DECEMBER 2020

DISCUSSION PAPER SERIES NO. 2020-59

Giving Cash to the Poor:

A Study of Pantawid Pamilya Cash Grants
Generosity, Frequency, and Modality

Kris Ann M. Melad, Nina Victoria V. Araos, and Aniceto C. Orbeta Jr.

The PIDS Discussion Paper Series constitutes studies that are preliminary and subject to further revisions. They are being circulated in a limited number of copies only for
purposes of soliciting comments and suggestions for further refinements. The studies under the Series are unedited and unreviewed. The views and opinions expressed are
those of the author(s) and do not necessarily reflect those of the Institute. Not for quotation without permission from the author(s) and the Institute.

CONTACT US:
RESEARCH INFORMATION DEPARTMENT
Philippine Institute for Development Studies

18th Floor, Three Cyberpod Centris - North Tower publications@[Link]

[Link]
EDSA corner Quezon Avenue, Quezon City, Philippines (+632) 8877-4000
 Giving Cash to the Poor: A Study of Pantawid Pamilya
Cash Grants Generosity, Frequency, and Modality

Kris Ann M. Melad

Nina Victoria V. Araos
Aniceto C. Orbeta Jr.

PHILIPPINE INSTITUTE FOR DEVELOPMENT STUDIES

December 2020
Abstract

This study assesses the payment system of the Pantawid Pamilyang Pilipino Program (4Ps or
Pantawid Pamilya) in terms of the amount, frequency, and mode of payment delivery of the
cash transfers. The study is an early attempt of the Institute to examine the Pantawid Pamilya
implementation and recommend improvements in its payment system in light of the enactment
of RA 11310 which institutionalized the program and identified PIDS to conduct regular
assessments on the program’s implementation and impact. The underlying framework for the
study is based on the idea that the amount of cash grants and its delivery system should be able
to encourage desired behaviors among beneficiaries and help achieve the program’s target
outcomes. The study answers the research objectives using a combination of qualitative and
quantitative methods of analysis that include a review of the available literature on payment
systems of international CCTs; re-assessment of the cash grant values in relation to original
intended levels and devaluation due to inflation; assessment of the implementation of the
payment system including challenges, facilitating factors, and alternative ideas; and an
assessment of the impact of the design features of the payment system on select outcome
variables. The study uses administrative data of the Department of Social Welfare and
Development (DSWD), survey data from the program’s 3rd impact evaluation, interviews of
key program implementers in the payment system, focus group discussions with beneficiaries,
and an online survey focused on the payment delivery aspect of the Social Amelioration
Program (SAP) among Pantawid Pamilya beneficiaries. The results show that the program
needs to be more responsive in adjusting the benefit level of the cash grants vis-à-vis inflation
and threats to welfare of the beneficiaries. With regard to payment frequency, there is no strong
demand by the beneficiaries for more frequent payouts. Changes in frequency of payment
should also be carefully examined to know if benefits outweigh additional costs in the
operations. Although the payment delivery system has improved through the years primarily
due to the conversion of mode of payment to cash cards, some opportunities to improve the
payment delivery of the program are still present. Finally, the experience with the SAP
distribution demonstrated the ease and speed of distributing assistance in times of emergencies
when recipients (e.g., 4Ps beneficiaries) are identified and a payment system in place.

Keywords: Pantawid Pamilya, Conditional Cash Transfer, cash grants, payment

system
 Table of Contents

1. Introduction........................................................................................................ 1
1.1. Background of the study ...................................................................................... 1
1.2. Objectives ............................................................................................................ 3
1.3. Research questions ............................................................................................. 4
2. Review of Related Literature ............................................................................. 4
2.1. Role of payment system in success of CCTs all over the world ........................... 4
2.2. Local assessments of the 4Ps Payment system .................................................. 6
3. Research Design and Methodology ................................................................. 7
3.1. Conceptual Framework ........................................................................................ 7
3.2. Research Design ................................................................................................. 9
3.3. Profile of Respondents by Source...................................................................... 15
4. Results ............................................................................................................. 23
4.1. The 4Ps Program Payment System ................................................................... 23
4.2. Assessing the Benefit Level ............................................................................... 26
4.3. Payment Frequency........................................................................................... 31
4.4. Payment Modality .............................................................................................. 35
4.5. 4Ps Payments System and SAP Implementation ............................................... 46
5. Summary and Recommendations .................................................................. 53
6. Bibliography..................................................................................................... 56
7. Appendix: Differential impact of Pantawid Pamilya by mode of payment ... 58

List of Tables

Table 1. Distribution of study sites by island cluster ............................................................ 11

Table 2. Target respondents for FGDs and KIIs .................................................................. 11
Table 3. Distribution of FGD participants by island cluster .................................................. 15
Table 4. Profile of FGD participants .................................................................................... 16
Table 5. Household composition of FGD participants .......................................................... 16
Table 6. Experience and caseload profile of C/ML and MRB respondents in KIIs ............... 18
Table 7. Household composition and estimated income of households at baseline, by
treatment assignment.......................................................................................................... 19
Table 8. Distribution of IE3 treatment households by mode of payment .............................. 19
Table 9. Frequency of payout and amount of grants received by IE3 treatment households
included in the analysis (n=2,613) ....................................................................................... 20
Table 10. Descriptive statistics of respondents included in the analysis .............................. 22
Table 11. Illustration of the bimonthly payment cycle of Pantawid Pamilya ......................... 32
Table 12. Share of Cash Card and OTC payment modes among 4Ps households, 2016-
2020.................................................................................................................................... 35
Table 13. Share of Cash Card and OTC payment modes among 4Ps households, by region,
2020.................................................................................................................................... 35
Table 14. Summary of results on subgroup analysis of 4Ps impact on child health by mode
of cash grant payment......................................................................................................... 39
Table 15. Summary of results on subgroup analysis of 4Ps impact on education by mode of
cash grant payment ............................................................................................................ 40
Table 16. Summary of results on subgroup analysis of 4Ps impact on household expenditure
and income by mode of cash grant payment ....................................................................... 44
Table 17. Summary of results on subgroup analysis of 4Ps impact on labor outcomes by
mode of cash grant payment ............................................................................................... 45
Table 18. Status of enrollment and possession of LBP cash card among 4Ps online survey
respondents ........................................................................................................................ 46
Table 19. How respondents usually receive their 4Ps cash grants ...................................... 47
Table 20. Status of enrollment and possession of LBP cash card among 4Ps online survey
respondents ........................................................................................................................ 47
Table 21. Average amount of SAP grant received by frequency of receipt .......................... 48
Table 22. Correct SAP subsidy per region ......................................................................... 49
Table 23. How respondents received their SAP cash grants ............................................... 49
Table 24. Average and median transportation cost and time spent during receipt of SAP
grants.................................................................................................................................. 49
Table 25. Top expenditure items where beneficiaries used their SAP grant (n=759) ........... 50
Table 26. Problems experienced in SAP implementation (n=759) ....................................... 51
Table 27. Sources of information on SAP payout schedule for those who have not received
SAP (n=27) ......................................................................................................................... 51
Table 28. Problems experienced by non-4Ps in SAP implementation (n=886) .................... 52
Table 29. Proportion of 4Ps beneficiaries in the parent group that received SAP according to
Parent Leaders (n=612) ...................................................................................................... 53

List of Figures

Figure 1. Conceptual Framework .......................................................................................... 9

Figure [Link] Discontinuity Design ........................................................................... 13
Figure 3. Frequency distribution of online survey respondents by date of response ............ 20
Figure 4. Completion rate by question item in online survey................................................ 21
Figure 5. Distribution of online survey respondents by region ............................................. 22
Figure 6. Role of payment system in Pantawid Pamilya Program Cycle .............................. 23
Figure 7. Fund flow of the Pantawid Pamilya Payment System ........................................... 24
Figure 8. Selection process of payment conduits ................................................................ 25
Figure 9. Timeline of changes in value of cash grants ......................................................... 27
Figure 10. Share of the real value of grants over 2006 poverty threshold, by year .............. 28
Figure 11. Comparison of 4Ps and international CCTs in terms of generosity in bottom 20%
........................................................................................................................................... 28
Figure 12. How beneficiaries spend their 4Ps cash grants .................................................. 30
Figure 13. Beneficiaries’ preferred payment frequency based on the FGDs ........................ 33
Figure 14. Compliance Verification Cycle ............................................................................ 34
 Giving cash to the poor: A study of Pantawid Pamilya cash grants
generosity, frequency, and modality

Kris Ann M. Melad, Nina Victoria V. Araos and Aniceto C. Orbeta Jr.1

1. Introduction

1.1. Background of the study

The Pantawid Pamilyang Pilipino Program (Pantawid Pamilya or 4Ps) is a human capital
development program that aims to arrest the intergenerational transmission of poverty by
providing cash grants to poor households, conditional to practice of certain desired behaviors.
Specifically, the program provides cash grants based on the poor households’ compliance with
conditionalities on health, education, and family development. These conditions require that
children attend school 85% of the time and eligible household members regularly avail
appropriate basic health care services. In addition, parents in the household must attend
monthly Family Development Sessions (FDS) or learning seminars.

In April 2019, Republic Act 11310 was enacted paving way to the institutionalization of
Pantawid Pamilya. The Philippine Institute for Developments Studies (PIDS) has been tasked
to conduct periodic assessments of the program in terms of its impact on the beneficiaries and
status of program implementation. The Institute is also tasked to recommend adjustments in
the amount of cash grants provided to the beneficiaries every six years. This research is an
early effort of the Institute to assess the overall payment system of the program considering the
tasks provided in the law. The study looks at three aspects of the payment system: (1) benefit
level or the amount of cash grants received by beneficiaries; (2) frequency that the cash grants
are received; and (3) modality or the channel through which the grants are received by the
beneficiaries.

In terms of benefit level, prior to the passing of the law, compliant household beneficiaries
receive Php500.00 per month for health grants. For education, each compliant child in
preschool, kindergarten, and elementary (K-6) receives Ph300.00 per month while each
compliant child in high school receives Php500.00 per month. The education grant can only be
received by a maximum of three (3) children for 10 months each year for each household.
Starting only in 2017, an amount of Php600.00 per month is given to the households as “rice
subsidy” – an additional benefit supposedly aimed to improve food security among
beneficiaries.

The amount of program grants generally remained at the same levels since its pilot
implementation in 2008, apart from the addition of the rice subsidy and the differentiation of
cash grants for high school students in 20142. From the intended level of benefits in the
program’s first year of implementation, a cumulative reduction in the real value of cash grants
by around 30% by 2016 is expected due to inflation. The provision of the rice subsidy in 2017
helped recover this loss in value of the cash grants, but this could also mean that no real increase
in benefits was experienced by the beneficiaries for food security. The reduction in the real

1
Supervising Research Specialist, Research Analyst, and Senior Research Fellow, respectively, at the Philippine Institute for
Development Studies; The usual disclaimer applies.
2
From 2008 to mid-2014, only children 0-14 were eligible beneficiaries of the program. Coverage has been expanded to include
children 15 to 18 years old in 2014 and grants have increased to Php500 for children in high school

1
value of grants has been cited as one of the possible reasons for the lack of program impact on
some outcomes including total per capita expenditure, and prevalence of working children
based on the past impact evaluation studies of the program (DSWD, 2014; DSWD and World
Bank, 2014). A study done by Albert (2015) on the extension of age coverage of the program
also recommended increasing the amount of education grants for elementary students from
Php300 to Php350, and for high school students from Php500 to Php600, to help improve
completion and cohort survival rates among the beneficiary children.

With the passing of the law, the health grants increased from Php500 to Php750 and education
grants for children in Senior high school (SHS) increased from Php500 to Php7003. Meanwhile,
grants for children in elementary and junior high school remain at Php300 and Php500,
respectively. These amounts are expected to be adjusted every six years based on
recommendations of PIDS through the program National Advisory Council – its policy making
body composed of DSWD and other agencies.

On frequency of receipt of benefits, the program beneficiaries currently receive cash grants
every two months based on a cycle of compliance monitoring conducted by the program
implementers. This is relatively faster than during the first three years of program
implementation where payouts were conducted every quarter.

In terms of impact, evidence on other CCTs suggest that more frequent provisions of cash
grants contribute to better consumption smoothing of beneficiaries while less frequent and
lump-sum provision of grants result in higher investments and asset accumulation among
households (Bastagli, et al., 2016). Given that the main objective of the program is to address
gaps in the food consumption and education and health expenses of the household, these
findings suggest that a more frequent payouts should be desired. However, it is apparent that
increasing payment frequencies will incur additional operations cost for the program, and
whether these additional costs will be cost-effective in terms of the marginal benefits is still
unclear. For example, evidence from the systematic review done by Garcia and Saavedra
(2012) point that monthly provision of cash grants is associated with smaller impact on
educational impacts compared to the less frequent provisions.

The third aspect examined in this study is the payment modality or how beneficiaries receive
their program benefits. Cash grants are delivered to the beneficiaries by the Land Bank of the
Philippines (LBP) – an Authorized Government Depository Bank (AGDB). Based on the
payroll advise generated by the DSWD, the LBP pays the cash grants to the beneficiaries
through two modalities. The first modality is through the Cash Cards that can be used to
withdraw the money from Automated Teller Machines (ATM), while the second modality
involves subcontracting of payment conduits such as money exchange centers, cooperatives,
and other firms to deliver the grants in cash to beneficiaries “Over the Counter” (OTC).

Timeliness and predictability of the payments are affected by the modality of cash grant
payments. Beneficiaries with cash cards, on the average, receive their benefits earlier than those
that receive their benefits through OTC transactions. While cash card holders have immediate
access to their benefits once credited by LBP, OTC payouts require successful bidding of a
payment conduit, and numerous logistic requirements like OTC venue, scheduling, security,
and transportation of beneficiaries to and from the OTC venue, among others. Given this, the
DSWD and LBP have been trying to increase the proportion of household beneficiaries with

3
(RA 11310 : An Act Institutionalizing the Pantawid Pamilyang Pilipino Program)

2
cash cards. From the 44% of cash card holders in 2017, the DSWD reports that almost 75% of
the beneficiaries are already cash card holders as of July 2019 (Pantawid Pamilya NPMO, July
12, 2019). However, this still means that until one year ago, one million out of the four million
beneficiaries of the program were still vulnerable to delays and unpredictability of payments
through OTC transactions. The study aims to provide additional information on this challenge
of enrolling 100% of the beneficiaries to cash card accounts and the benefits that come with it.

In more recent events, the payment system of the program has been an essential tool in the
delivery of the financial assistance for the Social Amelioration Program (SAP) to the Pantawid
Pamilya beneficiaries as part of the government’s mitigation and recovery plan during the
COVID-19 pandemic. Using information on the experiences of the program beneficiaries in
relation with the SAP, the study gains insights on the current capacity of the program payment
system and briefly discusses its adaptability for other forms of social protection such as
emergency cash assistance.

1.2. Objectives

The study aims to assess the payment system of the Pantawid Pamilyang Pilipino Program in
terms its benefit level, frequency, and modality.

Specifically, the study aims to:

i. review international evidence on the level of benefits, frequency, and delivery
mechanisms of CCTs in other countries;
ii. assess the implementation of the payment system and inputs of key actors and identify
variations and challenges in implementation;
iii. review the appropriateness of cash grant amounts based on of the objectives of the
program, the intended value of benefits at program inception, the average expenses of
poor households and students, and compare current levels with other successful CCTs;
iv. identify and study other options of payment delivery and solicit feedback on these
options and alternative recommendations of program staff to improve the current
payment delivery system; and
v. determine the impact of the program on for subgroups of beneficiaries by type of
payment modality and frequency of payment.

3
1.3. Research questions

The research is a study of the overall payment system of the program covering portions of its
results chain from outputs to outcomes. The following research questions are covered:

i. What does the international evidence say about the level of benefits, frequency, and
delivery mechanisms of CCTs?
ii. How appropriate are the current benefit levels of the program in terms of the objectives
of the program, its initial intended value, the average expenses of poor households and
students, and in comparison, with other successful CCTs?
iii. How have policies, infrastructure, logistic, and technical inputs from the DSWD, LBP,
and other actors changed and affected the payment delivery system of the program in
the past years? What motivated the changes?
iv. How is the payment system implemented? Are there any variations or unique
challenges in the implementation?
v. What do program staff recommend to improve the current payment delivery system?
vi. What other options of payment delivery can the program explore to improve delivery
of cash grants in terms of timeliness and reliability under different circumstances? How
feasible are these options based on the perspective of the DSWD and other
stakeholders?
vii. How does the level of benefits, frequency and modality of grant payments affect the
following outcomes?
(a) Expenditure and income
(b) Labor outcomes
(c) Health, education outcomes

In addition to the above, the study also gathers insights on the capacity of the program payment
system using information on the experiences of the Pantawid Pamilya beneficiaries in receiving
their financial assistance for the Social Amelioration Program (SAP) of the government in line
with the COVID-19 mitigation plan.

In answering these research questions, the study is expected to generate recommendations and
options that will help improve the payment system of the program. Aside from generating
recommendations to enhance the payment system implementation, the research also delves into
the operationalization of these policy recommendations and determines whether these proposed
modifications are feasible based on the perspective of the program stakeholders. Ultimately,
the goal of the research is to provide improvements in the program implementation –
particularly on its payment system – to increase the chances of achieving its desired outcomes
in human capital formation and household welfare.

2. Review of Related Literature

2.1. Role of payment system in success of CCTs all over the world

Various studies and subsequent systematic reviews have shown that the design and
implementation features of conditional cash transfers affect the uptake and outcomes of these
programs on its beneficiaries (Bastagli, et al., 2016; Garcia & Saavedra, 2012; Davis, Handa,
Arranz, Stampini, & Winters, 2002; Fiszbein & Schady, 2009; Baird, McIntosh, & Ozler,
2009).In terms of the benefit mechanism of cash transfers, majority looked at the variations in
4
impact in terms of the amount of the transfers, while several looked at the impact of variations
in frequency of payment, timing, and predictability. This section presents a review of studies
on this topic.

In terms of poverty reduction and household welfare, studies generally find that higher transfer
levels result in bigger impacts on total household and food expenditure as well as poverty
reduction. Davis, [Link]. (2002) assessed the impact of the level of cash transfers by comparing
the impact of two programs in Mexico that provide considerably different levels of benefits.
The study reported positive impact of both programs on several indicators including household
welfare and health and education outcomes of children. But, more importantly, it found that
increases in the amount of grants have corresponding increases in the household expenditures
of beneficiaries. The same is observed for other social protection programs that are
unconditional in nature. In a two-year experiment of an unconditional cash transfer in Kenya,
Haushofer, and Shapiro (2016) found relatively higher expenditures for non-durable items
among households that received higher amount of transfers.

For child health outcomes, three studies of Mexico’s Progresa program observed that higher
transfer levels are associated with improvements in nutrition outcomes of children. Manley et.
al (2015) found statistically significant improvement in the height-for-age Z scores (HAZ) to
be associated with higher transfer levels. Esteva (2012) also observed higher HAZ and weigh-
for-age Z scores (WAZ) for beneficiaries of Progresa that received higher transfer levels, but
these were both not significant. Meanwhile, Fernald et. al. (2009) found that larger cumulative
transfers are associated with better child development in both physical and cognitive domains.

The results for education outcomes are more mixed. Manley et. al (2015) found a significant
positive relationship between the size of transfers in Progresa and improvements in the
cognitive and verbal development indices of children. The study also observed stronger effects
on cognitive development due to transfer levels compared with longer duration of exposure.
Baird et. al. (2009) found that higher transfers in a Malawi program are associated with
improved attendance and progression in school levels among schoolgirls, but this result is only
applicable when conditions are enforced with regard to the school attendance of the children.
However, other education outcomes examined in the study, including dropout rates, are not
responsive to increases in transfer levels. Filmer and Schady (2011) also found small and
often insignificant effects on education outcomes of higher transfer levels of the CSSP in
Cambodia. A meta-analysis done by Garcia and Saavedra (2012) found that transfer amounts
are not statistically correlated to effect sizes of program impacts on education outcomes.

For employment, although the effects of CCTs on employment and work-seeking behavior
have been widely studied, there are limited research that specifically looked into the
relationship between employment and the level of transfers. One of these few studies is done
by Dabalen et. al (2008) where the authors observed reduction in working hours due to increase
in transfer levels in a program in Albania. The authors explain that this result could have been
exacerbated by other environment conditions such as poor job generation, among others, but
they also mention that underreporting of employment is possible as it is part of the eligibility
criteria for the program. On the other hand, a re-analysis of six CCT programs in developing
countries reveal that there is no systematic evidence that the programs have discourage work
among beneficiaries (Banerjee, Hanna, Kreindler, & Olken, 2017)

Relative to studies on transfer levels, studies on the effect of the frequency, timing and
predictability of cash transfers are scarcer. In terms of household welfare, Haushofer, and

5
Shapiro (2016) experiment of a Kenyan cash transfer program showed that monthly provision
of cash transfers compared to lump-sum receipt is associated with reduction in expenditures of
non-durable items. In addition, lump-sum receipt of benefits was found to be associated with
higher savings and investment in beneficiary households. Bazzi et. al. (2012) also observed
that predictability and the corresponding timely delivery of benefits are associated with higher
increases household expenditure in Indonesia.

For education, the results of the meta-analysis done by Garcia and Saavedra (2012) indicate
that more frequent provisions of cash transfers, i.e., monthly versus less frequent, is associated
with negative impacts on effect sizes of program impact on education outcomes of children.
The authors explained that less frequent payments (i.e., less frequent than monthly) are able to
relax the saving constraints of beneficiaries due to partial delay or lumping of payments.

Regarding employment, Bazzi et. al. (2012) found reduction in the working hours of adult
members associated with the delay in the receipt of transfers. As mentioned earlier, the program
eligibility criteria may have influenced this observation as households anticipate receiving
transfers under low employment conditions. However, is also important to highlight the
reanalysis done in Banerjee et al (2017) that show there is no significant effect on work
incentives.
Banerjee et al. (2017) showed a graph (Figure 1) on the negative relationship between the
generosity of the cash transfer programs and the view of people on why the poor are poor.
Countries that tend to believe that the poor are lazy tend to provide less generous cash grants
than countries that tend to believe that poverty is because their society is unfair.

2.2. Local assessments of the 4Ps Payment system

Studies by CGAP (2013) and Catubig et. al. (2015) both assessed the payment system of
Pantawid Pamilya, although the latter had a limited scope and focused only the experience of
Davao Region. Both studies noted improvements in the payment system of the program since
its pilot implementation, especially after the entry of additional payment conduits that manage
the delivery of cash grants to beneficiaries. However, both studies also identified gaps in the
payment system.

The CGAP report mentioned as challenges the underdevelopment of the financial

infrastructures and the difficulty of catching up with the demand due to rapid increase in
number of beneficiaries through the years. This issue is further exacerbated by the limitation
of other conduits in developing solutions and innovations as the monopoly of the
implementation lies with the Land Bank of the Philippines. Another key issue pointed out by
the report is the “missed opportunity” for the LBP and the other payment conduits as they fail
to fully appreciate and maximize the potential of beneficiaries to be formal bank clients.

Meanwhile, Catubig et. al. noted that trade-offs exist for all types of payment delivery modes
in terms of opportunity costs for the beneficiaries, transaction costs for the government,
predictability, and security. This observation highlights the need to study all possible payment
delivery options in order to weigh trade-offs and arrive at the most cost-efficient mode. The
study also noted facilitating factors such as use of established financial infrastructures already
present in the areas (e.g. LBP, Rural Banks, Philpost, etc.), and competition between payment
service providers that encourage innovations. Lastly, the study highlighted that the efficiency
of the program payment system relies heavily on the commitment of these providers to fulfill
their obligations in the payment delivery.

6
3. Research Design and Methodology

This section presents the research design for the study. The study used a mix of qualitative and
quantitative methods of analysis using data from both primary and secondary sources. The
analysis is anchored on the conceptual framework presented in Section 3.1 which shows the
pathways through which the design features of the payment system of the program affect the
program’s theory of change, and the achievement of the desired outcomes among beneficiaries.
Section 3.2 presents the research design and data collection and analysis components that were
used, while Section 3.3 briefly describes the data sources in terms of respondent profiles.

3.1. Conceptual Framework

One of the key assumptions in the program theory of Pantawid Pamilya is that the cash grants
incentivize beneficiaries to perform their co-responsibilities meant to facilitate human capital
accumulation of children in the household (DSWD, 2014). Under such assumptions, the
amount of grants and the frequency and manner by which the grants reach the beneficiaries are
crucial. As presented in Section 2, evidence on other CCTs demonstrate that these design
features of the payment system have an influence on the achievement of desired outcomes of
the program.

Figure 1 shows a simplified illustration of the program logic. Poor households receive the
package of interventions which include cash grants, the imposition of program conditions, and
learning sessions through the FDS. With these program inputs, beneficiaries of the program
are expected to have increased income and higher productivity of children in the future.
Beneficiaries receive cash grants in exchange for their compliance with program conditions on
the education and health of children, and family development, which in turn result in higher
investments in the human capital of children. The grants are expected to increase the resources
of the household beneficiaries and allow them to afford their needs and smoothen their overall
consumption. The additional income is also expected to improve chances of households to cope
or withstand economic shocks. All these intermediate outcomes contribute to the overall goal
of the program to increase the productivity of children beneficiaries as adults and improve the
overall welfare of their households.

For these pathways and outcomes to be realized, it is assumed that resources for the program
are adequate and delivered on time and the enforcement of conditionalities and other support
interventions including the FDS are implemented effectively. Likewise, it is also important that
supply of quality educational and health services is available and accessible to the beneficiaries.
The final assumption key to the realization of the final program goal is that merit-based
employment and entrepreneurial opportunities are available to the beneficiaries. This allows
that the investments on the human capital of the children beneficiaries translate to higher
productivity in gainful employments and entrepreneurial activities.

From the figure, the three features of the benefit payment are expected to have influence the
realization of changes for the beneficiaries. Evidence from the literature (Section 3) generally
suggest that higher transfer levels result in higher impacts on the household expenditure and
other measures of immediate income effect of the grants. Likewise, higher transfer levels are
also expected to influence compliance behavior of beneficiaries, therefore increasing the

7
likelihood of achieving outcomes related to the human capital investment in children. The
frequency of the grants provision relates more to the availability of the grants to support the
ongoing consumption of the beneficiaries, while the modality of grant provision influences
access of the beneficiaries to the grants including the costs they incur.

It is important that the program can balance the effectiveness and cost-efficiency of the benefit
scheme. The grants should be large enough that the program is able to encourage desired
behaviors from the beneficiaries and effect improvements in their consumption, but at the same
time, it should also be not too large to foster dependency among beneficiaries. Likewise, the
grants should be provided in a timely manner such that they are available to the beneficiaries
in times of need, but not at the expense of cost efficiency (i.e. operational costs vis-à-vis
marginal benefit). Lastly, it is important that the grants are delivered at minimum costs for the
program implementers and the beneficiaries.

8
 Figure 1. Conceptual Framework

3.2. Research Design

The study used a mix of quantitative and qualitative data and analysis to answer the research
questions specified in Section. For simplicity, these components are grouped into four clusters
of analysis below.

3.2.1. Desk review of literature and administrative data

Desk review of available literature and evidence related to the payment policies of international
CCT programs was conducted to answer the first and second research questions posted in the
study. Literature on the design and assessment of successful international CCTs were used to
develop insights on potential improvements in the 4Ps payment system.

The study also reviewed program administrative data and reports related to the payment system
and program operations. These include data and reports on the cash grant disbursement,
payment delivery modes, compliance monitoring, and budget of the program.

3.2.2. Key Informant Interviews and Focus Group Discussions

The second group of analyses include qualitative research methods through focus group
discussions (FGD) with program beneficiaries and key informant interviews (KII) with

9
program implementers. These were conducted to gain insights from the experiences of the
stakeholders with regard to the payment system of the program.

The key informant interviews of stakeholders of the program were aimed to gain an
understanding of the following aspects of the payment system:
i. Payment policies (i.e., on benefit levels, frequency, and modality) and rationale
ii. General processes and implementation
iii. Variations in the implementation across the years and areas
iv. Challenges, and facilitating factors
v. Plans and policy directions (e.g. modernization)
vi. Suggestions/recommendation for improvement

The target respondents were key persons from the program management responsible for
program policies, operations, and monitoring, as well as key officials of the Land Bank of the
Philippines (LBP) as the primary AGDB.

On the other hand, the focus group discussions with beneficiaries were aimed to understand
issues that they are facing in accessing and using cash transfers. The key topics in the FGD
include:
i. Assessment of the appropriateness of benefit levels relative to their household
expenditures
ii. Experience in compliance monitoring and correctness of cash grant amounts
iii. Knowledge and experience on filing payment-related grievances
iv. Experience in payout procedure such as receipt of information on payment
schedules, challenges in accessing grants, support services related to the program
payouts, etc.
v. Expenditure of grants and tracking of expenses
vi. Suggestions for improvement of program implementation

Data collection was conducted in National Capital Region (NCR) and the three island clusters
of Luzon, Visayas, and Mindanao. The study areas were selected from the 3rd wave impact
evaluation sites and consisted of 16 barangays from 8 municipalities4. One FGD per barangay
was conducted totaling to 16 FGDs in all. For the areas outside NCR, one city and one
municipality is chosen for each region, while one rural and one urban (or “Poblacion”)
barangay were chosen for each city or municipality. The distribution of study sites per island
cluster is shown in Table 1.

4
The areas and base sample frame for this study are the same with separate research entitled “Qualitative Follow-up Study to
the 3rd Wave Impact Evaluation”. The data collection for the two studies were simultaneously conducted but the FGD
participants were different.

10
Table 1. Distribution of study sites by island cluster

Island Group/ Number of Cities/

Number of barangays
Cluster Municipalities
NCR 2 cities 4
Luzon 1 city, 1 municipality 4
Visayas 1 city, 1 municipality 4
Mindanao 1 city, 1 municipality 4

For each of the 16 barangays, the data collection included a focus group discussion with
beneficiaries of the program, and interviews with the Pantawid Pamilya City/Municipal Link
(C/ML) and Municipal Roving Bookkeeper (MRB) assigned in the areas whenever possible.
The C/ML is the frontline staff of the program who oversees that the operations at the
beneficiary level and serves as the primary linkage of the program with the beneficiaries and
other local stakeholders. The MRB on the other hand is the staff primarily involved in the
payment operations at the city or municipal level. The MRB serves as the primary link between
the C/ML, the LBP or payment conduits, and higher level staff involved in the payment
operations. The number of C/MLs and MRBs assigned in the area usually depend on the
number of Pantawid Pamilya beneficiaries.

Data collection was conducted for majority of the Luzon areas (including NCR) from
November to December 2019 while data collection in Visayas and Mindanao areas were
conducted from January to March 2020.

Aside from the MRBs and C/MLs, other stakeholders interviewed are key officials involved in
the policy and design of the payment system. These interviews were scheduled after all of the
FGDs in order to validate and probe issues that were identified by the beneficiaries as well as
the frontline workers of the program. The complete list of target respondents for the FGD and
KIIs are presented in Table 2 below. The roles and number of target respondents are also
shown. All of the FGDs and most of the KIIs were conducted in person while KIIs with two
MRBs, the LBP, and the DSWD-FMS were conducted via phone and online video conference
due to COVID-19 travel restrictions. Due to scheduling difficulties, the interview with
Pantawid National Program Management office was conducted via email where the office
provided responses to guide questions.

Table 2. Target respondents for FGDs and KIIs

Respondent Number of Method Role in the program

respondents
Beneficiaries 10/ barangay FGD ▪ Program beneficiary
City/Municipal Link 1/city or KII ▪ Compliance monitoring
municipality ▪ Linkage with local actors
▪ Information channel to beneficiaries
▪ Facilitation of updates and grievances
Municipal Roving 1/city or KII ▪ DSWD staff assisting conduits for
Bookkeeper municipality payouts in OTC areas
Pantawid Pamilya 1-4* KII ▪ Program M&E, and key policies
National Program ▪ Grievance redress
Management Office ▪ Beneficiary updating, NPMO level
(NPMO) ▪ Compliance verification, NPMO level

11
 Respondent Number of Method Role in the program
respondents
DSWD Finance Director 1 KII ▪ In charge of payroll generation and
for 4Ps liquidation at the national level
Landbank Program PMO 1 KII ▪ Authorized government depository
bank
▪ Overall in-charge of payment of grants

Thematic analysis was used to analyze the qualitative data collected by the study team. This
was facilitated by the use of computer-assisted qualitative data analysis software (CAQDAS),
specifically NVivo. Transcripts of the interviews were systematically coded following themes
based on pre-identified areas of discussion during the questionnaire development. Sub-
categories were then generated based on the initial themes in order to further analyze the range
of insights provided by respondents.

Given validity and reliability issues inherent in qualitative analysis (Nowell et al. 2017), the
study team employed various methods in order to validate the study results. Namely, these are
triangulation, thorough documentation of interview proceedings and the analytical process, and
use of rich and thick description in the analysis text.

Triangulation is conducted by comparing responses of program beneficiaries and program

implementers, namely DSWD and Landbank staff. Cross verification is also done by
comparing qualitative analysis results with results of other methods employed by the study.
Interviews were documented with audio recordings, as well as notes taken during the interview.
Documentation of the analytical process was also practiced by the writing of notes or “memos”
using features provided by the CAQDAS utilized by the study team. Thick description is
provided in the analysis in order to provide a thorough context for interview responses.

Although measures are taken to support the validity and reliability, particularly the qualitative
component of the study, it must be noted that qualitative analysis on its own cannot be used to
generalize over a large population. The main purpose of the analysis in this case is to facilitate
triangulation by cross verification with the results of the other methods of analysis employed
by the study. This also supplements and provides depth the analysis by capturing insights not
covered by other methods.

12
3.2.3. Analysis of IE3 survey data

The study also included an analysis of the survey data collected for the 3rd impact evaluation
of the program to assess if target outcomes are affected by features of the payment system,
specifically the mode of delivery of receipt of cash grants.

The survey data was collected from November to February 2017 covering 6,775 households
that are beneficiaries and non-beneficiaries of the program and includes information on the
manner through which beneficiaries of the program receive their cash grants. Moreover, the
data includes information on the characteristics of the households and their individual
members, and indicators of desired program outcomes such health care utilization and
education of children, income and expenditure, receipt of social and government program
benefits, labor participation, anthropometric data, among others.

Consistent with the design of the 3rd impact evaluation of the program, the study applied an
RDD identification strategy was used to analyze the survey data. In RDD, poor and eligible
households – the beneficiaries – are compared with households that are just below the
eligibility cut-off for the program (i.e., poverty threshold). RDD is extensively discussed in
the works of Imbens and Lemiuex (2008), Cattaneo, Idrobo and Titiunik (2018a) and Cattaneo,
Idrobo and Titiunik (2018b).

Figure [Link] Discontinuity Design

TREATED CONTROL TREATED CONTROL

-h Cut-off + -h Cut-off +

Source: Impact evaluation concept note, World Bank (2010)

The estimation is constructed on the base equation

𝑦 − − 𝑦 + = lim 𝐸[𝑌𝑖 |𝑋𝑖 = 𝑥] − lim 𝐸[𝑌𝑖 |𝑋𝑖 = 𝑥]

𝑥↑𝑥̅ 𝑥↓𝑥̅

where Y is outcome of interest, X is the Proxy Means Test (PMT) score of the households
based on Listahanan targeting of the program, x is the eligibility cut-off or poverty threshold,
and T is the treatment assignment variable.

The study estimated the impact of the program on subsets of beneficiaries grouped by cash
grant modality (OTC versus cash card). This was done by estimating the impact of the program
separately on subsets of the sample based on the features of the payment system being
13
examined. Differences between impact estimates for the subgroups were tested using z-test of
equality of coefficients:
𝜏1 − 𝜏2
𝑍=
√(𝑠𝑒1 )2 + √(𝑠𝑒2 )2

where: 𝜏1 = coefficient (program impact) on the first subgroup

𝜏2 = coefficient (program impact) on the second subgroup
𝑠𝑒1 = standard error of the impact estimate on the first subgroup
𝑠𝑒2 = standard error of the impact estimate on the second subgroup

To improve precision of estimates, relevant covariates such as urban/rural classification of

areas, barangay characteristics, supply conditions, etc. were included in the estimation models.

The analysis of the IE3 data for this study is limited by two main issues: (1) the original
sampling of was not designed to measure the impact of such features of the payment system;
and (2) the RD methodology can only generate impact estimates among household
beneficiaries near the poverty thresholds and not of the poorer segments of the beneficiary
population. Ideally, a controlled experiment as in an RCT should be done to accurately measure
the impact of these payment system components; however, the results of the study should be
able to provide a reliable picture of the impacts of this distinction in the payment modality.

3.2.4. Online survey on SAP implementation

In line with the government’s social protection plan to alleviate loss of income and other shocks
due to the COVID 19 pandemic, financial assistance is given to Pantawid Pamilya beneficiaries
through the Social Amelioration Program (SAP). An online survey was launched to collect
information on the experiences of the program beneficiaries in receiving their financial
assistance. The information from the online survey was used to gain additional insight on the
current capacity of the program payment system, challenges, and its ability to adapt to wider
coverage of beneficiaries and available technologies in payment delivery.

The questions included in the survey are those that pertain to their experiences in accessing or
withdrawing their regular cash grants in Pantawid Pamilya and the SAP financial assistance.
Respondents were also asked regarding the platforms they used in accessing their benefits (e.g.,
ATMs, digital banking, etc.), as well as effects of the current pandemic on their sources of
income, expenses, and access to health services.

14
3.3. Profile of Respondents by Source

This section presents the profile of respondents in the data collection activities mentioned in
the preceding text. These include the description of participants in the FGDs and KIIs, as well
as respondents in the 3rd impact evaluation survey and the online survey on SAP experience.

3.3.1. FGD Respondents

A total of 156 individuals participated in the 16 FGDs that were conducted. Although the
research team targeted ten participants for each FGD, the number of participants varied across
sites with number of participants ranging from 5 to 18 due to availability of participants.
Among the island clusters, Mindanao FGDs had more participants than the average (Table 3).

Table 3. Distribution of FGD participants by island cluster

Island Cluster Freq. Percent

LUZON 36 23.08
MINDANAO 46 29.49
NCR 38 24.36
VISAYAS 36 23.08
Total 156 100

The base sample for the FGDs were taken from the sample of Pantawid Pamilya beneficiaries
in the IE3 survey5. In cases where the number of potential participants for the FGDs is lower
than 10, the city/municipal link were asked to invite other active beneficiaries of the program
in the barangay.
Table 4 shows the profile of the FGD participants. Majority of the respondents (83%) were
female, and most are the designated grantee of the household. The grantee, usually the primary
caretaker of the children, is the authorized representative to receive the cash grants and attend
the FDS. The rest of the participants were mostly the spouse of the grantee. Among the
participants, almost 20% were Parent Leaders (PL). A parent leader is a Pantawid Pamilya
beneficiary chosen as a leader or representative of a group of beneficiaries of around 25 to 40
members usually residing within the same barangay. The PLs are usually more well-versed
with the program operation details compared to other members of their parent group because
the C/MLs are usually relay information to the other beneficiaries through them. The average
age of participants is 47.

5
This excludes beneficiaries who were invited in FGDs for the research “Qualitative Follow-up Study to the 3rd
Wave Impact Evaluation”. The areas and base sample for these studies are the same but participants for the
FGDs were different to avoid respondent fatigue.

15
Table 4. Profile of FGD participants

Characteristic Freq. (n=151*) Percent

Female 125 82.78
Parent leader 29 19.21
Grantee 122 80.79
Age group
20-29 5 3.31
30-39 35 23.18
40-49 51 33.77
50-59 38 25.17
60 and older 22 14.57
Note: Five participants do not have profile information

Table 5 shows the average household composition of the FGD participants. From the table, not
all of the households have children 5 years old and below and pregnant women. Meanwhile,
on the average, there are two children age 3 to 18 years old in the households. These children
are within the age group eligible for the education monitoring of the program. However, the
average number of monitored children for education is lower which means not all of those
within the eligible age group are monitored for school attendance. For context, the program has
a limit of three children per household
Table 5. Household composition of FGD participants

Household member Average number Min. Max.

in the household
(n=151)
Pregnant woman 0.05 0 1
Children 0 to 5 years old 0.46 0 3
Children 3 to 18 years old 2.21 0 10
Monitored children
(for education condition) 1.49 0 3

In terms of mode of payment of cash grants, almost all of the respondents reported claiming
their cash grants through cash cards (96%) while only 5 participants reported receiving through
over the counter transactions in Land Bank branches. This is consistent with the KII responses
and recent program reports that indicate that most of the areas nationwide, except for areas in
Bangsamoro Autonomous Region in Muslim Mindanao (BARMM) are predominantly
receiving grants through cash cards. Among the participants, 145 were able to provide the
amount of the last grant they received. The average amount is PHP 3.400 pesos for two months
of compliance monitoring, while reported amounts range from PHP 1500 to PHP5200. Most
of the participants received their last cash grant one or two months prior to the interview but
five of the participants reported receiving their last grants in 2009 up to 2018. These
participants were later found to be inactive beneficiaries who have exited from the program
due to aging out of eligible children or other reasons for ineligibility.

3.3.2. KII respondents

16
As discussed earlier, City or Municipal Links (C/ML) are the program frontline staff who serve
as the direct link of the program to the Pantawid Pamilya beneficiaries. The C/MLs ensure that
systems and program operations are running smoothly at ground level and concerns of
beneficiaries regarding the program are addressed or referred to the proper stakeholders. Their
role include ensuring that information on the beneficiaries are up to date, compliance data are
being collected (with the help of health facilities, schools, and another program staff called
Social Welfare Assistant), grievances of beneficiaries are collected and properly endorsed to
appropriate channels, and FDS are being conducted and attended regularly. They are also
expected to be “case managers” of the beneficiaries and ensure that that other issues and needs
of the family are being addressed through linkages with local stakeholders including the Local
Government Units. Given the crucial role of the C/MLs in program implementation, they were
included as key informants in the study.
In total, there were 18 C/MLs that were interviewed. There were 5 each from NCR and Visayas,
and 4 each from Mindanao and Luzon. Table 6 shows the profile of the respondents in terms
of their experience as C/MLs and current caseload. The average years of experience as C/ML
is 6 years with the shortest at 1 year and longest at 9 years. In terms of experience as C/ML in
the study area, the average and range is lower since some of the C/MLs interviewed have been
previously assigned to a different area. The average number of barangays covered by one C/ML
is around 6, while number of beneficiaries on the average are around 780 and values range
from 678 to 980.
The Municipal Roving bookkeeper (MRB) is the staff in charge with the ensuring payment
operations in the program are running smoothly at the ground level. Their primary role is to
coordinate with the Land Bank of the Philippines and other payment conduits to ensure that
beneficiaries are able to receive their cash grants whether through cash cards or over the
counter. They also work with the C/MLs in helping beneficiaries enroll to a cash card account
and troubleshoot concerns regarding receipt of grants such as loss or damage of cash cards,
card capture, among others.
In total, there were nine MRBs assigned in the study areas that were interviewed. On the
average, the MRBs have had 5 years of experience as MRBs (Table 6). An MRB is usually
assigned to one or more municipality but in some areas, particularly in populated cities like in
NCR, more than one MRB can be assigned.

17
Table 6. Experience and caseload profile of C/ML and MRB respondents in KIIs

Respondent characteristics Mean Obs. Std. Dev. Min. Max.

CITY/ MUNICIPAL LINKS
Total years as C/ML 6.4 16 2.4 1 9
Total years as C/ML assigned in the study area 3.9 16 2.0 0.7 7
Caseload or number of Pantawid Pamilya 780.8 16 84.0 678 980
beneficiaries managed
MUNICIPAL ROVING BOOKKEEPER
Total years as MRB 5.1 9 2.4 2 10
Number of MRBs assigned in the same city or
1.3 9 0.7 1 3
municipality

In addition to the local KIIs with the C/MLs and MRBs assigned in the study sites, interviews
were conducted with officers of the Landbank of the Philippines’ Conditional Cash Transfer
Program Management Department (CCT-PMD), the DSWD Financial Management Service
for Special Programs (FMS-SP), and the National Program Management Office (NPMO) of
the program. These interviews were done to collect information on the payment process,
including rationale for past policy decisions, current, and future directions, from the perspective
of the national-level program implementers There were two KII respondents for the interview
with CCT-PMD and also two respondents for the interview with DSWD FMS-SP which was
done via video conferencing. The data collection for the NPMO was conducted through email
questionnaire through its Planning and Monitoring Evaluation Division.

1.1.1. IE3 data

As discussed in the previous section, the main sample of the 3rd wave impact evaluation survey
was based on a RDD evaluation design. The sample consisted of poor households eligible to
be enrolled in the program, and non-poor or non-eligible households. The total sample covered
in the survey were 6,775, coming from 180 barangays nationwide. The survey data was
collected from November 2017 to February 2018 through face-to-face interviews.

Of the total sample, 3,450 households formed part of the treatment group while 3,325
households were included in the comparison group. Table 7 shows the characteristics of the
treatment and comparison group at baseline and their estimated PMT scores at the time of
targeting. Based on the table, the households in the treatment and comparison groups, on the
average, are generally similar in terms of household composition and income. This supports
the validity of comparing the two groups to measure the impact of the program.

18
Table 7. Household composition and estimated income of households at baseline, by
treatment assignment

Treatment Comparison All

Total number of households 3,450 3,325 6,775
Average number of household members 5.17 5.10 5.13
Average number of HH members by age group:
0 to 5 years old 0.6 0.59 0.6
6 to 14 years old 1.28 1.21 1.25
15 to 18 years old 0.5 0.47 0.48
19 to 60 years old 2.48 2.49 2.48
Total no. of WRA (aged 15-49 years) 2,646 2,494 5,140
Ave. estimated income based on PMT PHP 14,466 PHP 15,596 PHP 15,017

The focus of this study component is the differential impact of the two modes of cash grant
payment on the key target outcomes of the program. Of the treatment households, data on the
mode of payment is only available to 3,011 Pantawid Pamilya beneficiaries. The distribution
of the households by mode of payment is shown in Table 8. Treatment households that did not
have information on the mode of payment were excluded in the analysis. In total, the analytical
sample for the treatment group is composed 1,315 household that receive grants through OTC
mode and 1,298 households that receive grants through cash card mode of payment.

Table 8. Distribution of IE3 treatment households by mode of payment

Mode of Payment Frequency Percentage

Over-the-Counter (OTC) 1,315 38.1
Cash card 1,298 37.6
No data 837 24.3
Total 3,450

Table 9 presents some information on the payment history of the 4Ps households in the
analytical sample. Based on the respondents reported frequency of payouts in the past 12
months, cash card payouts were more frequent compared to OTC payouts. This could be
expected because OTC payouts require more logistical preparations and schedules are
dependent on a number of factors that do not affect cash-card based delivery of grants. In terms
of the total amount of grants over the same period of 12 months, the difference between the
two mode of payments was not statistically significant although that of the cash-card group
was slightly higher possibly due to the slightly higher frequency of payouts.

19
Table 9. Frequency of payout and amount of grants received by IE3 treatment households
included in the analysis (n=2,613)

By Mode of Payment
All
OTC Cash Card Difference
Frequency of cash grant payouts in the past 5.3 5.1 5.5 -0.40**
12 months
Total amount of grants received in the past 15,357 15,060 15,658 -597.37
12 months (in PHP)

1.1.2. Online SAP survey

The online survey was launched starting April 2020 using the online platform Survey Monkey.
The data used in the analysis is as of December 2020, but majority of the responses included
in the analysis were those received towards the end of April 2020 until the first week of May
2020 as shown in Figure 3. During this period, the first tranche of cash grants was still currently
being distributed to its intended beneficiaries including 4Ps households.

Figure 3. Frequency distribution of online survey respondents by date of response

The main challenge in the data collection is the recruitment of respondents and completion of
the survey. Being an online survey, the response and completion rates depend on the access of
beneficiaries to internet-connected gadgets and requires a certain level of internet-literacy to
navigate through the survey. This was especially difficult because the target respondents – the
4Ps beneficiaries – are from low income households, and we expect that access to stable
internet connection and internet-capable devices is limited. Another issue observed in the data
is the imbalance in distribution of respondents in favor of urban areas where connectivity is
more available.

In total, there were 1,483 attempts to answer the survey and 1,325 of which (95.8%) were
eligible to answer the interview as they were current 4Ps beneficiaries at the time of data
collection. However, the number of responses dwindled as the questionnaire progressed due to

20
the respondents dropping out in the middle of the interview. Possible reasons could be unstable
internet connection, loss of interest, respondent fatigue, or errors in navigating through the
questionnaire. Of the 1,325 original respondents, only 488 beneficiaries reached the last section
of the questionnaire, equivalent to only 36% completion rate (Figure 4). Average completion
time for the whole questionnaire is 20 minutes.

Figure 4. Completion rate by question item in online survey

Completion rate (n=1325)
100.0%
98.2%
96.5%
90.8%
87.8%
83.8%
83.2%
66.9%
48.7%
46.2%
44.5%
41.9%
41.7%
41.5%
39.1%
38.8%
38.5%
38.3%
38.1%
36.8%
36.3%
Q18
Q10
Q11
Q12
Q13
Q14
Q15
Q16
Q17

Q19
Q20
Q21
Q8
Q1
Q2
Q3
Q4
Q5
Q6
Q7

Q9

Question item

To maximize the data collected, the responses of a total of 886 respondents were included in
the analysis. The data includes those of the respondents that were able to answer the questions
regarding their usual experience in receiving the 4Ps cash grants as well as the question asking
whether they have received their SAP grants. These are indicated by the completion rates for
Question 7 and 8 (Q7, Q8) in Figure 4. However, since the completion rate for the succeeding
questions decreases, these respondents have incomplete data for the latter sections which
include follow-up questions on the SAP and their profile characteristics. Because of this, profile
information are incomplete for some respondents.

Table 10 shows basic descriptive statistics on the respondents included in the analysis. The
number of observations with data on these characteristics are also presented for reference. Most
respondents are the assigned 4Ps “grantee” for their household. The grantee is the authorized
representative of the household to receive cash grant payments and is also the member required
to attend the Family Development Sessions. Almost all of the respondents are females which
is expected since most of 4Ps grantees are also the mothers of the children beneficiaries. Most
of the respondents reported that their spouse is the primary income earner or “breadwinner” of
the household. The top occupations reported were construction and factory work and driving
public utility transportation. The average age of respondents is 42 while the average household
size is 7. The average number of children 3 to 18 years old in the household is 3 and most, but
not all, of these children were enrolled in school before the quarantine.

21
Table 10. Descriptive statistics of respondents included in the analysis

Number of
Percentage of
Characteristic observations
respondents*
with data
Percentage of respondents who is the current grantee of the
98.0% 886
household
Percentage of respondents who is a 4Ps parent Leader 28.0% 612
Percentage of female 96.0% 552
Breadwinner of the household is the respondent’s spouse 60.2% 505
Mean*
Age 41.63 555
Household size 6.75 510
Number of children 3 to 18 years old 2.99 510
Number of children 3 to 18 years old currently attending school 2.72 507
Note: The percentage and mean values were computed among observations with data

More than half of the respondents are from the National Capital R. This is followed by
Region IV-A and Region III. The table below shows the distribution of the respondents in
NCR. Most of the NCR respondents are residents of Marikina City (66.0%) followed by
Makati City (16.7%).

Figure 5. Distribution of online survey respondents by region

Region VII Other regions

2% 2%
Region II
4%

Region IV-A
17%

NCR
Region III 60%
15%

Note: Other regions include Region I (4 respondents), Region V (3), Region VI (1),
Region VIII (1),Region XII (1), and BARMM (1); N = 550 respondents

When respondents were asked if a type of quarantine is being imposed in their area of
residence, majority responded that an Enhanced Community Quarantine (ECQ). This is
expected since most of the respondents are from NCR and other nearby regions and most of
the responses were collected in April and May 2020.

22
4. Results

This section presents the results of the study. The first subsection provides a detailed
description of the implementation of the payment system of the program as well as challenges
in the implementation that were collected from the FGDs with beneficiaries and KII with
implementers. The succeeding subsections focus on the three focus areas of the study, namely
assessment of the benefit level, the frequency of payment, and the modality of payment
delivery. The final sub-section discusses the analysis of the online survey on the SAP
implementation focusing on the experience of the beneficiaries in accessing their grants and
the role of the program’s payment system in its delivery.

4.1. The 4Ps Program Payment System

According to the key informants from the national program management, the payment system
of the program plays an especially important role in the success of the program as it ensures
the correct payment to the beneficiaries and timely release of grants which are expected to
encourage desired behaviors from the beneficiaries.

The payment system forms a key component of the program cycle starting from the first release
of unconditional grants to newly registered households and the succeeding release of grants
that are based on the compliance of beneficiaries (Figure 6).

Figure 6. Role of payment system in Pantawid Pamilya Program Cycle

Selection of Registration and 1st release of

Household Validation of grants
Beneficiaries Household (unconditional)

Verification of
Grievance redress systems

Pantawid
Beneficiary data updates

Supply-side compliance with

Assessment conditions Pamilya
Payment
Compliant households System
Selection of
Provinces/ Succeeding
Municipalities release of
grants

Source: Adapted from Figure 1 of Pantawid Pamilya Operations Manual (DSWD, 2015)
4.1.1. Current design of the payment system

Figure 7 shows the fund flow of the program payment system from the government allocation
in the General Appropriations Act (GAA) up to the payout to beneficiaries. The fund begins
with the approval of the GAA which provides authorization for the program to incur fund
obligations, followed by the release of Notices of Cash Allocation (NCAs) by the DBM to
DSWD for the payment of the expenses of the program. The cash grants are deposited by the
DSWD in the Land Bank of the Philippines as the program’s primary authorized government
depository bank (AGDB). The LBP then maintains the funds in takes care of the transfer of the

23
funds to its conduits that will pay the cash grants to the beneficiaries or directly to the accounts
of the beneficiaries with cash cards.

Figure 7. Fund flow of the Pantawid Pamilya Payment System

DBM DSWD LBP

General Disburse funds Open and

Appropriations under the maintain current
Act (GAA) Modified account under
Disbursement MDS for
System (MDS) Pantawid
Notice of Cash account Pamilya
Allowance maintained in Program fund
LBP

DSWD MDS
Account

LBP-Batasan DSWD Peso Current

Account Account maintained at
LBP Batasan for LBP
conduits
Credit cash grant to
individual cash cards of
Transfer funds to bank
beneficiaries
accounts of LBP conduits

Payout to Pantawid
Pamilya beneficiaries via
OTC

Source: Adapted from Pantawid Pamilya Operations Manual – FMS Sub-manual (DSWD, 2015)

Cash grants are delivered to beneficiaries through two modes of payment: Bank cash cards
where the grants are withdrawn via ATM; and Over-the-counter (OTC) transactions where the
grants are provided directly to beneficiaries and in the form of cash via payment conduits.

Payment conduits are hired by the Landbank to deliver the cash grants when there is no
available ATM or Point-of-Sale merchants in the area. The LBP procures the services of these
payment conduits following the existing government procurement law and selection is
generally done through competitive bidding of firms. Payment conduits include rural banks,
money remittance centers, cooperatives, money couriers, and postal services. The following
are the procedures highlighted in the program operations manual on the selection of conduits
that are eligible to provide their services in the payment delivery.

24
Figure 8. Selection process of payment conduits

Public or private conduit submits proposal to LBP

LBP invites potential conduits to present its plans and strategies in distribution and
liquidation of cash grants

DSWD shall ensure that the Information Systems of the potential conduits can readily
communicate with the information systems of the Pantawid Pamilya

LBP/AGDB shall negotiate for a reasonable fee to the potential conduits in consideration
of its services

DSWD issues a clearance that the potential conduit may now join the bidding process
undertaken by LBP

Source: Pantawid Pamilya Operations Manual – FMS Sub-manual (DSWD, 2015)

Payment conduits that have been involved in the delivery of cash grants include the PhilPost,
Globe GCASH Remit (merchant-based), Smart-NATCCO, MASSPEC, First Consolidated
Bank, Rural Bank of Gattaran, Philippine Veterans Bank, M Lhuillier, LBC, among others.

After the selection of the payment conduits, the payment system starts with the generation of
the Notice of Payroll Action (NAPA) based on the compliance data of beneficiaries. The NAPA
is prepared by the Compliance Verification Division program’s national program management
office and serves as the basis for the list of beneficiaries entitled for the cash grants in a specific
pay period. Using the NAPA and the report of the DSWD regional offices on the specific mode
of payment for the areas (in coordination with the LBP), the cash grants are computed. The
DSWD FMS facilitates the funding process for the payment of the cash grants through the
Landbank. The Landbank then facilitates the disbursement of the cash grants to the
beneficiaries – either through direct crediting to the cash cards of beneficiaries or the transfer
of funds to payment conduits. The conduits and LBP is likewise expected to facilitate the
liquidation of funds and submission of reports to DSWD.

To ensure quality assurance, the program management established some safeguards to ensure
correctness of the amount of grants and the intended beneficiaries. The following are the
measures mentioned by the DSWD key informants:
• The Pantawid Pamilya Information System (PPIS) which is maintained by the Planning
Monitoring and Evaluation Division (PMED) supports the core processes of the
program which are registration, updating, compliance verification, payment, and
grievance. It weeds out the need for manual processes which lead to human-errors.
• The program also conducts regular technical assistance and spot check activities (pre-
COVID).
• The Compliance Verification Division NPMO level conducts random checks per
household in all regions before and after NAPA generation.
• The Risk Management and Quality Assurance Division (RMQAD) of the NPMO helps
ensure that cash grants are released to correct households and in correct amounts and

25
 checks the payroll of beneficiaries in each pay period to check for any duplicates and
inconsistencies.
• At the regional and provincial level, the Compliance Verification Officers conducts
checking of the collected compliance forms from the facilities.

According to the NPMO, monitoring of the performance of the LBP as its ADGB is
continuously done. They mentioned having regular dialogues with LBP and other banking units
to address payment related issues and revisiting of the existing Memorandum of Agreement
with the LBP to respond to the encountered problems in the implementation.

4.1.2. Changes in the payment system

According to DSWD key informants, a combination of over-the-counter (Manual) and cash

card modes of payment were adopted during the early years of the program implementation.
The payment delivery was initially solely done by the Landbank of the Philippines. However,
during the early years of the program, due to the rapid expansion of the program, the LBP had
difficulty in catering to the program coverage. This prompted DSWD to request LBP to
contract out payment conduits that can do the payment delivery in their behalf especially in
areas where access to ATMs or banks are difficult.

According to the DSWD key informants, although the payment system had weaknesses in the
earlier years such as delays in payouts, it has improved through the years, especially starting
2017 when the DSWD management put pressure and attention to improving the payment
system. This is also supported by the feedback of beneficiaries during the FGDs. They
mentioned that delays in the payout were more frequent during the earlier years of the program
and they almost never experience significant delays in the payout of cash grants in the recent
years.

Shift towards full Cash Card mode of payment was also introduced in 2019 to facilitate faster
release of cash grants and ease beneficiaries’ access to it. The national program management
also mentioned that the program targets conversion of cash card account into transactional
accounts which allow savings and other banking transactions for beneficiaries.

According to the national program management respondents, the payment system also has
improved its accuracy due to various quality assurance procedures and safeguards that were
put in place. The full shift to EMV cash cards introduced faster obligation and crediting of cash
grants to the beneficiaries. However, the need for more infrastructure for access and wider
functionality of the accounts are lagging behind.

4.2. Assessing the Benefit Level

4.2.1. Changes in benefit level through the years

Since its implementation in 2008, the amount of grants amount of grants remained at same
levels until recently. Figure 9 below shows the timeline of the changes in the value of cash
grants from 2008 to 2020. From the figure we can observe that the cash grants remained the
same for most of the years since its implementation except for the increase in the high school

26
grants in 2014, the addition of the rice subsidy in 2017, and the very recent increase in grants
mandated by the institutionalization of the program through RA 11031.

Figure 9. Timeline of changes in value of cash grants

Source: DSWD, various years

Figure 10 shows the reduction in the real value of the cash grants from its original share of the
poverty threshold in the start of implementation in 2008 until 2017, excluding the addition of
the rice subsidy. From covering 20% of the poverty threshold during targeting, the real value
has been reduced by 5-percentage points after ten years of implementation. Note that this
simulation assumes a maximum cash grant amount for a household with three children in
elementary school who are all are fully compliant with conditionalities. Based on the program
administrative reports, beneficiaries are not always able to maximize the three-children limit
in the number of beneficiaries and compliance is not always perfect, so the shares may be even
lower in reality.

27
Figure 10. Share of the real value of grants over 2006 poverty threshold, by year

20%
19%
Share of grant over the 2006

19%
18% 17% 17% 16% 16%
Poverty Threshold

16% 15%

2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Source of basic data: DSWD, PSA

In a comparison made by Acosta &Velarde (2019) on the generosity (defined as of

international CCTs in bottom 20% or poorest households, the Pantawid Pamilya lags behind
other countries. Due to reduction in real value of the grants, the generosity of the program for
the bottom 20% has decreased from 11% in 2013 to only 9% in 2015. The slight rise in the
generosity observed in 2017 was explained by the authors as due to the addition of the rice
subsidy.
Figure 11. Comparison of 4Ps and international CCTs in terms of generosity in bottom 20%

27%

23%
22%
20%

12%
11% 11% 11%
10%
9%

Ecuador Mexico (2014) Argentina Brazil (2015) Peru (2014) Philippines Colombia Dominican Philippines Philippines
(2016) (2013) (2013) (2014) Republic (2017) (2015)
(2014)

Source: Acosta &Velarde (2019)

28
4.2.2. Adequacy of grant amounts

Beneficiaries and program implementers were asked to provide their assessment on the benefit
level or grant amounts paid by the program. Note that during the data collection, the
beneficiaries have yet to receive the increased rates for the cash grants under Republic Act
11310. The first payout with the new grant amounts started only in March 2020 after the
conduct of field activities.

[Link]. Assessment of beneficiaries

Even though the respondents reported that their household budget is not enough to cover
their expenses, most are reluctant to speak out on the optimal cash grant amount that
should be provided to them. They noted that an increase in the cash grant amount would
be able to help them cover cost of rising prices. They were not able to suggest a specific
amount for the increase in grants, but they mentioned that the grant for education and
rice could be increased – usually the grant for rice or education. Other beneficiaries also
said that there is no need to increase the grant amount but suggested that the program
cover up to college.

Some beneficiaries also said that they are already satisfied with current grant amount.
Although they acknowledged that the grant is not enough to cover their expenses, the
amount that they receive is already able to contribute to their household expenses. Most
beneficiaries said that they are satisfied with whatever amount the government chooses
to give them, also noting that the funding should go to programs for others who are also
in need. Many also mentioned that beneficiaries also need to work to earn income and
should not rely solely on the grant for their needs. A few of the beneficiaries, especially
parent leaders, cited that the increase of benefits under RA 11310 had already been
discussed with them and they are already satisfied with the amounts under the law.
Besides request for increase, grantees also suggest provision of other programs,
particularly livelihood programs, to support their income.

When asked how they spend the cash grants, beneficiaries report spend the cash grants
mostly for education of children, food for the household, clothing, and health expenses
including vitamins for children and check-ups (Figure 12). Some beneficiaries also
mentioned spending the grants for other household needs such as utilities and household
improvement and durables (e.g., tiles, kitchenware).

Grantees often report that they are not able to set aside a portion of the grant money, and
this is often spent by the time of the next payout. Although only a small number are able
to save, a few respondents shared that they use part of the grant as capital for their
businesses or long-term investments.

29
 Figure 12. How beneficiaries spend their 4Ps cash grants

Source: FGDs with beneficiaries

Unanticipated school expenses are often the reason cited why the cash grant is not
enough. Although 4Ps students are enrolled in government schools with no tuition fee,
parents cite miscellaneous fees such as PTA fees, and contributions for the improvement
of school facilities (i.e. electric fan, TV). Parents also cite costs of projects and
workbooks or learning modules. One parent shared that as 4Ps beneficiaries, they are
expected by the school administration/faculty to support their children’s participation in
voluntary activities such as field trips.

In cases where beneficiaries have insufficient funds for household expenses, their coping
mechanisms include scrimping on household expenses, adjusting consumption by
reducing food consumption or purchasing cheaper alternative foods (e.g. instant noodles,
vegetables), prioritizing of school expenses, borrowing money or food items from small
stores in the community, as well as working additional jobs such as doing laundry, selling
food/snacks, assisting in manual labor in farms.

[Link]. Assessment of program implementers

The C/MLs and MRBs interviewed are generally satisfied with the grant amounts as they
can see improvements in wellbeing of the beneficiaries. However, they also
acknowledged that an increase in amount would help beneficiaries because of increase
in prices of food and other commodities. They noted that any increase in the amount of
the grants will help the beneficiaries. They also mentioned that the grants have already
been increased in the law.

For the national program management, the increase in the cash grant amounts in the law
can help address the did address the inflation rates given the end of the implementation
of the unconditional cash transfer program which was originally meant to cushion poor
households due to the inflation

30
 Prior the enactment of RA 11310, the grant’s purpose is to be an immediate aid for
children stay in school and get regular health check-ups, have their growth monitored,
and receive vaccines. The World Bank used the term cushioned in explaining the
relief/aid done by the grants from the program. It cushioned the poor from the
adverse impacts of various shocks experienced by the country for the past six years
through this grant. Now, with the increased benefit, the program was able to address the
varying inflation rates from different regions. UCT did address the inflation rates but
the increased benefit was very appropriate in expanding the objectives and targets
because of shortened length of stay of the beneficiaries in the program. The program
increased the grants a little from the inception of the program

4.3. Payment Frequency

4.3.1. Current payment frequency

The cash grants of the program are currently being paid to beneficiaries on a bimonthly basis
following the schedule presented below. Each cycle begins with the compliance of
beneficiaries during the monitoring period of two months, followed by the collection of
compliance data and payroll preparation on the third month, and the payout on the fourth month
(Table 11). Based on this cycle, a compliant beneficiary is expected to receive cash grants every
two months.
The frequency of payment has already been revised from its previous quarterly cycle from 2008
to 2010 in order to provide the cash grants more frequently to the beneficiaries and promote
consumption smoothing through the grants.

31
Table 11. Illustration of the bimonthly payment cycle of Pantawid Pamilya

Period 1 Period 2 Period 3 Period 4 Period 5 Period 6

FEBRUARY Compliance
monitoring
MARCH
APRIL Payroll Compliance
preparation monitoring
MAY PAYOUT
JUNE Payroll Compliance
preparation monitoring
JULY PAYOUT
AUGUST Payroll Compliance
preparation monitoring
SEPTEMBER PAYOUT
OCTOBER Payroll Compliance
preparation monitoring
NOVEMBER PAYOUT
DECEMBER Payroll Compliance
preparation monitoring
JANUARY PAYOUT
FEBRUARY Payroll
preparation
MARCH PAYOUT

4.3.2. Feedback and opinion on payment frequency

[Link]. Opinion of beneficiaries

Based on the FGD with the beneficiaries, nine of the 16 barangays that responded stated
that they preferred the current bimonthly schedule of provision of grants to be retained.
Six of the 16 barangays leaned towards more frequent provision of grants, citing that they
preferred that grants be provided monthly. One barangay had 50-50 opinion on the
matter.

32
Figure 13. Beneficiaries’ preferred payment frequency based on the FGDs

According to members who preferred to retain the current bimonthly schedule of payouts,
the timing will allow beneficiaries to receive a larger amount compared to more frequent
payouts. They also mentioned that increasing the payout frequencies will incur additional
transportation costs for beneficiaries that do not have ATMs in their area or residence.
They also said that provision of the grant every two months is acceptable since the
purpose of 4Ps is to supplement the schooling needs of children and beneficiaries do not
rely on the program entirely for their financial needs.

Meanwhile, respondents that were in favor of monthly payouts, argue that more frequent
provisions of grants will allow them to cover emergency or sudden and urgent expenses
(i.e. school projects). They also said that this will allow them to avoid needing to take out
loans to tide over the household while waiting for the next provision of the grant. Several
also mentioned that it will be easier for them to include the grants in the household
budgeting since the timing coincides with their monthly budget preparation.

None of the respondents preferred a frequency of three months or more citing that this
interval would be too long and would cause them to incur debts.

[Link]. Opinion of program implementers

According to the program implementers interviewed, increasing the payment frequency
is more possible for areas that are already using cash cards as mode of payment and will

33
 be more challenging for areas such as BARMM that are still under OTC mode of
payment.

Another major consideration that they mentioned is that increasing the payment
frequency means added cost for the program operations and bank service fees. Following
the increase in payment frequency is the increase in the frequency by which the
compliance verification (CV) process is done. This increase in the CV processing eeds
commensurate increase in operations budget and staff. As shown in Figure 10, the
compliance verification process and payroll preparation is largely manual – i.e., forms
are printed and filled-out during compliance data collection. Meanwhile, bank service
fees are charged by the LBP to DSWD/NG for withdrawal transactions of beneficiaries
to non-LBP ATMs. In 2020, the bank service fees budget amounted to PHP 509 million.

Figure 14. Compliance Verification Cycle

Source: Pantawid Pamilya Operations Manual – FMS Sub-manual (DSWD, 2015)

Efforts to streamline the process are currently being explored by the program
management such as linking the DepEd BEIS and the program monitoring system. The
NPMO also explored means to reduce the duration of the distribution and collection of
compliance verification forms to improve the timeline of the payment cycle. The
Compliance Verification Division initiated decentralization of processes to reduce the
timeline. It also developed an offline version of encoding of compliance data. The NPMO
also mentioned that these improvements were supposed to be launched this year but
because of the pandemic the program postponed the use of this system/module.

34
4.4. Payment Modality

4.4.1. Current modes of payment

The share of cash card payment mode has continued to increase from less than half in 2016 to
almost 90% as of June 2020. The biggest increase was observed in 2019 when the DSWD has
started to pursue full conversion of the mode of payment to EMV-enabled cash cards.

Table 12. Share of Cash Card and OTC payment modes among 4Ps households, 2016-2020

YEAR Cash Card OTC

2016 45.0% 55.0%
2017 44.3% 55.8%
2018 56.1% 43.9%
2019 86.0% 14.0%
2020 (Q2) 87.8% 12.2%
Source: DSWD (Pantawid Pamilya Quarterly reports, 2016-2020)

Among the Regions, only BARMM is predominantly still under OTC mode of payment
(93.1%). According to the DSWD, the remaining areas under OTC in other regions are mostly
GIDA areas. There are also few beneficiaries who are currently under OTC but are only waiting
for completion of account enrollment process. However, there are further delays being
experienced due to the COVID-19 pandemic.

Table 13. Share of Cash Card and OTC payment modes among 4Ps households, by region,
2020

Region Cash Card OTC Total % Cash Card % OTC

NCR 213,140 7,443 220,583 96.60% 3.40%
CAR 59,431 680 60,111 98.90% 1.10%
I 204,683 932 205,615 99.50% 0.50%
II 103,198 2,042 105,240 98.10% 1.90%
III 287,135 6,561 293,696 97.80% 2.20%
IV-A CALABARZON 311,758 5,598 317,356 98.20% 1.80%
MIMAROPA 173,228 20,896 194,124 89.20% 10.80%
V 348,760 19,417 368,177 94.70% 5.30%
VI 312,573 8,775 321,348 97.30% 2.70%
VII 279,410 5,217 284,627 98.20% 1.80%
VIII 248,426 31,446 279,872 88.80% 11.20%
IX 291,860 6,837 298,697 97.70% 2.30%
X 245,809 19,908 265,717 92.50% 7.50%
XI 247,013 9,727 256,740 96.20% 3.80%
XII 238,144 9,040 247,184 96.30% 3.70%
CARAGA 178,723 9,808 188,531 94.80% 5.20%
BARMM 26,758 361,362 388,120 6.90% 93.10%
Source: DSWD (Pantawid Pamilya Quarterly reports, 2020-Q2)
According to DSWD, the direction of the program is towards implementing a 100% Cash Card
(EMV Cards) mode of payment, and eventually, conversion to transaction account where
beneficiaries shall be issued with an account in the form of deposit account or electronic

35
money/ wallet, which can be used to store money, send payments, and receive deposits. These
features are currently not available with the

The NPMO also mentioned that the program plans to engage all possible Point of Sale in these
areas to address the limitation of ATM supply in some areas. An example the respondents cited
is engaging sari-sari store which has digital money account to distribute cash grants.

4.4.2. Feedback and opinion on payment modalities

Generally, the beneficiaries and key informants preferred cash card mode of payment over
OTC due to the convenience and reliability it provides, but their experience is usually affected
by the availability of ATMs or local banks in the area and the costs that the beneficiaries need
to incur. The respondents noted challenges and gaps that need to be addressed to improve the
payment delivery. These are discussed in the succeeding subsections.

[Link]. Experience and opinion of beneficiaries

Most respondents stated that they prefer receiving the grants through their cash cards.
Respondents mentioned it was more convenient since they can choose when to withdraw
the cash grant, unlike OTC where they must follow the schedule of the payout to receive
the grants in person. Most respondents also mentioned that they have experienced more
frequent delays in payouts when they were still under OTC mode of payment. They also
mentioned that waiting time is longer compared to withdrawal via ATM.

Only respondents from one site preferred to receive the grant through OTC, citing that it
was faster when payouts were conducted at their barangay gym and they did not have to
transfer from one place to another to withdraw the grant. In this area, there was no
available ATM and beneficiaries need to travel to the town center to withdraw from
money exchange centers through POS machines and a rural bank which provides over
the counter withdrawal of their grants. They also noted that OTC mode of payment do
not incur fees on withdrawal (e.g., POS service fees) that reduce the amount of grants
that they receive. These group of beneficiaries also reported that the queue in
withdrawing through their cash cards take longer than OTC payouts due to their limited
options for merchants and banks in their area. One respondent also mentioned that chairs
are provided during OTC venues, so the queuing is less tiring. However, the beneficiaries
also acknowledged that the cash cards could be a convenient mode of payment once the
lack of ATMs and alternative means of withdrawal in the area is addressed.

Despite general preference for cash card mode of payment, the beneficiaries noted some
challenges they experience in this mode of payment. One major challenge was the
duration of the process of card replacement. This was not a big problem for some
beneficiaries, since in some municipalities, grantees were able to get their grant OTC
while card was in process, but some were not able to receive their grant from 6 months
up to more than a year. Most of the respondents also mentioned that they have to wait in
line at the ATM for more than an hour when the grants are released due to the volume of
beneficiaries that flock the ATMs for withdrawal. In areas without ATM or LBP
branches, beneficiaries either go to nearby town to withdraw their cash grants and incur
transportation costs or go to private establishments that allow withdrawal through Point-
of-Sale (POS) transactions/ withdrawal services. In one of the study areas, there was no
standardization of fees collected by private establishments that offer encashment using

36
POS machines. In the interview with the LBP key informants, they mentioned that an
accreditation process is currently being done to formally engage these establishments.
However, it is not certain whether the POS merchants accessed by the beneficiaries in
the FGDs are included in the accredited merchants of the LBP.

To address the challenges mentioned by beneficiaries, they suggested improving their

access to ATMs or withdrawal options. Some suggested that grants should be released in
the barangay so that beneficiaries do not have to spend money on transportation. Some
also suggested to have a fixed day (or week) of the month for release of the grant (e.g.
every first week of the month) for the payouts to be more predictable.

[Link]. Experience and opinion of local 4Ps staff

Most of the C/MLs and MRBs believe that the cash card payment is better as it is more
convenient for beneficiaries. However, there were also a few that wanted to return to the
OTC payment of specific challenges in cash card-based payments. These respondents
noted that there are still issues that they need to resolve even with cash cards (e.g., lost
cards, wrong pin entry, etc.) while they can immediately address queries on-site during
OTC payouts. They also mentioned that OTC payments minimize risk of pawning of cash
cards.

The C/MLs and MRBs also noted challenges that they experience in both modes of
payment. They mentioned that OTCs are more logistically challenging to conduct since
it requires coordination with local government units for set-up of venue and scheduling.
There were also more documentation and reportorial requirements such as physical
acknowledgement receipts compared with electronic bank records in cash card mode of
payment. Payout schedules are also more susceptible to delays due to issues in
procurement of conduits, weather disturbances, security threats, among others.

For cash card-based payments, the C/MLs and MRBs mentioned that their main
challenge is the need to frequently troubleshoot concerns of the beneficiaries with their
cash cards such as cases of “forced pin”, loss and damage of cards, card capture, and
disputes in the amount of grants received. They also mentioned that not all areas have
ATMs or banks which affects the access of beneficiaries to their cash grants. Similar to
what was mentioned by beneficiaries, the C/MLs and MRBs also mentioned that duration
of the processing of card replacement takes long and feedback loop on status has gaps.
In particular, some of the respondents lamented that they have difficulty providing
feedback to beneficiaries on the status of processes that are submitted to higher level of
program management (i.e., regional, and national level) as they do not have immediate
access to the status of these requests. A few also mentioned the risk of cash card pawning
as a challenge in using cash card mode of payment.

The respondents suggested the provision of satellite ATMs for areas with no LBP
branches the access of beneficiaries and improve the payment delivery system. With
regard to the administrative processes, the respondents recommended streamlining and
expediting of processing of card replacement and change grantee updates. They
suggested that the processing of card replacement should be done at the regional/local
level instead of at the central office of DSWD and Landbank. To improve feedback loops
to the beneficiaries, the respondents suggested that they should be provided access to
real-time status of update processing and grievance resolution that they facilitated. They

37
 also recommended the use of an online system or dashboard for the status of approval of
updates to avoid conflicting information that will be provided to beneficiaries.

[Link]. Opinion of national program staff

DSWD-FMS key informants believe that the conversion to cash cards for most of the
beneficiaries greatly improved the performance of the payment system. Regarding other
e-payment facilities (e.g., GCASH, PAYMAYA), respondents mentioned that there
might not be added value to this since LBP is already working to provide the same
features to beneficiaries once their cards are converted to fully transactional accounts as
promised by LBP. The conversion of confirmed by the interviews with LBP staff and
the National Program Management office. The key informants also noted that switching
to other modes of grants payment might pose difficulties because beneficiaries would
have a new system to learn and get comfortable with. They also added that many features
of these digital banking technologies are already offered in the LBP mobile and the
beneficiaries just need to be trained and taught how to access these features

In the response sent by the National Program Management Office, the program is also
currently exploring financial technology solutions in line with a general effort to shift to
digital delivery of social protection programs. They also noted that this has already been
communicated with LBP.

The NPMO also noted that there is a need to map out the areas and categorize each of
them according to readiness in implementing cash card mode of payment. Those which
are ready, shifted to full cash card release of grants while those areas that are not ready
can be classified as transition the areas that should adopt year-on-year plan to shift to
cash card.

They also noted the need to update the information systems of the program and upgrade
and IT equipment of the local staff so the data feedback can be more fluid. They also
mentioned the need strengthen the information systems of the program to self-check for
data errors. To streamline processes in the payment system, particularly card replacement
and enrollment of cash card accounts, the NPMO recommends that the LBP should
decentralize its processes and allow branches to open and manage the accounts of
beneficiaries in their areas.

4.4.3. Differential impact of program by payment modality

This section presents the results of the analysis of IE3 data that looked into the difference in
impact of the program on subgroups of beneficiaries with different modes of receiving their
cash grants namely through over-the-counter means and through cash card. The difference in
the impact of program on these subgroups could show advantages of pursuing specific design
features of the payment system. The results are presented for four groups of outcomes: (1)
children’s health; (2) education; (3) income and expenditure of the household; and (4) labor
outcomes. The full statistical tables of the results are shown in the appendix and only the
summary of results will be presented in the succeeding discussion.

In summary, the results show that the difference in program impact on health and education
outcomes for the two subgroups of beneficiaries do not significantly matter. The mixed results
in education suggest that both groups are favorably impacted by the program regardless of the
mode of payment. In terms of total expenditure and income, the results suggest that the

38
subgroup of beneficiaries with cash cards are more positively impacted by the program.
Although further study is needed to explain the results, a possible explanation could be the
more timely and predictable provision of grants through the cash cards relative to OTC means.
The results also indicate shifts in differences in expenditure patters where cash card
beneficiaries spend more on non-food expenditures, particularly health and education, while
OTC beneficiaries prefer food and clothing expenditures. Results on employment show
increased duration of work for cash card beneficiaries which could possibly explain the
observed positive impact on income excluding the grants.

[Link]. Differential impact on child health

In summary, the results of the subgroup analysis on the program impact on child health
show that for most of the outcomes examined, there is no significant difference in the
program impact among beneficiaries that receive their cash grants through OTC means
or through cash card. Of the outcomes, only the health visits of children and deworming
of children under six years old have shown significant difference in impact between the
two subgroups. These results, however, could be seen as conflicting in terms of the health
seeking behavior being illustrated by the beneficiaries. In the health visits outcome, a
negative impact was observed for the OTC group while an increase in deworming rate of
children was also observed. These results could be clarified by further investigating other
factors that could cause such discrepancy in addition to the possible effect of the mode
of payment.

Table 14. Summary of results on subgroup analysis of 4Ps impact on child health by
mode of cash grant payment

Outcome group and indicator Highlight of results

Growth monitoring
Regular weight monitoring for No significant impact observed for both Cash Card and
children 0 to <2 years old OTC subgroups
Frequency of weight monitoring for No significant impact observed for both Cash Card and
children 0 to 2 years old in the past OTC subgroups
six months
Child health services and practices
Vitamin A supplementation (6 Positive program impact observed within sampling
months to 6 years old) bandwidth only for both subgroups.
Full immunization at age 1 No significant impact observed for both Cash Card and
OTC subgroups
Visited a health facility or health Negative impact observed on OTC group, no
professional in the past 8 weeks significant impact on cash card group.
Difference between OTC and Cash card group is
significant.
Deworming
Deworming (under 6 years old) Positive impact on OTC group, no significant impact on
cash card group.
Difference between OTC and Cash card group is
significant
Deworming at least once (6 to 14 Positive impact observed in OTC group but not
years old) significantly different with Cash card group

39
 Outcome group and indicator Highlight of results
Deworming at least twice (6 to 14 Positive program impact observed for both subgroups.
years old)
Child nutrition outcomes
Underweight No significant impact observed for both Cash Card and
OTC subgroups
Severe underweight No significant impact observed for both Cash Card and
OTC subgroups
Stunting Increase in prevalence observed in OTC group within
sampling bandwidth only, but estimate is not
significantly different with Cash Card group
Severe stunting Increase in prevalence observed in OTC group within
sampling bandwidth only, but estimate is not
significantly different with Cash Card group
Wasting No significant impact observed for both Cash Card and
OTC subgroups
Severe wasting No significant impact observed for both Cash Card and
OTC subgroups

[Link]. Differential impact on education

Results of the subgroup analysis on the impact of education is mixed. For enrollment,
significant positive impact is more frequently observed for the subgroup that receives
their cash grants through cash card mode of payment, while no significant impact was
observed in most enrollment outcomes of the OTC subgroup, although, the differences
in estimates between the OTC and Cash card group are not significantly different based
on the estimated z statistics. On the other hand, for attendance rates, positive impact is
more commonly observed in OTC subgroup of beneficiaries while no impact is observed
for the cash card group. For dropout rates, impact is more evident in the cash card
subgroup while for child labor, the program impact is more observed in OTC subgroup
of beneficiaries. In terms of education expenditure, impact on total education expenditure
is observed in cash card subgroup while the impact on the OTC subgroup is positive for
clothing/uniform and school supplies or project materials, and negative for school fees.
With these results, it is not clear whether there is a significant advantage of a mode of
payment over the other in terms of education outcomes. Certain aspects of the mode of
payment, such as predictability and timeline of payouts vis-à-vis payment collection
schedules of schools may explain these results but need to be investigated using more
detailed information.

Table 15. Summary of results on subgroup analysis of 4Ps impact on education by

mode of cash grant payment

Outcome group and indicator Highlight of results

Enrollment
Enrollment of children 3 to 5 No significant impact for both subgroups
years old
Enrollment of children 6 to 11 Positive program impact observed for both subgroups.
years old
Enrollment of children 12 to 15 Positive impact observed in Cash Card group, but
years old difference with OTC group is not significantly different

40
Outcome group and indicator Highlight of results
Enrollment of children 16 to 17 Positive impact observed in Cash Card group, but
years old difference with OTC group is not significantly different
Enrollment of children 12 to 17 Positive impact observed in Cash Card group, but
years old difference with OTC group is not significantly different
Enrollment of children 6 to 14 Positive impact observed in Cash Card group, but
years old difference with OTC group is not significantly different
Enrollment of children 15 to 20 Positive impact observed in Cash Card group within
years old sampling bandwidth only, but difference with OTC
group is not significantly different
Attendance rates (in %)
Attendance rate of children 3 to No significant impact for both subgroups
5 years old
Attendance rate of children 6 to Positive impact on OTC group, no significant impact on
11 years old cash card group.
Difference between OTC and Cash card group is not
significant
Attendance rate of children 12 to No significant impact for both subgroups
15 years old
Attendance rate of children 16 to Positive impact on OTC group, no impact on cash card
17 years old group.
Difference between OTC and Cash card group is
significant
Attendance rate of children 12 to No significant impact for both subgroups
17 years old
Attendance rate of children 6 to Positive impact observed in OTC group within sampling
14 years old bandwidth only
Difference between OTC and Cash card group not
significant
Attendance rate of children 15 to No significant impact for both subgroups
20 years old
Attendance of at least 85%
Attendance of at least 85% No significant impact for both subgroups
among children 3 to 5 years old
Attendance of at least 85% Positive impact on Cash Card group, no significant
among children 6 to 11 years old impact on OTC group within MSE bandwidth.
Difference between OTC and Cash card group not
significant
Attendance of at least 85% No significant impact for both subgroups
among children 12 to 15 years old
Attendance of at least 85% Positive impact on OTC group, negative impact on cash
among children 16 to 17 years card group.
old Difference between OTC and Cash card group is also
significant
Attendance of at least 85% No significant impact for both subgroups
among children 12 to 17 years
old
Attendance of at least 85% No significant impact for both subgroups
among children 6 to 14 years old

41
Outcome group and indicator Highlight of results
Attendance of at least 85% No significant impact for both subgroups
among children 15 to 20 years
old
School level enrollment

Enrollment in daycare, nursery, No significant impact for both subgroups

preschool/kindergarten of
children 3 to 5 years old
Enrollment in preschool or No significant impact for both subgroups
kindergarten children 5 years old
Enrollment in elementary of No significant impact for both subgroups
children 6 to 11 years old
Enrollment in junior high school Positive impact on Cash Card group, no significant
of children 12 to 15 years old impact on OTC group. Difference between OTC and Cash
card group not significant
Enrollment in senior high school Positive impact on Cash Card group, no significant
of children 16 to 17 years old impact on OTC group. Difference between OTC and Cash
card group not significant
Dropout rate
Dropout rate among children 6 to Significant reduction in drop-out rate observed in Cash
11 years old Card group within sampling bandwidth only but not
significantly different with OTC group
Dropout rate among children 12 Significant reduction in drop-out rate observed in Cash
to 15 years old Card group within sampling bandwidth only but not
significantly different with OTC group
Dropout rate among children 16 No significant impact for both subgroups
to 17 years old
Dropout rate among children 12 Significant reduction in drop-out rate observed in Cash
to 17 years old Card group within sampling bandwidth only but not
significantly different with OTC group
Dropout rate among children 6 to Significant reduction in drop-out rate observed in both
14 years old subgroups
Dropout rate among children 15 No significant impact for both subgroups
to 20 years old
Child labor
At least 1 hour of work (with or Significant reduction in child labor incidence observed in
without pay) last month, 10-14 OTC group, but not significantly different with Cash Card
years old group
At least 1 hour of paid work last Significant reduction in child labor incidence observed in
month, 10-14 years old OTC group within MSE bandwidth, but not significantly
different with Cash Card group
Number of days worked (with or Significant reduction in drop-out rate observed in Cash
without pay) last month, 10-14 Card group within sampling bandwidth only but not
years old significantly different with OTC group
Education expenditures
Expenditures on tuition and other Negative impact on OTC group, no significant impact on
fees (per month) in the last school Cash card group. Difference between OTC and Cash card
year group is not significant

42
 Outcome group and indicator Highlight of results
Expenditures on school materials Positive impact on OTC group, no significant impact on
and supplies (per month) in the Cash card group. Difference between OTC and Cash card
last school year group is not significant
Expenditures on school uniform Positive impact on OTC group, no significant impact on
(per month) in the last school Cash card group. Difference between OTC and Cash card
year group is not significant
Expenditures on school allowance Positive impact on Cash Card group within sampling
(per month) last school year bandwidth only, no significant impact on Cash card
group. Difference between OTC and Cash card group is
not significant
Total school expenditures (per Positive impact on Cash Card group within sampling
month) last school year bandwidth only, no significant impact on Cash card
group. Difference between OTC and Cash card group is
not significant

[Link]. Differential impact on household expenditures and income

Results of the subgroup analysis on household income and expenditure some evidence
on differential impact of the two modes of payment. Results show positive impact on the
share of food and clothing expenditures of the beneficiaries in the OTC subgroup. The
shift in share of expenditures of beneficiaries under the OTC subgroup towards food and
clothing could be related to the emergence of flea markets in OTC payout venues as noted
by Adriano (2016) and other anecdotal evidence. Meanwhile, positive impact on non-
food expenditures is observed for beneficiaries in the cash card subgroup. Looking
further into the results, this increase in non-food expenditures could be due to the positive
impact on medical and education expenses. The increase in total expenditures for
education is also consistent with the results of the previous subsection.

The results also show that positive impact on total average per capita expenditure is only
observed for the cash card group of beneficiaries. This could be due to the relatively
faster delivery of grants through the cash card which increases the relative frequency of
payouts through this means compared to OTC payouts. As presented in Section 3, the
beneficiaries under cash card mode of payment were able to have payouts slightly more
frequently than the beneficiaries under OTC mode.

In terms of total per capita household income, positive impact is observed on both
subgroups of beneficiaries suggesting that the receipt of grants is able to augment the
household income regardless of the mode of delivery. But this positive impact is retained
only for the cash card group when the grants are excluded or when it is limited to income
from salaries and wages. Further study is needed to examine this result to see whether
aspects of the mode of payment lead to this result or some related intermediate outcomes.

With regard to self-rated poverty, an interesting result is the positive impact on self-rating
of cash card beneficiaries and negative impact on self-rating of OTC beneficiaries. This
could be studied further to understand the psychology of the beneficiaries related to the
mode of payment.

43
Table 16. Summary of results on subgroup analysis of 4Ps impact on household
expenditure and income by mode of cash grant payment

Outcome group and Highlight of results

indicator
Household expenditures: Share to total expenditures
Share of food to total Positive impact on OTC group, no significant impact on Cash
expenditures card group. Difference between OTC and Cash card group is
significant
Share of non-food to total Negative impact on OTC group, no significant impact on Cash
expenditures card group. Difference between OTC and Cash card group is
significant
Share of education to total No significant impact for both subgroups
expenditures
Share of clothing and Positive impact on OTC group, no significant impact on Cash
footwear to total card group. Difference between OTC and Cash card group is
expenditures significant within sampling bandwidth
Share of health to total No significant impact for both subgroups
expenditures
Share of alcohol and tobacco No significant impact for both subgroups
to total expenditures
Household expenditures
Average total per capita Positive impact on Cash-card subgroup but difference is not
expenditure significant with OTC group
Average total per capita food No significant impact for both subgroups
expenditure
Average total per capita non- Positive impact on Cash-card subgroup and slight negative
food expenditure impact on OTC group. Difference is significant
Average per capita non-food Positive impact on Cash-card subgroup and slight negative
expenditure (including other impact on OTC group. Difference is significant
disbursements)
Average per capita No significant impact for both subgroups
expenditure on vice goods
(e.g., alcohol, tobacco)
Average per capita No significant impact for both subgroups
expenditure on inpatient care
Average per capita No significant impact for both subgroups
expenditure on outpatient
care
Average per capita Positive impact on Cash Card group, no significant impact on
expenditure on medical OTC group. Difference between OTC and Cash card group
services and commodities not significant
Average per capita Positive impact on Cash Card group, no significant impact on
expenditure on education per OTC group. Difference between OTC and Cash card group
school age child not significant
Average per capita Positive impact on OTC group, no significant impact on Cash
expenditure on clothing and Card group. Difference between OTC and Cash card group is
footwear significant
Income
Per capita income including Positive program impact observed for both subgroups.
grants

44
 Outcome group and Highlight of results
indicator
Per capita income without Positive impact on Cash Card group, no significant impact on
grants OTC group. Difference between OTC and Cash card group is
significant
Per capita income from Positive impact on Cash Card group, no significant impact on
salaries and wages OTC group. Difference between OTC and Cash card group is
significant
Per capita income from No significant impact for both subgroups
entrepreneurial activities
Per capita income from other No significant impact for both subgroups
receipts (excluding grants)
Hunger and self-rated poverty
Incidence of hunger No significant impact for both subgroups
Number of days experienced Significant reduction in duration of hunger observed in OTC
hunger in the past 3 months group, but not significantly different with Cash Card group
Self-rated poverty status No significant impact for both subgroups
(Poor)
Self-rated poverty status Positive impact on Cash Card group, negative impact on OTC
(Not-Poor) group. Difference between OTC and Cash card group is
significant

[Link]. Differential impact on household expenditures and income

In terms of labor outcomes, the difference in impact between the two groups is observed
on the duration of working hours and looking for additional jobs. Positive impact on total
work hours is observed for beneficiaries in the cash card group while negative impact is
observed for OTC beneficiaries. While this result warrants further study to explain the
discrepancy, the increase in working hours for cash card beneficiaries could be a possible
explanation to the observed positive impact on salaries and wages of this subgroup of
beneficiaries which was discussed above. With regard to the negative impact on job-
seeking behavior, negative impact is observed for the subgroup of beneficiaries that
receive their grants through cash cards. However, the previous result on the longer
duration of working hours for this subgroup of beneficiaries justify this.

Table 17. Summary of results on subgroup analysis of 4Ps impact on labor outcomes
by mode of cash grant payment

Outcome group and indicator Highlight of results

Labor force participation No significant impact for both subgroups
Employment Negative impact observed in Cash Card group within MSE
bandwidth, but not significantly different with OTC group
Usual work hours per week in Positive program impact observed for both subgroups.
primary occupation
Other job or business besides No significant impact for both subgroups
primary occupation
Usual work hours per week in Negative impact observed in OTC group , no significant
other jobs impact on Cash card group. Difference between OTC and
Cash card group is significant

45
 Total usual work hours per Positive impact on Cash Card group, no significant impact
week on OTC group. Difference between OTC and Cash card
group is significant
Looking for additional work if Negative impact on Cash Card group, no significant impact
employed on OTC group. Difference between OTC and Cash card
group is significant
Unemployed and looking for No significant impact for both subgroups
work

4.5. 4Ps Payments System and SAP Implementation

This section presents the findings of the study related to the implementation of the Social
Amelioration Program and the role of the Pantawid Pamilya payment system in the
disbursement of the SAP grants. The study took advantage of implementation on SAP to learn
on the performance of the 4Ps payment system in times of emergencies. The results are mainly
from the online survey on the SAP among 4Ps beneficiaries but insights from the interviews of
the DSWD respondents are also included to provide more information from the implementors
perspective. The online survey run from April to December 2020.

In summary, the feedback of beneficiaries and key informants on the SAP implementation
suggest that the experience of 4Ps beneficiaries on the distribution of the SAP is instructive in
terms of ease and speed of reaching targeted beneficiaries. The beneficiaries did not report
significant challenges in withdrawing their grants which they were able to receive through their
cash cards. The Key informants cite that ease of the SAP distribution to 4Ps was due to the
existing registry of beneficiaries and established system of payment of grants to the households.

4.5.1. Online SAP Survey

[Link]. 4Ps Payment System

Almost all of the beneficiaries in the online survey data (96.8%) are currently enrolled
in an LBP cash card account as part of the Pantawid Pamilya payment system. Majority
of them also reported that their cash cards are currently in their possession. Two
beneficiaries reported that they have already been enrolled in a cash card account but
have not yet claimed or received their cash cards while three beneficiaries lost or
damaged their cards and are still waiting for these to be replaced. Six beneficiaries
reported other reasons why their cash cards are not in their possession including pending
applications for change of their household grantee and locked cash card due to incorrect
pin entry. Meanwhile, 31 respondents (3.5%) said that they are still not enrolled in an
LBP cash card account and 40 (4.5%) are unaware or uncertain of the status of their LBP
cash card enrollment.

Table 18. Status of enrollment and possession of LBP cash card among 4Ps online
survey respondents

Status of enrollment in LBP Cash card account Frequency Percentage

Enrolled 858 96.8%
Cash card in possession 847
Cash card not claimed 2

46
 Cash card lost or damaged 3
Cash card not in possession due to other reasons 6
Not enrolled 16 1.8%
Do not know/ Uncertain of status 12 1.4%

Almost all of the beneficiaries reported that they usually claim their cash grants by
withdrawing from ATMs of LBP or other banks. Seventeen respondents reported
receiving their grants through OTC means including five respondents who also claimed
they have been enrolled in an LBP cash card account. OTC means mentioned by the
respondents include withdrawing through LBP branch over the counter and encashment
of cheques through other banks. Eighteen respondents reported that they are currently
not receiving benefits from the program due to pending change grantee application, or
they are a newly registered beneficiary that have yet to receive grants.

Table 19. How respondents usually receive their 4Ps cash grants

Usual manner of
Number of beneficiaries by LBP enrollment status Total
receiving 4Ps cash grants
Enrolled Not enrolled Unknown status
ATM (LBP or other banks) 850 - - 850
Point-of-Sale (POS) 1 - - 1
machine
Over-the-counter (OTC) 5 10 2 17
Currently not receiving 2 6 10 18
TOTAL 858 16 12 886

[Link]. Receipt of SAP cash grant

At the time of answering the survey, around 85% of the respondents already received at
least one payment of the SAP financial assistance. Of those who already received their
SAP grants, 48 (6% of 759) of the beneficiaries reported receiving two payments of the
SAP subsidy, while the majority reported receiving only once. Note that at the time when
most of the responses were collected, the government has nearly completed disbursing
the first tranche of payment of SAP grants to 4Ps beneficiaries. Those responses claiming
two payments of SAP grants have already been provided to them were mostly received
during the latter period of data collection where the SAP distribution is on the 2nd tranche
of grants.

Table 20. Status of enrollment and possession of LBP cash card among 4Ps online
survey respondents

Receipt of SAP cash grants Frequency Percentage

Received 759 85.7%
Once (first tranche) 704
Twice (first and second tranche) 48
No response 7
Not yet received 127 14.3%

Of those who received their SAP subsidy once, the average amount they received is PHP
6,322. Of those who received their SAP subsidy twice, the average amount for the first
receipt is PHP 6,359 and PHP 807 on the second receipt. The low amounts of the second
47
receipt claimed by the respondents may be an indication that this is not a part of the SAP,
but of their regular cash grants from the 4Ps. The values reported for this item range from
400 and PHP 6,700. For reference, the correct values of SAP subsidy that should be
received by 4Ps beneficiaries is reflected in Table 22. According to the guidelines in the
SAP provision, the beneficiaries are entitled to PHP 5000 to PHP8000 SAP subsidy
depending on the region of residence. However, PHP1350 of the said amount will be
covered by the cash grants that they were supposed to receive during the period.

Table 21. Average amount of SAP grant received by frequency of receipt

Average Number of
Frequency of receipt of SAP cash grants (PHP) observations
Once (first tranche) 5,963 579
Twice (first and second tranche)
1st tranche 6,358 36
2nd tranche 807 36

48
Table 22. Correct SAP subsidy per region

Total amount of SAP Total amount of SAP excluding the

Region (in PHP) cash grants (PHP)
NCR 8,000 6,650.00
III, IV-A 6,500 5,150.00
VI, VII, X, XI 6,000 4,650.00
CAR, I, II 5,500 4,150.00
IV-B, V, VIII, IX, XII,
CARAGA, BARMM 5,000 3,650.00
Source: DSWD Memorandum Circular No, 14 s. 2020

[Link]. Mode of payment

Majority of the respondents were able to withdraw their SAP financial assistance
through Landbank and non-Landbank ATMs. Only around 3% of the respondents were
able to receive their financial assistance through other modes such as POS machines,
assemblies called by the DSWD or LGU, and house to house visits.

Table 23. How respondents received their SAP cash grants

Manner of receiving SAP cash grants Frequency Percentage

ATM 592 97.2%
POS 4 0.7%
OTC in bank branch 1 0.2%
Assembly called by LGU, DSWD or LBP 10 1.6%
House to house by DSWD 2 0.3%
Total 609

Most of the beneficiaries reported walking to the ATM, bank, or OTC venue where they
withdrew their SAP cash assistance. Of those who had to ride in their own vehicle or
public transportation, the median expense incurred is PHP 128, while highest value of
expense was reported at PHP2500. In terms of duration of travel, the median reported by
those that travelled through their own vehicle, public transportation, or LGU service
vehicle was 30 minutes. The longest travel time however is 300 minutes or equivalent to
5 hours. In terms of duration of queueing in ATMs, banks, or other payout venues for
SAP, the median time reported by the respondents is 60 minutes. The longest duration id
queueing however is 300 minutes or equivalent to 6 hours.

Table 24. Average and median transportation cost and time spent during receipt of
SAP grants

Number of Average Median Min Max

observations
Transportation expense (in PHP) 87 128.5 100 10 2500
Travel time (in minutes) 124 57.1 30 2 300
Queue time (in minutes) 505 73.9 60 1 600

49
Almost all of the respondents (97%) who received their SAP in venues outside their
homes reported that social distancing was followed all the time while they were claiming
their SAP financial assistance. Several reported that social distancing was followed only
partially.

[Link]. Correctness of Amount of Financial Assistance

Around two-thirds (66%) of the respondents think that the amount they received for SAP
is sufficient to cover their needs during the quarantine. The rest of the respondents think
otherwise.

Respondents were also asked whether they think the amount they received is correct
based on their knowledge of the SAP guidelines that they are supposed to receive.
Majority (81%) thinks that the amount they received is consistent with the guidelines that
they know. Of those who think the amount they received is incorrect according to
guidelines, their expected amounts range from PHP1,500 to PHP 50,000. The median
value they expect is PHP 8,000.

[Link]. Use of SAP Financial Assistance

The table below shows the expenditure items where the respondents spent their SAP
financial assistance. From the proportions, almost all of the respondents spent their
financial assistance on food for the household. More than half also spent the money on
medicine and other health-related expenditures. Around 12% of the beneficiaries also
reported using their SAP financial assistance to extend aid to other families. These other
families could be their relatives or neighbors in the community, as well as through other
humanitarian or relief operation modes.

Table 25. Top expenditure items where beneficiaries used their SAP grant (n=759)

Expenditure item Percentage

Food 71.0%
Health and Medicine 38.0%
Aid for others 12.0%
Non-food household expenses 6.0%
Savings 5.0%
Investment in business 3.0%

[Link]. Experience in SAP

In terms of their overall experience in the SAP implementation, 26% of the respondents
reported having no issues or challenges experienced in the implementation of the SAP.
Around 24% have reported that there were households they feel should not be given
financial assistance, while 19% think that there are those who should receive financial
assistance but were not given. Another common experience reported by the respondents
is receiving negative comments from neighbors and other people on their being SAP
beneficiaries. This was reported by 17% of the respondents. None reported experiencing

50
difficulty in withdrawing their SAP grants although this was included in the pre-coded
responses available in the questionnaire.

Table 26. Problems experienced in SAP implementation (n=759)

Problems experienced in SAP implementation Percentage

No problems experienced 26.0%
Some who are deserving did not receive SAP 24.0%
Some who are not deserving were able to receive SAP 19.0%
Disbursement of grants took a long time 17.0%
Insufficient SAC forms 12.0%
Lack of transportation to banks, ATM, or assembly venue 12.0%
The grants were shared by more than one family 6.0%
Confusing or frequently changing guidelines in implementation 6.0%
Barangay officials did not know what they should do 4.0%
The amount of grant is insufficient 3.0%
The amount of grant received is incorrect 3.0%
Negative comments directed (to respondents) by neighbors and other people 2.0%

[Link]. Source of information of those who have not received SAP

Among those who have not received their SAP financial assistance at the time of the
interview, majority (98 out of 125 respondents) still have not received any advice on the
schedule of their SAP assistance. Among those who have already received information
(27 respondents), their most common sources are the DSWD staff, LGU officials,
community associations (e.g., Home-owners’ Association), and/or neighbors.

Table 27. Sources of information on SAP payout schedule for those who have not
received SAP (n=27)

Source of information Percentage

DSWD Staff 44.0%
LGU Staff 30.0%
Home-owners' Association or similar organizations 22.0%
Neighbors or relatives 15.0%
Social media (Facebook, Twitter, Instagram) 11.0%
TV or radio 4.0%
[Link]. 5. Non-4Ps Experience
Majority of the respondents (62%) claimed that they know of a non-4ps beneficiary who
received the SAP financial assistance.

Beneficiaries were also asked what they observed, know, or perceived as the experience
of non-4Ps recipients of the SAP financial assistance. The same patterns in responses are
observed to when they were asked to report on their own experiences. Around 22% of
the respondents reported that they think there were no issues or challenges experienced
in the implementation of the SAP among the non-4ps. Around 28% have reported that
there were households they feel should not be given financial assistance, while 20% think
that there are those who should receive financial assistance but were not given. The

51
percentage of respondents that reported the issue on negative comments from neighbors
and other people is around 20% of the respondents.

Table 28. Problems experienced by non-4Ps in SAP implementation (n=886)

Problems experienced in SAP implementation Percentage

No problems experienced 22.0%
Some who are deserving did not receive SAP 28.0%
Some who are not deserving were able to receive SAP 20.0%
Disbursement of grants took a long time 20.0%
Insufficient SAC forms 17.0%
Lack of transportation to banks, ATM, or assembly venue 9.0%
The grants were shared by more than one family 8.0%
Barangay officials did not know what they should do 3.0%
Confusing or frequently changing guidelines in implementation 3.0%
The amount of grant is insufficient 2.0%
The amount of grant received is incorrect 1.0%
Negative comments directed to us by neighbors 1.0%

[Link]. Role of Parent Leaders

More than one-fourth of the respondents (173 of 612) are 4Ps Parent leaders. Based on
the responses, more than 80% of these parent leaders reported that almost all or all of
their 4Ps members have already received their financial assistance at least once. Among
those that reported at least half of their members not yet receiving SAP, the reasons they
reported include lack of cash cards for other members and failure to contact or
communicate with some members.

52
 Table 29. Proportion of 4Ps beneficiaries in the parent group that received SAP
according to Parent Leaders (n=612)

Proportion of members in parent groups that Percentage of responses

received SAP (by parent leaders)
All members 41.5%
Almost all 39.6%
More than 50% 7.9%
Half or 50% 4.9%
Less than 50% 0.6%
Almost none/ few of the members 1.8%
None of the members 3.7%

Most of the parent leaders play a role in the implementation of SAP among 4Ps
members. Their primary role is in information dissemination to their member
beneficiaries; this is done by almost all of the parent leaders who participated in the
survey. Several of them have also reported being involved in actual implementation
including distribution of the SAP financial assistance and identifying potential
beneficiaries.

4.5.2. Opinion of DSWD Key Informants on role of 4Ps Payment System in SAP
implementation

DSWD key informants noted that it was relatively easy for them to distribute the SAP benefits
to Pantawid Pamilya beneficiaries compared to non-beneficiaries of the program since this only
required them to provide the list of beneficiaries to the Landbank and request them to transfer
the funds of SAP to the accounts of the 4Ps beneficiaries. They acknowledged that this ease
was due to the existing database of beneficiaries and the wide coverage of the beneficiaries in
the cash card payment system.

When asked whether the current payment system appropriate or sufficient to handle a program
as large as the Social Amelioration Program, the national program management noted that the
current payment system, as it is, is only intended for Pantawid Pamilya beneficiaries and may
not be able to cater to other programs such as the SAP which has different coverage targets.
However, they also noted that the program’s payment system should be improved to readily
adapt and efficiently deliver benefits in times of emergency situations.

5. Summary and Recommendations

The study presents an assessment of the payment system of the Pantawid Pamilyang Pilipino
Program using a combination of qualitative and quantitative methodologies. Data sources
include program administrative reports, Focus Group Discussions with program beneficiaries,
interviews with program implementers, survey data of the 3rd impact evaluation of the program,
and an online survey that aims to collect information on the experiences of beneficiaries in
receiving their grants and the implementation of the Social Amelioration Program.

53
In terms of the benefit level, the results show that the amount of cash grants have remained at
their nominal levels starting 2008 up to 2016 even though the real value has already decreased
due to inflation. The amount of grants have only recently increased due to the Rice subsidy and
RA 11301 which started implementation in 2017 and 2020, respectively. In those years where
the grants have remained stagnant, the reduction in real value of the grants also reduced its
ability to augment the income of the beneficiaries and aid human capital investment of children.
Compared with other countries, the 4Ps grants are lagging in terms of its generosity.

When asked to give an assessment of the benefit level of the program, beneficiaries are hesitant
to demand increase in grant amounts but admit that their budget is barely enough to cover
needs. Most of the grants are spent on education expenses of children and food for the family.
Program implementers also acknowledged that an increase in the cash grants will be helpful to
cushion the increase in prices and possible economic shocks to the households.

Evidence on the perceived optimal frequency of payment of benefits are mixed. While more
frequent payments result in consumption smoothing, less frequent payments also resulted in
positive impact on savings and asset accumulation. Results from the FGDs with beneficiaries
also show that there is no strong demand for more frequent payments among the beneficiaries
and most prefer the current bimonthly payment of grants due to additional transportation costs
to be incurred and smaller amount of grants to be received every payout. As regards the
feasibility of increasing the payment frequency, program implementers noted that there are cost
considerations for the program. These include costs for operations in the compliance
monitoring, and cost for bank service fees.

In general, the program payment system have improved through the years despite gaps in the
initial years of its implementation. The conversion of the mode of delivery has been mentioned
by program implementors as the key improvement in the payment delivery. This is also
confirmed by accounts of beneficiaries saying that there are less delays in the payout of grants
in the recent years and noted that convenience offered by cash cards. However, the experience
of beneficiaries are largely dependent on the availability of ATMs or POS merchants in their
community or nearby areas. In addition, both beneficiaries and program implementers also
noted some issues that still need to be addressed even in the cash card mode of payment. This
includes difficulty in accessing ATMs and/or banks in some of the areas, and delays in the
approval of cash-card related processes such as replacement of cards in case of change of
grantee, and broken feedback loop on the status of some processes.

IE3 evidence suggest mode of payment do not create significant heterogeneity in impact of the
program for education and health outcomes. However, results for household income and
expenditure indicate more pronounced positive impact of the program for cash card holders.
While this needs further study to ascertain pathways for such discrepancy in impact, it is
possible that the predictability and reliability of payouts in cash cards relative to OTC means
have contributed to the positive effect of the program.

Results of the quick assessment of the Social Amelioration Program point to the ease of
providing emergency cash assistance to the Pantawid Pamilya beneficiaries due to its existing
registry of beneficiaries and established cash card system for most of the beneficiaries.

Based on the results of the study, the following recommendation are put forward:

54
• DSWD and PIDS should study the need to establish a principle for adjusting the grant
amount provided by the program ahead of the six-year schedule of reviewing the benefit
level stipulated in the 4Ps Act/RA11031. This is also important given that the country
will enter the recovery phase post-COVID. An an ex-ante analysis should precede the
increase in amount of grants.

• In addition to automatic adjustment of the amount grants, the program may consider
differentiating the amount of the assistance for specific target beneficiaries (e.g., by
gender, location) such as the recommendation of Reyes and Tabuga (2013), David,
Albert and Vizmanos (2018), and Paqueo and Orbeta (2019).

• If grant amounts cannot be adjusted proactively, supplementary interventions (other

programs or other cash assistance) should be pursued.

• More than increasing the frequency of payouts, reliability and predictability of payment
schedules appear to be more important. This can be done by ensuring payouts are
conducted according to an explicit declared schedule and beneficiaries have reduced
barriers to access the grants. To achieve this, improvements should be made in the in
processes and IT infrastructures of DSWD and LBP.

• Changes in frequency of payment should be carefully examined to know if benefits

outweigh additional costs. This can be piloted in a small area before full
implementation.

• LBP (or the relevant AGDB) should expand network of ATMs and local bank branches
in the country to reach all areas. LBP should also find alternative points of cash
withdrawal such as POS to cover areas without ATMs. POS establishments should be
monitored, and transaction fees should not be shouldered by beneficiaries (at least for
1st withdrawal) regardless of where the transaction is done.

• Processes of resolving payment and cash card related grievances (e.g., decentralized
process) should be streamlined and frontline staff should have access to real-time status.

55
6. Bibliography
Acosta, P., Avalos, J., & Velarde, R. (2019, September). Pantawid Pamilya 2017 Assessment:An
Update of the Philippine Conditional Cash Transfer’s Implementation Performance. WORLD
BANK SOCIAL PROTECTION POLICY NOTE.

Adriano, L. (2016). Pantawid Pamilyang Pilipino Program: Stimulus to Local Economic Growth.
DSWD.

Albert, J. (2015, January). Pantawid Extension to High School: Pathway for Accelerating
Improvements in Human Capital Among the Poor. Mandaluyong City: Asian Development
Bank.

Baird, S., McIntosh, C., & Ozler, B. (2009). Designing Cost-Effective Cash Transfer Programs to Boost
Schooling among Young Women in Sub-Saharan Africa. The World Bank.

Banerjee, A., Hanna, R., Kreindler, G., & Olken, B. (2017). Debunking the Stereotype of the Lazy
Welfare Recipient: Evidence from Cash Transfer Programs. The World Bank Research
Observer, 155-184.

Bastagli, F., Hagen-Zanker, J., Harman, L., Barca, V., Sturge, G., Schmidt, T., & Pellerano, L. (2016).
Cash Transfers: What Does the Evidence Say? A Rigorous Review of Programme Impact and
of the Role of Design and Implementation Features. London: Overseas Development
Institute.

Bazzi, S., Sumarto, S., & Suryahadi, A. (2012). Evaluating Indonesia’s unconditional cash transfer
programme, 2005–2006. 3ie.

Catubig, M., Villano, R., & Dollery, B. (2015). Payment Schemes in Conditional Cash Transfer
Programs: The Case of 4Ps in the Davao Region, Philippines. Administrative Sciences, 240–
259.

CGAP. (2013). Striving for E-payments at Scale: The Evolution of the Pantawid Pamilyang Pilipino
Program in the Philippines. CGAP/World Bank.

Dabalen, A., Kilic, T., & Wane, W. (2008). Social transfers, labour supply and poverty reduction: the
case of Albania. The World Bank.

Davis, B., Handa, S., Arranz, M. R., Stampini, M., & Winters, P. (2002). Conditionality and the impact
of program design on household welfare: Comparing two diverse cash transfer programs in
rural Mexico. Working Paper. Agricultural and Development Economics Division of the Food
and Agriculture Organization of the United Nations (FAO - ESA).

DSWD. (2014). Keeping Children Healthy and in School: Evaluating the Pantawid Pamilya Using
Regression Discontinuity Design.

DSWD and World Bank. (2014). Philippines conditional cash transfer program impact evaluation
2012. Washington DC: DSWD and World Bank.

56
Esteva, A. A. (2012, April). [Link]. Retrieved from AA Esteva:
[Link]
medium-term-effects-on-physical-and-cognitive-development-updated-april-2012/

Fernald, L., Gertler, P., & Neufeld, L. (2009). 10-year effect of Oportunidades, Mexico's conditional
cash transfer programme, on child growth, cognition, language, and behaviour: a
longitudinal follow-up study. The Lancet, 1997-2005.

Filmer, D., & Schady, N. (2011). Does more cash in conditional cash transfer programs always lead to
larger impacts on school attendance? Journal of Development Economics, 150–157.

Fiszbein, A., & Schady, N. (2009). Conditional Cash Transfers: Reducing Present and Future Poverty.
Washington DC: IBRD and WB.

Garcia, S., & Saavedra, J. (2012, February). Educational Impacts and Cost-Effectiveness of Conditional
Cash Transfer Programs in Developing Countries: A Meta-Analysis. RAND Working Paper
Series.

Gertler, P. (2004). Do Conditional Cash Transfers Improve Child Health? A Randomized Experiment.
The American Economic Review, 336-341. Retrieved October 6, 2014, from
[Link]

Haushofer, J., & Shapiro, J. (2016). The Short-term Impact of Unconditional Cash Transfer to the
Poor: Experimental Evidence from Kenya. The Quarterly Journal of Economics, 1973–2042.

Manley, J., Fernald, L., & Gertler, P. (2015). Wealthy, healthy and wise: does money compensate for
being born into difficult conditions? Apllied Economics Letters, 121-126.

Pantawid Pamilya NPMO. (July 12, 2019). Overview of the Pantawid Pamilyang Pilipino Program.
Orientation to 4Ps NAC Members and Partners. Quezon City.

57
7. Appendix: Differential impact of Pantawid Pamilya by mode of payment

Table 30. Growth monitoring

Outcomes Sharp RD Estimates

CER Optimal MSE Optimal Sample
Regular weight monitoring z-value 0.559 0.419 -0.372
for children 0 to <2 years old difference (impact) 8.580 8.579 -4.250
Cash impact 14.900 7.630 -7.130
card se 10.370 9.150 4.930
robust p-value 0.117 0.257 0.163
conventional p-
value 0.151 0.404 0.148
number of obs. 348 425 734
Pantawid 27.865 7.628 -7.130
non-Pantawid 12.965 24.632 19.467
bandwidth 356 470 3,008
OTC impact 6.320 11.230 -9.880
se 11.300 10.580 4.860
robust p-value 0.468 0.195 0.365
conventional p-
value 0.576 0.288 0.042
number of obs. 263 315 748
Pantawid 22.178 25.801 26.613
non-Pantawid 15.856 14.573 16.733
bandwidth 256 338 3,008
Frequency of weight z-value -0.024 -0.005 -0.216
monitoring for children 0 to 2 difference (impact) -0.021 -0.021 -0.256
years old in the past six Cash impact 0.292 0.113 -0.274
months card se 0.608 0.543 0.290
robust p-value 0.564 0.669 0.275
conventional p-
value 0.632 0.835 0.344
number of obs. 344 423 734
Pantawid 3.212 0.113 -0.274
non-Pantawid 2.920 3.125 2.991
bandwidth 350 461 3,008
OTC impact 0.313 0.354 -0.511
se 0.619 0.558 0.285
robust p-value 0.525 0.384 0.341
conventional p-
value 0.614 0.525 0.073
number of obs. 291 366 748
Pantawid 3.256 3.275 3.278
non-Pantawid 2.944 2.921 2.767
bandwidth 303 400 3,008
*** p<0.01, ** p<0.05, * p<0.10
Notes: Std. error presented is based on the conventional RD estimation while the p-value is from the robust version that corrects for bias
The conditional probability exceeds 100 due to the use of a linear probability model in the estimation

58
Table 31. Child health services and practices
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
Vitamin A supplementation (6 z-value -0.270 -0.015 -0.787
months to 6 years old) difference (impact) -1.640 -1.640 -3.650
Cash impact 4.870 5.770 5.600 *
card se 3.940 3.550 2.310
robust p-value 0.257 0.158 0.093
conventional p-value 0.217 0.104 0.015
number of obs. 1,172 1,423 2,460
Pantawid 84.446 84.742 5.611
non-Pantawid 79.579 78.972 84.086
bandwidth 361 479 3,008
OTC impact 6.510 5.890 9.600 **
se 4.610 4.170 2.430
robust p-value 0.217 0.263 0.012
conventional p-value 0.158 0.158 0.000
number of obs. 1,188 1,449 2,465
Pantawid 85.942 85.562 9.604
non-Pantawid 79.435 79.670 87.832
bandwidth 398 530 3,008
Full immunization at age 1 z-value -0.288 -0.062 -0.384
difference (impact) -2.240 -2.236 -5.440
Cash impact -0.840 -1.690 0.370
card se 4.720 4.210 3.110
robust p-value 0.944 0.844 0.339
conventional p-value 0.859 0.688 0.905
number of obs. 1,144 1,347 2,062
Pantawid 24.840 24.631 0.368
non-Pantawid 25.676 26.320 27.253
bandwidth 432 573 3,008
OTC impact 1.400 2.320 -0.610
se 6.190 5.610 3.500
robust p-value 0.786 0.636 0.750
conventional p-value 0.821 0.679 0.862
number of obs. 861 1,076 2,037
Pantawid 26.313 26.884 -0.603
non-Pantawid 24.912 24.561 25.434
bandwidth 332 441 3,008
Visited a health facility or z-value 2.286 ** 0.316 0.165
health professional in the past difference (impact) 18.740 18.738 4.570
8 weeks Cash impact 2.090 0.440 4.610
card se 4.750 4.360 2.790
robust p-value 0.765 0.940 0.221
conventional p-value 0.660 0.920 0.098
number of obs. 1,242 1,523 2,684
Pantawid 40.863 39.405 4.630
non-Pantawid 38.771 38.966 40.362
bandwidth 348 461 3,008
OTC impact -16.650 ** -10.520 * 7.630
se 6.680 5.990 3.230
robust p-value 0.012 0.061 0.201
conventional p-value 0.013 0.079 0.018
number of obs. 787 987 2,721
Pantawid 28.123 32.182 7.623
non-Pantawid 44.768 42.705 42.941
bandwidth 211 281 3,008
*** p<0.01, ** p<0.05, * p<0.10
Notes: Std. error presented is based on the conventional RD estimation while the p-value is from the robust version that corrects for bias
The conditional probability exceeds 100 due to the use of a linear probability model in the estimation
Table 32. Deworming
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
z-value -2.901 *** -0.359 -1.424

59
 difference (impact) -22.590 -22.598 -20.230
Cash impact -3.790 -2.950 -1.080
card se 4.840 4.410 3.280
robust p-value 0.441 0.503 0.443
conventional p-value 0.433 0.504 0.741
number of obs. 1,379 1,655 2,662
Pantawid 42.126 42.907 -1.084
non-Pantawid 45.919 45.856 45.692
Deworming (under 6 years
bandwidth 407 540 3,008
old)
OTC impact 18.800 *** 15.670 *** 9.380 **
se 6.100 5.370 3.220
robust p-value 0.002 0.004 0.019
conventional p-value 0.002 0.004 0.004
number of obs. 1,023 1,282 2,701
Pantawid 61.877 59.071 9.381
non-Pantawid 43.072 43.400 55.938
bandwidth 296 394 3,008
Deworming at least once (6 to z-value -2.090 ** -0.058 -1.202
14 years old) difference (impact) -7.240 -7.239 -9.900
Cash impact -3.390 -2.770 0.790
card se 2.590 2.310 1.350
robust p-value 0.223 0.280 0.292
conventional p-value 0.191 0.230 0.555
number of obs. 2,859 3,499 5,830
Pantawid 84.828 85.417 0.812
non-Pantawid 88.217 88.192 87.589
bandwidth 375 498 3,008
OTC impact 3.850 3.680 * 3.930 *
se 2.300 2.000 1.260
robust p-value 0.112 0.092 0.075
conventional p-value 0.094 0.066 0.002
number of obs. 2,868 3,539 5,801
Pantawid 90.892 90.825 3.928
non-Pantawid 87.043 87.145 91.028
bandwidth 426 569 3,008
Deworming at least twice (6 z-value -0.185 -0.034 -0.642
to 14 years old) difference (impact) -1.140 -1.140 -2.560
Cash impact 8.810 ** 9.000 ** 7.460 **
card se 3.810 3.460 2.320
robust p-value 0.024 0.013 0.019
conventional p-value 0.021 0.009 0.001
number of obs. 3,233 3,846 5,807
Pantawid 32.740 33.045 7.477
non-Pantawid 23.927 24.042 32.653
bandwidth 441 586 3,008
OTC impact 9.950 ** 11.040 *** 3.290
se 4.850 4.300 2.550
robust p-value 0.033 0.008 0.108
conventional p-value 0.040 0.010 0.197
number of obs. 2,510 3,096 5,777
Pantawid 33.240 33.820 3.288
non-Pantawid 23.287 22.780 30.163
bandwidth 355 474 3,008
*** p<0.01, ** p<0.05, * p<0.10
Notes: Std. error presented is based on the conventional RD estimation while the p-value is from the robust version that corrects for bias
The conditional probability exceeds 100 due to the use of a linear probability model in the estimation
Table 33. Child nutrition outcomes
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
Underweight z-value 0.224 0.056 0.048
difference (impact) 1.680 1.674 0.570
Cash impact 3.510 3.330 0.860
card se 4.570 4.190 2.660
robust p-value 0.435 0.413 0.457

60
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
conventional p-value 0.442 0.427 0.748
number of obs. 1,453 1,728 2,519
Pantawid 24.398 24.136 0.859
non-Pantawid 20.890 20.802 21.767
bandwidth 481 638 3,008
OTC impact 1.830 2.330 3.170
se 5.940 5.160 2.790
robust p-value 0.857 0.838 0.415
conventional p-value 0.758 0.652 0.255
number of obs. 1,077 1,342 2,539
Pantawid 23.453 24.137 3.170
non-Pantawid 21.618 21.809 24.128
bandwidth 332 443 3,008
Severe underweight z-value -0.059 -0.033 0.013
difference (impact) -0.250 -0.256 0.080
Cash impact 1.870 1.810 0.500
card se 2.270 2.040 1.260
robust p-value 0.453 0.433 0.530
conventional p-value 0.411 0.373 0.693
number of obs. 1,197 1,446 2,519
Pantawid 7.229 6.992 0.500
non-Pantawid 5.362 5.177 5.380
bandwidth 359 476 3,008
OTC impact 2.120 2.230 2.490
se 3.540 3.000 1.650
robust p-value 0.643 0.636 0.151
conventional p-value 0.549 0.457 0.132
number of obs. 1,375 1,657 2,539
Pantawid 7.848 7.777 2.487
non-Pantawid 5.724 5.545 7.124
bandwidth 464 618 3,008
Stunting z-value 1.192 0.210 0.054
difference (impact) 9.540 9.546 0.790
Cash impact 5.590 4.510 1.550
card se 4.510 4.140 2.810
robust p-value 0.225 0.279 0.311
conventional p-value 0.215 0.277 0.581
number of obs. 1,217 1,464 2,454
Pantawid 35.714 34.795 1.568
non-Pantawid 30.121 30.289 32.025
bandwidth 381 505 3,008
OTC impact -3.950 0.960 6.620 **
se 6.610 5.840 3.180
robust p-value 0.458 0.858 0.043
conventional p-value 0.550 0.869 0.037
number of obs. 854 1,079 2,506
Pantawid 30.092 33.856 6.617
non-Pantawid 34.045 32.893 37.308
bandwidth 256 341 3,008

Severe stunting z-value 0.232 0.096 0.043

difference (impact) 1.190 1.184 0.300
Cash impact 3.540 2.690 0.640
card se 2.970 2.770 1.740
robust p-value 0.247 0.345 0.512
conventional p-value 0.234 0.331 0.714
number of obs. 1,347 1,588 2,454
Pantawid 11.565 11.068 0.642
non-Pantawid 8.028 8.381 10.377
bandwidth 432 573 3,008
OTC impact 2.350 3.240 4.110 **
se 4.180 3.730 2.000

61
 Outcomes Bandwidth
CER Optimal MSE Optimal Sample
robust p-value 0.688 0.580 0.029
conventional p-value 0.573 0.384 0.040
number of obs. 1,045 1,303 2,506
Pantawid 11.790 12.764 4.097
non-Pantawid 9.437 9.520 14.052
bandwidth 329 438 3,008
Wasting z-value -0.416 -0.126 -0.055
difference (impact) -2.290 -2.294 -0.700
Cash impact -2.510 -1.040 0.920
card se 3.860 3.540 2.030
robust p-value 0.464 0.646 0.613
conventional p-value 0.515 0.770 0.650
number of obs. 1,074 1,302 2,207
Pantawid 10.835 11.359 0.928
non-Pantawid 13.348 12.397 11.217
bandwidth 370 491 3,008
OTC impact -0.220 0.040 1.370
se 3.930 3.500 2.040
robust p-value 0.912 0.910 0.957
conventional p-value 0.956 0.990 0.502
number of obs. 949 1,180 2,209
Pantawid 12.211 12.006 1.367
non-Pantawid 12.430 11.961 11.667
bandwidth 332 442 3,008
Severe wasting z-value -0.205 -0.088 0.075
difference (impact) -0.630 -0.635 0.760
Cash impact -2.910 -2.440 -1.210
card se 1.960 1.800 1.060
robust p-value 0.126 0.147 0.354
conventional p-value 0.138 0.174 0.257
number of obs. 1,115 1,331 2,207
Pantawid 2.240 2.310 -1.209
non-Pantawid 5.154 4.755 2.514
bandwidth 391 518 3,008
OTC impact -2.280 -2.100 -0.800
se 2.370 2.160 1.130
robust p-value 0.332 0.319 0.277
conventional p-value 0.336 0.331 0.480
number of obs. 1,038 1,268 2,209
Pantawid 2.782 2.591 -0.800
non-Pantawid 5.061 4.689 2.796
bandwidth 372 496 3,008
*** p<0.01, ** p<0.05, * p<0.10
Notes: Std. error presented is based on the conventional RD estimation while the p-value is from the robust version that corrects for bias
The conditional probability exceeds 100 due to the use of a linear probability model in the estimation

Table 34. Enrollment

Outcomes Bandwidth
CER Optimal MSE Optimal Sample
Enrollment of children 3 to 5 z-value 0.060 0.008 -0.082
years old difference (impact) 0.640 0.638 -1.910
Cash impact -2.210 -0.110 1.480
card se 7.160 6.720 4.290
robust p-value 0.780 0.991 0.875
conventional p-value 0.757 0.987 0.730
number of obs. 746 885 1,441
Pantawid 55.142 55.344 1.458
non-Pantawid 57.356 55.453 54.605
bandwidth 406 539 3,008
OTC impact -2.850 -0.650 3.260

62
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
se 7.880 7.210 4.130
robust p-value 0.699 0.867 0.509
conventional p-value 0.717 0.928 0.430
number of obs. 738 889 1,475
Pantawid 55.209 55.775 3.257
non-Pantawid 58.060 56.426 57.220
bandwidth 419 557 3,008
Enrollment of children 6 to 11 z-value 0.532 0.004 1.309
years old difference (impact) 0.550 0.555 0.650
Cash impact 1.800 ** 1.870 ** 0.510 **
card se 0.780 0.750 0.580
robust p-value 0.020 0.011 0.018
conventional p-value 0.021 0.012 0.374
number of obs. 2,312 2,733 3,959
Pantawid 99.422 99.376 0.517
non-Pantawid 97.620 97.505 98.502
bandwidth 474 629 3,008
OTC impact 1.250 * 1.380 * 0.590 **
se 0.680 0.700 0.530
robust p-value 0.076 0.062 0.030
conventional p-value 0.068 0.050 0.269
number of obs. 1,702 2,095 3,938
Pantawid 99.181 99.196 0.594
non-Pantawid 97.935 97.815 98.692
bandwidth 358 478 3,008
Enrollment of children 12 to 15 z-value 0.957 0.024 0.794
years old difference (impact) 3.210 3.207 4.460
Cash impact 2.800 3.980 3.400 ***
card se 2.180 2.000 1.260
robust p-value 0.279 0.120 0.001
conventional p-value 0.198 0.047 0.007
number of obs. 1,208 1,464 2,551
Pantawid 96.539 96.944 3.401
non-Pantawid 93.738 92.959 96.459
bandwidth 354 469 3,008
OTC impact -0.410 0.130 1.860
se 2.550 2.360 1.590
robust p-value 0.801 0.895 0.408
conventional p-value 0.873 0.958 0.243
number of obs. 1,445 1,771 2,534
Pantawid 93.057 93.400 1.863
non-Pantawid 93.463 93.275 95.053
bandwidth 508 677 3,008

Enrollment of children 16 to 17 z-value 1.062 0.086 1.809 *

years old difference (impact) 9.420 9.418 8.830
Cash impact 17.350 *** 14.920 ** 10.130 ***
card se 6.290 5.690 3.320
robust p-value 0.010 0.017 0.002
conventional p-value 0.006 0.009 0.002
number of obs. 584 698 1,157
Pantawid 93.467 92.091 10.130
non-Pantawid 76.120 77.173 89.052
bandwidth 352 466 3,008
OTC impact 7.930 9.050 11.050 *
se 6.250 5.450 3.230
robust p-value 0.259 0.178 0.068
conventional p-value 0.204 0.097 0.001
number of obs. 689 799 1,134
Pantawid 87.112 88.112 11.045
non-Pantawid 79.183 79.061 90.787
bandwidth 529 702 3,008
z-value 1.276 0.037 1.655 *

63
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
difference (impact) 4.630 4.637 5.900
Cash impact 7.700 *** 7.550 *** 5.570 ***
card se 2.540 2.270 1.330
robust p-value 0.006 0.005 0.000
conventional p-value 0.002 0.001 0.000
number of obs. 1,762 2,132 3,708
Pantawid 95.622 95.369 5.568
non-Pantawid 87.918 87.816 94.120
Enrollment of children 12 to 17
bandwidth 345 458 3,008
years old
OTC impact 3.070 2.930 5.080 **
se 2.590 2.300 1.430
robust p-value 0.327 0.364 0.025
conventional p-value 0.237 0.204 0.000
number of obs. 1,878 2,310 3,668
Pantawid 91.681 91.576 5.078
non-Pantawid 88.613 88.650 93.903
bandwidth 429 572 3,008
Enrollment of children 6 to 14 z-value 1.241 0.010 1.165
years old difference (impact) 1.410 1.409 1.750
Cash impact 2.550 *** 2.720 *** 1.930 ***
card se 0.680 0.640 0.440
robust p-value 0.001 0.000 0.000
conventional p-value 0.000 0.000 0.000
number of obs. 2,682 3,321 5,924
Pantawid 99.225 99.160 1.928
non-Pantawid 96.672 96.438 98.420
bandwidth 335 445 3,008
OTC impact 1.140 1.240 1.390 ***
se 0.910 0.860 0.570
robust p-value 0.273 0.249 0.007
conventional p-value 0.207 0.150 0.014
number of obs. 2,533 3,146 5,889
Pantawid 97.969 97.889 1.395
non-Pantawid 96.825 96.652 98.036
bandwidth 354 473 3,008

Enrollment of children 15 to 20 z-value 1.068 0.064 0.445

years old difference (impact) 6.750 6.753 4.320
Cash impact 6.650 6.870 8.380 ***
card se 4.150 3.680 2.170
robust p-value 0.146 0.120 0.005
conventional p-value 0.109 0.062 0.000
number of obs. 1,437 1,729 2,790
Pantawid 81.364 81.134 8.365
non-Pantawid 74.710 74.264 81.247
bandwidth 391 518 3,008
OTC impact -0.100 1.050 4.760
se 4.770 4.130 2.500
robust p-value 0.901 0.987 0.457
conventional p-value 0.983 0.800 0.057
number of obs. 1,511 1,824 2,716
Pantawid 75.501 76.132 4.765
non-Pantawid 75.601 75.084 79.029
bandwidth 475 632 3,008
*** p<0.01, ** p<0.05, * p<0.10
Notes: Std. error presented is based on the conventional RD estimation while the p-value is from the robust version that corrects for bias
The conditional probability exceeds 100 due to the use of a linear probability model in the estimation

64
Table 35. Attendance rates (in %)

Outcomes Bandwidth
CER Optimal MSE Optimal Sample
Attendance rate of children 3 z-value 0.170 0.008 -0.078
to 5 years old difference (impact) 0.930 0.927 -0.450
Cash impact 4.060 4.090 -1.010
card se 3.750 3.370 2.240
robust p-value 0.224 0.151 0.957
conventional p-value 0.280 0.226 0.653
number of obs. 396 463 771
Pantawid 88.834 89.011 -1.008
non-Pantawid 84.779 84.925 87.380
bandwidth 414 546 3,008
OTC impact 3.130 1.970 -0.250
se 3.990 3.580 2.300
robust p-value 0.414 0.527 0.680
conventional p-value 0.433 0.583 0.914
number of obs. 451 543 795
Pantawid 88.642 88.347 -0.235
non-Pantawid 85.515 86.381 87.880
bandwidth 535 708 3,008
Attendance rate of children 6 z-value -0.320 -0.003 0.542
to 11 years old difference (impact) -0.370 -0.374 0.170
Cash impact 1.200 1.090 0.120
card se 0.870 0.760 0.390
robust p-value 0.140 0.104 0.738
conventional p-value 0.169 0.153 0.760
number of obs. 1,787 2,210 3,887
Pantawid 97.385 97.369 0.117
non-Pantawid 96.187 96.279 97.095
bandwidth 343 456 3,008
OTC impact 1.570 * 1.320 * 0.800 **
se 0.760 0.650 0.350
robust p-value 0.054 0.071 0.012
conventional p-value 0.038 0.043 0.022
number of obs. 1,744 2,151 3,871
Pantawid 97.834 97.765 0.799
non-Pantawid 96.262 96.449 97.688
bandwidth 381 509 3,008
Attendance rate of children 12 z-value 0.897 0.010 0.297
to 15 years old difference (impact) 1.330 1.331 0.940
Cash impact 0.720 0.630 0.910
card se 0.730 0.650 0.410
robust p-value 0.352 0.384 0.261
conventional p-value 0.322 0.336 0.025
number of obs. 1,271 1,536 2,416
Pantawid 97.464 97.498 0.911
non-Pantawid 96.741 96.869 97.973
bandwidth 415 551 3,008
OTC impact -0.610 -0.250 0.430
se 1.290 1.050 0.470
robust p-value 0.608 0.716 0.616
conventional p-value 0.638 0.812 0.368
number of obs. 1,076 1,340 2,388
Pantawid 96.310 96.700 0.426
non-Pantawid 96.918 96.950 97.585
bandwidth 378 503 3,008

Attendance rate of children 16 z-value -2.656 *** -0.032 -0.642

to 17 years old difference (impact) -4.450 -4.454 -3.200
Cash impact -2.170 -1.640 -0.040
card se 1.450 1.320 0.720

65
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
robust p-value 0.121 0.172 0.361
conventional p-value 0.134 0.215 0.953
number of obs. 473 570 964
Pantawid 96.365 96.704 -0.043
non-Pantawid 98.535 98.343 97.773
bandwidth 346 458 3,008
OTC impact 2.280 *** 1.410 * 0.940
se 0.840 0.840 0.770
robust p-value 0.009 0.080 0.451
conventional p-value 0.007 0.093 0.222
number of obs. 407 498 955
Pantawid 99.736 99.044 0.945
non-Pantawid 97.452 97.638 98.555
bandwidth 334 442 3,008
Attendance rate of children 12 z-value 0.152 0.001 0.035
to 17 years old difference (impact) 0.180 0.184 0.090
Cash impact 0.000 0.090 0.640
card se 0.580 0.520 0.320
robust p-value 0.919 0.961 0.703
conventional p-value 0.998 0.866 0.042
number of obs. 1,870 2,244 3,380
Pantawid 97.235 97.356 0.645
non-Pantawid 97.233 97.268 97.920
bandwidth 437 580 3,008
OTC impact -0.180 -0.040 0.550
se 1.030 0.850 0.390
robust p-value 0.819 0.862 0.455
conventional p-value 0.859 0.962 0.164
number of obs. 1,475 1,836 3,343
Pantawid 96.980 97.167 0.546
non-Pantawid 97.162 97.207 97.844
bandwidth 362 483 3,008
Attendance rate of children 6 z-value 0.165 0.001 0.840
to 14 years old difference (impact) 0.170 0.171 0.590
Cash impact 1.030 0.940 0.330
card se 0.760 0.670 0.340
robust p-value 0.151 0.122 0.611
conventional p-value 0.177 0.165 0.326
number of obs. 2,559 3,140 5,767
Pantawid 97.318 97.326 0.333
non-Pantawid 96.287 96.390 97.326
bandwidth 323 429 3,008
OTC impact 0.860 0.790 0.720 **
se 0.700 0.600 0.300
robust p-value 0.272 0.277 0.027
conventional p-value 0.221 0.187 0.016
number of obs. 2,377 2,959 5,729
Pantawid 97.257 97.333 0.720
non-Pantawid 96.397 96.544 97.686
bandwidth 342 456 3,008

Attendance rate of children 15 z-value -0.698 -0.005 -0.574

to 20 years old difference (impact) -0.750 -0.747 -1.010
Cash impact -0.140 0.110 0.550
card se 0.770 0.670 0.400
robust p-value 0.785 0.969 0.859
conventional p-value 0.856 0.865 0.164
number of obs. 1,223 1,448 2,120
Pantawid 97.643 97.821 0.554
non-Pantawid 97.783 97.706 98.234

66
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
bandwidth 459 609 3,008
OTC impact 0.610 0.710 0.450
se 0.750 0.700 0.480
robust p-value 0.459 0.372 0.340
conventional p-value 0.418 0.311 0.345
number of obs. 1,111 1,364 2,074
Pantawid 98.151 98.190 0.454
non-Pantawid 97.543 97.480 98.081
bandwidth 465 619 3,008
*** p<0.01, ** p<0.05, * p<0.10
Notes: Std. error presented is based on the conventional RD estimation while the p-value is from the robust version that corrects for bias
The conditional probability exceeds 100 due to the use of a linear probability model in the estimation

67
Table 36. Attendance of at least 85%

Outcomes Bandwidth
CER Optimal MSE Optimal Sample
Attendance of at least 85% z-value -0.149 -0.018 -0.906
among children 3 to 5 years difference (impact) -1.690 -1.694 -6.620
old Cash impact 8.510 8.100 -1.830
card se 7.820 6.950 4.630
robust p-value 0.225 0.167 0.873
conventional p-
value 0.277 0.244 0.692
number of obs. 419 489 771
Pantawid 75.458 75.436 -1.827
non-Pantawid 66.952 67.337 71.760
bandwidth 445 587 3,008
OTC impact 10.200 8.700 2.190
se 8.260 7.290 4.780
robust p-value 0.204 0.205 0.310
conventional p-
value 0.217 0.233 0.646
number of obs. 397 469 795
Pantawid 77.978 77.456 2.223
non-Pantawid 67.778 68.757 75.433
bandwidth 431 570 3,008
Attendance of at least 85% z-value 0.197 0.005 1.204
among children 6 to 11 years difference (impact) 0.660 0.661 2.970
old Cash impact 3.430 3.620 * 1.690
card se 2.260 2.060 1.130
robust p-value 0.120 0.071 0.315
conventional p-
value 0.128 0.079 0.134
number of obs. 1,681 2,051 3,890
Pantawid 95.074 95.087 1.679
non-Pantawid 91.639 91.470 94.615
bandwidth 313 416 3,008
OTC impact 2.770 2.650 2.150 *
se 2.470 2.140 1.140
robust p-value 0.317 0.310 0.065
conventional p-
value 0.261 0.215 0.060
number of obs. 1,935 2,366 3,875
Pantawid 94.664 94.753 2.150
non-Pantawid 91.890 92.106 95.035
bandwidth 437 584 3,008
Attendance of at least 85% z-value 0.803 0.028 0.281
among children 12 to 15 years difference (impact) 3.760 3.764 3.710
old Cash impact -0.670 -0.840 0.970
card se 2.620 2.320 1.500
robust p-value 0.776 0.687 0.778
conventional p-
value 0.798 0.717 0.519
number of obs. 1,398 1,673 2,417
Pantawid 93.130 93.258 0.966
non-Pantawid 93.801 94.101 94.773
bandwidth 477 633 3,008
OTC impact -4.430 -3.140 -0.140
se 3.880 3.170 1.550
robust p-value 0.251 0.301 0.572
conventional p-
value 0.253 0.322 0.927
number of obs. 1,039 1,293 2,390
Pantawid 89.788 91.045 -0.145
non-Pantawid 94.223 94.187 93.921
bandwidth 359 477 3,008

68
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
Attendance of at least 85% z-value -2.592 *** -0.094 -0.566
among children 16 to 17 difference (impact) -12.810 -12.802 -8.170
years old Cash impact -6.520 * -5.280 -1.070
card se 4.010 3.680 2.330
robust p-value 0.099 0.129 0.176
conventional p-
value 0.104 0.151 0.646
number of obs. 511 626 966
Pantawid 90.897 91.563 -1.071
non-Pantawid 97.412 96.846 94.088
bandwidth 402 531 3,008
OTC impact 6.290 ** 3.490 1.850
se 2.890 2.730 2.270
robust p-value 0.031 0.160 0.806
conventional p-
value 0.029 0.201 0.417
number of obs. 355 443 958
Pantawid 100.905 98.684 1.848
non-Pantawid 94.618 95.193 96.309
bandwidth 286 378 3,008
Attendance of at least 85% z-value 0.137 0.004 0.094
among children 12 to 17 difference (impact) 0.520 0.519 1.060
years old Cash impact -2.170 -2.020 0.400
card se 2.330 2.040 1.160
robust p-value 0.339 0.297 0.319
conventional p-
value 0.351 0.321 0.731
number of obs. 1,797 2,171 3,383
Pantawid 92.521 92.762 0.400
non-Pantawid 94.692 94.784 94.587
bandwidth 416 552 3,008
OTC impact -2.690 -2.320 0.380
se 2.990 2.460 1.140
robust p-value 0.359 0.327 0.648
conventional p-
value 0.368 0.345 0.738
number of obs. 1,405 1,747 3,348
Pantawid 91.920 92.399 0.380
non-Pantawid 94.610 94.721 94.577
bandwidth 338 451 3,008
Attendance of at least 85% z-value 0.573 0.012 0.835
among children 6 to 14 years difference (impact) 1.620 1.623 3.220
old Cash impact 2.120 2.010 1.380
card se 1.880 1.690 0.950
robust p-value 0.228 0.186 0.587
conventional p-
value 0.258 0.236 0.145
number of obs. 2,576 3,169 5,770
Pantawid 94.222 94.216 1.376
non-Pantawid 92.099 92.210 94.613
bandwidth 327 435 3,008
OTC impact 0.500 0.770 1.350
se 2.110 1.790 0.910
robust p-value 0.879 0.806 0.223
conventional p-
value 0.812 0.667 0.137
number of obs. 2,471 3,052 5,734
Pantawid 93.057 93.413 1.346
non-Pantawid 92.557 92.646 94.662
bandwidth 356 476 3,008

69
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
Attendance of at least 85% z-value -0.363 -0.010 -0.219
among children 15 to 20 difference (impact) -1.270 -1.271 -1.390
years old Cash impact -0.280 0.070 1.030
card se 2.550 2.260 1.370
robust p-value 0.935 0.957 0.780
conventional p-
value 0.912 0.974 0.452
number of obs. 1,063 1,296 2,125
Pantawid 94.407 94.610 1.029
non-Pantawid 94.688 94.537 95.569
bandwidth 377 501 3,008
OTC impact 0.990 1.280 1.270
se 2.400 2.190 1.410
robust p-value 0.733 0.658 0.508
conventional p-
value 0.680 0.558 0.370
number of obs. 1,118 1,370 2,080
Pantawid 95.397 95.507 1.268
non-Pantawid 94.406 94.224 95.502
bandwidth 466 620 3,008
*** p<0.01, ** p<0.05, * p<0.10
Notes: Std. error presented is based on the conventional RD estimation while the p-value is from the robust version that corrects for bias
The conditional probability exceeds 100 due to the use of a linear probability model in the estimation

70
Table 37. School level enrollment
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
Enrollment in daycare, nursery, z-value 0.225 0.055 -0.001
preschool/kindergarten of difference (impact) 2.970 2.977 -0.040
children 3 to 5 years old Cash impact 1.150 1.700 3.040
card se 8.740 8.040 5.700
robust p-value 0.941 0.923 0.665
conventional p-value 0.895 0.833 0.593
number of obs. 589 693 922
Pantawid 38.279 37.989 3.015
non-Pantawid 37.125 36.288 36.626
bandwidth 556 735 3,008
OTC impact -1.820 -0.030 2.300
se 9.880 9.000 5.270
robust p-value 0.847 0.964 0.525
conventional p-value 0.854 0.997 0.662
number of obs. 442 541 933
Pantawid 38.156 38.582 2.289
non-Pantawid 39.979 38.616 38.044
bandwidth 393 521 3,008
Enrollment in preschool or z-value -0.872 -0.227 -0.915
kindergarten children 5 years difference (impact) -16.830 -16.834 -29.850
old Cash impact -2.840 -4.480 -1.530
card se 12.630 11.620 7.170
robust p-value 0.878 0.803 0.584
conventional p-value 0.822 0.700 0.831
number of obs. 250 299 519
Pantawid 50.864 50.419 -1.515
non-Pantawid 53.707 54.898 55.517
bandwidth 384 504 3,008
OTC impact 13.990 14.860 1.940
se 14.590 13.230 7.580
robust p-value 0.359 0.304 0.671
conventional p-value 0.338 0.262 0.798
number of obs. 249 303 542
Pantawid 65.181 65.484 1.939
non-Pantawid 51.190 50.627 56.767
bandwidth 383 504 3,008
Enrollment in elementary of z-value 1.433 0.025 0.436
children 6 to 11 years old difference (impact) 3.360 3.363 2.160
Cash impact 2.100 2.310 0.370
card se 1.490 1.390 0.980
robust p-value 0.162 0.102 0.145
conventional p-value 0.160 0.097 0.705
number of obs. 1,785 2,211 3,959
Pantawid 97.376 97.197 0.372
non-Pantawid 95.276 94.892 95.378
bandwidth 333 442 3,008
OTC impact -1.260 -1.190 -1.390
se 1.810 1.620 1.080
robust p-value 0.523 0.527 0.423
conventional p-value 0.484 0.462 0.201
number of obs. 2,095 2,571 3,938
Pantawid 94.340 94.295 -1.380
non-Pantawid 95.604 95.486 94.143
bandwidth 478 639 3,008

Enrollment in junior high z-value 0.787 0.036 0.995

school of children 12 to 15 difference (impact) 3.880 3.881 3.850
years old Cash impact 9.030 *** 9.310 *** 6.510 ***
card se 3.150 2.770 1.780

71
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
robust p-value 0.007 0.003 0.000
conventional p-value 0.004 0.001 0.000
number of obs. 2,061 2,462 3,708
Pantawid 86.177 86.405 6.505
non-Pantawid 77.144 77.091 85.904
bandwidth 433 574 3,008
OTC impact 5.150 5.580 5.030 *
se 3.790 3.260 1.920
robust p-value 0.219 0.151 0.097
conventional p-value 0.174 0.088 0.009
number of obs. 1,808 2,219 3,668
Pantawid 82.608 83.013 5.027
non-Pantawid 77.455 77.435 84.722
bandwidth 407 542 3,008
Enrollment in senior high z-value 1.102 0.209 0.983
school of children 16 to 17 difference (impact) 13.460 13.462 17.610
years old Cash impact 15.310 ** 13.780 * 7.710 **
card se 7.540 6.680 4.230
robust p-value 0.049 0.050 0.040
conventional p-value 0.042 0.039 0.068
number of obs. 655 771 1,157
Pantawid 59.595 58.587 7.693
non-Pantawid 44.285 44.809 56.096
bandwidth 422 558 3,008
OTC impact 1.850 1.760 2.340
se 9.610 8.110 4.770
robust p-value 0.891 0.913 0.751
conventional p-value 0.848 0.829 0.624
number of obs. 600 720 1,134
Pantawid 48.473 48.624 2.335
non-Pantawid 46.625 46.868 51.763
bandwidth 429 569 3,008
*** p<0.01, ** p<0.05, * p<0.10
Notes: Std. error presented is based on the conventional RD estimation while the p-value is from the robust version that corrects for bias
The conditional probability exceeds 100 due to the use of a linear probability model in the estimation

72
Table 38. Dropout rate
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
Dropout rate among children 6 z-value -0.023 -0.011 -0.130
to 11 years old difference (impact) -0.020 -0.018 -0.400
Cash impact -0.780 -0.890 -0.530 *
card se 0.710 0.660 0.360
robust p-value 0.318 0.253 0.084
conventional p-value 0.277 0.179 0.142
number of obs. 1,926 2,335 3,875
Pantawid 0.354 -0.885 -0.530
non-Pantawid 1.130 0.266 0.385
bandwidth 380 504 3,008
OTC impact -0.760 -0.850 0.060
se 0.530 0.540 0.340
robust p-value 0.196 0.171 0.281
conventional p-value 0.155 0.112 0.866
number of obs. 1,744 2,146 3,868
Pantawid 0.461 0.384 0.056
non-Pantawid 1.218 1.237 0.835
bandwidth 382 511 3,008
Dropout rate among children z-value -0.241 -0.111 -0.191
12 to 15 years old difference (impact) -0.550 -0.550 -1.330
Cash impact -1.930 -2.140 -1.790 **
card se 1.650 1.460 0.850
robust p-value 0.307 0.241 0.019
conventional p-value 0.242 0.142 0.034
number of obs. 1,103 1,339 2,472
Pantawid 1.415 1.375 -1.793
non-Pantawid 3.342 3.516 1.469
bandwidth 325 431 3,008
OTC impact -1.380 -1.040 -0.810
se 1.580 1.490 1.030
robust p-value 0.370 0.434 0.448
conventional p-value 0.384 0.482 0.431
number of obs. 1,193 1,467 2,445
Pantawid 2.293 2.577 -0.813
non-Pantawid 3.670 3.621 2.347
bandwidth 410 545 3,008
Dropout rate among children z-value -0.603 -0.374 -0.385
16 to 17 years old difference (impact) -4.130 -4.128 -5.310
Cash impact -4.130 -3.310 -2.800
card se 4.020 3.640 2.460
robust p-value 0.394 0.533 0.223
conventional p-value 0.304 0.363 0.255
number of obs. 514 626 1,022
Pantawid 4.151 4.733 -2.808
non-Pantawid 8.279 8.047 5.503
bandwidth 372 491 3,008
OTC impact 0.000 1.090 -2.780
se 5.540 4.620 2.410
robust p-value 0.926 0.698 0.821
conventional p-value 1.000 0.814 0.247
number of obs. 500 609 1,005
Pantawid 7.295 7.903 -2.784
non-Pantawid 7.296 6.816 4.599
bandwidth 412 545 3,008

Dropout rate among children z-value -0.439 -0.168 -0.364

12 to 17 years old difference (impact) -1.130 -1.130 -2.690
Cash impact -2.540 -2.520 -2.070 **
card se 1.860 1.640 0.890

73
 Outcomes Bandwidth
CER Optimal MSE Optimal Sample
robust p-value 0.228 0.215 0.015
conventional p-value 0.172 0.124 0.020
number of obs. 1,552 1,890 3,494
Pantawid 2.173 2.303 -2.067
non-Pantawid 4.709 4.825 2.674
bandwidth 319 423 3,008
OTC impact -1.410 -0.600 -1.440
se 1.780 1.540 0.960
robust p-value 0.473 0.742 0.382
conventional p-value 0.430 0.696 0.133
number of obs. 1,546 1,926 3,450
Pantawid 3.374 3.973 -1.437
non-Pantawid 4.780 4.572 2.977
bandwidth 371 494 3,008
Dropout rate among children 6 z-value 0.427 0.119 -0.094
to 14 years old difference (impact) 0.330 0.328 -0.350
Cash impact -1.300 ** -1.440 *** -0.930 ***
card se 0.500 0.480 0.320
robust p-value 0.013 0.005 0.002
conventional p-value 0.009 0.003 0.004
number of obs. 2,486 3,045 5,793
Pantawid 0.540 0.476 -0.928
non-Pantawid 1.838 1.915 0.660
bandwidth 312 414 3,008
OTC impact -1.630 *** -1.320 ** -0.350
se 0.590 0.590 0.420
robust p-value 0.005 0.017 0.170
conventional p-value 0.006 0.025 0.400
number of obs. 2,179 2,740 5,766
Pantawid 0.428 0.709 -0.352
non-Pantawid 2.053 2.033 1.165
bandwidth 306 408 3,008
Dropout rate among children z-value -0.396 -0.189 -0.166
15 to 20 years old difference (impact) -2.020 -2.021 -1.860
Cash impact -2.120 -2.110 -1.950
card se 3.640 3.250 1.640
robust p-value 0.663 0.701 0.120
conventional p-value 0.560 0.516 0.235
number of obs. 999 1,209 2,204
Pantawid 5.644 5.813 -1.948
non-Pantawid 7.766 7.927 6.068
bandwidth 318 422 3,008
OTC impact -0.100 -0.360 -1.760
se 3.580 3.060 1.600
robust p-value 0.941 0.924 0.412
conventional p-value 0.978 0.907 0.270
number of obs. 1,052 1,306 2,141
Pantawid 7.278 7.225 -1.762
non-Pantawid 7.379 7.584 5.562
bandwidth 408 542 3,008
*** p<0.01, ** p<0.05, * p<0.10
Notes: Std. error presented is based on the conventional RD estimation while the p-value is from the robust version that corrects for bias
The conditional probability exceeds 100 due to the use of a linear probability model in the estimation

Table 39. Child labor

Outcomes Bandwidth
CER Optimal MSE Optimal Sample
At least 1 hour of work (with z-value 2.222 ** 0.854 0.691
or without pay) last month, difference (impact) 7.040 7.041 6.250
10-14 years old Cash impact 3.440 3.110 2.340
card se 2.290 2.130 1.390
robust p-value 0.163 0.191 0.100

74
 conventional p-value 0.133 0.144 0.091
number of obs. 1,737 2,071 3,195
Pantawid 8.824 8.360 2.337
non-Pantawid 5.388 5.252 6.880
bandwidth 422 560 3,008
OTC impact -3.600 * -4.270 ** -0.820
se 2.190 1.800 1.210
robust p-value 0.081 0.012 0.284
conventional p-value 0.101 0.018 0.496
number of obs. 1,125 1,430 3,169
Pantawid 2.633 2.444 -0.823
non-Pantawid 6.237 6.715 4.192
bandwidth 278 370 3,008
At least 1 hour of paid work z-value 2.102 ** 0.856 0.737
last month, 10-14 years old difference (impact) 6.690 6.688 6.420
Cash impact 3.420 3.110 2.440 *
card se 2.250 2.090 1.380
robust p-value 0.165 0.197 0.078
conventional p-value 0.129 0.137 0.078
number of obs. 1,743 2,081 3,195
Pantawid 8.540 8.091 2.442
non-Pantawid 5.120 4.981 6.689
bandwidth 423 562 3,008
OTC impact -3.270 -3.910 ** -0.630
se 2.250 1.790 1.180
robust p-value 0.114 0.017 0.425
conventional p-value 0.146 0.029 0.593
number of obs. 1,095 1,407 3,169
Pantawid 2.636 2.458 -0.631
non-Pantawid 5.903 6.373 4.071
bandwidth 269 358 3,008
Number of days worked (with z-value 0.195 0.075 0.411
or without pay) last month, difference (impact) 0.604 0.604 3.689
10-14 years old Cash impact -1.390 -0.536 -0.232
card se 2.169 1.796 0.799
robust p-value 0.469 0.618 0.595
conventional p-value 0.522 0.765 0.771
number of obs. 69 78 147
Pantawid 4.564 4.938 -0.230
non-Pantawid 5.954 5.474 5.026
bandwidth 201 252 3,008
OTC impact -1.994 -2.073 0.186
se 2.213 1.960 0.840
robust p-value 0.352 0.281 0.794
conventional p-value 0.368 0.290 0.825
number of obs. 63 74 132
Pantawid 3.418 3.342 5.104
non-Pantawid 5.412 5.416 5.288
bandwidth 313 392 3,008
*** p<0.01, ** p<0.05, * p<0.10
Notes: Std. error presented is based on the conventional RD estimation while the p-value is from the robust version that corrects for bias
The conditional probability exceeds 100 due to the use of a linear probability model in the estimation
Table 40. Education expenditures
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
Expenditures on tuition z-value 0.531 0.015 0.230
and other fees (per difference (impact) 0.071 0.071 0.103
month) in the last Cash card impact -0.099 -0.060 0.081
school year se 0.101 0.086 0.049
robust p-value 0.315 0.427 0.955
conventional p-value 0.328 0.485 0.097
number of obs. 2,882 3,566 6,381
Pantawid 3.202 3.234 0.081

75
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
non-Pantawid 3.300 3.294 3.353
bandwidth 334 444 3,008
OTC impact -0.170 * -0.159 * -0.157 **
se 0.088 0.078 0.049
robust p-value 0.068 0.065 0.018
conventional p-value 0.054 0.043 0.001
number of obs. 2,910 3,629 6,415
Pantawid 3.080 3.092 3.255
non-Pantawid 3.250 3.251 3.099
bandwidth 383 511 3,008
Expenditures on school z-value -1.097 -0.018 -0.901
materials and supplies difference (impact) -0.089 -0.089 -0.068
(per month) in the last Cash card impact 0.042 0.043 0.082
school year se 0.055 0.051 0.030
robust p-value 0.504 0.508 0.217
conventional p-value 0.448 0.398 0.006
number of obs. 3,848 4,681 7,805
Pantawid 3.462 3.472 0.083
non-Pantawid 3.420 3.430 3.522
bandwidth 379 504 3,008
OTC impact 0.131 ** 0.114 ** 0.054 *
se 0.060 0.053 0.031
robust p-value 0.031 0.034 0.051
conventional p-value 0.028 0.030 0.081
number of obs. 3,719 4,619 7,751
Pantawid 3.523 3.517 3.443
non-Pantawid 3.392 3.403 3.497
bandwidth 412 551 3,008
Expenditures on school z-value -1.363 -0.019 -0.627
uniform (per month) in difference (impact) -0.100 -0.100 -0.061
the last school year Cash card impact -0.002 0.018 0.064
se 0.052 0.049 0.031
robust p-value 0.911 0.855 0.244
conventional p-value 0.976 0.714 0.042
number of obs. 2,693 3,296 5,957
Pantawid 3.823 3.837 0.064
non-Pantawid 3.824 3.820 3.872
bandwidth 329 438 3,008
OTC impact 0.099 * 0.082 * 0.054 *
se 0.052 0.045 0.029
robust p-value 0.064 0.079 0.055
conventional p-value 0.056 0.068 0.068
number of obs. 3,073 3,794 6,097
Pantawid 3.891 3.882 3.798
non-Pantawid 3.793 3.799 3.851
bandwidth 439 586 3,008

Expenditures on school z-value 1.293 0.023 0.548

allowance (per month) difference (impact) 0.175 0.175 0.134
last school year Cash card impact 0.121 0.139 0.099 **
se 0.092 0.082 0.052
robust p-value 0.228 0.149 0.028
conventional p-value 0.186 0.090 0.055
number of obs. 4,777 5,659 7,856
Pantawid 5.567 5.572 0.099
non-Pantawid 5.445 5.433 5.547
bandwidth 517 686 3,008
OTC impact -0.054 -0.054 -0.099
se 0.100 0.087 0.052
robust p-value 0.686 0.702 0.363
conventional p-value 0.591 0.531 0.057
number of obs. 3,933 4,840 7,796

76
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
Pantawid 5.377 5.365 5.406
non-Pantawid 5.431 5.419 5.308
bandwidth 437 584 3,008
Total school z-value 0.908 0.011 0.662
expenditures (per difference (impact) 0.089 0.089 0.076
month) last school year Cash card impact 0.094 0.100 0.118 ***
se 0.081 0.071 0.036
robust p-value 0.338 0.323 0.003
conventional p-value 0.241 0.157 0.001
number of obs. 3,806 4,650 8,062
Pantawid 6.011 6.016 0.118
non-Pantawid 5.916 5.916 6.029
bandwidth 351 467 3,008
OTC impact 0.006 0.019 -0.016
se 0.055 0.048 0.030
robust p-value 0.862 0.626 0.762
conventional p-value 0.916 0.683 0.609
number of obs. 4,659 5,665 8,000
Pantawid 5.898 5.902 5.887
non-Pantawid 5.892 5.883 5.872
bandwidth 532 712 3,008
*** p<0.01, ** p<0.05, * p<0.10
Notes: : Log-transformed values used in the estimation. Std. error presented is based on the conventional RD estimation while the p-value is
from the robust version that corrects for bias
The conditional probability exceeds 100 due to the use of a linear probability model in the estimation

77
Table 41. Household expenditures: Share to total expenditures
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
Share of food to total z-value -2.407 ** -0.054 -1.132
expenditures difference (impact) -4.830 -4.836 -5.990
Cash impact -1.880 -1.890 -1.640
card se 1.520 1.370 0.910
robust p-value 0.209 0.160 0.175
conventional p-value 0.215 0.167 0.069
number of obs. 1,723 2,091 3,749
Pantawid 62.073 62.021 -1.647
non-Pantawid 63.955 63.912 61.907
bandwidth 353 469 3,092
OTC impact 2.950 ** 2.880 ** 2.470 **
se 1.310 1.180 0.950
robust p-value 0.031 0.025 0.015
conventional p-value 0.024 0.014 0.009
number of obs. 2,191 2,673 3,793
Pantawid 66.669 66.495 2.476
non-Pantawid 63.715 63.619 66.135
bandwidth 537 717 3,092
Share of non-food to total z-value 2.407 ** 0.095 1.057
expenditures difference (impact) 4.830 4.836 5.990
Cash impact 1.880 1.890 1.640
card se 1.520 1.370 0.910
robust p-value 0.209 0.160 0.175
conventional p-value 0.215 0.167 0.069
number of obs. 1,723 2,091 3,749
Pantawid 37.927 37.979 1.647
non-Pantawid 36.045 36.088 38.093
bandwidth 353 469 3,092
OTC impact -2.950 ** -2.880 ** -2.470 **
se 1.310 1.180 0.950
robust p-value 0.031 0.025 0.015
conventional p-value 0.024 0.014 0.009
number of obs. 2,191 2,673 3,793
Pantawid 33.331 33.505 -2.476
non-Pantawid 36.285 36.381 33.865
bandwidth 537 717 3,092
Share of education to total z-value -0.244 -0.027 -0.305
expenditures difference (impact) -0.100 -0.093 -0.270
Cash impact -0.160 0.010 0.280
card se 0.290 0.250 0.150
robust p-value 0.527 0.827 0.176
conventional p-value 0.593 0.963 0.064
number of obs. 1,663 2,043 3,749
Pantawid 2.406 2.475 0.276
non-Pantawid 2.562 2.463 2.602
bandwidth 338 450 3,092
OTC impact -0.060 0.010 0.210
se 0.290 0.260 0.170
robust p-value 0.811 0.958 0.734
conventional p-value 0.831 0.973 0.211
number of obs. 2,019 2,490 3,793
Pantawid 2.349 2.389 0.208
non-Pantawid 2.412 2.380 2.515
bandwidth 481 643 3,092

Share of clothing and z-value -1.589 -0.225 -1.773 *

footwear to total difference (impact) -0.340 -0.340 -0.270
expenditures Cash impact 0.080 0.130 0.220 *
card se 0.130 0.110 0.070

78
 Outcomes Bandwidth
CER Optimal MSE Optimal Sample
robust p-value 0.587 0.380 0.053
conventional p-value 0.509 0.246 0.001
number of obs. 1,722 2,091 3,749
Pantawid 1.189 1.241 0.219
non-Pantawid 1.106 1.108 1.379
bandwidth 353 469 3,092
OTC impact 0.420 ** 0.370 ** 0.270 ***
se 0.170 0.150 0.080
robust p-value 0.020 0.021 0.002
conventional p-value 0.016 0.015 0.001
number of obs. 1,653 2,047 3,793
Pantawid 1.452 1.433 0.268
non-Pantawid 1.029 1.061 1.443
bandwidth 367 490 3,092
Share of health to total z-value -0.416 -0.126 -0.413
expenditures difference (impact) -0.150 -0.156 -0.120
Cash impact 0.210 0.170 0.110
card se 0.200 0.170 0.120
robust p-value 0.305 0.314 0.693
conventional p-value 0.303 0.322 0.342
number of obs. 1,858 2,269 3,749
Pantawid 1.111 1.105 0.114
non-Pantawid 0.903 0.932 1.041
bandwidth 400 531 3,092
OTC impact 0.360 0.290 0.000
se 0.300 0.260 0.130
robust p-value 0.211 0.222 0.881
conventional p-value 0.225 0.257 0.991
number of obs. 1,745 2,169 3,793
Pantawid 1.219 1.173 -0.002
non-Pantawid 0.855 0.880 0.919
bandwidth 396 529 3,092
Share of alcohol and z-value 0.839 0.175 0.469
tobacco to total difference (impact) 0.380 0.385 0.350
expenditures Cash impact 0.300 0.250 -0.120
card se 0.330 0.280 0.140
robust p-value 0.334 0.326 0.840
conventional p-value 0.357 0.370 0.392
number of obs. 1,553 1,903 3,749
Pantawid 1.824 1.763 -0.117
non-Pantawid 1.521 1.508 1.493
bandwidth 309 410 3,092
OTC impact -0.080 -0.090 -0.220
se 0.310 0.270 0.140
robust p-value 0.873 0.896 0.199
conventional p-value 0.787 0.741 0.109
number of obs. 1,661 2,068 3,793
Pantawid 1.516 1.503 -0.221
non-Pantawid 1.598 1.592 1.416
bandwidth 371 495 3,092
*** p<0.01, ** p<0.05, * p<0.10
Notes: Std. error presented is based on the conventional RD estimation while the p-value is from the robust version that corrects for bias
Table 42. Household expenditures
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
Average total per capita z-value 1.324 0.006 1.053
expenditure difference (impact) 0.093 0.093 0.122
Cash card impact 0.082 * 0.076 * 0.062 **
se 0.047 0.043 0.028

79
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
robust p-value 0.082 0.077 0.048
conventional p-value 0.081 0.077 0.025
number of obs. 1,885 2,290 3,749
Pantawid 10.331 10.331 0.062
non-Pantawid 10.250 10.255 10.326
bandwidth 405 539 3,092
OTC impact -0.012 -0.007 0.016
se 0.053 0.046 0.027
robust p-value 0.827 0.882 0.967
conventional p-value 0.824 0.886 0.553
number of obs. 2,069 2,540 3,793
Pantawid 10.254 10.262 10.271
non-Pantawid 10.266 10.269 10.287
bandwidth 496 662 3,092
Average total per capita z-value 0.225 0.001 0.414
food expenditure difference (impact) 0.017 0.017 0.032
Cash card impact 0.059 0.058 0.043
se 0.051 0.047 0.033
robust p-value 0.242 0.219 0.176
conventional p-value 0.243 0.222 0.193
number of obs. 1,890 2,307 3,858
Pantawid 9.837 9.838 0.042
non-Pantawid 9.778 9.780 9.829
bandwidth 394 524 3,092
OTC impact 0.042 0.048 0.060
se 0.056 0.049 0.029
robust p-value 0.494 0.405 0.197
conventional p-value 0.451 0.323 0.041
number of obs. 2,100 2,588 3,898
Pantawid 9.834 9.840 9.796
non-Pantawid 9.792 9.792 9.856
bandwidth 491 656 3,092
Average total per capita z-value 3.151 *** 0.023 1.283
non-food expenditure difference (impact) 0.300 0.300 0.341
Cash card impact 0.173 *** 0.161 *** 0.111 ***
se 0.064 0.058 0.037
robust p-value 0.007 0.005 0.006
conventional p-value 0.007 0.006 0.003
number of obs. 1,982 2,408 4,471
Pantawid 9.346 9.343 0.111
non-Pantawid 9.172 9.182 9.321
bandwidth 331 440 3,092
OTC impact -0.127 * -0.113 -0.079 *
se 0.071 0.063 0.044
robust p-value 0.096 0.112 0.082
conventional p-value 0.073 0.071 0.068
number of obs. 2,210 2,738 4,493
Pantawid 9.076 9.096 9.215
non-Pantawid 9.203 9.209 9.136
bandwidth 427 571 3,092

Average per capita non- z-value 3.191 *** 0.023 1.333

food expenditure difference (impact) 0.298 0.298 0.347
(including other Cash card impact 0.174 *** 0.166 *** 0.116 ***
disbursements) se 0.063 0.057 0.037
robust p-value 0.006 0.004 0.004
conventional p-value 0.006 0.004 0.002
number of obs. 2,047 2,490 4,471
Pantawid 9.328 9.328 0.116
non-Pantawid 9.155 9.163 9.306
bandwidth 349 463 3,092
OTC impact -0.125 * -0.111 -0.079 *

80
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
se 0.069 0.061 0.043
robust p-value 0.094 0.112 0.075
conventional p-value 0.071 0.071 0.068
number of obs. 2,243 2,769 4,493
Pantawid 9.058 9.079 9.196
non-Pantawid 9.183 9.189 9.118
bandwidth 435 582 3,092
Average per capita z-value 0.693 0.067 0.404
expenditure on vice difference (impact) 0.299 0.299 0.351
goods (e.g. alcohol, Cash card impact 0.127 0.122 0.072
tobacco) se 0.313 0.281 0.171
robust p-value 0.712 0.709 0.757
conventional p-value 0.684 0.664 0.672
number of obs. 1,851 2,248 4,125
Pantawid 3.246 3.217 0.073
non-Pantawid 3.118 3.095 3.230
bandwidth 341 453 3,092
OTC impact -0.172 -0.096 -0.152
se 0.297 0.258 0.167
robust p-value 0.588 0.735 0.535
conventional p-value 0.563 0.709 0.362
number of obs. 2,338 2,863 4,167
Pantawid 2.983 3.028 3.223
non-Pantawid 3.154 3.125 3.071
bandwidth 519 694 3,092
Average per capita z-value 0.264 0.117 0.523
expenditure on inpatient difference (impact) 0.066 0.066 0.151
care Cash card impact 0.193 0.161 0.113
se 0.166 0.149 0.096
robust p-value 0.251 0.283 0.420
conventional p-value 0.246 0.281 0.240
number of obs. 2,277 2,752 4,529
Pantawid 0.603 0.587 0.113
non-Pantawid 0.410 0.426 0.528
bandwidth 406 540 3,092
OTC impact 0.126 0.084 -0.038
se 0.189 0.168 0.090
robust p-value 0.449 0.494 0.772
conventional p-value 0.503 0.619 0.673
number of obs. 1,926 2,394 4,545
Pantawid 0.521 0.495 0.416
non-Pantawid 0.394 0.411 0.378
bandwidth 353 472 3,092

Average per capita z-value -0.690 -0.156 -0.115

expenditure on difference (impact) -0.146 -0.146 -0.060
outpatient care Cash card impact -0.148 -0.094 0.071
se 0.148 0.133 0.083
robust p-value 0.274 0.362 0.496
conventional p-value 0.318 0.481 0.397
number of obs. 1,980 2,416 4,495
Pantawid 0.537 0.575 0.070
non-Pantawid 0.685 0.669 0.616
bandwidth 332 441 3,092
OTC impact -0.002 0.013 0.065
se 0.150 0.131 0.088
robust p-value 0.953 0.995 0.497
conventional p-value 0.987 0.918 0.461
number of obs. 2,641 3,190 4,514

81
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
Pantawid 0.630 0.632 0.558
non-Pantawid 0.632 0.618 0.623
bandwidth 543 725 3,092
Average per capita z-value 0.682 0.058 0.562
expenditure on medical difference (impact) 0.262 0.262 0.358
services and Cash card impact 0.307 0.288 0.467 **
commodities se 0.249 0.231 0.155
robust p-value 0.282 0.326 0.040
conventional p-value 0.216 0.212 0.003
number of obs. 2,114 2,567 4,618
Pantawid 3.535 3.523 0.467
non-Pantawid 3.228 3.235 3.591
bandwidth 350 465 3,092
OTC impact 0.045 0.007 0.010
se 0.294 0.257 0.163
robust p-value 0.885 0.974 0.953
conventional p-value 0.877 0.979 0.949
number of obs. 2,354 2,879 4,634
Pantawid 3.239 3.198 3.140
non-Pantawid 3.194 3.191 3.150
bandwidth 447 598 3,092
Average per capita z-value -0.229 -0.013 -0.247
expenditure on education difference (impact) -0.099 -0.099 -0.155
per school age child Cash card impact 0.106 0.232 0.519 ***
se 0.308 0.278 0.146
robust p-value 0.887 0.676 0.006
conventional p-value 0.731 0.403 0.000
number of obs. 1,780 2,182 4,565
Pantawid 5.710 2,182 0.519
non-Pantawid 5.604 5.778 5.983
bandwidth 282 375 3,092
OTC impact 0.205 0.209 0.268
se 0.302 0.272 0.175
robust p-value 0.529 0.497 0.563
conventional p-value 0.497 0.441 0.126
number of obs. 2,093 2,581 4,578
Pantawid 5.643 5.615 5.477
non-Pantawid 5.438 5.406 5.744
bandwidth 389 520 3,092

Average per capita z-value -2.123 ** -0.100 -1.151

expenditure on clothing difference (impact) -0.651 -0.651 -0.546
and footwear Cash card impact 0.097 0.161 0.407 ***
se 0.227 0.201 0.109
robust p-value 0.813 0.685 0.004
conventional p-value 0.671 0.425 0.000
number of obs. 1,777 2,189 4,485
Pantawid 4.769 4.855 0.407
non-Pantawid 4.672 4.694 5.154
bandwidth 287 381 3,092
OTC impact 0.748 *** 0.673 *** 0.470 ***
se 0.206 0.188 0.114
robust p-value 0.000 0.001 0.000
conventional p-value 0.000 0.000 0.000
number of obs. 2,003 2,480 4,505
Pantawid 5.305 5.277 4.791
non-Pantawid 4.558 4.605 5.261

82
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
bandwidth 373 498 3,092
*** p<0.01, ** p<0.05, * p<0.10
Notes: : Log-transformed values used in the estimation. Std. error presented is based on the conventional RD estimation while the p-value is
from the robust version that corrects for bias
The conditional probability exceeds 100 due to the use of a linear probability model in the estimation

83
Table 43. Income

Outcomes Bandwidth
CER Optimal MSE Optimal Sample
Per capita z-value 1.551 0.023 0.726
income difference (impact) 0.300 0.300 0.295
including Cash card impact 0.658 *** 0.673 *** 0.609 ***
grants se 0.124 0.114 0.087
robust p-value 0.000 0.000 0.000
conventional p-
value 0.000 0.000 0.000
number of obs. 1,936 2,355 4,505
Pantawid 9.789 9.776 0.609
non-Pantawid 9.131 9.103 9.706
bandwidth 320 425 3,092
OTC impact 0.358 ** 0.422 ** 0.532 ***
se 0.148 0.133 0.093
robust p-value 0.034 0.011 0.000
conventional p-
value 0.016 0.001 0.000
number of obs. 1,661 2,098 4,531
Pantawid 9.557 9.588 9.097
non-Pantawid 9.199 9.166 9.629
bandwidth 298 398 3,092
Per capita z-value 2.117 ** 0.049 1.088
income difference (impact) 0.627 0.627 0.644
without Cash card impact 0.445 ** 0.420 ** 0.064
grants se 0.179 0.166 0.126
robust p-value 0.017 0.017 0.302
conventional p-
value 0.013 0.012 0.612
number of obs. 1,685 2,088 4,491
Pantawid 9.383 9.330 0.064
non-Pantawid 8.938 8.910 9.062
bandwidth 269 357 3,092
OTC impact -0.182 -0.079 -0.093
se 0.236 0.208 0.137
robust p-value 0.431 0.635 0.850
conventional p-
value 0.439 0.703 0.496
number of obs. 2,142 2,668 4,514
Pantawid 8.851 8.900 9.002
non-Pantawid 9.033 8.979 8.908
bandwidth 412 550 3,092
Per capita z-value 3.199 *** 0.194 1.174
income difference (impact) 1.894 1.894 2.032
from salaries Cash card impact 1.216 *** 1.115 *** 0.434 ***
and wages se 0.389 0.351 0.254
robust p-value 0.003 0.003 0.007
conventional p-
value 0.002 0.002 0.087
number of obs. 1,907 2,321 4,622
Pantawid 8.026 7.879 0.434
non-Pantawid 6.809 6.764 7.178
bandwidth 304 404 3,092
OTC impact -0.678 -0.698 * -0.532
se 0.446 0.411 0.291
robust p-value 0.134 0.098 0.143
conventional p-
value 0.129 0.089 0.067
number of obs. 2,017 2,491 4,638
Pantawid 6.294 6.229 6.680
non-Pantawid 6.972 6.927 6.148
bandwidth 364 487 3,092

Per capita z-value 0.478 0.013 0.319

income difference (impact) 0.152 0.152 0.199
from Cash card impact 0.053 0.103 0.009
se 0.205 0.188 0.125

84
Outcomes Bandwidth
CER Optimal MSE Optimal Sample
entrepreneur robust p-value 0.790 0.594 0.782
ial activities conventional p-
value 0.797 0.582 0.944
number of obs. 1,021 1,206 1,958
Pantawid 8.147 8.174 0.009
non-Pantawid 8.094 8.071 8.173
bandwidth 459 608 3,092
OTC impact -0.099 -0.071 -0.194
se 0.243 0.219 0.118
robust p-value 0.768 0.887 0.309
conventional p-
value 0.683 0.747 0.099
number of obs. 895 1,120 2,081
Pantawid 7.997 8.005 8.175
non-Pantawid 8.096 8.075 7.981
bandwidth 381 507 3,092
Per capita z-value -0.050 -0.001 -0.056
income difference (impact) -0.016 -0.016 -0.044
from other
Cash card impact -0.150 -0.131 -0.215
receipts
(excluding se 0.185 0.167 0.121
grants) robust p-value 0.501 0.585 0.149
conventional p-
value 0.417 0.433 0.077
number of obs. 1,029 1,236 1,887
Pantawid 7.867 7.880 -0.215
non-Pantawid 8.017 8.011 7.852
bandwidth 495 657 3,092
OTC impact -0.134 -0.077 -0.046
se 0.264 0.228 0.135
robust p-value 0.637 0.761 0.532
conventional p-
value 0.612 0.736 0.736
number of obs. 933 1,154 1,911
Pantawid 7.917 7.970 8.092
non-Pantawid 8.051 8.047 8.046
bandwidth 456 608 3,092
*** p<0.01, ** p<0.05, * p<0.10
Notes: : Log-transformed values used in the estimation. Std. error presented is based on the conventional RD estimation while the p-
value is from the robust version that corrects for bias
The conditional probability exceeds 100 due to the use of a linear probability model in the estimation

85
Table 44. Hunger and self-rated poverty
Bandwidth
Outcomes
CER Optimal MSE Optimal Sample
Incidence of hunger z-value 0.130 0.023 0.081
difference (impact) 0.600 0.607 1.260
Cash impact -4.260 -3.470 -2.690
card se 2.880 2.630 1.980
robust p-value 0.154 0.205 0.119
conventional p-value 0.139 0.186 0.176
number of obs. 2,729 3,250 4,622
Pantawid 14.295 14.350 -2.691
non-Pantawid 18.552 17.822 13.913
bandwidth 517 687 3,092
OTC impact -4.860 -4.530 -2.860
se 3.610 3.240 2.130
robust p-value 0.189 0.175 0.136
conventional p-value 0.178 0.163 0.180
number of obs. 2,287 2,830 4,638
Pantawid 14.232 13.922 -2.861
non-Pantawid 19.096 18.448 13.577
bandwidth 431 576 3,092
Number of days experienced z-value 0.584 0.179 0.246
hunger in the past 3 months difference (impact) 0.272 0.272 0.394
Cash impact -0.321 -0.319 -0.235
card se 0.343 0.316 0.216
robust p-value 0.332 0.280 0.559
conventional p-value 0.349 0.313 0.276
number of obs. 2,244 2,703 4,613
Pantawid 0.710 0.702 -0.235
non-Pantawid 1.030 1.021 0.621
bandwidth 381 507 3,092
OTC impact -0.593 * -0.496 * -0.027
se 0.317 0.298 0.233
robust p-value 0.062 0.090 0.404
conventional p-value 0.062 0.096 0.909
number of obs. 2,243 2,786 4,630
Pantawid 0.524 0.595 0.826
non-Pantawid 1.117 1.090 0.800
bandwidth 422 565 3,092
Self-rated poverty status (Poor) z-value 0.384 0.079 0.041
difference (impact) 2.340 2.337 0.540
Cash impact -0.100 -0.460 0.990
card se 4.100 3.740 2.370
robust p-value 0.906 0.781 0.340
conventional p-value 0.980 0.901 0.676
number of obs. 1,714 2,131 4,609
Pantawid 20.408 20.042 0.991
non-Pantawid 20.512 20.506 20.672
bandwidth 266 354 3,092
OTC impact -2.440 0.420 4.080
se 4.500 3.810 2.200
robust p-value 0.534 0.891 0.164
conventional p-value 0.587 0.912 0.064
number of obs. 1,860 2,316 4,626
Pantawid 19.034 20.793 4.078
non-Pantawid 21.475 20.371 23.436
bandwidth 329 440 3,092

Self-rated poverty status (Not- z-value 3.681 *** 0.815 1.313

Poor) difference (impact) 14.670 14.671 15.090
Cash impact 8.180 *** 7.570 ** 3.640 **
card se 2.950 2.760 1.760
robust p-value 0.008 0.010 0.012

86
 Bandwidth
Outcomes
CER Optimal MSE Optimal Sample
conventional p-value 0.006 0.006 0.039
number of obs. 2,010 2,436 4,609
Pantawid 20.295 19.554 3.645
non-Pantawid 12.114 11.982 16.258
bandwidth 324 430 3,092
OTC impact -6.490 ** -5.840 ** -4.600 ***
se 2.680 2.300 1.610
robust p-value 0.022 0.022 0.002
conventional p-value 0.016 0.011 0.004
number of obs. 2,449 3,014 4,626
Pantawid 6.817 7.284 -4.595
non-Pantawid 13.308 13.123 7.948
bandwidth 475 635 3,092
*** p<0.01, ** p<0.05, * p<0.10
Notes: Std. error presented is based on the conventional RD estimation while the p-value is from the robust version that corrects for bias
The conditional probability exceeds 100 due to the use of a linear probability model in the estimation

87
Table 45. Employment
Bandwidth
Outcomes
CER Optimal MSE Optimal Sample
Labor force participation z-value 0.817 0.026 0.352
difference (impact) 2.120 2.122 2.340
Cash impact 0.580 0.940 0.850
card se 1.780 1.520 0.780
robust p-value 0.771 0.607 0.197
conventional p-value 0.744 0.534 0.273
number of obs. 7,980 9,564 15,308
Pantawid 58.540 58.549 0.846
non-Pantawid 57.957 57.607 58.456
bandwidth 410 545 3,008
OTC impact -1.540 -1.920 -0.650
se 1.890 1.640 0.710
robust p-value 0.349 0.174 0.688
conventional p-value 0.414 0.244 0.358
number of obs. 4,708 6,039 15,013
Pantawid 56.987 56.670 -0.656
non-Pantawid 58.526 58.586 56.984
bandwidth 235 314 3,008
Employment z-value -1.064 -0.017 -0.177
difference (impact) -2.190 -2.189 -1.110
Cash impact -2.270 -2.260 * 0.330
card se 1.450 1.280 0.710
robust p-value 0.106 0.064 0.651
conventional p-value 0.116 0.079 0.644
number of obs. 3,573 4,379 8,819
Pantawid 90.972 90.753 0.327
non-Pantawid 93.245 93.008 92.093
bandwidth 288 383 3,008
OTC impact -0.080 0.120 1.100 *
se 1.460 1.270 0.720
robust p-value 0.863 0.881 0.092
conventional p-value 0.954 0.927 0.128
number of obs. 3,528 4,382 8,590
Pantawid 92.932 92.708 1.101
non-Pantawid 93.017 92.592 92.917
bandwidth 323 432 3,008
Usual work hours per week in z-value 1.109 0.051 0.605
primary occupation difference (impact) 2.770 2.771 1.629
Cash impact 6.196 *** 5.158 *** 1.436 **
card se 1.511 1.396 0.804
robust p-value 0.000 0.000 0.014
conventional p-value 0.000 0.000 0.074
number of obs. 3,356 4,089 8,032
Pantawid 44.367 43.745 1.439
non-Pantawid 38.171 38.587 41.628
bandwidth 300 399 3,008
OTC impact 3.426 * 3.005 ** -0.909
se 1.989 1.718 0.940
robust p-value 0.065 0.047 0.397
conventional p-value 0.085 0.080 0.333
number of obs. 2,428 3,099 7,860
Pantawid 41.665 41.393 39.978
non-Pantawid 38.240 38.388 39.072
bandwidth 229 306 3,008

Other job or business besides z-value 0.867 0.277 0.109

primary occupation difference (impact) 2.430 2.428 0.430
Cash impact 2.400 2.100 1.080
card se 2.040 1.930 1.050

88
 Bandwidth
Outcomes
CER Optimal MSE Optimal Sample
robust p-value 0.212 0.224 0.552
conventional p-value 0.240 0.276 0.301
number of obs. 3,108 3,832 8,003
Pantawid 8.320 8.058 1.085
non-Pantawid 5.925 5.961 7.553
bandwidth 274 365 3,008
OTC impact -0.030 0.350 -0.890
se 1.920 1.730 0.980
robust p-value 0.990 0.862 0.598
conventional p-value 0.987 0.839 0.364
number of obs. 3,293 4,085 7,820
Pantawid 6.404 6.706 -0.889
non-Pantawid 6.436 6.354 6.167
bandwidth 335 449 3,008
Usual work hours per week in z-value 4.371 *** 0.704 0.808
other jobs difference (impact) 18.290 18.291 15.273
Cash impact 5.038 5.101 3.168
card se 3.111 2.893 2.185
robust p-value 0.157 0.157 0.102
conventional p-value 0.105 0.078 0.147
number of obs. 231 285 548
Pantawid 21.414 21.112 3.159
non-Pantawid 16.377 16.011 19.066
bandwidth 347 455 3,008
OTC impact -13.253 *** -11.641 *** 0.941
se 2.799 2.659 2.136
robust p-value 0.000 0.000 0.527
conventional p-value 0.000 0.000 0.660
number of obs. 160 193 550
Pantawid 6.912 8.226 16.844
non-Pantawid 20.165 19.867 17.784
bandwidth 235 307 3,008
Total usual work hours per z-value 2.073 ** 0.098 0.911
week difference (impact) 5.408 5.407 3.885
Cash impact 7.860 *** 6.934 *** 1.857 ***
card se 1.722 1.560 0.808
robust p-value 0.000 0.000 0.003
conventional p-value 0.000 0.000 0.022
number of obs. 2,765 3,431 8,033
Pantawid 46.749 46.078 1.859
non-Pantawid 38.889 39.144 43.040
bandwidth 232 308 3,008
OTC impact 2.453 2.305 -1.015
se 1.959 1.677 0.900
robust p-value 0.164 0.103 0.256
conventional p-value 0.211 0.169 0.259
number of obs. 2,421 3,092 7,861
Pantawid 41.976 41.921 41.137
non-Pantawid 39.523 39.615 40.124
bandwidth 228 305 3,008

Looking for additional work if z-value -1.972 ** -0.409 -0.546

employed difference (impact) -5.800 -5.800 -4.140
Cash impact -3.200 * -3.110 ** 0.930
card se 1.820 1.620 0.970
robust p-value 0.067 0.040 0.377
conventional p-value 0.079 0.054 0.340
number of obs. 3,656 4,440 8,135
Pantawid 7.276 7.196 0.926
non-Pantawid 10.475 10.304 9.374

89
 Bandwidth
Outcomes
CER Optimal MSE Optimal Sample
bandwidth 332 441 3,008
OTC impact 2.600 1.940 -0.030
se 2.310 1.990 0.930
robust p-value 0.268 0.335 0.627
conventional p-value 0.261 0.331 0.971
number of obs. 3,301 4,104 7,950
Pantawid 12.140 11.209 -0.034
non-Pantawid 9.540 9.271 8.464
bandwidth 330 442 3,008
Unemployed and looking for z-value 0.440 0.168 0.059
work difference (impact) 9.580 9.579 3.680
Cash impact -0.690 -3.560 3.550
card se 13.950 12.530 6.280
robust p-value 0.977 0.910 0.977
conventional p-value 0.960 0.776 0.572
number of obs. 338 406 721
Pantawid 37.823 34.390 3.586
non-Pantawid 38.517 37.949 35.046
bandwidth 333 438 3,008
OTC impact -10.270 -15.410 -0.800
se 16.740 14.680 7.110
robust p-value 0.462 0.230 0.744
conventional p-value 0.539 0.294 0.911
number of obs. 270 333 675
Pantawid 31.901 26.317 -0.779
non-Pantawid 42.174 41.724 30.530
bandwidth 296 388 3,008
*** p<0.01, ** p<0.05, * p<0.10
Notes: Std. error presented is based on the conventional RD estimation while the p-value is from the robust version that corrects for bias
The conditional probability exceeds 100 due to the use of a linear probability model in the estimation

90