FACULTY OF THE SOCIAL SCIENCES - Home - University Of Nigeria …S THESIS... ·  · 2015-08-29trending unemployment in Nigeria. ... which taken together can be said to constitute

ANYANWU, KELECHI CLARA

PG/M.SC/10/52414

Ogbonna Nkiru

Digitally Signed by: Content manager’s Name

DN : CN = Webmaster’s name

O= University of Nigeria, Nsukka

OU = Innovation Centre

FACULTY OF THE SOCIAL SCIENCES

DEPARTMENT OF ECONOMICS

GOVERNMENT LABOUR POLICY, GRADUATE

UNEMPLOYMENT AND LABOUR PRODUCTIVITY IN

GOVERNMENT LABOUR POLICY, GRADUATE

UNEMPLOYMENT AND LABOUR PRODUCTIVITY IN

NIGERIA

BY

ANYANWU, KELECHI CLARA

PG/M.SC/10/52414

AN M.Sc THESIS SUBMITTED TO THE DEPARTMENT OF

ECONOMICS, FACULTY OF THE SOCIAL SCIENCES

UNIVERSITY OF NIGERIA, NSUKKA

IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE

AWARD OF MASTER OF SCIENCE DEGREE IN ECONOMICS

SUPERVISOR: PROF IKPEZE, N. I.

SEPTEMBER 2014

i

TITLE PAGE

GOVERNMENT LABOUR POLICY, GRADUATE

UNEMPLOYMENT AND LABOUR PRODUCTIVITY IN

NIGERIA

ii

CERTIFICATION

This is to certify that ANYANWU KELECHI CLARA, a post-graduate student of

the Department of Economics, University of Nigeria, Nsukka, and whose

registration number is PG/M.Sc/10/52414 has satisfactorily completed the

requirements for the award of Master of Science (M.Sc) Degree in Economics.

PROF. N. I. IKPEZE PROF C. C. AGU

Supervisor Head of Department

iii

APPROVAL PAGE

This project has been read and approved as meeting the requirements for the award

of the Degree of Master of Science (M.Sc) of the Department of Economics,

University of Nigeria, Nsukka.

PROF. N. I. IKPEZE PROF C. C. AGU

Supervisor Head of Department

PROF. IGNATIUS, A. MADU

Dean, of Social Sciences External Examiner

iv

DEDICATION

This work is dedicated to the almighty God for his mercies, protection and grace

throughout this programme.

v

ACKNOWLEDGEMENT

I am elated to appreciate the efforts and contributions of all the staff of Economics

Department for their various roles towards the completion of this programme. My

deep appreciation goes to my lecturers: Professor Agu C.C, Professor Ona F, Rev

Fr Professor Ichoku H.E, Prof Chidebelu S.A.N.D; Dr Oleru J.O, Dr Nwosu

Emma, Dr Ukwueze Ezebilo, Dr Mrs Aneke, Dr Ifelumini (SPG), Dr Oduh

Moses,Mr Jude Chukwu, Dr Urama Nathaniel.

My special thanks to my former supervisor Prof Onyukwu E. Onyukwu and my de

jure supervisor Prof Ikpeze N.I for your guidance.

To those that have read and will read this work, thanks for their assistance.

I am greatly indebted and in no small measure to my parents, Mr and Mrs

A.A.Anyanwu, for their constant prayers and advise.

To my family; the Anyanwu’s, Okwulehie’s , Uneze’s, Chimazuru and Chikamara

for their encouragement, moral and financial support.

Finally, to my friends: Andrew, Nemerem, Fidel, Parson, Chika, Steve, Ugo,

Doris, Uche, Akudo, Nnamani, Kelechi, Pat, Casmir and Peter. Thanks for

everything; I really enjoyed your company.

vi

ABSTRACT

The study evaluates government labour/employment policy, graduate unemployment and

labour productivity in Nigeria. It spanned the period between 1987 and 2013. The

model was built on the combination of Keynesian framework and modern labour market

theory and the various channels through which labour policy can impact on graduate

unemployment and labour productivity in Nigeria. The study employed Ordinary least

square method of estimation, co-integration and error correction techniques in

estimating the model. Data were sourced from Central Bank of Nigeria annual statistical

bulletin/reports (various years) and National Bureau of statistics annual statistical

reports. The econometric software used for the study was E-views. The result from trend

analysis showed that labour productivity proxied by growth in Gross Domestic Product

(GDP) maintained steady and near horizontal movement while graduate unemployment

has been on the increase over the years. Co-integration result showed that there exists a

long-run linear relationship among the variables used for the study, hence adopting an

error correction model. The result clearly indicated that government employment policy

proxied in this study by the total number of graduates that have benefited (TNA) from

National Directorate of Employment (NDE) since its inception in 1986 have positive and

significant influence on labour productivity and negative significant influence on

graduate unemployment. The study also found insignificant relationships between

graduate unemployment, gross secondary school enrolment and total factor productivity

in Nigeria. The study went further to establish unidirectional causal link between

government labour policy, graduate unemployment and labour productivity in Nigeria.

The study recommends that holistic programme should be established to arrest high

trending unemployment in Nigeria. The programme should capture those who are

heavily underutilized and grossly underpaid, in order to have a fair representation of the

unemployment situation in the country. Government spending pattern should be checked,

since the rising government total spending has not translated to increase in labour

productivity.

vii

TABLE OF CONTENTS

Page

Title Page .. .. . i

Certification Page .. .. .. ii

Approval Page .. .. .. iii

Dedication .. .. .. iv

Acknowledgement .. .. .. v

Abstract vi

Table of Contents .. .. . vii

List of Tables .. .. . x

List of Figures xi

CHAPTER ONE: INTRODUCTION

1.1 Background to the Study ... .. 1

1.2 Statement of the Problem ... .. . 3

1.3 Objectives of the Study ... .. .. 5

1.4 Research Hypotheses 5 .

1.5 Policy Relevance of the study ... .. 6

1.6 Scope of the Study ... .. .. 6

CHAPTER TWO: LITERATURE REVIEW

2.1 Conceptual frame work ... . . 8

2.2 Theoretical Literature ... .. .. 10

viii

2.3 Empirical Literature ... .. .. 17

2.4 Literature Gap and Value Addition ... . . 22

CHAPTER THREE: METHODOLOGY

3.1 Theoretical Frame work ... .. .. 24

3.2 The Model 26

3.3 Model Specification 27

3.4 Estimation Procedure 30

CHAPTER FOUR: PRESENTATION OF RESULTS

4.1 Analysis of the Data Generating Process .. .. .. 34

4.1.1 Trend Analysis of the Core Variables .. .. 34

4.1.2 Unit Root and Co-integration Analyses .. .. 35

4.2 Presentation and Analysis of ECM Models .. .. 38

4.2.1 Endogeneity Test .. .. .. 38

4.2.2 Test for Validity of Instruments 40

4.2.3 The ECM Result for the Models 41

4.2.4 Total Factor Productivity Result 44

4.2.5 Granger Causality Result 46

4.3 Evaluation of Research Hypotheses 47

CHAPTER FIVE: SUMMARY AND CONCLUSION

5.1 Summary of Research Findings 48

5.2 Policy Recommendations ... .. .. 49

5.3 Conclusion ... .. .. 50

ix

5.4 Limitations of the Study ... .. .. 51

References ... .. .. 52

Appendix 57

x

LIST OF TABLES

Tab 1: Summary of Literature Reviewed 19

Tab 2: ADF and PP Unit Root Test Results for Individual Variables 36

Tab 3: Co-Integration Tests 37

Tab 4. Endogeneity Result for Graduate Unemployment (UG) Model 38

Tab 4.1: Endogeneity Result for Labour Productivity Model 39

Tab 4.2: ECM result for Graduate Unemployment (UG) Model 41

Tab 4.3: ECM Result for Labour Productivity Model 42

Tab 4.4: Result of the Total Factor Productivity Model 45

Tab 4.5: Pairwise Granger Causality Tests 46

xi

LIST OF FIGURES

Fig 1: NATIONAL UNEMPLOYMENT RATE ( 2000-2009) 19

Fig 2: TREND ANALYSIS OF THE CORE VARIABLE 35

1

CHAPTER ONE

INTRODUCTION

1.1 Background of the Study

There is no doubt that unemployment is one of the major challenges facing economies of

the world (developed and developing). This exerts more distorting impacts on the developing

economy. According to Ekpo (2008), a developing economy such as Nigeria’s is faced with

poor growth performance which manifests in the rising incidence of poverty, massive graduate

unemployment, skyrocketing inflation, worsening balance of payments disequilibrium,

monumental external debt burden, widening income disparity and growing fiscal imbalances,

which taken together can be said to constitute major causes of underdevelopment. In all, rising

graduate unemployment poses the most pervasive and devastative effect which threatens the

productivity of the Nigerian economy (Ake, 2010).

Graduate unemployment is defined as the unemployment among people with academic

degrees. It is a situation where tertiary institution graduates do not get jobs after going through

the academic ladder successfully. It is the greatest component of aggregate unemployment.

According to the International Labour Organization (1982), the unemployed are persons that are

available and willing to work but without work in the past 39 weeks. One is forced to ask how

many Nigerians are willing and available to work but are currently without job. On the other

hand, Frank and Bernard (2001) noted that the rising unemployment rate in a nation is too

significant to be ignored as it is necessary in assessing the level of economic activity in such

nations. Thus, besides real GDP, unemployment and growth in labour productivity remains

economic statistics that receives a great deal of attention from both economists and the general

public.

The unemployment rate is a sensitive indicator of the conditions of the labour market. When

the unemployment rate is low, jobs are secured and relatively easier to find. Low unemployment

is often associated with improving wages and working conditions as well as employers

competing to attract and retain workers.

In the recent past, Nigeria has experienced low labour demand and productivity, a sign

which is widely blamed on the failure of government policies and programmes over the years.

Specifically, since the mid 1980s, there has been an alarming increase in the rate of graduate

unemployment, low labour productivity and its attendant’s social and economic consequences.

Just like some other developing nations in Africa, the Nigerian government and policy makers

are increasingly finding it difficult to deal with graduate unemployment successfully.

2

The high rate of graduate unemployment in Nigeria according to Adeyeye et. al. (2012),

can be associated with lack of adequate provision for job creation in the development plans, the

ever expanding educational growth and the desperate desire on the part of youths to acquire

tertiary education irrespective of the social and economic reality. Consequently, a number of

skills acquired from these tertiary institutions appear dysfunctional and irrelevant since most of

the skills and knowledge acquired in tertiary institutions are kept redundant through

unemployment and sometimes skills are not fully utilized.

On the other hand, as the graduate population in Nigeria increases without being

absorbed in the active market, labour productivity does not increase at its full potentials, this

could be as a result of the failure of government to control this phenomenon over the years in

spite of numerous programmes and policies on this issue. Without denying the impact of other

factors, unemployment has exacerbated social ills and delinquent behavior among youths (most

of whom are graduates) especially armed robbery, political thuggery, advanced fee fraud and the

recent spate of terrorism (Boko Haram) in Nigeria, which has been on rampage for the past 4

years, forcing the government to be spending a lot of money on crime control. Recent statistics

released by the National Bureau of Statistics (NBS) showed that about 10 million Nigerians

were unemployed. In fact, this number swells by 120,000 graduates each year, which are

produced with little or no jobs waiting for them. (NBS 2009).

In a bid to address the menace of unemployment, various policies have been put in place

by the Federal Government of Nigeria. Notably, the Small and Medium Enterprise Development

Agency of Nigeria (SMEDAN) was established in 2003 to promote the development of Micro,

Small and Medium Enterprises (MSME) sector of the Nigerian economy. It is to source, process

and disseminate business information, develop policy, establish business support programmes,

build capacity and promote services, enhance MSME access to finance. Others are the Nigerian

Agricultural Co-operative and Rural Development Bank (NACRDB) Limited which is dedicated

to financing agriculture at both micro and macro levels. They are to provide affordable financial

and advisory services to the farm and non-farm enterprises of the Nigerian economy using well

trained and highly motivated staff, backed by appropriate technology. Those that were

established but later scrapped include: the Directorate for food Roads and Rural Infrastructures

otherwise known as (DFRRI), Mass Mobilization for Self Reliance and Economic

Reconstruction (MAMSER) and the National Agricultural Land Development Project

(NALDA). These were created by the Babangida regime but scrapped by the Abacha regime.

Among all, the most innovative of these programmes is the National Open

Apprenticeship Scheme (NOAS) introduced by the Federal Government in 1987. The NOAS is

3

an attempt to link education with training and labour demand. It was managed by the National

Directorate of Employment (NDE) under the Ministry of Labour and Productivity. The NDE is a

policy document that addresses the provision of graduate level of employment, the essence of

this is to know how government policy should be used effectively in the reduction of graduate

unemployment and enhance productivity in Nigeria. This was also meant to provide vocational

education and training to unemployed youth in over 100 occupations. The main objective of

these government programmes and policies is to provide means of livelihood to able but idle

army of unemployed youths, especially the graduates and give assurance to the private sector

through the services of the youths.

Unfortunately, the observable low labour productivity and high level of graduate

unemployment over the years in the country have shown that no meaningful progress has been

made by these programmes and policies. It is against this background that this study is designed

to examine the nexus between government policy, graduate unemployment and labour

productivity in Nigeria and its attendant consequences to the social and economic wellbeing of

the nation.

1.2 Statement of the Problem

Tackling graduate unemployment and low labour productivity in Nigeria has remained

twin problems facing the country for some decades now. The Nigerian total labour force is made

up of all persons aged 15-64 years excluding students, home keepers, retired persons and stay-at-

home to work or those that are not interested. Nigeria’s unemployment can best be defined as the

proportion of labour force that was available for work but did not work in the week preceding the

survey period for at least 39hours (Asoluka & Okezie, 2009). The 2009 official figures from the

Bureau of Statistics put the unemployed figure at 19.70 per cent, about 30 million. Though, this

figure still did not include more than 40 million other Nigerian youths (some graduates of various

degrees) captured in World Bank statistics in 2009. By implication, it means that if Nigeria’s

population is 167 million, according to World Bank figure, then 50 percent of Nigerians are without

employment (unemployed). Viewing this from the perspective of the recent events in the Middle

East where unemployment and poverty among others played a key role in the uprising, one can only

conclude that Nigeria’s unemployment poses a threat to not only productivity and output growth, but

its security and peaceful co-existence.

The rising trend of unemployment assumed a doomsday scenario in Nigeria a decade

after political independence. As noted by Akintoye (2008), between 1970 and 1980, national

unemployment rate rose from 4.3 to 6.4% and further rose to 7.1% in 1987. This, according to

Akintoye, is attributed to the economic depression which engulfed the nation from 1980,

4

resulting in massive closure of businesses and retrenchment of workers. This was followed by

the placement of embargo on recruitment which further worsened the unemployment situation.

The Structural Adjustment Programme (SAP) established in 1986 had on its part, a

positive effect on job creation leading to a sharp fall in unemployment rate from 7.1% in 1987 to

1.8% in 1995. Thereafter, unemployment figure hovered around 4% between 1996 and 2000.

One worrisome trend in the Nigeria labour market of recent has been the growing incidence of

unemployment among professionals such as accountants, engineers and other graduates from

universities and other tertiary institutions. Graduate unemployment accounted for 32% of the

total unemployed labour force between 1992 and 1997 (Akintoye,2008). This growing incidence

of graduate unemployment in the face of acute skill shortages presents a paradox which further

complicates the analysis of labour market distortions in Nigeria. (Dabalen et al., 2000).

Expectedly, unemployment reduction has remained the central focus of macroeconomic

goals in Nigeria. It is a continuing policy and responsibility of the federal government to use all

practical means to promote higher level of employment, production and purchasing power

(Essien & Atan, 2006). The most critical factor explaining the rising unemployment in the

country is the failure of government policies to consciously tackle graduate unemployment

among others. The need to avert the negative effects of unemployment and improve labour

productivity through effective government policies will make the tackling of unemployment

problems to feature prominently in the development objectives of the Nigerian government.

The primary cause of graduate unemployment and low productivity is the absence of an

appropriate and well articulated government policy to guide the strategies and programmes of

the various institutions operating in all sectors of the economy (Asoluka & Okezie, 2011).

However, efforts by different regimes in Nigeria such as Federal Ministry of Labour and

Productivity, the National Directorate of Employment (NDE), Small and Medium Scale

Enterprise (SMEs) and National Poverty Eradication Programme (NAPEP) in 2001 to generate

more employment and improve labour productivity to a large extent have not yielded the

required results (Essien & Atan, 2006).

Apart from representing a colossal waste of a country’s manpower resources, graduate

unemployment and low productivity generates welfare loss in terms of lower output thereby

leading to lower income and wellbeing (Ibrahim, 2006). In 2005 the unemployment rate as

recorded by the NDE was 11.9% in 2006, 14.6% in 2007, and 10.9% in 2008 while as at

December 2009 the Bureau of Statistics gave the unemployment rate as 19.7%. Predictably, this

has also been accompanied by a high rate of social vices in Nigeria.

5

In view of the unfolding reality coupled with the protracted debates in the literature, this

study is built to examine the nexus between government policy, graduate unemployment and

labour productivity in Nigeria. In doing this, the study will be guided by the following research

questions.

1.3 Research Questions

1. What is the nexus between government labour policy, graduate unemployment and

labour productivity in Nigeria?

2. What is the long-run impact of graduate unemployment on total factor productivity in

Nigeria?

3. What are the causality between government policy, graduate unemployment and labour

productivity in Nigeria?

Based on the above research questions the following objectives are drawn for this study.

1.4 Objective of the Study

The broad objective of this study is to examine government labour policy, graduate

unemployment and labour productivity in Nigeria. In view of this, the specific objectives of the

study are;

1. To determine the nexus between of government labour policy, graduate unemployment

and labour productivity in Nigeria.

2. To determine the long-run impact of graduate unemployment on total factor productivity

in Nigeria.

3. To determine the causality between government labour policy, graduate unemployment

and labour productivity in Nigeria.

1.5 Research Hypotheses

Based on the objectives of this study, the following research hypotheses are formulated:

1. Ho: Government policy on labour has no significant impact on the rate of graduate

unemployment and labour productivity in Nigeria.

2. Ho: There is no long-run relationship between graduate unemployment and total factor

productivity in Nigeria.

3. Ho: There is no causal link between government labour policy, graduate unemployment

and labour productivity in Nigeria.

6

1.6 Policy Relevance of the Study

The findings of this study will aid the government efforts in addressing the protracted

unemployment, especially graduate unemployment and low labour productivity in Nigeria.

Firstly, identification of the nexus between government labour policy, graduate unemployment

and labour productivity in Nigeria will assist policy-makers in formulating the right policies to

addressing protracted rising unemployment trend in Nigeria. Policy-makers will therefore find

the study very relevant for drawing policy issues in line with the challenges of ensuring an

improved labour productivity and reducing general unemployment rate, irrespective of gender

and level of education, in line with the development agenda of the present administration.

Secondly, investigating the long-run impact of graduate unemployment on total factor

productivity will be informative in predicting how total factor productivity will change if policy

makers are to change in reducing the rising trend of graduate unemployment status. Finally,

providing an insight towards understanding the causal relationship between graduate

unemployment, government labour policy and low productivity in the Nigerian economy, will

be an important tool in designing effective policy interventions that addresses graduate

unemployment and labour productivity issues and achieving the new policy framework (mass

employment for the youths).

1.7 Scope of the Study

This study is an empirical analysis of the nexus between government labour policy,

graduate unemployment and labour productivity with evidence from the Nigerian economy. It is

important to note in this study that the connection between government policy and

macroeconomic phenomenon such as unemployment is not well developed. In principle, a

computable general equilibrium model (CGE) should be developed; however, unemployment

sensitive CGE models have not been developed and face many problems (Fontana & Wood,

Lofgren et al., 2008). Hence these are beyond the scope of this study. Instead, this study

addresses three graduate unemployment issues that are thought to affect both the demand as

well as supply of labour in Nigeria. These are the interaction between employment rate, labour

demand and the number of graduates produced in the economy every year.

This study considered the effective government labour/employment policies instruments

targeted at graduate unemployment in National Directorate of Employment (NDE) scheme.

These include the start your Own Business (SYOB) under the small Scale Enterprise programme

(SSE), the sensitization of National Youth Service Corps (NYSC) on Entrepreneurship

development, the Graduate Attachment Programme (GAP) and the Solar Energy Training

7

Scheme (SETS) which was recently introduced. However, the study estimated the government

labour policy with the total number of persons that benefited from NDE graduate programmes

annually (TNA).

Total Factor Productivity (TFP) is a notion linked to the aggregate production function.

Hence total factor productivity was measured in this study as the ratio of output and weighted

input factors (to be specified in the methodology). Productivity, on the other hand, is a technical

concept which refers to a ratio of output to input and a measure of efficiency. (see Classical

Ricardian labour theory of value).

Finally, graduate unemployment is referred to as graduates who were available for work

and looking for jobs, but unable to find employment. Again, this study covers a reasonable range

of observations, ranging from 1987 to 2013 in order to have a clear picture of the nexus between

government labour policy, the rate of graduate unemployment and labour productivity growth in

Nigeria. This period was chosen due to the fact that government labour/employment policy

considered in this study is captured from the National Directorate of Employment which was

established on November 22, 1986 and its initial core programmes were formally launched on

30th January 1987.

8

CHAPTER TWO

LITERATURE REVIEW

2.1 Theoretical Literature

Theories of Unemployment

Scholars have propounded various theories relating to employment, underemployment

and unemployment. These include those of the Classical theory of unemployment, Innovations

theory of unemployment and Effective Demand theory of unemployment.

Classical Theory of Unemployment

The classical theory, as analyzed by Pigou (1933) and Solow (1981), argues that the

labour market consists of demand and supply of labour. Demand for labour is a derived demand,

obtained from the declining portion of the marginal product of labour. The demand curve is a

negative function of real wage in that if wages increase, the quantity demand for labour will

decline and the opposite is correct. The supply of labour is derived from worker's choice

whether to spend part of their time working or not working (leisure). Supply of hours worked is

a positive function of the real wage, because if the real wage rises, workers supply more hours of

work. In equilibrium, demand and supply of labour are intersected at a clearing point that

determines the equilibrium real wage rate and full employment.

Essentially, for Wicksell the cyclical unemployment was due to the wrong investment of

capital. Capital was invested in areas where rates of return were low. He concluded that public

works is the best measure to fight cyclical unemployment. After World War I, Wicksell thinks

that the boom and the rise in prices induced by the war would come to an end. Thus,

unemployment would rise. Workers would have to accept lower wages. He also thought that

government should provide financial support to the unemployed who could not find jobs. After

1921, Wicksell turns to Malthus. He thought that the causes of the unemployment are the surplus

people, shortage of capital brought about by the war, and the disorganized state of the monetary

system. For the third cause, after the war prices were falling and producers decided to produce

lower amounts of production because they knew they would receive lower prices for their

products. Thus, they let their money lie idle in banks and workers became unemployed. These

causes suggest that emigration became one of the important policies for solving the

unemployment problem.

Wage reduction is not a competent policy to increase employment. The increase in

wages is most likely due to increased labour productivity and wage reduction will reduce work

9

intensity and productivity. Wage reduction will not force some capital intensive firms to switch

to labour intensive techniques in the short run. Higher wages should stimulate the substitution

effect by employing more machines for labour. And this substitution will increase labour

productivity and employment in the long-run.

Hayek contends that unemployment is due “to a discrepancy between the distribution of

labour and industries, and the distribution of demand among their producers. This discrepancy is

caused by a distortion of the system of relative prices and wages.” In other words, the

unemployment is caused by “a deviation from the equilibrium prices and wages which would

establish them with a free market and stable money.” This is actually a mismatch between

demand and supply of labour, which is usually caused by expansionary monetary and fiscal

policies and powerful trade unions. These policies create economic dislocation and structural

changes in an economy which misdirect labour and other economic resources to other

alternatives. Unions are also able to set higher wages compared to market wages, which generate

unemployment, particularly in industries that become less profitable. In short, for Hayek the

unemployment problem is caused by resources being in the wrong places at the wrong time and

can be corrected if wages and prices are determined by the equilibrium of supply and demand.

(Nishhiyama and Leube 1984)

In line with Hayek theory of unemployment, Trehan (2001) provides an important

explanation of the search theory of unemployment. Firms search for the productive workers and

workers search for high-paying jobs, so both agents continue searching until matches are

reached at the point a worker will leave the unemployment pool. But if a worker realizes later on

that her productivity is worth higher wages and firms are paying high wages on the average, then

the worker’s reservation wage will increase. Consequently, the unemployment rate will start

rising gradually, indicating that a mismatch has occurred again.

Innovations Theory of Unemployment

Originally, this theory was developed by the German economist Von Mangoldt in 1855

in a book of entrepreneurial profits which connected profits to risk but this theory was refined in

2007 by Ekelund and Hebert. They provided several ways by which the entrepreneur can make

profits. These ways are (1) finding particular markets, (2) acquisition of productive agents, (3)

skillful combination of factors of production, (4) successful sales policy, and (5) innovations. It

is a well understood proposition that entrepreneurial profits will increase employment

(Mohammed 2010).

Schumpeter (1934) does not provide explicitly a theory of unemployment but his theory

of the business cycle does demonstrate clearly how unemployment can be reduced. Innovation

10

(see also Vecchi 1995) which creates more jobs relative to job destruction is the basic force

beyond the increases in employment and the decreases in unemployment. When entrepreneurs

innovate something new such as the production of a new product, the finding of a new market,

the finding of a new method of production, and the introduction of new technologies and a new

organization they increase investments to materialize those innovations. Domestic investment

expenditures will increase demand on economic resources and will increase their prices. Other

entrepreneurs will imitate the leaders by adopting the new innovations. Labor and materials will

be employed to produce the new items. Consequently, wages will be increasing and

unemployment will be declining, assuming that employment creation will outweigh employment

destruction due to the new innovations (see also Mortensen and Pissarides 1998 and Manuelli

2000).

According to Schumpeter (1934: 64), economic development generates changes in the

socio-economic environment, including the existing equilibrium. As he puts it, “Development is

spontaneous and discontinuous change in the channels of the flow, disturbance of equilibrium,

which forever alters and displaces the equilibrium state previously existing.” The essential

driving force for generating development is innovations introduced by the entrepreneurs whose

leadership becomes the triggering device for the discontinuous dynamic changes. Innovations

start by “the producer [not consumer] who as a rule initiates economic change and consumers

are educated by him if necessary” (Schumpeter 1934: 65).

The concept of innovation which creates changes according to Schumpeter (1934: 66)

covers the following five areas of development: “(1) the introduction of new good...or of a new

quality of a good. (2) The introduction of a new method of production, (3) The opening of a new

market,(4) The conquest of a new source of supply of raw materials, or manufactured goods,(5)

The carrying out of the new organization of any industry, like the creation of a monopoly

position or the breaking up of a monopoly position.” The new combinations are usually

embodied in new productive enterprises which start by utilizing the unemployed working

people, the unsold raw materials, the new technologies, and the unused productive capacity.

Effective Demand Theory of Unemployment

The level of aggregate demand will provide the necessary increases in total revenues. On

the other side, the cost of production has to decline. If revenue rises and cost declines, then the

reasonable level of profits can be found. There are various forces in Veblen’s work that reduce

the cost of production. Technology increases production and reduce the cost of inputs used in

the production process, and enterprises cut wages and increase productivity in order to cut cost

per unit of output. Better technology can reduce the prices of capital goods, and government can

11

cut taxes. Banks can reduce the interest rates as well. Administrative and insurance cost can be

declined in order to stimulate business enterprises. The decline in costs, given rising revenues,

will increase the profit level according to Veblen. Consequently, higher profits will force the

business enterprises to expand and employ more workers. Thus, employment will increase and

the rate of unemployment will decline.

Keynes (1936) considers unemployment as an involuntary phenomenon. He thinks that

employment is cyclical, generated by the deficiency of aggregate demand (Mohammed, 2010).

Capitalists hire workers and invest to produce output when the expectations about the economy

and profits are favorable. If expectations about the future are supported by reality, investments

and employment continue rising until equilibrium is reached. This equilibrium is attained by the

intersection of the aggregate demand and supply--the point of the effective demand—which may

be less than the full employment equilibrium. If expectations about the future of the economy

are not favorable, capitalists invest less and employ less number of workers. Hence, the

equilibrium is achieved where cyclical unemployment exists. This unemployment is due to the

deficiency of the aggregate demand, particularly investment expenditures.

Graduate Unemployment and Government Policies

According to Olufemi and Adebola (2011), both government and policy makers are

increasingly finding it difficult to deal successfully with graduate unemployment, part of the

reasons was blamed on lack of adequate provision for job creation in the development plans of

the governments, thereby rendering the skills acquired dysfunctional and irrelevant to them and

the entire economy. But still more and more graduates are being turned out from the various

universities every year.

The Government of Nigeria had attempted to curb graduate unemployment. Firstly,

through the National Directorate of Employment (NDE), this is one of the steps taken by the

Nigerian government to reduce the problem of unemployment in Nigeria which was established

in November 22, 1986. The objective of NDE was to promptly and effectively fight

unemployment by designing and implementing innovative programmes, which are directed

towards the provision of training opportunities through the guidance and management support

services to graduate farmers and small scale entrepreneurs. The objectives of NDE spanned

across the following programmes: (i) Youth employment and vocational skills development

programme, (ii) Special public works, (iii) Small scale industries and graduate employment

programme, and (iv) Agricultural development programme.

12

The aim of the agricultural programme is to generate employment for graduates, non-

graduates and school leavers in the agricultural sector, with emphasis on self employment in

agricultural production and marketing. The programme is monitored by a team of agricultural

professionals in the agricultural department of the directorate. Chinedum (2009) shares the same

views as in Akintoye (2008), inadequate funding and late release of funds from the federation

account among others have impaired the effectiveness of the NDE agricultural programmes.

Also the National Economic Empowerment and Development Strategy (NEEDS) was

introduced in March 2004, in order to confront the various macroeconomic imbalances, social

challenges and structural problems in the Nigerian Economy.

To achieve this goal, the major target of NEEDS is to engineer wealth creation,

employment generation and poverty reduction. Sequel to this, Adebayo and Ogunrinola (2006)

pointed out that for NEEDS to achieve its objectives there is need to design many integrated

programmes that can generate employment for women and youths to enhance growth and

development. In combating unemployment problem, this further point out the need to seek help

in the informal sector in order to drastically reduces unemployment. Hence, unemployment and

government policy, according to Walterskirchen (1999) has positive relationship with economic

growth. GDP and unemployment are both rising in the long run. Employment will only increase

when GDP is increasing faster than productivity, On the other hand, the greater the amount of

goods and services produced, the greater the labour required for production; because economic

growth and employment go hand in hand.

Labour Demand and Graduate Unemployment

The gap that exists between the demand for and supply of university graduates in the

Nigerian labour market is a serious issue in Nigeria, and this has manifested in the prevailing

high level of graduate unemployment. This has serious adverse social and economic

consequences on the Nigerian economy, some of which are declining quality of education,

inadequate funding, insufficient and outmoded learning materials, poorly trained staff, irrelevant

curricula and inadequate information on job vacancies for job seekers in the market. However,

there is need for the establishment of labour market information system, a legal framework for

labour market information management and improved funding of university education to

effectively harness the products of the Nigerian universities for sustainable economic

development (Bassey and Atan, 2012). The high rate of graduate unemployment persuaded

unemployed graduates to demand that government should provide them with job because the

supply for labour is greater than the demand for labour. The demand for labour is derived from

13

production and distribution activities in the goods and service sectors. As a result, its size and

shape are sensitive to what happens in the national economy (Dabalen et al., 2000). The demand

for labour in Nigerian economy has been poor and volatile at best. It is perhaps the most difficult

challenge getting or securing accurate information on labour demand while collecting labour

market information. Accordingly, Dabalen et al. (2000), the reason for this is that, hiring

decisions by firms are typically uncoordinated and in many cases unannounced.

Additional labour analysis problem in Nigeria has the fact that no systematic collection

of labour market information takes place. In many cases, information on labour demand are

obtained through secondary data such as manpower surveys, the labour market studies and direct

interviews with major employers. One way of examining the contribution of tertiary education

to the demand for labour is by taking a look at unemployment in Nigeria by the level of

education. From the Nigerian labour market, it is important to observe two (2) conditions. First,

school leavers with more than secondary education experience significantly lower

unemployment rate than those with secondary education or less. The difference is very sharp

when secondary school leavers are compared with post-secondary graduates. This suggests that

graduates of tertiary institutions stand better chances of obtaining employment compared to

those with only secondary education or less.

Fajana ( 2000) and Standing( 1983) opined that unemployment can be described as the

state of worklessness experienced by persons who are members of the labour force who

perceived themselves and are perceived by others as capable of working. Unemployed people

can be categorized into those who have never worked after graduation from the university and

those who have lost their jobs thereby seeking re-entry into labour market. However, most of the

previous studies on unemployment especially graduates unemployment in developing countries

(Falae, 1971; Bhalla,1973; Diejomaoh,1979; Bear and Herve, 1966; Bhagwati, 1973; Diejomaoh

and Orimolade, 1971) have tended to ignore the special case of the university graduates that are

first time job seekers.

According to William (1976) the meaning of work to paid employment is the result of

the development of capitalist productive relations. However, according to Fajana (2002) the

concept of work has partly shifted from productive effort itself to the predominant social

relationship. For instance, it is only in the sense of social relationship that a woman running a

house and bringing up children can be said not to be working (Hayes and Nutman, 1981; Iyoha,

1987).

Awogbenle and Iwuamadi (2010) noted that the framework of potential efforts and

strategies to boost employment and job creation for young people, entrepreneurship is

14

increasingly accepted as an important means and a valuable additional strategy to create jobs

and improve livelihoods and economic independence of young people. Regrettably, problems of

unemployment as experienced by the educated youths and even the uneducated but skilled

youths have become more pathetic in many developing economies, despite the neo-liberal

strategies in addressing the issue of enhancing human capital. The aim of this paper is therefore

to examine the constraints that impede young people in search of non-existing jobs and the

urgent need to orient people of these affected economies particularly Nigerians on imbibing self-

employment and entrepreneurship through vocational and entrepreneurial training programmes

as a short-term intervention mechanism.

2.2 Empirical Literature

A number of empirical literatures on the drivers of employment and labour demand exist

both in country and cross-countries. Shapiro and Stiglitz (1984) state that unemployment plays

the role of a macroeconomic ‘‘discipline device’’ in order to induce employees to intensify their

efforts in their job. They are based on the shirking models, where the firm, differentiating from

the unemployment salaries, increases the dismissal cost for the employee, thus inducing him to

intensify his effort.

Turunen (1998) presents disaggregated wage curve results by individual characteristics,

occupations, industries and regions in the United States, using a panel data set of young workers.

The results suggest that instead of a strong aggregate wage curve there are a number of different

wage curves over time and for different workers groups. The slope of the aggregate wage curve

varies over time, with the strongest wage curves appearing in the late 1980s. Wage curves exist

for most labour market groups: the wages of the least educated, those in relatively low-skill

occupations or service industries are most sensitive to changes in unemployment. Wages of

government workers and those in the mining industry increase with unemployment.

Brayton, Roberts and Williams (1999) tried to investigate if the Phillips curve is valid for

the 90's low inflation and low unemployment coexist. They use quarterly data from the period

1967 to 1998. They suggest that the Phillips curve cannot explain the low inflation during the

last years. The degree of capacity utilisation gives the best results for the last years, but this does

not appear in the total sample. They also provide evidence for a significant decrease in the

ΝΑΙRU after 1995. Finally, they present an (error-correction) adjustment model of prices in

conjunction with the long run trend unit labour cost.

Marcelino and Mizon (1999) examine the relationship between wages, prices,

productivity, inflation and unemployment. In their paper quarterly data have been used for the

15

period 1965(1) to 1993(1) for Italy, Poland and the UK. They apply a cointegrated VAR model

with regime shifts. They analyse the labour markets of these three countries and conclude that

there have been significant changes in the structures of the relationships between wages- prices

and unemployment - inflation for the period 1979/80. According to the qualitative results they

suggest that although there have been some important changes in the labour markets of these

three examined economies taking into account a greater degree of flexibility, there are no

common characteristics among them. Indeed, this is rationale if someone takes into

consideration the different starting points and policies followed in the three examined

economies.

Chletsos, Kollias, and Manolas (2000) investigated the relationship between

employment, growth rate, labour productivity and wages rate in the case of Greece for the period

1970-93. This period is divided into two sub-periods 1970-1980 and 1981- 1993. In the first

period they indicate that the employment level is positively related to the growth rate and wages

rates are negatively related to the labour productivity. The reverse result is observed in the

second period, which is characterized by the restructuring of the Greek economy.

Hsing (2001), based on the augmented Phillips curve and the autoregressive conditional

heteroscedasticity model, studied the impact of the union wage increases to non-union wages

and found that the growth of non-union wages is positively associated with the expected

inflation productivity growth and negatively correlated with the unemployment rate

Puhani (2002) estimated the changes in the Polish wage and unemployment structures

between the years 1994 and 1998 in order to identify the labour market characteristics associated

with increasing and decreasing relative demand as well as relative wage rigidities. The evidence

from his paper showed that the relative demand for workers with a low level of education has

decreased.

Broersma and Butter (2002) examined the influence of labour market flows on wage

formation and they applied the Johansen multivariate cointegration analysis for Netherlands.

The estimation results suggest the combination of the outflow from employment to

unemployment and the outflow of vacancies as indicators of labour market tightness, qualifying

for inclusion into the wages equation.

The need to ascertain the extent to which the theoretical views on graduate

unemployment in Nigeria are linked with some evidence leads us to the direction of examining

the empirical findings of different researchers on the study area. As recorded in Akintoye

(2008), graduate unemployment in Nigeria accounted for less than 1 percent of the unemployed,

in 1974, by 1984, the proportion rose to 4 percent for urban areas and 2.2 percent in the rural

16

areas. He noted further that, in 2005, Nigerian’s unemployment rate declined to 11.9 percent

from 14.8 in 2003. This decline was attributed to the various government efforts aimed at

addressing the problem through poverty alleviation programmes. This decline also pointed to an

increased number of people who got engaged in the informal sector activities. Unemployment

increased sharply from 14.9% in March 2008 to 19.7 in March 2009 (see Fig 1 below).

According to the NBS (2010), the overall unemployment rate was 19.7% but when

disaggregated by sector, gave 19.2% for urban and 19.8% for the rural respectively. Some states

in the country recorded higher the composite unemployment rate. Example, Bayelsa (38.4%),

Katsina (37.3%), Bauchi (37.2%), Akwa-Ibom (34.1%), Gombe (32.1%), Adamawa (29.4%),

Borno (27.7%), Kano (27.6%), Yobe (27.3%), Taraba (26.8%), Jigawa (26.5%), FCT (21.5%)

and Imo (20.8%) while some recorded lower than the composite unemployment rate. Example,

Plateau state (7.1%.) In 2005, Niger state recorded the lowest rate of 0.2 while Zamfara recorded

the highest rate of 51.1 when the rate of unemployment in the country was 11.9.

Figure 1

Source: Author’s Plot from NBS, (2010).

According, to Oyebade, (2003), Nigeria’s unemployment can be grouped into two

categories: (1) The older unemployed who lost their jobs through retrenchment, redundancy or

bankruptcy. (2) The younger unemployed, most of who have never tasted what it is to be

employed. According to statistics from the Manpower Board and the Federal Bureau of

Statistics, Nigeria has a youth population of 80 million, representing 60% of the total population

of the country. 64 million of them unemployed, while 1.6 million are under-employed. The 1990

- 2000 data on youth unemployment showed that the largest groups of the unemployed are the

17

secondary school graduates. There are also 40% unemployment rate among urban youth aged 20

- 24 and a 31% rate among those aged 15 - 19. Moreover, the educated unemployed tends to be

young males with few dependents. There are relatively few secondary school graduates and the

lower job expectations of primary- school graduates. There is no consistent trend of

unemployment rates in Nigeria.

Also, Sandra et al (2006) used a sample of 360 firms in Kano and its environs to examine

whether or not, in comparison to large firms, small firms are relatively better at creation of

employment opportunities. Their results were positive in that small firms were found to be

relatively better, and the conclusion they derived was that a policy that gives special preference

to small firms is justified.

An empirical investigation by Bakare (2011) showed that the rising nominal wages and

the accelerated growth of population which affected the supply side through a high and rapid

increase in labour force relative to the absorptive capacity of the economy appear to be the main

determinant of high unemployment in Nigeria. The econometric results suggested the need for

the government to embark on direct measures capable of creating jobs through industrialization

and mechanization of agriculture. Bakare also recommended that programmes of integrated rural

development and re-orientation of economic activity and social investments towards the rural

areas need to be embarked upon to create an appropriate rural -urban economic balance.

Biobele Richards (2007) in his study to identify the problems of recruitment in Nigerian

federal civil service and determine the extent of the utilization of job description and job

specification in the recruitment process, used stratified random sampling method to draw a

sample of 190 from five federal ministries, which was grouped into four major categories of

grade levels. The descriptive statistics in the study identified factors such as increasing pressures

for employment, utilization of informal sources of recruitment, long military era, federal

character principle, lack of independence of the service commission and delegation of

recruitment functions as the prevalent problems affecting recruitment in Nigerian federal civil

service. As results of these problems, job description and standard personnel requirements were

not adequately used in the recruitment process, especially at the lowest category of grade levels.

Based on the findings, it was recommended that the federal government should promulgate laws

that will protect employees in the private sector in order to reduce consistent pressures for

employment in the public sector.

Amupitan (2011) examined graduate unemployment and how the National Directorate of

Employment (NDE) has helped in curbing it in Kaduna state. Adopted statistical tables and

charts in the data analysis while the Average score method was used for the test of hypotheses

18

formulated. He discovered that inadequate awareness and poor funding of the activities of the

NDE in Kaduna State undermined its activity, and that skills acquisition is an effective tool in

reducing graduate unemployment. He recommended specific skills acquisition schemes as

elements that could empower unemployed graduates; that specific skills acquisition schemes

should be included in the curriculum of post-secondary schools. Organising symposia, seminars

and using the media were some of the recommendations made for improvement on the

awareness of the activities of the National Directorate of Employment.

Asoluka and Okezie (2011) examined the relationship between unemployment and

growth in Nigeria (1985-2009). One major findings of the study was that the economy grew by

55.5 percent between 1991 and 2006; and the population increased by 36.4 percent. All things

being equal, this should have resulted to a decrease in the rate of unemployment but rather,

unemployment increased by 74.8 percent. The study also found out that the average contribution

of the oil sector to the GDP between 1991 and 2006 was 30.5 percent while agriculture which

was the main source of gainful employment in the country contributed 36.7 percent just a

difference of 6.1 percent from that of oil that employed less than 10 percent of the labour force.

The study recommended that the agricultural sector as a medium of reducing unemployment in

Nigeria should be harnessed and advises that Government and all relevant stakeholders continue

in their quest towards reducing unemployment, as well as give their support in ensuring that the

agricultural sector is not downtrodden but embraced in this task.

Ekkehard Ernst (2011), made use of newly developed information on unemployment

dynamics, and adopted a matching model of the labour market to analyze the economic,

institutional and policy determinants of unemployment in- and outflows. He found standard

determinants to be significant with the expected sign. In particular adverse productivity shocks,

higher user costs of capital, stronger real wage growth, heavier tax burden, larger unionization

rate and more stringent employment protection legislation can be shown to depress

unemployment dynamics. Moreover, the impact of the degree of wage bargaining centralization

confirms the original Calmfors-Driffill insight also in the flow context. The paper also identifies

the impact of policy interventions through fiscal and labour market spending using a

macroeconomic, simultaneous equation set-up. The paper assess the relative contribution of

these policies in stimulating unemployment outflows and analyzes the effectiveness of different

policy instruments at different time horizons, stressing the importance of passive labour market

measures to stimulate unemployment outflows and to limit unemployment inflows.

Godwin and Johnson (2012) studied the existing gap between the demand for and supply

of labour in the Nigerian labour market. The study found out that serious distortions exist in the

19

market for university graduates in Nigeria, giving rise to unacceptable high level of graduate

unemployment. The phenomenon of rising graduate unemployment is bound to have serious

adverse social and economic consequences on the Nigerian economy. The study traces the

problem to the declining quality of education, resulting from inadequate funding, insufficient

and outmoded learning materials, poorly trained staff, irrelevant curricula and inadequate

information on job vacancies for job seekers in the market. They recommended the

establishment of labour market information system, a legal framework for labour market

information management and improved funding of university education to effectively harness

the products of the Nigerian universities for sustainable economic development.

Table 1: Summary of Literature Reviewed

Author / Date Location Nature of Data Method Research Findings

Shapiro & Stiglitz

I984

Greece Cross -Sectional OLS Unemployment occurs as a result of low salary income

Mertz et al 1995 Germany Cross-Sectional OLS Labour productivity directly affect unemployment rate

Gordon, 1997 United

Kingdom

Time-Series OLS Productivity growth increases unemployment rate and

vice-versa

Tunumen,1998 United

States

Panel OLS It indicated a strong aggregate wage curve in determining

unemployment.

Brayton et al,

1999

United

States

Quarterly Time

Series

OLS Adoption of an effective Philip curve measure

Marcelino and

mizon, 1999

Italy,

Poland

and Uk

Quarterly Time

Series

VAR Existence of significant changes in wage price-

unemployment structure

Chletsos et al

2000

Greece Time-Series ECM Employment level is positively related to growth rate but

wage rate is negatively related to labour productivity

Akintoye 2008 Nigeria Time-Series OLS Unemployment is caused by urban saturated job seekers

Dabalen et al 2000 Nigeria Survey data OLS Significant number of graduate employee are majorly the

public sector

Hsing 2001 Hong-

kong

Time-Series ARCH Growth of non-union wage rate negatively affects

productivity and unemployment.

Broersma and

Butter 2002

Netherland Time Series OLS Significance combination of outflow from employment to

unemployment to be included

Oyebade 2003 Nigeria Time Series OLS Human capital has a direct impact on youth employment

Ajibefun &

Daramola 2003

Nigeria Survey data OLS Evidence of a wide variation between unemployment and

economic growth.

Sandra et al 2006 Nigeria Cross Sectional OLS Good government labour policy significantly enhance

productivity

Biobele Richards

2007

Nigeria Time Series OLS Job description and standard personnel requirements were

not adequately used in the recruitment process

Bakare, 2011 Nigeria Time Series 2SLS government should embark on direct measures capable of

creating jobs through industrialization

Amupitan 2011 Nigeria Cross seasonal Descriptiv

e statistic

Suggested that skills acquisition schemes could empower

unemployed graduates

Asoluka & Okezie

2011

Nigeria Cross seasonal OLS Agricultural sector should be harnessed as a way of

reducing unemployment.

Godwin &

Johnson 2012

Nigeria Time Series OLS Sees declining quality of education and inadequate

information on job vacancies as the cause of

unemployment.

20

2.3 Conceptual Frame work

Concept of Unemployment

As noted by Bello (2003), from time immemorial, the subject of unemployment has

always been an issue of great concern to the economists, policy makers and economic managers

alike given the devastating effect of this phenomenon on individuals, the society and the

economy at large. The classical school of thought that provided the earliest thinking on

economic issues did not fail to give a central point of reflection on the undesirability of

unemployment. The Keynesian revolution of the 1930’s, which commandeered the explosive

attack on economic orthodoxy apparently, treated unemployment as a central issue of great

concern. Following the path of the predecessors, economists at all times and in all ages have

expressed various degrees of concern over the threat of the monster called unemployment.

Thus, the population of every economy is divided into two categories, the economically

active and the economically inactive. The economically active population (labour force) or

working population refers to the population that is willing and able to work, including those

actively engaged in the production of goods and services (employed) and those who are

unemployed. Whereas unemployed refers to people who are willing and capable to work but are

unable to find suitable paid employment. The next category to the economically inactive

population refers to people who are neither working nor looking for jobs. There seems to be a

consensus on the definition of unemployment.

The International Labour Organization (ILO) defined unemployment as the people who

are out of work, want a job, have actively sought for work in the previous four weeks and are

available to start work within the next fortnight; or out of work and have accepted a job that they

are waiting to start in the next fortnight (ILO, 2005). The unemployment definition can differ

from one country to another according to how it is measured and implemented. Unemployment

can also be described according to region, sex, educational level, age, and economic conditions.

Here our focus is on the classification of unemployment according to educational level (graduate

unemployment).

However, graduate unemployment is unemployment among people with academic

degrees. It is a situation where tertiary institution graduates do not get jobs after going through

the academic ladder successfully. One of the major causes of this is the mismatch between the

aspirations of graduates and employment opportunities available to them. According to Juan

Ramón (2011), graduate unemployment is an evidence of serious shortcomings in educational

system and labour market in developing economy, which explains the country's relative high

rate of youth unemployment and the imbalance between job supply and demand at the different

21

educational levels attained, which also complicates graduates access to the labour market and

has a negative impact on their professional career.

Confirming this, three out of ten graduates of higher education cannot find work, the

reason being that high education does not increase the chance of finding job; many graduates of

higher education who find work are not usually gainfully employed. They are forced to accept

marginal jobs that do not use their qualification, for instance, in sales, agriculture and manual

labour. Therefore, graduate unemployment requires coordinated action between education and

the labour market.

Concept of Labour Productivity

Productivity measures the relationship between the quantity and quality of goods and

services produced and the quantity of resources needed to produce them (i.e. factor inputs such

as labour, capital and technology) (Simbeye, 1992; Okojie 1995; Roberts and Tybout, 1997).

Mali (1978) defines it thus: "The measure of how resources are being brought together in

organizations and utilized for accomplishing a set of results. It is reaching the highest level of

performance with the least expenditure of resources". Productivity is viewed as the instrument

for continuous progress, and of constant improvement of activities. It is often seen as output per

unit of input. Hence, higher productivity connotes achieving the same volume of output with

less factor inputs or more volume of output with the same amount of factor inputs. Thus,

increased productivity could result from the reduction in the use of resources, reduction in cost,

use of better methods or improvement in factor capabilities, particularly labour. Two variants of

productivity measurements have been cited in the literature: total factor productivity (TFP),

otherwise known as multifactor productivity, and partial productivity.

Roberts and Tybout (1997) and Tybout (1992), assuming a neo-classical production

function at the sectoral or industry level, define total factor output to be a concave function of

the vector of inputs and time (a proxy for shift in technological innovation). To them, the

elasticity of output with respect to time is the total factor productivity. In a more general sense,

TFP = Total Output ie Weighted Average of all inputs. Critical among these factor inputs are

labour, capital, raw materials and purchase of spare parts, and other miscellaneous goods and

services that serve as inputs in the production process. In a more practical sense, these factor

inputs are reduced to the weighted average of labour and capital (Okojie, 1995; Roberts and

Tybout, 1997). The second variant, partial productivity (PP) is defined as:

PP = Total Output. That is, partial productivity is equated to total output.

22

Partial Input

The partial input could either be labour or capital. This can be measured at the national level,

sectoral level, industry or factory level. Existing studies on productivity measurement show a

predilection for productivity per labour input. Several reasons have been put forward for the

choice of labour as against other factors of production. According to Ilyin and Motyler (1986)

labour is the "means and end of production". Labour is the only factor that creates value,

influences its prices and those of other factors and sets the general level of productivity. Second,

it is the most easily quantified factor of production (Okpechi, 1991). Finally, given the low

technological base of developing countries' economies, the quest for improved managerial

capability and effectiveness should give the human factor appropriate recognition and attention.

While labour productivity seems to be the most convenient to use, it is however important to

note that this approach has an important limitation. It treats labour as being homogenous instead

of differentiating it according to age, sex, education, application of skills, aptitude, among

others. Nevertheless, this study applies productivity per worker as opposed to per capital or total

factor productivity.

Government Labour/Employment Policy

The government labour policy is the policies and programmes put in place by the

government over the years to reduce unemployment, especially graduate unemployment. Here in

this study we look at the policy design in National Directorate of Employment (NDE) that

targets graduates. Hence, the effective government labour policies instruments targeted at

graduate unemployment in National Directorate of Employment (NDE) scheme include the start

your Own Business (SYOB) under the small Scale Enterprise programme (SSE), the

sensitization of National Youth Service Corps (NYSC) on Entrepreneurship development, the

Graduate Attachment Programme (GAP) and the Solar Energy Training Scheme (SETS) which

was recently introduced. However, the study will estimate the government labour policy with the

total number of persons that benefited from NDE graduate programmes annually (TNA).

2.4 Literature Gap and Value Addition

There is no doubt that great deal of literature exists on unemployment and its determinants both

in Nigeria and across the world . The recent studies of unemployment and its drivers carried out in

Nigeria (see Sandra et al 2006, Biobele Richards 2007, Bakare, 2011, Asoluka & Okezie 2011

and Godwin & Johnson 2012) focus more on the causes and consequences of unemployment in

Nigeria. For instance, Asoluka & Okezie (2011) presented time series data of the rate of

unemployment and gross domestic product in Nigeria, for the period of 1985 to 2009 to show

23

the percentage changes in both unemployment and GDP without applying any quantitative

measures to ascertain the significant nature of these figures. They only analyzed the

consequences of the growing trend of unemployment on economic growth with some

recommendation. Godwin & Johnson (2012) that studied related subject to this study again

showed all the trend of unemployment and school enrolment rate in Nigeria. Most of these

studies analyzed unemployment rate based on the figure, no further empirical examination was

carried out and those that did so, used OLS technique to examine the significant impact of

unemployment rate on economic growth in Nigeria.

Therefore, this study will bridge this huge research vacuum by building on the

theoretical framework suggested by ‘Keynes and Modern theory of labour market’ to establish a

labour market productivity model of the nexus between government labour policy, graduate

unemployment and labour productivity in Nigeria. As a matter of fact, a study of intra-

relationship as in this study will pose serious, endogeneity problems, just as found in some other

studies carried out in other countries; hence this study will adopt various techniques to

checkmate this endogeneity and other estimation biasness that may result in a study of this kind,

in order to ascertain the true intra-connections in this phenomena.

24

CHAPTER THREE

METHODOLOGY

3.1 Theoretical Framework

The framework for analyzing the relationship between government policy, labour

productivity and unemployment is a complex issue. Starting with the classical economists view,

the framework for employment and output is a one-way relationship that goes from the input of

labour to output. In the classical model's steady state (conditions where the growth rate of capital

stock and output are equal) the rate of growth of labour force and technical progress ultimately

determine the growth rate of output.

The premise of the classical model therefore is that the growth rate of employment is

exogenous to the growth rate of output. In this framework, the supply of labour is positively

related to the level of real wage, while the demand exhibits a negative relationship with real

wage, but a positive relationship with productivity (Fashola, 1983; Todaro, 1990). As pointed

out by these authors, if there is some `involuntary' unemployment at or below the current real

wage, the real wage would fail to induce employers to take more labour until all involuntary

unemployment is eliminated. However, if increases in labour productivity translate to increased

wages and such increases induce the substitution of capital for labour the effect on

unemployment will be positive (Fajana, 1983; Krugman, 1994).

The policy implications of this have been viewed as misleading particularly, to

developing countries such as Nigeria (Todaro, 1990; Hussain and Nadol, 1997). Also as argued

by Hussain and Nadol (1997:3), the policy implication of the neoclassical approach to primary

commodities-producing countries like Nigeria is that, given the existence of says Law, whatever

that was produced is automatically sold irrespective of the characteristics of the goods produced

and the demand for them. Recent developments in the world market for primary commodities

have proved this to be wrong.

In contrast, Keynesian theory explains the determination of output or productivity and

aggregate demand. This approach sees demand for labour as a derived demand. Productivity

growth (a la Verdoorn's Law), should increase the demand for labour thereby reducing

unemployment. The Keynesian framework, as examined by Thirlwall (1979), Grill and Zanalda

(1995) and Hussain and Nadol (1997), postulates that increase in employment, capital stock and

technological changes are largely endogenous. Thus, the growth of employment is demand

determined and that the fundamental determinants of long term growth of output also influence

the growth of employment.

25

Contrary to the strong belief of the neoclassical that equilibrium wage rate, price, interest

rate and real cash balances guarantee the quality of national output and full-employment level,

the Keynesians strongly believe in the efficacy of aggregate demand. Rather than the workings

of the real wage, price, interest rate and real cash balances, what could guarantee the attainments

of full employment is additional government spending. The Keynesian prescription for reducing

unemployment is increase in aggregate total demand through direct increases in government

spending or policies that encourage more private investment. As argued by the Keynesians, as

long as there is unemployment and excess capacity in the economy, the supply of goods and

services will respond automatically to this higher demand. A new equilibrium will always be

established with higher income and lower level of unemployment.

However, the extension of the Keynesian model dominated development theorizing in

the 1950s and beyond. Such extensions could be found in Okun's Law and the Harrod-Domar

model. For instance, Arthur Okun developed the relationship between the actual and potential

output and between the actual and benchmark unemployment in an equation called the "Okun's

Law" (Dernburg, 1985). Thus:

( )*

**

Q QU U

Q

−= −

Where Q* is potential output, Q is actual output, U is the unemployment rate, U* is the

benchmark unemployment rate, and " is Okun's coefficient (Akinboyo, 1987). The implication

of Okun's coefficient is that a 1 percentage rise in unemployment causes the economy to lose

"percent of its output. Okun's Law clearly gives a direct relationship between output and

unemployment and indirectly between productivity and unemployment (a la Verdoorn Law).

In a similar vein, the neo-Keynesians, in their efforts to provide reasons why

employment growth lags behind growth of industrial output, came out with a typical variant of

the Harrod-Domar unemployment equation

[ ][ ]

( ) [ ][ ]

** *Y

Y NN

YY NN

− =

Thus, the essence of this equation is that the rate of output growth (Y) minus the rate of

growth in labour productivity (Y/N) approximately equals the rate of growth of employment

(N). The implication is that the gap between growth rate of output and the growth of labour

productivity accounts for the rate of labour absorption. As had been argued hypothetically by

Todaro (1990), if output is growing by 8 percent per year while employment is expanding by

only 3 percent, the difference is due to the rise in labour productivity, and vice versa. By

implication, rapid economic growth could generate lagging employment creation. This tends to

26

support Essenberg's (1996) argument that if the reduction in labour demand resulting from

productivity increases is more than compensated by overall increases in output, then both

productivity and employment can increase together. This is particularly so when higher

productivity leads to increased profit and higher rate of investment, which in turn results in

higher rate of growth.

In conclusion, the neo-classical approach posits that the rate of growth of employment

(unemployment) is exogenous to the rate of growth of output (productivity). In contrast, the

Keynesian argument is premised on the fact that it is the strength of demand that determines the

amount of resources utilized. As such, employment is demand determined and the rate of output

growth is itself an important determinant of the rate of growth of employment. Thus, output,

productivity and employment are determined endogenously.

3.2 The Model

The analytical ideas employed for this study is built on the combination of Keynesian

framework and modern labour market theory which provides at least three competing models to

explain equilibrium unemployment (Gordon, 1997). These models are; (i) the union models,

which view wages as being determined by a bargain between unions and firms; (ii) the search

models, where the wage is determined by a bargain between individual workers and firms; and

(iii) the efficiency wage models, where firms set wages above the competitive level to increase

workers efforts. Though some of the reasoning behind these models are different, there are

noticeable similarity among them: the first is that the equilibrium rate of unemployment is

determined by institutional settings, such as the size and power of unions, the bargaining system,

and by the generosity of the unemployment insurance system. Secondly, the equilibrium rate of

unemployment is independent of production and productivity growth.

On its own, Keynesian prescription for reducing unemployment is to increase the

aggregate total demand through direct increases in government spending policy that encourage

more private investment. As argued by the Keynesians, as long as there is unemployment and

excess capacity in the economy, the supply of goods and services will respond automatically to

this higher demand. A new equilibrium will always be established with higher income and lower

level of unemployment. Thus, while labour market institutions can potentially explain cross-

country differences today, they do not appear able to explain the general evolution of

unemployment over time” Nickell et al. (2005). Despite conventional wisdom, high

unemployment does not appear to be primarily the result of things like overgenerous benefits,

trade union power, taxes or wage inflexibilities. Nickell et al. (2005) provide, the most thorough

27

(econometric) and eclectic relationship of unemployment and government policy when he

pointed that rise in the unemployment rate, one-half can be attributed to institutional changes in

the labour market, such as the unemployment compensation payment system, the system of

wage determination, employment protection, labour taxes, and barriers to labour mobility, and

the other half to demand deficiency. Hence this idea of theoretical model will be invoked in this

study to explain the observed government policy, graduates unemployment and labour market

behavior in Nigeria.

Therefore, following the Keynesian framework, the three labour union models and

numerous literature reviewed, government policy can impact on graduate unemployment

through the following variables, government spending, output gap, money supply, interest rate

real wage rate, marginal product of labour, external debt management, education and training,

price level and lag of unemployment. Not all these variables will be included in this study due to

unavailability of some of the data. The theoretical relationship underpinning these variables will

be explained below after the specification.

3.3 Model Specification

Model 1: The relationship between Government Labour Policy, Graduate Unemployment

and Labour productivity

To analyze the significant relationship among government labour policy, graduate

unemployment and labour productivity, as in Objective 1 of this study, we made use of the

system of labour productivity model. The channels through which labour policy can impact on

graduate unemployment and labour productivity in Nigeria are presented in the following

equations:

Ug = f(TNA, INV, LP, GXP, Yg, GS) ------------------------------------------------- (1)

LP = f(TNA, INV, Ug, GXP, Yg, M2,IR) ---------------------------------------------- (2)

where

Ug = graduate unemployment rate.

TNA = government labour policy captured by the total number of persons that

benefited from NDE graduate programmes annually.

INV = investment rate proxied by gross fixed capital formation.

LP = labour productivity, proxied by GDP per person employed.

GXP = total government expenditure

Yg = output gap, (potential output ‘p’ minus actual output ‘y’).

GSE = gross secondary school enrolment

28

IR = real interest rate

M2 = broad money supply

The choice of the above listed indicators is in line with government labour policy

response variables implicated in the literature. According to Tejvan Pettinger (2011), there are

two main strategies for reducing unemployment - demand side policies to reduce demand-

deficient unemployment (unemployment caused by recession) and supply side policies to reduce

structural unemployment (the natural rate of unemployment).

However, these variables are justified below; let us first transform equations 1 & 2 to an

econometric format in the following equations.

1 2 3 4 5

6 1

t t t t t

t t

U g T N A I N V L P G X P Y g

G S E

α φ φ φ φ φ

φ ε

= + + + + +

+ + --------- 3

1 2 3 4 5

6 7 22

t t t t t

t t t

L P I N V I N V U g G X P Y g

M I R

α φ φ φ φ φ

φ φ ε

= + + + + +

+ + + ----------4

where

φI = the slope coefficients of the vector of explanatory variables defining the interaction

among government policy, graduate unemployment and labour productivity. However, a

simultaneous equation such as in 3 & 4 is bound to suffer problem of endogeneity, but this study

adopted Hausman Endogeneity method to take care of this during the estimation.

Theoretical Justification of the variables in Model 1

Government Expenditure: Fiscal policy can decrease unemployment by helping to

increase aggregate demand and the rate of economic growth. The government will need to

pursue expansionary fiscal policy; this involves cutting taxes and increasing government

spending. Lower taxes increase disposable income and therefore help to increase consumption,

leading to higher aggregate demand (AD).

Output Gap: With an increase in AD, there will be an increase in real GDP (as long as

there is spare capacity in the economy). If firms produce more, there will be an increase in

demand for workers and therefore lower demand-deficient unemployment. Also, with higher

aggregate demand and strong economic growth, fewer firms will go bankrupt meaning fewer job

losses. Keynes was a strong advocate of expansionary fiscal policy during a prolonged

recession. He argues that in a recession, resources (both capital and labour) are idle, therefore

the government should intervene and create additional demand to reduce unemployment.

Interest rate: Monetary policy would involve cutting interest rates. Lower rates

decrease the cost of borrowing and encourage people to spend and invest. This increases AD and

29

should also help to increase GDP and reduce demand deficient unemployment. Also lower

interest rates will reduce exchange rate and make exports more competitive. In some cases,

lower interest rates may be ineffective in boosting demand. In this case, Central Banks may

resort to Quantitative easing. This is an attempt to increase money supply and boost aggregate

demand.

On the other hand, the supply side policies deal with more micro-economic issues. They

do not aim to boost overall aggregate demand, but seek to overcome imperfections in the labour

market and reduce unemployment caused by supply side factors. Supply side unemployment

includes – Frictional, Structural and Classical (real wage).

Education and Training captured in this study by gross secondary school enrolment:

The aim is to give the long term unemployed new skills which enable them to find jobs in

developing industries, e.g. retrain unemployed steel workers to have basic I.T. skills which helps

them find work in service sector. – However, despite providing education and training schemes,

the unemployed may be unable or unwilling to learn new skills. At best it will take several years

to reduce unemployment.

Real wage rate: If unions are able to bargain for wages above the market clearing level,

they will cause real wage unemployment. In this case reducing influence of trades unions (or

reducing Minimum wages) will help solve this real wage unemployment. To the neo-classicists,

the supply side of the labour market is infinitely elastic at what they called “natural price” of

labour.

Improve Labour Market Flexibility: It is argued that a higher structural rate of

unemployment is due to restrictiveness of a labour market which discourages firms from

employing workers in the first place. For example, abolishing maximum working weeks and

making it easier to hire and fire workers may encourage more job creation. However, increased

labour market flexibility could cause a rise in temporary employment and greater job insecurity.

Model 2: Graduate unemployment and Total factor productivity

In order to estimate objective two of this study, this examines the impact of graduate

unemployment on Total Factor Productivity (TFP). This was measured in this study as the ratio

of output and weighted input factors.

t

i

YTFP

Weighted X

=

---------------------------------------------------------------- (5)

Where

Y = total output in the economy

Xi = aggregate input factors (weighted).

30

For simplicity, let ( X ′ ) be expressed as a vector of all the input factors, while (a) defines

a vector of factor input prices. TFP is then calculated as the ratio of output and weighted input

factors (weighted with corresponding input factors) as:

aY

TFPX

= ′

---------------------------------------------------------------------------- (6)

where

Superscript ‘a’ = a vector of factor input prices

Expressing equation (6) in an implicit and log form, the following equation is

formulated:

1 2 3 4

5 6 2

t t l t t t

t t t

T F P w M P U g G X

Y g I R

α φ φ φ φ

φ φ ε

= + + + +

+ + + --------------------- (7)

From equation (7) we can derive the concept of Total Factor Productivity Growth (TFPg)

by taking the log difference of TFP as:

( ) 100TFPg TFP= ∆ ∗ ------------------------------------------------------------------------------ (8)

Model 3: Causal relationship between Graduate unemployment and government labour

policy variables

We also investigated the causal relationships existing between government labour policy,

graduate unemployment and labour productivity; we therefore specify our model by adopting

the Granger Causality Tests. This model focuses on the relationship between two time series.

Granger (1969) defined the causality in terms of predictability, based on the fact that the effect

cannot come before the cause. The model is specified thus as;

1

1 1 1

n n m

t i t i j t j k t i t

i i i

U g X U g Xα β λ µ− − −

= = =

= + + +∑ ∑ ∑ ---------------- (9)

2

1 1 1

n n m

t i t i j t j k t i t

i i i

L P U g X Xλ β λ µ− − −= = =

= + + +∑ ∑ ∑ --------------- (10)

Where it is assumed that the disturbances µ1t, µ2t and µ3t are uncorrelated.

Equations 9 and 10 were used to capture objective three of the study which examines the causal

relationship existing between government labour policy, graduate unemployment and labour

productivity.

3.4 Estimation Procedure

This study employed the ordinary least square technique of estimation, co-integration

and error correction techniques in estimating the models of the study. Most economic time series

31

data are non-stationary (Gujarati 2004), hence correct and appropriate specifications of the time

series model requires that we determine whether such time series are stationary or not. In this

case, all variables of the study will be subjected into unit root test of stationarity to avoid a

spurious result (Granger 1969). Here, coefficient of determination (R2) together with the value

of Durbin-Watson will be determined. In the small way, the standard significance test

(traditionally measured by t-test), Overall significance of the model measured by the F-test and

the coefficient of the parameter of the variables used in determinining the conformity of apriori

expectation of the economic theories will be analyzed in the study.

Unit Root Tests

Many macroeconomic time series contain unit roots dominated by stochastic trends as

developed by Nelson and Plosser (1982). Unit roots are significant in examining the stationarity

of a time series because a non-stationary regressor invalidates many empirical results. The

presence of a stochastic trend is determined by testing the presence of unit roots in time series

data. In this study, unit root test was carried out using Augmented Dickey (1979) and Fuller

(1981). Augmented Dickey Fuller test (ADF) (1979) refers to the t-statistic of δ2 coefficient on

the following regression:

∆Xt = δ0 + δ1t + δ2Xt-1 + Σ αi∆Χt-i + µt…………………………………….... (11)

The ADF regression tests for the existence of unit root of Χt, namely in the logarithm of

all model variables at time t. The variable ∆Χt-i expresses the first differences with k lags and

final Ut is the variable that adjusts the errors of autocorrelation. The coefficients δ0, δ1, δ2, and αi

are being estimated. The null and the alternative hypothesis for the existence of unit root in

variable Xt is:

Ηο : δ2 = 0 (unit root) Ηε : δ2 < 0

Τhis study follows the suggestion of Engle and Yoo (1987) using the Akaike information

criterion (AIC) (1974), to determine the optimal specification of Equation (11). The appropriate

order of the model is determined by computing Equation (11) over a selected nexus of values of

the number of lags k and has been found that value of k at which the AIC attains its minimum.

The distribution of the ADF statistic is non-standard and the critical values tabulated by

Mackinnon (1991) are used.

The results of ADF test are compared with critical values at 1%, 5% and 10% (usually

the 5% conventional approach), which we have obtained from Mackinnon (1991) tables. The

results of absolute value of ADF statistic for the examined time series should exceed the critical

values, before the null hypothesis of a unit root is rejected. Taking first differences renders each

32

series stationary, with the ADF statistics in all cases being less than the critical value at the 1%,

5% and 10% level of significance.

Co-integration Test

After establishing the existence of unit root and their respective order of integration

identified then it will be necessary to evaluate the variables that have the same order of

integration or if their linear combination is integrated at level form. If the variables are

integrated of the same order then the presence of co-integration is established as well as their

linear combination (Enders, 1995). Equation below represents co-integration equation.

∆Yt=α1Yt-1+X1ᵠ + β1∆Yt-1 + β2∆Yt-1 + βp∆Yt-p +εi.......................................... (12)

where X is the optimal exogenous regressor which may consist of constant or a constant and

trend α, ᵠ and β are parameters to be estimated.

If there is the presence of co-integration in the model, equations 3&4 will be transformed into an

error correction model (ECM).

1 2 3 4

5 6 1 1

t t l t t t

t t t t

U g w M P L P G X

Y g G S E C M

α φ φ φ φ

φ φ ν−

∆ = + ∆ + ∆ + ∆ + ∆

+ ∆ + ∆ + ∆ + ------------ (13)

1 2 3 4

5 6 1 2

t t l t t t

t t t t

L P w M P U g G X

Y g I R E C M

α φ φ φ φ

φ φ ν−

∆ = + ∆ + ∆ + ∆ + ∆

+ ∆ + ∆ + + ------------ (14)

Equations 13& 14 stated above are ECM formulation which indicates the speed of adjustment of

variables that were in a disequilibrium state into equilibrium.

The error correction parameter indicated above shows how disequilibrium in the

explained variables are corrected into equilibrium in each periods but if it is statistically

significant, it implies that the disequilibrium is corrected at different periods but if otherwise

stated, then it is corrected at the same period. Where ∆=difference operator.

The ECM analysis simply provides the short-run dynamic adjustments of the

explanatory variables towards the Long-run equilibrium.

The significance levels of the F – statistics for the lagged variables and the t – statistics for

the coefficient δ of ECt-1 are used to test for Granger causality i.e. there are two channels of

causality (Granger 1988). These are called channel 1 and channel 2. If lagged values of a

variable (except the lagged value of the dependent variable) on the right hand side in equation 3

are jointly significant then this is channel 1. On the other hand, if the lagged value of the error

correction term is significant, then this is channel 2. The numbers in parentheses are the lag

lengths determined by using the Akaike criterion.

33

3.5 Justification of the Model

The analytical framework of the model is such that it incorporates the major features of an

econometric analysis in a systematic manner. For instance, in the model specifications,

provisions were made to ensure numerical accuracy and the stationarity of the variables used in

the study. Also, provisions were made for the elimination of co-integration by the application of

the error correction model (ECM) which helps in capturing the long-run behavioral pattern of

the economic variables under study.

3.6 Sources of Data

The data that will be used in the analysis are annual time series data, covering the period

1987 – 2013. It is a secondary data sourced from Central Bank of Nigeria annual statistical

bulletin/reports (various years) and National Bureau of Statistics annual statistical reports. If

these variables share a common stochastic trend and their first differences are stationary, then

they can be co integrated. E-view was the econometric software used in the analysis.

34

CHAPTER FOUR

PRESENTATION OF RESULTS

This section presents the results of the specified models. It starts with the results of data

generating process (trend analysis of the core variables, the unit root or stationarity test and the

co-integration statistics of the data series), the results of the main models and evaluation of the

research hypotheses of the study.

4.1 Analysis of the Data Generating Process

We start in this section with the presentation and analyses of the descriptive statistic,

trend analysis of the core variables, stationarity or unit root test and the co-integration test of the

data series. This will enable us to carry out some predetermine operations, if required, on the

variables, so as to have a more précised analysis and interpretation.

4.1.1 Trend Analysis of the Core Variables

The trend analysis of the core variables was conducted first to establish the trending

patterns of these variables, hence we present below the trending of graduate unemployment rate

(UG), government labour policy captured by the total number of persons that benefited from

NDE graduate programmes annually (TNA), and labour productivity (LP), proxied by GDP per

person employed and total factor productivity (TFP).

The results show that while the total number of persons that benefited from NDE

graduate programmes annually and labour productivity proxy, in this study, by the growth in

GDP per person employed maintained steady and near horizontal movement, graduate

unemployment rate has been on the increase over the years given other considerations (see Tab

4.1). This implies that these variables are trending, showing a typical sign of unit root. Also

observed in the trend analysis is that total factor productivity exhibits upwards trending as

graduate unemployment.

35

Figure 2: Trend Analysis of the core variable

0

20

40

60

80

100

1990 1995 2000 2005 2010

TNA UG LP

5.000E+10

5.000E+11

5.000E+12

5.000E+13

1990 1995 2000 2005 2010

TFP

Source: Author’s calculation

4.1.3 Unit Root and Co-integration Analysis

In an attempt to normalize our data from unit root problem, we test for the presence of

unit root in the variables and obtain their integrating order. If the dependent variable associated

to each model is found to be integrated of the same order with the explanatory variables

36

included in that model, the co-integration test will be carried out to ascertain their long-run

relationships.

Testing for Stationarity

This is necessary in order to ensure that the parameters are estimated using stationary

time series data. Also this study seeks to avert the occurrence of spurious results. To do this,

both the Augmented Dikky-Fuller (ADF) and Phillips-Perron (PP) unit root tests are carried out.

The essence of these tests is to verify the null hypothesis of unit root or non-stationary stochastic

process. To reject this, the ADF statistics must be more negative than the critical values and

significant. On the other hand, the Phillips-Perron test differs because it provides a more robust

test for serial correlation and time dependent heteroskedasticities of the stochastic process. The

results of ADF and PP test statistics for the levels and first differences of the annual time series

data for the period under investigation are presented in Table 2.

Table 2: ADF and PP Unit Root Test Results for Individual Variables

Source: Computed from the Output of E-views software.

The asterisk (*) denotes rejection of the unit root hypothesis at the 5%, while the asterisk

(**) denotes rejection of the unit root hypothesis at the 1% level respectively. The ADF statistics

were generated with a test for a random walk against stationary AR (1) with drift and trend at

VARIABLE ADF PP

LEVEL 1ST

Diff, Prob. Level 1st Diff. Prob.

GS -1.282402 -6.060380** 0.0000 -1.282402 -6.222578** 0.0000

GXP 2.510561 -4.245044** 0.0031 3.414083 -4.274490** 0.0029

INV 3.880293 3.512078 1.0000 5.075342 -3.412299** 0.0206

IR -4.524767** -6.217123** 0.0000 -4.523883** -18.09608** 0.0001

LP -9.186530** -2.049518* 0.2653 -4.846324** -1.705559 0.4159

M2 -0.331316 1.146277 0.9959 3.530804 -1.351276 0.5884

TFP 4.294228 0.924351 0.9940 4.672281 -3.305201** 0.0260

TNA -5.602038** -5.203217** 0.0003 -12.81987 -5.808393** 0.0001

Ug 1.020143 -5.222694** 0.0003 0.113228 -8.584703** 0.0000

Yg -2.016543* -5.176852** 0.0003 -2.217760* -5.171846** 0.0003

where,

Ug = graduate unemployment rate.

TNA = government labour policy captured by the total number of persons that benefited

from NDE graduate programmes annually.

INV = investment rate proxy by gross fixed capital formation.

LP = labour productivity, proxy by GDP per person employed.

GXP = total government expenditure

Yg = output gap, (potential output ‘p’ minus actual output ‘y’).

GS = gross secondary school enrolment

IR = real interest rate

M2 = broad money supply

Note: ** indicates significance at 5% and 1% level.

37

the maximum lag length of 9. While the PP test uses the automatic bandwidth selection

technique of Newey-West.

From the result in table 4.1, the ADF indicated that the dependent variables for the three

objectives of this study are integrated of order one (∆ = 1) along side with some of the

explanatory variables. The real interest rate, labour productivity, the government labour policy

captured by the total number of persons that benefited from NDE graduate programmes annually

and output gap show stationary at level (∆ = 0), in ADF as well as PP results. This shows a

prerequisite for the presence of long-run linear combination among them, and to avoid mistake

of analysis of a long-run relationship in short-run analysis, we conduct co-integration test for the

variables.

Results from Co-Integration Test

Given the unit root properties of the variables, we proceed to implement the Engle-Granger co-

integration procedure. The explanatory variables that have the same order of integration (∆ = 1)

with the dependent variable; we estimate their linear combination at their level form without the

intercept term and obtain their residual which is then subjected to co-integration test as shown

in Tab 3.below.

Table 3: Co-Integration Tests

Null Hypothesis: The Residual has a unit root

Model Critical Value T – Statistic Prob.

1% 5% 10%

Model 1 -3.737853 -2.991878 -2.635542 -4.538538 0.0016

Model 2 -3.737853 -2.991878 -2.635542 -4.035495 0.0051

*MacKinnon (1996) one-sided p-values. Source: EView output

From the table, since the t–statistic associated with two models (-4.38538 and -4.035495) are

respectively less than their1%, 5% and 10% critical values as shown in table 3 above. This

means that the residual is not stationary and hence there is long-run linear relationship or co-

integration among the variables. Consequently, the study adopts the Error Correction Model

which was specified in Chapter Three as co-integration was noted among the variables.

38

4.2 Presentation and Analysis of ECM Models

Based on the underlying properties of the models specified in this study, such as; linear

in parameters, random sampling of the exogenous variables, homoskedasticity (equal variance),

and finally, the no-serial correlation, we analysed the validity and endogenous test to avoid

spurious and biased results. We therefore, start by implementing the endogeneity properties

before presentation of the main models.

4.2.1 Endogeneity Test

This study adopted Hausman (1978) method of endogeneity test. He suggested the direct

comparing of the OLS and 2SLS estimate and determining whether the differences are

statistically significant. If any appreciable differences are found between the two methods, (OLS

and 2SLS), we conclude that the endogenous explanatory variable(s) must be endogenous.

Hypothesis

H0: δi = 0 (no problem of Edogeneity)

Against

H1: δi ≠ 0 (there is a serious Edogeneity problem)

Decision Rule: Reject H0 if tcal > ttab, and accept otherwise at 5% level of significance.

The endogeneity results for the two models are presented as:

Table 4. Endogeneity Result for Graduate Unemployment (UG) Model

Variable Coefficient Std Error t - value t - prob

C 190.7218 83.04028 2.296738 0.0332

TNA 2.591692 0.917797 2.823817 0.0108

INV -9.02E-12 3.86E-12 -2.338469 0.0304

LP -5.311456 2.222258 -2.390117 0.0274

GXP 2.74E-05 4.45E-06 6.162319 0.0000

YG -4.41E-06 1.42E-06 -3.105822 0.0058

GS -0.083457 0.407366 -0.204870 0.8399

Structured estimate

C 70.60203 55.46384 1.272938 0.2192

TNA 1.590465 0.593273 2.680831 0.0153

INV -1.08E-11 2.40E-12 -4.486972 0.0003

LP -2.408248 1.464044 -1.644929 0.1173

GXP 2.82E-05 2.75E-06 10.24496 0.0000

YG -4.76E-06 8.79E-07 -5.410996 0.0000

GS 0.280486 0.259429 1.081168 0.2939

RESID01 1.033048 1.24674 1.650245 0.1235

F – Stat. = 169.5667; P – value = 0.0000

39

Table 4.1: Endogeneity Result for Labour Productivity Model

Variable Coefficient Std Error t - value t - prob

TNA 1.158564 0.078943 14.67595 0.0000

INV -5.20E-13 4.79E-12 -0.108512 0.9147

UG -0.054811 0.188153 -0.291310 0.7740

GXP -9.96E-07 7.76E-06 -0.128418 0.8992

YG 9.96E-10 1.36E-06 0.000733 0.9994

M2 5.82E-13 2.09E-12 0.278870 0.7834

IR -0.018228 0.064248 -0.283707 0.7797

Structured estimate

C 2.957904 0.014095 209.8596 0.0000

TNA 0.022042 0.000291 75.85171 0.0000

INV -9.95E-15 9.28E-16 -10.72223 0.0000

UG -0.000988 3.59E-05 -27.47677 0.0000

GXP -1.93E-08 1.51E-09 -12.78671 0.0000

YG 9.24E-11 2.55E-10 0.362267 0.7216

M2 1.10E-14 4.37E-16 25.18405 0.0000

IR -0.000351 1.72E-05 -20.42850 0.0000

RESID02 0.019022 0.020389 1.84280 0.0923

F – Stat. = 7695.298; P – value = 0.0000

Source: EViews Output

The above estimated results as presented in Table 4 for the Graduate Unemployment

(UG) Model and Table 4.1 for the Labour Productivity (LP) model shows the estimates of

reduced form equation and their structured form estimate with the inclusion of the residual

estimates of the reduced form as one of the explanatory variable in the structured form equation.

A clear picture from the results shows that there are no significant differences between the two

estimates in each model as noted in Table 4 and 4.1 respectively.

However, judging from the hypothesis and its subsequent decision rule, the residual in

the structured form equation has no correlation with the residual in the reduced form equation.

This decision was arrived at following a 2-t Rule of Thumb; a variable is statistically significant

if its t-value is greater than 2 in absolute value at 5% level of significance. On the other hand, it

is statistically insignificant if its t-value is less than 2 in absolute value at any 5% level of

significance (Gujarati, 2004). This is because the residual variables in the structure estimates in

the two models are not statistically significant, and for this reason, we conclude that there are no

endogeneity problems in the models.

4.2.2 Test for Validity of Instruments

This work adopted a validity test designed by Danis Sargan as advised by Liviaton

(1963) cited in Gujarati (2004) who suggested a way to choose Instrumental Variable(s) (IV)

40

and the test for validity of such instruments called Dubbed SARG Test. This test can also be

done with the help of the F – statistic since we have a multiple instrumental variables such as

total government expenditure (GXP) and gross secondary school enrolment (GS) in the graduate

unemployment model (Tab 4), the broad money supply (M2) and real interest rate (IR) in labour

productivity model (Tab 4.1).

Hypothesis

H0: π1 = π2 = π3 = π4 = 0 (the instrumental variables chosen did not associate with the

endogenous explanatory variable in the model)

Against

H1: π1 ≠ π2 ≠ π3 ≠ π4 ≠ 0 (the instrumental variables chosen did associate with the endogenous

explanatory variable in the model)

Test statistic is given as;

SARG = (n – k)R2

Where n = 26 (the number of observations); and

k = 7 (the number of coefficient in the original regression equation).

This test follows a chi-square distribution with r degree of freedom, and

r = s – q = 3 – 1 = 2

where q = 3 (the lag length of explanatory variables correlated with error

s = 1 (the lag length of the instrumental variable).

Decision rule: reject H0 if chi-square calculated (χ*cal) is greater than chi-square tabulated (χ*

tab)

and accept otherwise.

SARG (χ*cal) = (26 – 7) x 0.127977 = 19*0.127977 = 2.43, and the chi-square tabulated (χ*

tab) =

0.103.

Comparing the two results (chi-square calculated and tabulated), we noticed that χ*cal =

2.43 > χ*tab = 0.103, and for this, we reject the null hypothesis and conclude that the

instrumental variables chosen did associate with the endogenous explanatory variable in the

model.

Also judging from the F – statistic from the result, the first model gave F – statistic value

of 169.5667 with the P – value of 0.0000, while the second model gave F – statistic value of

7695.298 with P – value of 0.0000 with 6 and 19 degrees of freedom and at 5% level of

significance. With high significant nature of these results, we can undoubtedly claim that the

ranked order condition for identifying an instrument or variable is satisfied.

Another way of interpreting these results is that for the first model, the instrumental

variables chosen for the graduate unemployment (UG) model, such as GXP and GS, are

41

uncorrelated with the error terms εt, but highly correlated with the endogenous explanatory

variable LP. Interpreting the second model results also means that the instrumental variables

chosen for the labour productivity model are uncorrelated with the error terms µt but highly

correlated with the endogenous explanatory variable, UG. The result clearly shows that the

stochastic error terms for the two models have zero expected values, which is without loss of

generality when the equation contained an intercept.

Having satisfied these two conditions of simultaneous specification such as equations 1,

2, 3 and 4 in Chapter Three above, we can now present the long-run and short-run estimation

results for the study.

4.2.3 The ECM Result for the Models

The empirical results from modeling the nexus between of government labour policy,

graduate unemployment and labour productivity in Nigeria are presented in Tab 4.2 and 4.3

below.

Table 4.2: ECM result for Graduate Unemployment (UG) Model

Variable Coefficient Std Error t - value t - prob

C 70.60203 55.46384 1.272938 0.2192

TNA -1.590465 0.593273 -2.680831 0.0153

INV -1.08E-11 2.40E-12 -4.486972 0.0003

LP -2.408248 1.464044 -1.644929 0.1173

GXP 2.82E-05 2.75E-06 10.24496 0.0000

YG -4.76E-06 8.79E-07 -5.410996 0.0000

GS 0.280486 0.259429 1.081168 0.2939

ECM-1 -0.953048 0.168674 -5.650245 0.0000

R-squared = 0.985062; Adjusted R-squared = 0.979253; F-statistic = 169.5667

Prob(F-statistic) = 0.000000; Durbin-Watson stat = 1.957611

Source: EViews Output

42

Table 4.3: ECM Result for Labour Productivity Model

Variable Coefficient Std Error t - value t - prob

C 0.958966 0.654852 1.464400 0.1613

LOG(TNA) 0.624447 0.130336 4.791044 0.0002

LOG(INV) 0.000558 0.013167 0.042383 0.9667

UG -0.001936 0.000524 -3.697218 0.0018

LOG(GXP) -0.021862 0.011237 -1.945471 0.0684

YG -2.10E-10 2.66E-09 -0.079003 0.9380

LOG(M2) 0.035870 0.017299 2.073592 0.0536

IR 6.71E-05 0.000181 0.371317 0.7150

ECM-1 -0.011471 0.004482 -2.559430 0.0203

R-squared = 0.960247; Adjusted R-squared = 0.941540; F-statistic = 51.33034

Prob(F-statistic) = 0.000000; Durbin-Watson stat = 1.898457

Source: EViews Output

Economic Interpretation of Results

In this context, the estimated parameters were subjected to test based on economic theory

so as to ascertain whether they are well behaved. In other words, the coefficients derived from

the models are checked to ascertain its conformation with ‘a priori’ expectation underlying each

variable.

The results in Table 4.2 and 4.3 above shows that government labour policy variable

captured by the total number of graduate that benefited (TNA) from National Directorate of

Employment (NDE) since its inception in 1986 have positive coefficients in both models. This

shows that the variables TNA conformed to the economic theory. Following the neoclassical

tradition, the theoretical connection between the government employment policy and

unemployment are in two sides. The one based on the “subsistence wage theory” (policy that

affect the supply side) and the “wage fund” theory (policy that affect the demand side).

On the supply side, the subsistence wage theory posits that every policy that naturally

multiplies the proportion of new individual into labour market will negatively affect

unemployment. This implies that there exists a strong connection between the government

policy and demand and supply of labour (Iniodu and Ukpong, 2001). The demand for labour

depends on the wage policy designed for the payment of wages, which consists of the revenue

necessary for its maintenance and the labour force necessary for full employment (Iniodu and

Ukpong, 2001).

43

Also the signs shown in the result is in line with the NDE mandate which was to design

and implement programmes to combat mass unemployment; to articulate policies aimed at

developing programmes with labour-intensive potentials; to obtain and maintain a data bank of

employment and vacancies in the country with a view to acting as a clearing house to link job

seekers with vacancies, and implement any other policies as may be laid down from time to time

by the Board established under Section 3 of its Enabling Act.

These results clearly prove that government employment policy has high influence over

employment condition of the economy. In the graduate unemployment model, a unit increase in

the number of graduates benefiting from NDE programmes will reduce graduate unemployment

and increase labour productivity by -1.59 and 0.624 units respectively. Following the 2-t rule of

thumb, a variable is statistically significant if the t-value of the variable is greater than 2 in

absolute term at a given 5% level of significance. Similarly, a variable is not statistically

significant if its t-value is less than 2 in absolute term at a given 5% level of significance.

Therefore, the government labour policy tagged in this study as TNA in both models were

statistically significant having their t – values as -2.680 and 4.7910 respectively.

Also investment variable proxy as gross fixed capital investment has the expected

negative sign of -1.08E-11 in the graduate unemployment model and positive sign of 0.000558

on the labour productivity model, conforming to ‘a priori’ expectation, though, with low

transmission effects compared with that of government labour policy variable. The investment

variable is statistically significant in the graduate unemployment model having t – value of -

4.486, but statistically insignificant in the labour productivity model with t – value of 0.042

respectively. By implication, this result shows that gross fixed capital formation has not

adequately induced labour productivity, probably due to high level of graduate unemployment in

the country.

On the coefficient of the total government expenditure variable, the theoretical expected

signs were not satisfied. The result shows that total government expenditure variable has

positive relationship with graduate unemployment and negatively related to labour productivity.

By implication, as the total government spending increases, the graduate unemployment also

increases. Though, this can be said to be the true picture of the Nigerian economy, for instance,

our raw data shows that as total government spending rose from N701, 059.40billion in 2000 to

N3, 536,624.90 trillion in 2010, graduate unemployment also rose from 40.3% in 2000 to

whopping rate of 88.3% in 2010. To further confirm the effect of total government spending on

graduate unemployment and labour productivity in Nigeria, the t –values of 10.244 and -1.945

44

associated with the variable in the two models show a robust statistical significance of the

variable.

Again, the influence of total government spending on graduate unemployment and

labour productivity show some high level misalignment between government spending and

stability of macroeconomic environment in the country. This implies that increasing government

total spending on the contrary distorts macroeconomic factors such as unemployment, real

interest rate and output gap does. However, output gap shows no significant effect on labour

productivity but significantly and negatively influences graduate unemployment.

The long-run error correction mechanism variable show significant correction of the

disequilibrium at lag one in the two models. It proves that 95% correction is made every one

year to position graduate unemployment to its equilibrium root. On labour productivity, a little

error (1.1%) is corrected to adjust it disequilibrium state. The two long-run factors in the models

has the right negative sign , showing that at every disequilibrium in the graduate unemployment

and labour productivity instruments, there is positive adjustment mechanism at every one year to

put them back to equilibrium.

Evaluating the fitness of the two models analyzed above, the coefficient of multiple

determinations R2 of 0.98 and 0.96 respectively suggests that 98% of the variation in the

graduate unemployment model was explained by the selected variables in that model, while 96%

variations in the labour productivity model was also explained by the factors included for that

estimation, a proof of goodness of fits. In the same vein, F – statistics of 169.56 and 51.33

respectively shows that the models are well specified and as a result maintains good fit.

Another interest measure of the precision of this analysis is the Durbin-Watson (DW)

statistic. A rule of thumb shows that when the DW statistic is less than R2 in a model, not

minding the significant level, such model is said to suffer from multicollinearity, positive first

order autocorrelation and spurious regression. Therefore, with the DW statistic being greater

than the R2 in this study, and with reasonable number of the significant factors, these models are

said to be free from multicollinearity, positive first order autocorrelation, estimation bias

emanating from wrong specification of model and spurious regression.

4.2.4 Total Factor Productivity Result

The estimated result show that factors, such as, gross fixed capital formation, used as

investment variables in the study, total government expenditure and output gap exact significant

and appreciable influence on the total factor productivity of labour in Nigeria (see table 8).

45

Table 4.4: Result of the Total Factor Productivity Model Variable Coefficient Std. Error t-Statistic Prob.

C 7.321950 1.309652 5.590758 0.0000

LOG(INV) 0.560979 0.108395 5.175312 0.0001

D(UG) 0.001635 0.004959 0.329673 0.7453

LOG(GXP) 0.563929 0.105412 5.349778 0.0000

D(YG) 1.37E-07 4.31E-08 3.178559 0.0049

LOG(GS) -0.240307 0.174899 -1.373977 0.1854

R-squared 0.994951 F-statistic 748.8309

Adjusted R-squared 0.993622 Prob(F-statistic) 0.000000

S.E. of regression 0.145763 Durbin-Watson stat 1.735370

Source: E-Views Output

The robust coefficients and significant levels of investment, government total spending

and output gap variables in the TFP model proved that they are the major factors affecting

labour productivity in Nigeria. As indicated in the result, a unit increase in gross fixed capital

formation (investment parameter) will have 56% increases in the total factor productivity of

labour in the economy, provided other factors are kept constant. Also the t-value of 5.17

confirms the factor that investment is a major driver of total factor productivity of labour in

Nigeria.

In a similar way, a unit increase in total government expenditure will increase labour

productivity of labour by 56% provided other factors are kept constant. This is in-line with

scholars view on total factor productivity and investment. According to Mali (1978) total

productivity is a measure of how resources are being brought together in organizations and

utilized for accomplishing a set of results. It is reaching the highest level of performance with

the least expenditure of resources". Productivity is viewed as the instrument for continuous

progress, and of constant improvement of activities. It is often seen as output per unit of input.

Hence, higher productivity connotes achieving the same volume of output with less factor inputs

or more volume of output with the same amount of factor inputs. Thus, increased productivity

could result from the reduction in the use of resources, reduction in cost, use of better methods

or improvement in factor capabilities, particularly labour.

Roberts and Tybout (1997) and Tybout (1992), assuming a neo-classical production

function at the sectoral or industry level, define total factor output to be a concave function of

the vector of inputs and time (a proxy for shift in technological innovation). To them, the

elasticity of output with respect to time is the total factor productivity. In a more general sense,

TFP = Total Output ie Weighted Average of all inputs. Critical among these factor inputs are

46

labour, capital, raw materials and purchase of spare parts, and other miscellaneous goods and

services that serve as inputs in the production process.

The output gap variables, though, have a significant effect on total factor productivity; its

coefficient is not robust when compared with investment and total government expenditure

influence. The result shows that a unit increase in the output gap will have 1.37E-07 influence

on total factor productivity, a less than 1% effect. Graduate unemployment and gross secondary

school enrolment variables exhibit some influence on total factor productivity, but their effects

are statistically not different from zero, meaning they have insignificant effect on total factor

productivity. This result confirmed the classical theory of unemployment that says the supply of

labor is derived from worker’s choice whether to spend part of time working or not working

(leisure). Supply of hours worked is a positive function of the real wage, because if the real

wage rises, workers supply more hours of work. In equilibrium, demand and supply of labor are

intersected at a clearing point that determines the equilibrium real wage rate and full

employment.

4.2.5 Granger Causality Result

On the Pairwise Granger Causality between government labour/employment policy,

graduate unemployment and labour productivity, the result shows that unilateral causal

relationship exist between government employment policy and graduate unemployment (see Tab

4.5).

Table 4.5: Pairwise Granger Causality Tests

Null Hypothesis: Obs F-Statistic Probability

UG does not Granger Cause TNA 24 4.07030 0.03379

TNA does not Granger Cause UG 0.20811 0.81394

LP does not Granger Cause TNA 24 63.4267 3.9E-09

TNA does not Granger Cause LP 5.90022 0.01016

LP does not Granger Cause UG 24 2.27625 0.12996

UG does not Granger Cause LP 5.42773 0.01366

Source: E-Views Output

Table 4.5 Granger Causality results shows that graduate unemployment granger cause

government labour/employment policy, but later does not granger cause the former. On the other

hand, a bilateral causal relationship exists between labour productivity and graduate

unemployment. Finally, the result in Table 4.5 above shows that unilateral relationship exists

47

between the labour productivity and graduate unemployment. It shows that where graduate

unemployment granger cause labour productivity, and labour productivity does not granger

cause graduate unemployment. This finding is also in-line with classical theory of

unemployment, where Wicksell, after 1921 World War I, argued that the causes of the

unemployment are the surplus people, shortage of capital brought about by the war, and the

disorganized state of the monetary system. For the third cause, after the war prices were falling

and producers decided to produce lower amounts of production because they knew they would

receive lower prices for their products. Thus, they let their money sit idle in banks and workers

became unemployed. These causes suggest that emigration became one of the important policies

for solving the unemployment problem.

4.3 Evaluation of Research Hypotheses

The three hypotheses of this study were evaluated based on the estimation results and the

findings drawn from it. Thus, the first hypothesis, which states that government

labour/employment policy has no significant impact on the rate of graduate unemployment and

labour productivity in Nigeria was rejected based on the results, and conclusion was drawn

that government policy on employment has significant negative impact on the rate of graduate

unemployment and positive significant impact on labour productivity in Nigeria. Also the

second hypothesis, which states that there is no relationship between graduate unemployment

and total factor productivity in Nigeria was as well rejected, having noticed that long-run

relationship exist between graduate unemployment and total factor productivity (see the result in

Table 4.5).

Finally, in the third hypothesis there is no causal link between government labour policy,

graduate unemployment and labour productivity in Nigeria was also rejected when we found

that graduate unemployment granger cause government labour/employment policy, but later

does not granger cause the former. Again, graduate unemployment granger cause labour

productivity, and labour productivity does not granger cause graduate unemployment.

48

CHAPTER FIVE

SUMMARY AND CONCLUSION

5.1 Summary of Research Findings

The study evaluated the effect of government labour/employment policy on graduate

unemployment and labour productivity in Nigeria. In order to avoid running spurious regression

emanating from misspecification and omitted variables bias, we carried out several preliminary

tests, aimed at implementing all the data processing techniques. Among them are the trend

analysis of the data series, unit root test using Augmented Dickey-Fuller and Phillip Perron

stationary test. The unit root result showed that the variables exhibit similar integrating order (∆

= 1). Also, co-integration test was conducted using Engle – Granger approach to ascertain the

long-run linear combination among the variables. In line with the rule guiding the Engle-

Granger Co-integration approach, the long-run error correction model was finally adopted for

the estimation.

In line with the estimated models and their results, a number of findings were made. It

found that government labour/employment policy variable captured by the total number of

graduates that have benefited (TNA) from National Directorate of Employment (NDE) since its

inception in 1986 have positive coefficients in both models. The TNA variable significantly

affects graduate unemployment and labour productivity in Nigeria. While the government

employment policy negatively influences graduate unemployment, it positively influences

labour productivity, and this is in conformity with economic theory. The study found that a unit

increase in the number of graduates benefiting from NDE programme will reduce graduate

unemployment and increase labour productivity by -1.59 and 0.624 units respectively.

Also found is that increasing government total spending negatively affect labour

productivity and positively affect graduate unemployment. By implication, it means that the

incremental spending by the government has no employment or labour productivity target in

Nigeria. However, output gap shows no significant effect on labour productivity but

significantly and negatively influences graduate unemployment.

The long-run error correction mechanism variable show significant correction of the

disequilibrium at lag one in the two models. It proves that 95% correction is made every one

year to position graduate unemployment to its equilibrium root. On labour productivity, a little

error (1.1%) is corrected to adjust its disequilibrium state. The two long-run factors in the

models has the right negative sign, showing that at every disequilibrium in the graduate

49

unemployment and labour productivity instruments, there is positive adjustment mechanism at

every one year to put them back to equilibrium.

Also found in the study is that a unit increase in gross fixed capital formation

(investment parameter) will have 56% increases in the total factor productivity of labour in the

economy, provided other factor is kept constant. Also the t-value of 5.17 confirms the factor that

investment is a major driver of total factor productivity of labour in Nigeria. A unit increase in

total government expenditure will increase labour productivity of labour by 56% provided other

factors are kept constant. This is in-line with scholars’ view on total factor productivity and

investment. A unit increase in the output gap will have 1.37E-07 influence on total factor

productivity, a less than 1% effect. Graduate unemployment and gross secondary school

enrolment variables exhibit some influence on total factor productivity, but their effects are

statistically not different from zero, meaning there have insignificant effect of total factor

productivity.

On Granger causality test, the study found that graduate unemployment granger cause

government labour/employment policy, but later does not granger causal the former. On the

other hand, a bilateral causal relationship exists between labour productivity and graduate

unemployment. Also unilateral relationship exists between the labour productivity and graduate

unemployment. It found that graduate unemployment granger cause labour productivity, and

labour productivity does not granger cause graduate unemployment.

Finally, the three hypotheses of the study were rejected and conclusion was drawn that

government policy on labour has significant impact on the rate of graduate unemployment and

labour productivity, there is long-run relationship between graduate unemployment and total

factor productivity, and there is causal link between government labour policy, graduate

unemployment and labour productivity in Nigeria.

5.2 Policy Recommendations

Based on the findings in this study, the following policy recommendations are proffered

for urgent consideration:

� Based on the findings in this study that government employment policy, though

significant to graduate unemployment, in reality has not reduced raging unemployment

figure among school graduates. We then recommend that there should be some

restructuring or reforms on the existing government labour/employment programmes,

such as National Directorate of Employment (NDE) in order to redirect efforts towards

achieving its mandate.

50

� Since our result shows that graduate unemployment causes government employment

policy but the later does not cause the former show a clear evidence of government

policy failure to tackling unemployment especially graduate unemployment, we

therefore recommend that new employment programs should be established with the sole

aim of tackling graduate unemployment. Since unemployment cuts across various

barriers (gender, class, age) and the National Directorate of Employment which is

saddled with combating unemployment cannot effectively manage graduate

unemployment as its resources (time, finance) are divided among other classes. Thus,

giving more attention to graduate unemployment by establishing a body with the

mandate of combating graduate unemployment will go a long way.

� Any holistic programme to arrest unemployment in Nigeria should also capture those

who are heavily underutilized, grossly underpaid, in order to have fair representation of

the unemployment situation in the country.

� Also the importance of skills acquisition cannot be ignored. Hence, we recommend that

learning of specific skills should be inculcated into the curriculum of schools, right from

primary to tertiary institutions in Nigeria and made compulsory for every student

irrespective of the course of study. Every student must be made to undertake any of the

following vocational courses like fashion designing, automobile repairs, traffic control,

animal husbandry, typesetting, catering, horticulture, swimming, memo writing, satellite

installation, wood work and even cooking should be included as compulsory electives

thereby creating the spirit of entrepreneurship, as practiced in some developed countries.

� The spending pattern of the government, whether state or federal government should be

checked, since the rising government total spending has not translated to increase in

labour productivity. If not checked, it is capable of distorting the macroeconomic growth

of the nation.

5.3 Conclusion

In this study the research evaluated the effect of government labour/employment policy

on graduate unemployment and labour productivity in Nigeria. Unemployment in the views of

the International Labour Organization (ILO) “occurs when a person is available and willing to

work but currently without work.” This situation especially the graduate unemployment which is

very much prevalent in Nigeria has become a menace to the society with its consequences on

political and socio-economic effects. These consequences cannot be over-emphasized as

51

unemployment especially graduate unemployment can lead to political instability and economic

quagmire.

Economically, graduate unemployment hinders the graduate from contributing their

acquired skills/knowledge to building a sound economic productivity. Although, government

have made several efforts by establishing some organizations and institutions saddled with the

task of tackling unemployment, such as National Directorate of Employment (NDE). On this

basis, the question that was evaluated in this study is the effects of this employment programme

on graduate unemployment and labour productivity in Nigeria. Among the findings of the study

is that government employment policy has significantly affected graduate unemployment and

labour productivity in Nigeria. While the government employment policy negatively influences

graduate unemployment, labour productivity is positively influenced. By implication, it implies

that as more graduates benefit from these programmes, graduate unemployment reduces and

labour productivity will increase. The study found that a unit increase in the number of graduate

benefiting from NDE programme will reduce graduate unemployment and increase labour

productivity by -1.59 and 0.624 units respectively.

Contrary to economic theory, the study found that increase in government total spending

negatively affect labour productivity and positively affect graduate unemployment. By

implication, it means that the incremental spending by the government has no employment or

labour productivity target in Nigeria. However, output gap shows no significant effect on labour

productivity but significantly and negatively influences graduate unemployment.

The long-run error correction mechanism variable show significant correction of the

disequilibrium at lag one in the two models. It proves that 95% correction is made every one

year to position graduate unemployment to its equilibrium value. We conclude that reducing

graduate unemployment requires clear political will by the government irrespective of party

affiliation, religion and ethnicity.

5.4 Limitation of the Study

The major limitations of this study are non-existence and non-availability of some

important data that are important for a study like this. For instance, variables such as the number

of employment opportunities created by the government annually and number of graduates

actually employed over the years in those jobs created, are needed in this type of study. Salary or

wage rate was needed also, but due to limitation of data, was not included. We then suggest for

further study, that research on this subject should try to include these variables to identify their

influence on graduate unemployment and labour productivity.

52

53

REFERENCES

Adeyeye, J.Olufemi, Aina Sola, and Ige,A.Kolapo (2012). An Analysis of Factors Influencing

Hiring/Placement of University Graduates in Nigeria: Pakistan Journal of Business and

Economic Review, 3, 1.

Ake, (2010). “What does the future holds for today’s graduate: in the Guardian Newspaper of

Tuesday March, 30.

Akinboyo, G. B. (1987). "Job Creation as a Productivity Measure for Employment Problems in

Developing Countries: The Nigeria and Holland Experience", Increasing Productivity in

Nigeria, National Productivity Centre, Lagos, pp 404-411.

Akintoye, I. R., (2008). Reducing unemployment through the informal sector: A case study of

Nigeria. Europ. J. Econ. Finan. Admin. Sci., 11, 97-106.

Amupitan , O. (2011). An Assessment of the Role of the National Directorate of Employment

(Nde) In Reducing Graduate Unemployment In Kaduna State.

Asoluka, N. & Okezie, I. (2009), “Unemployment and Nigerian Economic Growth”, Applied

Economics, 34,15 83-92 of Econometrics, .39,199 – 211.

Awogbenle, A. C. & Iwuamadi, K. C. (2010). “Youth unemployment”: Entrepreneurship

development programme as an intervention mechanism.

Bakare, (2011). The determinant of urban unemployment crises in Nigeria: an econometric

analysis. 2, 184-192,3. Journal of emerging trends in economics and management

science.

Bassey & Atan, (2012). Labour Market Distortions and University Graduate Unemployment in

Nigeria: Issues and Remedies.74-75.

Bassey, G. E. & Atan J. A. (2012). “Labour market distortions and university graduate

unemployment in Nigeria. Issues and Remedies”: Maxwell Scientific Organization.

Journal of Economic Theory 4,3, 67-76.

Biobele Richards. (2007). “Problems of recruitment in civil services”: case of Nigeria civil

service. African journal of Business management.1,6,142-153.

Brayton, F., Roberts, J. M., & Williams J. C. (1999). “What’s happened to the Phillips curve?”

Division of Research and Statistics, Finance and Economics Discussion Series, 49.

Broersma, L.&. Butter ,F.A .(2002). An explorative empirical analysis of the influence of labour

flows on wage formation, Applied Economics, vol. 34 (13), pp. 1583-1592.

Chletsos, M., Kollias, Ch. &. Manolas, G. (2000). “The relationship between wages,

productivity and labour demand in Greece: An error correction approach.”RISEC, 47,1,

95 – 110.

Dabalen,A,B.Oni and O.Adekoka (2000). Labour market prospects for university graduates in

Nigeria. High. Educ. Manage., 14: 1-36.

54

Dickey, D. A., and Fuller, W.A. (1979). “Distribution of the Estimators for Autoregressive Time

Series with a Unit Root”, Journal of the American Statistical Association, Vol.74, pp.

427 – 31.

Dickey, D. A., (1979) and Fuller, W.A. (1981), “Likelihood Ratio Statistics for Autoregressive

Time Series with a Unit Root”, Econometrica, Vol.49, pp.1057 –72.

Ekkehard, E. (2011). Determinants of unemployment flows. Labour market institutions and

macroeconomic policies, . No. 209.

Ekpo, A. H. (2008). The Nigerian economy: Is it at the crossroad? In: Presidential Address at the

NES Annual Conference, pp: 67-105.

Essenberg, H. M. (1996). Determinants of unemployment duration for men and women in

Turkey. ERF Working Paper Series No: 0332, Cairo, Egypt: Economic Research

Forum.

Essien, E. B and Atan J. A. (2006). Generating employment in Nigeria: The role of rural

agriculture. Afri. J. Stabil. Dev., 2,1, 111-136.

Fajana, S. (2000). Functioning of the Nigerian labour market, and company, Lagos, Nigeria.

Fontana, Wood and Lofgren.(2008).“Microeconomic evidence on gender inequality and growth

in Nigeria”.

Frank and Bernard. (2001). “An assessment of the role of the national directorate of employment

in reducing graduate unemployment in Kaduna state”.

Godwin E. and Johnson. F. (2012). “Re-Designing University Curricular as a Long – Term

Solution to Graduate Unemployment in Nigeria.” In Unemployment Problems in

Nigeria; CBN Bullion Vol. 25, No. 4; Oct/Dec.

Gordon, R. J., (1997). “Is there a Trade-Off between Unemployment and Productivity Growth.”

In: D. J. Snower and G. de la Dehesa, eds Unemployment Policy, Government Option

for Labour Market.. Cambridge University Press, Cambridge, 433 - 463.

Granger, C. (1988). “Some recent developments in a concept of causality”. Journal of

conometrics.39,199-211.

Granger, C. W. J. (1969). "Investigating causal relations by econometric models and cross-

pectral methods." Econometrical 37: 424-438.

Grill, E. and Zanalda, G. (1995). "Growth and employment in developing countries: Where Do

We Stand?", Development Studies Working Papers No. 86, October, The International

Development Centre.

Gujarati, (2004). Basic econometrics. 4th

edition.

Hausman, S. (1978). “Unemployment and Consumption: Why are Job Losses Less Painful than

Expected near the Mediterranean”, mimeo, CEMFI.

55

Hsing, Y. (2001). “A macroeconomic analysis of the impact of union wage increases on non

union wages”, Applied Economics Letters, 8,803 – 06.

Ibrahim, M. (2006). Empirical survey of children and youth in organized violence in Nigeria.

Nigeria, A Divided Country, Newswatch

Igbinosun, E. (2010). “Nigeria National Employment Policy”.

Ilyin S A, Motyler, (1986). “ABC of social and political knowledge: What is Political

Economy?” Progress publishers, Moscow.

Iniodu, P.U. and I.I. Ukpong, (2001). A Synopsis of the History of Economic Through: From

the Greek philosophers to John Maynard Keynes. Centre for Development Studies,

University of Uyo, Uyo.

International Labour Organization Publication. (2005). Youth: Pathway to Decent Work

(defining youth). 1450-2275. Issue 11.

Juan Ramón. (2011). “Youth unemployment in Spain”: causes and solutions.

Kayode, (2008). “The African symposium” Graduate unemployment in Nigeria: A blind spot in

the nation’s educational system.

Labour Force Survey March, (2009). Lagos, Nigeria.

Liviaton C. (1963). “Socio-economic dimensions of unemployment in India”, mimeo, Central

Statistical Organisation, India.

Mackinnon, J. (1991). “Critical values for co- integration tests in long-run economic relationship

in readings in cointegration (eds) Engle and Granger, review, 22, 2,137-153.

Mali, P. (1978). Improving Total Productivity, John Wiley and Sons Inc. New York.

Manuelli, T. (2005). Unemployment duration and economic incentives – a quasi random

assignment approach. European Economic Review, 49(7), 1799-1825.

Marcelino and Mizon, (1999). “Modeling shifts in the wage-price and unemployment-inflation

relationships in Italy, Poland and the U.K”, IGIER working paper,145, bocconi

university.

Mohammed, A.H. (2010). "An entrepreneurial strategy for the prosperity of the American

economy," Journal of Applied Business and Economics,11, 4, 33 – 43.

Mortensen, Dale T. and Christopher Pissarides, (1994). Job Creation and Job Destruction in the

Theory of Unemployment," Review of Economic Studies, 61 (0), 397{415.

Mortesen, S. .J. and Pissarides (1998). Estimating the probability of leaving unemployment.

Econometrica, 47(5), 1249-1266.

National Bureau of Statistics. (2010). Statistical News: Labor Force Statistics No. 476. Abuja:

The NBS Publication.

56

National Bureau Statistics (2009). Labour Force Survey Abstract statistics.

Nelson and Plosser. (1982). “Trend and Random walks in macroeconomic Time series: Some

evidence and Implications”, Journal of monetary economics, 10, 139-62.

Nickell,S.(2005). Unemployment in the OECD Since the 1960s: What do we Know?, Economic

Journal, January.

Nishhiyama and Leube (1984). Does unemployment compensation affect unemployment

duration? Economic Journal, 113(484), 190-206.

Njoku,A and Ihugba, O. (2011). Unemployment and Nigerian economic growth. Proceedings

of the 2011 International Conference on Teaching, Learning and Change (c)

International Association for Teaching and Learning (IATEL).

Okojie, C.E.E. (1995). "Human capital investment for productivity growth in Nigeria", The

Nigerian Economic and Financial Review, 1, 1, 43-68.

Okpechi, S. O. (1991). "Private Sector And Productivity Improvement In Nigeria", Productivity

for Self-Reliance and Excellence, National Productivity Centre, Lagos, 40-58.

Olufemi, and Adebola, (2011). “An Analysis of factors influencing hiring / placement of

University graduates in Nigeria”. Oxford University Press, New York, 267 - 276.

Oyabade (2003). “Youth Unemployment And Implications For Stability Of Democracy In

Nigeria” Journal of Sustainable Development in Africa 13,1, 2011.

Pauw,k,M Oosthuizen and C.van Der westhuizen (2008). “Graduate unemployment in the face

of skills shortages”. A labour market paradox. S. Afr. J. Econ.,76(1): 45-57.

Pigou, A. C. (1933). & Solow, R. M. (1980). Theory of Unemployment, London: Macmillan.

Piguo, A.C. (1927). “Wage policy and unemployment” economic journal 37 1927 pp355-68

republished by Dimand, R.W. (2002), “The origin of macroeconomic,8. London Rout

ledge.

Puhani, P .(2002). “Relative wage and unemployment changes in Poland: micro econometric

evidence”, Economic System,26,99 – 126.

Puhani, P. A. (2004). The Contribution of Degree Subject to the Gender Wage Gap Among

Graduate, Research Report to the Anglo-German Foundation 1429, London. Report.

Roberts, M.J and Tybout J. R. (1997). "Producer Turnover and Productivity Growth in

Developing Countries", The World Bank Research Observer,12,1, 1-18.

Sandra, (2006). “Reducing unemployment Through the Informal Sector: European Journal of

Economics, Finance and Administrative Sciences.

Shapiro, C. and Stiglitz, J. (1984). “Equilibrium unemployment as a discipline device”,

American Economic Review, 74, 3, 433 – 44.

57

Shumpter, C. (1934). The Consequences of Rigid Wages in Search Models," Journal of the

European Economic Association, 2 (2/3), 469{479.

Standing, G. ( 1983). “The notion of structural unemployment’ international labour.

Tejvan Pettinger. (2011). “ Definition: Types and strategies for reducing unemployment.

Thirlwill, H. L. (1979). Estimating the determinants of the unemployment duration for college

graduates in Taiwan. Applied Economics Letters, 6(10), 677-681.

Todaro, M. P. (1990), A model of labor migration and urban unemployment in less developed

countries," American Economic Review, 59, 138-148.

Trehan, M. (2001). Unemployment duration and the interactions between unemployment

insurance and social assistance. Labour Economics, 13(6), 773-798.

Turunen, J. (1999). “ Disaggregated wage curves in the United States: Evidence from panel

data of oung workers”. Applied Economics,30,1665-77.

Vecchi H.E. (1995). Early identification of low employability graduates in Malaysia: the use of

proportional hazard model. Jurnal of Teknologi Maklumat & Sains Kuantitatif, 7(1), 41-

48.

Walterskirchen, E. (1999). The Relationship Between Growth, Employment and

Unemployment in the EU. European Economist for an Alternative Economic Policy

workshop, Barcelona, Spain.

William, H. M. (1976), “Theory of firm: managerial behavior. Agency of costs and ownership

structure. 3,4,305-366.

World Bank. (1999). African Development Indicators 1998/99, the World Bank, Washington,

DC.

58

APPENDIX

Endogenuous Estimates

Dependent Variable: LP

Method: Least Squares

Date: 06/11/14 Time: 17:54

Sample: 1987 2012

Included observations: 26

Variable Coefficient Std. Error t-Statistic Prob.

C 35.96825 0.943191 38.13466 0.0000

TNA 0.417975 0.021391 19.53969 0.0000

INV 5.81E-14 5.45E-13 0.106737 0.9162

UG -0.037714 0.021379 -1.764050 0.0947

GXP 4.86E-08 8.82E-07 0.055141 0.9566

YG 2.83E-09 1.54E-07 0.018318 0.9856

M2 8.79E-14 2.37E-13 0.370453 0.7154

IR 0.013081 0.007345 1.781063 0.0918

R-squared 0.960880 Mean dependent var 55.23846

Adjusted R-squared 0.945667 S.D. dependent var 2.346074

S.E. of regression 0.546858 Akaike info criterion 1.878403

Sum squared resid 5.382958 Schwarz criterion 2.265510

Log likelihood -16.41924 F-statistic 63.16056

Durbin-Watson stat 2.294514 Prob(F-statistic) 0.000000

Dependent Variable: LP

Method: Least Squares

Date: 06/04/14 Time: 12:40

Sample: 1987 2012

Included observations: 26

Variable Coefficient Std. Error t-Statistic Prob.

TNA 1.158564 0.078943 14.67595 0.0000

INV -5.20E-13 4.79E-12 -0.108512 0.9147

UG -0.054811 0.188153 -0.291310 0.7740

GXP -9.96E-07 7.76E-06 -0.128418 0.8992

YG 9.96E-10 1.36E-06 0.000733 0.9994

M2 5.82E-13 2.09E-12 0.278870 0.7834

IR -0.018228 0.064248 -0.283707 0.7797

R-squared -2.199686 Mean dependent var 55.23846

Adjusted R-squared -3.210113 S.D. dependent var 2.346074

S.E. of regression 4.813805 Akaike info criterion 6.205657

Sum squared resid 440.2817 Schwarz criterion 6.544375

Log likelihood -73.67354 Durbin-Watson stat 0.401641

59

Dependent Variable: LOG(LP)

Method: Least Squares

Date: 06/11/14 Time: 18:10

Sample: 1987 2012

Included observations: 26

Variable Coefficient Std. Error t-Statistic Prob.

C 2.957904 0.014095 209.8596 0.0000

TNA 0.022042 0.000291 75.85171 0.0000

INV -9.95E-15 9.28E-16 -10.72223 0.0000

UG -0.000988 3.59E-05 -27.47677 0.0000

GXP -1.93E-08 1.51E-09 -12.78671 0.0000

YG 9.24E-11 2.55E-10 0.362267 0.7216

M2 1.10E-14 4.37E-16 25.18405 0.0000

IR -0.000351 1.72E-05 -20.42850 0.0000

RESID02 0.019022 0.000389 48.84280 0.0000

R-squared 0.999724 Mean dependent var 4.010727

Adjusted R-squared 0.999594 S.D. dependent var 0.044846

S.E. of regression 0.000904 Akaike info criterion -10.91295

Sum squared resid 1.39E-05 Schwarz criterion -10.47746

Log likelihood 150.8684 F-statistic 7695.298

Durbin-Watson stat 1.304843 Prob(F-statistic) 0.000000

-.03

-.02

-.01

.00

.01

.02

1990 1995 2000 2005 2010

LOG(LP) Residuals

60

Estimated long-run labour productivity models

Dependent Variable: LOG(LP)

Method: Least Squares

Date: 06/04/14 Time: 13:36

Sample: 1987 2012

Included observations: 26

Variable Coefficient Std. Error t-Statistic Prob.

C 0.958966 0.654852 1.464400 0.1613

LOG(TNA) 0.624447 0.130336 4.791044 0.0002

LOG(INV) 0.000558 0.013167 0.042383 0.9667

UG -0.001936 0.000524 -3.697218 0.0018

LOG(GXP) -0.021862 0.011237 -1.945471 0.0684

YG -2.10E-10 2.66E-09 -0.079003 0.9380

LOG(M2) 0.035870 0.017299 2.073592 0.0536

IR 6.71E-05 0.000181 0.371317 0.7150

ECM-1 -0.011471 0.004482 -2.559430 0.0203

R-squared 0.960247 Mean dependent var 4.010727

Adjusted R-squared 0.941540 S.D. dependent var 0.044846

S.E. of regression 0.010843 Akaike info criterion -5.943157

Sum squared resid 0.001999 Schwarz criterion -5.507662

Log likelihood 86.26104 F-statistic 51.33034

Durbin-Watson stat 1.898457 Prob(F-statistic) 0.000000

Endogeneity Estimate for UG model

Dependent Variable: UG

Method: Least Squares

Date: 06/11/14 Time: 17:26

Sample: 1987 2012

Included observations: 26

Variable Coefficient Std. Error t-Statistic Prob.

C 190.7218 83.04028 2.296738 0.0332

TNA 2.591692 0.917797 2.823817 0.0108

INV -9.02E-12 3.86E-12 -2.338469 0.0304

LP -5.311456 2.222258 -2.390117 0.0274

GXP 2.74E-05 4.45E-06 6.162319 0.0000

YG -4.41E-06 1.42E-06 -3.105822 0.0058

GS -0.083457 0.407366 -0.204870 0.8399

R-squared 0.958567 Mean dependent var 46.63814

Adjusted R-squared 0.945483 S.D. dependent var 27.31800

S.E. of regression 6.378444 Akaike info criterion 6.768529

Sum squared resid 773.0063 Schwarz criterion 7.107248

Log likelihood -80.99088 F-statistic 73.26213

Durbin-Watson stat 1.631026 Prob(F-statistic) 0.000000