Rural Out-migration and Its Impact on Crop Production Efficiency of Producers in Southern Ethiopia

Abstract

The research delt with the effect of labour out-migration on output per hectare and efficiency of wheat and teff producers in Kembata-Tembaro and Hadiya zones of Ethiopia based on primary data gathered from 415 random sample of rural households. The multinomial endogenous switching model was used as an analytical model. According to the stochastic frontier model, the mean efficiency of wheat and teff are 82.98 and 66.43 per cent, respectively. According to the econometric results, rural–urban and international migration reduces wheat productivity by 110.94 and 179.11 kg, respectively. The average treatment effect on treated (ATT) also revealed that participation in international migration reduces Teff producers’ technical efficiency by 5.51 per cent. However, teff productivity is reduced by 382.94 and 747.49 kg due to rural urban and international migration, respectively. The lost labour channel of the modern theory of migration is maintained by the result of the study. So as to minimise the negative impact of out-migration in the study area, policymakers must focus more on promoting access to credit, irrigation, land, off-farm employment and public services for rural households.

Keywords

Migration stochastic frontier model switching regression technical efficiency

Introduction

According to World Bank (2020), people move from labour abundant lagging areas to capital abundant leading areas. In the world, the level of remittances from emigrants increased from 128 to 751 billion United State dollars, while international migrants increased from 173 to 281 million people (UNDESA 2021) in the period 2000–2020. To put it another way, while migrants move from lagging to leading areas, remittances move from migrant-receiving destination areas to migrant-sending origin areas. For example, the percentage of remittances to developing countries increased from 57 to 79 per cent between 2000 and 2020 (UNCTAD 2021). Further, the number of internal migrants in developing countries reached 1.3 billion in 2016 (FAO 2020).

Rural out-migration refers to the movement of the two production factors, labour and capital, between rural agricultural and urban non-agricultural sectors. Although the ongoing movement of labour from the rural primary sector to the urban secondary and tertiary sectors, the effect of out-migration on the economies of origin regions has remained a contentious issue (UN 2016). While the pro-migration perspective on rural out-migration contends that remittances boost crop production of households in rural regions (Stark 1985), the pessimistic view on migration sees movement as a loss of human capital to the locations where migrants are sent (Lucas 1987).

Ethiopia is the second and 12th most populous country in Africa and the world, with an estimated 115 million people in 2020 (World Bank, 2021). According to Adugna (2021), migration takes various forms depending on the regime in Ethiopia. Between 1941 and 1974, both internal and international migrations had been minimal in Ethiopia during the era of the emperor (Lyons 2009; Terrazas 2007). However, government policy restricted rural–urban migration, while international migration increased in Ethiopia during the military regime (1974–1991). However, internal and international movements of labour have risen in Ethiopia since 1991 under the present administration. In the period 2000–2020, the number of emigrants from Ethiopia climbed from 611,000 to 1.1 million people, according to World Bank (2021). But, between 1999 and 2021, the share of rural to urban migrants went from 21.6 to 32.2 per cent (CSA 2021).

Moreover, Saudi Arabia, South Africa and United Arab Emirate are the destinations of 31, 12 and 9 per cent of migrants from Ethiopia, whereas rural areas of Oromia, Amhara and SNNP regions are the sources of 42, 27 and 26 per cent of international migrants from Ethiopia, respectively (CSA 2021). Three migratory routes, the eastern route, the northern route and the southern route, are often used by Ethiopian international migrants. Ethiopians use the northern migration route to transit through Sudan to reach Libya and Europe, while they use the eastern migration corridor to get to the Middle East (Massey et al. 1998). From the Horn of Africa to South Africa, there is a southern channel for migration. Southern Ethiopian ethnic groups the Kembata and Hadiya move in large numbers to South Africa (Zewdu 2015). Beginning in 2000, people began migrating from the Hadiya and Kembata zones to South Africa (Kanko et al. 2013). The rate of out-migration is quite high (Tsedeke & Ayele 2017), despite the fact that the movement of labour from Kembata-Tembaro and Hadiya to South Africa just recently started.

Previous studies on the consequence of migration on agricultural output and technical efficacy yielded contradictory results. On the one hand, the stochastic frontier function was applied to measure the effect of migration on output per hectare and efficiency of producers and discovered a direct and significant relationship (Beyene et al. 2017; Mesfin et al. 2021). Odozi et al. (2020) used simple inferential analysis to examine the effect of migration on producer technical efficiency in Nigeria and discovered a positive and significant relationship. Similarly, some authors have quantified the impact of migration on efficiency and output of crop producers by applying the primary data and multiple linear regression method, and discovered an inverse and significant relationship (Adaku 2019; Goldsmith 2017; Huy 2016; Imran et al. 2016; Khanal et al. 2015; Sauer & Gorton 2013). Similarly, Iheke et al. (2013) discovered that migration and technical efficiency are inversely and significantly related using the ordinary least square method.

As a result, while previous studies have looked at the effect of migration on production and efficiency of producers, they found dichotomous results, do not have control endogeneity biases as a result of unobserved characteristics, and they also concentrated on rural–urban migration. However, the study employed the contemporary theory of migration as a theoretical basis and the newly developed endogenous switching model as an analytical model to investigate the link between migration, crop productivity and efficiency of wheat and teff producers in Kembata-Tembaro and Hadiya zones of Ethiopia. The multinomial endogenous switching model can control self-selection biases caused by both known and unknown factors. The remaining sections are listed below. The theoretical and conceptual frameworks are presented in the next section. The subsequent section then includes the materials and methods. This is followed by the section that deals with the results and discussion, and finally conclusion is presented.

Theoretical and Conceptual Frameworks

There are several migration theories, including the gravity theory of migration (Ravenstein 1885), the two-sectors labour migration theory (Lewis 1954), the push and pull factors migration theory (Lee 1966), the Harris-Todaro (1970) migration theory, the modern migration theory (Stark 1985) and the network theory of migration (Taylor & Wyatt 1996), which explain the sources of migration, the effect of migration on migrants, consequences of migration on destinations and source areas. According to the gravity theory of labour mobility, people migrate from low-opportunity areas to high-opportunity areas, and the volume of rural out-migration is determined by the physical distance between the location of sources and destinations. The Lewis’s (1954) two-sectors migration theory states that economic development entails an unlimited transfer of people from primitive primary sector to modern secondary sector.

Lee (1966) divides the causes of rural out-migration into four categories: push factors, pull factors, personal factors and intervening factors. Access to land, non-farm employment, education and basic public services are the drivers of the movement of people from low concentration of economic activities to high concentration of economic activities. Furthermore, the occurrence of drought, crop failure, large family sizes and the presence of returning migrants in the village continues to encourage rural households to migrate. According to the wage differential theory or the Harris-Todaro (1969) migration theory, income differential from employment in agricultural activities and non-agricultural activities is the driving force of labour migration. Furthermore, the Harris-Todaro migration theory asserts that the decision to emigrate is made on a personal level and focuses primarily on the reasons for migration and how it affects migrant-receiving urban areas (Todaro 1969). However, according to the modern theory of migration or the contemporary migration theory (Stark 1985), migration decision is determined at the family level in order to maximise household welfare. Altruism, insurance contracts, loan contracts, investments and inheritance are the four main reasons why migrants send money home (Stark & Bloom 1985).

The main topics of the NELM theory are the causes of labour migration and its effects on welfare and production of farmers in source regions. On the one hand, the lack of capital and insurance markets in remote and backward areas are the causes of labour movement from lagging to leading areas, according to the NELM theory. By placing a family member in the migrant labour market, a rural family can create a new financial intermediary in the form of migrants (Stark & Bloom 1985). The transfer of labour from agricultural sector to non-agricultural sectors has an effect on migrant-sending origin regions via both the lost labour channel and the remittance channel as predicted by the contemporary migration theory. The remittances are expected to improve the well-being of migrant-sending households. However, by reducing human capital and agricultural output in the areas of origin, the lost labour channel may have a negative impact on the welfare of households that send migrants. As a result, migration does not occur in a vacuum, but rather involves giving and taking. According to the modern theory of migration, how migration affects rural areas that send migrants is determined by the relative strength of the remittance effect and the lost labour effect. Finally, network migration theory (Taylor & Wyatt 1996) links rural out-migration to social networks. Connections between migrants, return migrants and non-migrants, according to this migration theory, encourage rural outmigration. As a result, the NELM theory or the modern theory of migration is used as a theoretical framework in this study. The conceptual framework of this study, which shows the causes and links between rural out-migration and crop production of migrant-sending households is presented in Figure 1.

Figure 1.

Conceptual Framework for Migration, Welfare and Production of Rural Households.

Research Methodology

Description of the Study Area

The research was conducted in one of the third largest region in Ethiopia, Southern Nations Nationalities Peoples (SNNP). This region contributes about 10 per cent of Ethiopia’s land area and 20 per cent of its population, respectively. Specifically, the research was carried out in Kembata-Tembaro and Hadiya zones of Ethiopia. The density of population is also higher in Hadiya and Kembata-Tembaro. According to Degelo (2015), Kembata and Hadiya are the major rural to urban and international migrant sending areas in Southern Ethiopia. The total population of Hadiya was 1.7 million people, while the total population of Kembata-Tembaro was about 1 million people. Besides, the total land area of Kembata-Tembaro and Hadiya are 1.4 and 3.6 thousand skm (CSA, 2021). Kembata-Tembaro and Hadiya zones contain 7 and 11 woredas, respectively. Purposively, Angacha woreda was included from Kembata-Tembaro, whereas Soro and Lemo woredas were chosen from Hadiya for the study. The main sources of livelihood in sample districts include crop production and animal husbandry.

Data Types and Collection Tools

This study employed a survey questionnaire to collect quantitative primary data from sample households in southern Ethiopia. After receiving adequate training on the data collection tools and study objectives, 11 enumerators collected survey data. Qualitative data were also gathered through focus group discussions, and key informant interviews from village leaders, school directors, religion leaders, development agents and elders. Besides, data from Ethiopian Statistical Services, the UNDP, WB, FAO and unpublished sources were obtained to provide context for the research.

Sampling Procedures and Sample Size

First, the Kembata and Hadiya in southern Ethiopia are purposefully chosen for the research. According to Zewdu (2015), the Kembata-Tembaro and Hadiya are the major origins of rural–urban and external migrants. There are seven districts in Kembata-Tembaro zone, and 11 districts in Hadiya zone of the SNNP region. Second, while Angacha district was chosen from Kembata-Tembaro, Lemo and Soro woredas were included from the Hadiya purposively for the current study. Nonetheless, according to a study conducted by Kanko et al. (2013) in Kembata-Tembaro and Hadiya, these sample woredas are the primary senders of international emigrants. Furthermore, Kebele is another administrative level under district in Ethiopia. There are 33 kebeles in Lemo, 17 in Angacha and 33 kebeles in Soro woreda. Third, 11 sample kebeles were randomly selected for this research. Accordingly, Shara, Sundusa, Bona and Sonda kebeles were drawn from Soro, while Garba Fandide, Kerekicho and Bobicho kebeles were used from Angacha. Likewise, Haise, Shurmo, Jawe and Sena kebeles were included from Lemo. In Ethiopia, got is the lowest level of administration under Kebele, and there are 147 Gots in the sample Kebeles. Fourth, a random sample of 36 gots was chosen, and a sampling frame, which contains households without migrants, rural to urban and international migrants, was developed. Finally, sample households were included in the study using the proportionate random sampling technique. Although there are various techniques of sample size determination in quantitative research, the Cochran (1963) equation is utilised in this study.

n = \frac{Z^{2} p q N}{e^{2} (N - 1) + Z^{2} p q}

where N, Z, n, p, e and q refer to the total households in sample districts, standard normal distribution, sample size, variability level in the population, precision level and one minus variability level in the population, respectively. In this study, the level of variability, the precision level and the total households (N) were 0.5, 0.05 and 69277, respectively. The size of the sample, which is determined from the above formula, is 383. However, to account for incomplete responses, 10 per cent of this figure was added, and data from 415 completed questionnaires were used in this study. There are 193 households without migrants, 85 households with rural to urban migrants and 137 households with international migrants in this investigation.

Analytical Models

Stochastic Frontier Model

The output and the inefficiency equations can be quantified by applying the two-stage method and the one-stage method of estimation. The two-stage method involves regressing the production frontier to obtain inefficiency values and then regressing the inefficiency values on the various independent variables. The assumption about the inefficiency term in the first-stage estimation technique is the limitation of the two-step method. However, Battese and Coelli (1995) proposed the one-step estimation method, which estimates the output equation and the producer’s inefficiency equation in the same step. As a result, the one-step method of quantifying the production equation and the inefficiency function of farmers, which employs the Cobb–Douglass production function for the output equation and the truncated normal function for the inefficiency equation. Therefore, the SFM, which includes the output equation and the technical inefficiency equation, is as follows:

l n Y_{i} = β_{0} + l n \sum_{j = 1}^{n} β_{j} X_{i j} + ε_{i}

(1)

ε_{i} = v_{i} - u_{i}

v_{i} ~ N (0, σ_{v}^{2})

u_{i} ~ N^{+} (μ_{i}, σ_{u}^{2})

u_{i} = δ_{0} + \sum_{j = 1}^{n} δ_{j} Z_{j i} + ω_{i}

(2)

where $l n$ is the natural logarithm, $Y_{i}$ is the output of wheat or teff per hectare of the $i^{t h}$ household in kilograms, $X_{i}$ is a vector of inputs, $β_{i}$ and $δ_{j}$ are unknown parameters, and $ε_{i}$ is the composite disturbance term, $Z_{j i}$ is a set of independent variables that impact inefficiency and $ω_{i}$ is a random term. Based on Equations (1) and (2), the following empirical Cobb–Douglass production function, and the truncated normal inefficiency function are specified.

l n Y_{i} = β_{0} + β_{1} l n L A N D_{i} + β_{2} ln S E E D_{i} + β_{3} l n F R T L_{i} + β_{4} l n L A B O R_{i} + β_{5} l n O X E N_{i} + β_{6} l n C H E M_{i} + v_{i} - u_{i}

(3)

u_{i} = δ_{0} + δ_{i} Z_{i} + φ_{i}

(4)

where ln is the natural logarithm, $Y_{i}$ is wheat or teff output of the $i^{t h}$ households in kilogram, $L A N D_{i}$ is the land size in hectares, $F R T L_{i}$ is the amount of fertiliser in kilogram, $L A B O R_{i}$ is the labour days, $O X E N_{i}$ are the oxen days, $C H E M_{i}$ is the values of chemicals, $v_{i}$ is disturbance term, $u_{i}$ is the level of inefficiency, $Z_{i}$ is independent variables and $φ_{i}$ is a random disturbance term. Agza et al. (2021) and Ndubueze-ogaraku et al. (2021) discovered that household head’s age and years of schooling are significantly and positively related to technical inefficiency. However, Dessale (2017) obtained that years of schooling of household head has a significant and negative effect on inefficiency of crop producers. Previous research by Belete (2020) and Francis et al. (2020) found that the frequency of extension visits and tropical livestock units has an indirect and significant impact on inefficiency of producers.

However, the empirical literature shows that credit has a mixed effect on technical inefficiency (Francis et al. 2020; Getachew et al. 2020). Abera and Email (2019) discovered an indirect and significant association between farm size and technical inefficiency, and this study hypothesised a negative relationship as well. Similarly, row planting is expected to have an indirect effect on wheat and teff technical inefficiency (Birhanu et al. 2022), whereas the number of plots used to measure land fragmentation is assumed to have a direct effect on technical inefficiency (Ayele & Tarekegn 2021). The dependency ratio, distances from the market, road and health centre are hypothesised to have a direct effect on wheat and teff producers’ technical inefficiency (Dessale 2017).

Furthermore, saving, land fertility and training are thought to have a negative impact on wheat and teff technical inefficiency (Kusse et al. 2019). Furthermore, Lema et al. (2022) discovered that age and age-square are positively and significantly related to producer technical efficiency. Lema et al. (2022) conducted research on determinants of technical efficiency, and found a direct and significant association. But Ndubueze-ogaraku et al. (2021) examined the sources of technical inefficiency of producers, and obtained an indirect and significant relationship between family size and producer technical inefficiency. The description of variables and expected signs are summarized in Table 1.

Table 1.

Description of Covariates, Measurements and Expected Relationship.

Variables	Description	Level of Measurement	Sign
Age of HH	Household Head’ Age	Continuous	$\pm$
Age2	Age Square	Continuous	$-$
EDUC	Education level in years	Continuous	$\pm$
FS	Family size	Continuous	$\pm$
LS	Land size	Continuous	$-$
CREDIT	Dummy variable for credit users	1 for users, & 0 otherwise	$-$
PSNP	Productive safety net users	1 for users, & 0 otherwise	$\pm$
RP	Dummy variable row planting	1 for users, & 0 otherwise	$-$
EXTN	Frequency of extension visits	Continuous	$-$
LF	Dummy for land fertility	1 for fertile, & 0 otherwise	$-$
TLU	Tropical livestock unit	Continuous	$-$
S	Saving of household head	Continuous	$-$
AP	Asset per capita of household	Continuous	$-$
DMRT	Distance from the market, in minutes	Continuous	$+$
DROAD	Distance from the road, in minutes	Continuous	$+$
DHC	Distance from health centre, minutes	Continuous	$+$
LR	Dummy variable for land renting	1 for rent, 0 otherwise	$\pm$
TRAINING	Dummy variable for training	1 for training, 0 otherwise	$-$
PLOT	Number of plots	Continuous	$+$
DR	Dependency ratio	Continuous	$+$
SICKNESS	Dummy for the sickness of a family	1 for sickness, 0 otherwise	$+$
SA	Dummy for sales of animals	1 for sales, 0 otherwise	$\pm$
ANGACHA	District dummy	1 for Angacha, 0 otherwise	$\pm$

Source: Authors’ Compilation (2023).

Multinomial Endogenous Switching Model

The impact of migration on output and efficiency of teff and wheat producers is measured by utilising the newly introduced switching regression in this research. The multinomial endogenous switching model simultaneously estimates the participation and outcome equations. The multinomial logit model that specifies the likelihood of selecting alternative j is given in the first stage by:

P_{i j} = \Pr (M_{i} = j / Z_{i}) = \frac{e^{β_{i} Z_{i}}}{\sum_{k = 1}^{M} e^{β_{i} Z_{i}}}

(5)

A simulation-based switching model is applied to quantify the influence of out-migration on wheat and teff producers’ technical efficiency in the second stage estimation. In this study, the base category is the rural household without a migrant, j = 0. As a result, technical efficiency is defined as the m regime:

Regime 0 : T E_{i 0} = X_{i} γ_{0} + ε_{i 0}, if j = 0

(6)

Regime 1 : T E_{i 1} = X_{i} γ_{1} + ε_{i 1}, if j = 1

(7)

Regime 2 : T E_{i 2} = X_{i} γ_{2} + ε_{i 2}, if j = 2

(8)

where $T E_{i j}$ is the technical efficiency of the i^th household in regime $j$ , $i = 1, 2, 3 \dots N$ , $j = 0, 1, 2 \dots M$ , $X_{i}$ is the vector of other covariates, $ε_{i j}$ is the unobserved factor. Based on Equations (6), (7) and (8), the following selection bias-corrected outcome equations are given.

Regime 0 : T E_{i 0} = X_{i} β_{0} + δ_{0} [ρ_{0} m (P_{i 0}) + \sum_{j} ρ_{j} m (P_{i j}) (\frac{P_{i j}}{P_{i j} - 1})] + ε_{i 0}, i f j = 0

(9)

Regime 1 : T E_{i 1} = X_{i} β_{1} + δ_{1} [ρ_{1} m (P_{i 1}) + \sum_{j} ρ_{j} m (P_{i j}) (\frac{P_{i j}}{P_{i j} - 1})] + ε_{i 1} i f j = 1

(10)

Regime 2 : T E_{i 2} = X_{i} β_{2} + δ_{2} [ρ_{2} m (P_{i 2}) + \sum_{j} ρ_{j} m (P_{i j}) (\frac{P_{i j}}{P_{i j} - 1})] + ε_{i 2} i f j = 2

(11)

where $P_{i j}$ is the probability that the i^th rural household chooses the j^th alternative, $ρ_{j}$ is the degree of correlation between the disturbance term of the participation function, and the disturbance term of the outcome function. Following the work of Bourguignon et al. (2007) and assuming households without migrants, $j = 0$ as the base category, the technical efficiency of producers with migrants is given by:

E (T E_{i 1} / j = 1) = X_{i} β_{1} + δ_{1} [ρ_{1} m (P_{i 1}) + \sum_{k = 1}^{M} ρ_{k} m (P_{i k}) (\frac{P_{i k}}{P_{i k} - 1})]

(12)

E (T E_{i 2} / j = 2) = X_{i} β_{2} + δ_{2} [ρ_{2} m (P_{i 2}) + \sum_{k = 1}^{M} ρ_{k} m (P_{i k}) (\frac{P_{i k}}{P_{i k} - 1})]

(13)

In addition to this, once the actual mean values of technical efficiency of producers are determined using the above two equations, the mean technical efficiency of producers from the counterfactual data is given by:

(T E_{i 0} / j = 1) = X_{i} β_{0} + δ_{0} [ρ_{0} m (P_{i 1}) + ρ_{1} m (P_{i 0}) (\frac{P_{i 1}}{P_{i 1} - 1}) + ρ_{1} m (P_{i 1}) (\frac{P_{i 3}}{P_{i 3} - 1})]

(14)

(T E_{i 0} / j = 2) = X_{i} β_{0} + δ_{0} [ρ_{0} m (P_{i 2}) + ρ_{2} m (P_{i 1}) (\frac{P_{i 1}}{P_{i 1} - 1}) + ρ_{1} m (P_{i 0}) (\frac{P_{i 3}}{P_{i 3} - 1})]

(15)

Finally, by subtracting Equations (14) and (15) from Equations (12) and (13), the conditional average treatment effect on treated (ATT) could be calculated. In estimating the multinomial endogenous switching regression, some variables are used as control variables. As instrumental variables, three variables were used: return migrant, urban house ownership and household religion (Kefelegn 2020; Zhang et al. 2019).

Data Presentations and Discussion

Results of Descriptive and Mean Difference Test

The average technical efficiency of wheat is 82.98 per cent, as it is indicated in Table 2. This suggests that with the current inputs and technologies, rural households can increase wheat production by 17.02 per cent. Minilik (2019) discovered that the mean efficiency of wheat is 78 per cent in Ethiopia’s Amhara region. Similarly, teff’s mean technical efficiency is 66.43 per cent. This also implies that with the current inputs and technologies, households can increase teff production by 34 per cent. According to Fisseha et al. (2021) research in central Ethiopia, the average technical efficiency of teff is 56 per cent. The mean age of household head is 51.3.6 whereas the mean family size and level of education are 7.5 and 4.5, respectively. The mean land size and dependency ratio of households are 0.53 and 0.95 hectares, respectively.

Table 2.

Basic Statistics Continuous and Discrete Attributes in the Study.

Attributes	Average	Std Error	Minimum	Maximum
Technical efficiency of wheat	82.98	14.34	30.71	99.02
Technical Efficiency of teff	66.43	17.35	17.66	93.75
Kilocalorie per adult equivalent	2111.01	1191.57	491.61	7519.76
Age of household head	51.36	9.20	27.00	76.00
Education level of household head	4.46	3.78	0.00	16.00
Household size	7.47	1.40	3.00	12.00
Dependency ratio	0.53	0.59	0.00	3.60
Adult equivalent	6.77	1.40	1.62	9.88
Land size (in hectares)	0.95	0.60	0.13	4.00
Tropical livestock unit	3.84	2.30	0.00	20.06
Extension visits per year	5.06	2.84	0.00	12.00
Gender of households head	Frequency		Per cent
Male	374		90.12
Migration status
With no migrants	193		46.51
Rural–urban migrants	85		20.48
International migrants	137		33.01

The mean kilocalories per capita per day is 2,111.01, which is less than the 2,200 kilocalories per adult equivalent per day required for a household to be food secure. Furthermore, rural–urban migrants and international migrants account for 20.48 and 33.01 per cent of sample households, respectively. Tsedeke and Ayele (2017) observed that 39 per cent of Kembata-Tembaro and Hadiya families include at least one emigrant. Furthermore, in the Hadiya and Kembata-Tembaro zones, 75 and 97 per cent of rural out-migrants are male and unmarried, respectively. Male migrants made up around 72.5 per cent of all emigrants in Kembata-Tembaro and Hadiya, according to Abire and Sagar’s (2016) study on the factors influencing emigration in the research area.

The one-way ANOVA test result showed that the mean efficiency of wheat for households with international migrant members is significantly higher than for households without migrants, as shown in Table 3. Furthermore, at a 1 per cent significance level, the average technical efficiency of teff for households with rural–urban migrants is significantly lower than for households with international migrant members.

Table 3.

One-way ANOVA Test for Technical Efficiency of Wheat and Teff Production.

		Technical Efficiency of Wheat		Technical Efficiency of Teff
Migration (A)	Migration (B)	Mean Difference (A–B)	Std Error	Mean Difference (A–B)	Std Error
No migrants	Rural–Urban	–2.51	1.71	1.29	2.08
No migrants	International	–13.32^***	1.48	–15.96^***	1.81
Rural–Urban	No Migrants	2.51	1.71	–1.29	2.08
Rural–Urban	International	–10.81^***	1.81	–17.25^***	2.21
International	No Migrants	13.32^***	1.48	15.96^***	1.81
International	Rural–Urban	10.81^***	1.81	17.25^***	2.21

Note: ^{***, **} and ^* refer to level of significance at 1%, 5% and 10%.

Estimation Results of Stochastic Frontier Model (SFM)

This study also computed the effect of out-migration on the productivity of wheat and teff growers in rural regions that send out migrants. To that end, the SFM is estimated to measure the value of wheat and teff technical efficiency for sample households. However, prior to estimating the stochastic frontier model, some hypothesis tests were performed using the likelihood ratio and Wald statistics, the results of which are shown in Table 4. To begin, the likelihood ratio test is used to select between OLS and the SFM for the estimation of the production equation. According to Table 4, the calculated likelihood ratio statistics for the production of wheat are 6.07 and 4.41, respectively, at one degree of freedom. The mixed chi-square table’s critical value, however, is 2.71 at the 5 per cent level of significance, suggesting that the null hypothesis supporting the specification of the OLS technique is rejected. In other words, the null hypothesis that there is no inefficiency is not true. Further evidence that the SFM specification is more suited than OLS is provided by the skewness test result utilising residuals from OLS estimate, as shown in Table 4.

Table 4.

Regression Results of the Production Function for Wheat and Teff.

Log of Covariates	Wheat (408)			Teff (389)
	Wald Chi2 (6): 562.62(0.000^***)			Wald Chi2 (6): 215.25(0.000^***)
	Coefficient	Std Error	Z	Coefficient	Std Error	Z
Land	0.2368	0.0447^***	5.30	0.0566	0.0473	1.20
Seed	0.0906	0.0531^*	1.70	0.2221	0.0371^***	5.98
Chemicals	0.0392	0.0270	1.45	0.1413	0.0402^***	3.51
Labour days	0.5560	0.0593^***	9.36	0.3288	0.0695^***	4.73
Fertilisers	0.2281	0.0615^***	3.70	0.1661	0.0539^***	3.08
Oxen days	0.1681	0.0404^***	4.16	0.2700	0.0384^***	7.02
Constant	2.9074	0.3224^***	9.31	2.2740	0.4232^***	5.37
Hypotheses Test Results		Test Statistic	Critical Value		Test Statistic	Critical Value
OLS versus SFM ( $H_{0} : σ_{u}^{2} = 0$ )		6.07	2.71		4.41	2.71
Cobb–Douglass versus translog ( $H_{0} = β_{7} = β_{8} = \dots β_{27} = 0$ )		25.56	39.53		26.76	39.53
Test for inefficiency factors ( $H_{0} = α_{0} = α_{1} = \dots α_{n} = 0)$ )		96.62	33.33		87.87	33.33
Skewness test for OLS residuals		Chi2 = 28.21 (0.003)			Chi2 = 5.10 (0.025)
Wald test for constant return to scale CRS ( $H_{0} : \sum_{i = 1}^{6} β_{i} = 1$ )		Chi2 = 26.81 (0.000)			Chi2 = 3.91 (0.0480)

Note: ^***, ^** and ^* refer to the level of significance at 1%, 5% and 10%.

Second, there are two commonly used SFM specifications: the Cobb–Douglass production function and the translog production function. At a 5 per cent level of significance, the computed likelihood ratio statistics are 25.56 for wheat production and 26.76 for teff production. However, given the critical value of the Kodde and Palm (1986) chi-square distribution is 39.53, the hypothesis that favours the Cobb–Douglass production function is accepted. Third, the likelihood ratio test is used to check the existence of inefficiency in the production of wheat and teff. The calculated likelihood ratio statistics for wheat and teff production functions are 96.62 and 87.87, respectively. At 22 degrees of freedom, the critical value of the mixed chi-square distribution is 33.33, indicating that at least one independent variable affects the inefficiency of wheat and teff producers.

Fourth, the Wald test is applied to know the return to scale in the production of teff and wheat in the research, and the results show that the null hypothesis of constant return to scale is rejected for both wheat and teff production. As a result, the Cobb–Douglass production function was used for the output equation, and the truncated normal distribution was employed for the equation of inefficiency. This is because the use of truncated function for the second equation in the SFM allows for simultaneous estimation of the equations of output and inefficiency using the sfcross Stata command (Battese & Coelli 1995). The unknowns of the output equation and the inefficiency function are estimated simultaneously in a one-step approach. As a result, the Cobb–Douglass production equation and the truncated normal inefficiency equation are estimated concurrently in Stata 16, and the results are shown in Tables 4 and 5. The hypothesis of no inefficiency is rejected since the Wald test is statistically significant at 1 per cent for both wheat and teff production.

Table 5.

Estimation Results of the Truncated Normal Function for Wheat and Teff.

Covariates	Wheat (408)	Teff (389)
Covariates	Coeff. (Std Error)	Covariates	Coeff. (Std Error)
Age of HH	–0.0045 (0.0048)	Age of HH	0.1411 (0.019)^***
Family size	–0.0243 (0.0298)	Size of Family	0.0685 (0.0410)^*
Education of HH	–0.0196 (0.0143)	Education of HH	0.0410 (0.0160)^**
Land size	0.2016 (0.1029)^**	Land size	–0.2148 (0.1243)^*
Extension visits	–0.0509 (0.0215)^**	Extension visits	–0.0545 (0.0216)^**
Row planting	–0.0496 (0.0878)	Row planting	–0.3471 (0.126)^***
Asset per capita	0.00009 (0.00006)	Asset per capita	–0.00003 (0.00002)
Angacha	–0.0203 (0.0993)	Angacha	–0.3092 (0.1520)^**
Land renting out	–0.4336 (0.1665)^***	Number of plot	0.1408 (0.0629)^**
Sales of crop	–0.2115 (0.1177)^*	Sales of Animals	0.2199 (0.1089)^**
Productive SNP	0.3436 (0.1470)^**	Land fertility	–0.2684 (0.1221)^**
Credit	0.2659 (0.1151)^**	Distance from Road	0.0106 (0.0055)^*
Distance from market	0.0089 (0.0037)^**	Distance from Health	0.0057 (0.0034)^*
Tropical livestock	–0.0789 (0.0384)^**	Training	–0.1465 (0.1051)
Sickness	0.4771 (0.3176)	Age-square	–0.001 (0.0002)^***
Saving of HH	–0.2802 (0.1435)^**	Dependency ratio	0.2043 (0.075)^***
Usigma constant	–3.2368 (0.7532)^***		–1.6053 (0.287)^***
Vsigma constant	–2.3064 (0.1108)^***		–2.4669 (0.221)^***
Sigma u	0.1982 (0.0746)^***		0.4481 (0.064)^***
Sigma v	0.3156 (0.0175)^***		0.2913 (0.032)^***
Lambda	0.6280 (0.0867) ^***		1.5385 (0.084)^***

Note: ^***, ^** and ^* refer to level of significance at 1%, 5% and 10%.

Table 4 shows that, at a 1 per cent level of significance, land size, labour days, fertilisers and oxen days are all positively and significantly associated with wheat output. In Hadiya and Kembata-Tembaro zones, for example, a 1 per cent increase in fertiliser use increases wheat output by 0.228 per cent, whereas a 1 per cent increase in labour days increases wheat output by 0.556 per cent. This suggests that labour and fertiliser are being used inefficiently per hectare in the study area. Similarly, at a 1 per cent level of significance, seed, labour days, fertilisers, chemicals and oxen days are positively and significantly related to teff output. A 1 per cent increase in u-days, fertiliser-days and oxen-days, for example, increases output of teff by 0.329, 0.166 and 0.270 per cent in the study area, respectively. Labour, land, fertiliser and oxen-days are all positively and significantly associated with output per hectare (Francis et al. 2020; Getachew et al. 2020; Minilik 2019). Some studies, on the other hand, discovered a negative and significant impact of seed, chemicals, labour and fertilisers on crop production (Isyanto et al. 2021; Ndubueze-ogaraku et al. 2021).

The results of the estimations of the factors contributing to the technical inefficiency of wheat and teff producers are presented in Table 5. The frequency of extension visits has an indirect and significant 5 per cent. This suggests that a rise in the number of development agents’ visits reduce the technical inefficiency of citrus and wheat producers alike. The fact that the tropical livestock unit’s coefficient is also indirect and significant at 5 per cent suggests the output of wheat and livestock in the study area complements one another. Frequency of visits by development agent and the number of tropical livestock units were indirectly and significantly associated with inefficiency of producers in previous researches on factors affecting the technical efficiency (Belete 2020; Francis et al. 2020).

Similar to this, the household saving dummy’s coefficient is negative and significant at 5 per cent. The slope of credit use is also significant and positive at 5 per cent. This might occur if credit users use the money they borrowed for consumption or other unproductive activities. Additionally, because productive safety net programmes have a positive and significant coefficient, it is possible that users of these programmes will use the resources they receive from them for non-productive purposes. Ayele and Tarekegn (2021) and Getachew et al. (2020) obtained an indirect and significant relationship, whereas a study by Francs et al. (2020) found a direct and significant relation between credit and technical inefficiency. The land size coefficient is positive and significant at 1 per cent. A study by Abera and Email (2019) found a negative and significant correlation between the total area under cultivation and producers’ technical inefficiency, in contrast to studies by Belete (2020) and Francis et al. (2020), which found a positive and significant association.

Additionally, the dummy coefficient for renting out land is negative and significant, whereas the dummy coefficient for distance from the local market is positive and significant at 1 per cent. Similar to how extension visits, land size, row planting, district dummy and land fertility are factors that reduce teff inefficiency, Table 5 shows that age of household head, education, number of plots, dependence ratio, distance from road and distance from health centre are factors that increase teff inefficiency. The positive and significant coefficient of education level of household suggests that more educated farmers are more likely to work outside the farm. This conclusion agrees with research done by Agza et al. (2021) and Ndubueze-ogaraku et al (2021). Teff’s technical inefficiency of producers is directly and significantly related with the number of plots that demonstrate the degree of land fragmentation, even though the use of row planting significantly reduces technical inefficiency. This is consistent with research from Birhanu et al. and Ayele and Tarekegn (2021, 2022). As anticipated, the inefficiency of teff producers is positively and significantly correlated with dependency ratio, distance from the neighbourhood market and distance from the health centre. The information and knowledge obtained from development agents may reduce the inefficiency of teff producers, according to the significant and negative coefficient of the number of visits by development agent. Abeje (2021) also discovered a negative and significant correlation between producers’ technical inefficiency and extension visits.

Estimation of Results of Multinomial Endogenous Switching Model

The regression outputs of the consequence of migration on efficiency of wheat and teff are shown in Table 6. By participating in rural–urban migration or international migration, the productivity of wheat is reduced by 110.94 and 179.11 kg, respectively. These reductions are statistically significant at 5 per cent. In a similar vein, the average treatment effect on treated (ATT) revealed that taking part in international migration reduces the technical efficiency of Teff producers by 5.51 on average and is statistically significant at 1 per cent. To put it another way, if they had not participated in international migration, teff producers’ technical efficiency would have increased by 5.51 per cent.

Table 6.

Effect of Migration on Productivity and Efficiency of Wheat and Teff Producers.

		Choices	Migration Status		Treatment Effect
Outcomes			Migrate	Not Migrate	$ATT$	Std Error
			Actual	Counterfactual
Wheat	Efficiency	Rural–urban	80.59	79.01	1.58	1.83
	Efficiency	International	91.37	92.55	–1.18	1.71
	Productivity	Rural–urban	1,767.974	1,878.918	–110.94	72.75^***
	Productivity	International	2,373.047	2,552.154	–179.11	89.18^***
Teff	Efficiency	Rural–urban	60.45	62.67	–2.22	1.84
	Efficiency	International	77.25	82.75	–5.51	1.38^***
	Productivity	Rural–urban	1,236.29	1,619.23	–382.94	99.41^***
	Productivity	International	1,713.27	2,460.768	–747.49	81.61^***
Base Heterogeneity			$B H_{1}$	$B H_{0}$	$T H$	Std Error
Wheat	Efficiency	Rural–urban	4.10 (1.69)^***	1.06 (1.74)	3.04	0.50^***
	Efficiency	International	12.85 (0.94)^***	14.60 (1.77)^***	–1.75	1.15
	Productivity	Rural–urban	352.56 (55.5)^***	283.06 (68.7)^***	69.50	60.25
	Productivity	International	574.9 (71.5) ^***	956.29 (72.5)^***	–381.42	87.40^***
Teff	Efficiency	Rural–urban	3.09 (1.74)^***	1.17 (1.59)	1.92	1.13^***
	Efficiency	International	14.2 (1.15)^***	21.25 (1.42)^***	–7.02	0.63 ^***
	Productivity	Rural–urban	345.3 (101.6)^***	489.26 (67.1)^***	–143.95	66.64 ^***
	Productivity	International	401.3 (51.32)^***	1,330.8 (71.3)^***	–929.46	66.72 ^***
Wald Test		$chi 2 (6)$	$= 82.88$	$Prob > χ^{2} = 0.000$
LR Test		$LR Chi 2 (6)$	$=$ 230.27	$Prob > χ^{2} = 0.000$

Note: ^***, ^** and ^* refer to level of significance at 1%, 5% and 10%.

Additionally, the ATT shows that involvement in international migration and rural urban migration decreases teff productivity by 747.49 and 382.94 kg, respectively. These reductions are statistically significant at 1 per cent. One the one hand, the technical efficiency of participants minus the technical efficiency of non-participants if they would have participated makes up the base heterogeneity for participants ( $B H_{1}$ ). Therefore, non-participant households’ technical efficiency of wheat and teff would have been lower by 4.10 and 3.09, respectively, and statistically significant, had they participated in rural–urban migration.

Similarly, the non-migrant households’ efficiency of wheat and teff in Hadiya and Kembata-Tembaro would have decreased by 12.85 and 14.24 per cent, respectively, if they had taken part in international migration. This difference is statistically significant. This implies that the negative effects of participation in rural out-migration are greater for non-participants than for participants. Contrarily, the base heterogeneity for non-participants ( $B H_{0}$ ) is the difference between the technical efficiency of participant households if they had not participated minus the technical efficiency of non-participant households. In comparison to non-participants, participants’ technical efficiency in producing wheat and teff would have been 14.60 and 21.25 per cent higher if they had not participated in international migration. This difference is statistically significant at 1 per cent.

Studies by Khanal et al. (2015) and Adaku (2019) found that rural–urban migration has an indirect and significant impact on crop productivity and efficiency. However, rural urban migration has a direct and significant effect on agricultural production (Beyene et al. 2017; Mesfin et al. 2021; Odozi et al. 2020). As a result, the study’s findings are consistent with the labour migration theory of the new economy’s lost labour hypothesis. Last but not least, a falsification test was applied to check whether the instruments were appropriate in accordance with Manda et al. (2021), and the test result revealed that the chosen instruments are valid.

Conclusion

Numerous earlier studies have looked at the effects of labour migration on urban destination areas, but there are few studies looking at the effects on crop productivity and producer technical proficiency in migrant-sending rural areas. As a result, this study used the switching regression as an analytical model and the NELM theory as a theoretical framework to quantify influence of migration on the output per hectare and deviation of output from the production frontier of wheat and teff producers. A sample of 415 households was selected for the purpose of the research using a stratified random sampling method.

According to the stochastic frontier model’s estimation results, producers of wheat and teff in the study area have average technical efficiencies of 83 and 66 per cent, respectively. This suggests that farmers in the study area can increase their output of wheat and teff by 17 and 34 per cent, respectively, given the current level of inputs and technological mix. According to the study’s findings, taking part in international migration reduces teff producers’ technical efficiency by an average of 5.51 per cent, which is significant at 1 per cent. Similar to domestic migration, participation in international migration lowers wheat productivity by 5.90 and 7.02 per cent at 1 per cent, respectively. Similar to domestic migration, participation in international migration lowers teff productivity by 30.38 and 23.65 per cent, respectively. These reductions are significant at 1 per cent level. These imply that productivity and efficiency of farmers in Hadiya and Kembata-Tembaro are significantly reduced due to migration. This is due to the fact that migration is a two-way transaction in which one hand gives and the other takes. While the lost labour channel may reduce crop productivity and efficiency of producers in source areas, the capital channel tends to improve crop productivity, technology adoption and welfare of remittance-receiving households. In light of this, the finding of this study supports the contemporary migration theory, which postulates that the continual withdrawal of labour from rural areas lowers productivity of the primitive sector by reducing the labour supply, work efforts of remittance-receiving households and rural human capital.

To at least lessen the rate out-migration in the study area, policymakers should encourage rural youth to pursue careers in agriculture through commercialisation, mechanisation and encouragement to remain in their home communities. In addition, increasing agricultural productivity, off-farm employment opportunities, credit availability and the provision of essential public services in rural areas where migrants are sent will lessen rural outmigration. Dependency on primary data and dearth of sufficient secondary data on rural–urban and international migration are two of this study’s limitations. The impact of rural outmigration on the labour market, welfare, vulnerability, land market, income inequality and the adoption of agricultural technology in migrant-sending origin areas are additional future research possibilities.

Footnotes

Acknowledgements

The authors are highly grateful to all data collectors, and administrators of sample Kebeles of the Angacha, Soro and Lemo Districts for their kind contributions for the completion of this study. The authors would also like to appreciate Haramaya University and Arba Minch University for providing the required materials and financial supports for this research work.

Data Availability Statement

Data related to this study will be provided by the authors on request.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Ethical Approval

This study was approved by the Directorate for Postgraduate Program of Haramaya University and School of Agricultural and Agribusiness in Haramaya University.

Funding

African Economic Research Consortium (AERC) provided financial supports for this research.

ORCID iD

Fassil Eshetu

References

Abebaw

, Admassie

, Kassa

, and Padoch

. 2019. Does rural outmigration affect investment in agriculture? Evidence from Ethiopia. Migration and Development 10, no. 1, 144–168. https://doi.org/10.1080/21632324.2019.1684046

Abeje

Aschalew

. 2021. Causes and effects of rural–urban migration in Ethiopia: A case study from Amhara Region. African Studies 80, no. 1, 77–94. https://doi.org/10.1080/00020184.2021.1904833

Abera

, and Email

. 2019. Technical efficiency of smallholder teff production. Open Access Journal of Agriculture Research 2019, no. 01, 1–6.

Abire

, and Sagar

. 2016. Determinant factors of illegal migration to South Africa and Its impacts on the Society in case of Gombora district, Hadiya zone in Ethiopia. IOSR Journal of Mathematics 12, no. 3, 51–65.

Adaku

Abigail Ampomah

. 2019. The effect of rural–urban migration on agricultural production in the northern region of Ghana. Journal of Agricultural Science and Applications 2, no. 4, 193–201.

Adugna

Girmachew

. 2021. Migration patterns & emigrants’ transnational activities: comparative findings from two migrant origin areas in Ethiopia. Comparative Migration Studies 1, 1–28.

Agza

, Alamirew

, and Shibru

. 2021. Crop producers’ technical efficiency and its determinants in the Gurage zone, Ethiopia: A comparative analysis using rural–urban migration as a parameter. Cogent Social Sciences 7, no. 1. https://doi.org/10.1080/23311886.2021.1995996

Assefa

, Meded

, Alemayehu

, Mulubrhan

, Ibrahim

, and Abduselam

. 2020. Determinants of productivity and efficiency of teff production in southwest Ethiopia. Research Square 1–26.

Ayele

, and Tarekegn

. 2021. Comparative analysis of technical efficiency of wheat production in row planting and broadcasting methods in southern Ethiopia African Journal of Science, Technology, Innovation and Development 14, no. 3, 708–717.

10.

Battese

G. E.

, and Coelli

T. J.

. 1995. A model for technical inefficiency effects in a stochastic frontier production function for panel data. Empirical Economics 20, no. 2, 325–332.

11.

Belete

Aschalew Shiferaw

. 2020. Analysis of technical efficiency in maize production in Guji Zone: Stochastic frontier model. Agriculture and Food Security 9, no. 1, 1–15. https://doi.org/10.1186/s40066-020-00270-w

12.

Beyene

A. D.

, Mekonnen

, Kassie

, and Di

. 2017. Determinants of adoption and impacts of sustainable land management and climate smart agricultural practices (SLM-CSA) . Discussion Paper, Ethiopian Development Research Institute, Addis Ababa, Ethiopia.

13.

Bhandari

, and Reddy

B. V. C.

. 2015. Impact of out-migration on agriculture and women work load: An economic analysis of hilly regions of Uttarakhand, India. Indian Journal of Agricultural Economics 70, no. 3, 396–404.

14.

Birhanu

F. Z.

, Tsehay

A. S.

, and Bimerew

D. A.

. 2022. Cereal production practices & technical efficiency among households in major “ teff ” growing mixed farming areas of Ethiopia. Cogent Economics & Finance 10, no. 1, 2012986. https://doi.org/10.1080/23322039.2021.2012986

15.

Bourguignon

, Fournier

, and Gurgand

. 2007. Selection bias corrections based on the multinomial logit model. Journal of Economic Surveys 21, no. 1, 174–205.

16.

Degelo

2015. Socio-economic impacts of migration of Ethiopians to the South Africa. The case of Kembata-Tembaro and Hadiya zones . MSc Thesis, Addis Ababa University, School of Graduate Studies, Department of Political Science and International Relations.

17.

Dessale

Moges

. 2017. Technical efficiency in teff production: The case of smallholder farmers in Jamma District, South Wollo Zone, Ethiopia. Journal of Economics and Sustainable Development 9, no. 21, 86–93.

18.

Kusse

, Jema

, Bosena

2019. Sources of technical inefficiency of smallholder farmers in sorghum production in konso district of southern Ethiopia. International Journal of Agricultural Extension Education 5, no. 1, 180–196.

19.

FAO (Food and Agriculture Organization). 2020. Migration Framework – Migration as a Choice and an Opportunity for Rural Development. Rome, Italy.

20.

Fisseha

, Abrham

, and Dawit

. 2021. Cereal production practices and technical efficiency among farm households in major “teff” growing mixed farming areas of Ethiopia. Cogent Economics and Finance 10, no. 1, 1–34. DOI: 10.1080/23322039.2021.2012986

21.

Francis

L. A.

, Samuel

C. M.

, and Samuel

N. N.

. 2020. Technical efficiency and its determinants in sugarcane production among smallholder sugarcane farmers in Malava sub-county, Kenya. African Journal of Agricultural Research 15, no. 3, 351–360. https://doi.org/10.5897/ajar2020.14703

22.

Getachew

, Ketema

, Goshu

, and Abebaw

. 2020. Technical efficiency of wheat producers in North Shewa Zone of Amhara Region, Central Ethiopia. Sustainable Agriculture Research 9, no. 3). https://doi.org/10.5539/sar.v9n3p77

23.

Girma

Haile

. 2014. Estimation of technical efficiency of dairy farms in the central zone of Tigray National Regional State. Heliyon 5, 1–24.

24.

Goldsmith

, and Ndarishikanye

. 2017. Rural–urban migration and agricultural productivity in Senegal. Agricultural Economics 31, 33–45.

25.

Houngue

, and Nonvide

G. M. A.

. 2020. Estimation and determinants of efficiency among rice farmers in Benin. Cogent Food & Agriculture 6, no. 1, 1819004. https://doi.org/10.1080/23311932.2020.1819004

26.

Huy

Huynh Truong

. 2016. Effect of labor migration on technical efficiency of rice in Vietnam . School of Economics & Business Administration, Tho University, Vietnam.

27.

Iheke

O. R.

, Nwaru

J. C.

, and Onyenweaku

C. E.

. 2013. The impact of Migrant Remittances on the Technical Efficiency of Arable Crop Farm Households in South-Eastern Nigeria. Paper presented at the 4th International Conference of the African Association of Agricultural Economists, September 22–25, 2013, Hammamet, Tunisia.

28.

Imran

, Khuda

, and Hassan

. 2016. Rural to urban migration & crop productivity in Pakistani Punjab. Akdeniz Üniversitesi Ziraat Fakültesi Dergisi 29, no. 1, 17–19.

29.

Isyanto

A. Y.

, Sudrajat

, and Yusuf

M. N.

. 2021. Determinants of technical in-efficiencies in swamp rice farming - Ciamis district, Indonesia. E3S Web of Conferences 226. https://doi.org/10.1051/e3sconf/202122600005

30.

Jondrow

, Lovell

K. C. A.

, Materov

I. S.

, and Schmidt

. 1982. On the estimation of technical inefficiency in the stochastic frontier production function model. Journal of Economics 19, 233–238.

31.

Kefelegn

Habtamu.

2020. Determinants of youths rural to urban migration in Ethiopia: The case of Addis Ababa). Academic Journal of Research and Scientific Publishing 2, no. 15, 5–7.

32.

Kanko

T. D.

, Bailey

, and Teller

. 2013. Irregular migration: Causes and consequences of young adult migration from Southern Ethiopia to South Africa . Paper presented at XXVII IUSSP International Population Conference, Busan, South Korea, pp. 26–31.

33.

Khanal

, Victoria

, Alam

, and Regmi

P. P.

. 2015. Implications of out-migration in rural agriculture: A case study of Manapang. Journal of Developing Areas 49, no. 1, 331–352.

34.

Lee

Everett S.

1966. A theory of migration. Demography 3, 47–57.

35.

Lema

T. Z.

, Masresha

S. E.

, and Neway

M. M.

. 2022. Analysis of the technical efficiency of barley production in North Shewa Zone of Amhara regional state, Ethiopia. Cogent Economics & Finance 10, no. 1. https://doi.org/10.1080/23322039.2022.2043509

36.

Lewis

1954. Economic development with unlimited supplies of labor. Manchester School of Economic and Social Studies 22:139–191.

37.

Lucas

Robert.

1987. Emigration to South Africa’s mines. The American Economic Review 77, no. 3, 313–330.

38.

Lyons

Terrence.

2009. The Ethiopian Diaspora and Homeland Conflict. Proceedings of the 16th International Conference of Ethiopian Studies. https://www.semanticscholar.org/paper/The-Ethiopian-Diaspora-and-Homeland-ConflictLyons/5c7f75935d60b8126dad6a933811dae241935ee0

39.

Massey

D. S.

, Arango

, Hugo

, Kouaouci

, Pellegrino

, and Taylor

J. E.

. 1998. Worlds in Motion. Understanding International Migration at the End of the Millennium. Clarendon Press.

40.

Mesfin

, Bamlaku

, and Admasu

41.

Minilik

Tsega

. 2020. Analysis of levels and determinants of technical efficiency of wheat-producing farmers in the Amhara region of Ethiopia. International Journal of Development Research 9, no. 11, 1–5.

42.

Ndubueze-ogaraku

M. E.

, Adeyoola

O. A.

, and Akuchinyere

. 2021. Determinants of technical efficiency of small-holders yam farmers in Nigeria. Review of Agricultural and Applied Economics 24, no. 1, 13–20. https://doi.org/10.15414/raae.2021.24.01.13-20

43.

Odozi

, Adeniyi

, and Yusuf

. 2020. Production efficiency in agriculture in Nigeria. Economics of Agriculture 67, no. 2, 2315–327.

44.

Ogundari

Kolawole

. 2013. Crop diversification and technical efficiency in food crop production: A study of peasant farmers in Nigeria. International Journal of Social Economics 40, no. 3, 84–91. https://doi.org/10.1108/03068291311291536

45.

Ravenstein

Ernest George

. 1885. Laws of migration. Journal of the Statistical Society of London 48, 167–235.

46.

Reifschneider

, and Stevenson

. 1991. Systematic departures from the frontier: a framework for the analysis of farm inefficiency. International Economic Review 32, no. 3, 71523. https://doi.org/10.2307/2527115.

47.

Sauer

, and Gorton

. 2013. Migration and agricultural efficiency – Empirical evidence for Kosovo. Agricultural Economics 46, no. 1, 629–641.

48.

Stark

Oded

. 1985. The new economics of labor migration. The American Economic Review 75, no. 2: 173–178.

49.

Sun

, Hu

, and Zhang

. 2021. Impact of rural–urban migration experience on rice farmers’ agricultural machinery expenditure. Agriculture 11, no. 764, 1–14.

50.

Zhang

, Hill

, Tao

, Sciences

, and Huang

. 2019. Determinants of out-migration in rural China: Effects of payments for ecosystem services. Population-Environment 40, no. 2, 182–203. https://doi.org/10.1007/s11111-018-0307-5