i
YOUTH JOB-SEARCH BEHAVIOR AND SOCIOECONOMIC PREDICTORS IN
KENYA: AN EMPIRICAL ANALYSIS OF SEARCH METHODS AND LABOR-
MARKET MATCHING
Arizona State University
SOC 324 - Sociology of Work and Organizations
Summer 2023
ii
TABLE OF CONTENTS
TABLE OF CONTENTS ................................................................................................................................2
LIST OF TABLES .........................................................................................................................................4
LIST OF FIGURES .......................................................................................................................................4
ABSTRACT ................................................................................................................................................5
CHAPTER ONE: INTRODUCTION ................................................................................................................1
1.1 Background to the Study ..................................................................................................................1
Figure 1.1: Trends in Age composition of Working-Age Population, from 1979 to 2019 ...............................2
Distribution of Kenya’s Labour Force (15-64) by Activity Status from 1989-2019 .........................................2
Table 1.1: Trends in Labour Force, Employment and Unemployment by Gender from 1989 to 2019 ............2
Figure 1.2: Labour Force Participation Rates by Gender from 1989 to 2019 ...............................................3
Distribution between the Working-Age (15-64) Population and the Youth (15-34) by Activity Status from
2019-2022................................................................................................................................................3
Table 1.2: Comparison of Activity Status between the Working Age-Population and Youth from 2019 to
2022 ........................................................................................................................................................4
Figure 1.3: Distribution of unemployed persons by gender in 2021 ............................................................4
1.1.2 Job Search Initiatives in Kenya ..................................................................................................5
1.2 Research Problem Statement ..........................................................................................................6
1.3 Research Questions .........................................................................................................................7
1.4 Research Objectives ........................................................................................................................7
1.4.3 Specific Objectives ....................................................................................................................7
1.5 Significance of the Study .................................................................................................................8
1.6 Outline of the Research Paper .........................................................................................................8
CHAPTER TWO: LITERATURE REVIEW ........................................................................................................9
2.1 Introduction ....................................................................................................................................9
2.1.1 The Concept of Job Search ........................................................................................................9
2.1.2 Choice of job search method ................................................................................................... 10
2.2 Theoretical Review........................................................................................................................ 11
2.3 Empirical Literature Review ........................................................................................................... 13
2.4 Overview of the Literature ............................................................................................................ 15
CHAPTER THREE: RESEARCH METHODOLOGY .......................................................................................... 17
3.1 Introduction .................................................................................................................................. 17
3.2 The model .................................................................................................................................... 17
3.3 Data and selection of variables ...................................................................................................... 18
Table 3.4: Variables definitions and measurement .................................................................................. 19
Table 3.4: Explanatory Models Included in the Research Theoretical Model ............................................. 19
iii
3.5.1 The Hausman-McFadden test / IIA Assumption ....................................................................... 21
CHAPTER FOUR: EMPIRICAL RESULTS ...................................................................................................... 22
Table 4.1: A descriptive summary of the job search methods used by Kenyan youths ............................... 22
Table 5.2: A descriptive summary of the Covariates by Job Search Channel .............................................. 22
Table 6.3: Model Summary ..................................................................................................................... 24
Registered at Employment Agency.......................................................................................................... 29
Placed or Answered a Job Advert ............................................................................................................ 29
Referrals from Relatives ......................................................................................................................... 30
Waited at the Street Side ........................................................................................................................ 30
CHAPTER FIVE: CONCLUSION AND RECOMMENDATION ........................................................................... 32
REFERENCES ........................................................................................................................................... 34
iv
LIST OF TABLES
Table 1.1: Trends in Labour Force, Employment and Unemployment by Gender from 1989 to
2019 ................................................................................................................................ 2
Table 1.2: Comparison of Activity Status between the Working Age-Population and Youth
from 2019 to 2022 .......................................................................................................... 4
Table 3.4: Explanatory Models Included in the Research Theoretical Model .............................. 19
Table 4.1: A descriptive summary of the job search methods used by Kenyan youths ................ 22
Table 4.2: A descriptive summary of the Covariates by Job Search Channel .............................. 22
Table 4.3: Model Summary ........................................................................................................... 24
Table 4.4: Summary of the Hausman-McFadden Test Results ..................................................... 25
Table 4.5: Parameter Estimates ..................................................................................................... 25
Table 4.6: Relative Risk Ratios (RRR) ......................................................................................... 27
LIST OF FIGURES
Figure 1.1: Trends in Age composition of Working-Age Population, from 1979 to 2019 ............. 2
Figure 1.2: Labour Force Participation Rates by Gender from 1989 to 2019 ................................ 3
Figure 1.3: Distribution of unemployed persons by gender in 2021 ............................................... 4
v
ABSTRACT
Youth unemployment in Kenya continues to be a major challenge over the past few years despite
labor market reforms and employment promotion initiatives that have been implemented. Knowing
how the youth search for jobs is fundamental to improving job matching and reducing labor
underutilization. This research work investigates the patterns and factors that influence the choice of
job-search methods among youth in Kenya based on nationally representative data from the 2015/16
Kenya Household Integrated Budget Survey (KHIBS). The study examines the use of five job-search
channels: registration at an employment agency, workplace enquiries, placing or responding to job
advertisements, referrals from relatives, and waiting at street-side locations. The study employs a
multinomial logit model to look at how different factors such as age, gender, education level, access to
information, duration of job searches, and location can determine the choice of a particular search
method. The findings reveal that employees are most likely to inquire at the workplaces, whereas
registration at employment agencies is the least preferred option. Formal search channels users
significantly are influenced by education level whereas those with limited networks and information
who may be waiting at street-side for jobs heavily depend on the informal sector. Results of the
research demonstrate that youth job-search behavior is diverse and therefore it is very important for
labor market interventions that target specific groups to continue through Kenya to further improve
access to information, the skills of employability, and the efficiency of job-matching systems.
1
CHAPTER ONE: INTRODUCTION
1.1 Background to the Study
Youth employment continues to be a priority in the development agenda not only in Africa as a
region but globally through the Sustainable Development Goals (SDGs, 2030). The SDGs
acknowledge youth unemployment to be a major global concern thus priority has been given for
countries to enhance labour market participation of young people through skills development,
access to full and productive employment and reducing the proportion of youths in
unemployment (UN DESA., 2022).
The ILO estimates world unemployment rate at 5.8% and the LFPR at 59.7% in 2022.The global
youth unemployment is more than three times the adult rate, estimated at 15.6 percent in 2021
and 12.7 percent in Africa, indicating the tough situations young men and women endure in the
labour market. In 2022, the LFPR for the youth aged between 15-24 in Africa region was 43.9
percent compared to 40.1 percent globally. The young males had higher labour force participation
rates at 46.9 percent than the young females at 40.9 percent. Youth employment to population
ratio in Africa was higher at 38.3 percent compared to global estimates at 34.1 percent with the
young male at 41.1 percent higher than that of the females at 35.4 percent (ILO, 2023).
The young persons aged between 15 to 35 years are found mostly in Africa, estimated to be more
than 400 million young men and women. To harness the potential of Africa’s youthful
demographic dividend, the African Union (AU) is implementing several youth development
policies. These include the AU Agenda, the AYC, YDPA, and the MDYE. These policies
recognize the critical role of participation of young persons in the labourforce and have clearly
outlined the strategic interventions on promoting youth employment and skills revolution (Africa
Union, 2023).
Kenya has a young population and labour force. The population of youth (aged 15-34) in Kenya
was slightly below five million in year 1979 rising to slightly over seventeen million in year 2019
as illustrated in Figure 1. In 1979, the youth were 57.9 percent of the working –age-population of
close to 7.5 million. By year 2019, young persons were 63 percent of the total working-age
population of approximately 27 million.
2
Figure 1.1: Trends in Age composition of Working-Age Population, from 1979 to 2019
Source: Author (2023), Data from KPHC, 2019 Report
Distribution of Kenya’s Labour Force (15-64) by Activity Status from 1989-2019
Labour force increased significantly from 7.8 million in 1989 to 19.8 million persons in 2019 as
shown in Table 1.1. The labourforce and persons employed are higher amidst the male compared
to females, while unemployment rate is higher among the female. However, in 2019 the male had
a higher unemployment rate of 15.4% in comparison with the females at 12%. The number of
total employed persons has significantly increased reaching 17 million in 2019 reflecting the
growing number of working-age (15-64) population shown in Figure 1.1. Unemployment rates
keep on fluctuating throughout the years despite the constant increase in employment, in 2019
the country recorded a high unemployment rate of 13.7% compared to 9.7% in 2009 and 6.5% in
1989.
Table 1.1: Trends in Labour Force, Employment and Unemployment by Gender from
1989 to 2019
1989
1999
2005/06
2009
2015/16
2019
Labour Force
(000,000)
Total
7.8
12.4
14.6
15.8
19.3
19.8
Male
4.1
6.4
8.2
9.7
10
Female
3.7
6
7.5
9.6
9.7
Employed
(000,000)
Total
7.3
11.1
12.7
14.2
17.9
17
Male
3.8
5.8
7.4
9.2
8.5
Female
3.5
5.3
6.8
8.7
8.6
Unemployed
(000,000)
Total
0.5
1.3
1.9
1.5
1.4
2.7
Male
6.5
0.5
0.8
0.5
1.5
Female
6.6
0.7
0.7
0.9
1.2
Unemployment
rates (%)
Total
6.5
10.1
13.1
9.7
7.4
13.7
Male
6.5
8.4
9.9
5.2
15.4
Female
6.6
11.9
9.4
9.4
12
Source: Author (2023), Data from the KNBS Labour Force Reports 1989-2019
2019
2009
1999
1989
1979
0
10000000
20000000
30000000
40000000
50000000
0-64+ 15-64 15-34
3
LFPR estimates the proportion of the population aged 15 to 65 years, that is actively engaged in
a specific labour market, either employed or searching for work (KNBS, 2019). The statistics
from the KNBS labour force reports illustrated in Figure 1.2. indicate that since 1989 to 2019,
Labour Force Participation Rates in the country has been constantly above 70%, with the male
recording a higher LFPR compared to the total LFPR and that of the females throughout the years.
Figure 1.2: Labour Force Participation Rates by Gender from 1989 to 2019
Source: Author (2023), Data from the KNBS Labour Force Reports 1989-2019
Distribution between the Working-Age (15-64) Population and the Youth (15-34) by
Activity Status from 2019-2022
From 2019, the KNBS has been releasing labour statistics through the Quarterly Labour Force
Surveys (QLFS) as indicated in details as Annex 1. As shown in Table 1.2 and figure 1.1, the
youth are most represented in the labourforce in Kenya. The rate of unemployment is higher
among the youth (nearly twice) compared to the total employable age population , in Quarter
Four (Q4) of 2022 the youth recorded unemployment rate of 8.5% in comparison to 4.9% of the
total labour force. Labor Force Participation Rates among the youth is lower with an average of
56% compared to LFPR of 67.5% among the total working-age population. The segment of the
Youth NEET was 21.5% in Q4 of 2021 decreasing to 19% in Q4 of 2022, while persons aged 20
– 24 years continue registering high proportions of youth NEET approximated at 27.5%. (KNBS,
QLFR 2022).
Labour underutilization is the mismatch between demand and supply of labour, due to an
unfulfilled need for decent employment among jobseekers, as indicated in Table 1.2 labour
underutilization in Kenya is relatively high with youth being most underutilized.
LFPRs by Gender, 1989-2019
100
80
60
40
20
0
1989
1999
2005/06
2009
2015/16
2019
Year
TOTAL MALE FEMALE
LFPR Rate (%)
4
Table 1.2: Comparison of Activity Status between the Working Age-Population and Youth
from 2019 to 2022
Q4 (Oct-
Dec. 2019)
Q4(Oct-
Dec. 2020)
Q4 (Oct-
Dec. 2021)
Q3(July-
Sept. 2022)
Q4(Oct-Dec.
2022)
Age Cohorts
15-64
15-34
15-64
15-34
15-64
15-34
15-64
15-34
15-64
15-34
Population (000,000)
27.1
17.1
27.8
17.6
28.3
17.9
28.9
18.3
29.1
18.3
Labour Force
(000,000)
19
10.3
19.1
10.1
18.7
9.5
19.1
9.4
19.4
10
Employed (000,000)
18.1
9.5
18.1
9.3
17.6
8.6
18.1
8.6
18.4
9.2
Unemployed
(000,000)
0.93
0.8
1
0.8
1.1
0.9
1
0.8
0.9
0.8
Rate of unemployment
(%)
4.9
7.5
5.4
7.7
5.6
8.9
5.3
8.7
4.9
8.5
Long-term rate of
unemployment (%)
2.2
3.5
2.8
4.1
3.4
5.9
2.2
3.7
3.2
5.7
Not in LabourForce/
Inactive (000,000)
8.1
6.9
8.7
7.5
9.6
8.4
9.8
8.6
9.7
8.4
LabourForce
Participation
Rate (%)
70.2
63.1
68.7
57.6
66.1
53.3
66
51.6
66.7
54.5
Labour
Underutilization (%)
7.9
10
13.5
15.8
11.5
14.8
10.4
17.5
9
13.6
NEET Rate
21.5
20.3
19
Source: Author (2023), Data from the KNBS Quarterly Labour Force Reports 2019-2022
The young persons between the age 20-34 years’ experience higher rates of unemployment than
the remaining labourforce as indicated in figure 1.3, showing that the largest proportion of
unemployed youths in Kenya is among those of age of 20-29 years.
Figure 1.3: Distribution of unemployed persons by gender in 2021
Source: Author (2023), Data from KNBS Quarterly Labour Force Surveys, (2021)
To address the high youth unemployment rates, the Government of Kenya has created various
institutional, legal, and policy frameworks and initiatives. The most recent is KYDP (2019)
indicating the Government commitment in promoting sustainable development through a more
inclusive and equitable approach (GoK, 2019). The Policy provides an inclusive framework for
200000
0
15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64
Unemployed Male
Unemployed Female
5
harnessing the potential of the young in their participation in the development and growth of the
country.
1.1.2 Job Search Initiatives in Kenya
The process of searching for a job has been aided by the internet (Suvankulov et al., 2012; Kuhn).
According to Pew Research Center survey conducted in early 2015, 54 percent of job seekers use
the internet. Posting open positions online is also simpler and less expensive for employers.
Internet-based technology is also used to assess a person's qualities and match them with the most
likely suitable job that is currently open (Mugambi, 2022). In Kenya, 15 percent of job seekers
use social media sites like LinkedIn, Fuzu, and Brighter Monday to share information about them
and look for opportunities that fit their needs (Nyaata, 2021). Due to the internet, it is now
significantly cheaper to obtain information about jobs and complete the application process
(Marí-Klose, 2020).
The National Employment Authority (NEA) was established in 2016 through the NEA Act (2016)
to provide public employment services. The NEA has 28 county employment offices. So, 19
counties do not have such offices. The mandate of NEA includes; the registration and placement
of job seekers, provision of vocational guidance and counseling, the dissemination of job
openings, internships, the management of labour migration among others (NEA ACT, 2016). To
this end, it has developed the NEA Integrated Management System (NEAIMS) to ease access to
public employment services.
The Kenya Labour Market Information System (KLMIS), was established in 2014 under the
Ministry responsible for Labour and Social Protection with the technical assistance from the
World Bank. KLMIS is a web-based portal whose purpose to ease access to and ensure quality
of LMI to help different labor market actors use evidence in decision making, by serving as labour
market intelligence/observatory that provide information on the supply of and demand for skills
and on career prospects (KLMIS, 2022). Such information is crucial for addressing skill gaps and
mismatches, and therefore for reducing structural unemployment and raising productivity of
Kenyan workers. Key Labor Market Indicators have been developed to provide information that
is relevant and useful for main labor market actors, especially students (prospective workers),
jobseekers, workers as well as education and training institutions and public employment services
(KLMIS, 2022).
The KYEOP, is another effort to address the difficulties young people face when looking for
work (Menya, 2020). In partnership with the Government of Kenya, training providers and
employers training and work experience were provided to 70,000 targeted youths in 17 counties.
6
Although KYEOP targeted unemployed youth with high school qualification or lower, majority
of applicants were university graduates (Menya, 2020). Another initiative by the Government of
Kenya, is the Ajira Digital Project. The project aims to facilitate job searching for youth in
addition provide the young people with soft skills and an introduction to accessible digital online
work, especially for those who have not been in a stable employment facility (KYDP, 2019).
There are labour market interventions by GoK and NGOs aimed at special population groups.
Persons with disabilities (PWDs) are supported in to search for jobs by the NCPWD in
collaboration with the Innovation-to-Inclusion (i2i) program. This is done through a digital
platform that promotes career growth for PWDs. The platform offers PWDs the chance to learn
useful skills and discover jobs that suit their personality.
1.2 Research Problem Statement
According to the ILO, the young men and women are the majority in Africa’s population in
comparison with the rest of the world with over 400 million youths aged between 15 and 35 years.
(ILO, 2023). Kenya has a young population where youth aged between 15-34 is slightly over
seventeen million. The youth represent 63 percent of the working age population of
approximately 27 million people. The LabourForce has significantly increased by 153.8%,
approximately 7.8 million in 1989 to 19.8 million persons in 2019. The number of total employed
persons has significantly increased reaching 19.1 million in 2019. The rates in unemployment
keep on fluctuating throughout the years despite the constant increase in employment, in 2019
the country recorded a high unemployment rate of 13.7% compared to 9.7% in 2009 and 6.5% in
1989 ((KNBS Economic Survey, 2023).
Unemployment and labour underutilization rates among the youth continue to be a major
challenge in the country, where unemployment rate is nearly twice compared to the working age
population. In 2022, unemployment rate of the youth was 8.5% compared to 4.9% of the total
labour force. Labor Force Participation Rates among the youth is lower with an average of 56%
compared to LFPR of 67.5% among the total working-age population. The rate of the Youth
NEET was 21.5% in 2021, while those aged 20 – 24 years recorded the highest proportion of
young men and women NEET reaching 27.5% (KNBS, 2022).
The high unemployment and labour underutilization remains a critical area of policy concern for
the Government of Kenya, with the youth aged 18-25 years bearing the greatest burden of
unemployment as majority of them are NEET. These particular age-cohort are at risk because
they are neither improving their future employability nor gaining work experience thus exposing
them to the danger of being shut out of the labour market and excluded in society, and are likely
7
to be caught in poverty trap making them vulnerable to radicalization and violent extremist
activities. Further, the country has inadequate available decent jobs to absorb the increasing
number of youths graduating from education and training institutions each year (Labour and
Employment Sector Plan, 2023-2027).
The Government of Kenya is implementing various mechanisms to facilitate creation of a million
new jobs every year through spurring economic growth and tapping into potential sectors such as
the digital and the creative economy (BETA, 2022-2027). The ML&SP is promoting access to
employment opportunities through Labour Migration to countries of destination. This is being
driven through signing of bilateral labour agreements, targeted skills development programmes
and registration of Private Recruitment Agencies who facilitate job-matching and placement
(National Policy on Labour Migration, 2023). Therefore, with the on-going employment
promotion strategies by the Government of Kenya, reduction in unemployment and labour
underutilization will partly depend on how efficient the matching of jobseekers with employment
opportunities is. The Unemployed youth through their search effort and methods to use can to a
great extent influence effectiveness of the matching process. However, empirical evidence on
youth job search behaviour is lacking.
1.3 Research Questions
This research was guided by the following research questions:
i. What was the incidence of job search methods among the youth in Kenya?
ii. What were the individual, household and locational characteristics associated with
choice of different methods of job search among the youth in Kenya?
1.4 Research Objectives
1.4.2 Main Objective
The main objective of this research was to investigate the incidence and determinants of job
search methods among the youth in Kenya.
1.4.3 Specific Objectives
This study had the following specific objectives.
i. To determine the incidence of methods used by Kenyan youth in search of
employment.
ii. To identify demographic and socioeconomic factors associated with use of
different job search methods among Kenyan youth.
8
iii. To draw implications for youth labour market programs on job search in Kenya.
1.5 Significance of the Study
The prompt and efficient matching of job-searching youth with available positions is a
fundamental requirement for the smooth functioning of the labor market. The research findings
have demonstrated social, geographical, and demographic aspects that influence the manner of
choosing methods of young people in Kenya for searching employment.
Understanding prevalent job search channels and factors influencing their use, youth
unemployment interventions resulting from the study findings serve as an evidence base for
policy changes in Kenya. Policymakers through targeted programs can match the needs and
constraints of young job seekers thus enhancing the effect of policy initiatives and youth
employment outcomes.
In a resource-constrained environment, policymakers and organizations need to be very prudent
in their resource allocation if they are to successfully address youth unemployment, among other
pressing issues. Decision-makers informed by the results of this study can therefore reposition
resources towards the most efficient job search channels thus ensuring that interventions are
directed at the youth where they can have the greatest impact in connecting them with suitable
employment opportunities.
Understanding the job search channels preferred by youth and the determinants behind these
choices can help address the mismatch between available jobs and the skills possessed by young
job seekers. By aligning job search strategies with the job market demands, young individuals
can make better-informed decisions about skill development and training, thus reducing the
overall skills gap.
A thorough analysis of job search channels can contribute to a more efficient labor market.
Matching job seekers with suitable vacancies promptly benefits both employers and employees.
The study can pave the way for smoother labor market operations by identifying barriers and
facilitators to efficient job matching.
1.6 Outline of the Research Paper
The rest of the research paper was organized into chapters. Chapter two discussed the literature
review for the study. It considered both the theoretical review and the empirical review of the
study. Chapter three discussed the research methodology. Chapter Four provided an in-depth
discussion of the research findings and empirical results. Chapter Five discussed the conclusion
and reccommendations from the Study.
9
CHAPTER TWO: LITERATURE REVIEW
2.1 Introduction
This chapter entailed literature from scholars and researchers that connect to labour markets and
how they influenced job search. Lastly, an overview of the empirical literature was discussed in
brief.
2.1.1 The Concept of Job Search
Job search encompasses all activities that encourage residents to seek employment by identifying
jo opportunities, applying for jobs, and attending job interviews (Nivorozhkin, 2006). The
concept of job search is part of job-search theory, which incorporates a person's economic
judgments to consider employment opportunities accorded to them as well as finding a job and
their remunerations expectations (Green et al., 2011).
Job search may be due to discontentment in the present job or, by virtue, being unemployed.
Unemployment may occur because of termination from previous employment, no renewal of job
contracts, or the need to change jobs (Van Hooft et al., 2021). The job search process aims to
match job seekers with appropriate job openings (Ceniza-Levine & Thanasoulis-Cerrachio,
2011). Thus, high preference in the job search is a consequence of high rate, especially by the
high number of graduates who complete studies every year.
Tymon (2013) posits that job search is mainly due to data asymmetry/imperfection and
uncertainty, which usually exists in the labour market. To that end, the job qualifications for those
seeking jobs must coincide with the minimum requirements of a real work environment offered
by employers. Therefore, it is vital that when looking for job openings job seekers gets acquainted
with the data concerning the job offers available and the best technique toapply in job search.
This is because a job seeker can manipulate the result of the search through the smartness used
in finding jobs visa vis the strategies available.
Once a worker decides to search for a job, the next decision is about the focus they put in doing
so. This entails the number of hours they put in job search as well as their consistency in sending
applications to different hiring firms (Faberman et al., 2020).
10
2.1.2 Choice of job search method
Traditionally, job hunt was classified into looking for formal and informal jobs. White collar or
formal jobs are accessed through various advertisement platforms like publications, periodicals,
personal contacts and digitally through the Internet. The latter means is achieved mainly through
referrals by close associates either close friends or family. Job search can alternatively be
classified as 'active' or 'passive,' with the former implying greater action than the latter (Okafor,
2011).
During the process of job search, the applicants collect information regarding the jobs they are
interested in then proceed to strategize on the best methodology they are going to apply to
succeed. The probability on whether the search for a job is going to succeed or not is dependent
on the determination put in applications as well as using various channels. Job search is likened
to fishing as knowing where many jobs are situated is key in landing a job faster compared to
those hunting in areas with limited openings (Osberg, 1993). The latest form of job search is
through use of internet which can match employers and employees from different scopes within
a very short period of time. Rees (1966) noted that it’s easier to get a job through referrals than
any other method. The author further reiterated that it takes minimal time to obtain informal jobs
than formal ones. It’s also vital to note that informal jobs are less costly as the costs associated
with advertising are eliminated due to the fact that referrals are the main form of advertisement.
The main assumption is that individuals can only refer people with similar or close characteristics
to theirs hence reducing the chances of firm incurring costs on screening new employees
(Fernandez et al., 2000).
Developed nations have a public employment office whose main aim is to link workers
particularly those looking for formal jobs to their potential employers. This is important
especially to those who don’t have contact persons or people to connect them to job openings.
The services are usually free of charge and the state job provider functions as a bridge between
employees and employers in many instances. However, it has been noted that those jobs obtained
through the state provider are characterized by low remunerations and low rate of offer
acceptance by job seekers (Blau & Robins, 1990; Holzer, 1988; Osberg, 1993). Subsequently,
majority of job seekers view these state corporations for matching them with potential employers
as bureaucratic. Some countries have suggested that inoder to address the above issue, they have
brought private sector on board to help achieve their target which is to match employees with
potential employers (Martin & Grubb, 2001).
11
Job seekers who have no clue on the exact level of remuneration may as well attain employment
by going directly to the employers without connections required. This is achieved through “blind”
application which sometimes bear fruits. According to Kahn and Low, (1990), in order to
circumnavigate the issue of low remunerations as well as offer rejections, many prefer getting
jobs through referrals. The internet has opened up numerous methods of job search in recent years
which are faster and more efficient (Kuhn & Skuterud, 2004). Furthermore, it gives applicants
the liberty of choosing which jobs are closer or meets their wage expectations hence reducing the
number of offer rejections.
The quest for employment among the youth, understanding the incidence and determinants of job
search methods is crucial. Holzer (1988) sheds light on this subject, revealing insights that can
empower unemployed youth in their pursuit of meaningful work. The incidence of job search
methods refers to the frequency with which individuals utilize different methods to search for
employment opportunities. In other words, it is a measure of how job seekers go about finding
job openings and applying for positions. Job search methods can vary widely and may include
approaches such as submitting online applications, networking, attending job fairs, using
recruitment agencies, and directly contacting companies. Determinants of job search methods
pertain to factors that influence an individual's choice of specific approaches. These determinants
include educational attainment, prior work experience, geographic location, personal preferences,
and available resources. By understanding these determinants and their impact on job search
behaviour, unemployed youth can make informed decisions about which methods are most likely
to yield positive outcomes.
2.2 Theoretical Review
Social Network Theory (SNT) is a theory that focuses on explaining the relationships and patterns
of interactions among individuals living in a social group. The theory notes that all social groups
are characterized by mutual independence and interpersonal interaction between members
(Reinders, 2011). It contends that the social structure determines the pattern of interaction and
relationship among individuals within the social group (Sih, Hanser, & McHugh, 2009). It is a
generic theory that has been developed by different personalities. Jacob Moreno made the earliest
contribution towards the development of this theory after he developed the first sociogram in the
1930s. This theory is applied in numerous social science disciplines, including psychology,
business, and development studies.
In 1973, Gronovetter applied the social network theory to examine how people look for jobs.
According to Gronovetter (1973), a person’s network comprises two groups of people: friends
and acquaintances. Friends refer to people who have very strong ties to the person, while
12
acquaintances are people with whom the person has weak ties. In his research, Gronovetter (1973)
found that a person who relies on acquaintances to obtain information about jobs is more likely
to find employment than a person who depends on friends. He explained that a network of
acquaintances is easy to develop; hence, they reduce the time and path lengths for securing
employment. Furthermore, friendship ties take long to develop as individuals in this network
gauge each other’s intentions as well as develop common norms, values, and expectations.
Consequently, friendship networks tend to be small and compact when compared to the network
of acquaintances.
Gronovetter (1983) also argues that acquaintances ties act as bridges between people’s friendship
networks. These connections link one friendship network to another creating large social systems
that grant a person access to information from distant parts. They provide individuals with
resources that are outside their friendship circle. A person with few acquaintances tends to form
social systems that are fragmented and incoherent, separated by geography and ethnicity, and
spread ideas slowly. However, Gronovetter's work does not imply that friendship ties are not
useful.
In fact, Gee et al. (2017) found that in closed societies with large income inequality, friendship
ties matter more than acquaintances when seeking job. Older adults have higher chances of a
larger network of acquaintances than youth by virtue of their experiences, which explains why
older adults find it easier to secure employment when compared to youths (Barnett, 2011). Even
among the youths, the networks of friends vary from one person to the next. Some youths have
greater networks of acquaintances than others. Youths with strong acquaintance connections have
higher probability of landing jobs compared to the ones with weak connections. Information,
referrals, access to hidden opportunities, social capital, and cultural influences provided by a
person's network can shape their decisions on how to approach their job search effectively and
efficiently.
The Human Capital Theory (HCT) was developed in 1960 as an advancement of work pioneered
by Theodore Schultz and Gary Becker. In their articles, Schultz (1961) and Becker (1962) argued
that investing in human capital leads to greater economic productivity. Prior to the development
of this theory, economic productivity was largely viewed as a function of physical assets such as
land and equipment. Human capital generally refers to skills, knowledge, abilities, and
experiences possessed by an individual that are essential to economic production. The role of
formal education in human capital development is largely emphasized in Becker (1962), where
it was found that, on average, individuals with higher education levels had higher incomes than
their counterparts with lower education. On the other hand, human capital can be created through
13
informal experiences. The theory also claims that it is possible for a person to create human
capital when he or she is involved in exchanges that lead to the transfer of skills and capabilities.
According to Schultz (1961), some forms of human capital, such as athletic talent, are not
acquired but are rather transferred through genes. However, these innate abilities must also be
identified and nurtured to translate into economic output. Generally, youth from high
socioeconomic backgrounds have higher chances of getting college education and higher
compared to those from poor background. Youth from families with high socio-economic status
are more likely to develop human capital because they are more likely to interact constantly with
highly skilled and educated people. Some families tend to involve children in their family
businesses at an early age, giving them an upper hand when it comes to the acquisition of
competencies and experiences that are requisite to securing employment, especially self-
employment.
HCT reiterates that there exists a direct correlation linking education and skills to employability
and consequently economic development. This theory is vital when analyzing youth
unemployment in Kenya as majority of them are fresh from school and lack skills desired by
many employers, explaining the main reason as to why most face unemployment particularly in
developing nations in the Sub African region like Kenya.
Therefore, Human Capital Theory predicts that individuals will select job search methods based
on their skills, education, experience, and personal attributes. The theory emphasizes the
importance of matching one's human capital to the chosen method to augment the likelihood of
getting suitable employment and advancing in the career.
2.3 Empirical Literature Review
Holzer (1988) analysed various levels of methodology applied in job seeking youths in the USA.
A sample of no enrolled and none listed young males (age 16 through 23) was taken from the
1981 panel data from the National Longitudinal Survey (NLS) data and was analyzed using OLS
and probit. The finding revealed that the main factors in establishing the methodology preferred
by different individuals was dependent on the costs associated with it as well as the speed of
receiving a response regarding the jobs applied for. It further revealed that the most popular
method in job seeking is through referrals. Additionally, the researcher established that through
this method, there are few offers rejections accounting for over 65 percent of main channel among
the youth.
14
Blau and Robins (1990) carried out a study on the relationship linking job search options to their
results insisting on the importance of separation between those employed and unemployed. The
study results revealed that various job search methods bear varied fruits on the job search. It
further revealed that although different job search methodology yields different results, the main
difference is the time lag between finding new jobs. Weber and Mahringer (2008) carried out
research to establish the relationship linking job search options among youth in Austria. The
results show that different methods used by the youth in searching for employment and their
success was dependent on the level of education as well as skill set. The research further shows
that in the informal sector, the above-mentioned factors are not much important as they can gain
experience and skills on job.
Osberg (1993) analysed the impact of different strategies used in public sector job search
methodology. The study utilized data collected for a 7-year period (1981-9586) where it revealed
that methods of looking for jobs are affected by different economic cyclicals. This implies that
when there is economic boom and many firms are hiring, the methods applied are different from
those used during economic downturns and firms are downsizing. The study of job-search
strategies within specific economic contexts is crucial for understanding how individuals adapt
to changing conditions. Works by Kuhn and Skuterud (2004) have highlighted the impact of
economic downturns and shifting industry demands on the efficacy of different search methods.
Given the emphasis on the early 1980s in Osberg's study, it is relevant to consider the job-search
behaviour of youth. Studies by Holzer (1988) have shown that youth face distinct hurdles in job
search owing to limited experience and networks, potentially leading to the exploration of
unconventional search methods.
Caliendo et al. (2011) examines the impact of social networks in the labour markets for search of
employment. Their analogy shows people with varied network for their job search have higher
probability of attaining new jobs and the levels of remuneration as well.
The study by Guillemyn and Horemans (2023) delves into the job search behaviour of older
individuals and examines whether age-related differences impact the size and effectiveness of
networks when securing employment. Feldman and Ng (2007) explored how age influences
individuals' job search strategies, taking into consideration factors such as motivation,
experience, and perception of job opportunities. Furlong and Cartmel (2007) highlighted that
older job seekers face distinct challenges, including potential gaps in skills and networks. Given
that social networks can be pivotal in job search, it is pertinent to explore whether older
individuals require larger networks to achieve comparable outcomes to younger job seekers.
15
Research by Feldman and Bolino (2000) suggested that older workers might rely more on internal
networks.
Wahba and Zenou's study (2005) investigates the complex interplay between methods of
searching for jobs, spatial density and social networks within the context of Egypt. The study
applied time data series to analyse study variables where the study results revealed that those who
have a large network of connection have a greater chance of getting employment than their
counterparts who have less. Granovetter (1973) reiterated that friends who have weak relation to
the job seekers are likely to assist them in landing employment than close relatives to the job
seekers. Mouw (2003) carried out similar research where the study findings revealed there exists
no correlations among the study variables.
Schaer and Leibbrandt (2006) study examines factors impacting search for work strategies in the
context of the Khayelitsha and Mitchell's Plain regions in South Africa. The study utilized logit
model to analyse the various job search techniques in increasing job seeker chance of getting a
job. The researchers reiterated that, there exists various demographic factors influencing the
chances of job seeker in getting a job. It further revealed that males are more aggressive in job
seeking compared to their female counterparts. The method used in job search varies from one
household to the other; when one member uses social network, its likely other will follow suit
and use the same.
Regarding gender, study by Sacky and Osei (2006) analysed the probability of male and female
job seekers in informal methods. The study findings revealed that the latter is more likely to use
informal techniques which is contradictory to Garcia and Nicodemo (2013) study which
established that there is no correlation among study variables. Garcia and Nicodemo (2013),
carried out a study to establish whether the neighborhood has an impact on job seekers probability
of success in landing a job. The study shows existence of an inverse correlation among the
variables under observation.
Wambugu et al. (2012) studied the behaviour of job search personnel in Kenya. The multinomial
logit model was applied in the analysis. The results showed that 83.15% of active job seekers
prefer referrals to land jobs. Lastly, the study showed that the job search channels used were
influenced by demographic factors.
2.4 Overview of the Literature
It was evident that most studies have analyzed various demographic factors that influence job
search among job seekers. There was an indication that educational attainment was a crucial
factor that largely dictated the choice of search for job channel employed in searching for work
16
among various groups of job searchers. Equally, the extent of social network ties was critical in
influencing the frequency with which an individual can easily get a work opportunity. To that
end, this current study endeavored to look at varied factors that determined job search among the
youth in Kenya by incorporating other factors that many studies fall short of by focusing solely
on job search channel employed in searching for work among youth in Kenya.
Despite the existing literature, there were still several research gaps that needed to be addressed,
especially on determinants of job search methods amidst the youth. This was due to the fact that
job search may not be homogenous and different population groups may behave differently, thus
heterogeneity in job search behaviour. Moreover, there was a need to explore the role of non-
traditional job search channels, such as social media, in the job search process, as there was
limited data among the youth. Additionally, there was need to investigate the long-term
implications beyond the scope of the study on various channels used by youth in searching for
work, including those with disabilities employment outcomes.
Although Holzer (1988) reiterated why referrals are vital in job search, it doesn’t come out clear
on whether this method gives a job seeker higher chance of landing a job or not. The reason as to
why this method is more popular among job seekers is because its less costly and ensures the
reputation of the hiring company is kept intact as the employee of a firm referring someone it
means they have close or similar character. Furthermore, the neighborhoods with high level of
employed individual is likely to experience low levels of unemployment as information flow from
the employed to unemployed regarding job openings.
17
𝑖
𝑖
𝑖
𝑖
𝑖
CHAPTER THREE: RESEARCH METHODOLOGY
3.1 Introduction
This chapter entailed; the model, data selection variables, data analysis techniques and diagnostic
tests which were applied in the study.
3.2 The model
The research aimed at investigating the demographic as well as the influence of choices applied
by job seeker youth in Kenya. Multinomial logit model was preferred to analyse the study
variables. The model was preferred due to its ability to be used where the study has more than
two groups. The categories are; π1, π2,…,πJ where π1+π2+…+πJ=1.
Consider an individual youth i whose utility of using job search method j is given by
𝑈
𝑖𝑗
= 𝑉
𝑖𝑗
+ 𝜀
𝑖𝑗
(1)
where
𝑉
𝑖𝑗
is a deterministic component, and
𝜀
𝑖𝑗
is the unobservable random component. The
functional form of the deterministic component of utility function is specified as 𝑉𝑖𝑗 = 𝑋′ 𝛽𝑗.
The utility function can be written as
𝑈
𝑖𝑗
=
𝑋
′
𝛽
𝑗
+
𝜀
𝑖𝑗
, j=1,…J
(2)
Where X is a vector of individual and household attributes such as education level, gender and
location, represents vectors of unknown parameters to be estimated and is defined as in
equation (1).
It was assumed that a young person will chose job search method j if it yield the highest utility,
that is, if 𝑈𝑖𝑗 > 𝑚𝑎𝑥𝑈𝑖𝑛 where n is the set of job search methods. The probability that the ith
youth chooses the job search method j is:
𝑃𝑖𝑗=𝑃𝑟[𝑈𝑖𝑗 > 𝑈𝑖𝑛; 𝑗 ≠ 𝑛, 𝑛 = 1, … , 𝐽] (3)
This study considered choice among four job search methods thus, j=1, 2….,4. The search
methods are dependent variables. Therefore, the model of job search methods:
𝑃
𝑖𝑗
=
𝑒𝑥𝑝(𝑋
′
𝛽
𝑗
)
, 𝑓𝑜𝑟 𝑗 = 1,2,3,4 (4)
5
𝑗=1
𝑒𝑥𝑝(𝑋
′
𝛽
𝑗
)
Where 𝑃𝑖𝑗 is the probability of the alternative that individual i is observed to choose from the
j set of choices 1,2,3 and 4. The Probability of being in the reference group was given as follows:
𝑃 = 1 , 𝑓𝑜𝑟 𝑗 = 0 (5)
𝑖0
4
𝑗=1
𝑒𝑥𝑝(𝑋
′
𝛽
𝑗
)
1+∑
1+∑
18
The natural logarithm of combining equations (4) and (5) gave the estimated equation as
suggested by (Greene, 2012) as follows:
𝑙𝑜𝑔
𝑃
𝑖𝑗
= 𝑋
′
(𝛽
− 𝛽
)
(6)
𝑃
𝑖0
𝑖
𝑗
0
The equation modeled the log odds of choosing job search method j over the reference category
𝑗′, given the values of the explanatory variables Xi. The coefficients β1, β2, ..., βk represented the
change in the log odds of choosing method j for a one-unit change in the corresponding
explanatory variable Xi, holding all other variables constant. The intercepts αj represented the
baseline log odds of choosing method j, which are specific to each job search method. The
explanatory variables used in the model were location (1=rural, 0=urban), age-group (1=15-19,
1=20-24, 1=25-29, 1=30-34, and 1=35 and above and 0 if others), search duration (1=1-24
months, 0=25+ months), education level (1=primary, 1=secondary, 1=vocational, 1=university
and 0 if others), access to information (1=No, 0=yes), and gender (1=male, 0 = female). These
variables were included in each of the four logit regression models to determine their significance
in predicting the choice of a specific job search method. By comparing each job search method
with the reference category (enquiring at workplaces), the model was able to estimate the
probability of choosing each job search method based on the values of the explanatory variables.
From equation (6) above, the relative probabilities of other job search options were relative to the
reference group. The choice of reference group is arbitrary. The relative risk of choosing
alternative j instead of choice j’ is given as:
𝑃𝑟(𝑌
𝑖
=𝑗)
= 𝑒𝑥𝑝(𝑋
′
𝛽 )
(7)
𝑃𝑟(𝑌
𝑖
=𝑗
′
)
𝑖 𝑗
From equation (7) above, yi were the risks associated with the model due to the changes in the
study variables (Viitanen, 1999).
3.3 Data and selection of variables
This study utilized secondary data sourced from the KHIBS 2015/16. The data was collected
using structured questionnaires administered at household level in all the 47 counties in Kenya.
The study focused on the "main method for seeking work in the last week." The options were
registering at employment agency, enquiring at workplaces, placing or answering a job advert,
referral from relatives, and waiting at street-side, hence the researcher aggregated into five job
search channels. The explanatory variables included age, gender (sex of the individual),
household characteristics (access to internet); education level, duration of job search and region
of residence (urban or rural).
19
The survey used NASSEP (V) structured as (C1, C2, C3 and C4). The statistical software SPSS
was used to analyse the collected data. The clusters totaled to 2,400 which were achieved by
using EPSEM. The methodology categorized the clustered data in terms of country, urban/rural
and geographic where a sample of 6 households picked from each cluster. A total of 23,852
households were sampled during the survey.
Table 3.4: Variables definitions and measurement
Based on the evidence in Chapter 1, the literature review in Chapter 2 of this study and the data
available, the following explanatory models were included in the Model; Age, Gender, Location,
education level, and duration of job search.
Table 3.4: Explanatory Models Included in the Research Theoretical Model
Variable
Operation Definition
Measurement
Job search Methods
Job search methods refer
to the various strategies
and approaches that youths
use to search for
employment opportunities
0 if registered at employment
agency (y=0)
1 if enquired at Workplaces
(y=1)
2 if placed or answered a job
advert. (y=2)
3 if Referral from
relatives(y=3)
4 if waited at street-side.
(y=4)
Level of Education
Educational level refers to
the highest level of formal
education completed by
the youth in Kenya
1 if Primary; 0 if others
1 if Secondary; 0 if others
1 if Vocational; 0 if others
1 if University; 0 if others
Age
Age refers to the
chronological age of the
youth participants, which
was measured in years.
1 if 15 - 19 Years, 0 if others
1 if 20 - 24 Years, 0 if others
1 if 25 - 29 Years; 0 if others
1 if 30 - 34 Years; 0 if others
1 if 35 years and above, 0 if
others
20
Variable
Operation Definition
Measurement
Location
Location refers to the
geographical area where
the youth population
resided, which was
categorized as either urban
or rural.
1 if Urban; and
0 if Rural
Gender
In this context, gender was
used to distinguish
between individuals who
identified as male or
female.
1 if Male; and
0 if Female
Duration searching
for Job
In this context, search
duration refers to the time
period an individual had
been actively looking for
employment. It was
initially expressed in
months, but later
categorized into two
groups: 1-24 months and
25 months and above.
1 if 1 - 24 Months; and
0 if 25+ Months
Access to
Information
(internet, computer,
television, radio)
In the context of the study,
"Access to Information"
referred to whether
individuals had access to
resources such as the
internet, computer,
television, and radio,
which aided in their job
search process. This
variable was categorized as
either "Yes" or "No,"
indicating whether or not
1 if Yes; and
0 if No
21
Variable
Operation Definition
Measurement
the individual had access
to these resources.
3.5 Diagnostics Tests
Diagnostic tests were used to specify the stochastic properties of the model. Its significance was
to validate the estimation results of the parameters obtained in empirical analysis. Diagnostic tests
were carried out in the study to check efficiency of the coefficient estimates and the
comprehensiveness of the data and information collected.
3.5.1 The Hausman-McFadden test / IIA Assumption
The test was used to verify whether the Independence of Irrelevant Alternatives (IIA) assumption
holds in a logit model. For example, if a young Kenyan preferred registering at an employment
agency to waiting on the street, this choice would have remained consistent even if additional job
seeking options, such as inquiring at workplaces, were added or withdrawn.
This IIA assumption proposed that the relative probabilities between any two search techniques
remained similar, regardless of the presence or absence of other methods, in the study looking at
the determinants of methods of searching for jobs among young Kenyans. Ensuring the validity
of IIA assumption in the context of this study was critical. If the assumption was violated, the
relative preference between two methods might have been influenced by other available choices,
which could lead to misleading interpretations of the data. To empirically validate the IIA
assumption for the data collected, the Hausman test was employed. This method compared the
estimated coefficients from a reduced model (excluding one of the alternatives) to those from the
full model. A significant discrepancy between the coefficients could have indicated a violation
of the IIA assumption (Hausman, 1978; Greene, 2018).
The test was used if one of the null hypotheses of the model gave efficient and consistent results
and the other, inefficient but consistent and under the alternative hypothesis the first model gives
inconsistent results and second consistent. The Hausman test results offered insights into the
validity of the IIA assumption for this dataset. Ensuring its validity was crucial as it affirmed the
appropriateness of the Multinomial logit model in analyzing the determinants influencing Kenyan
youth's job search methods.
22
CHAPTER FOUR: EMPIRICAL RESULTS
The study aimed at finding out the extent to which age, gender, education level, duration of job
search, access to information and location affected the young people in their choice for job search
methods. To allow for the modeling of choices among more than two alternatives, the study used
the Multinomial logit model to capture the complexity of decision-making and gain insights into
the factors influencing young people's decisions. The Study identified five job search methods
that is: registering at employment agency, enquiring at workplaces, placing or answering a job
advert, referral from relatives, and waiting at street-side. The model was preferred due to its
ability to be used where the study has more than two groups. The categories are; π1, π2,…,πJ
where π1+π2+…+πJ=1.
The Multinomial logit model was used to estimate the probability of choosing each job search
method based on the values of the independent variables.
A descriptive summary of the job search methods used by the Kenyan youth was analysed as
presented in Table 4.1 below.
Table 4.1: A descriptive summary of the job search methods used by Kenyan youths
Job
search
methods
Registered at
employment
agency
Enquired at
workplaces
Placed or
answered a
Job advert
Referral
from
relatives
Waited at
streets-
side
Total
Freq.
15
160
47
47
36
305
Percent
4.92
52.46
15.41
15.41
11.8
100
Source: Author (2023), Data from the KNBS KHIBS 2015/206 Reports
Table 5.2: A descriptive summary of the Covariates by Job Search Channel
Covariates
Job Search Methods
Register
Employment
Agency
Enquired
at
Workplaces
Placed or
Answered a
Job advert
Referral
from
Relatives
Waited at
Steet-side
Information
Access
Yes
15
142
47
44
34
No
18
3
2
Gender
Male
8
64
22
23
6
Female
7
96
25
24
30
23
Covariates
Job Search Methods
Register
Employment
Agency
Enquired
at
Workplaces
Placed or
Answered a
Job advert
Referral
from
Relatives
Waited at
Steet-side
Search Duration
1-24(Months)
12
130
42
42
30
25+ (Months)
3
30
5
5
6
Location
Urban
7
90
18
29
24
Rural
8
70
29
18
12
Age
15-19
4
39
9
12
11
20-24
1
58
19
17
15
25-29
7
25
14
8
4
30-34
3
38
5
10
6
Education level
Primary
2
77
1
22
14
Secondary
11
73
44
17
11
Vocational
&University
2
10
2
8
11
Source: Author (2023), Data from the KNBS KHIBS 2015/206 Reports
A total of 305 respondents were surveyed and categorized into five groups based on their job
search methods. The majority of respondents (52.46%) reported enquiring at workplaces as their
preferred job search method, followed by similar proportions of 15.41% from referrals from
relatives or placed/answered job adverts. 11.80% of the young people preferred waiting at the
street-side (11.80%) while 4.92% reported registering at an employment agency. These findings
however contradict study by Wambugu et al. (2012) which studied the behaviour of job search
personnel in Kenya and the results showed that 83.15% of active job seekers prefer referrals to
land jobs. This therefore shows that preference of job search methods of young people in the
labour market is different from the rest of the working-age population. In fact, Holzer (1988)
argues that youth face distinct hurdles in job search owing to limited experience and networks
which are important when using referrals as strategy for looking for work.
The study concluded that inquiring at workplaces was the most commonly used job search
method among the youth in both urban and rural Kenya, and thus, chosen as the reference
category for the multinomial logit model to allow for comparison with the other job search
methods.
Multinomial logit regression models were estimated to compare each of the four job search
methods (registered at an employment agency, placed or answered a job advert, referral from
relatives, and waited at street-side) with the reference category (enquiring at workplaces). The
24
purpose of each model was to determine which variables were most significant in predicting the
choice of a particular job search method, by comparing each of the four job search methods with
the reference category.
The fitted model from the collected data was according to tables 4.3 and 4.4 below.
Table 6.3: Model Summary
Number of observations
305
LR chi2(36)
133.25
Probability (p-value) of chi-squared test
0.0000
Log likelihood
-334.50162
Pseudo R-squared
0.1661
Table 4.3 reveals that the Multinomial Logistic Regression model was based on 305 observations.
A likelihood ratio test was utilized to assess the overall significance of the model's coefficients.
In this evaluation, the chi-squared value for the likelihood ratio was 133.25, and with a p-value
of p < 0.0001, this confirms that the model is significant compared to one without any predictors.
The Hausman-McFadden test was conducted to verify the Independence of Irrelevant
Alternatives (IIA) assumption. This assumption posits that the odds ratio between any two
choices remains constant, regardless of the presence or absence of other choices. Table 4.5
provides a quick overview of the Hausman-McFadden Test results. The rank of the differenced
variance matrix was 18, indicating adequate variability for the test. The chi-squared statistic was
computed to be 21.43, and with a corresponding p-value of 4.23%, this provides strong evidence
against the null hypothesis at a 5% significance level. This indicates that the IIA assumption
holds, suggesting that the model was reliable.
25
Table 7.4: Summary of the Hausman-McFadden Test Results
Test Component
Hypothetical Figures
Test Used
Hausman-McFadden Test (IIA Assumption)
Differenced Variance Matrix
Rank: 18
Number of Coefficients
28
Coefficients Difference
Not systematic across parameters
Chi-Squared Statistic
21.43
Probability (prob>chi2)
4.23%
Implication
IIA Assumption holds. Model is reliable
Table 8.5: Parameter Estimates
Job_search_methods
Coef. Std. Err.
z
P>z
[95% Conf.
Interval]
Registered at Employment Agency
Information Access
YES
14.93433 1237.867
0.01
0.990
-2411.24
2441.109
Gender
male
.5315426 .5799855
0.92
0.359
-.6052081
1.668293
Search Duration
1-24 Months
-.0685005 .7754233
-0.09
0.930
-1.588302
1.451301
Location
Urban
-.1609385 .5847157
-0.28
0.783
-1.30696
.9850833
Age Group
20-24
-2.390722 1.168971
-2.05
0.041
-4.681863
-.0995811
25-29
.4859682 .7546229
0.64
0.520
-.9930654
1.965002
30-34
-.670752 .8814412
-0.76
0.447
-2.398345
1.056841
Education level
Secondary
2.027606 .8126857
2.49
0.013
.4347714
3.620441
Vocational & University
2.582132
1.133922
2.28
0.023
.3596848
4.804579
_cons
-18.30704 1237.867
-0.01
0.988
-2444.482
2407.868
Enquired at Workplaces
(base outcome)
26
Placed or answered a Job advert
Information Access
YES
14.5373 693.7464
0.02
0.983
-1345.181
1374.255
Gender
Male
.2400481 .3785109
0.63
0.526
-.5018196
.9819159
Search Duration
1-24 Months
.7823673 .5766097
1.36
0.175
-.347767
1.912502
Location
Urban
-.5402204 .3789158
-1.43
0.154
-1.282882
.202441
Age Group
20-24
-.2457495 .5255553
-0.47
0.640
-1.275819
.78432
25-29
.4892397 .5951034
0.82
0.411
-.6771415
1.655621
30-34
-.7472602 .6839635
-1.09
0.275
-2.087804
.5932837
Education level
Secondary
3.910425 1.035362
3.78
0.000
1.881152
5.939697
Vocational & University
3.145133 1.311396
2.40
0.016
.5748432
5.715422
_cons
-19.25466 693.7474
-0.03
0.978
-1378.975
1340.465
Referral from relatives
Information Access
YES
.9899365 .7062231
1.40
0.161
-.3942353
2.374108
Gender
male
.2596501 .3452704
0.75
0.452
-.4170675
.9363678
Search Duration
1-24 Months
.5635924 .5288047
1.07
0.287
-.4728457
1.600031
Location
Urban
.2331534 .3515387
0.66
0.507
-.4558498
.9221566
Age Group
20-24
-.0556122 .4598546
-0.12
0.904
-.9569105
.8456862
25-29
-.1514989 .5443371
-0.28
0.781
-1.21838
.9153822
30-34
-.2430015 .5155216
-0.47
0.637
-1.253405
.7674022
Education level
Secondary
-.1832668 .3792258
-0.48
0.629
-.9265358
.5600022
Vocational & University
1.249598 .5782732
2.16
0.031
.1162031
2.382993
_cons
-2.824825 .9012379
-3.13
0.002
-4.591219
-1.058431
Waited at street side
Information Access
YES
2.028647 .8969869
2.26
0.024
.2705854
3.78671
Gender
male
-1.484188 .5120318
-2.90
0.004
-2.487751
-.4806237
Search Duration
27
1-24 Months
-.1720675 .5498805
-0.31
0.754
-1.249813
.9056785
Location
Urban
.4827786 .4207943
1.15
0.251
-.341963
1.30752
Age Group
20-24
-.2095032 .5237526
-0.40
0.689
-1.236039
.817033
25-29
-1.309353 .7358824
-1.78
0.075
-2.751656
.1329502
30-34
-1.434723 .6697688
-2.14
0.032
-2.747446
-.1220003
Education level
Secondary
-.2502158 .4790723
-0.52
0.601
-1.18918
.6887487
Vocational & University
2.482332 .6374112
3.89
0.000
1.233029
3.731635
_cons
-2.88062 1.035358
-2.78
0.005
-4.909884
-.851356
Table 9.6: Relative Risk Ratios (RRR)
Jobs_search_methods
RRR
Std. Err.
z
P>z
[95%Conf.
Interval]
Registered at Employment Agency
Information Access
YES
3061224
3.79e+09
0.01
0.990
0
.
Gender
Male
1.701555
.9868774
0.92
0.359
.5459608
5.30311
Search Duration
1-24 Months
.933793
.7240849
-0.09
0.930
.2042721
4.268666
Location
Urban
.8513444
.4977945
-0.28
0.783
.2706415
2.678035
Age Group
20-24
.0915636
.1070351
-2.05
0.041
.0092617
.9052166
25-29
1.625748
1.226827
0.64
0.520
.3704394
7.134926
30-34
.5113239
.450702
-0.76
0.447
.0908682
2.877267
Education level
Secondary
7.595881
6.173064
2.49
0.013
1.54461
37.35403
Vocational & University
13.2253
14.99646
2.28
0.023
1.432878
122.068
_cons
1.12e-08
.0000139
-0.01
0.988
0
.
Enquired at Workplaces
(base outcome)
Placed or answered a Job advert
Information Access
YES
2058123
1.43e+09
0.02
0.983
0
.
Gender
Male
1.27131
.4812048
0.63
0.526
.605428
2.669566
Search Duration
1-24 Months
2.186642
1.260839
1.36
0.175
.7062634
6.770003
Location
Urban
.5826199
.2207639
-1.43
0.154
.2772372
1.224388
Age Group
20-24
.7821181
.4110463
-0.47
0.640
.2792022
2.190917
28
25-29 1.631076
.9706587 0.82 0.411 .5080672 5.23633
30-34
.4736625
.3239679
-1.09
0.275
.123959
1.809922
Education level
Secondary
49.92014
51.68543
3.78
0.000
6.561057
379.82
Vocational & University
23.22276
30.45424
2.40
0.016
1.776852
303.5123
_cons
4.34e-09
3.01e-06
-0.03
0.978
0
.
Referral from relatives
Information Access
YES
2.691064
1.900491
1.40
0.161
.6741954
10.74143
Gender
Male
1.296476
.447635
0.75
0.452
.6589764
2.5507
Search Duration
1-24 Months
1.756973
.9290955
1.07
0.287
.6232262
4.953184
Location
Urban
1.262575
.443844
0.66
0.507
.633909
2.514708
Age Group
20-24
.9459059
.4349791
-0.12
0.904
.3840777
2.329576
25-29
.8594188
.4678136
-0.28
0.781
.2957088
2.49773
30-34
.7842703
.4043083
-0.47
0.637
.2855308
2.154163
Education level
Secondary
.832546
.315723
-0.48
0.629
.3959229
1.750676
Vocational & University
3.48894
2.01756
2.16
0.031
1.123224
10.83729
_cons
.059319
.0534606
-3.13
0.002
.0101405
.3469998
Waited at street side
Information Access
YES
Gender
7.603795
6.820505
2.26
0.024
1.310731
44.11101
Male
.2266864
.1160707
-2.90
0.004
.0830966
.6183976
Search Duration
1-24 Months
.8419223
.4629567
-0.31
0.754
.2865582
2.47361
Location
Urban
1.620571
.681927
1.15
0.251
.7103745
3.696994
Age Group
20-24
.810987
.4247566
-0.40
0.689
.2905326
2.263773
25-29
.2699947
.1986844
-1.78
0.075
.0638221
1.142193
30-34
.2381813
.1595264
-2.14
0.032
.0640913
.8851481
Education level
Secondary
.7786327
.3730214
-0.52
0.601
.3044707
1.991222
Vocational & University
11.96915
7.629268
3.89
0.000
3.431609
41.74732
_cons
.0561
.0580835
-2.78
0.005
.0073733
.4268358
Table 4.5 provides the coefficient estimates, standard errors, z-values, p-values, and 95%
confidence intervals for each of the explanatory variables for each job search method relative to
the reference category of enquiring at the workplace. The log odds (coefficients) provide a
summary of the likelihood of unemployed youth choosing the "registered at an employment
agency" method over the reference category.
29
Table 4.6 provides for the relative risk ratios (rrr) of choosing other job search options to the
reference group of enquiring at workplaces. According to Mc Fadden 1984, the choice
probabilities which are dependent on a set of alternatives are non-negative and sum to one. Thus,
when the relative risk is greater than 1 (RR>1) then the probability of the event is more likely to
occur but when the relative risk is less than 1 (RR<1) then the probability of the event to occur is
less likely.
Below are the findings from the research study on how various determinants affect the job search
method of a young Kenyan seeking for work.
Registered at Employment Agency
The level of education of the youth was statistically significant in their preference for registering
at employment agencies over enquiring at workplace with p-value less than 0.05. The relative
risk for choosing registration at work places over enquiring at workplaces for youth with
secondary education is 7.6 and those with university is 13.2 relative to those with primary
education. A similar study by Weber and Mahringer (2008) show that different methods are used
by the youth in searching for employment and their success was dependent on the level of
education as well as skill set. The Youth aged 20-24 years was also statistically significant in
their preference for registering at employment agencies over enquiring at workplace, with a
relative risk of 0.1 compared to those between 15-19 years.
Youth with access to information would 14.93 times more prefer registering at an employment
agency than inquiring at workplaces, compared to those without such access. The relative risk is
greater for the youth with information access in opting for registering at workplaces than those
without. The males were found to be 0.532 times more likely than females to use register at an
employment agency to enquire at workplaces. The relative risk of the male in choosing registering
at employment agency over enquiring at workplaces is 1.7 compared to the young females.
However, the statistical analysis however indicates no significance level with a p-value greater
than 0.05. The duration of job search, gender and the youth aged 25 and above years were not
statistically significant.
Placed or Answered a Job Advert
Only the level of education was statistically significant when it comes to youths answering a job
advert compared to enquiring at workplaces. Youths with university education were 3.145 times
more likely than those with primary education to answer a job advert than enquiring at work
30
places. The relative risk of answering a job advert to enquiring at work places was 49.9 for
Secondary and 23.2 for university compared to the those with primary education. This indicates
that the level of education is critical in answering job advertisements during pursuits of seeking
for employment by young persons. Explanatory variables on access to information, location, age,
gender, and those with vocational training were not significant in their association with placing
or answering an advert compared to enquiring at workplaces. The young males were 0.240 less
likely than the young females to answer an advert compared to enquiring at workplaces.
Referrals from Relatives
Education level at vocational and university is the only variable statistically significant with a p-
value of 0.031, for young persons to use referral from relatives than to enquiring at workplaces
with a relative risk of 3.5 compared to those with primary education level. This shows that at the
university level, the young person would have started to create a network of friends and
acquaintances. This is line with Gronovetter (1983) who argues that acquaintances ties act as
bridges between people’s friendship networks and connections link creating large social systems
that grant a person access to information from distant parts. Also, Studies by Holzer (1988) have
shown that referrals are important for effective job search and that youth face distinct hurdles in
job search owing to limited experience and networks.
Gender, age, location, access to information, and search duration were not significant with p-
values greater that 0.05 when it came to choosing referral from relatives compared to enquiring
at workplaces.
Waited at the Street Side
Access to Information, Gender, youth at the age group 30-34, and those with university &
vocational level of education were statistically significant with a p-value less than 0.05. The
relative risk for waiting at a street side over enquiring at workplace was 7.6 for those with access
to information to those with no access. The youth with university education, the relative risk in
opting at a street side over enquiring at workplaces was 11.97 relative to those with primary
education.
The young males were 1.484 less likely to wait at the street sides than enquire at workplaces, with
a relative risk ratio of 0.23 compared to the female youth. This finding is consistent with the study
by Sacky and Osei (2006) who analysed the probability of male and female job seekers in
informal methods and findings revealed that the male is more likely to use informal techniques.
31
The duration of searching for work, location, access to information, age groups between 20-29,
and secondary education level were not statistically significant in waiting at the street side
compared to enquiring at workplaces.
32
CHAPTER FIVE: CONCLUSION AND RECOMMENDATION
The objective of the study was to determine the incidence of methods used by Kenyan youth in
search of employment and to identify demographic and socioeconomic factors associated with
use of different job search methods among Kenyan youths. The study used the multinomial logit
model to analyse the variable given its ability to investigate more than two categories.
The results show that 52.46% of the young people preferred enquiring at workplaces compared
to other job search methods. Further, the results indicate that the level of education was
statistically significant for the young people who either registered at workplaces or answered job
adverts, with vocational or University education level also being statistically significant for those
choosing referrals or waiting at the street-side. These findings are consistent with Weber and
Mahringer (2008) who carried out research to establish the relationship linking job search options
among youth in Austria, where the different methods used by the youth in searching for
employment and their success was dependent on the level of education as well as skill set.
Gender and access to information were significant factors associated with youth preferring to wait
at street sides when looking for work. Young people between the age of 20-24 was significant
factor for registering at Employment Agencies while those between 30-34 was significant for
those opting to wait at the street sides. Location either rural or urban and the age group between
25-29 had no significance to any of the four job search methods compared to enquiring at
workplaces. The relative risk of the young people in choosing other job search strategies over
inquiring at workplaces was higher for young people with access to information and higher
education levels. This indicates information access and education level is important in
determining job search strategies by the young women and men.
From the findings of the descriptive analysis, only 4.92% of the young people registered at an
employment agency during their search for work. Therefore, this research recommends the
government through the public employment agencies such as National Employment Authority to
enhance accessibility of the existing county employment offices and the NEAIMS online portal
to the young men and women in the country. The National Employment Authority Intergrated
Management System is the official government portal for registering all jobseekers in the country
and declaration of job vacancies by all employers in the country as per NEA Act,2016.
With the Government of Kenya plan to create a million new jobs annually either locally or abroad,
policy makers and the private recruitment and/or employment agencies based on the research
findings need to develop strategies on how best to reduce the cost of job search for young people
33
especially when it comes to enquiring at workplaces as it’s their most preferred mode of job
search followed by referrals and answering job adverts.
Private Employment Agencies (PEAs) play an important role in supporting the Government in
the recruitment and placement of jobseekers; therefore, the findings of this research provide PEAs
and employers in general in developing efficient recruitment strategies for the young people to
ease their job search and enhance their participation in the labour market.
This research shows entry of young people into the labour market is highly dependent on their
job search strategies, therefore providing an area for further study by other researchers in studying
the behaviour of the young men and women in their entry and participation in the labour market.
34
REFERENCES
Africa Union. (2023). Youth Development 2023. Retrieved from https://au.int/sw/node/35020
Blau, D. M., & Robins, P. K. (1990). Job search outcomes for the employed and unemployed.
Journal of Political Economy, 3, 637–655.
Betcherman, G., Godfrey, M., Puerto, S., Rother, F. & Stavreska A. (2007). A Review of
Interventions to Support Young Workers: Findings of the Youth Employment Inventory, World
Bank Social Protections (SP) Discussion Paper No. 0715, Washington DC: World Bank.
Caliendo, M., Schmidl, R., & Uhlendorff, A. (2011). Social networks, job search methods and
reservation wages: Evidence for Germany. International Journal of Manpower, 32(7), 796–824.
Escudero, V. & López Mourelo, E. (2013). Understanding the drivers of the youth labour market
in Kenya. ILO Research Paper, 8. Geneva: ILO.
Fernandez, Roberto & Castilla, Emilio & Moore, Paul. (2000). Social Capital at Work: Networks
and Employment at a Phone Center. American Journal of Sociology - AMER J SOCIOL. 105.
1288-1356. 10.1086/210432.
Faberman, R. J. Mueller, A. & Sahin, Aysegul. (2022). Has the Willingness to Work Fallen
During the COVID Pandemic?. SSRN Electronic Journal. 10.2139/ssrn.4114632.
Feldman, D.C. & Ng, T.W.H. (2007) Careers: Mobility, Embeddedness, and Success. Journal of
Management, 33, 350-377. http://dx.doi.org/10.1177/0149206307300815
Feldman, D. & Bolino, M. (2000). Career Patterns of the Self-Employed: Career Motivations and
Career Outcomes. Journal of Small Business Management. 38.
Furlong, A. & Cartmel, F. (2006). Young People and Social Change: New Perspectives.
García, C. G. & Nicodemo, C. (2013). Job Search Channels, Neighborhood Effects and Wages
Inequality in Developing Countries: The Colombian Case. IZA Discussion Paper No. 7336,
Available at
SSRN: https://ssrn.com/abstract=2254189 or http://dx.doi.org/10.2139/ssrn.2254189
Granovetter, M. S. (1973). The strength of weak ties. American Journal of Sociology, 78, 1360–
1380.
Government of Kenya (2023). The Bottom-Up Economic Transformation Agenda (BETA 2022-
2027).
Government of Kenya (2023). Labour and Employment Sector Plan 2023-2027. Ministry of
Labour and Social Protection.
Government of Kenya (2023). Sessional Paper No. 5 of 2023, National Policy on Labour
Migration. Ministry of Labour and Social Protection.
Government of Kenya. (2023). Economic Survey. Kenya National Bureau of Statistics
35
Government of Kenya. (2019). Kenya population and housing census. Kenya National Bureau of
Statistics
Government of Kenya (2019-2022). Kenya Quarterly Labour Force Survey Reports. Kenya
National Bureau of Statistics
Government of Kenya (2018). Labour Force Basic Report. 2015/16 Kenya Integrated Household
Budget Survey (KIHBS) Reports. Nairobi, Kenya: Kenya National Bureau of Statistics.
Government of Kenya (1989-2019). Kenya Labour Force Reports. Kenya National Bureau of
Statistics.
Greene, W.H. (2012), Econometric Analysis. New York: Pearson Education Limited.
Green, A. E., De Hoyos, M., Li, Y., & Owen, D. (2011). Job search study: Literature review and
analysis of the Labour Force Survey. Department for Work and Pensions Research Report, 726.
Guillemyn, I. & Horemans, J. (2023). Age-Related Differences in Job Search Behavior: Do Older
Jobseekers Need a Larger Social Network?. Work, Aging and Retirement.
10.1093/workar/waad003.
Hausman, J. & McFadden, D. (1984) Specification Tests for the Multinomial Logit Model.
Econometrica, 52, 1219-1240. http://dx.doi.org/10.2307/1910997
Holzer, H. J. (1988). Search method use by unemployed youth. Journal of Labor Economics,
6(1), 1–20.
International Labour Organization (ILO). (2023). World Employment and Social Outlook: Trends
2023. Geneva: International Labour Office, 2023.
Ioannides, Y. & Loury, L. (2004). Job Information Networks, Neighborhood Effects, and
Inequality. Journal of Economic Literature. 42. 1056-1093. 10.1257/0022051043004595.
Kuhn, P. & Skuterud, M. (2004). Internet Job Search and Unemployment Durations. American
Economic Review, 94 (1): 218-232.
Martin, J. P. & Grubb, D. (2001). What Works and for Whom: A Review of OECD Countries'
Experiences with Active Labour Market Policies. Swedish Economic Policy Review, 8(2),9-56.
Mlatsheni, C. & Rospabé, S. (2002). Why is youth unemployment so high and unequally spread
in South
Africa?, Development Policy Research Unit Working Paper 02/65 (Cape Town, University of
CapeTown).
Mouw, T. (2003). Social Capital and Finding a Job: Do Contacts Matter? American Sociological
Review, 68(6), 868–898. https://doi.org/10.2307/1519749
Mussida, C., & Zanin, L. (2020). Determinants of the Choice of Job Search Channels by the
Unemployed Using a Multivariate Probit Model. Social indicators research, 152(1), 369–420.
https://doi.org/10.1007/s11205-020-02439-z
36
Okafor, C.E. (2011). The Role of Vocational and Technical Education in Manpower
Development and Job Creation in Nigeria. Journal of Research and Development, 2, 152-159.
Osberg, L. (1993). Fishing in Different Pools: Job Search Strategies and Job-Finding Success in
Canada in the Early 1980s. Journal of Labour Economics, Vol. 11, (2): 348-86.
Rees, A. (1966). Information Networks in Labour Markets. American Economics Review, 56:
559-66.
Sackey, H., & Osei, B. (2006). Human Resource Underutilization in an Era of Poverty Reduction:
An Analysis of Unemployment and Underemployment in Ghana. http://lst-iiep.iiep-
unesco.org/cgi-bin/wwwi32.exe/[in=epidoc1.in]/?t2000=023878/(100). 18. 10.1111/j.1467-
8268.2006.00140.x.
Schaer, V. & Leibbrandt, M. (2006). Determinants of Job Search Strategies: Evidence from
Khayelitsha/Mitchell's Plain Survey. South African Journal of Economics 74: 4: 702-24.
Suvankulov, Farrukh & Lau, Chi Keung & Chau, Frankie. (2012). Job search on the internet and
its outcome. Internet Research. 22. 298-317. 10.1108/10662241211235662.
UN DESA. (2022). The Sustainable Development Goals Report 2022 - July 2022. New York,
USA: UN DESA. © UN DESA. https://unstats.un.org/sdgs/report/2022/
van Hooft, E. A., Kammeyer-Mueller, J. D., Wanberg, C. R., Kanfer, R., & Basbug, G. (2021).
Job search and employment success: A quantitative review and future research agenda. Journal
of Applied Psychology, 106(5), 674.
Viitanen, T.K. (1999). Estimating the Probability of Youth Unemployment: An Extended Essay
for a BSocSc in Economics and Statistics.
Wahba, J., & Zenou, Y. (2005): Density, social networks and job search methods: Theory and
Applications to Egypt. Journal of Development Economics, 78, 443–473.
Wambugu, A., Onsomu, E., & Munga, B. (2012). Discussion Paper No. 130 of 2012 on Job
Search in Developing Countries: Empirical Evidence from Kenya.
Weber, A., & Mahringer, H. (2008). Choice and success of job search methods. Empirical
Economics, 35, 153–178.
World Bank (2010). Active Labour Market Programmes for Youth. A Framework to Guide Youth
Employment Interventions, World Bank Employment Policy Primer No. 16.
World Bank (2010). Demographic Transition and Growth in Kenya. World Bank Opinion.
Retrieved from https://www.worldbank.org/en/news/opinion/2010/04/28/demographic-
transition-growth-kenya.
37
