1 Do They Look for Informal Jobs? Migration of the Working Age in Indonesia Elda L. Pardede Faculty of Economics University of Indonesia [email protected]Rachmanina Listya PT. Bank BNI Syariah Abstract It is widely accepted assumption that migrants in developing countries who are not absorbed by the modern sector may enter informal sector as a transitional phase. The purpose of this study is to assess whether this assumption is valid for Indonesian case using the Indonesia Family Life Survey 2007 data. By taking advantage of the longitudinal data of employment and migration histories of 15 years and old individuals, multinomial regression with correlated random intercept is employed to study whether migrants are more likely to work informal sector than non migrants and to what extent can working in informal sector be linked with migration motives. The results show that migrants in Indonesia are more likely to work in formal sector than non migrants. In connection with migration motives, migrants with work-related motives are less likely to work in informal sector than migrants with motives related to family-related reasons. Among migrants with work-related motives, those who move due to being unemployed and who lack of employment opportunity in origin locations have higher likelihood to work in informal sector than formal sector compared with those who moved for other job-related reasons such as job transfers and closer to job. Paper prepared for the 27th IUSSP International Population Conference, Busan, Korea, 26-31 August, 2013.
23
Embed
Do They Look for Informal Jobs? Migration of the Working … · 2013-08-11 · 1 Do They Look for Informal Jobs? Migration of the Working Age in Indonesia Elda L. Pardede Faculty
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
It is widely accepted assumption that migrants in developing countries who are not absorbed
by the modern sector may enter informal sector as a transitional phase. The purpose of this
study is to assess whether this assumption is valid for Indonesian case using the Indonesia
Family Life Survey 2007 data. By taking advantage of the longitudinal data of employment
and migration histories of 15 years and old individuals, multinomial regression with
correlated random intercept is employed to study whether migrants are more likely to work
informal sector than non migrants and to what extent can working in informal sector be
linked with migration motives. The results show that migrants in Indonesia are more likely to
work in formal sector than non migrants. In connection with migration motives, migrants
with work-related motives are less likely to work in informal sector than migrants with
motives related to family-related reasons. Among migrants with work-related motives, those
who move due to being unemployed and who lack of employment opportunity in origin
locations have higher likelihood to work in informal sector than formal sector compared with
those who moved for other job-related reasons such as job transfers and closer to job.
Paper prepared for the 27th IUSSP International Population Conference, Busan, Korea, 26-31 August,
2013.
2
Introduction
Indonesia has been characterized by high proportion of informal jobs in its economy
within the last decade, which is around 60 to 70 percent1 (BPS). This high proportion of
informal jobs is one of the characteristics of many developing countries, especially in urban
areas (Todaro & Smith, 2006). High propotion of informal jobs in developing countries is due
to the inability of the formal economy to absorb excessive labor supply as a result of high rate
population growth (Handayani, 1991; Sahlan, 1996). In other words, economic growth fails
to create adequate job opportunities in formal modern sector for the growing supply of
working age population. Within the demand side of labor market, inadequate formal jobs
could also be explained by institutional factors such as tight regulations and rigidity in formal
sectors (formal contracts, workers’ protection, pension, etc) that may have induced
employers’ reluctance to hire formal workers (Nazara, 2010).
Growing supply of working age is usually attributed to migration of labors in addition
to population growth within an area, especially rural to urban migration. Rural-urban
migration that increases the urban pool accompanied by insufficient creation of modern
sector opportunities, may further enhance the growth of informal economy. It is even stated
that informal sector exist and is growing in many large cities in developing countries is
largely due to rural-urban migration (Cole & Sanders, 1985, p.482). Within the framework of
job search for rural-urban migration process by Todaro (1969), working in informal economy
in developing countries is viewed as a transitional phase for migrants before entry into the
formal sector. Although this view has been challenged by some because some people actually
move to expect long-term employment in informal sector (Cole & Sanders, 1985) and
because informal sector is not considered as solely ‘traditional’ but also a dynamic part of a
larger economy (GTZ and Bappenas, 2008, cf. Nazara, 2010), it is still worthy to investigate
whether migrants in Indonesia have higher tendency to work in informal sector (vis-á-vis
formal sector) than non migrants.
1The proportion around 70% is calculated by authors based on the proportion of workers whose main job status are self-employed, self-employed with unpaid and/or irregular workers, casual employee in agriculture and in non-agriculture sector, and family workers/unpaid workers. The data is the population of Indonesia 15 years and older based on the status of main job, 2004 – 2013 from Indonesia Labor Force Surveys published by Indonesia National Statistic Office or Biro Pusat Statistik, BPS (www.bps.go.id). Although the proportion of informal workers is lower using the most recent definition of informal sector used by BPS (Badan Pusat Statistik, 2012) as also pointed by Nazara (2010), the number is still considerably high, around 60%. The definition of informal sector that will be used in this study is similar with the number calculated by authors, which would be explained in the later section of this paper.
3
Considering the domination of informal sector in Indonesia and assumption of its role
as a transitional phase for migrants, it is plausible to presume that there are relatively high
proportion of migrants in Indonesia, especially the job seekers, who move and then work in
informal economy, either for temporary phase or intentionally for the whole period of stay.
The first objective of this study is thus to investigate the validity of this assumption,
especially given that, different than what one might expect, several studies have pointed out
that the probability of migrants to work in informal sector compared to formal sector is lower
than that for non-migrants in West Java and urban areas in Indonesia (Sahlan, 1996; Syahran
P., 2000). One of the contributions of this study is to use the more recent data, which is the
Indonesia Family Life Survey 2007 (IFLS4) by taking advantage its records on adults’
migration history and job history for the analysis. The availibility of longitudinal data
provides opportunity to perform panel data method, especially to include time-varying
explanatory variables in the model.
Although informal sector can be considered as an integral part that gives benefits to
overall economy and in some cases it perhaps has higher returns than formal sector (Meng,
2001), security including secure jobs is an integral part of human development. Furthermore,
if the concept of decent work would be employed by the policy makers, work with normal
regular hours and adequate earnings, which are usually not the characteristics of informal
sector, has also become the objective of development. Knowing whether migrants are more
likely to work in informal sector than non-migrants or not can be a valuable input for labor
market and human development policies.
The second objective of this study is connected with an important theoretical
rationale. The Todaro model was built in the frame of job search theory, which is not fully
appropriate to explain various migration types that were captured by the IFLSs data.
Theoretically, labors are assumed to move to search for jobs in the center of economic
activities (urban areas) while reasons to migrate can in fact be varied. Various studies on
migration and employment rarely distinguish motivations to migrate as determinants of labor
market outcomes of migrants. Little is known about how decision to migrate can be related to
the tendency to work in informal sector versus the formal sector. In IFLS, people move
because of job-related reasons, marriage, family related reasons, to follow the bread winners,
education, being independent, up to move due to natural disaster. Therefore, the second
objective of this study is to assess the link between working in informal sector or formal
sector and migration process by assessing reasons or motivations to migrate. If economic
reason is the basis of the modelling process of migration of labor, how is it related to the
4
probability to work in informal sector? To what extent can motivations be linked with and
can explain the likelihood of migrants to work in informal sector compared to informal
sector? Finding the answer to this question may improve our understanding on the migration
decision-making process and in what way it is linked to employment outcome. If certain
specific reasons explain the likelihood to work in informal sector, we may understand better
how migration motivations affect the labor market outcome.
Informal Sector, Migration Decision-Making, and Previous Empirical Studies
One of the difficulties of studying the relationship of migration and informal sector
and comparing studies on this subject is that the definition and characteristics of informal
sector is much more complex than the usual measurements employed to capture it in most of
Labor Force Surveys (LFSs). For instance, Todaro and Smith (2006) describe the definiton
and characteristics of informal sector as:
“a large number of small-scale production and service activities that are usually
individually or family-owned using simple and labor-intensive technology; tend to operate like
monopolistically competitive firms with ease of entry, excess capacity, and competition driving
profits down to the average supply price of labor of potential new entrants; the workers usually
have little formal education, are generally unskilled, and lack access to financial capital;
therefore their productivity tend to be lower in the informal sector than in the formal sector and
they do not enjoy protection afforded by formal modern sector in terms of job security, decent
working conditions, and old-age pensions” (pp. 328-329)
However, as Nazara points out (2010, cf. GTZ and Bappenas, 2008), informal sector may
consist of economic activities that continuously exist and expand alongside and perhaps with
established connection with the formal sector and even some segments are probably very
dynamic and innovatively adapt to market condition and thus giving high returns to its
workers, especially the self-employed.
These contradicting views on informal sector may be related to two conditions that
influence the ups and downs of informal sector described by Flórez (2003). The first
condition is that informal sector is a result of pro-cyclical behavior, which implies that
informal sector is highly integrated with formal sector and influenced by economic boom.
Informal sector is growing and expanding because the economy is growing. The second one
is that the informal activities flourish as a result of counter-cyclical behavior. This type of
informal sector is a subsistence one and act more as a survival strategy than taking
advantage of economic opportunity. It is created due to disadvantaged situation, namely,
5
economic crisis and being laid off (p. 8). In this context, we may simplify and distinguish two
types of relatively opposite conditions faced by individuals that drive them to work in
informal sector, which are taking advantage of opportunity or merely a survival strategy.
The distinction between guided by opportunity or survival strategy in migration and
entry to informal sector is not incorporated in rural-urban migration framework by Todaro
(1969) and Harris and Todaro (1970). They assume that migration occur as a response to
difference in rural-urban expected earnings and that migrants mainly pursue manufacturing
jobs in urban areas. Consequently, urban subsistence (informal) sector is simply a temporary
phase. As pointed by Cole and Sanders (1985), manufacturing jobs require relatively high
level of education and that the urban ‘surplus labor’ pool of potential workers may not
comprise of all workers potential to be recruited in the modern sector. Furthermore, migrants
may wish to enter informal sector with its ‘relative ease of entry’ without considering
possibility to work in formal sector. Therefore, the rates of unemployment faced by migrants
in the formal sector and informal sector needs to be differentiated in order to calculate the
difference in expected rural-urban earnings by taking into account the probability of
employment. While this concept may sound simple, the measurement of this has been proven
somewhat more complex than the theory (Speare & Harris, 1986), especially if we want to
incorporate opportunity and survival strategy. In this case, we may need to integrate detailed
decision-making process as well than.
One of the line of inquiries chosen in this study to incorporate this notion of
opportunity vs survival strategy is not to expand economic modelling or rural-urban
migration such as elaborated by Todaro and others but to link the decision to work in
informal sector with the components of motivations to migrate. Most of the studies of
migration and labor market outcomes consider earnings and other human capital variables but
do not consider whether the elements of motivations may be strongly connected with decision
to work in any sector or occupation. Labor market outcome of migration is usually discussed
as if it is solely related only to labor migration. However, even tied migrants (wives, children,
relatives) who move to follow the main bread-winner may as well enter the labor market and
end up in either formal or informal sector. This is an important consideration as the most
cited reasons to migrate is ‘family’ and ‘economy’ while it should be noted that ‘family’
reason may conceal the economic reason2. Another feature of this study is also that while
2 For example, from the study of Wajdi (2010) using the Indonesia Inter-Censal Survey or SUPAS 2005, it is
found that the most cited reasons for inter-island migrations are first ‘family’ and then followed by ‘economy’
6
most studies on migration and informal sector is usually about rural-urban migration, this
study does not limit the informal sector solely in urban areas because as also stated by Nazara
(2010), informal sector is rural phenomenon in Indonesia. People may move between rural
areas to look for jobs (and perhaps mainly in informal sectors) as well as from rural to urban
areas. In this respect, we do not distinguish migration and its link with informal sector only
for rural-urban migration.
To understand the connection between migration motivations and informal sector, we
need to elaborate the motivations to migrate in migration decision-making. One of the
concepts of motivations that is used here is the one proposed by Mulder (1993). According to
Mulder, the element of motivations in migration decision-making process can be
distinguished into four components as preferences, resources, constraints, and opportunities.
Incorporating this into decision-making process called Value-Expectancy model by De Jong
& Fawcett (1981), preferences component mentioned by Mulder are related to values as they
are the concrete transformation of goals3 (p. 20). Opportunities, resources, and constraints are
the components that form the expectation (expectancy or subjective probability). The seven
categories of values proposed by De Jong and Fawcett (1981, pp. 49-51) are: (1) wealth: a
wide range of factors related to individual economic reward; (2) status: factors related to
social standing or prestige; (3) comfort: achieving better living and working conditions; (4)
stimulation: exposure to pleasurable activity; (5) autonomy: personal freedom or the ability to
live one’s own life; (6) affiliation: being with other persons; and (7) morality: related to
deeply held values and belief system that prescribe good and bad ways of living, such as
religious system. People are postulated to form high intention to move if in sum, the
calculation of expectancy and values is high or in other words if there is high probability of
meeting these values/goals in the destination areas. The migration of labor motivations, if we
assume it as migration for economic reason, thus may fall into but not limited to the
some luxuries, etc), status (having a prestigious job), comfort (having an ‘easy’ job),
autonomy (being economically independent). Other motives that may be highly connected
while studies seem to assume that the reason of migration of labor is main(Wajdi, 2010)ly or even only connected with economic reason.
3 The discussion of the differences between values, goals, motives, motivations, etc. is beyond the scope of this
paper. For simplification, in this study we use motives, values, and motivations interchangably and specifically use the values described by De Jong and Fawcett.
7
with motives purely for work perhaps having comfort housing (comfort) and having power
and influence (status) that can be achieved by having prestigious jobs and high earnings.
After listing these various motives to migrate, our proposition is that the use of these
motives may explain the relationship between decision-making process and working in
informal sector in a more detailed picture than simply considering differentials in expected
earnings. Certain motives may relate to certain types of migration and consequently certain
labor market outcome. Persons who move for family reasons or housing needs probably only
wants to move their dwellings but still maintain their economic activities while persons who
move for jobs or education may consider further distance because they require new activity
pattern (Mulder, 1993, p. 25). Professionals search span to find jobs that suited for them may
cover wider scale than unskilled workers(Willekens, 1985). Refering to the description of
values, people move for affiliation (living near family, friends; being with spouse) may
choose to move in a closer distance than people move for work. In this line of thinking, if
individuals who move for affiliation end up working, they may be more likely to work in
informal sector than the ones who move for working reasons. We may also modify further the
reasons to work by including the differences between moving for opportunity or for necessity
(survival). This will be elaborated further in the method section when explaining the
categorization of motives to migrate using the IFLS4 data.
Given that the studies on the link between motivations to migrate and participations of
individuals in labor market is lacking, review on previous studies on migration and informal
sector for this study is mainly focused on several empirical results. Some studies show the
relationship between migration and characteristics of the migrants and whether the migrants’
jobs belong to formal or informal activity. Study by Koo and Smith (1983) in Manila and
several other major cities in the Philipines show that the proportion of recent migrants in the
informal sector is much higher than non migrants. However, for the city of Manila, when the
duration of staying in destination is raised to seven years, there are no signifficant differences
between migrants and non migrants in their involvement in the informal sector while in other
cities, the distinction remains quite large. It is also found that the pattern of entry to the
formal sector through the informal sector, as stated by some theories on migration, found
only on women migrant workers. Flórez (2003) also find that in Colombia, duration of stay in
destination areas has great effect on the migrant's risk to enter informal sector, formal sector
or to be unemployed. On the contrary for Indonesia, Handayani (1991) finds that there’s no
relationship between migrant status (being recent migrant) and working in informal sector. It
is more connected with other characteristics such as education, age, sex, and marital status.
8
She also finds that married migrants have higher probability to work in informal status than
non-married migrants. Other studies in Indonesia by Sahlan (1996) and Syahran P. (2000)
show that estimated proportion or probability of migrants to work in informal sector is lower
compared to non migrants in urban areas and in West Java, respectively. Furthermore, those
who are more likely to have informal jobs are the older age groups compared with the
younger ones and those with low education, especially among the non migrants. Considering
different findings previously elaborated, this research takes the view of studying this
phenomenon further by employing the most recent IFLS data using the panel historical
information on migration and work of Indonesian adults. In addition, component of
motivations to migrate will also be used in determining the likelihood to work in informal
sector among migrants.
Data and Method
Data used for this research is the 2007 IFLS or IFLS4, which is the fourth wave of
longitudinal surveys that has collected information about various aspects of lives of
Indonesians.The sampling scheme in the first wave in 1993 is the primary determinant of the
samples on the next wave. IFLS4 was held at the end of 2007 up to the beginning of 2008 to
trace the same households in 1993 and their descendants. In IFLS1, the respondents were
distributed in 13 provinces in Indonesia. There were 7,224 households interviewed, and
detailed individual-level data were collected from over 22,000 individuals. These households
and individuals were targeted for IFLS4 with recontact rate for the individual target
households was as high as 90.6% (Strauss, et al, 2009). In the fourth wave, the respondents
were scattered in 21 provinces in Indonesia4.
The observations for this research are individuals who were at least 15 years at any
starting point of observation from 2000-2007, meaning that if a person reached 15 years old
in 2003, that would be the starting point of observation for this person. The cut-off 15 years
old is used because it is the youngest age not only to be included in adult module of IFLS but
also to be included formally as labor force in Indonesia Labor Force Surveys. For this study,
we use two samples. The first one is 174,185 observations of yearly employment record from
22,847 individuals, who are observed in 2000-2007. The dependent variable for this study,
called employment status, is created from employment record. Employment status are
4There were 33 provinces in Indonesia during the 2007 survey. During the first survey, there were still 27 provinces in Indonesia and 13 provinces were chosen as sample. The locations of the respondents in 2007 is highly determined by the choice of provinces in the first wave.
9
divided into working in informal jobs, not working, and working in formal jobs as the
reference category5. As previously discussed, the definition of informal status can be varied,
but for this study the formal-informal status is determined by the use of question of job
status6. Individuals working in a particular year as employer with regularly paid employees,
as private or government employees are grouped as working in formal sector while the rest
are working in informal sector. The main explanatory variable is migration variable, whether
an individual migrate or not in the same year of observation of empoyment status. The
definition of migration in IFLS is movement at least across village level to stay in the
destination for at least six months. Other variables are individual characteristics such as sex,
age, highest level of education completed, marital status, and living in either urban or rural
area at the year of observation. Age, marital status, and area are allowed to be varied across
time while education variable is fixed using the information of level of education in 2007.
The second sample used for this study is 9,909 number of migrations of 1,487
individual migrants observed in 2000-2007. This sample is used to analyze the link between
migration motivations and employment status for migrants. IFLS4 has records on reasons to
migrate, varied from to work-related reasons up to natural or other disaster. The
categorization of these reasons is determined mainly based on the list of values from De Jong
and Fawcett (1981). Special attention should be paid on the way we group work-related
reasons. To get work, job transfer (including the military jobs), retirement, closer to job, and
other work-related reasons are grouped into one reason called ‘other work-related reasons’,
which is the reference category that will be used in regression model. To get work is a
separate category because it is more about looking for opportunity than survival strategy
while looking for work because there is not enough employment in the origin location and
work problem such as being laid off are grouped together as more about survival strategy
than looking for opportunity7.
5Note that the question for employment history in IFLS4 is not as detailed as standard Labor Force Surveys’ questions. The respondents were only asked whether they were working or not in the year observed. There is no additional information to distinguish whether they were out of labor force or unemployed.
6Although the definition of informal sector has been modified (see Nazara, 2010 and BPS, 2012), at this point we still use the earlier definition of informal sector. 7The decision to group these work-related reasons as survival strategy is also based on personal communication about the meaning of the question on reasons to migrate with Christine Peterson, the RAND corporation contact for IFLS data. See Appendix 1 for the categorization of reasons to migrate in IFLS4. The reasons to migrate is grouped mainly based on Value-Expectancy model from De Jong and Fawcett (1981) with some additional categories.
10
Because there is relatively high percentage of migrations for marriage, it is kept
separated from other reasons related to affiliation (living near family/friends, being with
spouse) and other moves following families/relatives/friends are considered as move for
affiliation. Events that occur within one’s household that trigger migration such as illness,
pregnancy, death, divorce, family problems, on the other hand, are treated as a separate group
of reasons because they are the trigger of migrations and not really connected with specific
values pursued by migrations. In addition, some reasons that we consider as somewhat
‘forced’ migration such as natural disaster, drought, political disturbance are also grouped
separately because the nature of such migrations is rather different than the rest. The
distribution of employment status by the final categories of motivations can be seen in Table
2 later in cross-tabulation analysis.
For statistical model, multinomial logit models with correlated random intercepts are
used for sample of adults and sample of individuals8. The IFLS4 data used is a longitudinal
data that consists of repeated observations on the same individual at different points in time,
namely the year 2000-2007. The repeated measurements are typically positively correlated
and thus require special methods of analysis beyond those traditionally used for cross-
sectional studies. One way of dealing with correlation among repeated observations for an
individual is to introduce random effects into the regression model (Haynes, et al, 2006).
Given the response variable, employment status consists of three categories, an appropriate
model is the multinomial logit model. Therefore, we utilize multinomial logit models with
correlated random intercepts in this study to capture unobserved heterogeneity among
individuals, i.e., spurious dependence.
Let Yit and Xit denote the t-th observation (t=1,…,T) for the i-th individual. If there are
J possible response states, then Pr(Yit = j | Xit), j=1,…, J, is the probability that individual i
has response j at time t given Xit, a column vector of explanatory variables for that
observation. The multinomial model is expressed as:
���� = Pr�� = �� ��� = ������∑ ����������
The logit model pairs each response category with an arbitrary baseline category. In this
study, the response has three states (J=3), i.e., formal (j=1), informal (j=2) and not working
8 The choice to use random effect is made the χ2 test show that the results of random effect models are
significantly different from the results of fixed effect models.
11
(j=3). We set “formal” as the reference category, so that β1=0. The multinomial logit model
can be written as follows:
��� �������� ! = ��# $�, where j=2,3.
This model assumes that error terms are iid (independent and identically distributed) if each
random variable has the same probability distribution as the others and all are mutually
independent) with homogenous variance, the regression coefficient remains the same for all
individuals i in job status j. If we also introduce individual-specific random effects αij and let
Zij denote a vector of coefficients for the random effects, then the logit model has the form as:
The random effects αi = {αi1,…, αiJ} capture the non-observable individual effects that are
specified to arise from a multivariate normal distribution. In this model, each individual i is
now considered as a cluster observations over time (t=1,…,8). The regression coefficient
remains the same for all individual i in job status j, but a random subject-specific intercept
term has been introduced to account for unobserved heterogeneity among individuals. This
model will be used to assess whether migrants are more likely to work in informal sector than
formal sector compared to non migrants and whether certain reason to migrate is related with
woking in informal sector. In the next section, cross-tabulation analysis of the two samples is
presented before proceeding with regression analysis.
Employment Status by Characteristic of All Adults and All Migrants
The distribution of employment status by individual characteristics is presented in
Table 1. From this cross-tabulation we can see that those who migrate have higher proportion
of individuals who are working in formal sector (50.6%) than the proportion of individuals
who are working in informal sector (30.6%). Those who do not migrate have higher
proportion of working in informal sector (50.6%) than in formal sector (27.6%). Interestingly,
migrants seem to have higher tendency to be not working compared to non migrants. This is
probably due to the fact that most of the migrants move not for work-related reasons. Other
feature of adults 15 years and older for this study is that males have higher percentage of both
working either in formal or informal sectors than females. However, females have higher
tendency to be not working, which confirm the common facts that females in general have
lower labor force participation than males. The finding that males have higher percentage to
12
be working in informal sector compared to males are contradictory with the general findings
from LFSs in Indonesia that usually females are more likely to work in informal
sectors(Nazara, 2010). However, it should be noted that the higher tendency of females to
work in informal sector in LFSs is the result of comparing only the females who are working.
In this sample, the not-working category is included. Therefore, if the non-working sample is
excluded, the percentage of females working in informal sector is consistently higher than the
percentage of males working in informal sector.
Table 1. Distribution of Employment Status by Individual Characteristics of Adults 15 Years and Older in 2000-2007, IFLS 2007
Individual Characteristics Employment Status Total
Formal Informal Not Working % N Migration status
Migrate 50.6 27.6 21.9 100 9,911 Do not migrate 30.6 50.6 18.8 100 164,274
Sex Male 37.2 50.7 12.1 100 93,347 Female 25.5 47.7 26.9 100 80,838
Age Group 15-24 33.3 28.1 38.5 100 49,742 25-44 36.9 51.7 11.4 100 83,365 45+ 19.4 70.1 10.5 100 41,078 Mean value 32.0 39.2 27.8 -
Education No School 8.5 78.9 12.6 100 13,236 Less than Primary School 15.8 72.2 11.9 100 32,020 Primary School 23.2 61.3 15.4 100 40,654 Junior High School 31.6 46.5 21.9 100 26,947 High School 46.6 28.2 25.2 100 45,851 Diploma/University 63.3 12.3 24.4 100 15,477
Marital Status Never Married 38.5 26.7 34.8 100 45,444 Married 30.1 56.8 13.1 100 118,705 Separated/divorced 29.5 56.5 14.0 100 3,654 Widowed 16.2 66.1 17.7 100 6,382
Age group 15-24 49.25 21.21 29.54 100 5,159 25-44 54.07 32.98 12.96 100 4,206 45+ 36.40 45.77 17.83 100 544 Mean value 26.20 29.20 23.80 -
Education No School and Less than Primary School 31.81 51.70 16.49 100 940 Primary School 42.61 38.05 19.34 100 1,758 Junior High School 47.29 33.77 18.94 100 1,922 High School 56.05 20.24 23.70 100 3,759 Diploma/University 62.03 10.78 27.19 100 1,530
Marital Status Never Married 55.71 18.59 25.71 100 3,960 Married 47.72 33.05 19.23 100 5,622 Separated/divorced 36.73 43.81 19.47 100 226 Widowed 40.59 36.63 22.77 100 101
Reason to Migrate To Get Work 70.00 22.25 7.75 100 2,040 Look for Work 56.42 31.05 12.53 100 950 Other Work-Related Reasons 78.04 14.72 7.24 100 428 Marriage 49.57 29.36 21.06 100 1,747 Affiliation 36.91 31.44 31.65 100 2,376 Household Events/Disturbances 37.87 38.96 23.16 100 367 Education/Training 21.61 9.43 68.96 100 509 Disturbances and Other Reasons 38.34 37.38 24.28 100 313 Being Independent, Like the Destination and Housing Reasons 51.15 29.69 19.17 100 1,179
Total 50.59 27.55 21.86 100 9,909
Regression Analysis
The result of multinomial regression with random effect for the first model is
presented in Table 3. To interpete the results, we use the exponential value of the regresson
coefficient (Exp(B)), which is termed as relative risk rasio (RRR). From Table 3 we can see
that adults who migrate are 0.42 less likely to work in informal jobs than adults who do not
15
migrate. This result confirms previous studies on the relationship between migration status
and informal economy in urban areas and in DKI Jakarta, Indonesia (Sahlan, 1996; Syahran
P., 2000). The higher tendency of migrants in Indonesia to work in informal sector compared
to non migrants is not statistically proven, or in other words, the tendency of migrants to enter
the transitional phase in informal jobs before finding formal jobs is not confirmed.
One explanation is that the possibility that less permanent migrations such as circular
migrations is more connected with informal sector. The definition of migration in IFLS,
SUPAS, and Census, which is a move that includes staying in destination for at least six
months, excludes the less permanent moves. There is a possibility that those who move and
then work in informal economy are circular migrants, as revealed in a study by Hugo (1982)
in several provinces in Indonesia. Circular migrants are those who make the shift from the
origin but with less permanent stay than migrants (not to settle in destination). Examples are
farmers who move to urban areas waiting for the harvest and then work in destination as
builders, hawkers and others, and then return when the harvest comes. Thus, it is possible that
in the case of Indonesia, more permanent moves is more connected to more permanent,
secure, formal jobs, which is not in line with Todaro’s assumption under the framework of
rural-urban migration that the rural unskilled worker initially spending some period of time in
the ‘urban-traditional’ sector (Todaro, 1969; Harris & Todaro, 1970), although this study
does not differentiate rural-urban migration from other types of migration. Other explanation
on this result can be more connected to the reason of migrations that would be analyzed
further in the next regression result in Table 4.
Contrary to the finding by cross-tabulation in Table 1, after including other variables
in the regression model as shown in Table 3, it is found that females are 1.57 more likely to
work in informal sector than formal sector compared to males. This result is consistent with
the notion that informal sector is usually entered by women due to its characteristics,
especially its flexible working hours(Todaro & Smith, 2006). If we look at the location
variable, adults who live in rural areas are 3.67 more likely to work in informal jobs
compared to adults who live in urban areas. This result is in line with the statement that
informal economy is actually a rural phenomenon in Indonesia(Nazara, 2010) because the
development in the countryside lags behind urban development so that livelihood in rural
areas is usually more traditional than the urban livelihoods. Traditional livelihoods jobs, such
as irregular agriculture workers for example, mostly fall into the category of informal sector.
This regression result also confirms that working in informal sector is more connected
with older age and a U-shaped pattern of age and informal sector is found. Entry to informal
16
sector is declining at younger age but then increase at older age. Those with low education
are more likely to work in informal jobs than in formal jobs compared to those with higher
education due to lower skill and less opportunities in formal labor market. According to
marital status, those who are never married have the lowest likelihood to work in informal
jobs than formal jobs compared to rest. This finding is in line with the findings of Sahlan
(1996) and Syahran P. (2000) that married individuals are more likely to work in informal
sector than in formal sector. Perhaps it is due to the fact that it is easier for single individuals
to look and to enter jobs in formal sector as they are more flexible than the married ones in
terms of dependence, time, etc., and also that in Indonesia, employers tend to put that being
single as one of the requirements to apply for formal jobs.
Table 3. Multinomial Logistic Regression Results, Employment Status by Migration
Informal Sector Not Working B SE Exp(B) B SE Exp(B)
Migrate -0.868 0.035** 0.42 -0.794 0.036** 0.45 Female 0.448 0.021** 1.57 1.653 0.023** 5.22 Living in Rural Areas 1.300 0.031** 3.67 0.646 0.031** 1.91 Age -0.170 0.006** 0.84 -0.436 0.006** 0.65 Age2 0.002 0.000** 1.00 0.005 0.000** 1.01 Married 0.670 0.033** 1.95 0.221 0.033** 1.25 Divorced/Separated 0.306 0.081** 1.36 -0.022 0.090** 0.98 Widowed 0.801 0.082** 2.23 0.556 0.088** 1.74 Less than Primary School -0.388 0.065** 0.68 -0.207 0.072** 0.81 Primary School -1.086 0.065** 0.34 -0.858 0.072** 0.42 Junior High School -1.803 0.069** 0.16 -1.402 0.076** 0.25 High School -2.797 0.070** 0.06 -1.754 0.076** 0.17 Diploma/University -4.188 0.077** 0.02 -2.248 0.082** 0.11 Constant 4.291 0.123** - 8.010 0.127** -
Log likelihood -117958.11 LR Chi-square (d.f.) 30469.28 (26) Pseudo R2 0.1921 Total Observations 174185 **Significant at 1 percent level
To analyze further the relationship between migration and informal sector, the link
between motivation to migrate and the employment status is investigated, which is the unique
contribution of this study. The result in Table 4 show the statistically significant relationship
between reasons to migrate and either working in formal sector, informal sector, or not
working. The highest likelihood to work in informal sector than in formal sector is found
among migrants who move for their education or training compared to migrants who move
17
for other work related reasons such as job transfer or move closer to their jobs. It is quite
logical as the process of education requires most of migrants’ time, which does not allow
them to work full time as usually happen in formal sector jobs.
The interesting part of this finding, nevertheless, is the fact that migrants who move
for the reasons related to affiliation, household events/disturbance (such as death, pregnancy,
illness, divorce, etc.) and those who move for the reasons related to other kinds of reason
(such as political disturbance, natural or other disasters, drought, etc) are around six to seven
times more likely to work in informal sector than in formal sector compared to those who
move for other work related reasons. It is rather remarkable to find that household events
such as death, pregnancy, illness, reasons related to affiliation and other events connected to
the environment of the migrants are strongly related to working in informal sector, which
proves the connection between the labor market consequences of migration and household
events/disturbances that trigger migration.
Table 4. Multinomial Logistic Regression Results, Employment Status of Migrants
Informal Sector Not Working B SE Exp(B) B SE Exp(B)
Female -0.297 0.073** 0.74 1.186 0.079** 3.27 Living in Rural Areas 0.476 0.071** 1.61 0.113 0.078** 1.12 Age -0.058 0.022** 0.94 -0.331 0.023** 0.72 Age2 0.001 0.0003** 1.00 0.004 0.0003** 1.00 Married 0.595 0.101** 1.81 0.133 0.109** 1.14 Up to Junior High School 1.276 0.080** 3.58 0.346 0.087** 1.41 Get Work 0.683 0.202** 1.98 -0.264 0.249** 0.77 Look for Work 1.275 0.213** 3.58 0.816 0.260** 2.26 Marriage 1.274 0.206** 3.58 1.481 0.249** 4.40 Affiliation 1.934 0.200** 6.92 2.425 0.239** 11.30 Household Events/Disturbances 1.910 0.251** 6.76 2.184 0.295** 8.88 Education/Training 2.402 0.286** 11.05 3.901 0.281** 49.47 Other Kinds of Reasons 2.025 0.262** 7.57 2.417 0.302** 11.21 Being Independent, Like 1.261 0.210** 3.53 1.789 0.253** 5.98 the Destination and Housing Reasons Constant -2.388 0.394** - 2.025 0.428** -
Log likelihood -8437.89 LR Chi-square (d.f.) 296.63 (28) Pseudo R2 0.1527 Total Observations 9909 **Significant at 1 percent level
18
Other interesting finding is that the RRR values of ‘get work in the destination’(1.98)
is lower than ‘look for work’ (3.58). If we see these two work-related reasons as looking for
opportunity and moving due to necessity, respectively, it is worthy of note that moving due to
necessity has higher likelihood to be related with working in informal sector than formal
sector compared to moving due to opportunity. This result shed some explanation on how
migration is linked to informal sector. Migration for work may not be treated as a
homogenous reason as different types of work-related reasons may result in different labor
market outcome. Migration for work-related reasons with higher certainty such as job
transfer, being closer to jobs, and looking for opportunity is more related to formal sector
than migration for work-related reasons with lower certainty as moving to find jobs due to
unemployment or lack of opportunity in locations of origin. This lower certainty work-related
reasons may be more connected to the informal sectors and probably also with less permanent
migrations as stated previously. It may imply further that migration phenomenon in Indonesia
with longer stay may be in fact more connected with less risky move than the shorter stay. In
other words, migrants in Indonesia are perhaps more risk-averse than we expected and that
they form migration decision for work based on the risk that would be faced9.
According to non work-related reasons, those who move for marriage also have the
same magnitude of RRR with those who move to look for work. Contrary to what one might
expect, marriage migration does not seem to operate at the same level as migration with
affiliation, which has twice RRR value as marriage migration. Although adults who move for
marriage are more likely to be working in informal sector than migration for other work-
related reasons, they have lower likelihood to work in informal sector than those who move
for affiliation. We may see this as an indication that marriage migration is probably more
connected with work related reasons than for ‘affiliation’ than what we usually expected. It is
probably because marriage is related with transition to adulthood, into being responsible and
independent (i.e. working in the more stable jobs such as jobs in formal sector), and thus
9 In connection with this implication, several things should be noted. First, Harris and Todaro do not state
specific length of stay in the destination to be defined as migration. So it would be possible that less permanent migrants confirm their theory that some migrants enter informal sector for temporary phase if we include less permanent migrants. Second, as elaborated by Speare and Harris (1986), representative sampling is usually household sampling while in some cases, migrants who are most likely to work in informal sector and possibly stay for long period of time in destination such as prostitutes, pedicab drivers, street artists, homeless beggars, etc., are excluded from the sample. Therefore, it is still possible that migrants are more likely to enter informal sector especially for this type of migrants.
19
migration for marriage has lower likelihood for working in informal sector than migration for
affiliation.
One last interesting finding to be discussed from Table 4 is that the RRR value of the
last category of reasons, being independent, like the destination, and housing reasons is
similar with move for marriage and look for work and lower than the RRR values of move
for affiliation, household events/disturbance, and other kinds of reasons. It seems like
migration for this reason somehow operate similarly with move for marriage and look for
work reasons if we compare it with more certain work-related reasons. Being independent
and like the destination are two reasons that are usually related more to individual decision
than family or household decision while moving for housing reasons is usually more
connected to household decision. In combination, it is possible that the move connected with
these two reasons are more likely to be related with informal sector than formal ones
compared to other work-related reasons.
Discussion and Further Study
The findings of this study show that adults who migrate have lower tendency to work
in informal jobs than those who do not migrate and it confirms the findings of previous
studies in Indonesia but contradicts the general assumption of relationship between migration
and informal sector. Because migrants are more likely to work in formal sector than the
informal sector compared to non migrants, it seems like migrants in Indonesia are probably
more risk-averse than we expected because they prefer to have secure jobs. Further
investigation into the motives to migrate shed a light on the explanation of this phenomenon.
After analyzing the migrants by disaggregating the reasons to migrate, two main findings of
this study can be highlighted. The first one is the significant proof that although migrants are
more likely to work in formal sector than non migrants, among migrants themselves, the
likelihood to work in informal sectors is varied with different motivations to migrate. Work-
related reasons are strongly connected with working in formal sector while affiliation,
household events/disturbance, and ‘forced’ migrations are strongly related with working in
informal sector. The second important finding is that, among the work-related reasons,
reasons connected to pursue for opportunity is more related with working in formal sector
while loss of employment and lack of employment in area of origin is more related with
working in informal sector.
The main conclusion we draw from these findings is that although migrants in
Indonesia tend to work in formal sector than the informal one, the likelihood of working in
20
informal sector is determined by the motives to migrate and higher when the motives are not
related with work. We also propose that studying labor migration needs to incorporate the
possibility that non labor migrants may also enter the labor market and to distinguish
migration for work whether it is looking for opportunity or merely a survival strategy.
However, we need to note that migration in this study is defined as moving across at least
village level to stay in destination for at least six months. This bears concequences of
excluding less permanent moves and also as the case in most household surveys, unusual
cases are excluded such as homeless or those with no permanent residence while they may be
the ones who are more likely to work in informal sector. Furthermore, it may be needed to
explore other variables available to understand this phenomenon such as differentiating the
type of migrations as rural-rural, rural-urban, urban-urban, and urban-rural and also includes
the number of movers, distance, etc.
Other limitation of this study is that according to Meng (2010), difference in results
regarding the relationship betwen migration and informal jobs may occur due to the
difference in the concept of informal economy. Although the division of informal and formal
activities or economy is used extensively in the study of economic development, there is no
standard theoretical definition of informal economy (Koo and Smith, 1983). The challenge to
improve this study, therefore, is an attempt to capture various aspects of informal sector that
are not considered in this study such as working hours, wage, size of the company and to use
more accurate definition of informal jobs, especially the latest one developed adopted by the
Indonesia statistical bureau, Badan Pusat Statistik (BPS, 2012) to observe whether the results
still show the same relationship between migration and informal jobs and.
21
Reference
Badan Pusat Statistik. (2012). Indikator Pasar Tenaga Kerja Indonesia [Indonesia Labor Market
Indicators], February 2012. Jakarta: Badan Pusat Statistik.
BPS. (n.d.). Last visited: July 22, 2013, from Badan Pusat Statistik Website: http://www.bps.go.id/
Cole, W. E., & R. D. Sanders (1985). Internal Migration and Urban Employment in the Thirld World.
The American Economic Review, Vol. 75, No.3 , 481-494.
De Jong, G. F., & J. T. Fawcett (1981). Motivations for Migration: An Assessment and a Value-
Expectancy Research Model. In G. F. De Jong, G. R. W., eds., Migration Decision Making.
New York: Center for Cultural and Technical Interchange between East and West, Inc.
Flórez, C. E. (2003). Migration and the Urban Informal Sector in Columbia. Paper presented at
Conference on African Migration in Comparative Perspective, Johannesburg, South Africa, 4-
7 June 2003.
Handayani, T. (1991). Migran dan Sektor Informal di DKI Jakarta (Analisis Data Supas 1985)
[Migrants and Informal Sector in DKI Jakarta: Analysis of SUPAS 1985 Data] . Master
Thesis. Depok: Population and Labor Studies Graduate Programme, University of Indonesia.
Haynes, M., M. Western, L. Yu, & M. Spallek (2006). Methods for Categorical Longitudinal Survey
Data: Understanding Employment Status of Australian Women. Paper prepared for the
Methods for Longitudinal Surveys (MOLS2006) Conference, University of Essex,
Colchester, United Kingdom, on 12-14 July 2006.
Hugo, G. J. (1982). Circular Migration in Indonesia. Population and Development Review, No. 1, Vol.
8, 59-83.
Koo, H., & P. C. Smith (1983). Migration, the Urban Informal Sector and Earnings in the Phillipines.
The Sociological Quarterly, Vol. 24, No. 2.
Listya, R. (2012). Migrasi dan Sektor Informal: Analisis Data IFLS 2007 [Migration and Informal
Sector: Analysis of the IFLS 2007 Data]. Bachelor Thesis. Depok: Faculty of Economics
University of Indonesia.
Lucas, R. E. (1997). Internal Migration in Developing Countries. In M. R. Rozensweig, & O. Stark,
Handbook of Population and Family Economics, 1st Edition, Vol. 1, No. 1. Elsevier.
Meng, X. (2001). The Informal Sector and Rural-Urban Migration: A Chinese Study. Asian Economic
Journal, No.1, Vol.15.
Mulder, C. H. (1993). Migration Dynamics: A Life Course Approach. Amsterdam: Thela Thesis
Publishers.
Nazara, S. (2010). Ekonomi Informal Indonesia: Ukuran, Komposisi dan Evolusi [Informal Economy
of Indonesia: Measurement, Composition, and Evolution]. Jakarta: International Labor
Organization.
Sahlan, E. (1996). Partisipasi Kaum Migran dalam Ekonomi Informal di Daerah Perkotaan: Suatu
Analisis data IFLS (Indonesia Family Life Survey 1993) [Migrants’ Participation in Informal
22
Economy in Urban Areas: An IFLS 1993 Data Analysis]. Master Thesis. Depok: Population
and Labor Studies Graduate Programme, University of Indonesia.
Speare, J. A., & J. Harris (1986). Education, Earnings, and Migration in Indonesia. Economic
Development and Cultural Change, Vol. 34, No.2 , 223-244.
Strauss, J., F. Witoelar, B. Sikoki & A. M. Wattie (2009). The Fourth Wave of the Indonesia Family
Life Survey: Overview and Field Report, Volume 1. Santa Monica: RAND.
Syahran P., T. (2000). ‘Migran dan Pekerja Sektor Informal di Povinsi Jawa Barat (Analisis Data
Sensus Penduduk Tahun 1990) [Migrants and Informal Sector Workers in Province of West
Java]. Master Thesis. Depok: Population and Labor Studies Graduate Programme, University
of Indonesia.
Todaro, M. P. (1969). A Model of Labour Migration and Urban Unemployment in Less Developed
Countries. American Economic Review, 59 (1) , 138-148.
Todaro, M. P. & J. R. Harris (1970). Migration, Unemployment and Development: Two-Sector
Analysis. The American Economic Review, Vol.60 No.1 , 126 – 142.
Todaro, M. P. & S. C. Smith (2006). Economic Development. Pearson Addison Wesley.
Wajdi, M. N. (2010). Migrasi Antar Pulau di Indonesia: Analisis Model Skedul Migrasi dan Model
Hybrida [Inter-island Migration in Indonesia: Migration Schedule and Hybrid Models
Analysis]. Master Thesis . Depok: Population and Labor Studies Graduate Programme,
University of Indonesia.
Willekens, F. J. (1985). Migration and Development: A Micro Perspective. Working Paper No. 62 .
Voorburg: NIDI.
23
Appendix 1 Reasons to Migrate, IFLS4
Category/List of Values Reason to Move Frequency Percent To get work (opportunity) To get work 2,040 20.59
Look for work (survival Looking for work (not enough employment) 577 5.82 strategy) Work problem (being laid-off, etc) 373 3.76
Other work reason Job transfer 138 1.39 Pension 16 0.16 Work, other reasons 57 0.58 Closer to jobs 118 1.19 Own work, no details 4 0.04 Military work (job transfer) 95 0.96 Affiliation: Being with spouse Marriage 1,747 17.63
Affiliation: Other people’s work (non family) 281 2.84 Living near family/friends Other people’s work (whose work unknown) 2 0.02 Education/training of family members 29 0.29 Education/training (whose education unknown) 1 0.01 Military work of family members 5 0.05 Move with family 647 6.53 To be closer with family 929 9.38 Live with family members 482 4.86
Household events/ Pregnancy 83 0.84 disturbances Spouse' death 19 0.19 Own or spouse's illness 52 0.52 Death of other people 20 0.20 Illness of other people 22 0.22 Family problem 100 1.01 Divorce 71 0.72 Status: Obtaining a good education Own education/training 509 5.14
Other disturbances Political disturbance 3 0.03 (include forced migration) Demolition 15 0.15 Transmigration 2 0.02 Drought 2 0.02 Natural and other disasters 35 0.35 Other reasons 256 2.58
Autonomy Being independent/separated from parents 484 4.88 Location preference Like the destination 204 2.06 Housing reason New housing opportunity 491 4.96