The Tobit model is a censored regression model used when the dependent variable is continuous but limited within a range. In this case efficiency scores are bounded between 0 and 1. Tobit helps analyze relationships while accounting for these limits, making it ideal for data with natural cutoffs. https://hdl.handle.net/10568/132209 https://hdl.handle.net/10568/141718 https://hdl.handle.net/10568/141717 https://hdl.handle.net/10568/148697 https://hdl.handle.net/10568/148697 9 The data used also encompasses key agricultural inputs such as land, measured in hectares, la- bor, fertilizers, seeds, and pesticides represented in monetary terms to reflect their associated amount used. This approach not only reflects the economic investment involved in crop production but also offers a more comprehensive view of resource allocation across different types of inputs. Building on the methodologies of Latruffe (2017), Bonfiglio (2019), and Tenaye (2020), crop out- puts and inputs are measured in monetary terms to enable consistent comparisons. To further en- hance accuracy, these values are adjusted for regional price variations and seasonal fluctuations, ensuring they represent local market dynamics and conditions accurately across various produc- tion environments. This framework thus provides a robust basis for assessing efficiency while ac- counting for localized economic factors that influence agricultural productivity. The Tobit model analysis integrates diverse data sources to capture the effect of household, pro- grammatic, and farm-specific factors on technical efficiency. This approach provides a deeper un- derstanding of how household demographics, access to agricultural programs, and farm attributes contribute to efficiency variations. Household factors, including gender, age, literacy, size of the household head, and access to wage income, are considered social dynamics influencing effi- ciency. Programmatic factors, such as access to extension services, the Smart Nkunganire Sys- tem, farmer field schools, participation in cooperatives, and land-use consolidation programs, are hypothesized to improve resource use, knowledge transfer, and market access, thereby enhancing efficiency. Farm characteristics, like livestock production, monocropping, land ownership, altitude, rainfall variability, and market proximity, account for environmental and logistical challenges affect- ing performance. This comprehensive analysis illustrates how social, economic, and environmental factors interact to maximize smallholder farmer agricultural output value. The selection of these variables is grounded in established literature, including studies on technical efficiency and factors affecting it across different regions and crops in Rwanda, as seen in the works of Ngango and Hong (2021), Ubarijoro et al. (2016), Bizimana et al. (2018), Tenaye (2020), and Bonfiglio et al. (2020). Furthermore, these variables align with key indicators and objectives outlined in Rwanda’s Fourth Strategic Plan for Agricultural Transformation (PSTA 4), supporting the development of evidence-based policies and strategic initiatives aimed at driving agricultural transformation in Rwanda. This alignment is intended to address essential areas of growth, effi- ciency, and sustainability within the sector. 3. RESULTS AND DISCUSSION 3.1 Descriptive statistics of variables On average, each household cultivates an average of four different crops across its plots during the 2021–2022 agricultural year. This study aggregates the total crop production per household by calculating the total output value of all cultivated crops. Staple crops, including bananas, maize, cassava, sweet and Irish potatoes, and beans, constitute a significant portion of household produc- tion. In addition, farmers often grow fruits and vegetables, with avocados and tomatoes being prominent, though they are cultivated less frequently, they are typically grown for commercial pur- poses. Irish potatoes, maize, and cassava serve both consumption and commercial needs. In con- trast, beans, sweet potatoes, and bananas are mainly cultivated for household consumption, high- lighting their importance as subsistence crops. In Rwanda, monocropping is the dominant agricul- tural practice, followed by mixed cropping, as noted by Warner et al. (2023). 10 Based on the variables4 used in this study, the average total crop output value per household was 190.6 USD, highlighting the economic contribution of smallholder farming. The primary inputs ex- amined include land and it is the basic asset of smallholder farmers in Rwanda with each house- hold cultivating an average of 0.4 hectares, underscoring the small-scale nature of Rwandan agri- culture. Households spent an average of 10 USD on seeds and 33 USD on fertilizers, pesticides, and fungicides. While 48.5% of households used inorganic fertilizers and 83.2% used organic ferti- lizers, only 16.4% applied pesticides or fungicides. These inputs are vital for enhancing crop out- put, increasing farm income, and ensuring food security, which is why the Rwandan government supports their use through subsidies, particularly for improved seeds and inorganic fertilizers, in line with the Strategic Plan for Agricultural Transformation (PSTA) policy frameworks. Fertilizers make up the largest portion of the input subsidy budget and are the most significant cash expense for farmers (Spielman et al., 2022). Labor is also a significant production input, with households spending an average of 33.3 USD on labor. This expenditure highlights the importance of labor in smallholder farming, where manual work is essential to maintain and cultivate the land effectively. Overall, this analysis offers a detailed overview of these inputs that shape household crop production values efficiency in Rwanda. Factors hypothesized to influence farmers' technical efficiency in crop production are listed in Ta- ble 5. For a detailed description of these variables, refer to the Rwanda Smallholder Commerciali- zation Survey Overview, which offers comprehensive insights into each selected variable. 3.2 Key drivers of crop output using stochastic frontier analysis Table 1 provides maximum likelihood (ML) estimates for translog production functions, identifying key drivers of crop output value among Rwandan smallholder farmers. This study shifts the focus from crop quantity to examine crop output value, providing critical insights for policymakers in Rwanda aiming to help smallholder farmers go beyond simply producing a higher quantity of crops and focuses on how effectively they can turn their inputs into the highest possible monetary value of crop output. The findings highlight opportunities to boost crop production, emphasizing land size, seeds, labor, fertilizers, and pesticides as key inputs. The results show a significant positive impact of fertilizers and pesticides on crop output value, em- phasizing their crucial role in boosting economic returns for smallholder farmers in Rwanda. This finding aligns with research by Ngango and Hong (2021), who observed that inorganic fertilizers significantly increased maize production in Rwanda and is further supported by REMA (2016) and Muriithi et al. (2022), who also found that improved agricultural inputs, including fertilizers and pes- ticides, substantially boost crop value outputs. For smallholder farmers in less mechanized sys- tems, effective input management plays a critical role in enhancing crop output. These studies un- derscore the importance of fertilizers and pesticides in improving agricultural production and eco- nomic viability. Enhancing access to these inputs could further elevate crop output value, contrib- uting to food security and sustainable agricultural development in Rwanda. That is consistent with broader research, such as Croppenstedt et al. (2013), which highlighted the importance of access to inputs, particularly fertilizers, in boosting crop production, especially for female-headed house- holds. However, it is important to acknowledge that excessive use of fertilizers and pesticides can lead to nutrient imbalances and environmental concerns, as noted by Zhang et al. (2015) and Chen, 4 Descriptive statistics for the study variables are presented in Table 4 in the appendix. https://hdl.handle.net/10568/132209 https://hdl.handle.net/10568/132209 11 Huffman, and Rozelle (2009). Therefore, while these inputs are beneficial, optimizing their use and monitoring potential environmental impacts is essential. An integrated land management strategy, incorporating organic matter enhancement, improved moisture retention, and careful use of inor- ganic fertilizers, is crucial for sustainable production gains. Furthermore, targeted extension ser- vices can help farmers refine their input management strategies, ensuring both increased output and long-term environmental sustainability, as suggested by Sakho-Jimbira and Hathie (2020). The results reveal a significant positive linear relationship between land size and crop output value. However, the negative squared term indicates diminishing returns beyond an optimal threshold, revealing that land size expansion initially raises relative crop output value, but their impact be- comes less pronounced with increased hectares of land, suggesting that expanding land size helps up to a point, but after that, it does not appear to enhance productivity and may even cause a slight reduction in crop output value. Understanding these dynamics is crucial for improving farm policy. The result highlights the need for strategic land management. This finding aligns with studies by Ali and Deininger (2015), Gollin (2019) and Giller et al. (2021) who observed that while increased land size improves productivity, larger land holdings can result in diminishing resources and labor re- turns. Furthermore, Byiringiro and Reardon (1996) and Jayne, Chamberlin, and Headey (2014) ob- served that Rwandan farmers, and those in Sub-Saharan Africa more generally, often face dimin- ishing returns on larger landholdings due to complexities in managing resources and labor effec- tively. These findings underscore the need for targeted agricultural policy to optimize land use while en- suring environmental sustainability. Promoting agroecological practices such as intercropping, ag- roforestry, and conservation agriculture can enhance productivity, maintain soil health, and protect the environment. For instance, initiatives that encourage crop diversification are essential for max- imizing land utility and mitigating risks associated with monocropping, thereby strengthening resili- ence to market and climatic shocks. Additionally, integrating GIS technology and advanced data analytics for land-use monitoring and policy assessment can facilitate adaptive, evidence-based interventions tailored to the evolving needs of farmers and local contexts. Table 1: Key determinants of crop output value from maximum likelihood estima- tions Inputs Variables coefficient Standard error Seed cost 0.01 0.02 Fertilizer and pesticide cost 0.31*** 0.07 Labor cost -0.03 0.06 Farm size 0.64*** 0.03 ½ (Seed cost)2 0.09** 0.04 ½ (Fertilizer and pesticide cost)2 -0.01 0.02 ½ (Labor cost)2 0.06*** 0.01 ½ (Farm size)2 -0.18*** 0.05 Seed cost * Fertilizer and pesticide cost -0.01 0.02 Seed cost*labor cost -0.02 0.02 Seed cost*farm size -0.002 0.02 Fertilizer and pesticide cost*labor cost -0.003 0.03 Fertilizer and pesticide cost*farm size -0.06* 0.03 Labor cost*farm size 0.07** 0.03 Constant 5.4 *** 0.1 Observations 2,006 Source: Authors’ calculations. 12 The results also indicated that while the linear relationship between seed use and crop output value is not significant, the positive and significant squared term reveals a non-linear effect. This suggests that seed use only begins to substantially enhance crop output once a certain threshold is reached. In other words, underuse or overuse of seeds is inefficient—optimal seed application is essential for maximizing production efficiency. This finding underscores the importance of precision seed management in improving crop productivity. Achieving higher agricultural returns requires not only the correct amount of seed, but also a broader, integrated approach that includes factors such as soil quality, complementary inputs (fertilizers, pesticides), and efficient farming practices. This conclusion is consistent with Kumbhakar et al. (2015), who stressed that seed input alone cannot address low agricultural productivity. Rather, it must be optimized within a framework that consid- ers the interplay of various inputs. Moreover, McGuire and Sperling (2016) highlighted that the benefits of high-quality seeds are significantly amplified when combined with effective manage- ment and appropriate input levels. Therefore, providing farmers with better access to extension services and input becomes crucial in helping them fine-tune their seed use and, in turn, increase crop output value. These findings reinforce the need for targeted agricultural interventions that fo- cus on optimizing input use to achieve sustainable productivity gains. The findings revealed that while labor alone does not significantly affect crop output value, its ef- fectiveness increases beyond a certain threshold, highlighting the importance of optimal labor utili- zation. This suggests that improved management practices and the integration of labor with com- plementary inputs are essential for increasing the economic value of produced crops. Moreover, the positive correlation between labor use and farm size implies that labor productivity increases with larger land sizes, likely because labor is better utilized on larger plots, enabling economies of scale or better allocation of tasks. This finding aligns with Haggblade et al. (2010), who observed that in Sub-Saharan Africa, productivity on larger landholdings increases with expanded labor allo- cation. Additionally, Hazell (2017) underscores the importance of strategically integrating labor with other inputs to boost crop output value. For smallholder farmers in Rwanda, leveraging extension services to optimize labor use and farm management practices is crucial for unlocking crop produc- tion gains. 3.3 Technical efficiency in smallholder farmers: Insights from stochastic frontier analysis of crop output value in Rwanda This section presents the results of a stochastic frontier analysis to assess smallholder farmers' ability to maximize the economic value derived from their crop production. Instead of focusing on physical crop output only, we evaluate the monetary value of output, considering both the quantity and the market value of the crops produced, with input use represented by household expenditure on inputs. This approach offers a nuanced perspective on how effectively smallholders can convert their resources—such as land, labor, and capital—into the highest possible monetary value of pro- duction. By focusing on output value efficiency, this analysis addresses land and income con- straints while supporting greater commercialization and sustainable growth. Table 2 presents the technical efficiency scores for smallholder farmers in Rwanda, assessing how effectively they utilize inputs such as land, seeds, labor, fertilizers, and pesticides to maximize crop output value. These scores reveal performance gaps and highlight critical areas for targeted inter- ventions, informing strategies to enhance resource efficiency and promote sustainable growth in the smallholder farming sector. 13 Table 2: Technical efficiency of smallholder farmers in Rwanda: Maximizing crop output value Mean Std Min Max % farms with TE > 0.5 Observation Technical Efficiency-TE All households 0.45 0.205 0.002 0.890 43.4 2,006 Source: Authors’ calculations. Figure 1: Distribution of Technical Efficiency Source: Authors’ calculations. The analysis shows that smallholder farmers in Rwanda have an average technical efficiency (TE) score of 0.45, indicating farms are operating at just 45% of their potential and only 43% of house- holds achieving a TE above 0.5, there is substantial room for improvement in resource use and output generation. This finding aligns with broader agricultural trends across Africa, where small- holder farmers consistently struggle to reach their full production potential due to various con- straints such as limited access to inputs, financial resources, and modern farming techniques. For example, Adom and Adams (2020) found that the average TE among smallholder farmers across 49 African countries was only 38.2%, indicating that approximately 62% of their potential agricul- tural output remains unrealized. Similarly, Chirwa (2007) observed a TE of 46.2% among small- holder maize farmers in Malawi, largely attributed to limited access to modern inputs and technolo- gies. The study underscores the need for targeted interventions, such as improved access to ex- tension services, technology adoption, and policy support, to enhance crop production value and close the efficiency gap. In Rwanda, lower efficiency scores may be partly due to the focus on subsistence or low-value crops rather than market-oriented or high-value crops (Mugabo and Warner, 2024; Warner et al., 2024). This focus on subsistence farming restricts farmers from optimizing input use and achieving higher crop output values, as Sell (2018) noted that smallholder farmers are often constrained by household labor demands. The study by Ngango and Hong (2021) further highlights that access to subsidies and improved inputs significantly boosts technical efficiency, suggesting that many Rwandan farmers need these crucial supports. 14 3.4 Factors influencing crop output technical efficiency Understanding the determinants of technical efficiency is essential for enhancing the economic re- turns of crop production among smallholder farmers in Rwanda. Table 3 presents the Tobit model results, identifying critical factors that influence technical efficiency in terms of crop value. The analysis reveals that household size, access to extension services, participation in the Smart Nkunganire System, irrigation practices, land use consolidation, land ownership, and monocrop- ping positively impact technical efficiency. Conversely, the gender of the household head, wage income, and distance to market exert a significant negative effect on technical efficiency. These findings are pivotal in shaping the ability of smallholder farmers to maximize crop value. By focus- ing on these determinants, policy makers and stakeholders can develop targeted strategies to opti- mize resource allocation, thereby increasing technical efficiency and enhancing economic returns in the agricultural sector. The results in Table 3 reveal a positive and significant association between larger household sizes and higher technical efficiency (TE) in maximizing crop output value. This underscores the critical role of family labor in farm activities, particularly in mitigating financial constraints that may limit the hiring of additional labor and likely reflects family labor's role in enhancing farm management and optimizing input utilization, thereby boosting technical efficiency. This is supported by the findings of the stochastic frontier analysis which highlight the positive synergy of labor with land expansion in boosting crop output value. The finding is also consistent with Kamau (2007), who noted that in smallholder farming systems, where hiring external labor is challenging, family labor plays a vital role in offsetting financial limitations. Similarly, Mechri and Cachia (2017) observed that larger households can generally allocate labor effectively, enhancing resource use. However, the results contrast with Alene and Coulibaly (2009), who found that smallholder farmers with limited landholdings and large family sizes strug- gled with efficiency due to poor labor allocation and restricted access to inputs. These mixed find- ings emphasize the critical role of both effective labor management and sufficient input access in improving technical efficiency in smallholder farming systems. When family labor is well-managed and adequately supported by inputs and subsidies, it can significantly enhance efficiency in maxim- izing crop output value among smallholder farms. The analysis also demonstrates that access to extension services significantly enhances technical efficiency among smallholder farmers in Rwanda, directly influencing farmer’s optimization of crop output value. This can be explained by the fact that agricultural extension services encompass training and practical guidance in areas such as agronomy, crop protection, efficient input use, soil management, and various other activities to enhance farming practices. This aligns with Ahmadzai (2017), who noted that effective extension services improve farming outcomes through knowledge transfer and resource optimization. Similar findings by Amaza and Iheanacho (2006) and Bravo- Ureta and Pinheiro (1993) further support the positive relationship between extension services and technical efficiency in food crop production and is supported by additional research (Alwarritzi et al. 2015, Solís et al. 2009, Tipi et al. 2009, Seyoum et al. 1998). To maximize these benefits, it is essential for governments and stakeholders to invest in agricul- tural extension programs. This includes enhancing the training of extension agents to provide tai- lored, user-friendly guidance for less literate farmers (Sakho-Jimbira and Hathie 2020). Addition- ally, promoting the Customized Agricultural Extension System (CAES) can play a key role in ensur- ing that farmers across different value chains receive tailored advisory services. This initiative en- courages collaboration among the government, private sector, and research institutions to facilitate 15 knowledge transfer and innovation. Given the proven link between extension services and im- proved efficiency, sustained investment in these programs is crucial to maximizing their impact and boosting agricultural productivity. Table 3: Projected socio-economic effects on technical efficiency Effect Socio-economic variables Positive effect Household size Extension services access Smart Nkunganire System (SNS) Irrigation application Monocropping pattern Participation in Land Use Consolidation (LUC) Household own land Negative effect Access to wage income Distance from home to market Household head gender Source: Authors’ calculations. Note: Smart Nkunganire Services: Rwandan digitalized agriculture input supply chain; A detailed table is available in Appendix Table 5. The analysis indicates that participation in the Smart Nkunganire System significantly enhances technical efficiency among farmers, particularly regarding crop output value. This improvement re- flects the SNS's effectiveness in digitalizing the agricultural input supply chain and increasing ac- cess to subsidies, financial services, and insurance. These findings align with the objectives of the Ministry of Agriculture and Animal Resources, emphasizing the SNS's role in enhancing efficiency and transparency within the agro-input subsidy program, ultimately boosting agricultural returns. Supporting literature further substantiates these results. Chirwa and Dorward (2013) highlight the essential role of subsidies in increasing technical efficiency among smallholders, while Choruma et al. (2024) and Abiri et al. (2023) demonstrate that input supply platforms like the SNS empower farmers to make informed decisions and allocate resources effectively. This aligns with broader Af- rican contexts where digital agricultural platforms enhance input distribution and farm management (Islam, Muzi, & Rodriguez Meza, 2018; Bångens & Söderberg, 2011). However, while the expansion of digital technologies has progressed significantly, challenges such as high connectivity costs, limited infrastructure, and digital illiteracy in rural areas continue to hin- der widespread access. Kropff et al. (2021) underscores the importance of addressing these barri- ers to fully leverage digital solutions and ensure that all farmers can benefit from these advance- ments. Overall, this study underscores the critical role of digital innovations, such as the SNS, in improving agricultural efficiency and crop production returns in Rwanda. To capitalize on these successes, the Government of Rwanda should expand the SNS's reach and increase subsidies, thereby further enhancing efficiency and supporting sustainable agricultural development. The results indicate that irrigation significantly enhances technical efficiency, highlighting its crucial role in optimizing crop output value. This finding aligns with the Ministry of Agriculture and Animal Resources (MINAGRI) 2023 report, which emphasizes the importance of irrigation considering Rwanda's vulnerability to climate variability. Expanding irrigation infrastructure is essential for en- hancing agricultural efficiency in the country. Furthermore, Warner et al. (2024) found that small- holder farmers who utilize irrigation are more likely to engage in high-value crop production and achieve greater levels of commercialization. Supporting this, Morais et al. (2021) and Kalli (2024) demonstrate that farms implementing irrigation practices consistently attain higher technical 16 efficiency. Given these insights, it is imperative for policymakers to prioritize the adoption of irriga- tion technologies. Implementing supportive policies will not only enhance crop production efficiency but also ensure that smallholder farmers can maximize the economic value of their output in Rwanda. Monocropping demonstrates a positive and significant effect on technical efficiency in this study. This approach, which involves concentrating on a single crop on a plot of land, facilitates more ef- fective resource management and ultimately contributes to higher crop output value. This result suggests that households practicing monocropping can allocate resources more strategically, streamline input use, and concentrate their expertise on one crop, leading to greater efficiency in crop production. By reducing the operational complexity of managing multiple crops, it enhances labor allocation and simplifies farm management. This finding aligns with previous research show- ing that monocropping can lead to more specialized agricultural practices and better input optimi- zation. For instance, Ngango and Kim (2019) found that the adoption of monocropping systems al- lowed for precise application of inputs, reducing inefficiencies in crop production. Their study on coffee farmers in Rwanda similarly demonstrated that improved cropping systems contribute to higher technical efficiencies. Moreover, Muimba-Kankolongo (2018) highlighted the advantages of monocropping for tasks like weeding, harvesting, and pest control, which are made easier when resources are concentrated on a single crop. However, it is crucial to recognize that while mono- cropping offers operational advantages, it also introduces risks such as increased vulnerability to pests, soil degradation, economic instability, and climate-related uncertainties. Thus, balancing monocropping's positive effects on technical efficiency with strategies to mitigate these risks is es- sential for long-term sustainability. The positive and significant effect of land use consolidation on technical efficiency highlights the benefits of reducing land fragmentation and increasing operational scale. Consolidated plots allow for more efficient integration of inputs and services, leading to higher crop output value. This finding aligns with Hakorimana and Akcaoz (2018), who demonstrated that land consolidation enhances efficiency by optimizing resource use and streamlining farm operations in Rwanda. Moreover, studies by Asiama et al. (2021) and Nilsson (2019) confirm that consolidated land im- proves crop management and operational efficiencies in Sub-Saharan Africa. These insights rein- force the critical role of land consolidation as a strategy to boost technical efficiency among small- holder farmers in Rwanda, by enabling better resource allocation and higher returns on inputs. The analysis also reveals that secure land ownership, particularly through formal titles, significantly enhances technical efficiency in maximizing crop output value among smallholder farmers in Rwanda. This finding aligns with Atwood (1990), who demonstrated that secure land tenure en- courages investment in land improvements and fosters effective management practices. Similarly, Mdoda and Gidi (2023) reported that full land ownership is associated with higher returns com- pared to rented land, suggesting that ownership incentivizes farmers to maximize the economic po- tential of their land. Byamugisha (2016) further supports this notion, indicating that formalizing land rights enables previously low-productive farmers to utilize their land more effectively, whether through improvements or rental arrangements. These findings underscore the critical role of land title security in enhancing technical efficiency in Rwanda's agricultural sector. While Rwanda's land title policy has successfully increased efficiency in maximizing returns from crop output among smallholder farmers, further efforts are essential to reinforce policies that promote formal land own- ership like formalizing land leasing and establishment of community land banks where unused land can be temporally allocated for farming. 17 The negative relationship between wage income and technical efficiency in crop output value, sug- gests that households with wage income may divert resources away from farming, reducing their agricultural efficiency. This shift in focus from farming to wage labor likely leads to less attention to crop management and suboptimal use of inputs such as seeds, labor, and fertilizers. Conse- quently, households relying on wage income may not fully capitalize on agricultural support pro- grams like subsidies and extension services, undermining their potential for maximizing crop out- put. These findings align with studies by Abdulai and Eberlin (2001) and Abdulai and Huffman (2000), which show that non-farm employment can reduce production efficiency by shifting priori- ties away from agriculture. The implications of these findings are critical, particularly in regions where agricultural subsidies and support programs are central to smallholder farming strategies. Our results emphasize the need for policies that balance the encouragement of diverse income streams with efforts to maintain agricultural efficiency. Households must be incentivized to optimize crop output despite having alternative income sources, ensuring the benefits of government sup- port programs are fully realized, ultimately enhancing technical efficiency in crop production. In addition, the results reveal that greater distance to markets significantly reduces the technical efficiency of smallholder farmers. This finding highlights the considerable barriers these farmers face in accessing both input and output markets. Long distances lead to higher transportation costs, logistical delays, and increased transaction costs, which hinder efficient farming operations. Farmers located farther from markets experience delays in purchasing necessary inputs, which disrupt optimal planting schedules and diminish potential yields. For households primarily focused on crop production, the adverse effects of market distance are even more pronounced. The ex- tended time and costs associated with transporting crops to market reduce their ability to quickly respond to market demand, affecting income generation and overall efficiency. Increased inefficiency created by market distance is not only a function of increased operational costs but also of reduced motivation to engage in profit-maximizing activities, as Abdulai and Huff- man (2000) observed. Their work aligns with our findings, demonstrating that farmers with limited access to markets tend to underperform in terms of efficiency. Furthermore, the negative impact of market distance on efficiency is consistent with other studies, such as those by Sibiko et al. (2013) and Binam et al. (2004), which highlight the importance of market proximity in boosting efficiency through timely input use and better market responsiveness. These findings point to the need for policies that improve rural infrastructure and market access for smallholder farmers in Rwanda. By addressing these logistical challenges, policymakers can enhance the technical efficiency of farm- ers, ensuring more timely input usage, higher crop output value, and ultimately, better livelihoods for smallholder households. Additionally, the results indicate a negative and significant effect of female household heads on technical efficiency, highlighting persistent gender disparities in agricultural efficiency. Female- headed households face significant socio-economic challenges that hinder their agricultural productivity. They often bear a heavy workload while having limited access to essential resources such as inputs, credit, and extension services, all of which are critical for effective farm manage- ment. These constraints, compounded by additional household responsibilities, can limit their abil- ity to optimize farm management and reduce their technical efficiency. The finding aligns with Qui- sumbing and Maluccio (2003), who emphasized that gender disparities in access to productive re- sources contribute to lower technical efficiency among female-headed households in developing regions. These results underscore the critical need for targeted interventions and tailored agricul- tural support to address gender-related barriers, improve resource access, and enhance productiv- ity and resilience. 18 The findings from the stochastic frontier analysis suggest that fertilizers, pesticides, and seeds are key drivers of crop output value, while labor also plays a significant role, especially on large farms. However, the diminishing returns to land expansion highlight the need for more efficient land man- agement strategies. Policymakers should focus on optimizing resource use and labor allocation to sustain and enhance crop output value while managing land expansion more effectively. However, the results suggest substantial inefficiencies in current farming practices. The average technical efficiency across farmers is only 0.45, indicating that farms are achieving less than half of their maximum possible crop output value given the same input levels. In fact, only 43% of households achieved a TE greater than 0.5, highlighting significant room for improvement. Moreover, the Tobit model highlights key strategies to enhance smallholder farmers' technical effi- ciency. Strengthening family labor capacity, improving access to extension services, and expand- ing initiatives like the Smart Nkunganire System are vital. Promoting irrigation, land consolidation, secure land ownership, and monocropping can further maximize crop output value. Addressing barriers such as efficiency gaps among female-headed households, labor diversions due to wage income, and limited market access is equally critical. These findings call for inclusive, targeted poli- cies that empower farmers, address structural inefficiencies, and sustainably boost agricultural re- turns. 4. CONCLUSION Rwanda's agricultural sector has implemented various interventions aimed at enhancing crop out- put value; however, there has been limited examination of how these efforts impact the overall technical efficiency of smallholder farmers. This study seeks to fill this gap by employing stochastic frontier analysis to evaluate the technical efficiency of smallholder farmers in Rwanda, with the aim of maximizing crop output value through the optimal use of available inputs. This study goes be- yond simply increasing crop output and explores strategies to enhance the value of crop output produced. The goal is to better inform the Rwandan government how to increase farm output value among smallholder farmers. This could ease the issue of small landholdings production efficiency as more valuable crops will lead to high output value, increasing the returns from crop production and improving smallholder income, a fundamental driver of agricultural transformation and the ba- sis of PSTA 5’s theory of change. The analysis identifies critical inputs influencing crop output, including land size, seeds, labor, and fertilizers and pesticides use. Additionally, the findings indicate that the contributions of labor and seeds become increasingly significant with higher utilization levels. Conversely, the results suggest that exceeding a certain threshold in land expansion may lead to reduced crop output value, high- lighting the importance of optimizing land use to avoid diminishing returns. The study also reveals that Rwandan smallholder farmers operate at an average TE score of 0.45, highlighting significant inefficiencies in achieving optimal production, with only 43% of households attaining a TE above 0.5. A Tobit model analysis evaluates the influence of various socioeconomic factors on TE, identi- fying critical determinants such as household size, access to extension services, participation in the Smart Nkunganire System, irrigation practices, land consolidation, and land ownership. Con- versely, negative impacts on TE are associated with the gender of the household head, wage in- come, and distance to markets. These findings underscore the importance of targeted interventions to enhance technical efficiency among smallholder farmers in Rwanda. Policymakers should strengthen access to agricultural 19 extension services to educate farmers about efficient practices and optimize resource allocation, which can positively impact TE. Promoting participation in support programs like the Smart Nkunganire System can empower farmers with resources and information that enhance productiv- ity and efficiency. Additionally, addressing gender disparities in access to resources and decision- making is vital, with policies aimed at empowering female-headed households, which often face additional barriers to achieving technical efficiency. Improving market access through infrastructure investment will help mitigate the negative effects of distance on efficiency, enabling improved ac- cess to inputs and for farmers to sell their products at fair prices. Finally, encouraging sustainable land management practices and land consolidation can help maximize efficiency, particularly given the non-linear relationships observed with land size. By adopting these strategies, stakeholders can create an enabling environment that enhances tech- nical efficiency, ultimately improving agricultural productivity and economic returns for smallholder farmers. Understanding the influence of socioeconomic factors on TE is essential for developing effective policies that support the growth and sustainability of the agricultural sector in Rwanda. This study provides a foundational analysis using data from the 2022 Smallholder Commercializa- tion Survey, contributing valuable insights to inform future agricultural policy and interventions. 5. RECOMMENDATIONS AND IMPLICATIONS 1. Optimize household labor utilization o Enhance labor productivity: Implement training programs focused on efficient farming practices and task management to maximize the productivity of larger household sizes. Introduce incentives or subsidies for hiring additional labor during peak seasons. o Family-focused policies: Design agricultural policies that support family-based farming practices by providing targeted subsidies for labor-saving technologies tai- lored to assist households facing labor shortages such as older female-headed households. o Community labor support: Develop cooperative labor-sharing initiatives and com- munity labor days to pool resources and reduce individual burdens. Introduce sim- ple labor-saving technologies for older farmers. 2. Optimizing labor use for higher productivity o Promote appropriate mechanization and labor-saving technologies to improve labor productivity and reduce reliance on inefficient manual farming practices. o Encourage youth engagement in agriculture through incentives, skills training, and entrepreneurship programs to create a more efficient agricultural workforce. 3. Strengthening agricultural extension services o Enhance the Implementation of the Customized Agricultural Extension Sys- tem (CAES): Strengthen coordination among the government, private sector, and research institutions to improve farmer access to tailored, demand-driven advisory services across value chains, with a focus on specialized support for different household types, such as youth-led farms, by promoting market-oriented practices and efficient resource use. o Promote high-value crop production: Encourage the cultivation of high-value crops per hectare by providing farmers with the knowledge and resources to reduce risks associated with market fluctuations and climate variability. This could include 20 offering insurance options, improved seeds, and access to markets, ensuring that farmers can produce more profitable and resilient crops. o Promote Public-Private Partnerships (PPPs): Encourage investment from private sector players to scale up innovative extension approaches and increase farmer outreach. o Enhance Digital, Entrepreneurial, and Training Approaches: Integrate ICT- based solutions and market-driven extension models to improve accessibility and efficiency. Launch innovative training programs on advanced crop management and sustainable techniques, partnering with farmer networks to expand outreach and adoption. 4. Enhancing access to improved inputs and knowledge o Broaden input subsidies: Scale up the Smart Nkunganire System to ensure equi- table access to essential inputs, particularly for less commercialized households. o Timely delivery and access: Improve the digital infrastructure for better data col- lection and input delivery. Consider subsidies for modern agricultural technologies to boost productivity. o Strengthening farmer cooperatives to enhance collective input procurement, re- duce costs, improve accessibility, and increase bargaining power for better market opportunities. 5. Facilitate access to irrigation o Invest in infrastructure: Expand irrigation facilities by developing small-scale sys- tems and providing water management training. Offer subsidies for installation and maintenance. 6. Improving efficiency in mixed farming systems o Promote farm specialization programs that help farmers optimize either crop or livestock production rather than splitting resources inefficiently. Develop training programs that focus on optimizing resource allocation and crop specializa- tion. o Develop integrated farming models that provide technical guidance on balancing crop-livestock interactions to enhance productivity. o Expand livestock extension services to ensure that households engaged in com- mercial livestock farming receive adequate support. 7. Strengthen land use and security for smallholder farmers o Promote group landownership models: Encourage cooperatives or associations to formalize group landownership and usage rights for small-scale farmers, particu- larly for communal resources like irrigation systems. o Land leasing program: Establish a matchmaking service to connect underutilized landowners with those seeking additional land, facilitating mutually beneficial agree- ments. o Establishing community land banks: To enhance land ownership security by pooling underutilized or disputed land for redistribution or productive use, providing farmers with equitable access to land resources, and fostering sustainable agricul- tural development. o Link land security to access to finance: enabling farmers to secure credit for agri- cultural investments, while also developing government-backed guarantee schemes to mitigate risks for lenders. 21 o Encourage agroecological practices such as crop rotation, intercropping, and conservation agriculture to improve land productivity. o Technical support for land management: Strengthen land use planning by lever- aging digital mapping and geospatial technology to optimize land allocation and sus- tainability. o Establish regional training hubs to provide workshops on large-scale farming, sustainable land management, and cooperative governance while offering financial and technical incentives to encourage best practices in land consolidation and re- source optimization. 8. Enhance seed strategies o Educate improved seed utilization: Inform farmers about optimal seed usage thresholds and provide access to high-quality seeds through subsidies. 9. Promote gender-inclusive policies o Targeted interventions for women: Implement gender-sensitive extension ser- vices and facilitate access to microcredit for female farmers. Support the formation of women-led cooperatives and provide childcare services to alleviate burdens on female farmers. 11. Integrate wage income with farming o Flexible farming support: Develop programs that allow for flexible work arrange- ments to balance wage employment with agricultural activities and offer financial lit- eracy programs to reinvest wage income into farming. 12. Strengthening market access and climate resilience o Improve market linkages by investing in rural infrastructure, roads, and storage facilities to reduce post-harvest losses and market access costs. Establish digital platforms to link farmers directly with markets. o Enhance climate-smart agriculture initiatives to help farmers mitigate risks asso- ciated with climate variability, ensuring sustainable and stable production. o Strengthen financial services such as microfinance, insurance, and credit pro- grams to enable farmers to invest in productivity-enhancing inputs and technology. 13. Implement monitoring and evaluation systems o Improve data-driven adjustments by establishing regular monitoring systems to assess the effectiveness of interventions. Utilize data to refine policies and address inefficiencies in productivity and resource use. 14. Farmer engagement in policy dialogue and implementation o Facilitate and empower farmers, to actively participate in agricultural policy dia- logues, provide informed feedback on implementation challenges, and collaborate with policymakers to develop and refine effective, inclusive, and sustainable farmer- centered policies. These recommendations maintain a focus on practical interventions that can lead to increased crop output and improved technical efficiency, while addressing the unique challenges faced by smallholder farmers in Rwanda. 22 6. REFERENCES Abdulai, A., & Eberlin, R. 2001. Technical efficiency during economic reform in Nicaragua: evidence from farm household survey data. Economic systems, 25(2), 113-125. Abdulai, A., & Huffman, W. 2000. Structural adjustment and economic efficiency of rice farmers in northern Ghana. Economic Development and Cultural Change, 48(3), 503-520. Abiri, R., Rizan, N., Balasundram, S. K., Shahbazi, A. B., & Abdul-Hamid, H. 2023. Application of digital tech- nologies for ensuring agricultural productivity. Heliyon 9(18),e22601. https://doi.org/10.1016/j.heli- yon.2023.e22601 Adom, P. K., & Adams, S. 2020. Decomposition of technical efficiency in agricultural production in Africa into transient and persistent technical efficiency under heterogeneous technologies. World Develop- ment, 129, 104907. Aigner, D., Lovell, C. A. K., & Schmidt, P. 1977. Formulation and estimation of stochastic frontier production function models. Journal of Econometrics, 6(1), 21-37. Ahmadzai, H. (2017). Crop diversification and technical efficiency in Afghanistan: Stochastic frontier analy- sis. CREDIT Research Paper (No. 17/04). Ali, D. A., & Deininger, K. 2015. Is there a farm size–productivity relationship in African agriculture? Evidence from Rwanda. Land economics, 91(2), 317-343. Alene, A. D., & Coulibaly, O. 2009. The impact of agricultural research on productivity and poverty in sub- Saharan Africa. Food policy, 34(2), 198-209. Alwarritzi, Widya, Teruaki Nanseki, and Yosuke Chomei. 2015. Analysis of the factors influencing the tech- nical efficiency among oil palm smallholder farmers in Indonesia. Procedia Environmental Sciences 28: 630–38.Amaza, P. S., Bila, Y., & Iheanacho, A. C. 2006. Identification of Factors that Influence Technical Effciency of Food Crop Production in West Africa: Empirical Evidence from Borno State, Nigeria. Journal of Agriculture and Rural Development in the Tropics and Subtropics, 107(2), 139. Atwood, D. A. 1990. Land registration in Africa: The impact on agricultural production. World develop- ment, 18(5), 659-671. Asiama, K. O., Voss, W., Bennett, R., & Rubanje, I. 2021. Land consolidation activities in Sub-Saharan Af- rica towards the agenda 2030: A tale of three countries. Land Use Policy, 101, 105140. Bångens, L., and Söderberg, B. 2011. Mobile Money Transfers and usage among micro-and small busi- nesses in Tanzania. SPIDER, The Swedish Program for Information and Communication Technology in Developing Regions Battese, G. E., & Coelli, T. J. 1995. A Model for Technical Inefficiency Effects in a Stochastic Frontier Pro- duction Function for Panel Data. Empirical Economics, 20(2), 325-332. Benimana, Gilberthe; Warner, James; and Mugabo, Serge. 2024. Synopsis: Identifying farm typologies in Rwandan agriculture: A framework for improving targeted interventions. Rwanda Strategy Support Pro- gram Policy Note 10. Washington, DC: International Food Policy Research Institute. https://hdl.han- dle.net/10568/141776 Bizimana, C., Nieuwoudt, W. L., & Ferrer, S. R. D. 2018. Measuring the Technical Efficiency of Smallholder Maize Farmers in Rwanda: A Stochastic Frontier Analysis. Agrekon, 57(1), 22-36. Bravo-Ureta, B. E., & Pinheiro, A. E. 1993. Efficiency Analysis of Developing Country Agriculture: A Review of the Frontier Function Literature. Agricultural and Resource Economics Review, 22(1), 88-101. Binam, J. N., Tonye, J., Nyambi, G., & Akoa, M. 2004. Factors affecting the technical efficiency among small- holder farmers in the slash and burn agriculture zone of Cameroon. Food policy, 29(5), 531-545. Byamugisha, F. 2016. Securing land tenure and easing access to land. African Center for Economic Trans- formation (ACET) and Japan International Cooperation Agency Research institute (JICA-RI). African Transformation Report 2016, 1-36. Byiringiro, F., & Reardon, T. 1996. Farm productivity in Rwanda: Effects of farm size, erosion, and soil con- servation investments. Agricultural Economics, 15(2), 127-136. Chirwa, E., & Dorward, A. 2013. Agricultural input subsidies: The recent Malawi experience. Oxford univer- sity press (p. 320). Chirwa, E. W. 2007. Sources of technical efficiency among smallholder maize farmers in Southern Malawi. African Economic Research Consortium (AERC). https://hdl.handle.net/10568/141776 https://hdl.handle.net/10568/141776 23 Chen, Z., Huffman, W. E., & Rozelle, S. 2009. Farm technology and technical efficiency: Evidence from four regions in China. China economic review, 20(2), 153-161. Coelli, T. J., Rao, D. S. P., O'Donnell, C. J., & Battese, G. E. 2005. An Introduction to Efficiency and Produc- tivity Analysis. Springer Science & Business Media. Choruma, D. J., Dirwai, T. L., Mutenje, M., Mustafa, M., Chimonyo, V. G. P., Jacobs-Mata, I., & Mabhaudhi, T. 2024. Digitalisation in agriculture: A scoping review of technologies in practice, challenges, and oppor- tunities for smallholder farmers in sub-Saharan Africa. Journal of Agriculture and Food Research, 101286. https://doi.org/10.1016/j.jafr.2024.101286 Christensen, L.R., Jorgenson, D.W., & Lau, L.J. 1973. Transcendental Logarithmic Production Frontiers. The Review of Economics and Statistics, 55(1), 28-45. Croppenstedt, A., Goldstein, M., & Rosas, N. 2013. Gender and agriculture: Inefficiencies, segregation, and low productivity traps. The World Bank Research Observer, 28(1), 79-109. Food and Agriculture Organization (FAO). 2020. Productivity and Efficiency Measurement in Agriculture. FAO. Gollin, D. 2019. Farm size and productivity: Lessons from recent literature. IFAD Research Series, (34), 1- 35. Giller, K. E., Delaune, T., Silva, J. V., van Wijk, M., Hammond, J., Descheemaeker, K., ... & Andersson, J. A. 2021. Small farms and development in sub-Saharan Africa: Farming for food, for income or for lack of better options? Food Security, 13(6), 1431-1454. Greene, W.H. 2008. The Econometric Approach to Efficiency Analysis. In The Measurement of Productive Efficiency and Productivity Growth (eds. Fried, Lovell, & Schmidt). Oxford University Press. Hazell, P. B. R. 2017. Africa agriculture status report: the business of smallholder agriculture in Sub-Saharan Africa. Alliance Green Revolut Afr AGRA, 5, 3-19. Haggblade, Steven & Hazell, Peter & Reardon, Thomas. 2010. The Rural Non-farm Economy: Prospects for Growth and Poverty Reduction, World Development, Elsevier, vol. 38(10), pages 1429-1441, October. Hakorimana, F., & Akcaoz, H. 2018. The functional analysis of maize production and the effect of land con- solidation on the productivity in Rwanda. The Journal of Animal & Plant Sciences, 28(1), 10. Islam, A., Muzi, S., & Rodriguez Meza, J. L. 2018. Does mobile money use increase firms’ investment? Evi- dence from Enterprise Surveys in Kenya, Uganda, and Tanzania. Small Business Economics, 51, 687- 708. Jayne, T. S., Chamberlin, J., & Headey, D. D. 2014. Land pressures, the evolution of farming systems, and development strategies in Africa: A synthesis. Food Policy, 48, 1-17. Kanyamibwa, J. M., Mulenga, B. P., & Chapoto, A. 2021. Technical Efficiency of Coffee Production in Rwanda: Evidence from the Northern Province. Agricultural Economics (Zemědělská ekonomika), 67(10), 421-430. Kamau, M. 2007. Farm household allocative efficiency: a multi-dimensional perspective on labour use in Western Kenya. Wageningen University and Research. Kalli, R., Jena, P. R., Timilsina, R. R., & Sonobe, T. 2024. Effect of irrigation on farm efficiency in tribal vil- lages of Eastern India. Agricultural Water Management, 291, 108647. Kropff, W.; Mabaya, E.; Afun-Ogidan, O.; Girvetz, E.; Jarvis, A. 2021 Digital Agriculture Profile: Rwanda. Rome (Italy): FAO. 18 p. https://hdl.handle.net/10568/113552 Kumbhakar, S. C., Wang, H.-J., & Horncastle, A. P. 2015. A practitioner's guide to stochastic frontier analy- sis using Stata. Cambridge University Press. Kumbhakar, S.C., & Lovell, C.A.K. 2000. Stochastic Frontier Analysis. Cambridge University Press. Mdoda, L., & Gidi, L. S. 2023. Impact of Land Ownership in Enhancing Agricultural Productivity in Rural Ar- eas of Eastern Cape Province. South African Journal of Agricultural Extension, 51(2), 1-23. Mechri, A., Lys, P., & Cachia, F. 2017. Productivity and efficiency measurement in agriculture: Literature re- view and gaps analysis. FAO technical report series GO-19-2017. Meeusen, W., & van den Broeck, J. 1977. "Efficiency Estimation from Cobb-Douglas Production Functions with Composed Error." International Economic Review, 18(2), 435–444. MINAGRI. 2020. Irish Potato Strategic Plan (2020-2025). Ministry of Agriculture and Animal Resources, Rwanda. MINAGRI. 2023. Annual report 2022/23. Ministry of Agriculture and Animal Resources, Rwanda MINECOFIN. 2020. Rwanda Economic Update: Edition No. 16. Ministry of Finance and Economic Planning (MINECOFIN), Republic of Rwanda. https://doi.org/10.1016/j.jafr.2024.101286 https://ideas.repec.org/a/eee/wdevel/v38y2010i10p1429-1441.html https://ideas.repec.org/a/eee/wdevel/v38y2010i10p1429-1441.html https://ideas.repec.org/s/eee/wdevel.html https://hdl.handle.net/10568/113552 24 Missiame, A., Irungu, P., & Nyikal, R. A. 2021. Gender-differentiated stochastic meta-frontier analysis of pro- duction technology heterogeneity among smallholder cassava farmers in Ghana. African Journal of Agri- cultural and Resource Economics, 16(2), 140-154. Mishra, A. K., Louhichi, K., Genovese, G., & Gomez y Paloma, S. 2023. Insights into land size and productiv- ity in Ethiopia: What do data and heterogenous analysis reveal? Frontiers in Sustainable Food Systems, 7, 1057674. McGuire, S., & Sperling, L. 2016. Seed systems smallholder farmers use. Food security, 8, 179-195. Mather, J. W. 1980. Cooperative benefits and limitations. Farmer Cooperatives in the United States Cooper- ative Information Report 1 Section 3, United States Department of Agriculture Rural Business - Coopera- tive Service Ministry of Trade and Industry,.2018. Toward Private Cooperative Enterprises and Business Entities for So- cio-Economic Transformation-National Policy on Cooperatives in Rwanda, Ministry of Trade and Industry Ministry of Agriculture and Animal Resources (MINAGRI). 2013. Land Husbandry, Water Harvesting and Hillside Irrigation (LWH) Project, Final report. Environmental And Social Management Framework (ESMF), MINAGRI Morais, G. A. S., Silva, F. F., Freitas, C. O. D., & Braga, M. J. 2021. Irrigation, technical efficiency, and farm size: the case of Brazil. Sustainability, 13(3), 1132. Muimba-Kankolongo, A. 2018. Food crop production by smallholder farmers in Southern Africa: Challenges and opportunities for improvement- Pre- and Postharvest Field Operations Chapter 7, Academic Press, p.59-71 Mugabo, Serge; and Warner, James. 2024. Costs and returns in Rwandan smallholder agricultural produc- tion: Gross margins and profitability analyses. Rwanda Strategy Support Program Working Paper 14. Washington, DC: International Food Policy Research Institute. https://hdl.handle.net/10568/148697 Ngango, J., & Hong, S. 2021. Impact of Hybrid Maize Varieties on the Technical Efficiency of Smallholder Farmers in Rwanda. Journal of Agricultural Science, 13(3), 87-98. Ngango, J., & Hong, S. 2021. Improving farm productivity through the reduction of managerial and technol- ogy gaps among farmers in Rwanda. Agriculture & Food Security, 10, 1-14. Ngango, J., & Kim, S. G. 2019. Assessment of technical efficiency and its potential determinants among small-scale coffee farmers in Rwanda. Agriculture, 9(7), 161. NISR. 2019. Agriculture Sector Report 2019. National Institute of Statistics of Rwanda. Nilsson, P. 2019. The role of land use consolidation in improving crop yields among farm households in Rwanda. The Journal of Development Studies, 55(8), 1726-1740. Rwanda Environment Management Authority (REMA). 2016. Effectiveness and efficiency of fertilizer use in rwanda, REMA and the Rwanda’s Ministry of Agriculture and Animal Resources (MINAGRI) report. De- cember 2016. Sakho-Jimbira, S., & Hathie, I. 2020. The future of agriculture in Sub-Saharan Africa. Southern Voice Policy brief, 2(3), 18. Sell, M., Bäckman, S., TettehAnang, B., & Niemi, J. K. 2018. The unequal efficiency gap: Key factors influ- encing women farmer’s efficiency in Uganda. Cogent Food & Agriculture, 4(1), 1551750. Seyoum, E. T., Battese, G. E., & Fleming, E. M. 1998. Technical efficiency and productivity of maize produc- ers in eastern Ethiopia: a study of farmers within and outside the Sasakawa‐Global 2000 project. Agricul- tural economics, 19(3), 341-348. Sibiko, K. W., Owuor, G., Birachi, E., Gido, E. O., Ayuya, O. I., & Mwangi, J. K. 2013. Analysis of determi- nants of productivity and technical efficiency among smallholder common bean farmers in eastern Uganda. Current Research Journal of Economic Theory, 5(3), 44-55. Tipi, Tolga, Nural Yildiz, Mehmet Nargeleçekenler, and Bahattin Çetin. 2009. Measuring the technical effi- ciency and determinants of efficiency of rice (Oryza sativa) farms in Marmara region, Turkey. New Zea- land Journal of Crop and Horticultural Science 37: 121–29. Tobin, J. (1958). Estimation of relationships for limited dependent variables. Econometrica: journal of the Econometric Society, 24-36. Ubarijoro, J., Hirwa, E., & Ntirenganya, C. 2016. The Impact of Agricultural Technology on Productivity: The Case of Improved Seeds and Organic Fertilizers in Rwanda. African Journal of Agricultural and Resource Economics, 11(4), 303-315. Warner, James, Benimana, Gilberthe, Mugabo, Serge, and Ingabire, Chantal. 2024. Crop commercialization in Rwanda: Current market participation and drivers. Rwanda SSP Working Paper 11. Washington, DC: International Food Policy Research Institute. https://hdl.handle.net/10568/141718 https://hdl.handle.net/10568/148697 https://hdl.handle.net/10568/141718 25 World Bank. (2020). Rwanda Economic Update: Navigating Challenges, Building Resilience. World Bank Group. Quisumbing, A. R., & Maluccio, J. A. 2003. Resources at marriage and intrahousehold allocation: Evidence from Bangladesh, Ethiopia, Indonesia, and South Africa. Oxford bulletin of economics and statis- tics, 65(3), 283-327. Zhang Chao, Z. C., Guan Ming Shi, G. M. S., Shen Jian, S. J., & Hu RuiFa, H. R. 2015. Productivity effect and overuse of pesticide in crop production in China. Journal of Integrative Agriculture 14.9 (2015): 1903-1910. 7. APPENDIX Table 4: Summary of input-output data used in the stochastic production frontier Variable Mean Min Max Crop output value (USD) 190.6 0 3,069.5 Seed cost (USD) 10.0 0 133.9 Fertilizer and Pesticide cost (USD) 33.0 0 264.6 Labor cost (USD) 33.3 0 1,774.8 Farm size (ha) 0.3 0.015 6.6 Observations 2,006 Table 5: Summary of principal variables of technical efficiency used in the Tobit model Household characteristics Household size (mean) 4.4 Household head age (years) 49 Femal7e household head (%) 30.2 Household head literacy (%) 66.6 Access to wage income (%) 62 Program participation Extension service access (%) 73 Twigire Muhinzi (%) 6 Smart Nkunganire Services (%) 29 Farmer field school (%) 8 Cooperative member (%) 14 Participation in Land Use Consolidation (%) 17 Farm characteristics Livestock production (%) 74.7 Irrigation application (%) 9.4 Monocropping pattern (%) 97 HH own land (%) 89.3 Distance home to market (minutes) 71.8 Altitude (100 m.) 1,734.8 Variation of rain (5-years average) (mm) 39.7 Observations 2,006 Note: HH: Household; * significant at 10%; ** 5%; *** 1% 26 Table 6: Determinants of technical efficiency in crop output across Rwanda's farm types: Tobit model insights Socio-economic factors Coefficient Standard error Household size 0.06* 0.03 Female household head -0.1*** 0.04 Household head age -0.007 -0.05 Household head literacy 0.007 -0.03 Extension service access 0.08** 0.04 Twigire Muhinzi 0.05 -0.07 Smart Nkunganire Services 0.09** 0.04 Farmer field school 0.01 -0.06 Cooperative member 0.06 -0.05 Livestock production 0.02 -0.04 Access to wage income -0.07** 0.03 Irrigation application 0.15*** 0.05 Monocropping pattern 0.51*** 0.08 Distance home to market -0.04** 0.02 Participation in Land Use Consolidation 0.07* 0.04 HH own land 0.21*** 0.05 Altitude (100 m.) -0.06 -0.05 Var. of rain (5-yr avg.) -0.1 -0.09 Constant -0.79 -0.58 Observations 2017 Source: Authors’ calculations. Note: * significant at 10%; ** 5%; *** 1% ; Twigire Muhinzi:Farmer-to-farmer extension approaches-farmer promoters and farmer field schools; Smart Nkunganire Services (SNS): Rwandan digitalized agriculture input supply chain. 27 ABOUT THE AUTHORS Gilberthe Uwera Benimana was, at the time of this research, a Senior Research Analyst with Development Strategies and Governance (DSG) unit of IFPRI based in Kigali, Rwanda. James Warner is a Research Fellow with Development Strategies and Governance (DSG) unit of IFPRI and Program Leader of Rwanda’s SSP, based in Kigali, Rwanda. Arnold Kwesi Missiame is a PhD candidate at the Department of Agricultural Economics, Univer- sity of Nairobi, Nairobi, Kenya. ACKNOWLEDGMENTS We thank the Ministry of Agricultural and Animal Production (MINAGRI) for their support in this study, including providing valuable insights, and facilitation that contributed to our research. INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE 1201 Eye St, NW | Washington, DC 20005 USA T. +1-202-862-5600 | F. +1-202-862-5606 ifpri@cgiar.org www.ifpri.org | www.ifpri.info IFPRI-RWANDA KG 563 Street #7, Kacyiru P.O. Box 1269 | Kigali, Rwanda IFPRI-Rwanda@cgiar.org www.rwanda.ifpri.info The Rwanda Strategy Support Program (Rwanda SSP) is managed by the International Food Policy Research Institute (IFPRI). Funding support for Rwanda SSP is provided by the European Union (EU); and the CGIAR Research Program on Policies, Institutions, and Markets. This publication has been prepared as an output of Rwanda SSP. It has not been independently peer reviewed. Any opinions expressed here belong to the author(s) and do not necessarily reflect those of IFPRI, EU, USAID, or CGIAR. © 2025, Copyright remains with the author(s). This publication is licensed for use under a Creative Commons Attribution 4.0 International License (CC BY 4.0). To view this license, visit https://creativecommons.org/licenses/by/4.0. IFPRI is a CGIAR Research Center | A world free of hunger and malnutrition mailto:ifpri@cgiar.org http://www.ifpri.org/ http://www.ifpri.info/ mailto:IFPRI-Kigali@cgiar.org https://creativecommons.org/licenses/by/4.0/