Content uploaded by Thia Jasmina
Author content
All content in this area was uploaded by Thia Jasmina on Jun 09, 2024
Content may be subject to copyright.
The Effect of Educational Cash Transfer for Students From Low-Income Families on
Students' Dropout Rate in Indonesia
Nisma Anggara Samalo, Universitas Indonesia, Indonesia
Thia Jasmina, Universitas Indonesia, Indonesia
The Asian Conference on Education & International Development 2024
Official Conference Proceedings
Abstract
The Indonesian government provides an educational cash transfer program for low-income
families with children aged 6-21 so that the children can complete their education up to
secondary level (Program Indonesia Pintar – PIP program). The Covid-19 pandemic from
2020 to 2022 might hamper the success of this program. Hence, this study aims to see how
the PIP program affects the dropout rate of students from low-income families at the primary,
junior, and senior secondary levels before and during the Covid-19 pandemic. This study uses
cross-sectional data from the National Socioeconomic Survey (SUSENAS) of 2019 and 2021
and applies the propensity score matching (PSM) method. The results show that before the
pandemic in 2019, the PIP program decreased the probability of dropping out for students
from low-income families at primary and junior secondary schools but not for students at
senior secondary schools. However, during the pandemic in 2021, the PIP program decreased
the probability of dropping out for students in junior secondary and senior secondary schools
but not for students in primary schools. The PIP program only significantly reduces the
probability of dropping out for junior secondary students, both before and during the Covid-
19 pandemic.
Keywords: PIP, Educational Cash Transfer, Dropout Rate, Low-Income Students, Propensity
Score Matching
iafor
The International Academic Forum
www.iafor.org
Introduction
Education represents an investment in the future and a cornerstone for creating high-quality
human resources. Inclusive education that fosters skills development is obligatory for every
child, regardless of legal status, nationality, or citizenship (UNESCO, 2019). Education is
advocated through Sustainable Development Goal 4 to achieve accessible, equitable, and
quality primary and secondary education by 2030. To achieve this goal, every child must
complete their education without dropping out (UNICEF, 2017). Despite education being a
top priority on the global agenda, the dropout rate is still relatively high. According to
UNESCO (2019), the total number of dropout children in 2018 amounted to 258.4 million.
The dropout probability for high school students is four times higher than for elementary and
twice as high as for junior high school students due to lack of early schooling opportunities,
non-compulsory upper secondary education, and high school-aged children opting to work
rather than continue their education (UNESCO, 2019).
Indonesia is among the countries facing the challenge of student dropout. According to the
Ministry of Education, Culture, Research, and Technology of Indonesia (MoECRT), in the
2021/2022 academic year, there were 75,303 school dropouts in the country, with the highest
number of dropouts being at the elementary school level (Figure 1). There was an increase in
dropout rates at all levels of education during the 2019/2020 academic year due to the
outbreak and spread risk of Covid-19. The number of dropout students has been relatively
higher at the elementary school level since 2018/2019. However, the dropout rate was the
lowest at elementary schools, with 0.24%. The dropout rate of junior high school students
was 0.39%, senior high school was 0.55%, and vocational high schools was 0.65%.
Source: MoECRT (2021), https://databoks.katadata.co.id (2022)
Figure 1: The number of school dropouts in Indonesia from 2017/2018 to 2021/2022
One of the main factors contributing to the increase in dropout rates in Indonesia is poverty.
Most families (76%) stated that their children dropped out of school due to economic factors,
with 67% unable to afford school fees and 8.7% needed their children to help support the
family financially (Ali, 2009). According to UNICEF, 938 school-aged children in Indonesia
dropped out due to the pandemic, with 74% due to lack of funds
(https://databoks.katadata.co.id, 2021).
From 2018 to 2020, the population living below the national poverty line in Indonesia was at
its highest in 2020. Figure 2 shows a significant increase of 11.13% of the population living
below the poverty line from 2019 to 2020. In 2022, nearly 26.4 million people were living
below the poverty line. The number of child workers aged 10-17 also reached 1.17 million,
an increase of 320,000 compared to 2019 (https://databoks.katadata.co.id, 2021). Children
face a trade-off between working or going to school.
Source: BPS-Statistics Indonesia (2022)
Figure 2: The number of poor people in Indonesia, 2017/2018 to 2021/2022 (in millions)
One strategy for reducing dropout rates is the implementation of a conditional cash transfer
program (CCT), which aims to improve the welfare and human capital investment to alleviate
poverty (Edo & Marchionni, 2019; Glewwe & Kassouf, 2012; Mo et al., 2013; Azevedo &
Robles, 2013). The Indonesian government provides conditional cash assistance to school-
aged children (6-21 years old) from low-income families who can not access education
named Program Indonesia Pintar (PIP program). The program provides assistance to the
respective families so their children can complete their secondary education. It is interesting
to examine the impact of the PIP program on dropout rates for students from low-income
families, especially during the pandemic.
Several quantitative studies have discussed the impact of conditional cash assistance on child
education in Indonesia (Anindita & Sahadewo, 2018; Listiyanto & Qibthiyyah, 2022; Purba,
2018; Setyadharma, 2018; Sihombing et al., 2022; Yulianti et al., 2015). Moreover, there
were some qualitative studies have evaluated the effectiveness of the PIP program in several
schools in Indonesia (Hafrienda et al., 2023; Kaidah & Ruslan, 2021; Uriyalita et al., 2020;
Zahimu, 2019). Most of the studies found that the PIP program can support the education of
children from low-income families. Still, there was limited evidence on how the program
affects dropout rates, especially during the Covid-19 pandemic. Hence, how the PIP program
affects the probability of student dropouts from low-income families needs to be further
elaborated, especially during the pandemic. This study contributes to understanding how the
PIP program prevents school dropout among students from low-income families, both before
and during the Covid-19 pandemic.
Briefly on the Program Indonesia Pintar (PIP) Program
Cash transfer is one policy to enhance educational demand by providing cash assistance to
support children in low-income families to attend schools. Unconditional cash transfer (UCT)
assists low-income families without specific conditions, whereas the cash conditional
transfers (CCT) assist with predefined requirements. Program Indonesia Pintar (PIP
program) is one of Indonesia's conditional cash transfer (CCT) programs that ensures students
from low-income families attend schools and can finish their education up to the secondary
level. The PIP program is targeted at children aged 6-21 from poor and vulnerable families.
The PIP program was the enhancement of the previous program called the Poor Student
Assistance (Bantuan Siswa Miskin-BSM program) in 2014. The PIP program allocates a fixed
amount of cash for eligible students. Elementary students receive the assistance of IDR
450,000/year (approx. USD 28), junior high school students receive IDR 750,000/year
(approx. USD 47), and senior/vocational high school students receive IDR 1,000,000/year
(approx. USD 63) (MoECRT, 2020). The students can use cash from the PIP program to
purchase books and stationery, school uniforms, school supplies, transportation to school,
students' pocket money, additional course fees for formal education participants, as well as
additional practice and internship fees or work placements (MoECRT, 2023).
The highest number of PIP program recipients was in 2018, with 18,699,376 students (Table
1). During the pandemic in 2020, the total number of PIP recipients decreased. However,
while the recipients at other school levels decreased, the number of program recipients at the
elementary school increased.
School Level
Year
2018
2019
2020
2021
2022
Elementary School
10,379,253
10,364,266
10,434,330
10,411,608
10,360,614
Junior High School
4,751,246
4,562,347
4,411,680
4,401,653
4,369,968
Senior High School
1,516,701
1,464,712
1,412,212
1,419,438
1,393,519
Vocational High School
2,052,176
2,007,074
1,834,669
1,852,279
1,829,167
Total
18,699,376
18,398,399
18,092,891
18,084,978
17,953,268
Source: MoECRT (2022)
Table 1: Number of PIP recipients from 2018 to 2022
There are challenges in the implementation of the PIP program. The delayed disbursement of
the funds resulted in the students being unable to utilize the transferred funds (Zamjani et al.,
2019). There are also some deviations in the utilization of the PIP funds. The cash was used
for non-educational purposes, such as buying food supplies and paying family debts
(MoECRT, 2017). Since 2021, the government has implemented the classification of students
who are eligible to receive PIP. The students from households with family incomes in the
lowest income categories, including very poor, poor, nearly poor, and vulnerable poor (decile
classification of 1 to 4), and who are both recorded and not recorded in the Integrated Social
Welfare Data (DTKS) of the Ministry of Social Affairs are eligible for the PIP program.
Literature Review
A number of previous studies have investigated the impact of conditional cash transfers
(CCT) on dropout rates. For instance, Brazil's Bolsa Escola/Familia program has been shown
to increase school participation and reduce the number of children working long-term
(Peruffo & Ferreira, 2017; Glewwe et al., 2020). Similarly, Mo et al. (2013) confirmed the
positive impact of CCTs on dropout rates in rural China. In contrast, Churchill et al. (2021)
found that Pakistan's Benazir Income Support Programme (BISP) positively and significantly
impacted school enrollment and grade progression but did not affect dropout rates in the short
term. In contrast to these findings, Canelas & Niño-Zarazúa (2019) found different results for
Bolivia's Bono Juancito Pinto education assistance program, where the program successfully
increased school participation rates.
Several studies in Indonesia have also examined the impact of educational assistance on low-
income families. Applying data from the Indonesian Family Life Survey (IFLS) and
Intention-to-Treat (ITT) analysis methods, Kharisma et al. (2017) found that JPS scholarships
were effective in reducing dropout rates in primary education. However, they concluded that
the impact of these scholarships could be further enhanced. The PIP program significantly
increased educational expenditure. In this case, government cash assistance can reduce the
likelihood of these students dropping out (Setyadharma, 2018). The BSM program can reduce
dropout rates at every educational level for children from poor households (Yulianti et al.,
2015). Other studies have indicated that PIP is more effective than BSM in reducing dropout
rates, with the most significant impact occurring at the junior high level (Listiyanto &
Qibthiyyah, 2022). Despite the diverse results of the impact of CCT on educational outputs,
the most recent studies prove that educational assistance can reduce dropout rates.
Previous studies also show that the primary factors driving children's failure to complete their
education can be attributed to the characteristics of the households in which they reside
(Khiem et al., 2020; Mo et al., 2013; Wils et al., 2019). In these households, the role of
parents, especially the household head (KRT), is of particular significance as the KRT is the
primary decision-maker in every household decision, including education. A higher level of
parental education is associated with a lower likelihood of children dropping out of school
(Alcaraz, 2020). Households with higher poverty levels are more likely to be in households
with primary employment status in the informal sector, where economic shocks are more
prevalent compared to the formal sector. Moreover, the economic condition of a household
can be gauged by the household per capita expenditure, with lower per capita expenditure
indicating a smaller family needs fulfillment capacity. Another crucial aspect of a household's
background is the number of children, which has a negative or inverse relationship with the
availability of household resources to be distributed (Al-Samarrai & Peasgood, 1998).
Methods and Data
The theoretical framework used in this study is the education production function, which
describes how educational outcomes, such as cognitive abilities and knowledge, are
generated from "raw inputs." It is a framework for understanding how various education
policies can influence student achievement. Children's knowledge and skills are not only
"produced" by school inputs and educational policies but also by other factors outside schools
(Hanushek, 1979; Lovenheim & Turner, 2018). Hanushek (1979) formulated a model by
considering several factors affecting the educational achievement of students as follows:
𝐴𝑖𝑡 = 𝑓(𝐵𝑖𝑡, 𝑃𝑖𝑡, 𝑆𝑖𝑡,𝐼𝑖) (1)
𝐴𝑖𝑡 represents the achievement of student i at a given time, 𝐵𝑖𝑡 is the vector of the family
background of student i at a given time, 𝑃𝑖𝑡 is the vector of peer influences of student i at a
given time, 𝑆𝑖𝑡 is the vector of school inputs at a given time, and 𝐼𝑖 is the vector of student
characteristics.
This study employs quantitative analysis using the Propensity Score Matching (PSM) method
to mitigate selection bias caused by non-random controls. PSM will find similarities in the
characteristics of two populations:
− the treatment group, which is students who received the PIP program;
− the control group, which is students who did not PIP program.
Randomization will be conducted on both populations by PSM by considering the similar
characteristics of the two groups and hence allowing for a direct comparison between the
groups. The sample in this study is divided into two groups, the treatment group and the
control group, which exhibit similar characteristics, as illustrated in Figure 3.
Source: Authors
Figure 3: Design of Propensity Score Matching (PSM) Method
First, the test for differences in mean covariates is conducted to check the imbalances
between treatment and control groups. Second, the PSM method will estimate probability
scores for PIP acceptance to each individual in the sample based on a logit regression as
follows:
(2)
where,
: Probability of PIP acceptance of the individual i
: Confounding variables in the form of household background of the individual i in
household j consisting of household expenditure per capita, head of household education,
leading occupation status of head of household, and number of children in the household
: Control variables considered for PIP acceptance in household j
Based on the education production function, this study focuses on the household background
as the primary factor that might affect students' dropout. Therefore, the variables included in
the logit estimations consist of household backgrounds, such as the education level and
employment status of the head of the family, per capita household income, and the number of
children in the households. In addition to PIP, low-income families also receive other
government transfers of Program Keluarga Harapan (PKH) and Kartu Keluarga Sejahtera
(KKS). Hence, other government transfers are also included in the logit estimation. The logit
estimation will provide the common characteristics of the recipient of PIP.
Third, following the logit estimation, the subsequent steps are to verify the presence of
common support by implementing the matching method. Common support refers to the
overlap in the distribution among treatment and control groups to ensure both have similar
propensities for treatment. The matching method using the Nearest Neighbor (NN) with the
caliper is applied because the number of observations for the treatment group was
significantly lower than the control group.
Lastly, the outcome of this study is the Average Treatment Effects on Treated (ATT) value as
the following equation:
(3)
Where ATT is the value of the effect of PIP on the probability of dropping out of school, T=1
represents program participants (recipients of PIP), and is the outcome of program
participants. Meanwhile, T=0 represents non-program participants (non-recipients of PIP),
and is the outcome of those who are not program recipients. Figure 4 summarizes the steps
in applying the PSM method in this study.
Source: Authors
Figure 4: Steps in Applying the PSM Method
This study utilizes secondary data from the National Socioeconomic Survey (SUSENAS) of
2019 and 2021 with a cross-sectional approach. The sample consists of school-aged children,
6-21 years old, from poor and vulnerable households based on expenditures per capita. In
2019, 112,004 individuals were identified as the sample, with 63,204 in elementary school,
29,248 in junior high school, and 19,552 in senior high school. Meanwhile, 121,163
individuals were identified as the sample in 2021, with 66,177 in elementary school, 30,458
in junior high school, and 24,528 in senior high school. Summary statistics of the
characteristics of the sample are presented in the Appendices.
Results and Discussion
Table 2 presents the dropout status of students from low-income families who were recipients
and non-recipients of the PIP program in 2019 based on their education level. In general, the
dropout rate was higher among poor and vulnerable students who did not receive PIP. It can
be seen that the percentage of students who dropped out in both sample groups (recipients of
PIP and non-recipients of PIP) was significantly low. However, the dropout rate of the
sample who received PIP is relatively lower (0.61%) compared to the ones who did not
receive PIP (0.91%). This trend is consistent in 2021, as depicted in Table 3. The dropout rate
of the sample who received PIP is relatively lower (0.71%) compared to the ones who did not
receive PIP (1.01%). A summary descriptive of the sample can be seen in the Appendices.
Education Level
2019
2019
Total
Sample
Recipients of PIP
(Treatment Group)
Non-Recipients of PIP
(Control Group)
Dropout
Not
Dropout
Total
Dropout
Not
Dropout
Total
Elementary School
63,204
34
17,883
17,917
177
45,110
45,287
0.19%
99.81%
100%
0.39%
99.61%
100%
Junior High School
29,248
78
8,243
8,321
281
20,646
20,927
0.94%
99.06%
100%
1.34%
98.66%
100%
Senior High School
19,552
78
4,636
4,714
281
14,557
14,838
1.65%
98.35%
100%
1.89%
98.11%
100%
Total
112,004
190
30,762
30,952
739
80,313
81,052
0.61%
99.39%
100%
0.91%
99.09%
100%
Source: National Socioeconomic Survey 2019 (processed by authors)
Table 2: The Number of Dropouts of Students from Low-Income Families, 2019
Education Level
2019
2019
Total
Sample
Recipients of PIP
(Treatment Group)
Non-Recipients of PIP
(Control Group)
Dropout
Not
Dropout
Total
Dropout
Not
Dropout
Total
Elementary School
66,177
32
15,576
15,608
156
50,413
50,569
0.21%
99.79%
100%
0.31%
99.69%
100%
Junior High School
30,458
90
7,885
7,975
360
22,123
22,483
1.13%
98.87%
100%
1.60%
98.40%
100%
Senior High School
24,528
84
5,212
5,296
414
18,818
19,232
1.59%
98.41%
100%
2.15%
97.85%
100%
Total
121,163
206
28,673
28,879
930
91,354
92,284
0.71%
99.29%
100%
1.01%
98.99%
100%
Source: National Socioeconomic Survey 2021 (processed by authors)
Table 3: The Number of Dropouts of Students from Low-Income Families, 2021
The subsequent step is to estimate probability scores for PIP by applying the logit regression
as outlined in equation (2). The estimation of the PSM model using logit is presented in Table
4, which indicates that, for all education levels, certain characteristics are associated with an
increased likelihood of receiving PIP. The characteristics include the low education level of
the household head (lower than senior high school level), having a lower number of children,
and being recipients of other government transfers of Program Keluarga Harapan (PKH) and
Kartu Keluarga Sejahtera (KKS).
Elementary
School
2019
(1)
Elementary
School
2021
(2)
Junior High
School
2019
(3)
Junior High
School
2021
(4)
Senior High
School
2019
(5)
Senior High
School
2021
(6)
PIP
PIP
PIP
PIP
PIP
PIP
cap
-0.099***
0.018
-0.049
-0.015
-0.143**
-0.027
(0.032)
(0.032)
(0.048)
(0.047)
(0.065)
(0.057)
educ_KRT
-0.265***
-0.104***
-0.22***
-0.087***
-0.237***
-0.133***
(0.023)
(0.022)
(0.035)
(0.032)
(0.045)
(0.039)
work_KRT
-0.016
0.027
0.002
0.009
0.018
-0.03
(0.022)
(0.021)
(0.033)
(0.032)
(0.042)
(0.038)
totalchild
-0.076***
-0.093***
-0.035***
-0.05***
-0.066***
-0.023*
(0.007)
(0.007)
(0.01)
(0.01)
(0.013)
(0.012)
PKH
1.292***
1.055***
1.337***
1.149***
1.352***
1.223***
(0.023)
(0.024)
(0.033)
(0.033)
(0.043)
(0.039)
KKS
0.827***
0.659***
0.8***
0.701***
0.928***
0.72***
(0.023)
(0.024)
(0.033)
(0.033)
(0.042)
(0.039)
_cons
-0.106
-1.719***
-1.021
-1.432**
0.037
-1.679**
(0.422)
(0.433)
(0.637)
(0.636)
(0.868)
(0.77)
Observations
63204
66177
29248
30458
19552
24528
Pseudo R2
0.135
0.085
0.14
0.103
0.152
0.112
Standard errors are in parentheses
*** p<.01, ** p<.05, * p<.1
Source: calculated by authors
Table 4: Results of the Propensity Score Estimation
Finally, the study employs the Nearest Neighbor oversampling (2-NN) matching method to
address the imbalance between treatment and control groups. A caliper matching method is
combined with NN oversampling as it can decrease the percentage bias by setting a
maximum tolerance level for propensity score distance (Gertler et al., 2016). Consequently,
for all models of this study, the matching method has small bias percentages of 1% to 3%.
The Average Treatment Effect on Treated (ATT) value, as in equation (3), represents the
average difference between the treatment and control groups, often called risk difference
(Austin & Stuart, 2017). It is expected that there is a significant average difference between
the treatment group (receiving the PIP program) and the control group (not receiving the PIP
program). The results of ATT are presented in Table 5.
Education Level
Treatment
Mean of
Matched
Treated
Mean of
Matched
Controls
ATT
Standard
Error
t-stat
Elementary School 2019
PIP
0.002
0.004
-0.002
0.001
-3.17***
Elementary School 2021
0.002
0.003
-0.001
0.001
-1.64
Junior High School 2019
0.009
0.017
-0.007
0.002
-3.60***
Junior High School 2021
0.011
0.02
-0.008
0.002
-4.09***
Senior High School 2019
0.016
0.021
-0.005
0.003
-1.56
Senior High School 2021
0.016
0.022
-0.006
0.003
-2.13**
*** p<0.01 **p<0.05, *p<0.1
Source: calculated by authors
Table 5: Results of the Average Treatment on Treated (ATT)
Table 5 shows that at the elementary school level, the PIP program significantly affected the
probability of dropping out in 2019. However, the program did not exert the same effect in
2021. At the junior high school level, the PIP program affected the probability of dropping
out both in 2019 and 2021. In contrast, for the senior high school level, the PIP program did
not affect the dropout probability in 2019 but significantly affected the dropout probability in
2021. Notably, the PIP program exerted its strongest influence on dropping out at the junior
high school level, persisting before and during the pandemic.
The dropout probability varies by student's educational level. One of the main factors
influencing dropping out is educational expenses. High educational expenses are associated
with an increased likelihood of dropout, particularly at higher educational levels. The lowest
educational expenses are observed among elementary students, while the highest are among
senior/vocational high school students. In 2019, the estimated personal cost for elementary
students in Indonesia was IDR 3,147,000 (approximately USD 197), for junior high students
was IDR 4,245,000 (approximately USD 265), and for senior high students was IDR
7,457,000 (approximately USD 466) (Zamjani et al., 2020).
The PIP program is government education assistance from the demand side to lower personal
educational expenses. In addition to the PIP program, the government provides educational
assistance from the supply side through the School Operational Assistance (BOS). This
program provides support to all-level public schools with operational expenses, including
maintaining school facilities, purchasing teaching aids, the payment of honorariums for non-
permanent teachers and staff, and other school operational expenses. However, according to
Zamjani et al. (2020), the allocated operational assistance was insufficient to cover the
school's non-personnel operational expenses. For instance, in 2019, the government allocated
IDR 800,000 per student (approximately USD 50) for elementary schools. However, the
average school non-personnel operational expenses were IDR 996,000 per student
(approximately USD 62). A similar discrepancy is observed in the case of junior high
schools. While the allocated operational assistance was IDR 1,000,000 per student
(approximately USD 63), the average operational expenses were IDR 1,539,000 per student
(approximately USD 96). For senior high school, the allocated operational assistance was
IDR 1,400,000 per student (approximately USD 88), while the average operational expenses
were IDR 1,651,000 per student (approximately USD 197). This suggests that both PIP and
BOS may not fully cover the expenses incurred by students.
The PIP program affects differently to the probability of dropout rate at every educational
level. At elementary school, the dropout rate was generally relatively low, at 0.37% in 2019
and 0.12% in 2021, due to several reasons (Direktorat Statistik Kesejahteraan Rakyat, 2021;
Statistik, 2019). First, the government has implemented a mandatory education program for
children aged 7-12. Second, the number of elementary schools is relatively higher compared
to junior and senior high schools, and hence, it is easier to access elementary schools. Third,
the personal unit costs for elementary school students are relatively low compared to other
educational levels, so there was less barrier for low-income families to send their children to
school. According to the Directorate of Population's Welfare Statistics (Direktorat Statistik
Kesejahteraan Rakyat), there was no significant change in school participation rates at the
elementary schools during the pandemic, which indicates the community's awareness that
elementary education is fundamental (Direktorat Statistik Kesejahteraan Rakyat, 2021).
Therefore, the PIP program has no impact on the probability of dropping out of elementary
school during the pandemic in 2021.
At the junior high school level, the PIP program affected the probability of dropping out both
in 2019 and 2021. In general, students at junior high schools were vulnerable to student
dropout, as this is a transition phase to high school education (Cameron, 2009). The dropout
rate at the junior high schools was higher compared to the elementary school level, with
1.07% in 2019 and 0.90% in 2021. Access to junior high school is more challenging than
elementary school due to the smaller number of junior high schools. Students from low-
income families are less likely to continue to high school (Direktorat Statistik Kesejahteraan
Rakyat, 2021; Statistik, 2019). The PIP program can alleviate the school expenses for
students from low-income families at the junior high school level, thereby reducing dropouts
before and during the pandemic. These findings are consistent with previous studies that
found educational assistance significantly affects dropout rates at the junior high school level
and has a more significant impact compared to the elementary school level (Cameron, 2009;
Yulianti et al., 2015; Listiyanto & Qibthiyyah, 2022).
At the senior high school level, the PIP program did not affect the dropout probability in
2019. However, it significantly affected the dropout probability in 2021. The dropout rate at
the high school level was higher than at the elementary and junior high school levels. In
2019, it was 1.76%, while in 2021 it was 1.12% (Direktorat Statistik Kesejahteraan Rakyat,
2021; Statistik, 2019). It is more challenging to access senior high school education than it is
to access elementary and junior high school education. The availability of senior high schools
is not uniform across Indonesia. There is an uneven distribution of high schools among
regions in Indonesia. The costs associated with attending senior high school extend beyond
the purchase of school equipment. Student personal expenses for senior high school are not
only for school equipment but also for transportation and boarding costs for students living
far from the schools (Baird et al., 2014). The transportation cost tends to increase as students'
education level is higher, and the location of schools for higher levels tends to be farther
away (Zamjani et al., 2020). The additional educational and transportation expenses costs can
cause negative school participation due to financial and non-financial access barriers
(Corrales-Herrero et al., 2021).
Although the PIP cash transfer is higher for senior high school students, the amount is
insufficient to cover the majority of high school education expenses. Furthermore, the
transferred cash for senior high school students does not guarantee that students will remain
in school (Churchill et al., 2021). The higher personal educational expenses, the location of
the schools, and the limited transferred cash for the PIP program may have resulted in an
insignificant influence of the PIP on the dropout probability among poor and vulnerable high
school students prior to the pandemic in 2019. However, during the pandemic in 2020, the
PIP program was significant in influencing the dropout probability among poor and
vulnerable high school students. High school students have a more significant opportunity to
attend school during the pandemic because there are no travel constraints compared to before
the pandemic in 2019. During the pandemic, students did not need to spend on transportation,
practical fees, excursions, and some additional costs related to face-to-face extracurricular
activities, thus reducing the financial burdens that students must meet. Consequently, PIP
assistance can markedly diminish the probability of students dropping out of high school or
vocational high school during a pandemic.
Conclusions
The findings of this study indicate that the PIP program can reduce the probability of school
dropout among recipients at each educational level. This study finds that there is a significant
negative relationship between the PIP program and the probability of school dropout among
students from low-income families at the elementary school in 2019, junior high school in
2019 and 2021, and senior high school in 2021. However, the results were found to be
insignificant at the elementary school in 2021 and at the senior high school in 2019. The
impact of PIP on the probability of school dropout among students from low-income families
varies across educational levels and periods. PIP has the most significant impact at the junior
high school level. The varying significance of results at elementary school may be attributed
to the relatively low dropout rates, both before and during the pandemic. In contrast, the
differing results at the senior high school level may be attributed to the considerable difficulty
in accessing education, higher education costs, and insufficient education assistance to meet
the needs of senior high school students, which may lead to a higher likelihood of dropout
among students. Consequently, the provision of PIP at the senior high school in 2019 did not
result in a significant reduction in the dropout rate.
This study is limited by its cross-sectional design, which precludes direct comparisons with
data from the same individuals in 2019 and 2021. A longitudinal analysis would be preferable
to determine the extent of the program's impact over a certain period. Furthermore, the PSM
method may lack specificity in determining propensity scores, making the resulting ATT
values sensitive to the covariates used in score determination. Accordingly, the authors
should consider the covariates utilized in the model, ensuring that confounding variables and
other controls are more representative. Moreover, this study was unable to encompass
variables from the supply side of education because the confounding variables that can link
treatment and outcome studied are limited to family background characteristics on the
demand side. Future research can combine variables from education's demand and supply
sides to obtain a comprehensive analysis.
Appendices
Variable
Observation
Mean
Standard
Deviation
Minimum
Maximum
2019
Dropout
(1=school dropout)
63,204
0.003
0.058
0
1
PIP
(1=receiving PIP)
63,204
0.283
0.451
0
1
Cap (per capita expenditure)
63,204
13.116
0.317
11.706
13.572
educ KRT
(1=senior high school & college)
63,204
0.291
0.454
0
1
work KRT
(1=formal)
63,204
0.297
0.457
0
1
Totalchild
63,204
3.131
1.446
1
18
PKH
(1=receiving PKH)
63,204
0.328
0.469
0
1
KKS
(1=having KKS)
63,204
0.269
0.444
0
1
2021
Dropout
(1=school dropout)
66,177
0.003
0.053
0
1
PIP
(1=receiving PIP)
66,177
0.236
0.425
0
1
Cap (per capita expenditure)
66,177
13.201
0.311
11.668
13.65
educ KRT
(1=senior high school & college)
66,177
0.325
0.469
0
1
work KRT
(1=formal)
66,177
0.313
0.464
0
1
Totalchild
66,177
3.056
1.404
1
15
PKH
(1=receiving PKH)
66,177
0.32
0.466
0
1
KKS
(1=having KKS)
66,177
0.234
0.423
0
1
Source: National Socioeconomic Survey (SUSENAS) 2019 and 2021, processed by authors
Appendix A: Summary Statistics for Elementary School
Variable
Observation
Mean
Standard
Deviation
Minimum
Maximum
2019
Dropout
(1=school dropout)
29,248
0.012
0.11
0
1
PIP
(1=receiving PIP)
29,248
0.284
0.451
0
1
Cap (per capita expenditure)
29,248
13.135
0.306
11.706
13.572
educ KRT
(1=senior high school & college)
29,248
0.261
0.439
0
1
work KRT
(1=formal)
29,248
0.276
0.447
0
1
Totalchild
29,248
3.184
1.447
1
12
PKH
(1=receiving PKH)
29,248
0.372
0.483
0
1
KKS
(1=having KKS)
29,248
0.301
0.459
0
1
2021
Dropout
(1=school dropout)
30,458
0.015
0.121
0
1
PIP
(1=receiving PIP)
30,458
0.262
0.44
0
1
Cap (per capita expenditure)
30,458
13.217
0.303
11.884
13.65
educ KRT
(1=senior high school & college)
30,458
0.302
0.459
0
1
work KRT
(1=formal)
30,458
0.286
0.452
0
1
Totalchild
30,458
3.181
1.446
1
15
PKH
(1=receiving PKH)
30,458
0.384
0.486
0
1
KKS
(1=having KKS)
30,458
0.28
0.449
0
1
Source: National Socioeconomic Survey (SUSENAS) 2019 and 2021, processed by authors
Appendix B: Summary Statistics for Junior High School
Variable
Observation
Mean
Standard
Deviation
Minimum
Maximum
2019
Dropout
(1=school dropout)
19,552
0.018
0.134
0
1
PIP
(1=receiving PIP)
19,552
0.241
0.428
0
1
Cap (per capita expenditure)
19,552
13.177
0.291
11.787
13.572
educ KRT
(1=senior high school & college)
19,552
0.278
0.448
0
1
work KRT
(1=formal)
19,552
0.281
0.45
0
1
Totalchild
19,552
3.18
1.437
1
12
PKH
(1=receiving PKH)
19,552
0.352
0.478
0
1
KKS
(1=having KKS)
19,552
0.291
0.454
0
1
2021
Dropout
(1=school dropout)
24,528
0.02
0.141
0
1
PIP
(1=receiving PIP)
24,528
0.216
0.411
0
1
Cap (per capita expenditure)
24,528
13.243
0.295
11.846
13.65
educ KRT
(1=senior high school & college)
24,528
0.308
0.462
0
1
work KRT
(1=formal)
24,528
0.275
0.447
0
1
Totalchild
24,528
3.169
1.441
1
12
PKH
(1=receiving PKH)
24,528
0.383
0.486
0
1
KKS
(1=having KKS)
24,528
0.285
0.451
0
1
Source: National Socioeconomic Survey (SUSENAS) 2019 and 2021, processed by authors
Appendix C: Summary Statistics for Senior High School
References
Al-Samarrai, S., & Peasgood, T. (1998). Educational attainments and household
characteristics in Tanzania. Economics of Education Review, 17(4).
https://doi.org/10.1016/S0272-7757(97)00052-6
Alcaraz, M. (2020). Beyond Financial Resources: The Role of Parents' Education in
Predicting Children's Educational Persistence in Mexico. International Journal of
Educational Development, 75, 102188.
https://doi.org/10.1016/j.ijedudev.2020.102188
Ali, M. (2009). Pendidikan Untuk Pembangunan Nasional. In Jakarta : Grasindo.
Anindita, A., & Sahadewo, G. A. (2018). Lighten The Burden : Assessing The Impact Of The
Indonesian For-Poor - Students Cash Transfer On Spending. SSRN Electronic
Journal. https://doi.org/10.2139/ssrn.3180062
Austin, P. C. (2010). Statistical criteria for selecting the optimal number of untreated subjects
matched to each treated subject when using many-to-one matching on the propensity
score. American Journal of Epidemiology, 172(9), 1092–1097.
https://doi.org/10.1093/aje/kwq224
Azevedo, V., & Robles, M. (2013). Multidimensional Targeting: Identifying Beneficiaries of
Conditional Cash Transfer Programs. Social Indicators Research, 112(2), 447–475.
https://doi.org/10.1007/s11205-013-0255-5
Badan Pusat Statistik. https://www.bps.go.id/
Baird, S., Ferreira, F. H. G., Özler, B., & Woolcock, M. (2014). Conditional, unconditional,
and everything in between a systematic review of the effects of cash transfer
programmes on schooling outcomes. Journal of Development Effectiveness, 6(1).
https://doi.org/10.1080/19439342.2014.890362
Berapa Jumlah Anak Putus Sekolah Di Indonesia?: Databoks. Pusat Data Ekonomi dan
Bisnis Indonesia. (n.d.). Retrieved February 2, 2023, from
https://databoks.katadata.co.id/datapublish/2022/03/16/berapa-jumlah-anak-putus-
sekolahdi-indonesia
Cameron, L. (2009). Can a public scholarship program successfully reduce school dropouts
in a time of economic crisis? Evidence from Indonesia. Economics of Education
Review, 28(3), 308–317. https://doi.org/10.1016/j.econedurev.2007.09.013
Canelas, C., & Niño-Zarazúa, M. (2019). Schooling and Labor Market Impacts of Bolivia's
Bono Juancito Pinto Program. Population and Development Review, 45(S1), 155–
179. https://doi.org/10.1111/padr.12270
Churchill, S. A., Iqbal, N., Nawaz, S., & Yew, S. L. (2021). Unconditional cash transfers,
child labour, and education: theory and evidence. Journal of Economic Behavior and
Organization, pp. 186, 437–457. https://doi.org/10.1016/j.jebo.2021.04.012
Corrales-Herrero, H., Him Camaño, M., Miranda-Escolar, B., & Ogando Canabal, O. (2021).
Anti-poverty transfers and school attendance: Panama's Red de Oportunidades.
International Journal of Social Economics, 48(2), 204–220.
https://doi.org/10.1108/IJSE-05-2020-0336
Dampak Pandemi, Mayoritas Anak Indonesia putus Sekolah Karena ekonomi: Databoks.
Pusat Data Ekonomi dan Bisnis Indonesia. (n.d.). Retrieved February 28, 2023, from
https://databoks.katadata.co.id/datapublish/2021/04/08/dampak-pandemi-mayoritas-
anakindonesia-putus-sekolah-karena-ekonomi
Dana Bos 2021: Syarat Pencairan Dana Bos tahap 2 Dan Besaran dana. Kompas.com.
https://www.kompas.com/edu/read/2021/05/22/083043971/dana-bos-2021-syarat-
pencairan-dana-bos-tahap-2-dan-besaran-dana
Direktorat Statistik Kesejahteraan Rakyat. (2021). Statistik Pendidikan 2021. Badan Pusat
Statistik. https://www.bps.go.id/
Edo, M., & Marchionni, M. (2019). The impact of a conditional cash transfer programme on
education outcomes beyond school attendance in Argentina. Journal of Development
Effectiveness, 11(3), 230–252. https://doi.org/10.1080/19439342.2019.1666898
Gertler, P. J., Martinez, S., Premand, P., Rawlings, L. B., & Vermeersch, C. M. J. (2016).
Impact Evaluation in Practice, Second Edition. The World Bank.
https://doi.org/10.1596/978-1-4648-0779-4
Glewwe, P., & Kassouf, A. L. (2012). The impact of the Bolsa Escola/Familia conditional
cash transfer program on enrollment, dropout rates, and grade promotion in Brazil.
Journal of Development Economics, 97(2), 505–517.
https://doi.org/10.1016/j.jdeveco.2011.05.008
Glewwe, P., Lambert, S., & Chen, Q. (2020). Education production functions: Updated
evidence from developing countries. In The Economics of Education: A
Comprehensive Overview. https://doi.org/10.1016/B978-0-12-815391-8.00015-X
Hafrienda, R., Candradewini, C., & Munajat, M. D. E. (2023). Efektivitas Program Indonesia
Pintar pada Jenjang SMA Negeri di Kota Bukittinggi. Jurnal Administrasi Negara
14(2), 697. https://doi.org/10.24198/jane.v14i2.45140
Hanushek. (1979). Conceptual and Empirical Issues in the Estimation of Educational
Production Functions. The Journal of Human Resources.
https://doi.org/10.1186/1478-7547-10-9
Kaidah, S., & Ruslan, R. (2022). Dampak program Indonesia pintar terhadap pendidikan
Anak Pada Keluarga Miskin di Desa Lokop kecamatan Serbajadi Kabupaten Aceh
Timur. Jurnal Geuthèë: Penelitian Multidisiplin, 5(3), 312.
https://doi.org/10.52626/jg.v5i3.193
Kharisma, B., Satriawan, E., & Arsyad, L. (2017). The impact of social safety net
scholarships program to school dropout rates in Indonesia: The intention-to-treat
analysis. The Journal of Developing Areas, 51(4), 303–316.
https://doi.org/10.1353/jda.2017.0103
Khiem, P. H., Linh, D. H., Tai, D. A., & Dung, N. D. (2020). Does tuition fee policy reform
encourage poor children's school enrolment? Evidence from Vietnam. Economic
Analysis and Policy, 66, 109–124. https://doi.org/10.1016/j.eap.2020.03.001
Listiyanto, H. A., & Qibthiyyah, R. M. (2022). The Impact of School Voucher Program on
School Dropouts in Indonesia. Jurnal Pendidikan Bisnis Dan Manajemen, 8(2), 113–
127. http://dx.doi.org/10.17977/um003v8i22022p113
Lovenheim, M., & Turner, S. (2018). Economics of Education. Worth Publishers. (Vol. 13,
Issue 1).
Ministry of Education, Culture, Research, and Technology of Indonesia. (2017). Efektifitas
Program Indonesia Pintar (PIP) dan kontribusinya terhadap kedisiplinan dan
prestasi belajar siswa. https://pskp.kemdikbud.go.id/
Ministry of Education, Culture, Research, and Technology of Indonesia. (2020). Biaya
Satuan & Lini Masa Pengelolaan Program Indonesia Pintar.
https://repositori.kemdikbud.go.id/
Ministry of Education, Culture, Research, and Technology of Indonesia. (2021). Statistik
Data SMA. Kementerian Pendidikan dan Kebudayaan.
http://publikasi.data.kemdikbud.go.id/
Ministry of Education, Culture, Research, and Technology of Indonesia. (2021). Statistik
Sekolah Dasar. Kementerian Pendidikan dan Kebudayaan.
http://publikasi.data.kemdikbud.go.id/
Ministry of Education, Culture, Research, and Technology of Indonesia. (2021). Statistik
SMK. Kementerian Pendidikan Dan Kebudayaan.
http://publikasi.data.kemdikbud.go.id/
Ministry of Education, Culture, Research, and Technology of Indonesia. (2021). Statistik
SMP 2020/2021. Kementerian Pendidikan dan Kebudayaan.
http://publikasi.data.kemdikbud.go.id/
Ministry of Education, Culture, Research, and Technology of Indonesia. (2023, January 4).
Inilah peserta Didik Yang layak Menerima Dana Bantuan Pip. Pusat Layanan
Pembiayaan Pendidikan. Retrieved March 6, 2023, from
https://puslapdik.kemdikbud.go.id/inilah-peserta-didik-yang-layak-menerima-
danabantuan-pip/
Mo, D., Zhang, L., Yi, H., Luo, R., Rozelle, S., & Brinton, C. (2013). School Dropouts and
Conditional Cash Transfers: Evidence from a Randomised Controlled Trial in Rural
China's Junior High Schools. Journal of Development Studies, 49(2), 190–207.
https://doi.org/10.1080/00220388.2012.724166
Peruffo, M., & Ferreira, P. C. (2017). The Long-Term Effects of Conditional Cash Transfers
on Child Labor and School Enrollment. Economic Inquiry, 55(4), 2008–2030.
https://doi.org/10.1111/ecin.12457
Program Indonesia Pintar. https://pip.kemdikbud.go.id/home_v1
Purba, R. H. F. (2018). Impact Evaluation of Indonesia Conditional Cash Transfer Program
(BSM) on Student Achievement. European Journal of Economics and Business
Studies, 10(1), 104. https://doi.org/10.26417/ejes.v10i1.p104-115
Setyadharma, A. (2018). Government's Cash Transfers And School Dropout In Rural Areas.
Jejak, 11(2), 447–461. https://doi.org/10.15294/jejak.v11i2.16125
Sihombing, P. R., Pratiko, W., & Alkahfi, C. (2022). Aplikasi Model Diffence in Difference
Pada Regresi Binomial Logistik Aplikasi Model Diffence in Difference Pada Regresi
Binomial Logistik ( Studi Kasus : Implementasi Program Indonesia Pintar Terhadap
Status Sekolah Anak ) untuk bersekolah atau putus sekolah. Jurnal Ilmu Ekonomi dan
Sosial. https://doi.org/10.22441/jies.v11i2.16615
Statistik, B. P. (2019). Statistik Pendidikan 2019. 73–92. https://www.bps.go.id/
UNESCO. (2019). New methodology shows that 258 Million children, adolescents, and
youth are out of school. http://uis.unesco.org
UNICEF. (2017). Improving Education Participation. https://www.unicef.org/
Uriyalita, F., Syahrodi, J., & Sumanta. (2020). Evaluasi program Indonesia Pintar (PIP)
Telaah Tentang Aksesibilitas, Pencegahan Dan Penanggulangan Anak putus Sekolah
di Wilayah Urban Fringe Harjamukti, cirebon. Edum Journal, 3(2), 179–199.
https://doi.org/10.31943/edumjournal.v3i2.69
Wils, A., Sheehan, P., & Shi, H. (2019). Better secondary schooling outcomes for adolescents
in low- and middle-income countries: Projections of cost-effective approaches.
Journal of Adolescent Health, 65(1). https://doi.org/10.1016/j.jadohealth.2019.03.024
Yulianti, N. R., Bedi, A. S., & Bergeijk, P. V. (2015). The Functioning and effect of a cash
transfer program in Indonesia. International Institute of Social Studies.
https://thesis.eur.nl/pub/33409/
Zahimu, H. (2019). Evaluasi Program Indonesia Pintar Pada Dinas Pendidikan Dan
Kebudayaan Kota Baubau Tahun 2017. Jurnal Studi Kepemerintahan, 2(1), 37–46.
https://doi.org/10.35326/kybernan.v2i1.469
Zamjani, I., Herlinawati, Perdana, N. S., Widiputera, F., & Azizah, S. N. (2020). Pendanaan
Pendidikan Dasar dan Menengah pada Satuan Pendidikan Formal.
https://repositori.kemdikbud.go.id/
Contact emails: nismaanggarasamalo@gmail.com
thia.jasmina@ui.ac.id
