1 Impact of Small Scale Irrigation on Resettled Household Food

Security: The Case of Laga Hora, Meda Jalela & Shenkora Small Scale
irrigations Sasiga District of Wollega, Ethiopia
Habtamu Tolla Shono1 & Shumet Asefa 2
M.Sc. in Development economics, DBE Nekemte District, Nekemte, Ethiopia
Email: tollahabtamu@gmail.com
Doctor of Philosophy in Applied Economics and V/P at Mekdella University
Former Research and Technology Transfer Head at Wollega University
Email: Shumetasefa@gmail.com
Abstract: Small Scale Irrigation plays substantial role in advancing agriculture by increasing productivity and
enhances household food security. The Objective of the study is to see factors affecting resettled farmers
participation decision in Small scale irrigation and to analyze the impacts of small Scale irrigation on the
household food security. The study used cross sectional data with multistage stage sampling technique of 262
and 77 non- irrigation and irrigation user resettled households were interviewed. Binary probit models,
Heckman sample selection model two step are employed in the analysis of the study. The study showed that
resettled households with strong perception to land productivity, utilize modern agricultural input, more
perception on soil fertility, more access of extension service and with previous land holding on irrigation site
have more likelihood of participation decision in small scale irrigation than those household with less access
and perceptions. Contrary to that resettled households of large land Size holding, more perception on
irrigation land repossession and more awareness on irrigation regulation are found to have less likelihood of
participation decision in small scale irrigation. The ordinary least square using adult Equivalent expenditure
on food as outcome variable revealed that as Age of resettled households, distance from the market, tropical
livestock holding and crop diversification increases the resettle household food security increases. While
Awareness of households on irrigation regulation & modern farm input utilization were found to have negative
impact on resettled household food security. This is due to farm input cost escalation, soil fertility deterioration
and poor land preparation and poor crop management practices. As per the study finding holding large land
size, less awareness on irrigation regulation and repossession of irrigation land are found to be the challenging
factors to influence resettled household participation decision in small scale irrigation and while as the farmers
adopt more input on rain fed land and develop less awareness on irrigation regulation copying capacity to food
security was challenged. The study recommended concerned body and stockholders to revisit irrigation
regulation for amendment, soil and water conservation practice to be taken as an intervention package and
further study to exceed the scope of the study. Key Words –Small Scale Irrigation, Food security, Heckman
Model, district Sasiga
1. Introduction

Agriculture plays indispensable role in Ethiopian economy, sources of food for the people,
supplies raw material to the industry and creates market for the other economic sectors.
However rain fed dependency, outdated methods of cultivation, low productivity, shortage of
rainfall and pest infestation is the common problem of the sector. A number of challenges
are confronting the Ethiopian agriculture among others drought, limited knowledge on
modern agricultural practices and poor infrastructural development are common. Rural
communities in drought prone area develop less copying capacity in challenging bad weather
conditions particularly during food shortage.

To tackle , the government has launched food security enhancing program since 2003 among
which resettlement program, productive safety net program, household asset building
program and complimentary community investment program are common (Guush B, 2013).
In the study area there are resettled communities who have had basic benefit packages set in
the program among others farm land and institutional services. Small scale irrigation
schemes are developed to utilize land during dry season, hence to enhance the resettled
household food security. This means the schemes are playing a significant role regardless of
certain draw backs in the management of irrigation land, water and utilization of improved
farm input and cultural practices.

Figure 1 Map of Sasiga District (Source: Ethiopian Mapping Agency 2018)

2. Methods and Materials

1.1 Study Area

The study area, Sasiga district is found in Oromia Regional state of Ethiopia 349 kms from
Addis Ababa to the West North direction. It is bordered on the south by Diga Leka , on the
west by the Benishangul-Gumuz region, on the north west by Limmu, on the north by
an exclave of the Benishangul-Gumuz Region and on the east by Guto Wayu. In the district
there are 27 peasant associations and 1 urban town. Among which 8 peasant associations
hosts of the resettled communities.
The district has a total of 16,787 households and population of male 41,326 (51.13%), female
39,458 (48.82%) of the total 80,814. 78,241 (96.82%) of these population living in the rural
and the rest 2,659 (3.29%) are living in urban area. 16,113 (21.9%) households and 78,136
(96.82%) of the population members are living in rural areas. The remaining 674(4%)
households 2,659 (3.29%) populations are living in urban area.

It has an area of 93,813 hectares with topographical features of more gentle sloped plains,
flat areas and small mountain chains. Large parts of the area are covered by organic coffee
plantation in the evergreen forest. The altitude ranges between 1500 – 1960 meter above sea
level, annual rainfall between 800mm – 1200mm and temperature 18 oc- 31 oc. The soil type
is characterized by sandy loam, silt and silt loam; the area have heavy rain fall that largely
eroded the soil and devastating soil fertility.
Inters of the command area irrigation schemes can be categorized in to:-
1. Small-scale irrigation system (<200 hectare)
2. Medium- scale irrigation system (200-300 hectare)
3. Large- scale irrigation system (>300 hectare).

The focus of the study is on small scale irrigation system.

Food Security: Food Security [Is] Situation That Exists When All People, At All Times, Have
Physical, Social And Economic Access To Sufficient, Safe And Nutritious Food That
Meets Their Dietary Needs And Food Preferences For An Active And Healthy Life
(FAO, 2002).
Household: A Household Is A Person Or Group Of People That Acts As A Single Decision-
Making Unit (Depken, 2006). In The Study Context Household Is An Individual
Holding Land, Home And Family Who Takes The Position Of Heading Family By
Providing Basic Needs From Local Agriculture.

Resettled household: Is food insecure households moved from drought prone areas to rainy
areas equipping them with benefit packages of land, food aid and residential land.

Impact Assessment:

1.2 Study Design

Resettled household based cross-sectional data was employed for the research study.

1.1 Study Population

The study population is households living in three peasants association where the three
irrigation schemes area located. These are identified as those possessing irrigated land and
those who do not have; they are commonly called irrigation users and non-irrigation users.

1.3 Sample size

The study used Cochran (1963:75) to yield a representative sample for proportions below:
2
z pq
n o= 2
… … … … … … … … … … … … … … … … … … . … … … … (1)
e

Where n o is the sample size, Z is the abscissa of the normal curve that cuts off an area at the
tails (1.96 - equals the desired confidence level, e.g., 95%), e is the desired level of precision,
p is the estimated proportion of an attribute that is present in the population, and q is 1-p. The
value for Z is found from statistical tables which contain the area under the normal curve.

As per data sources of the district agriculture development office (2018), the target
population of the study covered 654 irrigation user households and 2,185 households that are
not using small scale irrigation for farm practices. A maximum variably of P = 0.5
(maximum variability) is taken as an assumption. Furthermore, it was supposed that a 95%
confidence level and ±5% precision enables a sound sample size. Accordingly, the resulting
sample size is demonstrated below
 2
z pq
n o= 2 n0= ((1.96)2x0.5x0.5)/ (0.05)2 = 385
e
In order to slightly reduce the sample size and make a given sample size provide
proportionately more information for a small population than for a large population. The
sample size (no) can be adjusted using the following equation.
no
n= ……………………………………………... (2)
1+¿ ¿ ¿
Where n is the sample size and N is the population size., this adjustment (called the finite
population correction) can substantially reduce the necessary sample size for small
populations.
Accordingly, the sample size is adjusted below
385
no n=
n=
1+¿ ¿ ¿
= 1+
(385−1) = 339
2839
Where
n- Adjusted sample size for the target population
no- calculated sample size on the bases of three critical factors
N –is total population of both irrigation users and non-irrigation users of the three peasant
association. Thus the study will have 77 irrigation and 262 non irrigation user respondents

Sample Size Allocation Summary Table

N Total Number of HHs Sample allocation
o Name of PAs Non Irrigation user Non Irrigation user
Irrigation Irrigation Sample Sample Sample Sample
user user Weight Size weight Size
1 HoraWata 138 749 21% 16 34% 89
2 BekilteShenkor 358 645 55% 42 30% 78
3 MedaJalela 158 791 24% 19 36% 95
Total 654 2,185 100% 77 100% 262
Table 1- Sample Size Allocation Summary
Source: Own survey 2018

1.4 Sampling Techniques

The study employed multistage sampling techniques that resettled peasant associations &
operational small scale irrigation were purposively selected depending on existing
operational schemes, accessibility of the PAs and easy follows up of the research study.
Stratified sampling was employed by classifying the sample into two strata irrigation users
and non-irrigation users among which 77 HHs from irrigation user and 262 from non-
irrigation user were randomly selected.

1.5 Econometrics model

The study employed Heckman’s two-step procedure to investigate the impact of small scale
irrigation on resettled household food security.

Specification of the Heckman model

Following Heckman revised version (1979), the model is specified as follows. Let Z ik be a
group of variables K which represent the characteristics of household i which determine the
decision to participate in the irrigation scheme measured by a latent variable Y *I and γk are the
coefficients which reflect the effect of these variables on this decision; and X is is a group of
variables S which represent the characteristics of household i which determine household
food security Y2i and β’s are the coefficients which reflect the effect of these variables on
household food security. Thus, the Heckman model (1979) takes the following form:

k
Y 1i¿ =∑ γ k Z ik +U i … … … … … … … … … … … … … … … … .. 1
k−1

s
Y 2i=∑ B s X is +ε i … … . observed only if D¿i > 0 … … .2
s−1

Where: the disturbances ui and εi follow a bivariate normal distribution with a zero mean,
variance σu and σε respectively, and covariance σεu. Therefore, we define a dichotomous
variable Y1i which takes a value 1 when the latent variable Y*I > 0 and 0 otherwise. In this way
Y1i= 1 indicates the decision to participate in the irrigation scheme and Y1i= 0 that of not to
participate in the irrigation scheme, β and γ are the vectors of parameters which measure the
effect of variables x and z, respectively. Lambda (inverse Mills ratio) which is related to the
conditional probability that an individual household will decide to use (given a set of
independent variables) is determined by:

ϕ (−γZ )
λ (−γZ )=
¿¿
Where λ is inverse Mills ratio, φ and Ф are density and cumulative distribution function;
respectively of the standard normal distribution.

In the second stage of the outcome equation λ is included as an independent variable in the
OLS estimates to measures the impact of small scale irrigation on household food security.
The selection control factor (λ(-Zγ) reflects the effect of all the unmeasured characteristics
which are related to the small scale irrigation water use decisions. The model is specified as:

Y 2i= Xβ+σ εu (−γ Z ) +V i … … … … … … … … … … … … … … … … … … … … … … … … .. 4

Where νi is the distributed error term uncorrelated with X𝝱 and λ (−γz)

1.6 Data collection procedures

The study employed both qualitative and quantitative data type for statistical and
econometrical analysis. Both dummy and continuous variables were used because there are
some interesting variables that have to be expressed in qualitative terms which have to be
transformed into some proxy so that it has to be represented and used in a regression and
also there area numeric variables that have an infinite number of values between any two
values. Secondary sources are those which have been made available or collected to support
the findings with theories and empirical reviews.

1.7 Data Processing and Analysis

For the household survey, a semi-structured questionnaire that contained both open and
close-ended questions was prepared and translated into vernacular language (Afan Oromo).
Before conducting actual process of data using this questionnaire, pre testing of questionnaire
was done to ensure validity and reliability. Moreover, in order to minimize errors in data
collection and properly administer the questionnaire, two day training was given for five
enumerators including pretesting work. Finally, minor modification was done on
questionnaire based on the feedback gained and the final modified semi structured
questionnaire administered on selected sampled households residing in the three resettle
association. Original data sources were moved in to excel to the form needed in
specified econometric model, the data were imported to Stata- software to employ the
most widely used model for analysis.
3. Result and Discussion

This chapter discusses detail of the research findings dividing in to 2 sections. The 1 st section
describes the results of descriptive statistics on household demographic and socio economic
characteristics. The 2nd section discusses about the econometric results of the selection and
outcome model as its concern is to see the impacts of participation in irrigated farming on
house hold food security.

3.1 Descriptive Analysis

3.1.1 Demographic Characteristics
In order to asses data collected on sampled households, data were collected from each PA
where the irrigation schemes are operating. Hence, in the next sections the characteristics of
sample households such as sex, age, marital status, literacy, household size, livestock size
and size of cultivated land is described.
Sex and Marital Status of Household Head

Out of the sampled household heads, 85 percent of irrigation users and 87 percent of non-
irrigation users were male headed while the remaining15 and 13 percent were female headed
respectively. Regarding their marital status, about 87 percent irrigation and non-users were
married while 9 and 10 percent were divorced and 4 and 3 percent were widowed household
heads (Table 3).

Table 2- Sex and Marital Status of Resettled Household Head

Description Sex of House hold Head Frequency Percent

Male 64 85
Irrigation User Female 13 15
Male 228 87
Non Irrigation user Female 34 13
Description Marital Status
Married 67 87
Divorced 7 9
Irrigation User Widowed 3 4
Married 67 87
Divorced 8 10
Non Irrigation user Widowed 2 3
Source: Own Survey 2019

Age and Education Level Household Heads

The sampled household age distribution is shown below. The mean age of household head of
irrigation users was found to be 41.84 and 41.12 percent for non-irrigation users with
standard deviation of 9.6 and 8.88 percent respectively. The maximum and minimum age for
irrigation and non-irrigation users was found to be 26, 69, 23 and 68 respectively (Table 4).

Table 3 Ages of Surveyed Resettled Household Head

Description Average St. Deviation Minimum Maximum

Irrigation user 41.84 9.6 26 69

Non- irrigation user 41.12 8.9 23 68

Source: Own Survey 2019

Majority of both participants and non-participants didn’t attend education, 67 percent of

sampled participants and 82 percent of non-participants were found to be illiterate. The
remaining 19 percent of participants households can read and write while 13.6 percent were
attend primary school. With regard to non-participants, 4.52 percent of them had a primary
school education level, 13 percent can read and write while the remaining 0.38 have
secondary education level (Table 5).

Table 4 Education levels of Resettled Household Head

Description Level of education in percent

Illiterate Read and write Primary Secondary
Irrigation user 67 19 13.6 1.4
Non irrigation user 82 13 4.52 0.38
Source: Own Survey 2019

3.1.2 Socio Economic Factors

Occupation and Economic Activity of Households

The survey result revealed as all portion of irrigation participant household and non-
participant households engaged in agriculture in which all of them drive their livelihood from
farming. This implies that none of them participate in other economic activity to generate
income. Family member of both irrigation and non-irrigation user households were involved
in off farm income as daily laborers particularly during summer season. As depicted in the
table below 3 percent of household family engaged on off farm income sourcing activities
while the remaining 97 percent of the family members on farm activities in condition that
they fit to the labor force. Seven percent of the non-irrigator household family engaged in off
farm activities (Table 6).

Table 5 Resettled Household Family Engaged on off Farm Activity

Household Family Frequency Percent

Irrigation user Off farm participants 14 3

Non participants 371 97

Non irrigation user Off farm participants 89 7

Non participants 1221 93

Source: Own Survey 2019

Land Holdings

The settler households acquired 2.2 hectares of land of which 200m 2 was for residential
house while the remaining 2ha was granted for farm land. Moreover few of the settle
households expand land holding informally sourcing from host community activities. As
shown in the table irrigation householders were found to have 2.2 ha of landholding while
Non irrigation households have 2.73ha on average bases. Among 1.95, 0.25 1.7 and 0.25
hectare of land was used for crop production by irrigation and rain fed mode of operation
while the remaining land was used for animal grazing and related activities respectively.
Non-irrigation users hold more land size than irrigation users. On averaged they hold 2.73ha
average land size of 2.17haand 0.56 for crop cultivation and animal grazing.

Table 6 Land Holdings of the Sampled Resettled Households in Hectare

Description I Irrigation user Non- irrigation user

A Average t- value
 Total land area 2.2 2.73 2.5 3.7841*

Land used for cultivation 1.95 2.17 2.6 2.925*

Land used for irrigation 0.25 - 0.1 2.754*

Land used for rain fed 1.70 2.17 1.9 11.016*

L Land used for other purpose 0.25 0.56 0.4 -3.4148**

Source: Own Survey 2019

* , **, Significant at 1 & 5 percent

Livestock Holding

As shown in the table below the livestock holdings of the sampled households has a
significant difference among irrigators and non-irrigators. Non irrigators hold 2.9cattle
comparison to irrigators with average holdings cattle count to 1.42. The mean Sheep/Goat
holding of irrigators was found to be 0.25 while non-irrigator household was 0.45. Coming to
donkey and chicken both farmers holds insignificant amount on average it is found to be
0.045, 0.048 and 0.019 and 0.025 of the live stokes respectively (Table 8).

Table 7– Major Livestock Holdings of the Sample Resettled Household of the Study Area (TLU)

Description Irrigation user Non irrigation user t-value

Cattle 1.42 2.9 3.9733*
Sheep/Goat 0.25 0.45 1.6049**
Donkey 0.045 0.048 -7.7109*
Chicken 0.019 0.025 -3.0156*
Source: Own Survey 2019

 * Significant at 5 % significance level

3.2 Econometric Analysis

The regression model that analyses a combination of a quantitative and qualitative
information by including both continuous and discrete variables was selected; in this regard
discrete /binary outcome model among which probit model , sample selection model
/Heckman two step and linear regression model were used to analyze the organized data .
Probit model is used to see the relationship of individual, socio economic and land holding
characteristics of the household to participation decision in small scale irrigation. This
addresses the question of what happen to the likelihood of the household participation
decision in small scale irrigation as the marginal effect of the model changes. The objective
of using Heckman selection model is to correct selection bias and get lamda/ hazard
function/Mils ration so as it is used as an independent variable in the liner regression
model. Finally linear regression model was used to see the impact of small scale
irrigation on house hold food security.

3.2.1 Econometrics Diagnostic Test

Detecting Multicollinearity

One of the basic assumptions in Leaner regression model is existence linear relationship
among some or all explanatory variable of the model. Multicollinearity can be raised from a
lot of factors among others sampling over a limited ranges of values taken by response
variables in the population, constraints in the model or in the population, problem in the
model specification and also multi- co linearity can be happened when regressor in model
share a common trend particularly in time series data.

Since multi-co linearity is essentially a sample phenomenon, arising out of the large non
experimental data collected in social science, there is no unique method of detecting or
measuring its strength. The study applied Variance inflation factor (VIF) to detect strength
of Multi-collinearity. As indicated in (Gujarati, 2004)the larger the value of VIFj, the more”
troublesome” or collinear the variable Xj. As rule of thumb, if the VIF of a variable exceeds
10, which will happen if R2exceeds 0.9, that variable is said be highly collinear. The study
found on average VIF value of the variables to 2.59 and R 2 ranging 0.34 and found no
significant level of Multicollinearity.

In addition, existence of interaction between two qualitative variables was detected as if it

may lead to Multicollinearity or associated problems. In this regard no series problem was
observed in both discrete and continuous variables.

Normality Test

Among several methods of assessing data for normally distribution, statistical and graphical
methods are common. Graphical methods are typical not very useful when the sample size is
small (Gujarati, 2004).Under this assumption, the study used statistical method for assessing
normality distribution of the data under study. The study used Doorkin Hansen test for
normality distribution and found that positive chi-square (x 2) and Probability of chi2= 0.0000
showing for normal distribution of the data under study.

Test for Hetroscedesticity

The study used Breusch-Pagan- Godfrey test to detect for existence of heteroschedesticity.
BPG test avoid limitations of other tests demanding not only depending on the value of the
number of observations to be omitted but also on identifying the correct variable with which
to order the observation. (Gujarati, 2004). Under the assumption of BPG test if the computed
Chi-square (X2) exceeds the critical Chi- square (X2) at a choice level of significance let 1
and 5 percent assumption homoscedasticity is rejected. In this case at 19 degrees of freedom
and 1 and 5 percent significance level the critical values are 36.1908and 30.1435 respectively
while the observed Chi- square (X2) is 32.1093 implying for significance of the value at 5
percent confidence level.

Overall Model Specification Test

A vital issue in estimating program net impact is self-selection bias. Selection bias results
from estimation on a subsample of individuals who have essentially selected themselves for
estimation through their decision to participate in a particular program. A number of authors
have effectively demonstrated that evaluating the impact of an institution or a program on an
outcome variable using regression analysis could lead to biased estimate if the underlying
process which framework. Several techniques have been developed to correct for this bias,
most notably a two stage technique attributed to (Heckman, 1979). And hence, the study
applied Heckman two step procedure to identify the impact of small scale irrigation on
households’ food security from among possible other households food security influencing
factor. Even though irrigated agriculture contributes greater opportunity and reduce rain fed
inherited risks; there are farmers who didn’t participate in irrigated farming. The study
tailored a semi-structured questionnaire and asked why some farmers involved in irrigated
farming others weren’t. To realize the stated objective the study employed probit model here
under the following findings.
3.2.2 Empirical Results of the Binary Probit Model

The starting point for the Heckman selection model the selection function describing the
way resettle households decided to participate in small scale irrigation or not. The model was
fit and estimated using binary probit regression applied to analyze demographic
socioeconomic characteristics, farm and Land Holding specific and institutional factors that
determine participation and extent of participation in small scale irrigation farming. The
model was also used to see impacts of irrigation on house hold food security. The study
indicated that the model was appropriately specified with overall chi-square of 147.04 that
was significant at a 1% significance level / 99 percent significance. Five of the variables
were such as age, sex and family size of the household head, distance from the market, land
holding for replacement and live stock holding of the household were found to be
statistically insignificant with low p value. While the remaining eight variables were found
that highly and moderately significant.

Table 8 Maximum Likely Hood of Binary Probit Model

Coefficie Margina
Variable nts l Effects P - Value
Age of House hold Head- (AGHHD) - 0.008 -0.004 0.632

Sex of Households – (SEXHDD) + 0.35 + 0.012 0.464

Family Size -(FAMSIZ) + 0.10 0.040 0.309

Size of cultivable Land (LANDSIZ) -. -1.07 -0.010 0.016**

Perception of household on Soil Fertility-(SOILFERTI) + 0.54 +0.25 0.000*

Distance from the Water source-(WATDIST) - 0.98 -0.045 0.154

Access to extension Service - (ACCCEXT) 1.05 0.10 0.013**

Distance from the domestic market(DISTMKT)- +0.10 0.005 0.471

Awareness on irrigation regulation- (IREGULA) -0.94 - 0.030 0.094**

Land holding on command area (LHONSITE) 1.72 0.23 0.007*

Land for replacement (REPLND)- -0.26 -0.06 0.339

Livestock holding(LIVESTOCK)- + 0.24 0.01 0.157

Repossession (IREPOSES)- - 1.1 0.001 0.012**

Land productivity (LPDVTY) – +0.10 0.046 0.002*

Modern agricultural input (FARMINPUT)- 0.01 0.0001 0.037**

Source: Analytical Output of the research study, 2019

- *, ** and ***, are significant at 1%, 5%, and 10% level of significance
respectively.
- Number of observation = 339
- Pseudo R2=0.8953
 3.2.3 Outcome Equation
After analyzing determinants factors that affect the decision of sampled households to participate
in small scale irrigated agriculture by maximum likely hood binary probit model, Heckman
selection model is used to analyze the impact of irrigation on household food security. The
inverse mills ratio/hazard function /lamda estimated by Heckman selection model/ in the
selection equation, then it is added to the 2 nd equation as an independent variable so as to capture
the effect of the selection bias. It also brings consistency in estimation of the remaining
coefficient of the equation (Maddala, 1986). The model was found to be statistically significant
at probability of Chi- square /X2/ 0.0000 at 95 percent confidence interval and wild chi-square
1046.90 and lamda was negative in sign with value -1200.95 rho -1.00 & sigma +1200.95,
justifying for existence of selection problem and soundness of the model to proceed into OLS.

As per the model output five variables were found to be statistically significant among others
perception of households on soil fertility, house hold crop productivity per hectare , educational
level of the households and distance of the households from the water source and participation
decision of the households in small scale irrigations are common. The relation of these factors in
influencing households characters to participate on small scale irrigation farm and hence with
households on food security, the significance of the variables and their effect on the output
equation discussed here under the following.

1. Access to irrigation (PARTICIP-Y1): Access to irrigation a dummy variable is 1 if a household

participates in irrigation and 0 otherwise. Access to irrigation is hypothesized to have a positive
relationship with household food security. The study found positive relationship of the farmer’s
access to irrigation with household food security. It is found that as the likely hood of the
household participation decision increased by 1 percent the household food security will increase
by 2921 ETB. The variable was statistically significant at p value 0.000 which shows strong
influence of small scale irrigation on securing the household food demand. This implies that
participation in irrigation would enable farmers to diversify crop production, to enhance
livestock husbandry, practice multiple cropping, and supplement moisture deficiency.
 The study report by (Abonesh, 2007) and (Fekadu, 2012) found similar result supporting the
study finding. Farmers participating in irrigation are likely to increase production, income and
consumption than non-participant farmers. Abonesh (2007), Fekadu (2012) indicated that
irrigation reduces the risk of household food insecurity.
2. Distance from market center (DISMKT): This variable is a continuous variable measured in
kilometer. The study hypothesized negative relationship between distances of the resettled house
hold from the market center with house hold food security. Contrary to the hypothesis positive
relationship was found between distance of the house hold from the market and food security and
statistically significant with p-value 0.075 at 90 percent probability. As distance of the house
hold from the market center increases by 1 km food security of the resettled house hold was
found to increase by Birr 10 and 50 cents. This implies as distance of the resettled household
from the market destination increases the frequency of supplying food products to the market
decreases and the amount food from own production saved for own consumption increases. This
reduces expenditure incurred on food sources demanded from the market by reducing cost of
transportation and related. On the other hand, farmers reduce production of perishable crops and
increase production annual and perennial crops.
The study report by (Abonesh, 2007)disagree with this finding stating that as the further a market
center is the lesser the income from the sale of farm produce, especially for perishable
commodities. To avoid the risk of loss farmers who do not have easy access to markets due to
distance are likely to sell farm products at cheaper price in nearby markets or at farm gate
thereby negatively impacting household income and consequently household food security.
3. Age of the resettled household head (AGEHHD): house hold in rural area relies on agriculture
particularly for their lively hood. Age has a determinant factor to influence the character of
household individuals to pass decision to participate in a given program. By its feature
agriculture is a lobour intensive activity demanding active labour. So as age increases above a
given threshold the decision making ability of a farmer increases whereas lobour productivity
and efficiency declines. Age of household is a continuous variable measured in years and
hypothesized to negative relationship with household food security. The variable was statistically
significant but and positively influence food security of the households with p-value 0.000 at 1
percent level of confidence. As the age of the resettle household increase by 1 year the household
food security is expected to increase by 18.50 ETB. This implies as the age of households
 increase above average year of the respondents food security of the resettled house hold
increases. In the study area resettled household who arrived at host district in time the
resettlement program got the benefit package while the young generation family don not have
such an opportunity. They rely on the land holdings of their parents.
The same study finding was reported by (Nahusenay, 2015) stating that house holding age has
significant effect on household food security.
4. Modern agricultural input (FARMINPUT) - The variable included agriculture input such as
improved agricultural crop varieties, fertilizers and chemicals measured in kilogram and
converted to Ethiopian Birr. It is continuous variable measured in ETB. The variable was
hypothesized to have positive contribution to resettled house hold participation decision in small
scale irrigation and food security. As hypothesized the variable was found to influence the
behavior of resettled household participation decision in small scale irrigation while negatively
related to resettle household food security. It was found to be statistically significant with 99
probability with p value 0.007 Z- score -2.12 and coefficient of relation -1.44. This implies an
increase in ETB in household usage of improved agricultural input decreases the 1.44 Ethiopian
Birr. In the area as more the house hold participated in agricultural input utilization the more
they face a challenging condition in the coming harvest. This is due to failure of the revenue
generated from production recovering costs. As resettled households use more modern
agricultural input the more they decide to participate in a given program including irrigation.
Contrary to the hypothesis the variable was found to have negative impact on household food
security and statistically significant with p- value 0.007 at 99 percent probability. The site
selected for the resettlement program was state farm which highly deteriorated due to continuous
cultivation and utilization of chemicals. Farmers prefer sorghum production to maize production
using improved seed fertilizers such as DAP and UREA. They also prefer livestock husbandry to
grain production and said rather than utilizing cereal seed chemicals and fertilizer better to use
the money for bull and oxen fattening. As they stated the amount of grain harvested from these
input couldn’t cover costs of production and was becoming loss to the farmers by eroding their
asset base. This significantly affected the resettled household food security and adoption of new
technology.
5. Resettled Household Awareness on Irrigation regulation (IREGULA) – Irrigation regulation
is one of the irrigation management tools to manage and guide proper utilization of the irrigation
 water and its maintenance. The variable is dummy that takes 1 if the sampled household has
awareness on irrigation regulation and 0 otherwise. The variable was hypothesized to have
negative influence on the behavior of the households to participate in small scale irrigation and
hypothesized to have positive impact on house hold food security. As hypotheses the variable
have negative impact on the resettled household participation decision in small scale irrigation
and food security and statistically significant in case of the participation decision and household
food security. The study found that as the likelihood of the resettled household participation
decision in small scale irrigation due to awareness on irrigation regulation increases by 1 percent
the resettled household food security will decrease by 344 ETB. The more the household have
more awareness on irrigation regulation the less they secure food security less to his/her counter
part by 344 ETB. This implies as awareness on irrigation of an individual increased by 1 percent
it deducts 344 ETB from his/her pocket challenging food security. This implies that the more the
household develops awareness on irrigation the less they participate in small scale irrigation
farming. Thus, awareness on irrigation management regulation is not the determinant factor for
food security. See table 21 below
6. Crop Diversification: - The study hypothesized that resettled households diversifying crop
cultivation are more likely to participate in small scale irrigation and secure household food than
non-irrigation users. Compliance to the hypothesis the variable have a positive relationship with
the resettled household food security. An increase in one crop type will increase the resettled
household food security by Ethiopian Birr 124 and 50 cents. The variable is statistically
significant at 90 percent probability showing that crop diversification determines resettle
household food security. Resettled households of the study area stated that cultivation of
different crops on rainy season enables to improve soil fertility while in irrigation farm
diversifying crops enable them to increase to have early maturing crops before rainy season. In
doing so the resettled households secure food more than those who does not want to diversify
crops. Table 21 shows determinant factors of affecting the resettled household food security
including crop diversification.
Table 9 Determinant Factors affecting Resettled Household Food Security of the OLS Model

Variable Coefficients P-value

Constant 2,307.34 0.000*
Participation(Y1) 2,921.033 0.001*
AGHDD 18.51 0.005*
SEX -84.86 0.593
LIT 57.46 0.340
SIZELAND -9.54 0.586
NEWAT 217.37 0.241
ACCEXT 229.01 0.112
DMKTC 10.32 0.075**
REGARNESS -344.97 0.035**
LHOLDING 224.09 0.195
LIVESTO 44.22 0.077***
CROPDIV 124.50 0.011**
FARMINPUT -1.44 0.007**
PLOTDIST -0.01 0.592
DEPRATIO -34.27 0.897
Source: Researcher Own survey 2019

Significant at 1% *, ** at 5 %, *** 10% level and

Observation- 339

F(13,316) = 262

Prob= 0.0000

R-squared- 0.9128

Adjusted R –squared= 0.9093

Root MSE= 913.49

4. Summary, Conclusion and Recommendation
4.1 Summary

Agriculture plays indispensable role in Ethiopian economy, sources of food for the people,
supplies of raw material to the industry and creates market for the other economic sectors.
However rain fed dependency, outdated methods of cultivation, low productivity, shortage of
rainfall and pest infestation is the common problem of the Ethiopian agriculture. Regardless of,
significant contribution to the country’s GDP, employment creation and export earnings
Ethiopian agriculture at its infant stage. A number of challenges are confronting the Ethiopian
agriculture among others drought, lack of modern agricultural practices and poor infrastructural
development are common. Rural communities in drought prone area don’t develop strong
capacity copying challenging conditions particularly during food shortage.
To tackle , the government has launched food security enhancing program since 2003 among
which resettlement program, productive safety net program, household asset building program
and complimentary community investment program are common (Guush B, 2013). In the study
area there are resettled communities who have had basic benefit packages set in the program
including farm land and institutional services. With the objective of enhancing the food security
conditions the food insecure people was resettled by government in collaboration of donating
organizations, developing four irrigations schemes and that has been distributed to resettle
households.

Impact evaluation is essentially a problem of missing data, because one cannot observe the
outcomes of program participants had they not been beneficiaries. Without information on the
counterfactual, the next best alternative is to compare outcomes of treated individuals or
households with those of a comparison group that has not been treated.
The purpose of the study was thus, to assess factors that affect resettled household farmers’
participation decision to irrigation for crop production and assessing the impact of small scale
irrigation on resettled household food security. The study was hypothesized to address what
factors determine participation of resettled farmers in small scale irrigation water use? The path
ways through which irrigation contributed to food security and the impacts of small scale
irrigation on the resettle house hold food security.

The study was conducted in Oromia region East Wollega zone Sasiga district in three
resettlement peasant associations on resettled households in case of three small scale irrigation
schemes namely Shekora, Gorrea and Laga Hora small scale irrigation schemes. The study
employed multistage sampling techniques. Three resettled household peasant associations and
operational small scale irrigation schemes were purposively selected while a sample of resettled
households from both irrigation user and non- irrigation users were randomly selected.

Data were collected, organized, analyzed and interpreted adherence to statistical procedure and
existing relevant theories in relation to identified variables utilizing specified models. As stated a
number of factors affect the resettled household participation decision in small scale irrigation.
Since the impacts of small scale irrigation on resettled household food security is the compound
impact of participation in small scale irrigation and outcome equation the study used Heckman
sample selection two step model. As hypothesized, as specified continuous and response
variables affect the likelihood of the household participation decision in small scale irrigation
few of them have deviated from the hypothesis. On the side of the outcome equation the study
shadowed as participation in small scale irrigation had significant impact on resettled household
food security but there are others factors affecting food security status of the resettled
households.

After econometric diagnostic testing, the maximum likely hood of the sampled household
participation decision in small scale irrigation was estimated using binary probit model. The
objective is to see the sign of coefficients of the variables if either increase or decrease the
likelihood of the household participation decision in small scale irrigation. In probit model the
magnitude of the marginal effects of both the dummy and continuous independent variables
show the impact of a given variable on resettled household participation decision in small scale
irrigation. On the other side probit model is used to fit the model for the Heckman selection two
step model.

Eight variables were influence the resettled household participation decision in small scale
irrigation among 2 continuous and 3 dummy variables such as, land productivity and modern
agricultural input utilization and perception of resettled household on soil fertility, access to
extension service and previous land holding on the site selected for irrigation positively influence
the resettled household participation decision in small scale irrigation. Size of land holding and
two dummy variables perception of resettled household on repossession of irrigation land on
default and awareness on irrigation regulations were natively influence the resettled household
participation decision in small scale irrigation.

As the resettled household land productivity increases in quintal and as utilization of modern
agricultural input of the sampled resettled households increases by one Ethiopian Birr the
likelihood of participation decision in small scale irrigation will increase by a given percent
showing for positive influence of the variables due to positive instantaneous rate of change of the
continuous variables. But as the size of land holding of the resettled household increases by a
hectare the participation of resettled household in small scale irrigation will decreases.

After participation variable was estimated by binary probit model on the total sample of
participants and non-participants, the predicted participations were used to calculate the
correction term called inverse mills ration / λ. The correction term λ is the result of rho and
sigma in which sigma is a standard deviation taking on positive values while rho may have two
sign. The study found p-value of lambda (λ) to be 0.046 and negatively signed so that it is
significant at 5% level and the error term in selection and outcome equations are negatively
correlated. The correction term was again entered in the food security model as a predictor then
estimated via OLS on the total sample and used to simulate the net impact of the program.

In the outcome equation, access to small scale irrigation, age of resettled households, distance
from the market center, awareness on irrigation regulation, livestock holding and diversification
of crop, Farm input utilization and access to extension service were found to have impact on
resettled household food security while the remaining variables have no significant influence.
As hypothesized participation in small scale irrigation have a significant impact on resettled
household food security. Awareness of households on irrigation regulation have a negative
impact on resettled food security.

Continuous variables such as age of resettle households, distance from the market, livestock
holding and crop diversification have positively impact the resettle household food security.
While modern farm input utilization was found to negatively influence the resettled household
food security.

4.1.1 Conclusion

The purpose of this study was to investigate the impact of small scale irrigation on resettled
household food security other variables were investigated which also contributed the pass ways
through which households pass decision to either participate or not in small scale irrigations.

Constructive theories briefly emphasized as the behavior of individual to participate in a given

program can be affected by several factors among which demographic, socio- economic and
institutional factors are common. Resettled household participation decision in small scale
irrigation is also believed to be influenced by a number of factors. On the bases of the study
finding household demographic, socio- economic characteristics and institutional factors
influenced the resettled household participation decision in small scale irrigation. In addition,
resettled household with more access to extension, perception on soil fertility, perception on
default repossessing of irrigation land, previous land holding on the site selected for irrigation
and less awareness on irrigation regulation have more tendency of passing decision to participate
in small scale irrigation.

In addition it is identified that the more the resettled household believe that small irrigation
contributed to livestock husbandry the more they participate in small scale irrigation. And an
increase in land productivity, utilization of modern agricultural input have an impact on the
likelihood of the resettle household participation decision. This is an implication for the more the
resettled household expect the land to be productive the more they participate verifying for land
productivity to be a determinant factor on resettled household participation decision in small
scale irrigation. The household that adopt modern agricultural technology have more likelihood
to participate in small irrigation. Hence it is reasonable to say that modern agricultural input
utilization has a positive impact on resettle household participation decision in small scale
irrigation.

According to the outcome equation, access to small scale irrigation, Age of resettled household
, distance from the market center, awareness on irrigation regulation, livestock holding and
diversification of crop and Farm input utilization have a positive impact on resettle household
food security while the remaining variables have a negative impact . Thus, it is reasonable to say
that these variables are determinant factors to affect the likelihood of the household participation
decision in small scale irrigation

Participation in small scale irrigation and awareness on irrigation regulation positively and
negatively affect the resettle household food security respectively. Participation in small scale
irrigation improves frequency of production and harvest which resulted in increase in production.
Awareness on irrigation regulation has a negative influence on the resettle household food
security. This means the more they aware on irrigation regulation the more they hastate to
participate in small scale irrigation.

Continuous variables such as age of resettle households, distance from the market livestock
holding and crop diversification were found to positively impact the resettle household food
security. While modern farm input utilization was found negatively impacting the resettle
household food security.

The final regression output on the household food expenditure regression model revealed that
minimum amount of expense required to meet calorie recommended daily allowance was
Ethiopian Birr 1351.50. However resettled households participating in small scale irrigation
holds 2,899.07 Ethiopian Birr more than non-irrigation users. This shows households who
participated in small scale irrigation secure more food than non-participants.

4.1.2 Recommendation

On the bases of the study result and significance of the variables, recommendations are given
here under two sections. The first section presents a set of recommendations to concerned
government and non-government organizations. The second section offers a set of
recommendations providing suggestions for future researchers to exceed the scope of the study.
4.1.2.1 Recommendation and Policy Implication

All concerned organizations specifically agricultural development office, cooperative promotion

office, Irrigation development office have better strengthen the capacity of resettled farmers and
extension agents through provision of training and awareness creation particularly on modern
agricultural practices , utilization of appropriate input soil and water conservation practices.
Since awareness on small scale irrigation and access to irrigation service are significant in
determining the resettled household’s participation decision in small scale irrigation and resettle
household food security it is advisable to advocate the benefit of irrigated agriculture over rain
fed mode of operation.

Road network was one of the benefit package stated in the settlement program yet the package
was not fully implemented but the variable was significant in determining the resettled
household participation decision. In addition the significance of distances of households from the
market center was identified as it influence the resettled house hold participation decision in
small scale irrigation. Thus concerned unit shall integrally work on the construction and
development of road networks. The soil and water conservation practices has to be put in place
as an intervention package since perception of households on soil fertility and land productivity
found to determine the resettled household participation decision in small scale irrigation

4.1.2.2 Recommendations for Further Researchers

This study provided the first step towards the assessment of small scale irrigation on the food
security of resettled households participating and non-participating in the program. The study
identified and covered major variables determining the resettle household participation decision
in small scale irrigation and the impact small scale irrigation on house hold food security. It is
recommended to assess the overall livelihood of the resettled communities against other resettles
with in the program.

The study found that as the household utilization of modern agricultural input increases the
capacity of the household to meet calorie recommended per day decreases. Suitability and
mineral contents of the soil has to be investigated and recommendation has better given on the
type and rate of fertilizer required to increase land productivity. Contrary to the resettlement
benefit package the resettle household were landed on highly deteriorated land in addition to
susceptibility of the area regional conflict: These are a challenge to annual bread wining effort
of the resettle households: the study therefore recommended further researchers to conduct
detail study.