Environmental Economics, Volume 7, Issue 4, 2016

Mbulaheni Albert Dagume (South Africa), Agyapong Gyekye (South Africa)

Determinants of youth unemployment in South Africa: evidence

from the Vhembe district of Limpopo province
Abstract
The persistent high unemployment rate confronting South Africa, in particular among the youth, continues to be a
cause of concern for all stakeholders including academic researchers dealing with labor market issues. As a
contribution to efforts at finding solution to the unemployment challenge, the study sought to investigate the nature and
causes of rural unemployment amongst the youth using data on a sample of 580 randomly and systematically selected
youth from the four local municipalities in the Vhembe district of Limpopo province, South Africa. Binary logistic
regression model was estimated to determine the socio-demographic, as well as economic factors that influence youth
unemployment. Results of the binary logistic regression model showed that having received (skills) training and work
experience were associated with reduced odds of being unemployed. This study highlights the importance of skills
training opportunities, as well as apprenticeship in mitigating the unemployment challenge among the youth. The
provision of necessary training infrastructure and funding for skills training, as well as work integrated learning and
extended career specific internship programs to help to equip South African youth with work experience are major
recommendations from this study.
Keywords: unemployment, youth, binary logistic regression, Vhembe district, South Africa.
JEL Classification: J64.
Introduction ” South Africa, the country with the most advanced
economy in Africa and with three-quarters of poor
The problem of high youth unemployment is a global
phenomenon (Lam, Leibbrandt and Mlatsheni, 2008; people living in rural areas has a rural
Zinhnwe, 2012 and Yarima, 2014). This view has been unemployment rate of about 70 percent (du Toit,
echoed by International Labor Organization (ILO) 2003, Eastwood, Kirsten and Lipton, 2006). The
(2011), which points out that youth unemployment is 15 to 24 year-old age group constitutes 71 percent
among the major challenges facing both developed of the unemployed (Bhorat, 2006). The
and developing countries. ILO (2010) reports that the unemployment rates along gender groups also
global youth unemployment reached its highest level differ considerably. According to Statistics South
in 2010, pointing out that out of 620 million Africa (2011), unemployment affects females
economically active youth aged 15 to 25 years, 81 more than males. In support of this, in Limpopo
million were unemployed at the end of 2009. This is province, there were 376 101 (59.4%)
the highest number ever recorded and it is 7.8 million unemployed females and 256 860 (40.6%)
more than in 2007. Nighty percent of youth who live unemployed males during 2011 Statistics South
in developing economies are vulnerable to Africa survey.
underemployment and poverty (ILO, 2010).
According to Giri, as quoted in Pandey (2005),
Unemployment among the youth has been and unemployment is the problem of problems. This is
continues to be South Africa’s primary policy because unemployment contributes significantly to the
challenge (Bernstein, 2008). It has become a unacceptably high levels of poverty and income
contentious issue, as the rate is much higher than that inequality. The inability of jobseekers to secure jobs
of adults in most countries of the world (Dale, 2014). tends to create frustration among them, especially
Almost three quarters (70.8%, 3.4 million) of the 4.8 among the youth, resulting in increase in crime,
million people who are unemployed are youths suicide, violence, drug abuse and other social
between the ages of 15 and 34 (Stats SA, QLFSQ1, problems that can increase personal insecurity (Malik,
2013). According to Statistics South Africa (2015), 2014). From a national economic perspective,
youth unemployment rates for South Africa between unemployment results in the loss of output to the
2008 and 2015 were 32, 7% in 2008, 33.7% in 2009,
economy (du Toit, 2003).
35.7% in 2010, 36.1% in 2011, 35.8% in 2012, 36.2%
in 2013, 36.1% in 2014 and 36.9% in 2015. 1. Statement of the problem
Since the advent of the new democratic dispensation
in 1994, expectations have been high especially
” Mbulaheni Albert Dagume, Agyapong Gyekye, 2016.
Mbulaheni Albert Dagume, Ph.D., Lecturer of Economics, School of among the previously disadvantaged (black)
management Sciences, University of Venda, Thohoyandou, South population for better living conditions in post-
Africa.
Agyapong Gyekye, Professor, Department of Economics, University of
apartheid South Africa. In response to such
Venda, South Africa. expectations, the government promised to prioritize
59
Environmental Economics, Volume 7, Issue 4, 2016

poverty alleviation, income inequality through the district. Binary regression estimation
reduction of unemployment in its development technique is applied to investigate the causes of
strategies. Consequently, since 1994, the unemployment among the youth in the Vhembe
government introduced a plethora of development district from the supply side of the labor market.
policies designed to mitigate the challenges of
The paper is organized as follows: sections 3 discusses
unemployment, poverty and income inequality,
operational concepts together with theoretical and
including the Redistribution and Development
empirical literature. sections 4 focuses on the overview
Program (RDP), Growth, Employment and
of youth unemployment in South Africa. In section 5,
Redistribution Policy (GEAR), Accelerated and
the methodology employed for the econometric
Shared Growth Initiative of South Africa (ASGISA)
analysis of the cross-sectional data, results and
and the Joint Initiative for Priority Skills
discussions are presented, while final section presents
Acquisition (JIPSA).
conclusions and policy recommendations.
Despite the good intentions embedded in the
3. Literature review
above policy initiatives, unemployment still
remains stubbornly high in South Africa. Vhembe 3.1. The concept of unemployment. The
district of Limpopo province, which is the focus International Labor Organization (ILO) considers a
of the study, has a large population of person of working age fifteen years and older to be
unemployed youths. According to Census (2011), unemployed if during a specified reference period
youth unemployment rate in the district has (a week) if that person is not working for some kind of
increased from 53.4% in 2010 to 58.4% in 2011. financial compensation, but is: (a) willing to work; (b)
Despite these alarming figures, unemployment available to work; and (c) actively searching for work
among the youth in the Vhembe district has not (Statistics South Africa, 2011).
received much scholarly attention, apart from two
3.2. Youth. This study is on youth unemployment.
studies, one by Kyei et al. (2011) attempting to
The definition of youth varies from country to country.
find determinants of unemployment in Limpopo
For example, according to the United Nations, youth
Province, and another by Nemalili (2006)
refers to a person aged between 15-24. In South
investigating unemployment in Tshiheni Village
Africa, youth refers to persons falling within the age
within the Vhembe district. All these studies were
group of 15-34 years, which is the official definition of
trying to address unemployment on the whole,
youth according to Statistics South Africa (2011) and
generally for the working age group 15 to 64. The
which is the definition used in this study.
problem of unemployment of the working age
group 15 to 34 years (the youth) who constitutes 3.3. Types of unemployment. Several types of
the highest percentage of the entire population has unemployment may be experienced in an economy
yet to be addressed. As a social phenomenon, such as that of South Africa and they include:
unemployment is a multifaceted issue, because it frictional, seasonal, cyclical and structural
affects not only the unemployed individual, but unemployment. Frictional unemployment which,
also the entire society. An investigation of the amongst other things, deals with unemployment
factors of youth unemployment in the Vhembe arising from people who are in the process of moving
district, being the first of its kind, will, therefore, from one job to another. This may be common among
contribute to an understanding of the problems of the youth who prefer to move from one job to another,
youth job seekers from individual, social and with the aim of searching for greener pastures.
household perspectives, and how the problems Seasonal unemployment arises when workers are laid
can possibly be mitigated. off during off seasons. This type of unemployment is
common in sectors such as agriculture. Cyclical
2. Research objectives unemployment occurs when the economy slows down,
Drawing on data from representative sample of such as during times of recession, when people lose
youth, the main objective of the study is to their jobs. Structural unemployment occurs when there
investigate the causes of unemployment among is a change in the structure of an industry or economic
the youth in the Vhembe district of South Africa. activities due to say rapid changes in technology which
results in mismatch between the skills of workers and
The specific objectives are: to provide an
skill requirements of available jobs.
overview of trends in youth unemployment in
South Africa; to empirically analyze the 3.4. Theoretical literature review. A variety of
determinants of unemployment incidence among theories have been advanced in economic literature
the youth in the Vhembe district; and to to explain the phenomenon of unemployment.
recommend public policy measures for Under classical economic theory, unemployment in
employment creation to reduce unemployment in the economy is simply a short term phenomenon, and
60
 Environmental Economics, Volume 7, Issue 4, 2016

the workings of the free market forces would Human capital theory asserts that the quality of labor
eventually restore the economy to full employment. In force is vital for labor demand by companies and it is
other words, labor market always clears on the basic relevant to the decisions of these companies as to who
assumption of flexible wages and perfect information. is retrenched or dismissed. As a result, human capital
In situations where this assumption does not hold, due variables such as education level and training explain
to institutional rigidities such as minimum wage the probability of unemployment (Osberg et al., 1986;
legislation, the market may not clear, leading to the Byrne and Strobl, 2001; Lindley, 2005).
classical involuntary unemployment. As Goodwin
3.5. Empirical literature survey. Empirical studies
(2006) points out, involuntary unemployment can
on determinants of unemployment have focused on the
exist if minimum wages get in the way of market
phenomenon either from the supply side perspective or
forces. If employers are required to pay a minimum
from the demand side perspective. From the supply
wage that is above the equilibrium wages, fewer
side perspective, a number of empirical studies have
workers will be hired. In the classical theory
found that unemployment tends to be lower for
parlance, it is not possible to have involuntary
persons with higher educational attainment. In his
unemployment unless there are distortions in the
empirical analysis on unemployment, Elhorst (2003)
functions of the labor market. The theory maintains
found a direct relationship between the level of
that there is no such thing as cyclical unemployment
educational attainment and employability. He explored
and that those people who are not working have
four impacts education has on individuals’
made decisions not to work or that they are either
employment pattern/employability. First, he opines
part of frictional or structural unemployment.
that in an economy with continued technological
From the Keynesian school of thought, progress, labor demands tend to focus on skills
unemployment largely arises from deficiencies in exhibited mainly by better educated individuals.
aggregate demand over certain periods in the Second, the search behaviors of better educated
business cycle such that jobs created are inadequate individuals enable them to find jobs more quickly that
for everyone who wants to work (Keynes, 1936). of those of the less educated individuals. Thirdly, the
This type of unemployment is cyclical and higher a person’s educational attainment, the lower the
involuntary due to the constraint imposed by limited probability that the person will be laid off. Lastly,
availability of employment opportunities. When higher skilled workers crowd out workers with less
aggregate demand is low, firms hire little labor. The education by accepting jobs that require lower
idea underlying the theory is that lack of aggregate educational attainments than they possess. Kabaklarli,
demand would lead to falling sales which lead to a Hezeler and Buhus (2011), in an Organization for
fall in investments; this leads to another fall in Economic Cooperation and Development’s (OECD)
aggregate demand and more unemployment, study, found that employment rate of the youth aged
therefore, creating a vicious cycle. Keynesian school (15-29 years) with higher education level was higher
of thought ,therefore, calls for the use of appropriate than those with secondary education level and high
government policies, either fiscal or monetary illiteracy level. A study by Sackey and Osei (2006) in
policies, to eliminate involuntary unemployment Ghana similarly found that basic and senior secondary
in the economy. levels of education are associated with relatively more
unemployment in Ghana due to the relatively lower
The insider-outsider model of wage-setting
levels of such education compared with tertiary
behavior of firms is a model used to explain
education. In South Africa, a study by Msigwa and
institutional form of involuntary unemployment.
Kipesha (2013) found that youth with higher
In the model, there are two groups of potential
educational levels had lower levels of unemployment.
workers; insiders are workers who are already
The reason given was that young people who had
employed (e.g., unionized workers) and strive to
completed primary or secondary education did not
maximize the interest of their members only.
possess any skills required in the job market. A
Outsiders are workers, i.e., non-unionized
number of studies have also found that youth with job
workers who possess the same potential to be
training are in a better position of getting employment
hired as unionized workers. The model asserts
than their counterparts who had never gone for any
that unemployment arises when wages are
training. For example a study by Qayyum (2007)
determined by taking into consideration the
found that technical and vocational training
interest of only those employed (insiders) without
significantly decreases the probable chances of an
regard to the interest of those seeking to be employed
individual becoming unemployed.
referred to as outsiders (Bentolila et al., 2011). Thus,
firms and insiders bargain to lock the outsiders out of Studies by Layard and Nickel (1991), Qayyum (2007)
the job market and thereby create unemployment. and Echebiri (2005), have confirmed that age of a

61
Environmental Economics, Volume 7, Issue 4, 2016

youth can influence employment status due to the fact For example, Altman and Potgieter-Gqubule (2009), as
that employability goes with experience, implying that cited in Smith (2011), argue that the quantum of job
fresh graduates with no experience will always find it creation was too small to make a dent in youth
difficult to access jobs, since they lack relevant unemployment in South Africa, although Altman and
experience. This has been echoed by Dickens and Potgieter-Gqubule (2009) equally point out supply side
Lang (1995) who found unemployment to be highest factors such as mismatch between worker skills and
among the highly educated Sri Lankan youth (aged 15- employer needs, which causes available jobs to go
24), but declining thereafter. Sackey and Osei (2006) unfilled, may explain the slow employment growth in
conclude that younger people are more likely to be the South African economy. To correct the mismatch,
unemployed due to the fact that they possess lower it been variously been suggested that jobs should be
skills in comparison to their older counterparts. created in the private sector, in particular, small and
Gender disparities in unemployment have been reported medium enterprise (SME) sector, through the
by Qayyum (2007) using data from labor force survey development and promotion of entrepreneurial skills
(2003-2004) and probit model to investigate causes of for the sector. A study on youth employment by Kiiru,
unemployment among young people aged 15-29 years Onsomu and Wamalwa (2013) suggests that a youth
in Pakistan. Qayyum found higher unemployment rate fund be established in order to assist youth
among females than males. A study by the entrepreneurs.
International Labor Office (2004), pointed out that 4. Overview of youth unemployment in
youth unemployment in South Africa has a gender South Africa
dimension, because it is generally higher among
females than males. This has been confirmed recently Youth unemployment poses a formidable policy
by Msigwa and Kipesha (2013). The results of the challenge for South Africa. Figure 1 illustrates
same study by Quyyum (2007) relating to marital unemployment trends in South Africa (SA) by age
status showed that unemployment among unmarried group over the period 1995 to 2013. These figures
people was found to be considerably higher than reveal a strong relationship between age and
among married people. This seems to contradict the unemployment rates. Unemployment rates are
findings of a study by Gebere (2011) which found that increasingly lower in the higher age groups. Thus, the
the percentage of unemployed for married youths unemployment rate among persons aged 15-24 years,
(46.5 percent) was greater for never married youth as shown above, is higher than those in the 15-34 years
(36.9 percent). age group. In turn, the unemployment rate among
From the demand side, prolonged unemployment has persons aged 15-34 years is higher than among those
been attributed to weak demand for labor. in the 35-64 years age group.

80
70
60
percentage

50
15-24 years
40
15-34 years
30
35 -64 years
20
10
0
1995 2002 2003 2008 2009 2010 2011 2012 2013

Fig. 1. Unemployment rates by age group, SA, 1995-2013

Source: Stats SA (2014), QLFS, Q4 (2010), QLFS Q3(2013), Q1 (2014).

Youth unemployment in South Africa also seems by a number of empirical studies such as the
to have a gender dimension. The unemployment ones done by Fawcett (2002), Msigwa and
rate among the youth for the period 2001-2011 is Kipesha (2013) and Kyei and Gyekye (2011) that
consistently higher for females than males, as have found female unemployment rates to be
depicted in Fig. 2. This trend has been confirmed higher than those of males.

62
 Environmental Economics, Volume 7, Issue 4, 2016

70

60

percentage 50

40 Female
Male
30
Column1
20

10

0
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
Fig. 2. Unemployment rate by gender (15-24 years), SA, 2001 -2011
Source: SADC Statistical Yearbook (2011)

Fig. 3 depicts the impact of educational (with secondary school or tertiary qualification),
attainment on the unemployment situation in i.e., unemployment rate decreases as levels of
South Africa. Unemployment rate is the highest educational attainment increase. This supports
among those with lower than Matric (without the view that education, particularly the
secondary school qualification) compared to attainment of tertiary qualifications, enhances
those with education levels higher than Matric access to the labor market.

70
Percentage share of unemployed by

60
educational level

50

40

30

20

10

0
2008 2009 2010 2011 2012
Lower than Matric 63.8 60.9 59.7 60.6 61.3
Matric 31.1 32.8 33.9 32.3 32.1
Tertiary 4.4 5.4 5.7 6.3 6.2

Fig. 3. Share of the unemployed by educational level, South Africa, 2008-2012

Source: Stats SA, Q4 (2012).

5. Research methodology municipalities of the district. The data were collected in

June/July 2013 using a structured questionnaire. The
5.1. Data and model specification. In order to analyze
questionnaire enabled detailed data to be collected on
the relationship between socio-demographic factors and
respondents’ demographic information, as well as
unemployment among the youth in the Vhembe district,
information on their human capital assets.
this study uses a cross-sectional data on a representative
sample of 580 youth (defined as individuals falling Conceptually, the model of determinants of
within the age group of 15-34) in the four local unemployment among the youth is specified as:
63
Environmental Economics, Volume 7, Issue 4, 2016

ܻ݅ ൌ ߚߧ ൅ ߚ݅ߕ ൅ ߝ, (1) ߚଵ ൐ 0, ߚଶ ൏ Ͳ , ߚଷ ൐ Ͳ , ߚସ ൏ Ͳ , ߚହ ൏ Ͳ ,

ߚ଺ ൏ Ͳߚ଻ ൏ Ͳ , and ߚ଼ < 0.
where Yi = unemployment, iΧ is a vector of
explanatory variables, β represents a vector of The estimated regression coefficients represent the
parameters of explanatory variables to be “odds ratio” – exp (Bi) which measures the effect of
estimated, and ε is a vector representing the explanatory variables on the “odds ratio” (the odds
stochastic error term. ratio is the probability of being unemployed divided
In this study, the dependent variable (Y) is by the probability of being employed; p/1-p).
measured as a binary variable with an assigned For categorical explanatory variables, if the value of
value of 1 if the individual is unemployed and 0 if the odds ratio exp (B) for a certain category is
the individual is employed. In the context of this greater than 1, the chance of unemployment is
study, an individual is unemployed if he or she higher for a member of that group in relation to the
falls within the age group of 15-34 and was not in reference category. On the other hand, the odd ratio
paid employment or self-employment and has not of less than 1 implies the lower chance of
worked for five or more hours for a wage or unemployment in relation to the reference category.
salary or for profit or family gain during the For continuous explanatory variables, exp (Bi) is the
reference period; or was available for paid estimated multiplicative change in the odds for a
employment or self-employment during the unit of increase in the predictors, controlling the
reference week; or took specific steps during the effects of others. In addition, a positive predictive
four weeks preceding the interview to find paid coefficient (β>1) means that the predicted odds
employment or self-employment. increased, as the predictor value increased, and a
On the basis of the binary nature of the dependent negative coefficient (β< 1) indicates that the
variable, a binary logistic regression technique was predicted odds decreased, as the predictor value
adopted to analyze how each of the explanatory increased.
variables ( age, gender, marital status, educational
The maximum likelihood method was used to
level, ethnicity, training or skills, geographical location
estimate the numerical values of the parameters of
and work experience) influence the probability of an
the binary logistic model. This method has stronger
individual becoming unemployed. The logistic model
theoretical appeal compared to the method of OLS
to be estimated for this study is specified as follows:
(Bhat, 2000), due to its ability to estimate the non-
௉ linear equation associated with the binomial logit
Ž ቀଵି௉೔ ቁ ൌ ߚ଴ ൅ ߚଵ ܺଵ ൅ ߚଶ ܺଶ ൅ ߚଷ ܺଷ ൅ ߚସ ܺସ ൅
೔ model adopted in this study.
൅ߚହ ܺହ ൅ ߚ଺ ܺ଺ ൅ ߚ଻ ܺ଻ ൅ ߚ଼ ଼ܺ  ൅ ߝ௜ , (2)
5.2. Results and discussion. Table 1 presents the
where: Y = employment status (unemployed = 1; results of the estimated logistic model of the
employed = 0). Employment status is expressed as a determinants of youth unemployment in the
௉
probability functionቀଵି௉೔ ቁǡ with ܲ௜ denoting the Vhembe district of South Africa. We, first, analyzed
೔
probability of youth ith being unemployed and 1-Pi is whether the independent variables in our model
the probability that ith respondent is employed from have a significant relationship with the dependent
the survey data. X1 denotes age group of respondents variable. This was necessary for determining the
in years; X2 denotes gender of respondent, 1 if male ability of the model to predict the dependent
and 0 if female; X3 denotes race of respondent variable accurately. The Likelihood ratio test was,
consequently, performed to test the overall
measured according to categories in South Africa, 1 if
black, 0 if colored (the other racial groups, i.e. Indians significance of all the coefficients in the model and
and Whites declined to be interviewed); X4 denotes our results indicate that the overall model is
location of respondent, 1 if urban, 0 if rural; X5 denotes significant at the 0.01 level according to the Model
the highest level of education of respondent and chi-square, implying that, as a whole, the
independent variables have a significant
captured as dummy variable (1 if primary, 2 if
secondary and 3 if tertiary with no schooling as a contribution to predict the response variable (i.e.,
reference dummy); X6 denotes marital status, 1 married the unemployment status of the youth in the
and 0 single; X7 denotes prior training status of Vhembe district of South Africa). The logistic
respondent, 1 if trained, 0 not trained and X8 denotes regression coefficients, Wald test and odds ratio for
work experience, 1 experience and 0 no experience. each of the predictors are presented in this table.
The “sig” column reveals the significance (or p-
The parameters for to be estimated from equation 2 value) of each of the variables while B values
areߚଵ , ߚଶ , ߚଷ , ߚସ , ߚହ , ߚ଺ , ߚ଻ and ߚ଼ . Based on explain the direction of the relationship of the
theory and empirical literature, expected a priori particular independent variable, with the
signs of the parameters are: dependent variable.

64
 Environmental Economics, Volume 7, Issue 4, 2016

On the other hand, Exp (B) column represents the odds The table indicates that the coefficients for training and
ratio. Using 0.05 level of significance as a standard for experience are statistically significant at the 0.05 and
test of statistical significance, the coefficients of all the 0.01 level, respectively. This shows that the two
variables, with the exception of training and variables play a significant role in predicting the
experience were found to be statistically insignificant. probability of being unemployed.
Table 1. Parameter estimates of the logistic model of the determinants of youth unemployment
Wald Confidence interval 95%
Variables Category B S.E df Sig. Exp (B)
Chi-square Lower Upper
Age group 15-20 .181 .515 .123 1 .726 1.198 .437 .3.288
21-25 -.130 .318 .167 1 .683 .878 .471 1.637
26-30 .103 .329 .099 1 .753 1.109 .581 2.115
31-34 (RC)
Gender Male .124 .222 .310 1 .578 1.132 .732 1.751
Female (RC)
Marital status Single .180 .342 .278 1 .598 1.197 .613 2.341
Married (RC)
Education Primary .968 .771 1.575 1 .209 2.632 .581 11.931
Secondary .411 .362 1.288 1 .256 1.508 .742 3.064
Tertiary (RC)

Training Trained -5.03** .227 4.931 1 .026 .604 .388 .943

Not trained (RC)

Experience Experience -2.272*** .259 77.049 1 .000 .103 .062 .171

No experience (RC)
-2 Log Likelihood 547.507
-
Model Chi-square 136.185***

RC = reference category **p<0.05, ***p<0.01

Source: data survey.

From the results, training has a negative coefficient who constitute a large proportion of the unemployed
and is statistically significant (p = 0.026 < 0.05). in the Vhembe district municipality.
This result is in agreement with the prior
Our results also provide evidence to support the
expectation and indicates that the high level of
youth unemployment in Vhembe district arose as a view held by many that prior work experience is
result of a lack of training or skills. The sample odds likely to positively impact on the probability of
of a trained youth being unemployed were 0.604 unemployed youth finding employment. From Table
times lower than those of an untrained youth. This 1, provided all other variables are held constant, the
means that a youth who had some training was less odds of a youth with some work experience being
likely to be unemployed, compared to a youth unemployed are about 0.103 times lower than those
without training. Our results collaborate the findings of a youth without work experience. This means that
presented in a study by Msigwa and Kipesha (2013) a youth who has some work experience is less likely
in Tanzania, which also reported that being a youth to be unemployed, compared to a youth without job
without training made a person about 23 percent experience. The association between experience and
more likely to be unemployed. The results are also unemployment was statistically significant
in line with the studies done by Mahlwele (2009) (p = 0.000). The results agree with studies and
who found that untrained women were less likely to
findings by Altman and Gqulube (2009), as cited in
get employment than the trained women. Gebeyaw
Smith (2011), who found that individuals who had
(2011) in a study conducted in Addis Ababa
similarly found training to have a negative impact never held a job before were 35% more likely to
on unemployment, and was statistically significant remain unemployed than those who had prior work
at 1% level of significance. The implication of our experience. Similarly, the ILO (2004), as cited in
result is that training could be an important strategy Gebere (2011), notes that lack of work experience
to reduce youth unemployment, particularly among reduces the chances of getting employment in the
those with only primary and secondary education modern sectors of the economy.

65
Environmental Economics, Volume 7, Issue 4, 2016

Youth unemployment in Vhembe district in 2013 was essentially a supply side (of the labor market) one, used
largely found to be determined by training and work binary logistic regression model to examine the
experience. The other variables (age, gender, marital relationship between age, gender, marital status, race,
status and education level) were insignificant in education, geographical location, work experience,
determining youth unemployment. With regards to age, training and youth unemployment. It was hypothesized
even though it was not significant, the coefficients on that these factors exert a strong impact on youth
the other age categories (i.e., 15-20 years and 26-30 unemployment in the economy of Vhembe district over
years) were positive as per prior expectation indicating the period up to 2013. The results showed that two out
that youths in these age groups are more likely to be of eight explanatory variables tested were significant in
unemployed, as compared to those in the reference explaining youth unemployment; training and work
category (age 31-34 years). These results are in experience were found to have major influence on
conformity with Qayyum (2007) who found age to be youth unemployment. The results showed that having
significant and negatively related with unemployment received some training and having some experience are
indicating that additional increase in age reduced the associated with reduced odds of being unemployed.
probability of becoming unemployed by 0.2%. The Controlling for the level of training, the odds of a youth
coefficient of the gender variable was found to be
without experience being unemployed are about (0.103)
insignificant at the 0.05 level. The insignificance of the
times more likely to be unemployed, as compared to a
gender variable, in apparent contrast with the picture in
youth with experience. With regard to training or skills,
South Africa as a whole (see Fig. 2), may be attributable
the sample odds of a trained youth being unemployed
to a better compliance and implementation of the
national affirmative action legislation, known as were 0.604 times lower than those of an untrained
Employment Equity Act 55 of 1998 in the Vhembe youth. This means that a youth who had some training
district, our study area. The purpose of this legislations was less likely to be unemployed compared to a youth
is to achieve equity in the workplace, by promoting without training. Lack of work experience and training
equal opportunity and fair treatment in employment or skills are the key factors for youth unemployment in
through the elimination of unfair discrimination, while Vhembe district of Limpopo province. Based on these
implementing affirmative action to redress the findings, the recommendations are that the South
disadvantages in employment experiences by African government must strengthen the laws and
designated groups (black people, women and people policies which will enable the youth to acquire work
with disabilities). experience and more training. Thus, instead of
government and private sectors offering internships to
Conclusions and recommendations the youths for twelve months or less, this period should
The aim of the study was to investigate the determinants be increased to at least twenty four months, as many
of youth unemployment in Vhembe district of Limpopo companies require three years’ experience when
province South Africa and suggest way forward recruiting. In addition, those offered internships should
for mitigating the joblessness challenge. The study, do career-specific training.
References
1. Altman, M. and Potgieter-Gqubule (2009). The State of Youth: Labour Market Status and Policy Challenge,
Centre for Poverty, Employment and Growth, Human Sciences Research Council, Pretoria.
2. Bentolila, S.J., Juan, D. and Jimeno, J.F. (2011). Reforming an Insider-Outsider Labour Market: The Spanish
Experience IZA, Discussion Paper No. 6186, 53072, Bonn, Germany.
3. Bernstein, A. (2008). Youth unemployment interventions neither attacking root causes nor going to scale, Centre
for Development and Enterprise, Johannesburg.
4. Bhat, C.R. (2000). Flexible model structures for discrete choice analysis. In D.A. Hensher and K.J. Button (eds),
Hand Book of Transport Modelling, Elsevier Science, Amsterdam.
5. Bhorat, H. (2006). Youth employment and education in South Africa. Cape Town, Harold Wolpe Memorial Trust,
February 2.
6. Byrne, D. and Strobl (2001). Defining unemployment in developing countries: The case of Trinidad and Tobag,
Credit Research paper No.01/09, Center for Research in Economic development and International Trade,
University of Nottingham, Nottingham.
7. Dale, B.B. (2014). Unemployment Experience of Youth in Addis Ababa, Master of Arts in Development Studies
thesis. International Institute of Social Studies, Ethiopia.
8. Dickens, W.T. and Lang, K. (1995). Analysis of the Nature of Unemployment in Sri Lank, The Journal of
Development Studies, 31(4), pp. 620-636.
9. Du Toit, R. (2003). Unemployed youth in South Africa: The distressed generation? Paper presented at the
Minnesota International Counseling Institute (MICI), 27 July-1 August 2003.
10. Eastwood, R., Kirsten, J. and Lipton, M. (2006). Can land redistribution help reduce rural dependency in South
Africa? Brighton, Institute of Development Studies.

66
 Environmental Economics, Volume 7, Issue 4, 2016

11. Echebiri, R.N. (2001). Youth Unemployment and Job Perceptions among Youth in Rural Southeastern Nigeria.
Paper presented at the 2001 conference of Agricultural Society of Nigeria.
12. Elhorst, J.P. (2003). The Mystry of regional unemployment differentials; Theoretical and empirical explanations,
Journal of Economic Surveys, 17, pp. 709-748.
13. Fawcett, C. (2002). Latin American Youth in Transition: A Policy Paper on Youth Unemployment in Latin
America and the Caribbean. Inter-American Development Bank. Sustainable Development Department, Social
Development Division, Labor Markets Policy Briefs Series, America [Online]. Available at:
http://www.youthpolicy.org/library/wpcontent/uploads/library/2002_Latin_American_Youth_Transition_Policy_P
aper_Youth_Unemployment_Eng.pdf. Accessed on 12 June 2014.
14. Gebere, A. (2011). Demographic and Socio-economic determinants of youth unemployment in DebereBirhan
Town, North Showa Administrative zone, Ahhara National Regional State. Masters in Population Studies Thesis,
Addis Ababa University.
15. Gebeyaw, T. (2011). Socio-Demographic Determinants of Urban Unemployment: The Case of Addis Ababa,
Ethiopian Journal of Development Research, 33(2), pp. 79-125.
16. Goodwin, N.R., Nelson, J.A., Ackerman, F. and Weisskopf, T. (2006). Theories of unemployment [Online].
Available at: http://www.eoearth.org/article/Theories-of-unemploymentssami. Accessed on 26 April 2014.
17. ILO. (2010). Global Employment Trends for Youth: A generation at risk, International Labor Office
Geneva,Switzerland.
18. ILO. (2014). Global Employment Trends for Youth. International Labor Office – Geneva, Switzerland.
19. Kabaklarli, E., Hazeler, P. and Buhus, A. (2011). Economic Determinants of Turkish youth unemployment
problem: Cointegration Analysis, Proceedings of the International Conference on Applied Economics, pp. 267-
272.
20. Keynes, J.M. (1936). General Theory of Employment, Interest and Money, Macmillan Cambridge University
Press.
21. Kiiru, J.M., Onsomu, E. and Wamalwa, F. (2003). Does Education and Training matter for youth employment?
Journal of Emerging Issues in Economics, Finance and Banking, vol. 1 No. 2, pp. 128-149.
22. Kyei, K.A. and Gyekye, K.B. (2011) Determinants of Uemployment in Limpopo Province of South Africa: An
Exploratory Study, Journal of Emerging Trends in Economics and Management Sciences (JETEMS), 2(1), pp. 54-
61.
23. Lam, D., M. Leibbrandt and Mlatsheni, C. (2008). Education and Youth Unemployment in South Africa,
University of Cape Town, Southern Africa Labor and Development Research Unit.
24. Layard, R. and Nickel, S. (1991). Unemployment in the OECD Countries, Economics Series Working Paper
99130, University of Oxford, Department of Economics.
25. Lindley, J. (2005). Explaining Ethic Unemployment and Activity Rates: Evidence from the QLFS in the 1990s and
2000s, Bulletin of Economic Research, 57(2), pp. 185-203.
26. Mahlwele, C. (2009). Factors Associated with Women Unemployment in South Africa. Makerere University,
Statistics South Africa.
27. Malik, K. (2014). Human Development Report, Sustaining Progress: Reducing vulnerabilities and Building
Resilience. United National Development Program.
28. Msigwa, R. and Kipesha, E.F. (2013). Determinants of youth unemployment in developing countries: Evidence
from Tanzania, Journal of Economics and study sustainable development, 4 (14).
29. Nemalili, N.P. (2006). An investigation of Unemployment at Tshiheni Village: Limpopo Province. Master of
Development thesis, University of Limpopo, South Africa.
30. Osberg, L., Apostle, R. and Clairmont, D. (1986). The Incidence and Duration of Individual Unemployment:
Supply Side or Demand Side, Cambridge Journal of Economics, 10, pp. 13-33.
31. Pandey, A.P. (2005). The new employment guarantees scheme - a- radical deal for India’s poor.
32. Qayyum, W. (2007). Causes of Youth Unemployment, The Pakistan Development Review, 46(4/2), pp. 611-621.
33. Smith, J.M. (2011). Connecting Young South Africans to opportunity Literature Review and strategy, The DG
Murray Trust (DGMT).
34. Statistics South Africa (2011). Census. Pretoria.
35. StatsSA Labour Force Survey: Statistics South Africa. Statistical release Labour force survey 2013, 2015.
Available at: http://www.statssa.gov.za/
36. Yarima, Y.A. (2014). Socio-Economic effects of Unemployment among the Youth in Nigeria, Journal of
Contemporary Issues in Business Research, 3(5), pp. 240-249.
37. Zinhunwe, C. (2012). The Youth and Unemployment in Zimbabwe. Paper Presented at the National Conference
with the theme “Building a Future with Decent Work for Young People held at Rainbow Towers in Harare” 29
March 2012.

67