The Scientific World Journal

Wiley
The Scientific World Journal

Volume 2024, Article ID 9982796, 16 pages
https://doi.org/10.1155/2024/9982796
Research Article
Influence of Small-Scale Irrigation on Livelihoods of Rural
Farm Households in the Case of Legehida District, Ethiopia
Awol Hussen1 and Arebu Hussen 2
1
Department of Agricultural Economics, Wollo University, Dessie, Ethiopia
2
Department of Plant Science, Mekdela Amba University, Tuluawlia, Ethiopia
Correspondence should be addressed to Arebu Hussen; [email protected]
Received 8 December 2023; Revised 29 April 2024; Accepted 2 May 2024; Published 23 May 2024
Academic Editor: Difang Huang
Copyright © 2024 Awol Hussen and Arebu Hussen. Tis is an open access article distributed under the Creative Commons
Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is
properly cited.
Irrigation development, particularly small-scale irrigation, is one of the most important projects for improving agricultural
productivity in a country’s rural communities. Te extent to which small-scale irrigation has improved household livelihoods in
Ethiopia’s rural areas is not widely recognized. As a result, research on the infuence of small-scale irrigation on farmers’
livelihoods in the Legehida district will be sought. Te study took a “with” and “without” strategy, comparing farmers who used
irrigation against those who did not. For analysis, both quantitative and qualitative data were employed. Te survey’s respondents
were chosen using a random sample approach from both irrigation users and nonuser households. Quantitative data for the study
were collected from randomly selected 241 farm households, of which 113 were users and 128 were nonusers, using a semi-
structured questionnaire. Accordingly, the propensity score matching model was employed to examine the impacts of small-scale
irrigation on farmers’ livelihoods. Te logit model result indicates that cultivated land size, of-farm income, education level,
family size, dependency ratio, total livestock unit, and distance to the nearest agricultural extension ofce/FTC are determinant
factors in determining whether to practice irrigation when other factors remain constant. Te impact of irrigation on
a household’s income and food security (in terms of daily calorie intake) was evaluated using a propensity score matching model.
Te result shows that a positive and signifcant impact on farmers who use small-scale irrigation has increased the daily calorie
intake and annual income of households by 244.162 kilocalories and 5234.258 ETB, respectively, as compared to nonirrigation
users. Tis shows that households that participate in small-scale irrigation activities have a higher annual income and food security
status than comparable groups. In general, the study recommends that to reduce food insecurity and the socioeconomic problems
of rural households, irrigation farming is one of the viable solutions; therefore, the government and nongovernmental orga-
nizations should extensively focus on the enhancement of small-scale irrigation infrastructure, policies, strategies, and extension
services to increase productivity, income, and livelihood improvement in rural households.
1. Introduction raw material supply, 28% of government tax income, and

85% of export profts in 2011. However, it is largely based on
1.1. Background of the Study. Ethiopia is one of the world’s subsistence farming due to small and fragmented land-
poorest countries, with around 29.2% of its people living holdings, reliance on natural factors of production, envi-
below the poverty line [1]. Te majority of Ethiopians reside ronmental degradation, population growth, limited access to
in rural regions, where smallholder farmers generate around new agricultural technologies, traditional methods of cul-
95% of agricultural products [2]. Agriculture is the backbone tivation, and a lack of institutional support services [3].
of the Ethiopian economy, employing around 80% of the To address the issue of subsistence farming, economic
people directly or indirectly. Tis suggests that agriculture is actors created a national strategic plan in 1991 called Ag-
the leading contributor to GDP. Agriculture, for example, ricultural Development Led Industrialization (ADLI), which
supplied 40% of national GDP, 80% of employment, 70% of focuses on irrigation, cooperative societies, and agricultural
8086, 2024, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1155/2024/9982796 by ENDRIS ASSEN EBRAHIM - EBMG ACCESS - ETHIOPIA , Wiley Online Library on [28/07/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2 Te Scientifc World Journal
technologies to meet food demand and bring socioeconomic country’s development policy to boost and stabilize food
development to the country. As they lack agriculture in- production in the country. Small-scale irrigation develop-
surance and are less vulnerable to economic shocks, low- ment has been acknowledged as a key strategy for stimu-
income rural families rely heavily on government subsidies lating economic growth and household rural development.
to reduce potential losses [4]. One of the policies outlined in
this approach is small-scale irrigation development. Te
success of Asian countries such as China in accommodating 1.2. Statement of the Problem. Irrigation boosts agricultural
a growing population, achieving rapid economic growth, productivity by improving crop yields and allowing farmers
and increasing employment through irrigated agriculture in to increase cropping intensity and switch to high-value crops
the 1960s and 1970s has prompted the Ethiopian govern- [15]. Similarly, according to Lipton et al. [16], irrigated
ment to prioritize the development of irrigation schemes [5]. agriculture can reduce poverty by increasing production and
According to Bao and Huang [6], there is a greater income, as well as lowering food prices, which helps very
likelihood of acquiring loans from the FinTech business for poor households meet basic needs by improving overall
borrowers who reside in locations where there were more economic welfare, protecting against crop loss due to in-
COVID cases during the epidemic, have low incomes, and sufcient rainwater supplies, encouraging greater use of
are unemployed. Based on this, the federal and provincial yield-enhancing farm inputs, and creating additional em-
governments, in collaboration with other international and ployment, all of which allow people to move out of poverty.
local NGOs, have considerably aided rural farmers in par- Irrigation development, particularly small-scale irriga-
ticipating in and using irrigated farming. As a result, irri- tion, is one of the most important projects for improving
gated acreage, irrigation production, and the number of agricultural productivity in a country’s rural communities. It
irrigation farmers in the nation rose by up to 80% between assists poor households in overcoming rainfall shortages by
1990 and 2010 [7]. providing optimal water for irrigation agriculture and
Irrigation is one of the agricultural methods that is livestock, strengthening the foundation for sustainable ag-
described as the man-made application of water to ensure riculture, increasing food security in poor communities
double cropping and a consistent supply of water in loca- through irrigated agriculture, and contributing to human
tions where rainfall is inconsistent [8]. Small-scale irrigation nutrition improvement [9]. According to MoARD [2], the
is another key intervention to boost agricultural productivity importance of irrigation development, particularly at the
in rural areas of a country. Tis assists farmers in over- smallholder level, is critical for increasing productivity,
coming rainfall constraints by providing a steady supply of achieving food self-sufciency, and ensuring food security at
water for agriculture and livestock production [9]. Using the family and national levels.
small-scale irrigation is crucial to increasing production Traditional farm implements, unimproved crops and
during times of low rainfall. Trough higher incomes, food fertilizers, and inadequate animal breeds, as well as water use
security, meeting social needs, and reducing poverty, it can and availability, are important impediments to Ethiopian
help the rural population’s overall quality of life [10, 11]. agricultural growth. Te country’s capacity to sustain agri-
Irrigation development has been proposed as a crucial culture through irrigation development has been limited [17].
technique for increasing agricultural production and stim- Te Legehida district is one of the most drought-prone
ulating economic growth [12]. Irrigation in Ethiopia helps and erratic rainfall regions. Due to this, the government is
boost farmer income, household resilience, and bufer implementing diferent agricultural development programs
livelihoods against shocks and stresses by producing higher- to achieve livelihoods in rural farm households. Among
value crops for market and harvesting more than once a year. these programs, small-scale irrigation development is pri-
As a result, they were able to accumulate assets, purchase marily undertaken by the government, but due to economic
additional food and nonfood household items, educate their scarcity, it is not widely implemented.
children, and reinvest in improving their output by pur- Irrigation is signifcant for improving agricultural pro-
chasing farm supplies or cattle. However, advantages are duction and food security. However, the farmers in the study
dispersed relatively unevenly across families [13]. Large- area do not have enough knowledge to determine to what
scale irrigation programs and other associated technolo- extent using irrigation is vital to boost agricultural production,
gies are well recognized in Ethiopia, and the country’s household income, food security, and eradicate poverty;
governments actively promote them. Similarly, in recent rather, they depend on rainfall. Accordingly, most rural people
years, because small-scale irrigation systems are fnancially are highly vulnerable to poverty, socioeconomic problems,
viable, investment cost recovery operation, maintenance and food insecurity. As a result, in the district of Legehida
costs, and the ability to replicate investments, Ethiopia’s where this study was conducted, in-depth comparative
irrigation program has shifted from large-scale irrigation to analysis studies are scarce, due to diferent factors such as
small-scale irrigation [14]. It is feasible to completely un- irrigation technology, scheme and plot size, infrastructure, low
derstand the heterogeneity impacted by COVID-19 policies income, limited extension access, and marketing opportuni-
since rural households’ income levels have a signifcant ties. Terefore, the infuence of small-scale irrigation on
infuence on the degree and intensity of agricultural in- livelihood and determinant factors of household participation
tensifcation and operation [4]. Regional states and non- in the irrigation activity creates an empirical question in the
governmental organizations (NGOs) are pushing for the study area. Hence, the study was conducted to reveal the
construction of small-scale irrigation schemes by the seriousness of the problem and fll the gaps.
Te Scientifc World Journal 3
Tus, this study aimed to determine the factors afecting Among these which is cultivated, 27607 ha, covered with
the participation of small-scale irrigation activities in rural forest 1320 ha, bush, and shrubs 30,725 ha, grazing
farm households and to examine the infuence of small-scale 12,341 ha, and 2,992.02 ha of land in the district. Conse-
irrigation on household income and food security. quently, the district has a total population of 78845 that were
included both male 39,182 and female 39,659, respectively.
People living in the study region rely mostly on agriculture
1.3. Research Questions. Te study addressed the following for a living, and the area is noted for its low productivity owing
research questions to achieve the above objectives related to to soil degradation. Crop and animal production systems are
the study on income, food security, and livelihood of rural the primary source of income for the majority of the district’s
farm households on diferent determinant factors. residents. Among the crops planted are tef, barley, wheat,
(1) What are the elements infuencing farmer partici- bean, malt barley, fnger millet, haricot bean, and chickpea.
pation in small-scale irrigation? Diferent crops and fruits, such as tomato, potato, onion,
(2) How does irrigation afect food security signifcantly? lettuce, carrot, garlic, and apple, are grown in the region with
irrigation during both the wet and dry seasons.
(3) What are the main income sources for rural farm Irrigation was performed in the research region by
households? various methods such as microdams (earth and concrete),
(4) To what degree do irrigation programs afect the river diversion, Acre ware dams, shallow wells, and
livelihood of rural agricultural households? ponds [2].
1.4. Scope and Limitations of the Study. Tis study was un- 2.2. Data Source and Methods of Data Collection. Te in-
dertaken to evaluate the infuence of small-scale irrigation vestigators should employ both primary and secondary data
schemes on the livelihoods of rural farm households in the sources to appropriately meet the study goals. To capture all
Legehida district. Tis research focuses on two small-scale information on homes, primary data sources such as interviews,
irrigation schemes and the concept of livelihood is very broad; focus group discussions, and structured questionnaires with
therefore, it is not possible to capture all aspects in a single both open-ended and closed-ended questions might be
study due to limitations imposed by time and fnancial re- employed. Secondary data sources would be gathered from
sources and other related problems. Terefore, this study various ofces using either published or unpublished records.
examines only the infuence of small-scale irrigation schemes Agricultural agencies, environmental protection ofces, and the
on users’ income and food security. Furthermore, the in- district’s Ofce of Water and Mineral Energy might all be
formation was acquired using a semistructured questionnaire included in a comprehensive assessment of the research.
survey, and the quality of the information was dependent on
respondents’ willingness, knowledge, and recall capacity.
2.2.1. Primary Data. Structured surveys, key informant
However, every attempt was made to collect valid data by
interviews, and focus group discussions were used to obtain
convincing farm households of the study’s aims.
primary data. Before data collection, all enumerators were
instructed on data collection methodologies, survey content,
1.5. Signifcance of the Study. Te attainment of the and how to approach interviewees/best approach strategies.
abovementioned objectives is an essential instrument for Te questionnaire also covered household wealth mea-
agricultural growth; it increases the efciency of small-scale surements such as household assets, livestock, and house-
irrigation and family food security. Tis study is noteworthy hold agricultural production activities, as well as income
because it raises household awareness of the factors that quantities and sources. Te same questionnaire was utilized
determine small-scale irrigation involvement and its impact for both irrigators and nonirrigators but with additional
on food security. In general, the research’s importance stems parts to cover particular irrigation-related topics.
from its endeavour to provide actual information on the
overall concerns of small-scale irrigation development in the 2.2.2. Focus Group Discussion. One of the qualitative data
study region, as well as its infuence on increased income and gathering approaches for this study was focus group dis-
better rural farm household livelihoods. cussions (FGD) with farmers. FGD obtained qualitative data
through discussions with development agents, district ag-
2. Research Methodology ricultural and rural development ofce irrigation specialists,
and irrigating and nonirrigating farmers. Each focus group
2.1. Description of the Study Area. Legehida district is one of
consisted of nine people from the same village in the re-
the rural districts of the Amhara National Regional State. It
search region. FGD was used to create data that supple-
is in the northeastern part of the Amhara National Regional
mented the structured questionnaire by explaining and
State, South Wollo Zone, Ethiopia, lying between 10° 49′ 48″
addressing concerns raised by quantitative data.
N latitude and between 39°22′ 12″ E longitude. Te district
has an altitude that ranges from 1644 to 3409 m.a.s.l. and an
annual rainfall of 1250 mm with the minimum and maxi- 2.2.3. Key Informant Interview. Individuals in the research
mum temperatures of 10 and 22, respectively. Te total land who were deemed competent and experienced in irrigation
area of the Woreda is estimated to be about 516,074 ha. operations, rural livelihoods, and the socioeconomic
situations of the community were selected and interviewed. groups, variables that are likely to afect both the treatment
Fifteen key informants were questioned, four of whom were assignment and the result should be included. We want to
extension ofcers and eleven of whom were farmers with reduce the possibility of confounding bias by balancing the
knowledge of the scheme’s farming operations. distribution of variables between the groups through indi-
vidual matching based on propensity scores.
To execute the PSM, one must complete fve steps: es-
2.3. Sample Size and Sampling Technique. Two small-scale timate propensity scores, match treatment and control
irrigation (SSI) schemes, Berberti and Siba, were purpose- groups, verify common support conditions, assess matching
fully chosen based on population concentration in the quality, and perform sensitivity analysis [19]. However, we
schemes, the type of SSI they utilized, and the suggestions of have combined the stages and provided them as follows for
the district’s Agricultural and Rural Development Experts. A clarity and simplicity: the frst step in the PSM technique is
random sample of families should be drawn from each ir- the estimation of propensity scores. Tere are two options to
rigation project and peasant organization. Household heads consider when estimating the propensity score. Te frst one
from the two kebeles with their diferent irrigation schemes relates to the estimated model and the second one to the
were identifed and classifed into two groups: users and variables that will be part of this model [19]. In terms of
nonusers. A list of irrigating farmers was collected from model selection, a probit/logit model has been used in
scheme management, and farmers were stratifed according several studies to evaluate impact analysis to calculate
to their position on the scheme to guarantee a fair repre- propensity scores [20]. Using a probit or logit model yields
sentation of irrigation users. Using a simple random sam- comparable fndings when evaluating an individual’s pro-
pling approach, a sample of (113) irrigating farmers and pensity score to be an adopter or nonadopter [21]. None-
(128) nonirrigators in the same geographic region were theless, the logit model was used in this work to estimate the
picked from these substrata (Table 1). Because the district is propensity score of the studied homes because of its sim-
homogenous in terms of climate, resource endowment, and plicity. Users of irrigation receive a value of 1, while non-
other criteria relevant to the research, this sample size is users receive a value of 0. Te mathematical formulation of
believed to refect the population. Te study used a simpli- the logit model is as follows:
fed formula presented by Yamane [18], statistically calcu-
lated at the 95% confdence level, degree of variability � 0.05, 1
pi � F(zi) � F􏼐α + 􏽘βi xi 􏼑 � − (α+Σβixi)
, (2)
and level of precision of 95%, to include the abovementioned 1+e
sample household size.
where e is the base of the natural logarithm. Xi represents the
N ith independent variable. Pi is the probability that
n� , (1)
1 + N(e)2 a household belongs to the participant given. βi is the pa-
rameter to be estimated. (Pi) households will be participants
where n is the sample size, N is the population size (total (irrigators). (1-Pi) is the probability that a household will be
household size), and e is the level of precision. Based on the nonparticipants (nonirrigators).
number of total households in the sampling frame, the for-
mula equated and reached a minimum of 241 respondents ezi
pi � , (3)
from the two selected kebeles. 1 + ezi
where Pi is the probability of participation in small-scale
2.4. Data Analysis Techniques. Te data could be analyzed by irrigation and 1-Pi is the probability of the household be-
using both descriptive and inferential statistical tools for the longing to nonparticipants (nonirrigators) that is
STATA software version 14 that would be used. Te de-
1
scriptive analyses are mean, median, and standard deviation, 1 − pi 􏼁 � ,
and the econometric data are analyzed using the propensity 1 + ezi
score matching technique.
Pi 1 + ezi z
􏼠 􏼡 �􏼠 􏼡 � e i,
1 − pi 1 + e− zi
2.5. Specifcation of the Model
pi 1 + ezi
􏼡 � e(
α+Σβi Xi )
2.5.1. Propensity Score Matching (PSM) Methods. Te pro- 􏼠 􏼡 �􏼠 , (4)
1 − pi 1 + e− zi
pensity score matching method is a statistical technique used
to estimate the causal efect of a treatment or intervention in
pi
observational studies. Te study should be applied in a PSM Zi � ln􏼠 􏼡 � α + β1 X 1 + β 2 X 2 . . . . . . βm X m ,
technique, which has been widely applied as an impact 1 − pi
evaluation instrument in the absence of baseline survey data m
for impact evaluation. zi � α + 􏽘 βi χ i + υi .
Te justifcation for incorporating particular factors into i�1
the propensity score matching model is based on theoretical
understanding, empirical data, and subject matter expertise. In PSM, many matching algorithms may be employed to
To improve the comparability of the treated and control estimate the treatment impact on the treated. However, the
Table 1: Te target sample is taken from the population of each kebele.

Total sample size population
Total number of Total number of
Kebele Total population Number Number
irrigation users nonirrigation users Total sample
of irrigation users of nonirrigation users
Siba 320 146 174 66 74 140
Berberti 290 106 184 47 54 101
Total 610 252 358 113 128 241
most commonly used matching algorithms in PSM are member’s input, giving greater weight to those comparators
nearest neighbour matching, radius matching, and kernel that provide a better match. Te most frequent strategy is to
matching [22]. Tese matching approaches employ various utilize the normal distribution (with a mean of zero) as
ways of comparing the treated and control groups to as- a kernel, with the weight assigned to a specifc comparator
certain the average efect of a specifc participation irrigation proportional to the frequency of the distribution for the
activity or intervention. diference in scores observed [25].
According to Caliendo and Kopeinig [19], one disad-
vantage of this technique is that potentially poor matches are
2.6. Matching Estimators. Te program evaluator’s main
employed because the estimator contains comparator ob-
objective after estimating the propensity scores is to fnd
servations for all treatment observations. As a result, ac-
a suitable matching estimator. In theory, there exist various
curate enforcement of the common support condition is
matching estimators. Only the most commonly used
critical for the kernel matching approach. One practical
matching estimators are presented. Tey are as follows.
problem with its use is that it is not always evident how to
establish tolerance. According to Mendola [26], 0.25
2.6.1. Nearest Neighbour (NN) Matching. It is the simplest bandwidth kernel matching is the most commonly utilized.
basic matching estimator. In NN matching, a person from Te question remains as to how and which approach to
a comparison group is chosen as a matching partner for use. Obviously, there is no single answer to this topic. Te
a treated person who has the highest propensity score [23]. nature of the available dataset infuences the selection of
NN matching can be performed with or without replacement a matching estimator [25]. When the distribution of the
choices. In the case of NN matching with replacement, propensity score between the comparison and treatment
a comparison person can be matched to more than one groups overlaps signifcantly, most matching algorithms
treatment individual, resulting in higher match quality but produce comparable fndings [22].
lower estimate precision. Matching without replacement
increases bias, although it may enhance estimate precision. 2.7. Region of Common Support Condition. Figure 1 illustrates
Finding a suitable match via matching without replacement that by imposing a common support condition, any combi-
might be difcult in circumstances when the treatment and nation of features identifed in the treatment group can also be
comparison units are highly diverse [23]. observed in the control group [25]. Te common support
region is defned as the area containing the treatment and
2.6.2. Calliper Matching. As previously discussed, NN control group families’ least and greatest propensity scores,
matching is vulnerable to poor matches when the nearest respectively. It necessitates the deletion of all observations
neighbour is far away. To solve this difculty, the researcher with propensity scores less than or more than the minimum
employs a second alternative matching algorithm known as and maximum of treatment and control, respectively [22].
calliper matching. Caliendo and Kopeinig [19] defne cal-
lipers matching as selecting a person from the comparison 2.8. Testing the Matching Quality. Te balance test is a crucial
group as a matching partner for a treated individual who consideration to make while doing PSM. While variations in
falls within a certain calliper (propensity score range) and is variables are to be expected before matching, they should be
closest in terms of the propensity score. If the neighbour- avoided afterward. Te PSM’s principal function is to act as
hood dimension is confgured to be very tiny, likely, certain a balancing technique for covariates between the two groups.
treated units will not be matched because the neighbour- Propensity score estimation performance is thus measured
hood lacks a control unit. Conversely, the smaller the by the resultant balance rather than the ft of the models used
neighbourhood, the higher the quality of the matches [24]. to generate the estimated propensity scores [28].
One issue with calliper matching is determining which
option for the tolerance level is the most realistic.
2.9. Variable Modifcation
2.6.3. Kernel Matching. Tis is another matching approach 2.9.1. Dependent Variables. Participation in the small-scale
in which all treated units are matched with a weighted irrigation system with families having access to irrigation
average of all controls, with weights inversely proportional and others without access to the irrigation scheme in the
to the distance between the treatment and control pro- research region is the dependent variable of the frst stage of
pensity scores [24]. Kernel weights each comparison group this study.
Density Density of explained as follows (Table 2) that have a signifcant impact

comparison Density of treatment on annual income and food (daily calorie intake per adult
of households equivalent).
3. Results and Discussion

Tis chapter summarizes and examines the fndings on the
socioeconomic features of rural farm households, as well as
the infuence of small-scale irrigation on farmer livelihood.
As a result of the descriptive analysis of chosen demographic
and socioeconomic features of sample homes, the t-test is
used to evaluate the signifcance of continuous variables,
0 while the chi-square test is used to assess the signifcance of
Figure 1: Region of common support condition. Source [27].
possible dummy variables (Appendix 8).
2.9.2. Outcome Variable. To evaluate the average treatment 3.1. Age of Household Heads. Te average age of the sample
efect, food calorie intake per adult equivalent and annual household heads was 44.16 years. Te average ages of irri-
income of rural farm households are proxy variables to gation users and nonusers were 42.079 and 46 years, re-
examine the impacts of household food security and income. spectively (Table 3). In terms of age, the statistical analysis
Total annual household income: this variable is a con- revealed a substantial diference between irrigation users and
tinuous variable that represents the total annual household nonusers of homes. Dowd et al.’s [31] study stated that the
income in Ethiopian Birr. Evidence [29, 30] indicates that age structure of a household is crucial for performing
this variable has a favorable and signifcant impact on policies and engaging in tasks properly.
farmers’ engagement in small-scale irrigation practices.
Farmers are more likely to participate in small-scale irri-
gation activities if their overall family income is higher. Tis 3.2. Family Size of Households. According to the survey
might be the case if farmers with greater incomes can easily results, the average family size of sampled irrigation users
pay irrigation costs compared to lower-income households. and nonirrigation users was 5.088 and 4.039, respectively
As a result, this characteristic was theorized to infuence (Table 3). However, the t-test results revealed that there was
farmers’ decision to participate in small-scale irrigation. Te a signifcant mean diference between the two groups in
primary irrigation indicator refects the amount of money terms of mean family size between irrigation users and
earned (in cash) by the farmer or any household member nonirrigation users of the tested household. Te fndings
from farm production, nonfarm revenue, and of-farm in- revealed that the larger the family size, the more productive
come. It is determined by the amount earned each year in the workforce and the lower the labor cost for farm output.
birr from such sources.
Daily calorie intake per adult equivalent: daily calorie
consumption per adult equivalent is an essential statistic for 3.3. Size of Cultivated Land Size. Land is well understood to
measuring family food security. Individual dietary intake be one of the most important components in manufacturing.
may be assessed using a variety of approaches, including (i) As a result, it was projected that the likelihood of a family
24-hour recall; (ii) food frequency surveys; and (iii) food participating in irrigation activities would grow with farm size
records maintained by people or an observer. A reference holding and decrease with nonirrigation users. According to
period must be used in all dietary intake strategies. While the survey results, irrigation users owned more acreage than
some approaches rely on participant recollection (24-hour nonirrigation users, and this diference is statistically sig-
recall and food frequency questionnaire), others rely on the nifcant at a less than 1% probability level. As a result, the
research participant, a proxy, or an observer noting meals as average total cultivated area for irrigation and nonirrigation
they are consumed. To assess family food security, both users might be 1.221 hectares and 0.832 hectares, respectively
irrigation users and nonuser households provided in- (Table 3). As a result of this fnding, total cultivated land sizes
formation on the types and amounts of food consumed by difered signifcantly between users and nonusers.
their families in the 10 days preceding the survey. Following
that, the total amount of food ingested was totalled and
converted into calories (kcal) using the conversion factor of 3.4. Total Livestock Unit. Farm animals play a signifcant
each food item consumed each day. part in rural livelihood. Tey provide draught power, protein
supplements, status, income, animal dung for organic fer-
tilizer, and transportation. It was quantifed in terms of
2.9.3. Independent Variables. Te independent variables tropical livestock units, as in many previous similar research
hypothesized to infuence household participation in small- studies [32]. Te combined mean of respondents’ livestock
scale irrigation are the combined efects of various factors holdings was 5.568 TLU. Te mean comparison test scores
such as physical, socioeconomic, and institutional factors for both users and nonusers were 6.647 and 4.615,
Table 2: Types of variable, measurement, and hypothesis expectation.
Te Scientifc World Journal
Types of variable Description Measurement Hypothesis expectation

Dependent variable
PSSI/participation in small-scale
Dummy, the status of participation in small-scale irrigation 1 � Participation on irrigation +
irrigation
Outcome variable
Continuous, income earned from irrigation production and livestock
Income ETB
rearing
Food security Continuous, food calorie intake per day AE
Independent/explanatory variable
AHH Continuous, the age of household head of years old Year −
1 � male
SHH Dummy, gender of household head +
0 � otherwise
FSHH Continuous, family size of the household Number +
DPRHH Continuous, dependency ratio of household Determine in age +/−
1 � Literate
EDSHH Dummy, the educational status of a household +
0 � Illiterate
TCLHH Continuous, total cultivated land holding size Hectare +
1 � Fertile
SFHH Dummy, soil fertility status of land +
0 � Otherwise
TLUHH Continuous, total livestock holding size TLU +
1 � Participation in of-farm
EOFF Dummy, engage in of-farm +
0 � Otherwise
DMKTHH Continuous, distance of market to household home Km − /+
DFTCHH Continuous, distance of FTC to household home Km −
TRIIRHH Dummy, training on the issue of irrigation 1 � Access to train and 0 � Otherwise +
1 � Yes
ACRSHH Dummy, the level of access to credit service +
0 � Otherwise
7
Table 3: Descriptive statistics for continuous variables of the household characteristics.

Irrigation user Nonirrigation user Total
Variables t-value
Mean Standard deviation Mean Stade Mean Stade
AHH 42.079 9.629 46 10.274 44.161 10.147 − 3.044∗
FASIZE 5.088 1.655 4.039 1.513 4.531 1.663 − 5.140∗
DPRHH 4.747 1.507 4.559 1.472 4.647 1.489 − 0.980
TLU 6.647 1.828 4.615 1.722 5.568 2.040 − 8.876∗
SIZECUL 1.221 0.440 0.832 0.294 1.014 0.417 − 8.147∗
DMKTHH 7.146 2.406 7.617 2.595 7.396 2.514 1.454
DFTCHH 1.741 0.970 2.027 1.300 1.893 1.163 1.917
∗∗∗ ∗∗
Te signifcance level at 1%, 5%, and 10% was expressed in asterisks , , and ∗ , respectively.
respectively (Table 3). Te comparison test results revealed agricultural technologies more readily. As a result, farmers
a statistically signifcant diference in total livestock holdings employ irrigation to boost their revenue and improve their
between irrigation users and nonuser households. food security. Access to quality education and skill devel-
opment opportunities is critical for rural households to
enhance their income-earning potential.
3.5. Distance to Agricultural Ofce (FTC). Tis is the distance
in kilometers between the residential village and the nearest
agricultural extension ofce. Households with access to 3.8. Of-Farm Job Participation of Households. Te survey
agricultural extension services through development agents results indicated that 55.60% of the selected respondents had
may have a choice of alternatives for adopting various of-farm income involvement and 44.40% did not have of-
contemporary and traditional technologies that may help farm income participation. In terms of irrigation access, 37%
them increase agricultural productivity. Te average dis- of irrigation users and 18.67% of nonusers had of-farm jobs
tance to the agricultural ofce/FTC was (1.741) and (2.027), on treated farms, while 10.30% of irrigation users and
respectively (Table 3). Similarly, the distance between ag- 34.02% of nonusers had no of-farm jobs on control. As
ricultural ofces//FTC makes a statistically signifcant dif- a consequence, the chi-square test (42.760) and (pr < 0.0001)
ference between irrigation and nonirrigation users of rural results demonstrated a substantial relationship between ir-
farm households. rigation usage and access to of-farm income.
3.6. Sex of Household Head. Among the demographic factors, 3.9. Soil Fertility Status. According to the survey results, the
household head sex appeared to have a signifcant relation- chi-square test of soil fertility 30.309 and (Pr < 0.0001) ex-
ship with small-scale irrigation activity participation. Te plains the positive association/infuence on irrigation ac-
survey results revealed that the chi-square test was (12.704) tivity participation. Households with fertile land may
and (pr < 0.0001), indicating that the gender of the household generate a higher yield; moreover, if the households have
head is positively linked with the usage of small-scale irri- access to irrigation, the production may grow signifcantly
gation. It is worth noting that, when all other variables are for both domestic use and sale. Tis condition motivates
held constant, male-headed families are shown to be more people to engage in small-scale irrigation, which increases
likely than female-headed households to employ small-scale income and improves food security for rural communities.
irrigation. Tis fnding supports the premise that small-scale
irrigation is appealing to male-headed families. Tis dem- 3.10. Training on Irrigation Farm. For the complete obser-
onstrates that male farmers are more likely to participate in vation, about 57.26 percent of irrigation users received ir-
irrigation technology (adjusting for other variables in Ta- rigation instruction, whereas approximately 42.74% of
ble 4). Dillon [33] discovered, similar to this study, that the nonirrigation users received no irrigation training (Table 4).
gender of the head is a variable that statistically and sub- In terms of involvement in training on the chi-square result,
stantially explains involvement in irrigated agriculture. there was a highly signifcant correlation between irrigation
and nonirrigation users. Tis variable’s outcome suggests
that irrigation user families received more irrigation in-
3.7. Education Level of Household. It is commonly assumed struction than nonirrigation user households.
that the education level of household heads has a signifcant
impact on the adoption of irrigation technology and the
improvement of agricultural output. Household heads with 3.11. Access to Credit Service of Households. It is the primary
a higher education level were found to be more educated in source of money for impoverished farmers to acquire input
irrigating homes. Te chi-square test (39.430) found a sig- and, ultimately, to embrace new technologies. Amhara
nifcant connection between the education level and irri- Credit and Saving Institution (ACSI) and credit cooperatives
gation use at the 5% signifcance level. Tis suggests that are the primary sources of credit in the research region.
household heads with more years of training are more likely According to the study results, 60.17% of the sample homes
to see the benefts of irrigation and to embrace and employ utilize credit. In terms of irrigation access, the survey results
Table 4: Descriptive statistics of dummy variables of household characteristics.

0 � Otherwise 24 9.95 79 32.78 103 42.74 40.185
1 � Yes 89 36.93 56 23.23 145 60.17
Access to credit service 30.695
0 � No 24 9.95 72 29.87 96 39.83
Variable Category of variables Chi2-test/x2
Frequency % Frequency % Frequency %
Male 96 39.83 83 34.44 179 74.27
Sex hhh 12.704
Female 17 7.05 45 18.67 62 25.73
1 � Literate 87 36.09 47 19.50 134 55.60
Education level hhh 39.430
0 � Illiterate 26 10.78 81 30.61 107 44.40
1 � Fertile 88 36.51 55 22.82 143 59.34
Soil fertility land 30.309
0 � Unfertile 25 10.37 73 30.29 98 40.66
1 � Participate 88 36.51 46 19.08 134 55.60
Of-farm 42.760
0 � Otherwise 25 10.37 82 34.02 107 44.40
Training irrigation 1 � Access to train 89 36.93 49 20.33 138 57.26
indicated that 37.34% of irrigation users and 22.82% of 3.12. Estimation Procedures of Econometric Models. Te
nonusers had used credit, while 9.95% of irrigation users and propensity score is estimated as the initial step in de-
29.87% of nonusers did not. Improving access to fnancial termining the treatment impact. To obtain these propensity
services, including microfnance, credit, insurance, and scores for binary treatment cases, the logit or probit models
savings groups, can empower rural households and enable can be utilized since both models often produce similar
them to invest in income-generating activities. Te chi- results when estimating the likelihood of involvement vs.
square test found a signifcant link between loan availabil- nonparticipation [19].
ity and irrigation use. Tis fnding is controversial for Alemu Te model was calculated using the propensity score
[34]; there was no signifcant association between irrigation matching approach developed by Leuven and Sianesi [36]
participation and access to credit services. Te results of using the STATA 14.0 computer program. Data from the
a research by Alexander et al. [35] indicated that many small two groups, namely, participant and nonparticipant homes,
businesses are likely to collapse in the absence of various were pooled in the estimate so that the dependent variable
forms of fnancial support. has a value of 1 if the household has access to irrigation and
0 otherwise. Te variance infation factor (VIF) was used to
determine if there was a strong multicollinearity problem
3.11.1. Household Income and Food Security. Respondents in among the explanatory variables. Before running the logit
the research region rely on agriculture for a living, work, model to estimate propensity scores, the explanatory vari-
money creation, food, and nonfood production and con- ables were tested for the presence of severe multicollinearity
sumption. In the research region, agricultural revenue (ir- using several tests.
rigated and rain-fed crops) and of-farm and nonfarm To discover the problem of multicollinearity among
income were the main sources of income (Appendix 6). As continuous explanatory variables, the variance infation
mentioned in the research region, the survey fndings show factor (VIF) approach was used. As a consequence, the VIF
that the relative percentage of revenues from grain to total (Xi) result indicates that the data did not have a major
yearly household income is the highest. As a result, grain problem with multicollinearity (Appendix 1). Tis is be-
cultivation is the primary source of revenue in the study cause the values of VIF for all continuous explanatory
region. It is followed by livestock production, of-farm variables were signifcantly less than 10. Similarly, the
production, and nonfarm production. Increased food cal- contingency coefcients, which quantify the degree of
orie consumption and yearly birr income were substantially connection between distinct discrete variables, were esti-
connected with better family food security status. Te av- mated to determine the degree of link among the discrete
erage daily calorie consumption of the entire sample homes variables. Te values of the contingency coefcient vary
was found to be 2266.2. According to the survey results, the from 0 to 1, with zero indicating no correlation and values
average calorie consumption of irrigation user families and near 1 suggesting a signifcant degree of association. As
nonirrigation user households was 2408.85 and 1605.93 a consequence, the calculation results indicated that there
kilocalories per day, respectively (Table 5). Te average were no severe issues with correlations among discrete
yearly income of a household’s irrigation and nonirrigation explanatory variables since the contingency coefcient did
users was 29069.56 birr and 17110.53 birr, respectively not surpass 0.75, which is commonly used as a cut-of point.
(Table 5). As a result, the survey results revealed that irri- As a result, all six discrete variables were included in the
gation user families were in a better position than non- logistic analysis (Appendix 2). Furthermore, robust stan-
irrigation user households in terms of income and daily dard errors were used to overcome the problem of het-
calorie consumption for improving lifestyles. eroscedasticity and normality.
Table 5: Descriptive statistics for the outcome variable.

Variables t-value p value
Mean S. d Mean S. d Mean S. d
Annual income 29069.56 8313.035 17110.53 6302.97 22717.88 9507.46 − 12.5∗ 0.0001
Daily ADE 2408.85 422.01 1605.93 534.51 1982.40 628.89 − 12.8∗ 0.0001
∗∗∗ ∗∗
Te signifcance level at 1%, 5%, and 10% was expressed in asterisks , , and ∗ , respectively.
3.13. Econometric Result. Te logistic regression model Tis is because the extent of farmed land encourages people
revealed that seven variables, including land holding size, to employ irrigation. Te consequence is that the likelihood
of/nonfarm income, education level, family size, de- of being an irrigator increases with farm size (other factors
pendency ratio, total livestock unit, and distance to nearest remaining constant). According to the survey fndings, when
agricultural extension ofce/FTC, were signifcantly infu- the area of cultivated land increases, a household’s ability to
enced by participation in small-scale irrigation at diferent vary the quantity and kind of the crop produced improves,
statistically signifcant levels as shown in Table 6 and see also which may lead to greater consumption and household food
Appendix 4. Te key explanatory variables that infuence security. Tis fnding is in line with Alemu [34] who showed
farm household irrigation participation are listed as follows. a signifcant diference between irrigation-user households
and nonuser households.
3.13.1. Household Head Education Level. According to the
survey result, it is statistically signifcant at the 5% level and 3.13.5. Livestock Holding. It was favorably and statistically
has a positive substantial infuence on household food se- signifcant at the 1% signifcance level for irrigation activity
curity. As a consequence, if the household was educated, participation. As a result, families have more animals, and
involvement in irrigation activities increased by (0.2) per- irrigation usage has increased by (0.08) percent. Further-
cent; hence, educated families have a greater likelihood of more, having more livestock provides farm households with
adopting small-scale irrigation. Asayehegn et al. [37] dis- a greater opportunity to generate more cash through the sale
covered that education has a crucial impact on family de- of animals, which increases income and improves food
cisions about technology adoption. security through direct consumption of livestock products.
3.13.2. Family Size. Te survey results show that it was 3.13.6. Distance to Farmers’ Training Center. Te model
statistically signifcant at the 1% probability level, and result indicates that this variable’s coefcient has a negative
households with a high family size utilize irrigation by (0.12) association with irrigation participation, yet it is statistically
percent more. Similarly, the more people who participate in signifcant at the 1% signifcance level. Tis suggests that
irrigation activities, the more crops are produced for con- families in the study region that are closer to the farmer’s
sumption and sale, increasing revenue and improving food training center are more likely to participate in small-scale
security. Farmers with larger family sizes were found to irrigation. As a consequence, the variable’s coefcient was
participate in small-scale irrigation practices more than those negative, and at 1 km of FTC, the chance of households
with smaller family sizes [38]. Tis might be the situation if participating in irrigation activities increased by (0.11)
family members are utilized as laborers in irrigated farming. percent.
3.13.3. Dependency Ratio. Te survey results showed that 3.13.7. Of-Farm Job Participation. At the 5% signifcance
involvement in small-scale irrigation was statistically sig- level, it had a favorable and signifcant efect on irrigation
nifcant at the 5% signifcance level. As a result of the activity participation. As a result, farmers who engage in of-
variable’s coefcient increasing at one unit, household in- farm activities are less likely to participate in irrigation.
volvement in irrigation activity increased by (0.077) per- Participation in of-farm activities may limit the allocation of
centage. As a consequence, the dependence ratio shows that manpower to farm operations. However, an increase of
the working population (active labor force, i.e., 15–64 year) (0.05) percent in the marginal efect of this variable suggests
actively participates in irrigation activities to create revenue that household heads who participate in of-farm work
or enhance food security. Similar fndings suggest that the participation are more likely to participate in small-scale
coefcient of the variable dependence ratio was statistically irrigation and contribute positively to revenue generation.
diferent from zero and found to be signifcant to impact the Similar fndings were obtained by [40]; however, Ayana [41]
food security status of families [39]. and Hadush [38] discovered that of-farm activities posi-
tively increase the amount of irrigation involvement.
3.13.4. Cultivated Land Size. Te survey results showed that
it is signifcant at the 1% level and has a favorable association 3.13.8. Propensity Scores. Te logit model was used to
with irrigation participation. Te data show that increasing predict the likelihood (propensity score) of each irrigation
the size of the cultivated area by one time improves the user (participant) and nonuser (nonparticipant) home based
chance of becoming an irrigation user by (0.414) percentage. on observed household data (Table 7). Te fndings of the
Table 6: Results of logistic regression.

Marginal efect
Variables Coefcient Std. err z-value p value
dy/dx
SEXHHH 0.6988549 0.4662645 1.50 0.134 0.1662
AGEHHH − 0.0035608 0.0206931 − 0.17 0.863 0.0008
EDSHHH 0.8521269 0.4223542 2.02 0.044∗∗ 0.2059
DPRHH 0.5631012 0.1586323 − 1.98 0.047∗∗ 0.0775
FASIZE | 0.5631012 0.1513425 3.72 0.001∗ 0.1388
SIZECUL 1.681016 0.6226108 2.70 0.007∗ 0.4145
SFLAND 0.23163 0.1513425 0.54 0.591 0.0569
TLU 0.3496817 0.430924 2.91 0.004∗ 0.0862
EOFF 0.8273996 0.4230811 1.96 0.051∗∗∗ 0.2001
DMKTHH − 0.0091119 0.0806091 − 0.11 0.910 0.0022
TRAIRR 0.4589877 0.4232527 − 2.57 0.278 0.1122
DFTCHH − 0.4542078 0.1769482 − 2.57 0.010∗ 0.1120
ACRSHH 0.3397416 0.4567797 0.74 0.457 0.0832
Constant − 5.949574 1.551089 − 3.84 0.001
Number of observations �241
Prob > chi2 �0.0001
LR chi2 (13) �145.44
Pseudo− R2 �0.4365
∗∗∗ ∗∗ ∗
Te survey result shows , , and level of signifcance at 1%, 5%, and 10%, respectively.
logit model of sample home participation in small-scale Table 7: Schematic representation common support region.
irrigation were used to generate propensity scores for the
Standard
matching algorithms. Te model employed thirteen (13) Variables Observation Mean Min Max
deviation
matching variables as explanatory variables, including the
age of the household head, gender of household head, ed- Treated 113 0.7440055 0.2422463 0.0153 0.998
Control 131 0.248 0.225 0.0028 0.981
ucational level of the household head, family size, de-
All 244 0.468 0.356 0.0028 0.998
pendency ratio, size of cultivable land, total livestock
holding, household income in of-farm activities, distance to
the nearest market, distance to the farmer training center, 3.13.10. Matching Algorithm. To match the treatment and
irrigation training, and access to credit service (Appendix 5). control groups in the common support region, several
As a result, the dependent variable was a binary variable with matching estimators were used. Te fnal choice of
a value of 1 for household involvement in small-scale irri- a matching estimator was determined by several criteria,
gation and a value of 0 for nonparticipation. including the equal mean test, also known as the balancing
Once the covariate propensity score estimation is test, pseudo-R2 , and matching sample size. Specifcally,
complete, a common support area should be imposed on the a matching estimator that balances all explanatory variables
propensity score distributions of both sample groups. Im- (i.e., results in an insignifcant mean diference between the
portantly, common support was increased by excluding two groups), a low pseudo-R2 value, and a large matched
treatment observations with estimated propensity scores sample size are preferred. According to Table 8, the projected
larger than or less than the maximum or lowest of the results of the test of the matching quality are based on the
comparison group propensity scores. Similarly, comparison performance criteria mentioned above; hence, kernel
group observations with propensity scores less than or matching with a bandwidth of (0.25) is the best estimator for
greater than the treatment observations were dropped as the data in this matching result.
shown in Figure 2.
3.13.11. Testing the Balance of Propensity Score and
3.13.9. Schematic Presentation of the Common Support Covariates. After selecting the best performing matching
Condition. Te predicted propensity scores for user algorithm, the balancing test is performed to determine the
(treatment) homes range between 0.0153 and 0.998 signifcance of the mean diference between all variables used
(mean � 0.744) and between 0.0028 and 0.981 (mean � 0.24) for matching before and after matching [42]. Te goal of the
for nonuser (control) households, as shown in Table 7. Te balancing test is to confrm that treatment is independent of
zone of shared support would thus be between 0.0153 and unit characteristics after conditioning on observed charac-
0.9988. In other words, families with estimated propensity teristics (as estimated in the propensity score model), DX/
scores between 0.015 and 0.998 were excluded from the P(X), where X is the set of characteristics that are thought to
matching procedure. As a consequence, all of the treated satisfy the conditional independence assumption.
families’ approaches to 1 show that most irrigation users Any variation in the covariate mean between the two
raise their income and enhance their food security more groups in the matched sampling has been erased after
than the control groups (nonirrigation users). matching techniques, and hence, the pseudo-R2 should be
0 .2 .4 .6 .8 1
Propensity Score
Untreated: Off support Untreated: On support

Treated: On support Treated: Off support
3
kdensity ps
0
0 .2 .4 .6 .8 1
Propensity score (ps)
Treated ps
All household ps
Control ps
Figure 2: Prosperity score of treated, all households, and control.
Table 8: Performance of matching estimators.

Matching algorism Bandwidth pseudo- pseudo − R2 Insignifcant variable Matching sample size
Radius (0.01) 0.078 12 111
Radius (0.05) 0.042 10 217
Radius calliper (RC)
Radius (0.1) 0.035 11 224
Radius (0.25) 0.039 12 224
NN (1) 0.08 9 224
NN (2) 0.058 11 224
Nearest neighbour (NN)
NN (3) 0.055 10 224
NN (4) 0.045 10 224
Bw (0.01) 0.067 12 111
Bw (0.1) 0.036 11 224
Kernel matching (KM)
Bw (0.25) 0.031 13 224
Bw (0.5) 0.061 10 224
Source: own econometric result (2021).
low [19]. Te basic goal of propensity score estimation is to insignifcant, indicating that the disparities in the covariate
balance the distributions of important variables in both mean between the treatment and control groups were
groups, rather than to achieve a precise forecast of treatment abolished and balanced as shown in Figure 3.
selection. As indicated in Table 9, using the matching
method technique resulted in a decrease in overall bias. 3.13.12. Estimating Treatment Efect on the Treated. Tese
Before matching, certain characteristics were substantially sections investigate the efects of small-scale irrigation on the
diferent for the two groups of respondents, including land end variables of family income and food security status
holding size, of/nonfarm income, education level, family before and after irrigation usage to achieve the study’s main
size, dependence ratio, total livestock unit, and distance to and specifc stated objectives. Daily calorie intake per adult
farmer training center (Table 9). However, following equivalent (DCAE) and annual income of rural families were
matching, these signifcant variables were conditioned to be evaluated as outcome variables.
Table 9: Propensity score and covariate balancing.

Mean T-test
Variables Sample Reduction bias (%)
Treated Control t-value p value
U 0.84956 0.64844 3.65∗∗∗ 0.001
Sexhhh
M 0.84906 0.80573 78.5 0.83 0.406
U 42.08 46 − 3.04∗∗∗ 0.003
Agehhh
M 42.292 40.901 64.5 1.19 0.234
U 0.76991 0.36719 6.84∗∗∗ 0.001
Edshhh
M 0.75472 0.78374 92.8 − 0.5 0.618
U 4.7479 4.5595 0.98 0.328
Dprhh
M 4.7812 5.0177 − 25.5 − 1.15 0.252
U 5.0885 4.0391 5.14∗∗∗ 0.001
Fasize
M 5.0472 5.0167 97.1 0.13 0.893
U 1.2212 0.83203 8.15∗∗∗ 0.001
Sizecul
M 1.184 1.1117 81.4 1.29 0.197
U 0.77876 0.42969 5.86∗∗∗ 0.001
FPSFland
M 0.77358 0.79442 94 − 0.37 0.714
U 6.6473 4.6154 8.88∗∗∗ 0.001
Tlu
M 6.3906 6.2694 94 0.54 0.589
U 0.77876 0.35938 7.18 0.001
Eof-farm
M 0.77358 0.79159 95.7 − 0.32 0.752
U 7.146 7.6172 − 1.45 0.147
Dmkthh
M 7.217 7.1161 78.6 0.27 0.785
U 0.78761 0.38281 6.92∗∗∗ 0.001
Traing irri
M 0.77358 0.75329 95 0.35 0.73
U 1.7409 2.0273 − 1.92∗∗ 0.056
Dftchh
M 1.7426 1.7125 89.5 0.23 0.82
U 0.78761 0.4375 5.91∗∗∗ 0.001
Accredit
M 0.77358 0.83819 81.5 − 1.19 0.236
∗∗∗ ∗∗ ∗
, , and shows the level of signifcance at 1%, 5%, and 10%, respectively. Source: own econometric result (2021).
Unmatched
.04
.03
Density
.02
.01
0
-116 -87 -58 -29 0 29 58 87 116
Standardized % bias across covariates
Matched
.04
.03
Density
.02
.01
0
-116 -87 -58 -29 0 29 58 87 116
Standardized % bias across covariates
Figure 3: Propensity score and covariate balancing.
Table 10: Te average treatment efect on the treated.

Mean
Variables ATT Se t-stat
Treated Control
Income 28423.2075 23188.9491 5234.25845 1527.21004 3.43
Daily calorie/ADE 2388.30189 2144.14014 244.161748 116.621446 2.09
Table 10 presents the estimation results and supportive 2388.3 Kcal per capita per day, higher than the national
evidence of the statistically signifcant efect of small-scale average. Access to small-scale irrigation increased the yearly
irrigation on the abovelisted outcome variables based on the income of irrigation user families by 5234.258 ETB com-
already chosen matching algorithm. Accordingly, the small- pared with nonirrigation users.
scale intervention has resulted in a positive and statistically In conclusion, access to small-scale irrigation is a feasible
signifcant mean diference between irrigation users and solution for securing household food needs in the study area.
nonirrigation users of farm households in terms of daily Encouraging rural farmers to use small-scale irrigation is
calorie intake per adult equivalent and annual income. A critical for producing consumable food and marketable
positive value of ATT confrmed that the income and food crops, reducing socioeconomic problems, and achieving
security have been improved due to the participation of sustainable livelihood improvement. Te fndings suggest
irrigation activity in the study area. improving water access for irrigation, raising farmers’
After passing diferent steps of the matching technique, it awareness to enhance participation, and scaling up irrigation
was found that the use of small-scale irrigation increased the interventions to other irrigable land areas.
daily calorie intake and annual income of participant
households by 244.162 kilocalories and 5234.258 ETB, birr, Data Availability
and this diference was signifcant at the 1% probability level,
respectively (Appendix 3). Alemu [34] reported that the Te collected and analyzed data during the current study are
average increase in total calorie intake and consumption available upon reasonable request from the corresponding
expenditure per adult equivalent among the participating author.
families was 907.07 kcal and 2593.92 birr, respectively.
Conflicts of Interest
3.13.13. Sensitivity Analysis. In this study, sensitivity anal- Te authors declare that they have no conficts of interest.
ysis was carried out on the estimated average treatment efect
using alternative matching estimators for both outcome Authors’ Contributions
variables and checked the observed characteristic of vari-
ables for both the treated and control groups (Appendix 7). AwH initiated the research, wrote the research proposal,
Which were mentioned earlier, all matching estimators conducted the research, performed data entry and analysis,
resulted in statistically signifcant efects on the households and wrote the manuscript. ArH was involved in the
that use irrigation rather than nonirrigation users for the methodology, writing, reviewing, and editing of the research
comparison of the outcome of household calorie intake and proposal and manuscript. All authors have read and ap-
total annual income for the improvement of livelihoods. proved the fnal manuscript.
4. Conclusion Acknowledgments
Tis study investigates the impact of a small-scale irrigation Te authors would like to thank the expert’s agriculture
system on farmer livelihoods in the Legehida area of South ofce of the district for their patience and support in getting
East Ethiopia. Te research aims to reduce poverty and food the required supplementary data. Besides, the authors would
insecurity and enhance income by identifying socioeco- like to thank respondents for their dedicated willingness to
nomic issues. Data were collected from primary and sec- participate in this study.
ondary sources using a simple random sampling approach
and a quasiexperimental design. Te logit and PSM models Supplementary Materials
were applied to examine the infuence of small-scale irri-
gation on farmer livelihoods. Te study found that land Appendix 1: Livestock conversion factor. Appendix 2:
holding sizes, of-farm income, education level, family size Conversion factor for adult equivalent. Appendix 3: Food
dependency ratio, total livestock unit, and distance to farmer poverty line /conversion factor for each food item. Appendix
training centers signifcantly infuence farmers to practice 4: Results of the logistic model. Appendix 5: Result re-
small-scale irrigation and enhance food security in the area. gressions. Appendix 6: Variance infation factor. Appendix
Te propensity score matching model revealed that the 7: Sensitivity analysis result. Appendix 8: Shortened version
mean food intake of families on the treatment was of survey questionnaires. (Supplementary Materials)
References [18] T. Yamane, Statistics: An Introductory Analysis, New York

University, New York, NY, USA, 2nd edition, 1967.
[1] Photius, “CIA world factbook,” 2013, https://www.photius.com/ [19] M. Caliendo and S. Kopeinig, Some Practical Guidance for the
rankings/economy/population_below_poverty_line_2013_0.html. Implementation of Propensity Score Matching, University of
[2] MoARD, Small-Scale Irrigation Situation Analysis and Ca- Cologne, Cologne, Germany, 2005.
pacity Needs Assessment, MoARD, Addis Ababa, Ethiopia, [20] E. Geta, A. Bogale, B. Kassa, and E. Elias, “Productivity and
2011. efciency analysis of smallholder maize producers in southern
[3] MoARD, “Performance and challenges on the fve-year Ethiopia,” Journal of Human Ecology, vol. 41, no. 1, pp. 67–75,
strategic plan of growth and transformation: annual Report 2013.
of MoFED, addis ababa, Ethiopia,” 2012, http://www.mofed. [21] O. A. Fadare, D. Akerele, and B. Toritseju, “Factors infu-
gov.et. encing adoption decisions of maize farmers in Nigeria,” Te
[4] N. Li, M. Chen, H. Gao, D. Huang, and X. Yang, “Impact of International Journal of Food and Agricultural Economics,
lockdown and government subsidies on rural households at vol. 2, no. 3, pp. 45–54, 2014.
early COVID-19 pandemic in China,” China Agricultural [22] P. R. Rosenbaum and D. B. Rubin, “Te central role of the
Economic Review, vol. 15, no. 1, pp. 109–133, 2023. propensity score in observational studies for causal efects,”
[5] D. Bacha, R. Namara, A. Bogale, and A. Tesfaye, “Impact of Biometrika, vol. 70, no. 1, pp. 41–55, 1983.
small-scale irrigation on household poverty: empirical evi- [23] R. H. Dehejia and S. Wahba, “Propensity score matching
dence from the ambo district in Ethiopia,” Irrigation and methods for non-experimental causal studies,” Te Review of
Drainage, vol. 60, pp. 1–10, 2011. Economics and Statistics, vol. 84, no. 1, pp. 151–161, 2002.
[6] Z. Bao and D. Huang, “Shadow banking in a crisis: evidence [24] S. O. Becker and A. Ichino, “Estimation of average treatment
from FinTech during COVID-19,” Journal of Financial and efects based on propensity scores,” STATA Journal, vol. 2,
Quantitative Analysis, vol. 56, no. 7, pp. 2320–2355, 2021. no. 4, pp. 358–377, 2002.
[7] Csa (Central Statistical Agency), Agricultural Sample Survey [25] A. Bryson, R. Dorsett, and S. Purdon, Te Use of Propensity
for Strategic Planning: Annual Report Federal Democratic Score Matching in the Evaluation of Labor Market Policies,
Republic of Ethiopia Central Statistical Authority of Ethiopia: Department for Work and Pensions, London, UK, 2002.
Annual Report, CSA, Addis Ababa, Ethiopia, 2012. [26] M. Mendola, “Agricultural technology adoption and poverty
[8] T. Mutsvangwa and K. Doranalli, Agriculture and Sustainable reduction: a propensity Score Matching Analysis for Rural
Development, Te Hague University Press, Netherlands,
Bangladesh,” Food Policy, vol. 32, no. 3, pp. 372–393, 2007.
Europe, 2006.
[27] J. Jalan and M. Ravallion, “Estimating the beneft incidence of
[9] Fao, “Irrigation in africa South of the sahara,” FAO Investment
an antipoverty program by propensity-score matching,”
Centers Technical Paper 5, FAO, Rome, Italy, 2003.
Journal of Business and Economic Statistics, vol. 21, no. 1,
[10] Y. Jambo, A. Alemu, and W. Tasew, “Impact of small-scale
pp. 19–30, 2003.
irrigation on household food security: evidence from Ethio-
[28] W. S. Lee, Propensity Score Matching and Variations on the
pia,” Agriculture and Food Security, vol. 10, no. 1, 2021.
Balancing Test: Melbourne Institute of Applied Economic and
[11] B. Mehretie and B. Woldeamlak, “Traditional irrigation and
Social Research, the University of Melbourne, Melbourne,
water management practices in highland, Ethiopia: case study
in Dangila Woreda,” Irrigation and Drainage, vol. 52, 2013. Australia, 2006.
[12] M. Bhattarai, R. Sakthivadivel, and I. Hussein, Irrigation [29] K. Asayehegn, “Negative impact of small-scale irrigation
Impacts on Income Inequality and Poverty Alleviation: Policy schemes: a case study of Central Tigray regional state,
Issues and Options for Improved Management of Irrigation Ethiopia,” Wudpecker Research Journals, Agricultural Re-
Systems, International Water Management Institute (IWMI) search and Reviews, vol. 1, no. 3, pp. 80–85, 2012.
Working Paper, Colombo, Sri Lanka, 2002. [30] A. Abiyu, M. Tebeje, and E. Mekonnen, “Determinants of
[13] S. Eshetu, B. Belete, D. Goshu et al., Income Diversifcation small-scale irrigation utilization by smallholder farmers in rift
through Improved Irrigation in Ethiopia: Impacts, Constraints valley basin, Wolaita zone, Ethiopia,” Journal of Natural
And Respects for Poverty Reduction, Evidence from East Sciences Research, vol. 5, no. 21, pp. 2224–3186, 2015.
Harerghe Zone, Oromia Region, Ethiopia, Working Paper 14, [31] J. B. Dowd, L. Andriano, D. M. Brazel et al., “Demographic
Research-inspired Policy and Practice Learning in Ethiopia and science aids in understanding the spread and fatality rates of
the Nile region (TRiPPLE), Addis Ababa, Ethiopia, 2010. COVID-19,” Proceedings of the National Academy of Sciences,
[14] B. B. Awulachew, Experiences and Opportunities for Pro- vol. 117, no. 18, pp. 9696–9698, 2020.
moting Small Scale Micro Irrigation and Rainwater Harvesting [32] H. Storck, B. Emana, B. Adnew, A. Borowiecki
for Food Security in Ethiopia, IWMI (International Water S. W/Hawariat, “Farming systems and farm management
Management), Addis Ababa, Ethiopia, 2005. practices of smallholders in hararghe highlands,” Farming
[15] S. D. Zhou, T. Herzfeld, T. Glauben, Y. Zhang, and B. Hu, Systems and Resource Economics in Tropics, vol. 12, 1991.
“Factors afecting Chinese farmers’ decisions to adopt a water- [33] A. Dillon, “Te efect of irrigation on poverty reduction, asset
saving technology,” Canadian Journal of Agricultural Eco- accumulation, and informal insurance: evidence from
nomics/Revue canadienne d’agroeconomie, vol. 56, no. 1, northern Mali,” World Development, vol. 39, no. 12,
pp. 51–61, 2008. pp. 2165–2175, 2011.
[16] M. Lipton, J. Litchfeld, and J. M. Faurès, “Te efects of ir- [34] M. Alemu, “Impact of small-scale irrigation on farm
rigation on poverty: a framework for analysis,” Water Policy, households’ food security: the case of girana irrigation scheme
vol. 5, no. 5-6, pp. 413–427, 2003. in kobo district, north Wollo zone of Amhara region,
[17] M. Hulluka, Agricultural Sector Development and the Role of Ethiopia,” Haramaya University, Addis Ababa, Ethiopia,
the National Agricultural Extension Program; Centre for 2017, M.Sc. thesis.
Applied Studies in International Negotiation, Te food chain [35] A. W. Bartik, M. Bertrand, Z. Cullen, E. L. Glaeser, M. Luca,
in Sub-Saharan Africa, Geneva, Switzerland, 2000. and C. Stanton, “Te impact of COVID-19 on small business
outcomes and expectations,” Proceedings of the National

Academy of Sciences, vol. 117, no. 30, pp. 17656–17666, 2020.
[36] E. Leuven and B. Sianesi, “Psmatch2 stata module to perform
full mahalanobis and propensity score matching,” 2003,
http://ideas.repec.org/c/boc/bocode/s432001.html.
[37] K. Asayehegn, C. Yirga, and S. Rajan, “Efect of small-scale
irrigation on the income of rural farm households: the case of
laelay maichew district, central tigray, Ethiopia,” Journal of
Stored Products and Postharvest, vol. 56, 2011.
[38] H. Hadush, Adoption and Impact of Micro-irrigation on
Household Income: Te Case of Eastern Tigray, Doctoral
dissertation, Mekelle University, Addis Ababa, Ethiopia, 2014.
[39] G. T. Semere, Dimensions and Determinants of Rural House
Hold Poverty: the Case of Eastern Zone, Haromaya university,
Dire Dawa, Ethiopia, 2006.
[40] A. Abdulai and W. Hufman, “Te adoption and impact of soil
and water conservation technology: an endogenous switching
regression application,” Land Economics, vol. 90, no. 1,
pp. 26–43, 2014.
[41] A. Ayana, “Assessing the impact of small-scale irrigation
schemes on household income in bahir dar zuria Woreda,
Ethiopia,” Journal of Economics and Sustainable Development,
vol. 45, 2016.
[42] C. Heinrich, A. Mafoli, and G. Vazquez, A Primer for Ap-
plying Propensity Score Matching: Impacts Evaluation
Guidelines, University of Wisconsin-Madison, Madison,
Wisconsin, 2010.

The Scientific World Journal - 2024 - Hussen - Influence of Small Scale Irrigation On Livelihoods of Rural Farm Households

Uploaded by

Copyright:

Available Formats

The Scientific World Journal - 2024 - Hussen - Influence of Small Scale Irrigation On Livelihoods of Rural Farm Households

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Scientific World Journal - 2024 - Hussen - Influence of Small Scale Irrigation On Livelihoods of Rural Farm Households

Uploaded by

Copyright:

Available Formats

Wiley

Awol Hussen1 and Arebu Hussen 2

Correspondence should be addressed to Arebu Hussen; [email protected]

Academic Editor: Difang Huang

1. Introduction raw material supply, 28% of government tax income, and

Table 1: Te target sample is taken from the population of each kebele.

Density Density of explained as follows (Table 2) that have a signifcant impact

3. Results and Discussion

Types of variable Description Measurement Hypothesis expectation

Table 3: Descriptive statistics for continuous variables of the household characteristics.

Table 4: Descriptive statistics of dummy variables of household characteristics.

Table 5: Descriptive statistics for the outcome variable.

Table 6: Results of logistic regression.

Untreated: Off support Untreated: On support

Table 8: Performance of matching estimators.

Table 9: Propensity score and covariate balancing.

Table 10: Te average treatment efect on the treated.

References [18] T. Yamane, Statistics: An Introductory Analysis, New York

outcomes and expectations,” Proceedings of the National

You might also like