Jurnal Ekonomi dan Studi Pembangunan, 14 .
, 2022
ISSN 2086-1575 E-ISSN 2502-7115
Why Are Children Still Working? Heri Sandra1.
Taufiq Carnegie Dawood2.
Shabri Abd.
Majid2 Statistics Indonesia (BPS) Padang Pariaman Branch and Master of Economics.
Faculty of Economics and Business.
Syiah Kuala University.
Indonesia Faculty of Economics and Business.
Syiah Kuala University.
Indonesia E-mail: taufiq.
dawood@unsyiah.
Received: December 5, 2020.
Revised: April 5, 2022.
Accepted: April 5, 2022 Permalink/DOI: http://dx.
org/10.
17977/um002v14i12022p072 Abstract This study aims to determine the effect of household characteristics and children's living environment on child labor to reduce the likelihood of child labor occurrence.
The data source of this research is Sakernas (Indonesian National Labor Force Surve.
of August 2018, including school accreditation, household head gender, age, working sector, employment status, household size, dependency ratio, location, and school quality, towards the probability of child labor occurrence in Indonesia.
Results suggest that the government must prioritize welfare programs for women household heads, asses the minimum age for marriage, improve education facilities, and increase parental awareness to eliminate child labor.
The government needs to improve education infrastructure.
This infrastructure improvement should be accompanied by the convenience of school access and its costs.
The lack of education infrastructure, especialy in rural areas makes it difficult for children there to attend formal schooling.
Keywords: Child Labor.
Households.
Quality Education JEL Classification: J13.
J49.
I24
INTRODUCTION
Children should spend their time playing and learning new things to prepare them to become mature individuals in the future without being burdened with heavy However, many children are abandoned or forced to help their household economics, leading to the deprivation of their childhood.
There are 152 million child workers globally, equivalent to 9.
6 percent of the children population, wherein about 72 million are doing hazardous work for themselves and their future (ILO, 2.
Criteria for hazardous work are work that physically, sexually, and psychologically abuses children, works with dangerous equipment, substances, processes, and places, or works too long (ILO, 2.
Research findings suggest that child labor adversely affects children's education (He, 2.
and health (Ibrahim et al.
These findings reinforce the hypothesis that child labor can Jurnal Ekonomi dan Studi Pembangunan, 14 .
, 2022
ISSN 2086-1575 E-ISSN 2502-7115
hinder a child's development, the impact of which will become more pronounced as they grow up.
Therefore, we must support all efforts to reduce and eliminate child labor.
Child labor is common in low-income countries.
For example, there was 4% of child labor in low-income countries in 2016, whereas, in countries with higher GNI, the percentage of child labor was lower (ILO, 2.
Therefore, it indicates that welfare determines the number of child laborers.
Some previous studies also indicate this and found that poverty is one of the determinants of child labor (Hamenoo et al.
The higher the poverty rate of a country or region, the higher the number of child laborers.
On the contrary, in their study.
Basu et al.
used land ownership as a wealth indicator to see the wealth effect has on child labor.
The results show that land wealth will increase child labor but decrease when the household already has at least 4 acres (A 1.
6 h.
of land.
Basu concludes that the relationship between land wealth and child labor is a reversed U-shape.
Meanwhile.
Bhalotra .
uses fishpond ownership as a wealth indicator.
Through research in Pakistan and Ghana.
Bhalotra found that wealth was positively related to child labor.
Another study conducted by (Lima, et al.
also used indicators of land ownership as wealth.
The results show that wealth goes hand in hand with an increase in child labor, so they hypothesized that the so-called "wealth paradox" in child labor is driven by parents' preferences .
The results of this study are interesting to explore because children grow and develop in a household.
Therefore, the condition of the household will influence its development.
Children do not have a mature mindset, so the household acts as a decision-maker in the child's life.
It causes the household head to play a significant role in decisions concerning the lives of children, including in terms of Thus, the preference of the household head plays a more critical role in the child's decision to go to work than any other factor (Lima et al.
Therefore, it is vital to know the characteristics of the household head concerning their influence on the existence of child labor in the household.
Household characteristics, such as the household head's gender, age, occupation, and the number of household members, are known to influence child labor (Ali & Arabsheibani, 2017.
Chong & Yanez-Pagans, 2019.
Tang et al.
However, there are contradictions in the results of previous studies on the effect of some household characteristics on child labor.
Susanli et al.
and Cummings .
found that parentAos education affects child labor, while Tang et al.
and Afriyie et al.
found that parentAos education had no significant effect.
Furthermore, the household head gender has less influence on child labor, according to Tang et al.
, while Ali & Arabsheibani .
find that children are more likely to work if the household head is a woman, rather than man.
Regardless, we will place a position supported by solid evidence and argument from observations on the household level of 96,971 children aged 11-17 years recorded in Indonesia's August 2018 Sakernas regarding their influence characteristics on the likelihood of child labor occurrence.
In addition to the household, the child's home environment also affects the lives of children, including working, rural, and urban environments differ in terms of population density, infrastructure, and other socioeconomic conditions.
For example.
Webbink et al.
found that children were less likely to work in urban Jurnal Ekonomi dan Studi Pembangunan, 14 .
, 2022
ISSN 2086-1575 E-ISSN 2502-7115
It may be related to road infrastructure and educational facilities that are generally better, thus encouraging parents to send their children to school rather than work.
Education variables are often researched concerning child labor, where they are related to children's participation in schools and education subsidies (Canagarajah & Coulombe, 1999.
Ravallion & Wodon, 1999.
Tang et al.
, 2.
In general, this study aims to determine the effect of household characteristics .
ousehold headAos gender, age, education, working sector, employment status, number of household members, and household dependency rati.
and environmental characteristics .
ousehold location and KLP) on the risk of children becoming child laborers in Indonesia.
The urgency of this research is to provide information for the Indonesian government to fight child labor.
Reducing child labor is essential because Indonesia has vowed to fulfill the sustainable development goals (SDG.
, and one of the targets of the SDGs is to eliminate child labor by 2030.
METHOD
The Sakernas 2018, by Indonesian Statistics, provides data on child labor, child household characteristics, and residential location characteristics.
In addition, the KLP data in a regency or city area was obtained from the National Accreditation Board for Schools or Madrasa (BANSM).
This KLP data includes 54,007 primary and secondary school accreditation data throughout Indonesia in 2018, which BANSM has assessed.
Specifically, the measurement of this data is expressed by the percentage of primary and secondary schools that are accredited at least Good (B) in one regency or city, so the KLP data is in the form of aggregate data at the regency or city level.
On the other hand, the Sakernas data is in the form of individual child data.
Therefore, when these two data are combined, it will result in children's KLP data in the same regency or city.
We analyze the combined data using the Binary Logistic .
The logit model can estimate binary response opportunities based on a set of predictor variables (Putri & Prasetyani, 2.
The dependent variable in the logit model has two possible values .
, which are "1" if it meets specific given criteria or "0" if otherwise (Dawood et al.
Among the advantages of the logit model is that the relationship continuity between the dependent and independent variables is unnecessary.
Moreover, the dependent variable does not have to be normally distributed.
In addition, it does not have to be protected from heteroskedasticity (Yah, 2.
Previous researchers have widely used the logit model to determine the determinants of child labor (Afriyie et al.
Ali & Haszelinna, 2016.
Binci & Giannelli, 2016.
Chakrabarty, 2015.
Cummings, 2016.
(Oryoie et al.
The following equation shows our logistic regression model:
Ln!"# !! " = Zi = 0 1X1 2X2 3X3 , nXn Wherein Pi is a Bernoulli random variable that is defined as the probability that the ST variable is valued 1 with condition Z as follows:
Pi = E (ST = 1 iCZi ) = .
"% &"#$ Jurnal Ekonomi dan Studi Pembangunan, 14 .
, 2022
ISSN 2086-1575 E-ISSN 2502-7115
The 1 - Pi is the probability that the variable ST is zero and "# !! is the odds ratio value that a child falls into the category of child labor (ST = .
In addition, the random variable Ai defines the marginal change in probability, which is assumed to follow a normal distribution with a mean of zero and variance ' ! ("# ! ) where ! ! N is the number of children.
The object of this study is children aged 11-17 years recorded in the August 2018 Sakernas.
The dependent variable in this study is child labor (PA) which is defined as follows:
1 O yaEaycnycoycc ycoycaycaycuyc
PA = #
2: ycCycEayceyc yaycuycuyccycnycycnycuycu Meanwhile, the independent variables are .
the household head (JKKRT), which is a binary variable with Au1Ay representing male and Au0Ay representing female, .
the household head Age (UMRKRT) in years, .
the household head education (SEKRT) as measured by the length of the school year, .
the household head working sector (TANI) with Au1Ay representing the agricultural sector, while Au0Ay if otherwise, .
the household head employment status (BURUH) with Au1Ay representing the status as a worker/employee/free-worker and Au0Ay if others, .
Dependency ratio (DEP) with Au1Ay if the number of non-productive household members is more than the productive members .
igh dependenc.
and Au0Ay if other than that .
ow dependenc.
, .
the location of the household (KOTA) is Au1Ay if in urban areas and Au0Ay if rural, and .
Quality of educational institutions (KLP) the percentage of schools that are accredited at least Good (B) in the Regency or City where children live.
Thus, the research model is as follows:
PA = 0 1JKKRT 2UMRKRT 3SEKRT 4TANI 5BURUH
6JUM 7DEP 8KOTA 9KLP A .
RESULTS AND DISCUSSION
As many as 187.
420 households were sampled in August 2018 Sakernas.
There are 96.
971 children of 11-17 years from all the sample households.
However, several regions did not have KLP data.
thus, we omit as many as 688 children aged 11-17 years old.
Table 1 provides a comparative description of the characteristics of child households involved in child labor with those not child laborers.
Table 1 shows 3,652 child workers from 96,971 child data samples in Sakernas in August 2018, equivalent to 3.
77 percent of all registered children, and the rest are not child Of the child laborers, 60 percent are men, and 40 percent are women.
indicates that men still dominate child labor in Indonesia.
The household head age ranged between 41 to 64 years, around 71.
2 percent.
The household head gender is primarily male.
Jurnal Ekonomi dan Studi Pembangunan, 14 .
, 2022
ISSN 2086-1575 E-ISSN 2502-7115
Table 1.
Descriptives of Children Based on Household Characteristics Non-child labor
Household
Total
Men
Wome
Child labor
Total
Men
Sum
Wom
Total
Total
Men
Sum
Total
Wome
Sum
96,23
3,77
<= 40 years 21,82 1,02 22,85 41 - 64
68,70
2,52
71,22
>= 65 years 5,71 0,23 5,94 Househ
old head
Women
10,17
0,54
10,71
Men
86,07
3,23
89,29
Househ old head Househ old head field of <= Ele sch 45,88 2,33 48,21 High sch 41,56 1,30 42,85 University Agricultural 8,79 0,14 8,94 35,24 1,74 36,99 Others 60,99 2,02 63,01 Dependency High 80,65 2,88 83,53 Low 15,58 0,89 16,47 Househ Rural 53,07 2,46 55,53 Urban 43,16 1,31 44,47 Househ old head Source: Sakernas August 2018, data processed Meanwhile, the household head's education mostly graduated from elementary school or may not even finish whether the household is from child labor.
Even though Indonesia is an agricultural country, the household head field of work in the agricultural sector is smaller when compared to all non-agricultural sectors, which is 46,3 percent.
The majority of child households are in rural areas, and the most dependency ratio is low dependency.
Percentage 0 20 40 60 80 100120140160180200220240260280300320340360380400420440460480500520 No Regency/City KLP Regency/City KLP Indonesia Figure 1.
Quality of Educational Institutions in Indonesian Regencies/Cities 2018 Source: BANSM, data processed Jurnal Ekonomi dan Studi Pembangunan, 14 .
, 2022
ISSN 2086-1575 E-ISSN 2502-7115
The average quality of educational institutions in Indonesia is 82 percent, meaning 82 percent of schools or madrasah (Islamic school.
in Indonesia already have good quality .
inimum accreditation B) by the expected standards.
However, there are quite striking differences when viewed by Indonesia's regency or city area.
For example, in Figure 1, the regency or city number 1-151 is in Sumatra, 152-270 in Java, and 271-510 in Kalimantan and Eastern Indonesia.
The KLP is already good in the Java region, where the percentage of schools in the region is generally above average.
However, there are more regions with low and below average KLP rates in regions outside Java, especially in the eastern region.
It shows that there are still gaps in the quality of educational institutions in Indonesia.
As for seeing the relationship between household characteristics and regional characteristics with child labor, as seen in the logistic regression result in Table 2.
By considering the last three lines, the significance of the Hosmer and Lemeshow-Test is greater than 0.
05, so there is no difference between the model and its observational values significant.
On the other hand, the significance of the Omnibus-Test is smaller than 0.
05, meaning that there is at least one independent variable that significantly influences the dependent variable.
These two tests indicate that the model is appropriate for estimating the dependent variable.
The estimation accuracy as shown from the Overall Percentage value, which is 96.
Through the coefficients in Table 2, we can write the model as follows.
PA = - 1.
905 Ae 0.
297JKKRT Ae 0.
015UMRKRT Ae 0.
062SEKRT 0.
071TANI
- 0.
177BURUH 0.
04JUM 0.
352DEP Ae 0.
155KOTA Ae 0.
154KLP
A negative sign of intercept indicates that children tend to avoid child labor if they are not influenced by the characteristics of the household and the environment.
Whereas if influenced by household and environmental characteristics such as the male head of household, 48 years old, graduated elementary school .
, worked in agriculture as a laborer with five household members with high dependency in rural areas and 75% KLP is = - 1.
905 Ae 0.
Ae 0.
- 0.
- 0.
- 0.
- 0.
= -3,1185
= "% * %& = "% * "',))* = 0.
* %&
* "',))*
So the risk of children becoming child laborers with these characteristics is 4.
Through the Wald test's significance, independent variables that have a significant effect on child labor can be found.
All independent variables tested significantly affect child labor, but to different levels.
For example, in addition to the business activities of household heads and KLP, which are significant at the 10 percent level, other independent variables are significant at the 1 percent level.
The gender coefficient of household heads (JKKRT) is negative.
Because this variable is coded 1 for boys, this coefficient indicates that children in a household headed by a woman are more likely to be child labor by 1/0.
743 = 1.
34 times than children in the household that males head.
The head of the household is the person most Jurnal Ekonomi dan Studi Pembangunan, 14 .
, 2022
ISSN 2086-1575 E-ISSN 2502-7115
responsible in the family.
In a country where most people adhere to patrilineal, the household head is male, wherein in Indonesia, in 2018, the percentage of male household heads was 84.
83 percent (BPS, 2.
As for women, they do more work in the household.
It results in those men being more prepared to lead the household than women.
In a situation where a woman must be the head of the household, she needs to adjust to this role, especially in terms of meeting household needs.
Children are more likely to work if the woman, being the head household, does not meet their needs.
Table 2.
Variables in the Equation
Variable JKKRT
UMRKRT
SEKRT
TANI
BURUH
JUM
DEP
KOTA
KLP
Constant Coefficient (B) Wald P-Value Exp(B) Hosmer and Lemeshow-Test Omnibus-Test Overall Percentage Total Observation Source: AuthorsAo calculations 96,283 Older household heads reduce the risk of children in the household becoming child laborers.
One year increase in the household head age is in line with a decrease of 0.
985 times the risk of children becoming child laborers.
It is because older household heads have more life experience.
This experience can also make it wiser to make children's decisions, including for work.
Therefore, children in households with older household heads are less likely to become child laborers.
This result is in line with research by Ali & Arabsheibani .
, the age of the head of the household has a negative relation to the incidence of child labor.
The variable of the head of household education significantly influences child labor.
The sign of the education coefficient of household heads is negative, which is -0.
062, which means the length of schooling of the head of the household is in line with the reduction in child labor.
A one-year increase in the head of household education will reduce the risk of a child becoming child labor 0.
94 times as much as the odds ratio for the education of household heads by 0.
These findings confirm the research results from Cummings .
Ali & Arabsheibani .
, where the education of household heads is negatively related to the opportunities for children to become child laborers.
With better education, the head of the household will enable him to work or get a better income so that children no longer need to work.
Besides, education will affect the household head's knowledge Jurnal Ekonomi dan Studi Pembangunan, 14 .
, 2022
ISSN 2086-1575 E-ISSN 2502-7115
of the negative effects of child labor so that it will certainly be difficult to permit children in their households to work.
Children in agricultural households are at greater risk of becoming child 074 times than children in non-agriculture.
This result is in line with the findings of (Tang et al.
that the involvement of households in the nonagricultural sector is negatively related to child labor in China.
Children's participation at work in their household agriculture as unpaid labor is generally in light jobs such as at harvest.
However, the household is less aware that the work time is longer than the child deserves.
In line with these results, the variable of the work status of the head of the household as laborer/employee/free-worker is negatively related to child labor.
It means that if the household head has the status of outside employment, namely as an entrepreneur or not working at all, it will increase the risk of child labor in the household by 1/0.
838 = 1.
19 times.
The August 2018 Sakernas results found that most child workers .
8 percen.
were family or unpaid workers.
It informs that child workers are involved in their household businesses so that household heads who as entrepreneurs are more likely to involve their children to work than the head of the household as a laborer.
Agricultural households are primarily in rural areas, so children in rural areas are also more at risk of child labor than children in urban areas 1/0.
168 times.
Besides the factor of agricultural households, the increased risk of child labor in rural areas is due to the lack of educational facilities and more difficult access to achieve this than in urban areas.
Tang et al.
also found the same thing that the factor of residence of children in rural areas has the opportunity to increase child labor.
Large households tend to send their children to the workforce, as seen from the coefficient of the number of positive household members.
The addition of one household member is associated with an increase in the opportunity for children in the household to become child labor 1.
04 times.
The greater the number of household members, their needs will also If adults cannot meet these needs, they will send children to the workforce to help meet household needs.
A similar result found by Nengroo & Bhat .
suggests that the family's size forced children to enter the labor market at an early Correspondingly, many household members followed by high dependency increasingly force children in the household to work.
High dependency indicates that fewer people are more productive to work than non-productive people.
So they have a greater burden in meeting household needs.
Children are more likely to work if productive members can no longer bear the burden.
Children in high dependency households are 1.
422 times more likely to be child laborers than children in low dependency households.
These results confirm the findings of (Priyambada et al.
that the dependency ratio is in line with the incidence of child labor.
A good quality educational institution will encourage households to send their children to school (Bae, 2.
Meanwhile, the presence of children in school automatically reduces the time children do other activities, especially work, and will undoubtedly lead to a reduction in child labor (Canagarajah & Coulombe.
It is evident in this study that the quality of good educational institutions is related to the reduction of child labor, where an increase of one percent of schools that are accredited minimally both in regency or cities where children live affects reducing child labor by 0.
858 times.
What can we do with the results of this study? Formulating policies to reduce child labor is one of the more values expected in this Jurnal Ekonomi dan Studi Pembangunan, 14 .
, 2022
ISSN 2086-1575 E-ISSN 2502-7115
Government should prioritize Household welfare improvement programs, especially for households headed by single women.
The women became the household head due to several problems, such as the household head who had previously died, causing the children to be orphaned.
Novella .
found that orphans tend to work and not go to school.
The minimum age limit for marriage needs to be studied more closely.
Immature marriages will tend to make their children enter the workforce early, burdening the next generation.
Mature age is also marked by adequate education provision before marriage.
A good education will make it easier for them to get a better income.
Besides, they will have wisdom in making decisions in the household, including the possibility of working for their children later.
They also need to plan well the number of their children later.
The number of household members is closely related to the magnitude of the cost of meeting the needs of life.
The number of qualified large household members will be a potential for welfare, but if not qualified, it will be a burden, including for the children themselves.
Parental awareness is vital in preventing their children from the adverse effects of child labor.
Children's participation in domestic business needs to pay close attention to the child's age and working hours.
Therefore, it does not interfere with physical, mental, and spiritual development and does not interfere with their study time.
Children 12 years or younger may not work at all.
Older children may help with light work no more than 15 hours a week for children aged 13-14 years and no more than 40 hours a week for children aged 15-17 years.
Work that involves children should be work that contributes to the development of children and gives them additional skills and experience and helps them prepare to become productive members of society in the future.
The government needs to improve education infrastructure.
This infrastructure improvement should be accompanied by the convenience of school access and its costs.
The lack of education infrastructure, especialy in rural areas makes it difficult for children there to attend formal schooling, so it encourages them not to go to school, and work is a positive activity that they might do.
It is worth noting that through data processing in August 2018 Sakernas, it turns out that the majority .
4 percen.
of child laborers are children who are active in school.
They combine functions as students and economic actors.
Increasing KLP can at least make the portion of their function as students can be greater than as economic CONCLUSION The contribution of this research in the literature is to determine the opportunities for children to become child laborers based on household characteristics and their environment as a reference in efforts to eliminate child Using Sakernas Indonesia data for August 2018, it was found that household quality is in line with the opportunity to reduce child labor.
In contrast, the household head's working status, the number of household members, and high dependency in the household tend to increase child labor.
To reduce the occurrence of child labor, the government must prioritize welfare improvement programs for households led by women, assess the minimum marriage age, and control the Government should also improve the quality of education, and raise Jurnal Ekonomi dan Studi Pembangunan, 14 .
, 2022
ISSN 2086-1575 E-ISSN 2502-7115
parent awareness of the adverse effects of child labor.
Future studies can consider adding household welfare variables.
REFERENCES