In Ethiopia, where agriculture, particularly cereals, dominates the GDP, the multifaceted concept of food security is crucial. It encompasses physical and economic access to sufficient, safe, and nutritious food for an active and healthy life, with elements such as availability, accessibility, utilization, and stability. Understanding this complex nature is vital for formulating effective policies and interventions ( Ram, 1996).

Food security’s multidimensional nature poses challenges in measurement and policy targeting ( FAO, 2015b). This section navigates through various indicators, including food availability, expenditure, access, and composite indexes. Emphasizing the demand-side approach, the focus is placed on parameters such as income, consumption, nutrition, and diet diversity ( FAO, 2015c), providing a holistic framework for evaluating the impact of small-scale irrigation on rural household food security.

Water’s pivotal role in agriculture makes irrigation a linchpin for food. This section classifies irrigation into small-scale, medium-scale, and large-scale categories. It zeroes in on small-scale, community-based groundwater irrigation, its significance in agricultural production, and its control and management by users, highlighting the central role of community involvement in decision-making processes ( CoSAERAR, 2002b).

Investments in irrigation prove instrumental in fostering crop diversification, encouraging the use of modern inputs, and achieving higher yields (CoSAERAR, 2002). Small-scale irrigation, accorded priority in Ethiopia, emerges as a key player in ensuring food security by enhancing agricultural productivity and resilience. Studies include Bhattarai et al. (2007), highlighting how irrigation investments in India led to increased crop diversification and the adoption of chemical inputs. Similarly, Huang et al. (2006) demonstrated in China that irrigation facilitated a shift from low-value subsistence production to high-value market-oriented agriculture. Von Braun (2009) emphasized the role of irrigation in breaking the vicious circle of poverty and environmental degradation in areas with persistent rural poverty.

In the Ethiopian context, small-scale irrigation schemes are recognized as a policy priority for rural poverty alleviation and food security ( Tesfaye et al., 2008). Eshetu et al. (2010) and Fikade (2020) documented the significant role of irrigation in Ethiopia, indicating increased income, resilience, and crop diversification. Hagos et al. (2009) further highlighted how irrigation in Ethiopia contributed to higher yields, income, consumption, and overall food security. The studies conducted by Watto & Mugera (2014) on groundwater markets in Punjab, Pakistan, aim to assess whether these markets enhance water use efficiency. The research explores the belief that water markets improve water productivity by efficiently transferring water to users for the highest marginal returns, leading to increased water efficiency among participants in groundwater markets. In the Pakistani context, informal groundwater markets offer an additional advantage, particularly benefiting resource-poor farmers facing financial constraints, as they cannot invest in wells. This perspective aligns with the broader discussion on the impacts of excessive irrigation on resource use sustainability and water use efficiency.

Feng Ye et al. (2023) eexamines the impact of rural industrial integration on agricultural total factor productivity (ATFP) in China, emphasizing heterogeneity and spatial spillover effects. Findings indicate a positive correlation between rural industrial integration and ATFP growth, driven by technological advancements and efficiency improvements. Regional differences exist, with the western region experiencing the most significant impact. While the study suggests a marginal incremental effect in areas with higher ATFP, a negative spatial spillover effect is observed, implying a potential inhibition of ATFP growth in surrounding regions. Relating this to excessive irrigation, the study underscores the complex interplay between agricultural practices, industrial integration, and regional dynamics, emphasizing the need for nuanced approaches to enhance resource sustainability and water use efficiency.

Irrigation methods play a pivotal role in shaping the sustainability and productivity of smallholder farming, impacting resource utilization and overall agricultural practices. The two primary categories, surface irrigation and sub-surface irrigation, are employed based on factors such as water resources, topography, and crop types, influencing the delivery of water to fields.

Surface irrigation encompasses several methods, each tailored to specific conditions. Basin irrigation, the most common, involves creating enclosed land sections for water retention, suitable for all soil types but with a high initial cost ( Fikade, 2020). Furrow irrigation, utilizing channels along the field’s slope, provides uniform water application but demands substantial labor and investment ( Dupriez & De Leener, 2002). Border irrigation extends basin principles to sloping fields, emphasizing efficient water use but requiring uniform land grading and supervision. Flood irrigation, an ancient method, is cost-effective but inefficient, resulting in water wastage due to runoff and evaporation.

Innovative techniques like drip irrigation and solar-powered irrigation revolutionize water use efficiency, particularly in arid regions. Drip irrigation, delivering water directly to crop roots through pipes, minimizes evaporation, conserves water, and enhances yields significantly ( Dupriez & De Leener, 2002).

Solar-powered irrigation, harnessing solar energy to pump and distribute water, addresses the energy costs associated with traditional methods, contributing to sustainability. Solar-powered irrigation systems (SPIS) offer significant advantages, providing a reliable energy source for water pumping in remote areas without grid access. This proves crucial in regions facing economic water scarcity, helping mitigate drought effects and enhancing year-round cultivation. SPIS contribute to increased yields, diversification, improved food security, and income generation for small-scale farmers. Despite their benefits, SPIS adoption faces challenges such as high initial costs, requiring innovative financing and technical expertise. Unregulated use may lead to unsustainable water practices and over-abstraction of groundwater. On a national scale, SPIS contribute to rural development, reduce dependence on fossil fuels, and play a role in climate change mitigation by replacing traditional power sources with renewable energy. Addressing challenges requires comprehensive water accounting, smart management, and visionary policies ( FAO, 2018).

Another noteworthy approach is sprinkler irrigation, simulating rainfall by spraying water into the air. While ensuring uniform water distribution and efficient use, sprinkler systems come with higher initial costs and maintenance requirements ( Dupriez & De Leener, 2002). This is offset by their ability to control water application rates and provide temperature-related benefits like cooling crops and frost control during freezing temperatures. These diverse irrigation techniques underscore the intricate choices faced by smallholder farmers, highlighting the need for sustainable and efficient practices ( FAO, 2018).

Amar Razzaq et al. (2022) study on the impact of participation in the groundwater market on farmland, income, and water access in Pakistan, identify positive effects of water markets on farmers’ income and equity. However, the research concurrently raises concerns regarding the overextraction of groundwater and inefficiencies in water resource utilization. The study highlights the absence of inherent mechanisms to control overuse in the current water market system, noting that some tube well owners operate with lower efficiency, indicating challenges associated with excessive irrigation. The conclusion stresses the urgent need for policy options to improve water use efficiency, safeguard groundwater sustainability, and ensure farmers’ incomes and food security. While recognizing the Punjab Water Act of 2019 as a positive step, the study raises questions about its implementation, particularly regarding tube well licensing and abstraction. The researchers recommend promoting participatory irrigation management to jointly control overexploitation and formulate regulations to enhance groundwater irrigation efficiency for the overall welfare of the community.

Despite Ethiopia’s rich historical legacy in irrigation, challenges such as fragmented farmland, political instability, and financial constraints have impeded its comprehensive development ( CoSAERAR, 2002b). An exploration of these challenges provides context to the intricate landscape of irrigation initiatives in the country.

Ethiopia grapples with food security challenges stemming from population growth, deforestation, and land distribution issues ( World Bank, 2013a). Acknowledging small-scale irrigation as a pivotal solution underscores the government’s commitment to addressing these challenges and ensuring sustainable food security in the country ( World Bank, 2013b).

A plethora of empirical studies has delved into the intricate web of factors influencing household participation in irrigation and the ensuing impact on food security. Education, gender dynamics, landholding patterns, access to credit, farm size, livestock ownership, and the adoption of modern technologies have surfaced as pivotal elements shaping participation trends and determining the overall food security scenario ( Smith et al., 2017; Zerihun et al., 2020).

The realm of irrigation techniques spans from traditional surface methods to advanced systems like drip and sprinkler irrigation, each wielding a distinct influence on water use efficiency and crop yield ( Kumar and Singh, 2015; Khan et al., 2019). This section provides nuanced insights into the advantages and disadvantages associated with various methods, enriching the understanding of the diverse landscape of irrigation practices and their implications for agricultural productivity.

Raya Kobo Woreda, nestled in the Amhara National Regional State, Ethiopia, serves as the canvas for this study. Geographically diverse and economically dynamic, it covers an expanse from 11°51′45.63′′ to 12°19′24.97′′ N north latitude and 39°19′54.87′′ to 39°53′2.33′′ E longitude. With altitudes ranging from 1360 to 3000 meters, the agro-ecological zones exhibit a fascinating blend of 59% kola, 3% Dega, and 38% woynadega. Climate intricacies, population demographics, and economic activities weave a tapestry crucial for understanding the nexus between small-scale irrigation and household food security.

For a comprehensive understanding, both quantitative and qualitative data were harnessed from primary and secondary sources. Primary data, gleaned from sample farmers in Raya Kobo Woreda’s key kebeles, Aradum, Ayub, and Abuare, provides an on-the-ground perspective. Semi-structured questionnaires, thoughtfully designed, unravel economic and social factors shaping households. Secondary data, a qualitative treasure trove, emerges from local offices, governmental organizations, publications, and policy documents.

The study utilized a multi-stage sampling methodology to derive a representative sample from the population frame. In the initial phase, a study Woreda was purposefully selected, leveraging the researcher’s familiarity and the widespread implementation of modern small-scale irrigation projects. Subsequently, three specific irrigation sites (Aradum, Ayub, and Abuarie) were purposively chosen from the 43 rural kebeles within the study Woreda. The selection criteria included factors such as accessibility, proximity to the study Woreda, and the potential number of beneficiaries of irrigation projects. The sample size (n=348) was determined utilizing Yamane’s formula, ensuring a 95% confidence level and a 5% margin of error ( Yamane, 1967).

To establish the sample frame, the Water User Association members’ registry was employed for household selection, with Key Interview informants and Focus Group Discussion members identified through purposive sampling. In the final sampling stage, a simple random sampling approach, employing the probability proportionate to size (PPS) random sampling technique, was implemented to assure the representativeness of the population.

Table 1 illustrates the population and sample size distribution across different kebeles. The total population of the study area was determined to be 2701, as provided by the Raya Kobo Woreda Agricultural Office. Sample sizes for each kebele were calculated using proportional allocation, ensuring equitable representation across both beneficiary and non-beneficiary groups. The table comprehensively presents the total population, participant sample, and non-participant sample for each kebele, offering a transparent overview of the distribution strategy employed in the study.

3.4.1 Descriptive statistics

Descriptive statistics, including mean, standard deviation, and graphical representations, are employed to illustrate the socioeconomic, demographic, and institutional characteristics of the sampled households. The evaluation of household food security utilizes methodologies such as the Household Food Insecurity Access Scale (HFIAS), Household Dietary Diversity Score (HDDS), and Food Consumption Score (FCS) to offer a detailed understanding of food security conditions.

3.4.2 Econometric models

Econometric analysis, specifically logit regression, addresses factors influencing participation in small-scale irrigation. Propensity Score Matching (PSM) is employed to assess the impact of irrigation participation on productivity and income. The PSM model considers conditional independence assumption (CIA) and common support condition (CSC) to mitigate selection bias in observational studies and quasi-experimental evaluations. These econometric models contribute to a robust analytical framework for understanding the determinants and consequences of small-scale irrigation participation, crucial for formulating effective agricultural policies.

This study, titled “Exploring the Nexus between Small-Scale Irrigation and Household Food Security: A Comprehensive Study in Raya Kobo Woreda, Amhara Regional State, Ethiopia,” involved human participants and adhered to strict ethical guidelines. Informed consent was obtained from all participants after providing them with detailed information about the study, ensuring their right to withdraw at any time. Participant confidentiality and anonymity were rigorously maintained, with personal identifiers anonymized and data securely stored.

The study received approval from the Institutional Review Board (IRB) committee of the Faculty of Business and Economics at Woldia University on July 13, 2023 (Ref: FBE/RCSTT/216/2023). Field research was conducted respectfully, considering the cultural and social context of the Raya Kobo Woreda, and aimed to benefit the community by enhancing understanding of food security and asset-building strategies.

4.1.1 Descriptive statistics of categorical variables

Involving 348 surveyed households, 43.68% participated in small-scale irrigation, revealing significant differences between participants and non-participants. Gender-wise, 67.10% of participants were male-headed households, highlighting a strong association between gender and participation. Access to improved agricultural technology was pivotal, with 67.10% of participants having access, in contrast to 28.57% among non-participants. This disparity underscored the positive influence of technology access on irrigation participation. Notably, 57.5% of participant households received extension services, compared to 41.83% among non-participants, indicating a substantial divergence in service contact.

4.1.2 Descriptive statistics of continuous variables

Exploring demographic and socio-economic factors revealed significant differences. The average age of household heads was 51 years, with a notable mean difference between participants (42.34 years) and non-participants (57.77 years), suggesting that younger household heads were more likely to engage in small-scale irrigation. Livestock ownership, measured in Tropical Livestock Units (TLU), displayed a mean difference between participants and non-participants, indicating diverse livestock sizes.

4.1.3 Food security status of sample households

Table 2 shows that the assessment of household food security using the Household Food Insecurity Access Scale (HFIAS) revealed 63.51% of households were food secure, while 36.49% were food insecure. Among irrigation users, 71.71% were food secure, underscoring a significant positive correlation between participation in small-scale irrigation and household food security. Pearson correlation confirmed that participants were more food secure than non-participants, highlighting the potential positive impact of irrigation on food security.

4.1.4 Household Dietary Diversity Score (HDDS)

The Household Dietary Diversity Score (HDDS) for the previous 24 hours in this study showed that 7.18% of households consumed 1 to 3 food groups (low diversity), 31.61% consumed 4 to 5 food groups (medium diversity), and 61.21% consumed more than 5 food groups (high diversity), based on HDDS thresholds. The data revealed that about 69.74% of irrigation user households and 54.59% of non-user households had a high dietary diversity of 6 or more food groups. Additionally, 24.34% of irrigation users and 37.24% of non-users had a medium dietary diversity of 4 to 5 food groups. Low dietary diversity of less than 3 food groups was observed in only 5.92% of irrigation users and 8.16% of non-users ( Table 3). The average dietary diversity score for the study area was 6.37. Notably, irrigation user households had a higher mean dietary diversity score of 7.11 compared to 5.79 for non-irrigating households. This difference in mean diversity scores between irrigation users and non-users was statistically significant at the 5% significance level, indicating that irrigating households consumed a more diverse range of food groups than non-irrigating households.

4.1.5 Household Food Consumption Score (HFCS)

The Household Food Consumption Score (HFCS) is derived from the Household Dietary Diversity Score (HDDS) by incorporating frequency-weighted data on the consumption of eight distinct food groups: main staples, pulses, vegetables, fruits, meat and fish, milk, sugar, and oil. The HFCS is calculated based on the frequency of consumption of these food groups over a standard 7 day period, with each group assigned a specific weight. The weighted scores are then summed to obtain the HFCS, which is subsequently categorized into food consumption groups using predefined thresholds: 0-28 (food poor), 28.5-42 (borderline), and above 42 (acceptable) according to guidelines from the World Food Programme (WFP) and the International Food Policy Research Institute (IFPRI). The results, as shown in Table 4, indicate that 73% of irrigation users exhibited acceptable food consumption, with 25% classified as borderline and 1.97% with poor food consumption scores. For non-irrigation users, 61.22% had acceptable food consumption, 33.16% were borderline and 5.61% were classified as poor. Overall, the study revealed that 66.38% of households in the area had acceptable food consumption scores, while only 4.02% exhibited poor food consumption. These findings provide insights into the food security status of households, emphasizing the importance of irrigation in enhancing food consumption adequacy.

4.1.6 Challenges in implementing small-scale irrigation

Embarking on the path of implementing small-scale irrigation practices in the study area reveals a nuanced journey marked by economic and environmental challenges. These challenges not only mirror broader issues in regional irrigation development but also shed light on critical hurdles hindering the full realization of the potential benefits for rural households.

While small-scale irrigation holds promise in boosting rural incomes, the survey uncovers key constraints that demand immediate attention. Financial considerations, insufficient irrigation water supply, flaws in marketing systems, limited access to inputs and facilities, and persistent threats from pests and diseases collectively cast a shadow on the effectiveness of small-scale irrigation.

Table 5 provides a snapshot of specific challenges faced by participants in small-scale irrigation, offering a deeper understanding. A noteworthy 30.17% of respondents highlight water scarcity as a predominant issue. Additionally, concerns about administrative hurdles, the absence of an efficient marketing system, pests and diseases, suboptimal productivity in existing schemes, and a lack of awareness about irrigation water management make up 11.78%, 23.85%, 14.08%, 7.75%, and 12.36%, respectively.

These findings underscore the intricate web of challenges associated with small-scale irrigation, emphasizing the need for holistic strategies and interventions. Addressing these constraints is imperative for the sustainable and impactful implementation of small-scale irrigation practices in the region, ensuring enduring success amid the evolving landscape.

4.1.7 Financial landscape: Revealing household expenditure dynamics

Delving into the economic tapestry of households, the annual total expenditure per adult equivalent emerges as a pivotal metric, encapsulating both food and non-food expenditures. This financial gauge serves as the lifeline for households, ensuring the fulfillment of essential calorie intake per adult equivalent daily and meeting other fundamental needs. A robust income source empowers households to secure the required daily food intake. The total annual expenditure is underpinned by income streams stemming from diverse channels, including farm and non-farm activities like livestock and crop production, petty trade, employment, and the sale of natural products.

The comparative analysis of annual total expenditures for irrigation participants and non-participants unfolds intriguing insights. Table 6 shows the mean annual expenditure for irrigation participants stands at 12,096.85 ETB per adult equivalent, surpassing its non-participant counterpart at 9,069.73 ETB. This observed difference undergoes statistical scrutiny through a t-test, revealing its significance at the 1% level. The financial terrain, thus, underscores the noteworthy impact of small-scale irrigation participation on household expenditure patterns, portraying a compelling narrative of economic dynamics in the study area.

Embarking on the econometric voyage, our vessel charts a course through the intricate waters of model specifications and post-estimation tests. The earlier glimpses into descriptive statistics hinted at the prosperity of treatment households engaged in small-scale irrigation. However, for a more profound understanding, we now delve into a rigorous exploration in three acts: delineating model specifications and post-estimation tests, unveiling logit regression results for the treatment-dependent variable, and illuminating the results and discussions of propensity score matching for the outcome variable.

4.2.1 Setting the stage: Model specifications and tests

Before the variables join the model’s ensemble, rigorous tests ensure the model’s integrity. The journey begins with a scrutiny of multicollinearity, dispelling any whispers of interrelated explanatory variables. The variance inflation factor (VIF) stands as the sentinel, ensuring values remain below 10. Contingency coefficients, guardians against association problems, assure a harmonious relationship among dichotomous variables. Heteroskedasticity faces the Brush Pagan test, with results negating its presence. The model emerges fortified, with its foundations unshaken by specification violations.

4.2.2 Figuring out determinants: Logit regression quest

Embarking on small-scale irrigation implementation in our study area unveils a narrative entwined with economic potential and environmental complexities. These challenges, reflecting broader issues in regional irrigation, illuminate critical hurdles hindering the realization of benefits for rural households.

While small-scale irrigation holds promise for rural income upliftment, our survey spotlights immediate constraints. From financial considerations to water scarcity, marketing system glitches, limited inputs access, and persistent pests and diseases, these challenges collectively overshadow small-scale irrigation effectiveness.

These findings emphasize the need for scholarly interventions. Addressing these constraints is imperative for the sustainable success of small-scale irrigation practices in our region, firmly anchoring them in the evolving landscape of agricultural development.

4.2.3 Unveiling agricultural realities: What logit regression reveals

The logistic regression analysis in Table 7, examining factors influencing a binary outcome related to households (hhp), offers valuable insights into various predictors. Age of the household head (agehh) emerges as a significant factor, with a one-unit increase associated with a statistically significant decrease of approximately 0.39 in the log-odds of the binary outcome. Conversely, the gender of the household head (sexhh) and education level (edushh) do not exhibit statistically significant associations with the log-odds, as their respective p-values exceed 0.05. Family size (famsize) shows a marginally significant negative association, indicating that an increase in family size is associated with a decrease of approximately 0.51 in the log-odds of the binary outcome. Livestock units (tlu) exhibit a statistically significant negative association, implying that a one-unit increase in livestock units corresponds to a decrease of approximately 0.21 in the log-odds. The type of land ownership or use (culland), involvement in the workforce (actforce), and credit access (crdtacc) do not display statistically significant associations with the binary outcome.

Additionally, the extention service (exstenserv) shows a statistically significant negative association, suggesting that a one-unit increase is associated with a decrease of approximately 1.40 in the log-odds. The statistically significant negative association between extension service contacts (exstenserv) and irrigation participation in the logit model results suggests that, on average, an increase in extension service contacts is associated with a decrease in the likelihood of participating in irrigation. This unexpected finding prompts consideration of potential interpretations. One plausible explanation could be that the current structure or content of existing agricultural extension programs may not effectively address or promote irrigation practices among the targeted population. The negative coefficient may indicate that, despite engaging with extension services, farmers may not be receiving adequate guidance, information, or support related to irrigation methods or may encounter challenges that hinder their participation.

The importance of agriculture (impagri) exhibits a significant positive association, indicating that a one-unit increase leads to an increase of approximately 3.97 in the log-odds. Remittances received by the household (remitance) do not demonstrate a statistically significant association with the log-odds. The intercept term (-cons) represents the log-odds when all other predictor variables are zero, with a large positive value contributing significantly to the overall prediction of the binary outcome. The LR chi2, Prob > chi2, Log likelihood, and Pseudo R ² values collectively provide information on the model’s goodness of fit, while the statistically significant coefficients and associated p-values aid in evaluating the individual predictors’ significance in explaining the variation in the binary outcome.

4.2.4 Unveiling impact: Propensity score matching in small-scale irrigation analysis

In this section, we delve into the impact analysis of the study, providing an intricate exploration of the process employed to discern the program’s influence through the lens of the propensity score matching model. The journey encompasses the meticulous estimation of propensity scores, identification of the common support region, selection of matching algorithms, conducting balancing tests, and ultimately, calculating the treatment effect.

4.2.5 Estimation of propensity score: Navigating probabilities in irrigation practices

To address the study’s second objective evaluating the impact of small-scale irrigation on household total expenditure per adult equivalent we employed propensity score matching (PSM). This versatile model, hinging on binary variables and predictors, was constructed using a logit regression. The logit regression was chosen for its commonality in propensity score estimation. The process involved considering a myriad of observable characteristics to construct propensity scores, ensuring their exhaustiveness. However, particular attention was paid to exclude characteristics influenced by the treatment, such as household income, from the covariates in the propensity score estimation. The results revealed significant influences of four explanatory variables on small-scale irrigation practices.

4.2.6 Common support condition: Navigating the overlapping realms

The subsequent step in the propensity score matching technique is the establishment of the common support condition. This ensures that only observations within the common support region are considered for matching, while others are excluded from further analysis. The kernel density in Figure 1, estimate visually displayed the distribution of total sample households, highlighting the common support region. The participants predominantly occupied the right side, with non-participants concentrated on the left. This broad area of similarity in propensity scores between treated and control groups are aptly termed the common support region, forming the foundation for subsequent impact assessments.

4.2.7 Harmonizing realms: Crafting common grounds through propensity score matching

The final task involves discarding observations falling outside the common support region, refining the match. Table 8 displays the outcomes, with overall estimated propensity scores ranging from 0.006504 to 0.980515, encapsulating the common support essence. Among participants, scores ranged from 0.010272 to 0.980515, and for non-participants, from 0.006504 to 0.91518. Calibration excludes 16 outliers, bidding adieu to the analysis due to scores beyond common support boundaries.

With 332 resilient households within this realm, the stage is set to predict the impact of small-scale irrigation on annual total expenditure per adult equivalent. The average propensity score, dancing at approximately 0.436782, unveils a narrative of probability. On average, the probability of participation stands at about 0.637321 for sampled households, while non-participants echo a probability of 0.281261. In this statistical symphony, common threads are woven, laying the foundation for a robust impact prediction.

Figure 2 depicts the distributions of propensity scores and the common support region. The lower halves of the histogram display the propensity score distribution of non-participants in small-scale irrigation, while the upper halves showcase the propensity score distribution of small-scale irrigation participant households. The red-colored (Untreated on support) and green-colored (Treated on support) sections indicate observations in the non-participant group and participant group, respectively, that have suitable comparisons. On the other hand, the orange-colored (Treated off support) and blue-colored (Untreated-off support) segments represent observations in the small-scale irrigation participant and non-participant groups that lack a suitable comparison, with their propensity score values approaching one and zero, respectively. This visualization helps illustrate the overlap and common support region essential for an effective propensity score matching analysis.

4.2.8 Precision in propensity: Orchestrating impact with kernel matching

In the realm of propensity score matching, where precision is paramount, the art lies in selecting the right algorithm. A comprehensive evaluation of nearest neighbor, radius matching, and kernel matching unfolds, scrutinizing their efficacy under varied scenarios. Among these, kernel matching emerges as the virtuoso, harmonizing precision and information richness.

As the statistical orchestra’s conductor, the kernel matching algorithm, adorned with a calibrated caliper, takes center stage. Bestowing each treated individual with a weight of one maximizes precision, ensuring a symmetrical impact evaluation. As shown in Table 9 the spotlight then shifts to the Treatment Effect on the Treated (ATT), reaching a crescendo in the analysis. Findings resonate with significance at the 1% probability level, highlighting the transformative influence of small-scale irrigation on household expenditure a substantial 20.85% surge, totaling 2557.74 ETB annually.

In the meticulous dance of controlling pre-intervention disparities, the study unfolds a narrative of positive and statistically significant divergence between participant and non-participant households. This echoes the impactful strides observed in the Panganai irrigation scheme in Bikita district, Zimbabwe, as explored by Bernard Chazovachii in 2012. This study, akin to a well-conducted symphony, underscores the pivotal role of small-scale irrigation in orchestrating prosperity and resilience in rural landscapes.

Table 9 presents a comparison of total expenditure per adult equivalent (AE) between the treated group (those participating in small-scale irrigation) and the control group (non-participants). The unmatched figures indicate that the total expenditure per AE for the treated group is 12,096.85, while for the control group, it is 9,069.73, resulting in a difference of 3,027.12. The standard error (S.E.) associated with this difference is 343.26, and the t-test value is 8.82, indicating a statistically significant difference between the two groups.

In the case of the average treatment effect on the treated (ATT), the total expenditure per AE for the treated group is 12,266.28, and for the control group, it is 9,708.54. The calculated difference is 2,557.74, with a standard error of 446.65 and a t-test value of 5.73. This signifies a statistically significant positive impact of small-scale irrigation participation on household expenditure, reinforcing the economic benefits observed in the propensity score matching analysis. The results suggest that households engaged in small-scale irrigation exhibit higher total expenditures compared to their non-participant counterpart.

4.2.9 Harmony in equivalence: A ballet of covariate balance

In the delicate choreography of impact evaluation, achieving balance is akin to a ballet, where the matched units in treatment and comparison groups pirouette in statistical harmony. The litmus test for this synchronization lies in the t-test, a discerning maestro in determining if the means of covariates, included in the propensity score, orchestrate a statistically comparable performance.

As the curtains rise on the stage of statistical scrutiny, the t-test takes center stage, its preference evident when the gaze is set on the statistical significance of results. Solivas, Ramirez, and Manalo (2007) set the rhythm, guiding the evaluator through this symphony of equivalence. The results unfurl like a seamless ballet, revealing that the constructed treatment and comparison groups dance to the same statistical tune.

The absence of significant difference, as revealed by the t-test, paints a portrait of covariate balance. The narrative woven through this statistical ballet is one of satisfaction wherein the covariate balance criteria stand fulfilled. In this elegantly orchestrated dance of statistical equivalence, the study establishes a foundation of trust in the comparability of treatment and comparison groups, ensuring that the impact analysis unfolds on a stage set with the backdrop of balance.

This study, situated in Raya Kobo Woreda within the North Wollo Zone of the Amhara Regional State, aimed to comprehensively evaluate the influence of small-scale irrigation practices on the annual expenditures of households, encompassing both food and non-food aspects. The research design involved the collection of cross-sectional data in 2022, where 348 farmers were selected through a multistage stratified random sampling approach. Among this sample, 152 households actively participated in small-scale irrigation, while the remaining 196 did not engage in such practices. The analytical framework incorporated descriptive statistics, econometric methods, and propensity score matching to derive meaningful insights from the collected data.

The primary data, acquired through interviews, underwent rigorous statistical analysis, employing measures such as mean, standard deviation, percentage, T-test, chi-square test, and t-test. The study particularly focused on food security, revealing that 63.51% of households overall maintained food security, with noteworthy distinctions observed between irrigation participants (71.71% food secure) and non-participants (57.14% food secure). The investigation also highlighted significant variations in diet diversity and food consumption scores between irrigators and non-irrigators.

Identified constraints in small-scale irrigation practices encompassed challenges such as water shortages, administrative issues, marketing problems, pests and diseases, and low scheme productivity. The logit model scrutinized eleven variables, ultimately pinpointing four significant factors negatively affecting irrigation practices: the age of the household head, extension service, TLU, and improved agricultural technologies.

Furthermore, propensity score matching techniques were employed to assess the financial impact, revealing that small-scale irrigation participants annually spent 2557.74 ETB more than their non-participating counterparts, indicative of a positive influence on farm income. To ensure the robustness of these findings against hidden biases, average treatment effects were scrutinized using the Rosenbaum (2002) Bounding approach under varying gamma values in sensitivity analysis, affirming the resilience of the results against unobservable selection bias.

To promote sustainable small-scale irrigation practices and enhance food security, tailored recommendations proposed for policymakers, development practitioners, and farmers.

Policymakers are urged to prioritize investments in irrigation infrastructure, allocating substantial budgets for construction and maintenance. Additionally, enacting policies to address post-harvest challenges and formulating supportive regulations for small-scale irrigation practices are essential.

For development practitioners, the focus should be on enhancing extension services, conducting targeted capacity-building programs, and promoting collaborative initiatives among farmers. Specifically, increased engagement through workshops and field visits, specialized training programs for farmers, and facilitation of cooperative efforts recommended.

Farmers are encouraged to actively participate in training programs, establish irrigation cooperatives to collectively address challenges, and adopt sustainable irrigation practices. This holistic approach ensures that each stakeholder group plays a vital role in fostering sustainable small-scale irrigation, contributing positively to farmers’ livelihoods and overall food security.