Site Loader

LOCATION

510 5th Ave, New York, NY 10036, USA

AN ANALYSIS OF THE INFLUENCE OF YOUTH UNEMPLOYMENT ON ECONOMIC GROWTH: EMPIRICAL EVIDENCE FROM TANZANIA.
(1991:1-2017:4)
BY
FREDERICK A. MAHAWI
22017011/T.14

SUPERVISOR:
DR. CORETHA KOMBA

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

Table of Contents
CHAPTER ONE 4
INTRODUCTION 4
1.1 Background of the study 4
1.2 Statement of the Problem 6
1.3 Objectives 7
1.3.1 Main Objective 7
1.3.2 Specific Objectives 7
1.4 Hypothesis 7
1.5 Research Question 7
1.6 Significance of the Study 7
1.7 Scope of the Study 7
1.8 Limitations of the Study. 8
1.9 Layout of the Study 8
CHAPTER TWO 9
REVIEW OF THE RELATED LITERATURE 9
2.0 Introduction 9
2.1 Theoretical Literature Review 9
2.1.1 Marxist theory of unemployment 9
2.1.2 Traditional neoclassical growth theory 10
2.1.3 The Keynesian unemployment theory 10
2.1.4 Okun’s theory of unemployment 11
2.2 Empirical Literature Review 11
2.3 A Critique of Empirical Literature and Research Gaps 13
CHAPTER THREE 14
MODEL SPECIFICATION AND METHODOLOGY 14
3.0 Introduction 14
3.1 Functional Relationship and model Specification 14
3.2 Variables Description 16
3.3 Prior Estimation Time Series Tests 18
3.3.1 Descriptive Analysis 18
3.3.2 Unit Root Test 18
3.3.3 Johansen Co-integration Test 19
3.3.4 Vector Error Correction Model (VECM) 19
3.3.5 The Granger Causality Test 20
3.4 Diagnostic Tests 20
3.4.1 Normality test 20
3.4.2 Autocorrelation test 20
3.4.3 Heteroskedasticity test 21
3.4.4 Stability test 21
3.4.5 Multicollinearity Test 21
3.5 Data Analysis 21
REFERENCE 22

CHAPTER ONE
INTRODUCTION
1.1 Background of the study
Africa, being the world’s youngest region, continues to be threatened with high levels of unemployment. The unemployment challenge is even worse to some age groups especially the youth. Trends for youth shows that as a result, the global youth unemployment is expected to rise by half a million in 2016 to reach 71 million, which is the first such increase in 3 years’ time (ILO, 2016). The degree of severity of youth unemployment in the continent differs from one sub-region to another especially between the Sub-Saharan Africa and the Maghreb (North Africa) Africa. Although the youth population is projected to reach over 830 million by 2050 in the entire continent, the prevalence of youth unemployment in the Sub-Saharan Africa is expected to continue falling; the trend that started in 2012 reaching 10.9% in 2016 with a slight fall to 10.8% in 2017, (ibid). Moreover, there is a strong correlation between education and productivity levels; while the youth labour force participation rate is the highest in Sub-Saharan Africa reaching 54.2%, the sub-region’s enrolment rates in secondary and tertiary education is the lowest in the continent. This has a bearing impact on the economic growth processes.
The economies of many African countries are weaker today as compared to the past since we witness many African youth are joining a disheartened job market and worse still the continent has even the educated people struggling to secure a job. The above insights were shared by Mr. Tonny Elumelu who is the chairman of Heirs Holdings (Nigeria) during the Africa Finance Corporation Infrastructure Summit held in Abuja Nigeria during May 2017. Mr. Tonny Elumelu is among the largest youth employers in the continent; he said further that youth unemployment in the Sub-Saharan Africa runs between 12% and 14% while the problem is around 9% to 10% in South Asia, over the same period.
For the case of East African Community (EAC) sub-region youth unemployment is still a huge challenge too. The EAC comprises of Burundi, Kenya, Uganda, Rwanda and Tanzania. Among these countries, Kenya has the highest youth unemployment rate. It is projected that one in every five Kenyan youth of working age has no job, compared to one in every 20 youth in Uganda and Tanzania. That puts Kenya at roughly 20% youth unemployment, which makes Uganda and Tanzania’s unemployment problems (5%) seem insignificant in comparison. In fact they are not, since there’s nothing insignificant about youth unemployment across East Africa. There are various factors contributing to the region’s worrying employment figures namely: rapid population growth, skills gap, informal unemployment, and insecurity (The East Africa Monitor, 26 April 2017).
The origin of the unemployment challenge in Tanzania starts in 1970s when the country experienced economic crunch which was observed through a slowdown in economic growth rate from around 5% to merely 2.6% in the 1980s; the situation had worsened further to about 1% in early 1990s, Salim et al (2017). The job creation pace was far below than the available labour force during the same periods; hence the unemployment rate kept increasing within the same period, (WB, 2017).
The pervasiveness of unemployment in Tanzania as in other countries, is regarded as one among the key national developmental challenges. The government of Tanzania since the country got her independence is working smarter and harder to minimize the unemployment problem alongside with poverty reduction and reducing inequality issues in the country. The government therefore has been crafting various policies to deal with the issues which include: National Strategy for Economic Growth and Poverty Reduction (NSEGPR), Tanzania Development Vision 2025 (TDV 2025), amongst others.

Tanzania has claimed strong real GDP of over 6 percent for the past 15 years yet the growth rates have not translated into more jobs creation since the annual unemployment rate within the same period was at the rising rate from 2% in 2005 to 2.9 percent in 2013, despite the decline of the unemployment rate to 2.3% in 2016, (Salim et al, 2017). The real GDP growth rates and unemployment rates provide a paradoxical phenomenon since it was expected that the higher GDP growth rates would translate into lower unemployment rates.

The current youth unemployment crisis in Tanzania has escalated the academic and policy debate on its impact to economic growth. It is on the basis of this background that this study is carried out purposefully in seeking the responses to contribute to the debate since the issue of youth unemployment is still hot and indeed a burden to the Tanzanian economy.

1.2 Statement of the Problem
The economic growth and unemployment exhibit inverse relationship, meaning, as one variable increases the other decreases. This relationship is termed as Okun’s law. Okun postulated that a 1% increase in unemployment would result in more than 3% loss in economic growth, Ihensekhien (2017).

In the Sub-Saharan Africa (SSA) sub-region, economic growth rates of most countries are still not high enough to suppress the influence of unemployment in the societies. Therefore it will take longer time if the current growth rates persist for these countries of the SSA sub-region region to catch up with other developing countries in other regions.

Different studies have been conducted on the same topic using different approaches and covering different time spans with mixing results pertaining to the connection between economic growth and unemployment. The results of some studies applause while others contradict the Okun’s law.
For instance: (Gocer and Erdal: 2015), Ihensekhien (2017), and (Salim et al: 2017), conducted their researches with the aim of investigating the linkage between unemployment and economic growth in its generic terms. The results from all these researches are different because some affirm the Okun’s law while others contradict it.

These contradicting results pose challenges to policy makers in formulating appropriate policies to boost economic growth as well as reducing unemployment challenge especially youth unemployment. Moreover, a lot of changes occurred in the economy within the recent past which prompt for further studies to identify plausible relationship between the two variables so as to correct the abnormally.
It is these contradictions of the relationships between economic growth and unemployment that prompts the need for further research on the same in order that pertinent information is provided to help in the formulation of evidence-based policies to address the abnormally in Tanzania.
1.3 Objectives
This research will be guided by two types of objectives namely the main objective (aim) and specific objectives.
1.3.1 Main Objective
The main objective of this study is to analyze the Impact of Youth Unemployment on Economic Growth drawing a case from Tanzania.
1.3.2 Specific Objectives
i. To assess the extent to which youth unemployment affects the economic growth of Tanzania.
ii. To provide policy recommendations based on empirical findings.
1.4 Hypothesis
Youth unemployment affects economic growth negatively.
1.5 Research Question
To what extent does youth unemployment affect the economic growth of Tanzania?
1.6 Significance of the Study
The significance of this study is two-fold: Firstly, the findings of the study are expected to add to the existing fountain of knowledge about the impact of unemployment to the growth of an economy with the emphasis on youth unemployment; and secondly, the study will provide in concrete terms, policy recommendations geared towards addressing the abnormally in our economy.
1.7 Scope of the Study
The study has also two-fold coverage: Firstly on location, it covers the United Republic of Tanzania which entails the Tanzania mainland and the Island Tanzania (Zanzibar); and Secondly on contextual basis, it looks at Capital Stock (gross capital formation) and labour force as control variables, and youth unemployment, trade openness and foreign direct investment as experimental variables with great emphasis of youth unemployment.
1.8 Limitations of the Study.
The researcher faces two main limitations in the course of conducting this study. Firstly, time constraint since the researcher is a full time employee of the public sector agency. This situation compels him to strike a balance between study and work. Secondly, there is a great challenge on securing data and hence the reliability of the same. For instance, in Tanzania youth are people aged 18-35 but it has been like impossible to find data on that range hence researcher uses 15-24 as a proxy for all unemployed youth in the country.
1.9 Layout of the Study
The paper will be made of five chapters, one through five. Chapter one concerns an introduction of the paper. Chapter two evaluates the theoretical and empirical literature on the impact of youth unemployment on economic growth. Chapter three deals with data types and sources, methods of data presentation and analysis, model specification including both theoretical framework and empirical model and tests carried out in the study. Chapter four will contain the empirical findings and the discussion of results; and chapter five will present conclusions and policy recommendations.

CHAPTER TWO
REVIEW OF THE RELATED LITERATURE
2.0 Introduction
This chapter presents the review of the literature supporting the study area. The chapter chats in concrete terms about the relationship between economic growth and unemployment with emphasis on youth unemployment. The chapter is subdivided into three parts namely: theoretical literature review, empirical literature review and a critique of related empirical works and consequent research gaps.
2.1 Theoretical Literature Review
The main theoretical review supporting this study includes the Marxist theory of unemployment, traditional neoclassical growth theory, Keynesian unemployment theory, and Okun’s theory of unemployment. These theories explain the connection between unemployment and economic progress (GDP) in the development process. The narrative from these theories is that no country can claim to be developing while it is experiencing high level of unemployment, poverty and income inequality. The main insinuation is that unemployment level greatly affect economic growth process of an economy. These theories are explained briefly just to give insights on how they narrate the linkage between unemployment and economic progress (growth).

2.1.1 Marxist theory of unemployment
Karl Marx propounded this theory in 1863. The theory argues that unemployment in any economy is inherent due to insatiable nature of capitalist system. Capitalist unfairly manipulates the labour market by causing unemployment that in turn leads to low demand for labour and hence low wages. The theory therefore, suggests that the best way of reducing unemployment is by ending capitalism and shifting to socialist economic system. In view of these insights, it is vivid that Karl Marx was a proponent of socialism sort of economic system, (Onyebuchi et al, 2016).

2.1.2 Traditional neoclassical growth theory
This theory is the brain child of Solow (1956) and Harrod ; Domar (1957) growth models. The model centered on the significance of saving in an economy. In the model, growth model was expanded through Harrod-Domar postulation by including labour as another control variable besides capital stock and technology as additional variable in the growth equation. The Solow’s growth model demonstrated falling returns to labour and capital exclusively, and constant returns to both factors when they are jointly examined, (Onyebuchi et al 2016).

The main proponents of the neoclassical growth model especially Solow (1956) and Phelps (1961) considered the technological progress to be a residual factor which describes the long term growth, and its level was assumed to be determined exogenously. For instance, Solow (1956) claims that when production takes place there would be no disapproval between natural and superfluous rates of growth. The implication drawn from this scenario is that the system is self-regulating to any given rate of growth of labour (human capital) and eventually nears the state of steady proportionate expansion, (ibid).

2.1.3 The Keynesian unemployment theory
The Keynesian theory of unemployment is also known as cyclical or deficient demand theory of unemployment. The theory explains that ineffective demand in an economy is the primary cause of unemployment in which those that are willing to work at a given wage are unable to find job at a given time, (Obadan & Odusola: 2010) as cited in (Onyebuchi et al: 2016).

Furthermore, the theory argues that as demand for goods and services decreases, production level reduces and hence, few workers are needed in the production process. The theory also emphasized that since the number of unemployed work force would always exceeds job openings, so that even if full employment is achieved, some labour will still remain unemployed due to mismatch in the economy. Keynes therefore, perceived that absence of effective demand for jobs can be fixed by actions of government through deficit spending which can boost employment level and increases aggregate demand in the economy, (ibid).

2.1.4 Okun’s theory of unemployment
Arthur Okun is the American economist who explained the connection amid unemployment and economic progress (growth) in an economy. The theory postulates that unemployment is inversely related with economic growth in any given economy. Okun established that a 1% decrease in unemployment rate leads to 3% increase in economic growth (Okun, 1962) as cited in (Onyebuchi et al, 2016). This study is essentially performed to attest this theory by using data from the World Development Indicators database of the World Bank. For instance, in Tanzania within the recent past, the economic growth rate and unemployment rate have been exhibiting a positive relationship as opposed to inverse relation in accordance with the Okun’s law. This phenomenon is indeed a paradoxical in the sense it opposes the Okun’s law.

2.2 Empirical Literature Review
This subsection of the chapter presents the studies undertaken by different researchers within the area of study. In essence it looks at the methodologies, findings along with other issues that aim at attesting the theory. In this regard, the following are some of the empirical literature reviewed by the researcher.

Onyebuchi et al (2016) examined the relationship between unemployment and economic growth in Nigeria; with a special focus on the impact of unemployment on economic growth for the period of 34 years commencing in 1980 and ending in 2013. They deployed the Co-integration test, Vector Error Correction Model, and the Granger Causality tests to perform their analysis. They used real gross domestic product, unemployment rate and private consumption expenditure to substantiate the model. The co-integration test result exposed that long run association exists among the variables under study. Moreover, VECM result revealed that unemployment has inverse and significant impact on real gross domestic product. Lastly, the Granger Causality results showed a unidirectional relationship amid unemployment and real gross domestic product, with causality running from real gross domestic product to unemployment.

Hussain et al (2010) examined the relationship amid growth and unemployment, using time series data from Pakistan for the time span that ranged from 1972 to 2006. They applied the Johansen Co-integration test and the results showed that there is lengthy relationship between growth and unemployment. They also used Vector Error Correction Model (VECM) to ascertain the short run dynamics and causality, where the results of VECM indicated that there is short and long run causal relationship amid growth and unemployment plus capital, labor and human capital as explanatory variables.

Abraham and Ozemhoka (2017) examined the empirical relationship amid youth unemployment rate and economic growth rate in low-income countries particularly those which are found within the Sub-Saharan Africa. The authors had adopted two different techniques to perform model estimation namely: Panel Least Squares and Ordinary Least Squares. They used datasets which were generated annually from 1991 to 2013, which makes 23 years. The findings revealed that there exists inverse relationships between youth unemployment and economic growth variables in the panel result while in the specific country cases, some countries were found to have progressive relationship between the two variables signifying a case of non-existence of Okun’s law. In some other countries, it was found that there was an inverse relationship between the variables hence the existence of Okun’s law.

Nikolli (2014) examined the relationship between the economic growth and the unemployment rate within the context of Albanian economy. The study utilized data which were obtained for 14 years from 2000 to 2013 inclusive. This was the period when the country had experienced a very severe unemployment problem that posed a negative impact in the economy. The aim of the paper was to test the existence of the Okun’s law in Albania during that period. The paper just used a simple regression model where unemployment was regressed on GDP and the findings failed to ascertain the existence of the Okun’s law for Albania during the period of study.

Gocer and Erdal (2015) worked to ascertain the linkage between youth unemployment and economic growth, within the context of Okun’s law, where they used panel data analysis and executed co-integration test. They collected data for 8 countries from the central and eastern Europe based on the reason that their youth unemployment rate was above the EU member countries (28 countries) average of 25 % for the period of 7 years ranging from 2006 to 2012 inclusive. The results show that, if youth unemployment is quite severe; even an exclusive economic growth will not be enough to reduce the youth unemployment rate in the country.

Sodipe and Ogunrinola (2011) studied the employment and economic growth interactions in the context of the Nigerian economy. They formulated and estimated a simple regression model where they carried out estimations by using the Ordinary Least Squares (OLS) technique. The results for the econometric model analysis revealed that a positive and statistically significant association exists between employment and economic growth, both in level form, in Nigeria whereas an inverse relationship was detected between employment rate and the GDP growth rate in the same economy.
2.3 A Critique of Empirical Literature and Research Gaps
The reviewed related literature revealed that most studies were done by using annual data which are the annual aggregates. Indeed, there are instances when researchers prefer to use quarterly datasets of variables as opposed to the annual datasets. The variables which are used in this study especially GDP are computed on quarterly basis. Based on these situations the researcher has decided to convert from the annual data to quarterly data because through the latter it is easier to identify changes in drifts or even shocks, and it also stands a better chance for future projections.
Additionally, most studies have studied the impact of unemployment on economic progress (growth) by applying the labour force only or labour force and human capital as control variables on one hand; and unemployment or youth unemployment as experimental variable. By considering the traditional growth model worked on by Solow (1956) and its extensions, the control variables are capital stock and labour force. This study is going to utilize both capital stock (gross capital formation) and labour force as control variables which will be in consistency with neoclassical growth model.

CHAPTER THREE
MODEL SPECIFICATION AND METHODOLOGY
3.0 Introduction
This chapter concerns the research methodology that will be used in the study, various sources of data, variables that will be used, technique (s) for data analysis and various tests that will be carried out including the procedures for the interpretation of results.
3.1 Functional Relationship and model Specification
This study model specification is hinged in theoretical foundation of the augmented Solow and endogenous growth models for economic growth equation. The Solow growth model is preferred since it is designed to show how in the long time horizon, the growth in the capital stock, labour force (population) and advances in technology interrelate in an economy (Mankiw, 2010).
Thus, the growth function is expressed as:
…………………………………………………………………………………………………………………….. (1)
Where,
Y denotes Economic Growth
K denotes Capital Stock
LF denotes Labour Force
A denotes technology.
(Salim et al: 2017) explored the impact of Unemployment on Economic Growth. Based on their model, the unemployment was included in the growth model hence it became as follows:
Y= f(K, LF, UNEMPL, A)……………………………………………………… (2)
Where,
UNEMPL denotes total unemployment (population unemployed) in an economy. This paper however considers youth unemployment (YUNEMPL) as opposed to the total population unemployment as used by (Salim et al, 2017) since the whole paper hinges on youth unemployment. Moreover, this paper explores the impact of one more variable which is Consumer Price Index (CPI) as a proxy for inflation.
The modified economic growth function can be expressed as:
Y= f (K, LF, YUNEMPL, CPI) ……………………………………………………..(3)
Where,
Y denotes Economic Growth (GDP)
K denotes Capital Stock (Gross Capital Formation)
LF denotes Labour Force
YUNEMPL denotes Youth Unemployment
CPI denotes Consumer Price Index
Using equation (3) and expressing the variables in natural logarithmic form, an attempt will be made to look at the relative contribution (elasticity) of each variable to the growth process. Also, the use of natural log helps to easy interpretations in cases of the existence of the non-linear relationship among variables. Similarly, it is convenient to change a highly skewed variable into more approximately normal distribution.
The transformed model is thus becomes:
LnGDP = ?0 +?1lnK + ?2lnLF + ?3lnYUNEMPL +?4lnCPI + Ut………. (4)
Where,
lnGDP is the natural logarithm of Gross Domestic Product
lnK denotes the natural logarithm of capital stock
lnYUNEMPL denotes the natural logarithm of Youth Unemployment
lnCPI denotes the natural logarithm of Consumer Price Index
Ut denotes the error term that is normally distributed with mean and constant variance, (Gujarat, 2014).
?0 denotes an intercept
?1, ?2, ?3, and ?4 are the linear coefficients of the explanatory variables.
3.2 Variables Description
The study will use secondary data which will be acquired from the World Development Indicators Database (WDI) of the World Bank. The variables that will be used are GDP, Capital accumulation (K), Labour Force (15-64), Consumer Price Index (CPI) and Youth Unemployment (YUNEMPL).
The table below presents a description of the variables to be used, the source of the data as well as the expected sign of the coefficient.
Table 1: Description of the variables to be used, the source of the data and expected sign of the coefficient.
Variable Name Description Expected sign of the coefficient Source of data
Gross Domestic Product “GDP” (constant 2010 US$)
GDP at purchaser’s prices is the sum of gross value added by all resident producers in the economy plus any product taxes and minus any subsidies not included in the value of the products. It is calculated without making deductions for depreciation of fabricated assets or for depletion and degradation of natural resources. Data are in constant 2010 U.S. dollars.

Dependent variable WDI of the World Bank

Gross capital formation” K” (constant 2010 US$)
Gross capital formation (formerly gross domestic investment) consists of outlays on additions to the fixed assets of the economy plus net changes in the level of inventories. Fixed assets include land improvements (fences, ditches, drains, and so on); plant, machinery, and equipment purchases; and the construction of roads, railways, and the like, including schools, offices, hospitals, private residential dwellings, and commercial and industrial buildings. Positive WDI of the World Bank

Labour Force “LF”
Labor force participation rate for ages 15-64, total (%) (Modeled ILO estimate).
Labor force participation rate for ages 15-64 is the proportion of the population ages 15-64 that is economically active: all people who supply labor for the production of goods and services during a specified period. Positive WDI of the World Bank

Youth Unemployment (YUNEMPL))
Unemployment, youth total (% of total labor force ages 15-24) (modeled ILO estimate).
Youth unemployment refers to the share of the labor force ages 15-24 without work but available for and seeking employment. Negative WDI of the World Bank

Consumer price index (2010 = 100)
Consumer price index reflects changes in the cost to the average consumer of acquiring a basket of goods and services that may be fixed or changed at specified intervals, such as yearly. The Laspeyres formula is generally used. Data are period averages. Positive WDI of the World Bank

3.3 Prior Estimation Time Series Tests
The dataset will be tested in priory before the model is fitted in. Thus, the tests like stationarity, cointegration, and multicollinearity will be performed before the data are fitted in the model. The essence of perforimng the cointegration analysis is to inspect the problem of a spurious regression results and rectify it for the equation to yield meaningful results in economic terms, Engle and Granger (1987)
3.3.1 Descriptive Analysis
It is highly recommended to start with providing statistical descriptive analysis of the used variables in the study before doing anything. This involves checking the measures of central tendencies like mean, mode, median; measures of dispersion like standard deviation; as well as measures of peakedness of the dataset like kurtosis.
3.3.2 Unit Root Test
For the econometric analysis, it is highly recommended to subject your dataset for the stationarity test to ascertain whether they are stationary at their level form or they become stationary after being differenced. The reasons for performing the stationarity test at first place are: First, the non-stationary series can highly influence the conduct and features of time series data. Second, it helps to avoid spurious regression that results from non-stationary data. Third, the standard assumptions of distribution will not be valid for non-stationary variables. That is, the usual F-statistic and t-ratios will not follow an F and t distributions respectively (Brooks, 2008). The most popular strategy for testing the unit root is the ADF (Nkoro and Uko: 2016) and Philips–Perron tests (ibid). This study will apply both tests, the ADF and Philips– Perron (PP) tests to ascertain the degree of stationarity for the data. The Augmented-Dickey Fuller (ADF) test assumes the following hypotheses:
H0: time series has a unit root and
H1: time series has no unit root.
The essence of performing this test is to determine the order of integration of the dataset by applying either ADF or PP unit root test, postulated by Dickey ; Fuller (1981). Besides the order of integration, the test is also applied so as to find the long term properties of the variables in the study. In case the time series data are found to be stationary, it means that their variance, mean and covariance are constant overtime and that the result obtained from their analysis is reliable and can be used to predict future economic activities of the economy, (Wooldridge, 2004).
3.3.3 Johansen Co-integration Test
The classical linear regression model (CLRM) assumes that the dependent and independent variables are stationary over time, Gujarat (2004). However, most economic variables exhibit long time horizon trend movement and only become stationary after they are differenced, (ibid). Therefore any equilibrium relationship among a set of non-stationary variables implies that the variables cannot move independently of each other.
Therefore, this estimation procedure involves the test of the level of co-integration among the data series of the same order through the application of the Johansen Co-integration test. It therefore implies that if in the long run, two or more data series move strictly together, whether the data series itself is trended, the difference between them is constant. In theory, they can roam arbitrarily faraway from each other; hence achieving empirical result leading to establishing maximum-likelihood test procedure, Johansen ; Juselius (1990).

3.3.4 Vector Error Correction Model (VECM)
This test is carried out immediately after performing the Co-integration test, especially when the co-integration test showed evidence of long run relationship among the variables. The traditional vector error correction model (VECM) is employed to examine the short run dynamics and co-integrating equation among the data series, (Onyebuchi et al: 2016). The term ‘error correction term’ is estimated for the coefficients, such that when the series fails to co-integrate, it means that the short run model becomes the next estimation method. The concept of VECM is used to explain the relationship existing between short run dynamics and long run equilibrium relationship among the data series. The application of VECM is necessary as it is used to correct temporary short run deviation of series from the long run equilibrium relationship, (Onyebuchi et al: 2016). The method of vector error correction model is estimated to investigate the dynamic behaviour of the relevant variables of the study, following the confirmation of long run equilibrium relationship, (ibid).

3.3.5 The Granger Causality Test
At this juncture, the researcher will scan the causality between unemployment and economic growth through the application of the Granger causality test propounded by Engle ; Granger (1989). The test focuses on determining the nature of relationship between the two variables; that is, to ascertain whether the direction of the relationship is one-way, two-way, or no causation between the two variables.
3.4 Diagnostic Tests
These are tests done to check for the violations of the basic linear classical assumptions.
3.4.1 Normality test
This test is aimed at testing for specification errors or non-normality which violate the assumption that the disturbances are normally distributed and a histogram-normality test was used to test the normality of the residuals.
3.4.2 Autocorrelation test
This test will be done to check the classical linear regression assumptions that the errors entering the regression function were random. Breusch-Godfrey serial correlation (LM) test will be performed to establish whether serial correlation exist in the model or otherwise. The null hypothesis of no serial correlation (H0: No serial correlation), will be tested against the alternative hypothesis of serial correlation, (H1: There is Serial Correlation).
3.4.3 Heteroskedasticity test
Breusch Pegan Godfrey test was conducted in order to ascertain whether the disturbances or errors have the same variances such that the Ordinary Least Squares estimators are efficient or have minimum variance.
3.4.4 Stability test
This test aims at ascertaining for the omitted variables that is; the vector of the regressors does not include irrelevant variables or incorrect functional form since Ordinary Least Squares (OLS) estimators would be biased and inconsistent in case of incorrect functional form. Cummulative sum (CUSUM) and the Recussive Cummulative sum (RCUSUM) estimation techniques will be used to determine the stability of the model.
3.4.5 Multicollinearity Test
Aman (2016: pp41) argues that multicollinearity test is performed in order to identify correlation between explanatory variables for the purpose of avoiding double effect of the regressors within the same model. Multicollinearity in essence is the problem of degree of the association and not the lack of relationship. A correlation is a single number that describes the extent of association between two variables (ibid.).
3.5 Data Analysis
To aid the estimation of the experimental variables (youth unemployment, trade openness and FDI) as insinuated prior, the quarterly data for the period 1991:01 – 2017:04 will be obtained and used. The obtained data will be analyzed using STATA version 14 to perform the Ordinary Least Squares regression in order to ascertain whether the selected variables significantly inform or determine Economic Growth in Tanzania.
Annual data will be obtained and in order to create more data points, it will be converted to quarterly range using the technique called the quadratic match average.

REFERENCE
Abraham and Ozemhoka (2017) Youth Unemployment and Economic Growth: Lesson From Low-Income Countries in Sub-Saharan Africa.
Gocer and Erdal (2015). The Relationship between Youth Unemployment and Economic Growth in Central and Eastern European Countries: An Empirical Analysis
Hussain et al (2010) A Coherent Relationship between Economic Growth and Unemployment: An Empirical Evidence from Pakistan
Ihensekhien (2017). Youth Unemployment and Economic Growth: Lesson From Low-Income Countries in Sub-Saharan Africa
Ismet Gocer and Leman Erdal (2015) The Relationship between Youth Unemployment and Economic Growth in Central and Eastern European Countries: An Empirical Analysis Leman ERDAL
N. Gregory Mankiw(2010). Macroeconomics: Harvard University
Nikolli (2014) . Economic Growth and Unemployment Rate. Case of Albania
Nkoro and Uko (2016). Autoregressive Distributed Lag (ARDL) cointegration technique: application and interpretation
Onyebuchi et al (2016). The Relationship between Unemployment and Economic Growth in Nigeria. Granger Causality Approach
Salim Hamad Suleiman et al (2017). Unemployment and Economic Growth in Tanzania. An examination of the impact of unemployment on economic growth in Tanzania and causal relationship between unemployment and economic growth in Tanzania.
Sodipe and Ogunrinola (2011). Employment and Economic Growth Nexus in Nigeria

Post Author: admin

x

Hi!
I'm Chloe

Would you like to get a custom essay? How about receiving a customized one?

Check it out