Socioeconomic Determinants of Youth Unemployment in Ethiopia, the Case of Wolaita Sodo Town, Southern Ethiopia

According to UN Secretariat, UNESCO,UNICEF, WHO, UNFPA and ILO youth are defined these individuals whose age is in between 15-24years, while for UN Habitat (Youth Fund) and the African youth charter, the age range is from 15-32 and 15-35, respectively. For others like organization for Economic Co-operation and Development (OECD) defines youth as those men and women in between 15 and 29 years of age.. To achieve the objective of this paper is to examine the socio-economic determinants of urban youth unemployment logistic model was used. A simple random sampling technique employed to select 395 youth were used. Both quantitative and qualitative data were collected from sampled youth unemployed persons through questionnaire and interview. Both descriptive and econometric data analyses techniques were applied. From the econometrics result revealed that confidence ,education status , access to information human related factor, institutional factors ,socioeconomic and demographic factors were significantly affect the youth unemployment status in the study area . The result also indicates that the higher months spent on searching, the higher the probability of getting job. This shows the cost of job search is positively related with employment status of the youth, as cost of search increases the higher the probability of being employed. The result also indicates that educational level of the youth’s household head has a negative and statistically significant effect on employment status of the youth at 10 percent level of significance. This finding is inconsistent with the general fact and the hypotheses of this study. Since confidence level of the youth matters for employment status of the youth, the concerned bodies should develop the confidence level of the youth by providing different trainings for the youth. The concerned body, especially the government should provide facilities related with information access like mass media and magazines in the study area. : employment, Youth unemployment Binary logistic regression.

The research findings on the factors that determine youth unemployment may be due to a number of differences in these studies such as definitional differences, methodological differences, socio-economic difference and spatial differences. Different results can be obtained when youth population is defined differently in different studies (World Bank, 2005). Hence this study will consider only these individuals whose age is in between 15-35 years. Even when the definitions of youth age have been homogeneous, different results can still be obtained if different methodologies have been used. However, this research is intended to examine determinants of youth unemployment using cross sectional survey data rather than secondary data by specifying logit model approach. Once again, the issue of youth unemployment may differ from town to town, which calls for studying the situation for each town separately rather than making generalizations based on the studies in few urban centers.

Objectives of the Study 2.1 General Objectives
The general objective of the study was to identify the socioeconomic determinants of youth unemployment in case of wolaita sodo town. 2.2 Specific Objective 1. To identify socio-economic factors of youth unemployment in Wolaita Sodo town 2. To identify the demographic determinants of urban youth unemployment at the study area. 3. To assess the major characteristics (in terms of age composition, educational level, gender, migration status of the youth of youth unemployment in Wolaita Sodo town

RESEARCH METHODOLOGY 3.1. Description of the Study Area
The Southern Nations Nationalities of Peoples Region covering an area of 111,000km 2 accounting for 10percent of the total area of the country, SNNP region is home for more than 56 ethnic groups. It is located at the southern and southwestern part of the country. The region shares common borders with Sudan in the west, Kenya in the south, Gambella region in the North West and Oromia region in the east and North. The region which is the most diverse in ethnic and linguistic composition has a population of approximately 15 million, the average plots size is 0.4 household. There are 126 woredas in this region, of which 8 are Special Woredas. The region has diverse ecology and socio-economic profile. Out of the total area of land 56 percent is lowland, which accommodates all of the pastoral and agro-pastoral communities of the region. The proportion of land occupied by pastoral and agro-pastoralists is estimated to be 34 percent of the region scattered over three administrative zones (South Omo, Bench-Maji and Kefa) 4.6 percent of the total population of the region is found in these areas(CSA, 2013).

Types and Sources of Data
This study was conducted based on the procedures of mixed methods approach. This means, a mixture of data collection and analysis techniques used from both the quantitative and qualitative research methods. The mixed approach is preferred for this study since it enables to use different methods of data collection and analysis for addressing the problem effectively. Again, survey design is the kind of the research design used in the study. Specifically, cross-sectional survey design was employed for this study. The study was conducted based on both primary and secondary data sources. Primary data was collected through questionnaire, interview, focus group discussion and field observation; while the secondary data will be collected from review of related literature, government office reports, NGOs, research papers, books, journals and other related documents.

Sampling Techniques and Sample Size Determination
The sample size was systematically determined by using sampling technique and based on the simple random sampling method, from three sub-towns of the town.Namely, Mehal, Arada and Merkato sub-towns. Yamane's sampling computation method is adopted with 5percent of error term and 95percent confidence interval to determine 395 sample size from the total population 33,223 of youths dwelling at Wolaita Sodo town. Thus, according to Yamane's formula: = , Where, n = is the sample size N = is total population = 33,223 youth population e = is sampling error (sample error = 5%) d.f = degrees of freedom (confidence interval = 95%) Therefore, n= , , .
= 395 , thus the total sample size of 395 youth population was distributed proportionally among the three sub towns.

Data Collection Instruments
In this study, different instruments were used for gathering data about the determinants, status, socio economic and demographic characteristics status of youth unemployment. Primary and secondary data collection instruments were was used in order to capture these variables. Qualitative and quantitative data were collected through questionnaire; interviews and group discussion were used to collect the data.

Methods of Data Analysis
The study was employed both descriptive and econometrics data analyses to achieve the objectives. The data analysis was based on both qualitative and quantitative methods. The collected data was analyzed by descriptive statistics and econometrics method of analysis. In descriptive data analysis the researcher used percentage, mean, mode, standard deviation, frequency and chi-square test, F-test, charts and figures. In econometrics method of data analysis, the researcher adopted logit model by taking youth unemployment status as the probability of participation in the labor market. Hence participation assumes a binary response variable and taking a value 1 for these youths who are participated in the labor market and zero otherwise.

Econometric Analysis
The objective of this paper is to examine the socio-economic determinants of urban youth unemployment in Wolaita Sodo town. Thus, it needs to identify the probability of youth population in the participation of urban labor market. Thus, participation assumes binary outcome, whether participated in urban labor market or not (being unemployed. In other hand our dependent variable assumes youths dwelling in urban centers are either employed or unemployed in urban labor market. According to Gujarati (2004) the logistic or logit model could be written in terms of the odds ratio and log of odds ratio, which enable one to understand the interpretation of the coefficients. In this study, the odds ratio is the ratio of the probability that the youth will be unemployed (Pi) to the probability that he/she will be employed (1-Pi).
Since Zi is ∝+βiXiwe can rewrite the above equation as Taking the natural logarithm of equation (4) we have the following: Where, Yi is our dependent variable either the youth is employed or not in the labor market. K: the number of explanatory variables included in the model Xi: are vectors of all explanatory variables βi: the coefficient or the parameter to be estimated in the model Pi: is the probability that the youth is employed in the labor market 1-Pi: is the probability of failure or the probability that the youth is unemployed Ui: is the disturbance (error) term showing the effect of other variables (other than the included variables) on our dependent variable.

Independent variables Sex:
Dummy for sex of the youth: 1= male and 0 = female Age: D continuous for the age of the youth in years Monthly search : number of months the youth spent in searching job at sodo town Education: Continuous for the highest grade of formal schooling attained by the youth in years Occupation head: A dummy for the occupation of the youth's household head as 1=farmer, 2=government employee, 3= self employed and 4=others Aces to information : The level of access to vacancy position information as 1= low, 2=medium, and 3=high Confidence: Dummy for the youth being confident in his/her available skills: 1=yes and 0=no Drug: A category for either the youth is addict of any type of drug or not: 1=yes; 0=no Information access: Access to job information and vacancy announcements by youth (1=yes; 0=no) Experience: Previous experience in any type of job in years Education head: Continuous for the highest grade of formal schooling attained by the youth's household head in years.

Income of parents:
Monthly income of youth's parent in Ethiopian birr. Institutional factors: A likert for institutional factors that determines youth's employment status. Economic factor: A likert for economic factors that determines youth's employment status. Human capital factors: A likert for human capital related factors that determine youth's employment status. Demographic factors: A likert for demographic factors that determines youth's employment status. Youth related factors: A likert for youth's specific characteristics that determines youth's employment status.

Results and Discussion
In this section the data obtained of the study were presented by using descriptive statistics such as mean, percentage, standard deviation and t-test and chi-squares-test to obtain an overview of the respondents' socioeconomic and demographic characteristics of the households with respective to youth unemployment .

Descriptive Data Analysis
As displayed in table .2 employment statuses is collected on a total of 395 youths aged between 15-35 years old. Thus, the result shows that about 254 (64.30 percent) of the youth are unemployed and the remaining 141 (35.70 percent) of the youth are employed as shown in the table below in the study area.

Gender versus unemployment status of the study
As shown in table 3 indicates that unemployment status of the youth and gender of the youth. It indicates that out of the total female respondents 70percent and 30percent of them are unemployed and employed, respectively. While among male respondents 63.7 percent are unemployed and 36.3 percent of them are employed. This shows that females are more likely to be unemployed than males. Before entering the variables in to the model, the multi-co linearity problems were checked in terms of variance inflation factor (VIF) for continuous and contingency coefficients for dummy and discrete variables respectively. As a rule of the thumb, when the variables having VIF values less than the cut off value (10) are believed to have no multi-co linearity problems and those with VIF of above 10 are assumed to have a multi-co linearity problem. Therefore, since, in this study, the computational results of the VIF for continuous variables Journal of Economics and Sustainable Development www.iiste.org ISSN 2222-1700 (Paper) ISSN 2222-2855 (Online) Vol.10, No.23, 2019 confirmed the non-existence of association between the explanatory variables and were included in the model.
In addition to this, the threshold for contingency coefficients for dummy and discrete variables is 0.75. The values below 0.75 indicate the existence of weak association and above 0.75 indicates strong association of variables. However, the results obtained in this study regarding dummy and discrete variables were less than 0.75. Therefore, this indicated that there was no any multi-co linearity between dummy explanatory variables.

Socio-economic and demographic factors influencing youth unemployment
Socioeconomic and demographic variables that are expect affect and determine un employment status of youths were used to in order to attain the study objectives of the study .Further information of the variables were shown in table in line with the dependent variable which kids unemployment status. Two objectives are achieved in model result in table 4 and the rest objectives addressed as shown in descriptive statistics in table 1 2 and 3 .  (2019), n = 395 the stars; ***, ** and * indicates 1percent, 5percent and 10 percent level of significance respectively.
From total independent variables employed in analysis expected tom influence unemployment of youth, four variables are positively and significantly related with the dependent variable, employment status of the youth. These were educational level of the youth, access to information, income of parents and economic factor of the parents. Similarly, three variables are negatively and significantly related with the dependent variable, employment status of the youth. These were months spent at search, educational level of youth's household head, and infrastructural factors. The variable Education, stands for educational level of the youth, has a positively and statistically significant effect on employment status of the youth at 5 percent level of significance. With the coefficient of 0.1106493, keeping other factors remain constant, thus we would predict that the log odds for the youth being employed increases by 0.0003443 for every one year increment in educational level of the youth. This shows that years spent on education or investment on education acts as a better signaling of productivity of the youth, thereby it increases the probability of being employed. In other hand, it shows many people invest in education expecting long-run positive return. The finding of Abebe (2004) also shows that educated youths have higher probability of getting job in urban Ethiopia than these having relatively low level of education. This is also consistent with the finding by Bhorat (2008) for South Africa. The finding of Berhan (2013) also supplements the conclusion of the higher the level of education the better the youth being employed. The variable Confidence, stands for either the youth is confident on his/her available skill, has a positively and statistically significant effect on employment status of the youth at 10 percent level of significance. With the coefficient of 1.226189, keeping other factors remain constant, thus we would predict that the log odds for the youth being employed increases by 1.226189 for these youths who are confident on their skill as compared to nonconfident youths.
The variable Access to information, stands for youths access to any information related to vacant position announcement (medium access, high access), has a positively and statistically significant effect on employment status of the youth at 5 and 1 percent level of significance, respectively. With the coefficient of 1.075958, keeping other factors remain constant, thus we would predict that the log odds for the youth being employed increases by 1.075958 for these youths who have relatively medium access to information than youths having low access to information, which is slightly significant at 5percent level of significance. Similarly, With the coefficient of 2.220913, keeping other factors remain constant, thus we would predict that the log odds for the youth being employed increases by 2.220913 for these youths who have relatively high access to information than youths having low access to information, which is strongly significant at 1 percent level of significance. The result indicates the better access to information regrding vacant position as the indicator of playing field (market), the higher the youth being employed. The study by Amanuel (2016) also found that Access to market information is significantly associated with youths' occupational status. For Amanuel those who have access to information are 86 percent less likely to be unemployed compared to those who have not getting the access, keeping the other variables constant. The variable Occupation Head, stands for the occupation of youth's household head (government employee, self-employed), has a positively and statistically significant effect on employment status of the youth at 5 and 1percent level of significance, respectively. The coefficient 1.969101 youth's household head as government employee, keeping other factors remain constant, thus we would predict that the log odds for the youth being employed increases by 1.969101 for these youths whose household head is government employed as compared to the youth with household head as a farmer, which is slightly significant at 5percent level of significance. Similarly, The coefficient 4.216824 youth's household head as government employee, keeping other factors remain constant, thus we would predict that the log odds for the youth being employed increases by 4.216824 for these youths whose household head is self-employed as compared to the youth with household head as a farmer, which is strongly significant at 1percent level of significance. The variable Income of parents, stands for monthly income level of youth's parent, has a positively and statistically significant effect on employment status of the youth at 5 percent level of significance. With the coefficient of 0.0001996, keeping other factors remain constant, thus we would predict that the log odds for the youth being employed increases by 0.0001996 for every one birr increment in the monthly income his/her parent. This shows that as income and wealth related factors of the youth's family increases, we expect the youths likely hood of being employed also increases. In other hand, these youths from relatively poorer families are most likely unemployed as compared to youths from richer households. This result might be the fact that youths from relatively higher income families may have better inputs for searching jobs or else they can easily get initial capital to start their own business. This finding is consistent with the finding of Amanuel (2016) examining determinants of youth unemployment, evidence from Ethiopia using 2011 Ethiopian Demographic and Health Survey (DHS) data. The finding of Aynalem and Mulugeta in Gojjam zone of amhara region also supplements family income background matters for employment status of the youth. The variable Economic factor, stands for economic factors related with youth's family, has a positively and statistically significant effect on employment status of the youth at 1 percent level of significance. The coefficient for the economic factor of youth's parent is 0.9942784, this shows that keeping other factors remain constant, we would predict that the log odds for the youth being employed increases by 0.9942784 for every one unit increment in the economic factor of the youth's family. The variable Month search stands for the number of months that the youth spent on searching job, has a negatively and statistically significant effect on employment status of the youth at 5 percent level of significance. With the coefficient of 0.026, keeping other factors remain constant, thus we would predict that the log odds for the youth being employed decreases by 0.026 for every one month increment in search for job by the youth. This shows the cost of job search is positively related with employment status of the youth. As cost of search increases, the higher the probability of being employed is also. The variable Educational Head, stands for the educational level of the youth's household head, has a negative and statistically significant effect on employment status of the youth at 10 percent level of significance. With the coefficient of -0.168, keeping other factors remain constant, thus we would predict that the log odds for the youth being employed decreases by 0.168 for every one year increment in the educational level of the youth's household head. The variable Institutional factor, stands for the institutional factors that affect youth's employment status, has a negative and statistically significant effect on employment status of the youth at 5 percent level of significance. With the coefficient of -0.875, keeping other factors remain constant, thus we would predict that the log odds for the youth being employed decreases by 0.875 for every one unit increment in the infrastructural factors, which is slightly significant at 5 percent level of significant. The result indicates that lack of labour unions, lack absent of credit providing instituiosns, and lack or absent of effective working agents in the area of youth are not really hindering the probability of the youth being employed.

CONCLUSION AND RECOMMENDATION 5.1 Conclusion
The findings from the binary logit regression model taking the unemployment status of the youth, indicates that:  Educational level of the youth has a positively and statistically significant effect on employment status of the youth. It shows that years spent on education or investment on education acts as a better signaling of productivity of the youth, thereby it increases the probability of being employed.  Confidence level of the youth has positively related with the employment level of the youth. Accordingly, these youths who are confident about their skill and effort are more likely to be employed.  The result indicates the better access to information regrding vacant position as the indicator of playing field (market), the higher the youth being employed. Access to market information is significantly associated with youths' occupational status.  Keeping other factors remains constant, these youths whose household head is government employed and self-employed are more likely to be employed as compared to the youth with occupation of household head as a farmer.  The finding shows that as income and wealth related factors of the youth's family increases, we expect the youths likely hood of being employed also increases. In other hand, these youths from relatively poorer families are most likely unemployed as compared to youths from richer households.  The economic factors related with youth's family, has a positively and statistically significant effect on employment status of the youth showing that keeping other factors remain constant, we would predict that the log odds for the youth being employed increases by 0.9942784 for every one unit increment in the economic factor of the youth's family.  The result also indicates that the higher months spent on searching, the higher the probability of getting job.
This shows the cost of job search is positively related with employment status of the youth. As cost of search increases the higher the probability of being employed.  The result also indicates that educational level of the youth's household head has a negative and statistically significant effect on employment status of the youth at 10 percent level of significance. This finding is inconsistent with the general fact and the hypotheses of this study.  The result indicates institutional factors like that lack of labour unions, lack absent of credit providing institutions, and Lack or absent of effective working agents in the area of youth are not really hindering the probability of the youth being employed.

Recommendation
Based on the finding of the study the following recommendations are forwarded.  Since level of education matters for employment status of the youth, policies and strategies that promote education and create more job opportunities should be implemented. The concerned bodies should also provide some job-training for those youths who are illiterate or less educated.  Since confidence level of the youth matters for employment status of the youth, the concerned bodies should develop the confidence level of the youth by providing different trainings for the youth.  The concerned body, especially the government should provide facilities related with information access like mass media and magazines in the study area.  Government and other concerned bodies in the study area should intervene in community mobilization for these household heads whose occupation farmer is regarding the occupational status of their youths.  Enhancing income as well as wealth level of youth's family needs intervention from the concerned bodies thereby to reduce youth unemployment.  Since the higher months spent on searching, the higher the probability of getting job. Thus youths are expected to invest more outlays on search of jobs. Hence, any agency should enhance the capacity of the youth to cover searching costs.