1

Assessment of the Anti-Red Tape Act Implementation for the Ease of Doing Business Act: Lessons from a Mixed Methods Approach

Authors Czarina Medina-Guce1, Kidjie Saguin2, and Kathleen Jovellanos3 With Thinking Machines Data Science, represented by Kevin Go, Machine Learning Researcher4

Abstract In 2018, Republic Act 11032 or the Ease of Doing Business and Efficient Government Service Delivery Act was passed, which expands Republic Act 9485 or the Anti-Red Tape Act (ARTA) of 2007. The EODB Law takes off from the achievements of the 2007 ARTA as implemented by the Civil Service Commission through its ARTA Integrated Program, and further incorporates the lessons and momentum from the Ease of Doing Business initiatives of the Department of Trade and Industry, the National Competitiveness Council, and their partners. Against the backdrop of the preparation for transition and the finalization of the Implementing Rules and Regulations, an assessment was conducted on the efficiency, effectiveness, and relevance of the ARTA implementation, which utilized an integrated mixed methods approach: statistical analysis, case study method, and data science techniques. The conference paper shares the key findings of the assessment, including policy insights on leveraging on the ARTA standards that have become a normative reference for frontline government services, revisiting the anti-corruption goals of the anti-fixing campaign, and increasing efforts to account for capacity of the frontline offices to deliver according to standards. The discussion employs self-reflexivity, as to how the researchers used three different methods to address data robustness issues, national and subnational lenses, and triangulation towards a coherent assessment narrative. The discussion ends with a special focus on a data science approach using machine learning and visualization methods, with emphasis on understanding factors affecting customer satisfaction. Ways forward in using data science and machine learning in evaluation studies are also discussed. The conference paper is an off-shoot of the assessment as supported by the National Economic and Development Authority and United Nations Development Programme – Strategic M&E Project. ******

1 Overview

1 Ateneo de Manila University, mmedina@ateneo.edu 2 National University of Singapore, kidjie.saguin@gmail.com 3 Independent Evaluator, kathleenjovellanos@gmail.com 4 kevin@thinkingmachin.es

2

This paper is an off-shoot of the assessment of the Anti-Red Tape Act of 2007 (ARTA, or Republic Act 9485), conducted under the Strategic Monitoring and Evaluation (M&E) Project of the National Economic and Development Authority (NEDA) and the United Nations Development Programme (UNDP). The assessment project is largely shaped by the policy transition, when in 2018, Republic Act 11032 or the Ease of Doing Business (EODB) and Efficient Government Service Delivery Act (EODBA) was passed, which updates and expands the scope of the ARTA. The evaluation sought to unpack lessons in efficiency, effectiveness, and sustainability from the ARTA implementation, to inform policy and program decisions for EODBA. Taking off from the assessment (Medina-Guce, Saguin, Jovellanos, 2019), this conference paper shares our insights in three parts. First, we discuss key findings of the assessments, including overarching policy insights. Second, we discuss the statistical analysis component of the mixed methods deployed, with the findings and technical recommendations. And third, we share the data science method component, with its corresponding findings and recommendations on the RCS methodology.

1.1 Overview of the ARTA The 2007 ARTA’s Section 2, Declaration of Policy states:

“It is hereby declared the policy of the State to promote integrity, accountability, proper management of public affairs and public property as well as to establish effective practices aimed at the prevention of graft and corruption in government. Towards this end, the State shall maintain honesty and responsibility among its public officials and employees, and shall take appropriate measures to promote transparency in each agency with regard to the manner of transacting with the public, which shall encompass a program for the adoption of simplified procedures that will reduce red tape and expedite transactions in government.” (Highlights by authors)

As such, the policy positions itself as fundamentally a governance improvement initiative, by connecting efficiency in frontline government services as a means towards achieving integrity and accountability in government, and reducing corruption. ARTA’s governance provisions can be summarized in five points. First, ARTA provides a set of universal service standards that all clients can expect from all frontline services of the government, e.g., simple and complex transactions should be completed within five and ten days, respectively. Second, it mandates the implementation of various process improvement initiatives, such as process reengineering and signature reduction. These initiatives were legally mandated to ensure that efficiency is promoted by reducing excessive and unnecessary rules and regulations of these frontline services. Third, the law puts the publication of Citizen Charters (CCs) as its centerpiece provision (Mendoza, 2011). ARTA defines a CC as an official document, a service standard, or a pledge, that communicates, in simple terms, information on the services provided by the government to its clientele. Posting this information, which usually includes step-by-step procedures, time to complete the service, documentary requirements, among others, acts as the primary transparency mechanism of ARTA. Fourth, agencies are also expected to establish their own complaints and redress system as a means of empowering those accessing the services with recourse when the quality of service is less than expected. Lastly, the law puts criminal liability not only to operational “fixers” but also to those working in collusion with fixers, who may be public servants. As a means of promoting compliance with ARTA provisions, deviations from the published charters were made subject to criminal proceedings.

3

Owing to priorities allocated to specific provisions of the law, ARTA implementation took a more focused gaze into efficiency of frontline government service delivery. As the law and its IRR, CSC performed the primary role in operationalizing the ARTA implementation. To operationalize implementation of the ARTA, CSC launched the Integrated ARTA Program in 2010. The program has the following components:

● Report Card Survey (RCS), “a client feedback survey used to check government service offices’ compliance with the provisions of the Anti-Red Tape Act of 2007. It provides a quantitative measure of user perceptions on the quality, efficiency, and adequacy of public services. (CSC, 2014a)”

● Contact Center ng Bayan (CCB), “a feedback mechanism designated as the government’s main helpdesk where citizens can request for information and assistance on government frontline service procedures, and report commendations, appreciation, complaints, and feedback. (CSC, 2014b)”

● Service Delivery Excellence Program (SDEP), which is given to service offices obtaining a “Failed” rating from RCS and includes “conduct of service improvement workshops and facilitation for increasing the efficiency and effectiveness of the service office’s strategies, structure, staff, and systems. (CSC, 2014c)”

● Citizen’s Satisfaction Center Seal of Excellence (CSC-SE), “an award given to government service offices that have demonstrated exemplary frontline service delivery by meeting both conditions: (a) obtained an Excellent rating in the RCS, and (b) has no valid and unresolved ARTA-related complaint based on the Contact Center ng Bayan data. (CSC, 2014d)”

Through the implementation of the ARTA Integrated Program, CSC is able to gather information on agencies with most number of ARTA-related complaints (Table 1) and the types of complaints most frequently received (Figure 1). Table 1. Agencies with the Highest Number of ARTA-related Complaints (Agamata, 2018)

Figure 1. Top Five ARTA-related Complaints from 2013 to 2017, in Percentage of Total Reports Per Year (Data reference: Agamata, 2018)

4

The ARTA assessment takes on both the outcomes targeted by the policy declaration, and the CSC program for the scope of the evaluation. The authors deemed it compelling to take off from the information already available from the Integrated Program, but also expand the evaluation gaze to explore how the ARTA as a policy has contributed to its declared higher-level societal outcomes.

1.2 Overview of the Integrative Mixed Methods Deployed In conducting the assessment, the Strategic M&E Project put together several approaches to the evaluation, which comprised the integrative mixed methods approach deployed (Figure 2).

● Statistical analysis and data science methods were deployed on the RCS implemented by the CSC. Both methods unpacked the factors relating to client satisfaction of frontline government services, which was a variable included in the RCS.

● Case studies were also developed for four national government agencies (NGAs), and four local government units (LGUs) for the qualitative analysis component. The case studies were selected based on the results of the RCS, with inputs from CSC, Department of Trade and Industry (DTI) and NEDA -- the agencies that comprised the oversight group for the evaluation project. Data gathering through field visits, key informant interviews (KIIs), and focus group discussions (FGDs) were conducted from September 2018 to February 2019.

○ The case studies includes: Land Registration Authority (LRA), Bureau of Internal Revenue (BIR), Philippine Health Insurance Corporation (PHIC or PhilHealth), and Social Security System (SSS) for the NGAs; and Tagum City, Arayat Pampanga, Cainta Rizal, and Quezon City for the LGUs.

● Complementary interviews and focus group discussions were conducted with oversight agencies for insights on the effectiveness, efficiency, relevance, and sustainability of the ARTA policy implementation, especially in light with the transition to the EODBA.

5

Figure 2. Summary of Integrative Mixed Methods Approach

2 Summary of the Assessment Findings This second section briefly discusses the key findings from the assessment, which serves as a take-off point for the elaboration of the statistical analysis and data science method discussions in the succeeding sections. Interested readers may access the report (in working paper version) through http://bit.ly/ARTAAssessmentWorkingPaper. In summary, the assessment finds that:

1. Focusing on ARTA’s effect on government processes, there is substantiated evidence supporting how ARTA implementation (through the Integrated Program of the CSC) has been able to influence efficiency levels of frontline government services as evidenced by client satisfaction. Factors contributing to this success areas are leadership and management prioritization, agency-level initiatives, and technology investments; all of which are seen to be mutually reinforcing of each other. The relative successes of these initiatives, however, are still impeded when volume, density, and levels of complexity of transactions are high in frontline government offices. Any deviation from the prescribed processes, such as paying for hidden costs, tend to lower client satisfaction.

2. While there is overall improvement in efficiency of frontline services, not all ARTA program implementation components contribute to effectiveness-related outcomes. Those with evidence supporting effectiveness contribute to the frontline experience of quick response time and positive transaction outcomes, such as agency-level initiatives and leadership and management prioritization as earlier noted. However, an unintended consequence of the drive to meet ARTA standards is the stretch on the time and welfare of frontline government employees, as offices exhaust the limits of their manpower. Adding to the stretch are what employees describe as aggression from clients threatening to report them via the hotlines, and compromised time for lunch breaks. The frontline government employees hope that their own satisfaction, motivation, and welfare can be better accounted for in future ARTA/EODB implementation.

3. Meanwhile, evidence is weak and unsubstantiated on the effects of ARTA implementation to the integrity and anti-corruption outcomes that the law aspires to achieve. The Anti-Fixing

6

Campaign is found to have no correlation to client satisfaction from the RCS data. While technology investments (such as computerization of transactions) have been employed partly as an anti-fixing strategy, both big-time and small-time fixers are still prevalent in many frontline offices, with fixers themselves sharing how they have appropriated their strategies for conducting business despite the ARTA provisions. Data is available on administrative cases filed and resolved against erring government employees, as well as potentially important data on fixing from the RCS datasets, but all these data have not been consolidated and processed to generate insights to improve policy and program interventions. Inter-agency collaboration and prioritization for implementation is also a gap that is yet to be sufficiently addressed.

4. Furthermore, the evaluation also underscores the need to review the guidelines and indicators measuring the Citizen Charters (CCs), since the value of availability of information through CCs are compromised by the client perceptions of CCs being overwhelming, unreadable, and in effect, unhelpful. Preference of clients is still to interact with persons (e.g., PACD, guard-on-duty), to provide them the specific information they need. Hence the indicators for evaluating CCs in the RCS need to be reviewed to emphasize usability to clients, and appropriate support to innovations in presenting CCs are identified. On the process, guidance may be given to agencies how to make the development of CCs more inclusive of government employees and citizens alike.

5. The various evaluation tracks provide layered approaches into answering an improved implementation of ARTA in light of the EODB law. On short-term change management, communicating the transition plans and support to be given to frontline government offices and employees is seen to be important in light of growing concerns on the more stringent standards under EODB. In terms of a strategic and evaluative framework for EODB, the importance of articulating a Theory of Change and a results framework is again emphasized. These can be expressed in a Governance Roadmap that provides logic to the priorities of EODB implementation, and the improvements it seeks to achieve towards higher-level outcomes throughout the years.

3 Lessons from Statistical Analysis Specific to this Statistics Conference, this paper elaborates on the statistical analysis conducted for the ARTA assessment, as part of the integrative mixed methods approach. This section discusses the usefulness of statistics for impact evaluation, the challenges faced, and emerging recommendations for future ARTA/EODBA implementation.

3.1 Usefulness of statistics for impact evaluation The main purpose of this component of the study was to perform a “post-intervention analysis” or to understand whether ARTA, as a policy intervention, had any discernible effects on the quality of frontline services. More specifically, it sought to answer the following research questions: Does compliance to ARTA contribute to improving the quality of public service delivery? What specific components of ARTA increases citizen satisfaction? Does experience of ‘hidden costs’ influence citizen satisfaction? The analysis used only the comparative datasets generated during the 2014-2015 rounds of RCS. Although there were several rounds made available to the research team (2010-2016), 2014-2015

7

rounds were based on the same survey questionnaire, reducing the likely measurement error derived from the changes made to the questionnaire before and after these rounds. The 2014-2015 rounds also had similar scope with the following agencies having been surveyed for both years: HDMF, PHIC, SSS, LRA, BIR, and GSIS In contrast to the data science methods, the statistical analysis focused specifically on creating a relationship between how the agencies has implemented ARTA with client satisfaction. The focus is warranted because CSC placed a great amount of value on client-respondent satisfaction as it is the main outcome variables included in the RCS, on top of identifying hidden costs. The RCS also features a measurement of the compliance with ARTA provisions. A scoring mechanism is used to determine a numeric rating and its corresponding adjectival rating, which are then used as an input to knowing which agency has excelled or failed. The scores are derived from a survey of actual clients and evaluation of a CSC-employed RCS researcher. While the survey questionnaire asked respondents to provide answers relating to both satisfaction and ARTA compliance, the researcher serves as a mystery client and conducts an inspection or actual physical observation of the service office’s compliance to the requirements of ARTA and CSC’s integrated ARTA program. An OLS regression was used to examine the potential relationship between ARTA compliance (IC score) and client satisfaction (SQ score). Table X shows a positive and statistically significant association between improvements in ARTA implementation and client satisfaction. Service quality dimensions such as response time, attention and positive outcome, also matter in shaping satisfaction among clients.

Table 2. Effects of ARTA Compliance on SQ Earned Score, 2014-2015

Model 1 Model 2 Model 3

ARTA compliance 0.209 0.199 0.139

(0.020)** (0.019)** (0.014)**

2015 0.555 0.758

(0.249)* (0.150)**

Age 0.049 0.207

(0.031) (0.019)**

Age squared -0.001 -0.002

(0.000) (0.000)**

Sex (1=male, 0=female) -0.211 -0.003

(0.136) (0.078)

Civil Status (1=married, 0=never married)

0.399 0.146

8

(0.165)* (0.102)

Hidden cost (1=paid hidden cost, 0=otherwise)

-10.166 -4.674

(1.966)** (0.924)**

Agency

SSS -4.803 -2.180

(0.422)** (0.274)**

BIR -5.984 -1.687

(0.512)** (0.289)**

LRA -2.649 1.932

(0.534)** (0.299)**

PAGIBIG -3.525 -0.971

(0.514)** (0.339)**

GSIS -1.077 -0.667

(0.486)* (0.283)*

Service Quality

Timeliness – Response Time 5.845

(0.264)**

Timeliness – Attention 2.475

(0.268)**

Outcome 3.238

(0.130)**

Knowledge 1.987

9

(0.270)**

Competence 1.496

(0.282)**

Courtesy and extra mile 1.768

(0.123)**

Fairness and ethical treatment 2.228

(0.130)**

Education attainment No Yes Yes

Region No Yes Yes

_cons 58.00 59.531 14.276

(1.877)** (2.543)** (1.799)**

R2 /Adjusted R2 0.03 0.09 0.52

N 40,271 39,962 35,621

Reference agency is Philhealth. Clustered robust standard errors in service office are in parenthesis. * p

10

Table 3. Average Amount per Type of Hidden Costs (in PhP), 2014-2015

Variables Obs. Mean Standard deviation

Min Max

Don’t know why 12 317.92 553.86 15 2,000

Different from CC 9 1,105.56 1,943.81 50 6,000

Grease money 12 1,612.5 2,847.50 50 10,200

Food and transportation 7 95.71 94.67 20 300

Extra requirements 48 1,002.71 6,489.39 15 45,015

Extra services 47 139.60 123.82 3 300

Registration and penalties 28 1,415.25 2,306.80 32.02 12,000

Observations with 0 amount were excluded.

3.2 Challenges to using statistics The usefulness of the RCS dataset for statistical analysis has been constrained by various reasons. First, there were several threats of measurement error. The statistics generated from the RCS is inconsistently measured across the years and do not cover the same frontline services. The inconsistent measurement constrains the utilization of the RCS despite being available for multiple years. There is also a reason to suspect that the sampling approach may not have been enforced on the ground. While the RCS manual suggests systematic random sampling, the lack of clients and short-time period of the survey may have forced ARTA researchers to perform convenience sampling. If such is the case, any form of inference derived from the RCS should be qualified and nuanced. Second, while monitoring and evaluation were built into ARTA, how the data was collected constrains the use of other analytical methods to derive meaning from the RCS data. The RCS was used as a basis of deriving a score - a measure of the quality of frontline services. The score is a function of the results of the two instruments that capture client satisfaction (Survey Questionnaire or SQ) and compliance to ARTA (Inspection Checklist or IC). While a relationship between SQ and IC can be established, it would be difficult to relate the final RCS score with ARTA implementation because the relationship will be endogenous.

11

There are also variables that could offer a more comprehensive view about client satisfaction that were not captured by RCS, such as subjective expectations, transaction complexity and relative measure of efficiency. Expectations, theoretically, play an important role in shaping satisfaction with services. Coined as expectations disconfirmation model, the potential discrepancy between expected and actual level of performance have been shown to be a powerful determinant of citizen satisfaction (van Ryzin 2004, 2005, James 2009). Transaction complexity is about whether a service can be situated within a larger system of processes. For instance, a local government health permit could be part of an application for a business permit to operate. If a transaction feeds into the processes of other transactions, this must be taken into account in measuring satisfaction. A relative measure of efficiency can also be used as a proxy for quality of frontline services.

3.3 Recommendations

Revisions to existing RCS questionnaire

The following are suggested improvements to the existing RCS as a monitoring tool of ARTA:

1. Include qualitative measures of expectations. Existing theories of citizen satisfaction with public services point to the difference between expected and actual performance as a crucial determinant of satisfaction. High levels of expectations could explain low satisfaction levels despite improvements in the delivery of the service. A better understanding of expectations can allow prioritization of service improvement initiatives. Once service offices are confronted by a more discerning clientele, potentially a function of neighborhood or locality the office serves, agencies better channel resources towards these service offices to improve citizen satisfaction.

2. Develop a measure of transaction complexity. There is a heterogeneity of services availed in the agencies covered in this study. Low average satisfaction scores in BIR could in face be attributed to service complexity more than the qualitative features of service delivery. The citizen’s charter initiative forced all agencies with frontline services to catalogue the services they provide and the corresponding complexity (simple vs. complex) of these transactions. These information could be collected to generate a universe of public services in the country that describes the expected complexity of the service including service standards. Service complexity can be included in subsequent quantitative evaluation studies. It should also include information about the stage of the transaction relative to service being availed. For instance, payment of fees may be more complicated than the assessment of an application for services, which could take place on different periods.

3. Incorporate a relative measure of efficiency. One of the key outcomes of ARTA is to improve the relative efficiency of providing services vis-à-vis inputs. Ideally, measures of efficiency should measure the reduction in time, signatures and steps in delivering the process. However, existing methods of measuring process efficiency (e.g. time and motion studies) can be strenuous and expensive. In the absence of measures of process efficiency, input-output efficiency can be used as a proxy in order to ascertain the effect of efficiency gains on satisfaction. For example, monthly productivity as measured by number of transactions processed can be divided by number of staff or budget used to deliver the transactions.

Revisions to existing RCS methodology

There are also features of the RCS methodology that can be improved to facilitate and enhance the analysis that can be performed in the future. The following are suggested improvements to the existing RCS as a monitoring tool of ARTA:

12

1. Create a universe of public services. The unit of analysis of the RCS needs to be better clarified. At its current form, the RCS is conducted at the frontline service office level but given that several frontline services are delivered by the office, crucial information about the nature of services that might influence client satisfaction is ignored. It is recommended that rather than service office, future RCS can make frontline services as its main unit of analysis so that service complexity can be controlled for in the analysis. Random selecting services, however, can only be done once a list of frontline services is developed. All implementing agencies can be requested to submit a list of their frontline services that are subjected to CC.

2. Ensure consistency of measurement across time. The purpose of the RCS is allow for trends analysis across time but the inconsistent measurement of satisfaction inhibits a more rigorous comparison of the outcomes across time. Moreover, efforts should be made to construct a dataset that assesses similar units across time to allow future analysis to control for unit-level endogeneity.

4 Lessons from the Data Science Methods In this last section, we illustrate how data science methods can be serve as useful evaluation tools, and provide actionable insights to improve ARTA implementation.We then provide recommendations for improving the RCS methodology through the data gathering process, RCS scoring process, and data analysis process.

4.1 Machine learning methods for understanding the customer To understand the ARTA customer, our approach was to identify which features best explained overall customer satisfaction. We first constructed classification trees where our dependent variable was the overall satisfaction score from the RCS datasets, and our predictors were variables related to satisfaction criteria (e.g., response time, outcome, fairness, etc.) and respondent demographics (e.g., age, marital status, etc.). We then conducted feature importance analysis to identify which features were most useful in classifying respondents. Higher feature importance signifies that a particular variable was more useful in determining overall satisfaction scores. Figure 3. Relative feature importances from LGBM (Light Gradient Boosted Machine) classification of overall satisfaction.

13

Feature importance analysis reveals that response time, outcome, and overall transaction time are the three elements most important in driving customer satisfaction (Fig 3). This means that customers are most satisfied when government agencies respond quickly to their concerns, and they get what they request for at the soonest possible time. Based on these findings, some recommendations include: (1) for prioritization of training and improvement, focus on factors that are most important to customers, such as speed and outcome, rather than less important factors such as facilities; (2) attend to customers quickly, and keep them updated on the status of their requests; and (3) level off expectations with the customers based on the complexity of their requests, and have a clear system to prioritize which requests to complete first.

4.2 Data visualization method for understanding each unit’s performance drivers Data visualization can be useful in drawing insights as well. For example, we found that by breaking scores down across agencies per region, PhilHealth tended to bring mean scores up, while Municipality scores were typically below average (Fig 4). Figure 4. Mean RCS of all units across each region and agency. Colored circles represent the top 5 agencies in terms of RCS. The black circle represents the mean RCS across all agencies in a given region.

14

But how can we help lower-performing agencies? When ranking agencies across the top 8 satisfaction ratings, it is notable that the same agencies tended to stay within the same rank (Fig. 5), which implies relatively consistent levels of agency performance across key features. This implies that in terms of potential to leverage learning across offices, it would be valuable to conduct knowledge-sharing sessions with a proper mix of low-performing agencies to high-performing agencies. Fig. 5. Mean satisfaction scores per agency across all sub-satisfaction metrics included in the top 10 most important features. Note the change in X-axis range for ‘Mean Physical setup Score’ and ‘Mean Facilities Score’, due to mean scores below 4 for some agencies (i.e., SUCs, PRC).

15

4.3 Recommendations to improve the RCS methodology From the data science methods, the following recommendations are forwarded:

1. For the data collection process: In general, the type of data collected through the RCS process has significant potential for major analytics and insight-generation. However, because of the nature of certain parts of the data gathering process, this potential is significantly reduced. For example, some recommendations include keeping maintaining a standardized methodology and survey offices, avoiding aggregation, and limiting response types.

2. For the RCS scoring process: Results of feature importance analysis on customer satisfaction indicate that the RCS scoring mechanism should be weighted based on the features that more

16

closely drive customer satisfaction, in order to accurately portray RCS scores that most accurately depict customer sentiment.

3. For the data analysis process: For the Data Analysis Process, we have two main recommendations. First, shifting from reporting pure summary statistics to time-series analyses will enable evaluators to quantify how each one’s performance has improved or deteriorated over time. Second, gathering more information on most relevant features will allow for deep diving on sub-processes that may be helping or hindering aspects of customer satisfaction.

Works Cited Agamata, L. (2018). ARTA Overview - Presentation to the ARTA ERG. Quezon City: Civil Service Commission. CSC. (2014a, February). Report Card Survey. Retrieved from Civil Service Commission: http://www.csc.gov.ph/2014-02-21-08-16-56/2014-02-21-08-17-48/2014-02-28-06-38-42.html CSC. (2014b, February). Contact Center ng Bayan. Retrieved from Civil Service Commission: http://www.csc.gov.ph/2014-02-21-08-16-56/2014-02-21-08-17-48/2014-02-28-06-40-47.html CSC. (2014c, February). Service Delivery Excellence Program. Retrieved from Civil Service Commission: http://www.csc.gov.ph/2014-02-21-08-16-56/2014-02-21-08-17-48/2014-02-28-06-39-48.html CSC. (2014d, February). Citizen's Satisfaction Center Seal of Excellence. Retrieved from Civil Service Commission: http://www.csc.gov.ph/2014-02-21-08-16-56/2014-02-21-08-17-48/2014-02-28-06-40-13.html

Medina-Guce, C., Saguin, K., & Jovellanos, K. (2019). Assessment of the Anti-Red Tape Act (ARTA): Efficiency, Effectiveness, and Relevance, Towards the Expanded ARTA/Ease of Doing Business Act Implementation. Pasig City: National Economic and Development Authority and United Nations Development Programme.

Mendoza, M. (2011). Legislating Citizen's Charters. Makati: Development Academy of the Philippines.