56448 Using Public Expenditure Tracking Surveys to Monitor Projects and Small-Scale Programs /A Guidebook Margaret Koziol and Courtney Tolmie Using Public Expenditure Tracking Surveys to Monitor Projects and Small-Scale Programs A Guidebook Margaret Koziol and Courtney Tolmie Washington, D.C. ©2010 The International Bank for Reconstruction and Development/The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org All rights reserved 1 2 3 4 13 12 11 10 This volume is a product of the staff of the International Bank for Reconstruction and Development/The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgement on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www. copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN: 978-0-8213-8519-7 e-ISBN: 978-0-8213-8520-3 DOI: 10.1596/978-0-8213-8519-7 Library of Congress Cataloging-in-Publication data has been requested. iii Table of Contents Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 What is a Public Expenditure Tracking Survey? . . . . . . . . . . . . . . . . . . 3 Step 0. Defining the Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Step 1. Mapping Resource Flows . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Step 2. Collecting and Analyzing the Data . . . . . . . . . . . . . . . . . . . . . 33 Step 3. Identifying the Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Step 4. Recommending Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Step 5. Dissemination and Advocacy . . . . . . . . . . . . . . . . . . . . . . . . . 41 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Annex A: World Bank PETS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Annex B: CSO-Led PETS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 Annex C: Resources for Power Calculations . . . . . . . . . . . . . . . . . . . . 66 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 v Acknowledgments This guidebook is the product of a collaborative effort between the World Bank Human Development Network's Office of the Chief Economist and the Transparency and Accountability Program (TAP) of the Results for Development Institute. The goal of this partnership is to support the capacity of civil society organizations (CSOs) to monitor and evaluate government programs in the human development sectors. The World Bank contributes to this partnership by offering support to civil society organizations participating in the program, including con- tributing technical expertise on Public Expenditure Tracking Surveys (PETS) and absenteeism methodologies, preparing training materials, re- viewing research instruments and reports, and supporting workshops and courses for the researchers involved in the studies. The Bank also takes an active role in making connections between the CSOs supported by TAP and task teams to build links with operations. Established in 2006 with support from the William and Flora Hewlett Foundation, TAP focuses on strengthening the demand side of gover- nance by providing technical and financial support to civil society organi- zations (e.g., non-governmental organizations [NGOs], think tanks, and universities) involved in promoting the accountability of public services to civil society and to citizens through analysis and advocacy. TAP's activities aim to address the major barriers that civil society organi- zations face in turning actions into results: weak institutional capacity and skills, non-existent or adversarial relationships with government and poli- cymakers, and lack of access to information. TAP helps organizations over- come these barriers by supporting access to information, providing both the incentive and tools to develop analytic skills and capabilities, and pro- viding technical support and peer learning opportunities to help connect vi Using PETS to Monitor Projects and Small-Scale Programs CSOs with government and stakeholders so that their policy recommen- dations might be heard. This guidebook benefitted from comments by and conversations with many colleagues, including Zafar Ahmed, Deon Filmer, Ariel Fiszbein, Bernard Gauthier, Markus Goldstein, Charles Griffin, Kai Kaiser, Dena Ringold, and many current and former TAP program participants. Please send comments to mkoziol@worldbank.org and ctolmie@results- fordevelopment.org. 1 Using Public Expenditure Tracking Surveys to Monitor Projects and Small-Scale Programs Overview Most often, Public Expenditure Tracking Surveys (PETS) are referred to in the context of a large, nationally sampled public expenditure review, or in the context of similar studies conducted by organizations such as the World Bank. A tried and tested methodology, PETS have been shown to be effective in identifying delays in financial and in-kind transfers, leakage rates, and general inefficiencies in public spending. Recently, civil society organizations have successfully taken up this methodology as part of a push for budget accountability from the ground up, an effort supported in part by the creation of this guidebook. Civil society organizations' comparative advantage resides in their ability to "take the temperature" on the ground and to act on those issues that are most heated in the minds of the citizens whom they represent. In many cases, service delivery in the education and health sectors is a top priority. One way of improving service delivery, by keeping both govern- ments and service providers accountable, is through the monitoring of budgets and efficiencies in public spending. PETS, when used by civil society organizations, offer an opportunity to carefully monitor specific programs or public spending in targeted districts and regions. Moreover, this instrument may be used to monitor World Bank and other projects. This brand of accountability strengthens the voice of citizens and focuses the discussion on problems that may be micro-level in nature, or specific to a particular region of a country. 2 Using PETS to Monitor Projects and Small-Scale Programs The goal of this guidebook is to serve as a starting point for civil soci- ety organizations, as well as Bank teams interested in conducting Public Expenditure Tracking Surveys, both on a small and larger scale. It is de- signed to lead a research team from idea inception to results dissemination, while emphasizing the importance of utilizing evidence to influence poli- cy, regardless of whether it is on a macro or microlevel. Though the World Bank has been at the forefront of efforts to measure the effectiveness of service delivery, it is hoped that Bank teams and civil society alike will take the research one step further and empower citizen users to keep ser- vice providers accountable through information dissemination and citizen engagement efforts. The Power of PETS Experiences with PETS point to the strength of this particular methodology when used to monitor programs, projects, and public spending. Some illustrative examples of findings from these studies include the experiences of Ghana, where 50 percent of non-wage education expenditures and 80 percent of non-wage health ex- penditures did not reach intended beneficiaries; of Zambia, where leakage rates of discretionary non-wage expenditures reached 76 percent; and of Chad, where less than 1 percent of the non-wage budget officially allocated to regions actually reached health cen- ters.1 In Ghana, researchers found that incidences of leakage were more prevalent when the value of in-kind materials distributed to beneficiaries was unknown. Researchers in Zambia found evidence to suggest that a few select schools with greater bargaining power were able to capture the majority of the funds. And, in the case of Chad, it was estimated that, had all resources officially budgeted to regions actually reached service providers, the number of patients seeking primary health care would have more than doubled. These examples are only a small indication of what researchers can do by strategically utilizing the PETS methodology. 1 Ye and Sundharshan 2002, Gauthier and Wane 2006, and Das, et al 2004a. A Guidebook 3 What is a Public Expenditure Tracking Survey? Public Expenditure Tracking Surveys are tools in a methodology used to track the flow of public resources (including human, financial, or in- kind) from the highest levels of government to frontline service providers. PETS first identify a research objective, then employ an extensive map- ping exercise in order to understand the flow of funds through the dif- ferent levels of government. Once the resource flows are mapped, budget data are collected and analyzed, and often complemented with a facilities survey and qualitative research. PETS can help civil society and policy makers alike to understand funding flows, identify areas of leakage, and make informed policy decisions based on their findings. Goals and Uses PETS were developed to gather information beyond official data and administrative records to understand what actually happens to money that is appropriated for service delivery. We propose a two-tiered typol- ogy to describe how PETS can be applied in the human development sectors. Analytical PETS ­ PETS with an Applied Research Focus The first PETS conducted by the World Bank was a diagnostic tool to sup- port a Public Expenditure Review (PER) in Uganda. At the national lev- el, funding had been provided for non-salary expenditures to each school in the form of a capitation grant. Normally a PER would take note of the amount attributed to each school and how it compared to what was need- ed for non-salary inputs. In this case, the PETS allowed analysis of what actually happened in the allocation; the result was startling. On average, 87 percent of the grant did not reach the school and was diverted along the way. The Ugandan government subsidized an information campaign alerting service users to the findings, whose results were later correlated to a decrease in the amount of the grant being diverted in subsequent years. Outside agencies have come to see PETS as a good instrument for 4 Using PETS to Monitor Projects and Small-Scale Programs ascertaining what actually happens to money or goods that are supposed to reach service delivery points. PETS, and other similar methodologies,2 are important in showing sys- temic problems related to how a budget is executed, which--with auditing and evaluation--is typically the weakest element of government perfor- mance (as shown by PEFA surveys and the Open Budget Initiative).3 In the case of Uganda, the discovery of leakages and the public campaign to alert parents seem to have adequately addressed the problem.4 One way of using this approach to PETS, especially on a more micro level, might be to consider whether there are higher rates of leakage and capture in certain districts than others, or to focus the research by target- ing a specific question, program, or geographical location. Yet another design might examine different types of education or health facilities and then compare outcomes. When using PETS as an analytical tool, re- searchers or civil society organizations can utilize the findings to identify causes, which they can then use as evidence to inform policy decisions; advocacy supported by robust results can be a powerful tool, regardless of whether a research team is working at the grassroots or national level. In many cases consulting firms or universities are contracted to conduct analytical PETS because the function of the research is to derive gener- alizations at a fairly high level. However, developing capacity among civil society organizations interested in scaling up research efforts should also be considered. Monitoring PETS ­ PETS with a Governance Purpose There is another way to use PETS, which is the primary purpose of this guide. Monitoring PETS, which are also analytical in nature, ask whether 2 Quantitative Service Delivery Surveys, Public Expenditure Reviews, Public Bud- geting Analysis, Benefit-Cost Analysis, etc. 3 http://www.pefa.org/ and http://www.openbudgetindex.org/ 4 Reinikka and Svensson 2001. A Guidebook 5 an intervention is working as intended and other narrow questions about service delivery at the local level. Typically, they do not seek to solve national problems or devise new laws or regulations that could change incentives countrywide, though this sometimes ends up being the result. World Bank teams, for example, may have an interest in ensuring that the intended beneficiaries of a project are able to monitor whether, and what, the project is delivering. In that case, it would not make much sense to turn the job over to a consulting firm, a university, or a government agen- cy, because part of the purpose is to generate knowledge among those who have an interest in demanding better results--the users--and to increase their ability to monitor and advocate for improvements. Monitoring PETS can also amplify the voice of civil society organiza- tions and user-citizens by giving them an empirical tool to demand, on behalf of end-users, improvements in governance, management, and delivery of services, at least from the standpoint of resource provision and the reduction of financial waste. Because they are based on facts, PETS-derived lessons can help civil society organizations build cred- ibility through knowledge-based constructive engagement with govern- ments and seek feasible changes on the local level. CSOs are able to shift from pure advocacy--for causes, for higher spending, or for rights--to an analytically-based approach that demands accountability for results and for delivery on promises already made. An example of bridging the gap between analytical and monitoring PETS exists in Uganda where the first step was diagnosis through an analytical PETS, next was a public media campaign as a solution, and finally a follow-up survey to ascertain what occurred. In Uganda, using PETS identified a particular problem, action was taken to address the problem, and the result was monitored. The monitoring PETS combines diagnosis, action, and follow-up. If per- formed locally by a CSO that is interested in monitoring governmental performance, the PETS can also contribute to institutional development in this area.5 5 Reinikka and Smith 2004. Gauthier 2006. 6 Using PETS to Monitor Projects and Small-Scale Programs One example of a CSO-led PETS incorporating diagnosis, action, and follow-up--as well as discernable policy change--comes from Centro de Investigaciones Económicas Nacionales (CIEN) in Guatemala. Re- sponding to the call from a newly elected Minister of Education to im- prove education in her first 100 days in office, CIEN designed a PETS to identify existing problems in six primary school financing programs including support for textbooks and school meals. While CIEN found little evidence of leakage during its study, the organization identified sig- nificant delays in the distribution of resources to schools. Interviews and focus group discussions carried out in conjunction with the study allowed researchers to diagnose one possible cause for these delays: a school cal- endar that overlapped with the fiscal calendar and led to backlogs in ap- proving resource allocation and delivery to schools. CIEN took action by presenting findings and recommendations to a willing Minister of Edu- cation. CIEN continually followed up with the Minister to encourage adoption of its recommendations, which contributed to the Ministry's decision to shift the school calendar in 2009 in order to address the delays. It is hoped that this manual will serve as a guide for organizations and research teams alike; the general procedure is the same regardless of the purpose. Scaling it up for an analytical PETS would be straightforward, but here there is an emphasis on choosing a narrow purpose, developing a small sampling framework, carrying out the work inexpensively, and tak- ing the important final step of not shelving the study but of advocating for change, based on the findings. An important item to note is that, by focusing on a discrete and measur- able objective, as we suggest in this guidebook, researchers are likely to be looking at a very small portion of the overall government expenditure. This was the case with the first PETS conducted in Uganda, where the capitation grant tracked represented approximately 3 percent of the re- current expenditure on education. This does not in any way diminish or lessen the importance and impact that a small-scale exercise may poten- tially have. Nonetheless, understanding the limitations of these types of studies is essential. A Guidebook 7 Step 0. Defining the Objective The first step in conducting a PETS--defining its objective--can be the most significant step in determining whether the implementing orga- nization succeeds in conducting a valid and valuable study and whether the work has a short-term or long-term impact on public spending and service delivery. Defining a clear objective for a PETS will drive not only the scope of the study, but will also determine what an organization or research team can do with the study's results. That said, thinking about the way in which the research results will be used can also be a power- ful tool for determining the objective and scope of the study. For either category of PETS--analytical or monitoring--defining the objective involves making critical decisions about the study, including selecting the topic and scope, developing research questions, and identifying target audiences. Selecting the Topic and Scope The success of any research team in conducting a high-quality and high- impact study depends first on how well the topic and scope of the research are defined. A list of PETS conducted by the World Bank alone (Annex A) illustrates the breadth of possible study topics. Comparing these stud- ies with those conducted by CSOs (Annex B) further demonstrates the difference in scopes among successful PETS. It is critical that implement- ers choosing to conduct a PETS make decisions about the specific scope and focus of their study. Table 1 presents a non-exhaustive list of ques- tions and examples of decisions made in response. When answering these questions, research teams should consider what they are trying to achieve with the PETS project (both the study itself and the advocacy and dissemination of results) and how to design the project to best achieve these goals. While the project design and choice of topic should take into consideration a number of different factors, there are three questions that we recommend all research teams ask during the project design process. We call this the FIR criteria: 8 Using PETS to Monitor Projects and Small-Scale Programs Is the proposed topic FEASIBLE for your organization? CSOs are well- placed to implement PETS because of characteristics that they share-- namely, being based in-country and having an understanding of the local context. However, individual CSOs have differing strengths, and the most successful PETS are designed to complement an organiza- tion's strengths. For example, the Centre for Regional Information and Studies (PATTIRO),6 a research advocacy organization in Indonesia, designed a tracking study of education program grants that supple- mented basic quantitative findings with qualitative results from focus group discussions and interviews. On the other hand, the Romanian Academic Society (SAR),7 a think tank, conducted a more rigorous quantitative analysis of school spending in Romania. Both of these CSOs built upon the existing strengths of researchers rather than at- tempting to conduct a study that fell outside their areas of expertise. Is the proposed topic of INTEREST to the target audiences? A common mistake made with PETS and other public expenditure studies is that they are seen by project implementers as purely research exercises. To maximize the likelihood of improving policies and the effectiveness of public spending, CSOs and project teams should choose a topic with an eye towards getting the attention of stakeholders and using their results to influence positive change. One way of approaching this is for implementers to choose a topic that is of interest to the people that they hope to influence. For example, focusing the research on an area that is being focused on by the national government. Furthermore, and perhaps even more essential, research teams can increase the likeli- hood of successfully impacting policy change by identifying a policy entry-point or a well-placed champion for the cause, or by packaging the results in such a way that the research team is targeting the right audience and being heard by the right people. Is the proposed topic RELEVANT to the organization's mis- sion and long-term agenda? Target audiences such as policymak- ers, NGOs, and the public are not the only ones that should buy into 6 http://www.pattiro.org/ 7 http://www.sar.org.ro/ A Guidebook 9 the project design and topic; the implementing CSO or project team should choose a topic that fits into its mission and agenda for the future. While project funding can be attractive to CSOs, it can also have the unintended effect of taking an organization off its mission- designated path. Project topics should not be donor-driven; the most successful PETS are designed to address the mission and goals of the lead organization, as well as critical needs identified on-the-ground. Table 1. Selecting a PETS Topic Question Examples What facility level? Primary schools versus secondary schools What facility type(s)? Public, Private, NGO-led What category of spending or Salaries, Capital spending resource flow? What program or policy? Scholarship program, Textbook fund, Vaccine spending or in-kind transfers What is the geographic scope? Focus on entire country, specific province(s), specific district(s) A Case for Targeted PETS While the World Bank and individual country governments have under- taken large-scale, sector-wide PETS in the education and health sectors, the most successful and most frequently cited studies have focused on specific programs or policies. The inceptive PETS conducted in Uganda in 1996 followed the allocation and disbursement of capitation grants for non-wage spending in Ugandan primary schools. While many factors contributed to the success of this PETS in lowering the level of leakage in the Ugandan program, the value of focusing on one program cannot be overstated. Armed with unambiguous evidence of leakage points in the capitation grants program, the public and media could easily transition to monitoring improvements in grant distribution and holding the correct officials accountable for any continuing problems. 10 Using PETS to Monitor Projects and Small-Scale Programs Recent PETS conducted by civil society organizations have taken similar approaches and shown early signs of success in developing a sustainable system of monitoring spending in specific programs. Focusing on a specif- ic program or policy provides many advantages for smaller scale exercises: Implementation is less complicated. Rather than having to disentangle the large and often complicated web of sector-wide spending, a re- search team can focus on the small portion of the web that goes to its program of interest. Further, program-specific PETS generally require fewer data sources, increasing the likelihood that implementers will manage to gain access to all necessary data in a timely manner. Audiences are easier to target. A well-implemented PETS will gener- ally identify areas for improvement (if not points of leakage) at each stage of the expenditure chain. Successful dissemination of results and recommendations includes bringing results back to government offi- cials and service providers at each of those links in the chain. In a sec- tor-wide study, the number of individuals or departments that would need to be targeted for dissemination can be daunting. In a PETS that focuses on a particular program or policy, the audiences for dissemina- tion are smaller in number and generally easier for a team to target. Recommendations are often targeted, concrete, and more attractive to policymakers. Large-scale PETS take on the difficult task of evaluat- ing the effectiveness of public spending for an entire sector, and the recommendations that come out of the studies often reflect this. For example, without focusing on specific components of education spend- ing, project implementers generally identify system-wide problems and recommend solutions that reform the wider public expenditure management system. While these recommendations are often valu- able, they are not always as attractive to policymakers (especially those facing short political terms or upcoming elections). On the other hand, a PETS that focuses on a textbook spending program can identify very specific problems and develop tangible and frequently inexpensive so- lutions. When presented with this type of recommendation by a cred- ible CSO or team of researchers, policymakers are much more likely to be amenable to implementing the proposed solution. A Guidebook 11 Box 1. World Bank Health PETS Nigeria8 Motivation: To assess a decentralized delivery of health services un- der a federal system. Findings: · Large-scale leakage of public resources in Kogi State. · 42 percent of health staff had not received salaries in over six months, despite the fact that budget allocations were sufficient to cover costs. · To supplement salaries, public health providers charged for ser- vices illegally, while expropriating drug supplies and selling them to patients. Follow Up: Though the study suggested that providing service users and citizens with more information about budget resources so they could hold providers accountable, no documented post-PETS experi- ences currently exist. Developing Research Questions Like topic and scope, the list of possible research questions that a PETS can be used to answer is extensive. The most cited findings from expen- diture tracking studies are leakages in spending. Results of leakage also tend to command the most attention in the media and the public because they can explicitly point out instances of corruption. Budgeting and ex- penditure analysis are often described as unglamorous topics; however, demonstrating corruption in the form of stealing public money is likely to grab the attention of those who are not normally interested in public spending. Implementers can also take the next step to explore patterns in leakages across geographic areas or time of year and determine potential causes of these problems. 8 Das Gupta, Gauri, Khemani, 2004. 12 Using PETS to Monitor Projects and Small-Scale Programs In addition to leakages, though, PETS studies can uncover many other problems with the functioning of the public expenditure management system and service delivery. PETS, along with review of financial records at facility and government levels, can also uncover delays and inefficien- cies in resources reaching their intended beneficiaries. These issues can be due to poor planning, simple accounting errors, pharmaceutical expiration dates leading to medicines going unused, or inexplicable losses due to hu- man error. In other words, leakage does not necessarily mean corruption. Past PETS have also frequently been expanded to include a collection of data on service provider absenteeism, mismatches between facility needs and allocations, and user satisfaction with public services.9 Although these topics are not a part of traditional PETS, research teams and civil society organizations often bring one or more of these additional questions into their budget tracking studies. Before developing surveys and other methodological tools, it is valuable for the implementers to determine what questions they are hoping to an- swer with the study. Examples of questions (investigative and analytical) are listed in Table 2. We make the same case for research questions as we do for topic and scope; taking a targeted approach increases the feasibility of the study and the likelihood of successful dissemination and advocacy of the PETS results. While it can be tempting for implementers to attempt to answer all of these possible research questions with one study, narrowing the list of questions allows researchers to focus on the issues that are most im- portant and investigate those few issues in a much more rigorous way. Limiting the list of research questions also allows implementers to design and administer shorter surveys that are less burdensome to respondents, increasing the likelihood of a high response rate. 9 Information on methodology like quantitative service delivery surveys is available online at http://go.worldbank.org/MB54FMT3E0. A Guidebook 13 Table 2. Examples of Diagnostic and Analytical Research Questions for PETS Studies Investigative Questions Analytical Questions What is the level of leakage? How does the level of leakage vary across districts? At what stages of the expenditure Are schools with X characteristic chain do leakages occur? less likely to experience leakages? Are there delays in resources Are there institutional factors or reaching beneficiaries? characteristics of the expenditure system that may be causing delays in resource disbursement? What is the level of absenteeism? How does absenteeism vary between different facility types? Does the funding requested by facilities match the funding allocated to them? Many PETS implementers find it helpful to utilize a research matrix to carefully define their topic and research questions and to ensure a targeted and narrow set of objectives. The sample matrix in Figure 2 comes from ABANTU for Development,10 a Nigerian CSO currently implementing a PETS in the education sector, with the support of a grant from the Trans- parency and Accountability Program. A matrix such the one in Table 3 al- lows implementers to translate objectives into specific research questions, variables, and necessary sources of data. Identifying Target Audiences and Developing a Dissemination Plan The final critical component of defining project objectives is to identify the target audience for the study and develop a dissemination plan. A common criticism of PETS is that, while the study may identify critical 10 http://www.abantunig.org/ 14 Using PETS to Monitor Projects and Small-Scale Programs Figure 1. Conceptualizing the PETS Determine Research Questions Identify Target Define Audience for Objectives Dissemination Conceptualizing the PETS issues, the results are not utilized to improve the efficiency of public spending or service delivery. One way in which implementers can ensure that that PETS results gain traction after the project is completed is to begin the PETS process with a plan of how results will be used to enact change. Research teams should begin by identifying and consulting with the project's target audiences before it has been implemented. Gaining buy-in from stakeholders at the project design phase will secure the interest of audiences in the topic and increase the probability that they will implement the recommended changes following the completion of the study. In addition to identifying target audiences at the beginning of the PETS, implementers should also develop a strategy to keep these stakeholders in- formed and involved throughout the project's duration. This is particularly Table 3. Sample Research Matrix for PETS Method of Data Data Objectives Questions Variable of Interest Data Source1 Collection Collection Tool To understand What are the · What was budgeted? State budget Desk review Extraction the mechanism of procedures with · What was approved? LGA budget forms/checklist resource allocation in regard to resource · What was released to SUBEB budget the education sector allocation and the sector? MOEP at the state and LGC transfers? · What are the levels and identify methods of major issues in disbursement of budgetary allocation What are the spending funds at the two in 3 LGAs in Kaduna patterns at each level? levels? State · What is the funding gap to sector needs? To identify problems What are the · Actual funds Financial Extraction Extraction with resource procedures with received at each records of the forms/checklist allocation and regard to resource level? tier In-depth transfers transfers? · Efficiency in interview In-depth disbursement? Finance and Interview What are the spending · Budget discipline? administration schedule patterns at each level? · Transparency and officials at the Accountability of two levels funds at all levels? · Challenges and A Guidebook constraints? 1 LGA ­ local government area; SUBEB ­ State Universal Education Board; MOEP ­ Ministry of Economic Planning From Research Matrix developed by ABANTU for Development based on materials presented by Delius Aslimwe, July 2009. 15 16 Using PETS to Monitor Projects and Small-Scale Programs true for studies that are implemented over a long period of time. While stakeholders may show enthusiasm at the beginning of a project, this interest can wane if there are significant stretches of time during which the research team does not provide these individuals with updates. At the same time, researchers should be aware of and adapt to changes in their constituency and government counterparts, as contexts such as the political climate and government representation may change. While interacting with stakeholders, it is best for research teams to know what their dissemination and advocacy goals are. However, strategizing about what advocacy and communications activities will be undertaken, who will be targeted, and what will be achieved is a valuable exercise to undertake before the first pieces of data are collected. Choosing a Topic ­ Major Take-Aways · Consider whether your topic is feasible given the skills of your proj- ect team, of interest to policymakers and other key audiences, and relevant to the mission of your organization. · Narrowing the focus of the PETS can make implementation and subsequent advocacy easier and more effective. · In addition to selecting a topic, develop research questions that you hope to answer with your study and keep these questions in mind as you create your methodology and instruments. · Identify key audiences before beginning the work plan to ensure buy-in and willingness to listen to results and recommendations. Step 1. Mapping Resource Flows A core idea behind Public Expenditure Tracking Surveys is that govern- ment officials and facilities have an incentive to misrepresent the amount of resources going to service users if some fraction of these resources is being used for purposes for which the funds are not intended (including padding the pockets of politicians and diverting to other public spending A Guidebook 17 line items). To deal with this problem of incentives, PETS implementers triangulate budget data by looking at records from all of the places where money changes hands for particular resource flows. Viewing district level records of money allocated to a primary school by itself cannot identify potential corruption; comparing these records to ones kept by the primary school itself and seeing that the two records do not match, however, can point to a place where money may have leaked. Before triangulating budget data however, implementers must prepare to collect the necessary data. One critical step that research teams must take before actually gathering the data is determining what data are necessary. Most public social sector funds go through many hands before reaching the service user, and accurately determining both the level and location of leakages requires that project teams start with a complete and correct picture of what the expenditure chain looks like. This exercise in and of itself can be a major contribution of a PETS. Though various entities (government ministries and agencies) understand their own budgets and transfers systems, very few, if any, truly understand how frontline service providers are actually funded. A well-mapped PETS can serve as a vehicle for establishing this information. Mapping the flow of resources also provides a good list of sources from which to gather the data needed to conduct the PETS. Researchers should also note that different entities along the expenditure chain may call the same pot of money by different names, potentially making data collection and analysis more challenging if not noted and accounted for at early stages of research. The resource map may be a relatively simple flow or a complex web with multiple sources of funding and links in the chain. A 2008 tracking study by the Institute for Policy Analysis and Research (IPAR) in Kenya fol- lowed the flow of secondary education bursary funds from the central government to the secondary school level.11 IPAR's background work 11 Oyugi, Riechi and Anupi 2008. 18 Using PETS to Monitor Projects and Small-Scale Programs Figure 2. Flow of Funds for Secondary Education Bursary Scheme in Kenya Ministry of Secondary Education Schools · Charged with the responsibility of evaluating and · The main source · The custodian of awarding bursaries of bursary funds the Secondary to students at the · Recipient of the School Bursary · Ultimate beneficiary received in constituency level in funds secondary schools Scheme accordance with the guidelines issued by the Ministry of Education Central Constituency Bursary Students Government Committees A Guidebook 19 mapping the flow of funds demonstrated a simple flow through four links in a chain: the central Government, Ministry of Education, Constituency Bursary Committees, and the secondary schools (Figure 3). All of the funding for this scheme goes through the same channels, providing the research team from IPAR a clear list of data that it would need to collect. Another example of a reasonably easy to follow resource flow is found in the 2002 Zambia PETS conducted by the World Bank, as seen in Figure 3. In this analysis, six main flows of funds were identified and categorized by their sources, types, and discretionary powers in fund al- location. This PETS tracked non-wage financial flows from the Minis- try of Education and donors at the district, provincial, and facility levels. Unlike many other PETS, in-kind transfers and salaries were excluded from the analysis. The main objective of this study was to determine: Figure 3. Map of Resource Flows to Schools in Zambia1 Case IV Donors Ministry of Education (primarily the Program for Advancement of Province Girls Education) (Centralized) Province (Centralized) District (Centralized) District (Decentralized) Key to flows Schools Discretionary Flows Households Rule-based Flows 1 Das et al. 2004a. 20 Using PETS to Monitor Projects and Small-Scale Programs 1) whether schools received their allocated lump-sum payments; 2) if and how provinces and districts supported schools through discretionary spending; and 3) how decentralization affected fund allocation. Not all programs or sectors follow a relatively simple linear pattern like that of the Kenyan secondary education bursary scheme, or in the case of Zambia. Budget flows can be complex and tangled, as illustrated in the resource flow mapping for Peru's health sector from the Research Center of the University of the Pacific (Figure 4). A mapping of resource flows to Peruvian health care facilities demonstrates the myriad ways in which resource flows can prove to be complicated. First, funds originate from many different sources, unlike the bursary funds in Kenya that all start at Figure 4. Map of Resource Flows to Health Care Facilities Ministry of Key to flows Donors Finance Monetary (Treasury) Salaries Health Supplies and Insurance pharmaceuticals (SIS)(BU) Ministry PAAG of Health (BU) Administration Unit Health Networks (EU) Health Directorates (EU) Facility 1 SISMED Fund (DISA) Facility 2 Out-of-pocket payments Note: EU: "expenditure unit"; BU: "budget unit." See the section on CIUP in Chapter 2. Source: Alvarado, Betty, and Eduardo Moron. 2008. "The Route of Expenditures and Decision Making in the Health Sector in Peru." Centro de Investigacion de la Universidad del Pacifico: Lima. A Guidebook 21 the central Government level. Second, the direction of flows is not always "one way." Third, the resources that flow through the health expenditure chain are not only monetary but also in-kind. Despite the complexity of the resource mapping, an expenditure tracking exercise requires collecting data from each of these links and knowing how each link in the chain is connected in order to accurately identify leakages and other problems in the system. One example of the problems that can arise from an incorrect resource flow mapping is provided in Box 2. There is no algorithm for uncovering how resources are supposed to flow through a system, often making the exercise of mapping funds a tedious one. Some resource flows may be common knowledge to those working on public expenditure analyses. In other cases, research teams may have to interview facility-level administrators to uncover the source of their funding and follow the chain up to the funding source. Regardless of the method used to map the flow of funds, the importance of starting with an accurate mapping cannot be overstated. Box 2. The Importance of Mapping: Tanzania Not mapping the resources correctly can have detrimental effects on the accuracy and impact of a PETS. In 2003, an education expendi- ture tracking exercise in Tanzania found minimal leakages of funds (5 percent) for the Primary Education Development Project.12 However, further investigation uncovered that the actual level of leakage was much higher than that estimated in the 2003 PETS. The reason for this miscalculation was that the consultants carrying out the PETS had only looked at funds coming from the Ministry of Finance. By leaving out two additional main sources of funds for the program (Ministry of Education and Ministry of Local Government), the implementers had ignored a significant fraction of the expenditure chain, resulting in skewed findings. 12 Sundet 2007. 22 Using PETS to Monitor Projects and Small-Scale Programs Gathering the Budget Data After completing the map of resource flows, PETS implementers have a complete list of budget sources to check, compare, and triangulate in order to provide an accurate picture of inefficiencies in the expenditure chain. The research teams tracking the flow of funds described in Fig- ures 3 and 4 had to compile financial data from each of the sources of recipients appearing in the resource maps. Data for all links in the ex- penditure chain above the facility level will be secondary data, often ob- tained directly from the respective government office. Facility level data are generally gathered using survey instruments that the implementing team develops. Gaining access to necessary data can require some amount of finesse. On the one hand, engaging with government officials from the beginning can help research teams overcome the problem of uncooperative government offices; on the other, it can serve as a signal to government officials that it is time to start keeping an extra set of budget books. In the experience of TAP-supported CSOs, however, when officials see that the study is being conducted out in the open and not under the radar of policymak- ers, they sometimes feel more comfortable providing researchers access to data (Box 3 has additional strategies from CSOs). However, in many cases, and despite all efforts, research teams are faced with the challenge of lacking access to data. Gathering budget data itself, however, can also reveal interesting trends in public expenditure management. First, the challenges that civil so- ciety and research teams more generally may face in gaining access to data from government officials can demonstrate the lack of transpar- ency that exists in the system and the need for a remedy to opaque budgeting procedures (Box 3). Second, without looking at the numbers, many implementers find from the process of collecting budget records that there is significant miscommunication between different levels of policymakers about budgeting responsibilities and funding requests. Although uncovering these problems is not a goal of PETS, it is an A Guidebook 23 Box 3. Lack of Access to Data One of the major challenges that research teams face in conducting PETS is access to reliable data. This problem can occur in a number of different forms, including: · Data exists but the government does not want to provide it to re- searchers · Data exists but is of very poor quality and is unreliable or in an unusable format · Data just does not exist (a particularly common problem at the fa- cility level) Centro de Análisis y Difusión de la Economia Paraguaya in Paraguay (CADEP) encountered all of these problems when it conducted a PETS of the primary education sector in 2008. CADEP found that policymak- ers were very hesitant to provide the researchers access to data. When CADEP was finally able to obtain data from the government and the schools in their sample, they found that the financial records (espe- cially those kept by school administrators) did not provide adequate information on money received and money spent within schools in Paraguay. Although these data problems prevented CADEP from completing a traditional expenditure tracking exercise, they did provide the re- search team with evidence of the significant problems with transpar- ency of budget information in the country. The discovery also led CA- DEP to conclude that better tools were needed if civil society was to monitor the efficiency of public spending in educational facilities. At the current time, CADEP is working with active parent associations to develop ways to monitor transparency of budgeting within the schools and training modules to improve the financial record-keeping skills of school administrators. 24 Using PETS to Monitor Projects and Small-Scale Programs important finding that research teams can use when developing policy recommendations and advocating for change after the completion of their studies. Moreover, sharing this type of information with govern- ments can represent an easy-win for research teams; inefficiencies in the system may be easier to address than outright corruption. Additionally, an important consideration in implementing any PETS is timing. Ide- ally, data collection should include annual data and span a series of years for the sake of comparability and to maximize data quality. The key is to be certain to collect the necessary data before it is destroyed or archived in such a way that researchers will have a difficult time gaining access to it. Preparing to Follow the Money PETS is an excellent tool designed not only to examine budget alloca- tions, but more importantly to trace the flow of financial, human, and in-kind resources all the way down to frontline service providers such as teachers and health care workers. Mapping the landscape of budget al- locations, spending, and receipts alone is only the first step in determin- ing the quality and quantity of service delivery on the ground. We have begun to explore how PETS implementers can determine what budget data they need to collect and begin compiling that data; following the money generally requires the collection of secondary and primary data, including: National level budget data Sub-national level budget data Facility budget data Data on spending and service delivery on the ground The latter two types of data, in particular, are generally collected through the use of facilities surveys, interviews, and direct observation. Once the research team has mapped resource flows and determined the necessary sources of data, it can prepare to follow the money by selecting a sample for the facility survey and designing questionnaires. A Guidebook 25 Selecting a Sample While it would be valuable to identify all incidence of leakage in the se- lected program or system that the PETS is exploring, this is generally not feasible given the limited time and resources of the implementing team. Instead, research teams should select a sample of facilities that benefit from the program funds and could be suffering from inefficiencies and corruption in the program. Selecting a statistically valid sample is a critical part of any PETS, and this is especially true for micro-level PETS such as those often undertaken by civil society organizations. Some basic power calculations can help in- form research teams about how statistically significant their results will be; more subjects in a sample yields higher power. For civil society organi- zations, we suggest utilizing online resources and open-source programs that can help with the technical aspects of power calculations.13 PETS with small samples can still be informative, diagnose problems such as leakages, and point to possible ways to mitigate these problems. However, limiting the scope of the study decreases the likelihood that one or two non-typical facilities in a sample will drive the results, and allows the CSO to make more conclusive statements about problems with public spending. General rules for limiting the scope of a smaller study include: Select one type of facility for the sample. For small samples, the facilities included should be largely homogeneous but should exhibit one or two differing characteristics across which you are interested in comparing leakages. For example, if a research team wants to consider differences in leakages between urban and rural schools, then the team will need to select a sample containing both urban and rural schools. However, the sample schools should be otherwise homogeneous. For this reason, it is best to focus on one type of facility, for example primary schools, or primary health clinics. 13 See Annex C for Power Calculations Resources. 26 Using PETS to Monitor Projects and Small-Scale Programs Limit your study population. A sample of 30 schools out of 2,000 schools nationwide is not very representative and does not reflect the country's schools as a whole. Further, if the schools are scattered ran- domly throughout the country, it is impossible to determine if an es- pecially high leakage level in one school is an accurate representation of leakage in that school's district or if it is simply an extreme case or outlier. One solution may be to limit the study population by focusing on one district or province. After determining the scope of the sample, research teams can undertake one of a number of sampling strategies to finalize their sample. The strat- egy chosen will depend on the scope of the sample, the research ques- tions that the team develops, and the homogeneity of facilities within the population. The following are sampling strategies commonly used for PETS along with guidelines for when to use each strategy. Simple Random Sample. A simple random sample can be created by obtaining a complete list of facilities in your population (for example, all primary schools in a province) and randomly selecting facilities from that population. This strategy is only appropriate if the study population is relatively homogeneous and the sample is large relative to the population size. Example ­ A research team conducts a PETS that focuses on one dis- trict's secondary school scholarship program. They want to select a sample of 30 from the 75 primary schools in the district. The entire district is mostly rural and has similar literacy and enrollment rates. Stratified Sample. This strategy is used when it is important to get fa- cilities in a sample that are heterogeneous in one or more dimensions. To select a stratified sample, the population must first be broken into different sub-populations (or strata) and then facilities are randomly selected from each of the different sub-populations. The stratification should only be done on pre-determined or "exogenous" factors. This strategy is best used when you want to make a comparison across dif- ferent types of facilities. A Guidebook 27 Example ­ A CSO wants to conduct a PETS on spending in primary health facilities in one large district. In particular, the organization hy- pothesizes that clinics in rural areas of the district experience lower levels of leakage than clinics in urban areas. After obtaining a list of all clinics in the district, the CSO breaks the list of facilities into rural and urban (two sub-populations). The team then randomly selects 15 facilities from the urban list and randomly selects 15 facilities from the rural list. Multistage Sample. This strategy is often used for large populations (rela- tive to the sample size). Generally, a multistage sample involves selecting a certain number of districts from a province or country (stage 1) and then selecting an equal number of facilities from each of the chosen districts (stage 2). Selection at each of these stages can be random or stratified. Frequently, the first stage is stratified and the second stage is randomized. Example ­ The project implementer is interested in comparing the level of secondary school leakages in districts with different socio-economic profiles in a province. In the first stage of sampling, the team undertakes a stratified sample of districts in the state. Dividing the list into "high income" and "low income" districts, they randomly select two districts from each sub-population. The team then randomly selects 15 secondary schools from each of the four districts. Questionnaire Design While much of the data necessary to track expenditures can be taken directly from budget records at various levels of government, data from other links in the expenditure chain (such as facilities) will require the research team to develop and administer questionnaires to gather infor- mation about the facilities contained in the study sample. Surveys can provide critical information, such as demographic data on health workers or teachers, as well as data related to schools or health clinics. The questions that should be included in PETS questionnaires will be dictated by the research questions of the implementing team. A research matrix like the one described in Step 0 and found in Figure 5 can be help- 28 Using PETS to Monitor Projects and Small-Scale Programs ful to teams in deciding what questions are necessary and even what types of questionnaires should be developed for the PETS. After determining the questions and corresponding variables of interest, the research team can determine what sources of data are needed to collect each variable. Taking the first variable as an example, data on what is budgeted will generally be available in budget records from each link in the expenditure chain, circumventing the need for a questionnaire to collect data on that variable. However, information on what was released to the facility level may require a survey targeting school or health center administrators while answering the question of how well targeted funds are may require infor- mation from beneficiaries about access to textbooks or pharmaceuticals. Questions included in a facility questionnaire will generally fall into one of the six categories in Table 4. While every facility questionnaire need not include questions in each of the categories, some PETS instruments cover most or all of these issues. While research teams are well advised to limit the length of their survey so as to reduce the burden on respondents, implementers should also refer to the framework in Box 4 to ensure that they collect enough data to be able to answer the research questions that they pose. Table 4. Sample Research Matrix for PETS Method Data of Data Collection Objectives Questions Variable of Interest Data Source1 Collection Tool To understand What are the · What was budgeted? State budget Extraction the mechanism of procedures with regard · What was approved? LGA budget Desk review forms/ resource allocation to resource allocation · What was released to SUBEB budget checklist in the education and transfers? the sector? MOEP sector at the state · What are the methods and LGC levels of disbursement of and identify major What are the spending funds at the two issues in budgetary patterns at each level? levels? allocation in 3 LGAs · What is the funding in Kaduna State gap to sector needs? To identify problems What are the · Actual funds received Financial Extraction Extraction with resource procedures with regard at each level? records of the forms/ allocation and to resource transfers? · Efficiency in tier In-depth checklist transfers What are the spending disbursement interview patterns at each level? · Budget discipline? Finance and In-depth · Transparency and administration Interview Accountability of officials at the schedule funds at all levels? two levels · Challenges and constraints A Guidebook 1 LGA ­ local government area; SUBEB ­ State Universal Education Board; MOEP ­ Ministry of Economic Planning From Research Matrix developed by ABANTU for Development based on materials presented by Delius Aslimwe, July 2009 29 30 Using PETS to Monitor Projects and Small-Scale Programs Box 4. Facilities Questionnaire Framework 1. Facility Characteristics ­ This includes size of the facility, how many students or patients use the facility, what type of facility it is (non-profit, public, or private), range of services provided, and com- petition from other facilities. It may also be useful to collect some demographic information about the service users in the catchment area. 2. Inputs ­ This needs to be measured in monetary terms. In the case of health clinics wages, drugs and labor costs should be included here; in the case of schools, information on the head teacher and teachers should be included. 3. Outputs ­ Measurable outputs to calculate cost-efficiency include number of students enrolled in a school, number of patients seen, etc. In some cases test scores have been used to measure outputs in education. This was the case with a PETS conducted in Zambia. 4. Quality ­ In the case of PETS, quality can be measured through ob- servation and through focus group discussions. Quality can apply to issues of a facility's structural quality and its procedural tech- niques, as well as to educational and health outcomes. 5. Financing ­ Information should be collected on financing. Where is the money coming from? Government, donors, user charges? Amounts and type should also be recorded. 6. Institutional Mechanisms and System of Accountability ­ Because different types of facilities operate under different types of account- ability structures--for example, private schools are not beholden to the same rules that public schools are in many circumstances--it is important to collect data on management, reporting mechanisms, and record keeping. Source: Adapted from Gauthier, 2006. While developing a facility survey can be a daunting task for even a highly experienced research team, there are a few tips that implementers can fol- low to increase the likelihood of developing a clear and complete facility survey: A Guidebook 31 Utilize existing surveys.14 Research teams can benefit from question- naires used in other countries and contexts. Many PETS have been conducted on a diverse set of topics and locations, and new imple- menters need not reinvent the wheel to utilize the methodology in their own context. However, individual country context and issues of political economy cannot be ignored. The PETS methodology, though easily duplicated, must be designed to answer questions specific to each situation and research question. For example, a PETS questionnaire designed to study leakage in the health sector will look very different than a questionnaire focusing on a textbook procurement program, even within the same country. Limit survey questions to those needed to answer PETS research ques- tions. Research teams should remember that it is essential to collect a precise and rich set of data for analysis. That said, if too much data are collected it will become burdensome for research teams to sift through and pick out the pieces most important to their analysis. Furthermore, long questionnaires can become burdensome to respondents and re- sult in high non-response rates. Be particularly wary of "matrix-style" questions such as enrollment by grade, by gender, by year, for the last five years. This type of question can quickly end in an unmanageable amount of data points. For example, 6 grades, x 2 genders, x 5 years = 60 data points. In this way a small table in a questionnaire could end up actually containing 60 questions! Develop different instruments for different respondents if necessary. Depending upon the research questions, the project team may decide that it will need to collect different information from different types of stakeholders. When different questions need to be asked of different sources, it can be helpful to create a number of modules for the PETS questionnaires to be administered to various officials. For example, one module may be created for teachers, whereas another module could be created for principals or head teachers of the same school. Research teams should also consider whether they need to obtain non-budget 14 The World Bank is currently developing an online database of PETS tools, includ- ing instruments. 32 Using PETS to Monitor Projects and Small-Scale Programs information from outside of the facilities. Some PETS include surveys for district level officials (to collect information on district procedures for distributing funds, for example), beneficiaries (to collect informa- tion on delays in obtaining resources, for example), and others. It can be useful for research teams to start with one survey and adapt it for other types of respondents. This strategy can help ensure comparability across different types of respondents. Piloting the surveys can prevent costly and timely problems with the questionnaires. Even experienced survey developers often find that questions that make sense in the office do not make sense or collect the right types of data in the field. For this reason, it is of paramount importance to first adequately design questionnaires in advance, then field test them, and finally address any issues that may arise. While a pilot of a survey can add an extra week or more to the PETS timeline, it can also identify problems with the survey that could ruin the entire study if not caught early on. Note that the research team should field test its survey at each level--government, facility, and individual--to be sure that modifications to the survey are made, if needed, and that the questions they have designed are indeed returning unbiased and viable responses. Ideally, a pilot should not be conducted with units (facilities) that will ultimately be included in the study. Additionally, it is essential that key researchers participate in the pilot so they are aware of how questions are being asked and how individuals respond. Field testing, like training, can be a time consuming process. However, bad data will unravel good analysis and make any type of policy rec- ommendation infeasible. Preparing to Follow the Money ­ Major Take-Aways · For small sample sizes, it is important to limit the scope of the pop- ulation with respect to geographic coverage and heterogeneity of the facilities. · Facility surveys should be clear, should not include extraneous questions, and should be tested in the field before full PETS imple- mentation. A Guidebook 33 Step 2. Collecting and Analyzing the Data The first three steps outlined in this guidebook have dealt with prepar- ing for PETS implementation. Research teams are well advised to put significant time and effort into these earlier steps; carefully articulating objectives, developing a strong work plan, and crafting a clear set of in- struments will make the step of collecting and analyzing the data simpler and more successful. Once the research team has completed the prepara- tory steps, it can begin implementing the PETS itself, a step that involves several sub-steps including training the survey team, facility visits, enter- ing and cleaning the data, and data analysis. Recruiting and Training a Survey Team The first step in data collection is identifying a team of enumerators who understand the basics of the project and, preferably, have some experience with surveying. Many research teams face obstacles in ad- ministering the survey instruments themselves, including language and cultural barriers (often faced by external research organizations such as the World Bank) and small team size (a common obstacle faced by civil society organizations). As such, recruiting and training a team of tem- porary staff to serve as enumerators can be a cost-effective means of administering the survey. The quality of the data collection team will directly impact the quality of data; therefore, teams should carefully consider whom they hire to con- duct data collection and should undertake some spot-checking during the data collection process to ensure comparability and good standards. Once the research team has identified and selected enumerators, the next step is to adequately train them. Training on the survey instru- ments can be a timely process, but its importance cannot be overstated. If enumerators are unfamiliar with the questions they are asking, the data they will be collecting faces the possibility of being unusable. If possible, enumerators should be involved in the piloting of the survey; this allows 34 Using PETS to Monitor Projects and Small-Scale Programs the research team to observe the enumerators before the survey is actually implemented. Entering and Cleaning the Data While some research teams wait until completing the survey to enter the data into their statistical package of choice, there is a case to be made for entering data as they are collected. Although piloting a survey should uncover many of the potential problems with an instrument, additional problems with coding answers can arise when entering the data. Entering data can also uncover trends such as answer choices that should be added to the survey. Consider for example the following scenario: A research team is interested in how teachers deal with shortages of school supplies. In the teacher survey, one of the questions is "How do you deal with a shortage of textbooks?" and the answer choices are "Have students share textbooks," "Ask parents to pay for addi- tional textbooks," "Randomly distribute textbooks," and "Other." After all of the data has been collected, the research team enters the data and finds that 80 percent of teachers respond "Other." By entering the data as they came in, the research team could have uncov- ered after the first day of data collection that most teachers were answering "Other" and could have added alternative answer choices before collecting more responses. The enumerators would then only have to return to the first day's respondents to ask the additional question. After entering the data, the research team can then begin sorting and cleaning the data. This involves creating summary statistics and removing any outlying data that could skew the analytical results. Double data-entry is another method commonly used to ensure good quality data. In this case, data is entered once, and then a special program is used that allows you to enter the data again and checks each entry against the first. If there is a discrepancy, the program alerts the user and corrections can be made. A Guidebook 35 Data Analysis The analysis phase is arguably the most important, yet often takes far less time than the actual data collection and preparation. Analysis includes the identification of leakages of funds, delays (where did money get held up?) and areas where problems with priorities may exist in the expenditure chain. Leakages can be identified at each stage of the expenditure chain by comparing the amount disbursed by the higher link in the chain and the amount received by the lower link in the chain. In some cases, the research team may not be able to identify the exact source of leakage. However, the team should be able to at least find evidence of leakage. Take for example the case of a district education office (DEO) that pro- vides funding for textbooks for the five primary schools in the district. If the district office keeps disaggregated budget records (Table 5), the research team should be able to identify the level of leakage for each school. Alternatively, if the district offices keep only aggregated budget records (Table 6), the research team could estimate the overall level of leakage between the office and schools but would be unable to calculate per-school leakage. Table 5. Disaggregated Education Data DEO Amt. School Amt. School Disbursed Received Leakage 1 2,000 1,800 200 2 2,000 1,500 500 3 2,000 2,000 0 4 2,000 2,000 0 5 2,000 1,600 400 TOTAL 1,100 36 Using PETS to Monitor Projects and Small-Scale Programs Table 6. Aggregated Education Data DEO Amt. School Amt. School Disbursed Received Leakage 1 1,800 2 1,500 3 2,000 4 2,000 5 1,600 TOTAL 10,000 8,900 1,100 In addition to calculating leakages, the research team can also calculate the incidence of inefficiencies, such as delays in the arrival of funds or in- kind resources. It may also be possible to calculate correlations between inefficiencies and characteristics such as the location of the facility, char- acteristics of the facility proper, and other variables. While many analyses can be conducted, the research team should revisit its research objectives and overall project objectives when deciding how much time to spend on data analysis. PETS results can be interesting as stand-alone research results; however, implementers are well advised to remember that results can be more effective if they are used to identify issues, recommend solu- tions, and advocate for change. Data Collection and Analysis ­ Major Take-Aways · Well-trained enumerators and data entry personnel are essential to ensuring good data quality. · Good quality data are the key to high quality, robust analysis. Step 3. Identifying the Issues The preceding steps have provided the framework needed to identify and answer research questions regarding the efficiency of public spending and service delivery. However, in order to utilize research results to improve A Guidebook 37 policies, spending, and human development outcomes, the research team must translate numbers into actual answers. In particular, there are three types of questions that the quantitative results, combined with supporting qualitative evidence, can be used to answer: What are the problems? For example, leakage levels or delays in fund disbursement? What are the trends in problem occurrence? These are the "who," "where," and "when" of the problems. Where is leakage higher, rural schools or urban schools? When are school meals most likely to be delayed in ar- riving at schools, early in the school year or later? Why are these problems occurring? It can be difficult to answer this question, especially with quantitative data exclusively. However, in- sight into the reasons for inefficiencies such as leakage and delays can come from qualitative tools such as focus group discussions and key informant interviews. Researchers can choose to first conduct quanti- tative analysis, followed up by qualitative work designed to contextual- ize the quantitative results. Or, alternatively, researchers may start with qualitative tools, which can then inform the questionnaire, making the quantitative work more effective. Regardless of the order, a research team would ideally collect both types of evidence. 38 Using PETS to Monitor Projects and Small-Scale Programs Box 5. Common Qualitative Tools to Supplement PETS Findings Focus Group Discussions (FGDs) PETS implementers frequently incorporate focus group discussions with frontline service providers and other stakeholders to obtain infor- mation about the consequences and responses to problems that the PETS uncovers. In the case of the Center for Democratic Development, focus group discussions with representatives from parent teacher as- sociations illuminated the reactions and experiences of parents with regard to absenteeism as well as verifying some of the information provided by teachers and head teachers about school management and financing. Key Informant Interviews In cases where the research team is interested in getting the insights of a small number of major stakeholders, key informant interviews can be a valuable tool. Many implementers will conduct interviews with key government officials or opinion leaders in the community af- ter a preliminary analysis of the data has revealed certain issues with public spending and service delivery. Researchers can then ask stake- holders a small number of targeted questions to reveal why these is- sues may be occurring. Exit Surveys Although the PETS methodology is developed to follow money to the "ground," many PETS only follow money to the frontline service provider. However, stopping at this level misses a critical link in the chain: service users. Although service users often do not handle public spending (such as salaries), these stakeholders can provide valuable insight into the efficiency of public spending. Expenditure tracking ex- ercises targeting health facilities in particular often employ exit sur- veys for clients leaving a government-run facility. Users can be asked targeted questions regarding their perceptions: Was a health center well-stocked? How long did they have to wait to see a doctor? The results of data analysis can be used to identify (1) delays and/or leak- age and (2) trends in these issues. Once leakage or other inefficiencies have A Guidebook 39 been identified and the research team can point out some of the situations in which the problems are most prevalent, the implementers then must face the "so what" issue. While numbers are interesting, they also leave policymakers and beneficiaries alike with the two obvious questions:"Why are these problems occurring?" and "How can we solve them?" (a question investigated further in the next section). Combining quantitative results from teacher surveys with data gleaned from qualitative techniques such as focus group discussions, the research team can begin pinpointing possible causes of leakage and reasons for the trends uncovered. Particularly for small scale PETS, quantitative methods generally identify the problems but fail to explain the underlying issues. However, qualita- tive methods alone do not provide the statistical evidence that decision- makers seek before they are willing to move away from the status quo. CSOs in particular can bolster their evidence from standard PETS meth- ods by incorporating into their research design qualitative tools such as those presented in Table 4. By utilizing qualitative approaches in addition to standard PETS methods, CSOs can be better positioned to dissemi- nate their research results as well as ensure that the results reach not only service users, but also frontline service providers. Step 4. Recommending Solutions The final question that PETS must tackle is the question of how to rem- edy the problems uncovered in the previous four steps. Large-scale PETS take on the difficult task of evaluating the effectiveness of public spend- ing for an entire sector, and the recommendations that come out of the studies often reflect this. Without focusing on specific components of education spending for example, project implementers generally identify system-wide problems and recommend solutions that reform the wider public expenditure management system. While these recommendations are often valuable, they are not always as attractive to policymakers, es- pecially when considering contextual political constraints. On the other hand, a PETS that focuses on a pharmaceutical procurement program 40 Using PETS to Monitor Projects and Small-Scale Programs can identify very specific problems and develop tangible and frequently inexpensive solutions to these problems. When presented with this type of recommendation by a credible research team, policymakers are much more likely to be amenable to implementing the proposed solution. From the issues identified, PETS implementers can now develop specific recommendations. This part of the process is project-specific however. For CSOs conducting these studies, it is important to keep a few things in mind. Civil society organizations can be well-positioned (depending on country context); these organizations can target recommendations to specific audiences and take advantage of their local status to influence policy at a level that large organizations may not be able to accomplish. Recommendations can be targeted not only at the central Government and ministries, but also at local government, service users, service provid- ers, and stakeholders-at-large. Research teams and CSOs should target their recommendations according to the audience. In an effort to not only disseminate the PETS evidence but also recommend solutions to identi- fied problems, different strategies can be employed, ranging from press- releases to media campaigns. Recommendations do not necessarily have to be exclusively policy-based; some of the best recommendations can call for action related to specifically identified issues of leakage or delays, for example. Finally, CSOs and research teams can advantageously highlight qualitative findings to examine the underlying issues of leakages, delays, problems of priorities and service delivery issues. The best recommendations are: Feasible ­ They have taken into account the specific context of the project or program the PETS is targeting and are in the realm of pos- sibility given budget, time, human resources and political constraints. Concrete ­ They are concise and specific, while also being targeted and understandable to both policymakers and stakeholders. Inexpensive ­ They contrast the expense of a remedy with the benefit of implementing it (i.e., a recommendation may be expensive, but it could represent an inexpensive alternative to the amount of money be- A Guidebook 41 ing lost). Presenting recommendations based on cost-effectiveness can be a particularly powerful tool in exacting policy change. Box 6. The Value of Actionable Recommendations Of the many public expenditure analyses that have been conducted over the past several decades, many have proposed major over-arch- ing reforms of the entire financial and budgeting system. While such reforms might be effective if implemented, they are also expensive in terms of time and money. As such, they are inherently unattract- ive to policymakers who are dealing with limited resources and, in some cases, short terms in office. Further, sometimes a tangible and easily-implementable recommendation can be far more effective in improving public spending. The Institute for Policy Analysis and Re- search (IPAR) in Kenya found this to be true in its 2008 study of the efficiency of the Secondary Education Bursary Scheme. Triangulating data from the constituency bursary fund committees and secondary schools in Nairobi province, the organization uncovered significant in- efficiencies with the Secondary Education Bursary Fund. For example, it was discovered that 20 percent of schools were receiving bursaries for students no longer enrolled. In 27 percent of schools, students were receiving multiple bursaries totaling more than the school fees. This inefficiency was largely due to lack of communication between the government-run bursary scheme and private schemes in Kenya. Rather than recommend that the government invest more money into the program, the organization made the recommendation that all bur- sary funds (government and private) use standardized records, mak- ing it easier to compare recipients from the different funds. Further, IPAR made the low-cost recommendation that bursary funds share in- formation on proposed beneficiaries to avoid double or triple bursary allocations to one student. Step 5. Dissemination and Advocacy While many rigorous and statistically valid PETS studies have been com- pleted, many have failed to have any impact on the efficiency of public 42 Using PETS to Monitor Projects and Small-Scale Programs spending in education, health, and other social sectors. One possible rea- son for this failure is that many PETS implementers have overlooked the final and critical step: disseminating findings and advocating for change. Instead of looking at PETS as a means to an end, they have looked at PETS as an isolated research project. Other implementers do attempt to disseminate findings from the PETS but approach dissemination as an afterthought. Rather than beginning the study with a plan for how they will use the results to push for positive change, they only bring policy- makers and stakeholders into the discussion after they have designed the study, implemented the PETS, and compiled results. While a communications strategy will not be the same for a think tank as for a grassroots advocacy organization, there are several best practices that should be followed by all PETS implementers: Identify target audiences before starting the PETS. This issue was cov- ered in Step 0 of the guidebook, however its importance cannot be overstated. All PETS and related studies should be undertaken with the goals of (1) identifying problems and potential solutions and (2) changing the actions and decisions of stakeholders to improve the functioning of budgeting in the target sector or program. Both of these goals are better achieved if research teams design the PETS and the communications plan simultaneously. First, identifying stakeholders to whom the organization wants to communicate results allows the team to better design the study and ascertain what potential problems are of the most interest to these stakeholders. Second, stakeholders are more likely to be receptive to recommendations and advocacy if they have some amount of ownership in the PETS. The best way for an organi- zation to create this sense of ownership is to engage with stakeholders from the start of the project. Develop a dissemination and advocacy plan before starting the PETS. In addition to identifying target audiences before beginning the study, it is important to develop a communications plan before implementing the PETS. Knowing what its researchers want to do with the results of a PETS can help in designing survey instruments, sample strategy, and A Guidebook 43 collection of supporting qualitative data. Further, PETS data are most impactful immediately after they have been compiled and analyzed; having a communications plan before starting the study allows teams to immediately dive into their advocacy and dissemination work once they have preliminary results. Think about the impact you hope to have rather than the activities you want to undertake. Communications plans can fail because the peo- ple who develop them take it for granted that activities such as press conferences and policy briefs will raise awareness and spark dialogue about public expenditure topics. While well-implemented meetings and publications can affect policy and policymaker actions, these strat- egies are not guaranteed to have such an effect. Rather than begin with the communications activities or products, research teams and CSOs should start with their advocacy goals. Is the goal to change policies in an effort to lower leakage in a textbook financing program? Is it to im- prove the ability of school administrators to record the allocation and receipt of textbook funds? Or is it to develop ways for students to track how many textbooks are getting to their classrooms and compare their findings with students in nearby schools? By determining the goals of the PETS first, organizations are better equipped to determine the best means to achieve the desired impact. Keeping these best practices in mind, the research team can undertake the three stages of dissemination and advocacy: planning, evidence-based action, and evaluation.15 Planning the Dissemination and Advocacy Strategy The most successful CSOs and research teams begin by determining the goals of their study and related advocacy. Goals for PETS generally consist of changing the behavior or actions of participants in the pub- lic expenditure process to improve the efficiency of spending and service delivery. For example, take the case of a CSO tracking spending on common 15 Sandilya, Indira. Presentation. Delhi, India. July 2009. 44 Using PETS to Monitor Projects and Small-Scale Programs pharmaceuticals destined for primary health centers in one district in In- dia. The ultimate goal of the CSO is to improve the health of Indians in this district, and the intermediate goal is to improve the efficiency of the purchase and allocation of pharmaceuticals in the district. Project-specific goals, however, could take many forms. Examples include (but are not lim- ited to): To increase government oversight of the distribution of pharmaceuticals To improve the system used by facilities in reporting their pharmaceu- tical needs to district offices To provide a way for service users to report problems in accessing pharmaceuticals to government officials To develop a method for doctors and pharmacists to track whether all of the drugs that were allocated to their facilities actually arrived Box 7. Overcoming the Challenge of the Lack of Public Interest in Budgeting The budget is one of the most important tools that a government has to impact the lives of its constituents. However, corruption and inef- ficiencies in the budget sector do not tend to get as much attention or generate as much as interest as other development problems. Civil society organizations are well-positioned to increase public interest in budget problems by using creative dissemination methods and tools. An increase in public interest can, in turn, boost participation in the budget process and improve transparency. The Centre for Budget and Policy Studies (CBPS) in India offers an example. CBPS often uses films about governance and public expenditure management to introduce people to the ways in which budgeting has a direct impact on their daily lives. Methods and tools such as cartoons, posters, and videos can all be used to increase interest in budgeting. Each of these goals targets a different audience and requires different strategies and tools for ensuring the desired change. And each of these goals may require the research team to partner with different actors, in- A Guidebook 45 cluding the media and other civil society organizations. It is important for teams to consider these issues early in the process of implementing the PETS. For each goal, the organization should answer the following questions: Who is/are the target audience(s)? The efficiency of public spending and service delivery is in the hands of many different actors, and thus im- proving the impact of spending and service delivery can be achieved by changing the actions and decisions of an array of people. The most effective communications strategies will be tailored to the agents they are trying to affect. In general, though, results should always be brought back and presented to study participants, particularly frontline service providers such as teachers and medical professionals that were part of the study sample, as well as to the communities that they serve. What are the best products and/or activities to develop in order to achieve this goal? While a press conference may be a good way to influence the dialogue between policymakers, it is probably not a good way to reach citizens in rural communities. Likewise, posters in village health clinics may not be an effective way to carry messages to government officials. Rather than choosing the traditional communications and advocacy products, CSOs and research teams should think critically and creatively about what tools will catch the attention of their target audience and drive home their messages. Who are the strategic partners for these products and activities? While CSOs have the advantage of being based in-country, they may not always have eyes and ears in the schools, clinics, or government offices that they are trying to improve. Rather than expect to carry out an entire communications strategy on their own, CSOs should choose their partners with a view to maximizing the impact of their work. Potential partners might include the media, other CSOs, think tanks, service providers, etc. How will the research team measure success in achieving this goal? If a CSO or research team hopes to have a sustainable impact on public spending and service delivery, it needs to devise a means of measuring the effectiveness of its advocacy and communications strategies. 46 Using PETS to Monitor Projects and Small-Scale Programs Who will the CSO or research team partner with to achieve results? In addition to directing activities toward key stakeholders, many CSOs find it valuable to partner with key stakeholders in undertaking their dissemination and advocacy activities. Most commonly, CSOs partner with other civil society organizations to disseminate findings. How- ever, some research teams are able to find progressive individuals in government who gain an interest in the study and are willing to push the recommendations and results among their government colleagues (see Box 8). Undertaking Evidence-Based Action After developing the communications plan and compiling the results of the study to use in advocacy work, organizations and research teams actu- ally undertake actions and produce products to encourage the changes that they hope to see in social sector spending and service delivery. There is no one method or product that is best to undertake; a good action plan will be tailored to the audience and goals of the project. Rather than at- tempt to develop an exhaustive list of communications tools, we highlight in Annex D some of the tools and strategies developed by past grantees of the Transparency and Accountability Program. Evaluation of Communications and Advocacy Work A final step is needed to answer the "so what" question and to ascertain whether a PETS has had any policy impact. An effective way to improve the likelihood of having an impact with evidence-based action related to a PETS is to critically evaluate the success of each advocacy tool and activity. Evaluating communications work is not a simple activity, but it can be very useful in determining whether to continue engaging and spending money on a particular strategy or to test a new approach. For each activity, the team should have not only a goal that it is trying to achieve, but also a way to measure its success in achieving that goal. For example, a re- A Guidebook 47 Box 8. Building Government Ownership for Results For obvious reasons, policymakers (particularly those who are cor- rupt) have a disincentive to provide data and support to research teams conducting PETS. Even those who are not involved in the leak- age of funds may be concerned that a study of public expenditure could uncover inefficiencies or problems in the system that could in- crease public scrutiny of those policymakers. While this is a common challenge, civil society organizations have come up with creative ways of dealing with such resistance from policymakers. One example is the Center for the Implementation of Public Policies Promoting Equity and Growth (CIPPEC) in Argentina. With a long track record of producing high-quality policy research, CIPPEC's work is fre- quently cited and used by the Government in policymaking decisions. However, as an independent think tank, CIPPEC recognizes that it has an obligation to publicize findings that may in fact point out flaws in the Government, an obligation that could make the Government less inclined to consider its policy research and recommendations. One strategy that CIPPEC uses to deal with this conflict is to provide poli- cymakers with first access to study results before going public with them. CIPPEC still goes to the media with any results (positive or nega- tive), but fosters a constructive relationship with officials by prevent- ing them from being blind-sided by the results. Another example is the Centre for Budget and Policy Studies (CBPS) in India. Many politicians are not economists, accountants, or people with financial or budgeting experience. As such, CBPS produces products that inform not only the general public, but also government officials. By providing training and guides to policymakers on budget practices and processes, CBPS raises its visibility with government officials. search team with the goal of improving the ability of parent associations to monitor textbook procurement may choose to develop a training guide for parent associations on monitoring delays. The team could measure the success in achieving this goal in a number of ways, for example by counting 48 Using PETS to Monitor Projects and Small-Scale Programs the number of schools that request copies of the training guide or the number of complaints to government officials about lack of textbooks from parent associations that have access to the guide. Information such as this can be useful for a research team or organization that is deciding whether to increase its production of the guide, develop a new guide for measuring textbook procurement delays, or abandon this tool in favor of a new advocacy strategy. Disseminate and Advocate ­ Major Take-Aways · Develop an advocacy and communications plan before beginning your project, and involve target audiences from the outset · Think critically and creatively about what advocacy tools and strat- egies are most likely to help you reach your impact goals · Develop ways to measure your communication strategy's impact · Always bring your results back to frontline service providers and others that participated in your study · Think about how you can make the impact of your short-term study sustainable by developing tools and policies that can be adopted by people outside of the research team Conclusions When designed with the goal of looking toward solutions, PETS can be a particularly valuable tool. The objective of the research should go beyond the simple identification of a problem; its aim should be to understand potential causes and recommend viable, long-term, and sustainable solutions. This is especially true when an effort is made to design a research plan that includes targeted efforts at long-term results dissemination to government officials, service providers, service users, and community-based stakeholders. PETS are most successful if they fit the structure and characteristics of the organizations conducting them; civil society's comparative ad- vantage is not the same as that of large international organizations like A Guidebook 49 the World Bank. Rather than take on an analysis of a country's entire education sector, small CSOs can make large gains by narrowing the analytical focus, or by leveraging their own resources with those of other analytically strong organizations. Most importantly, small-scale analytical PETS can be just as effective in promoting policy change as large-scale national surveys. Prospective planning of research, from the survey design to final re- sults dissemination plans is essential. Understanding the big picture, and working with stakeholders--government, service providers, ser- vice users, and community leaders--will help to inform the process and increase the likelihood that the research results inform policy and improve sector service delivery and governance. Quantitative analysis associated with PETS is only a starting point. Incorporating qualitative work such as focus group discussions and targeted interviews can increase the robustness of research results. Ideally, research teams can incorporate qualitative work into survey design, as well as follow data analysis to better understand the results of the research. However, research teams and civil society organiza- tions conducting this type of research must be realistic about what PETS can and cannot do; PETS can tell us that pharmaceuticals are unavailable at a health facility, or that some percentage of a capitation grant is not reaching schools, but they cannot tell us conclusively why student outcomes are low or explain why health indicators are below average. 50 Annex A. World Bank PETS Using PETS to Monitor Projects and Small-Scale Programs Survey Year of Resource Flow Country Type Report Sector(s) Monitored Sample Size Albania PETS 2004 Health Different kinds of The scope of the study expenditure (salaries, includes the TRHA region, supplies, administrative Dürres, and a "control" costs, etc.) municipality outside of Tirana and Dürres. http://www-wbweb.worldbank.org/prem/prmps/publicsector/PETS%20Studies/Albania/Albania%20Health%20PETS.pdf Albania PETS 2004 Education Block grants, funds 20 municipalities and communes, and 201 schools http://www-wbweb.worldbank.org/prem/prmps/publicsector/PETS%20Studies/Albania/Albania%20PETS%20Education.pdf Brazil PETS 2006 Health Six states, 17 municipalities in those six states, and 49 hospitals and 20 outpatient units in the sampled municipalities http://intresources.worldbank.org/PUBLICSECTORANDGOVERNANCE/Resources/BrazilPETs.pdf Cambodia PETS Health Expenditures in 17 Provincial Health government health Departments (PHD), 30 facilities Operational Districts (OD), 200 Health Centers (HC), and 29 Referral Hospitals (RH) http://intresources.worldbank.org/PUBLICSECTORANDGOVERNANCE/Resources/CambodiaHealthPETSJan15.pdf Continued on next page Annex A. World Bank PETS (Continued) Survey Year of Resource Flow Country Type Report Sector(s) Monitored Sample Size Cambodia PETS 2005 Education School funding Two hundred schools in seven provinces http://www-wbweb.worldbank.org/prem/prmps/publicsector/Pets%20studies/Cambodia-education%20pets%20final.pdf Cambodia PETS 2008 Health Public health sector 17 Provincial Health expenditures Departments, 30 Operational Districts, 200 Health Centers and Health Posts, 29 Referral Hospitals Cambodia Health PETS: Public Expenditure Tracking Report. East Asia and Pacific Region. World Bank. Feb. 2008. Chad PETS 2010 Education Report forthcoming Honduras PER 2001 Poverty Government spending reduction and fiscal sustainability http://www-wbweb.worldbank.org/prem/prmps/publicsector/PETS%20Studies/Honduras/Honduras.pdf A Guidebook Continued on next page 51 52 Annex A. World Bank PETS (Continued) Using PETS to Monitor Projects and Small-Scale Programs Survey Year of Resource Flow Country Type Report Sector(s) Monitored Sample Size Indonesia QSDS 1998 Education To gather information 600 primary and junior about the magnitude of secondary schools in five the effect of the crisis provinces on education and on whom the impact was strongest as part of a larger effort to monitor the impact of the crisis on social sectors http://intresources.worldbank.org/PUBLICSECTORANDGOVERNANCE/Resources/survey5.pdf Indonesia QSDS 2001 Education 600 primary and junior secondary schools in five provinces http://intresources.worldbank.org/PUBLICSECTORANDGOVERNANCE/Resources/600schoolsurveyreport.pdf Lao PDR PETS 2008 Education and Financial flows such as 17 provinces, 54 districts, and Health salary payments 252 villages http://intresources.worldbank.org/PUBLICSECTORANDGOVERNANCE/ Resources/285741-1239741640129/6023127-1240595002422/LaoPDR.pdf Continued on next page Annex A. World Bank PETS (Continued) Survey Year of Resource Flow Country Type Report Sector(s) Monitored Sample Size Madagascar PETS, QSDS 2003 Education IPPTE, CRESED First survey: 36 districts and 326 primary schools. Second survey: 185 primary schools were surveyed, about half of the schools included in the first survey. Third survey: 24 district administrations (out of 111) Gauthier, Bernard. PETS-QSDS in Sub-Saharan Africa: A Stocktaking Study. 2006. Mongolia PETS 2006 Education 118 schools in urban and rural areas http://www-wbweb.worldbank.org/prem/prmps/publicsector/Pets%20Studies/Mongolia/MongoliaSept2006.pdf Mozambique PETS, QSDS 2004 Health 11 Provinces (out of 11), 35 Districts, 90 Public primary health centers, 167 workers, 679 patients Gauthier, Bernard. PETS-QSDS in Sub-Saharan Africa: A Stocktaking Study. 2006. Continued on next page A Guidebook 53 54 Annex A. World Bank PETS (Continued) Using PETS to Monitor Projects and Small-Scale Programs Survey Year of Resource Flow Country Type Report Sector(s) Monitored Sample Size Niger PETS 2008 Education and Selected government See page 7 of report Health expenditures in education (textbooks, notebooks and drawing books for students) and in health (food expenditure and hospital supplies, and essential medicines) http://intresources.worldbank.org/PUBLICSECTORANDGOVERNANCE/ Resources/285741-1239741640129/6023127-1240595002422/NigerVol1.pdf Nigeria PETS, QSDS 2004 Health 30 local government, 252 health facilities, 700 staff Gauthier, Bernard. PETS-QSDS in Sub-Saharan Africa: A Stocktaking Study. 2006. Papua New Guinea PESD 2004 Public School facilities 214 schools in 19 districts spending and environment, across 8 provinces and service financial flow of delivery resources, and teachers, issues, mainly administration and education students. sector http://www-wbweb.worldbank.org/prem/prmps/publicsector/PETS%20studies/PNG/PNG.pdf Continued on next page Annex A. World Bank PETS (Continued) Survey Year of Resource Flow Country Type Report Sector(s) Monitored Sample Size Tajikistan PETS 2008 Health Government 66 rayons and 12 cities expenditures http://intresources.worldbank.org/PUBLICSECTORANDGOVERNANCE/ Resources/285741-1239741640129/6023127-1240595002422/TajikistanPPER.pdf Uganda PETS 1996 Education and Education: capitation Education: 18 districts (out of Health grants, Health: non- 39), 250 public primary schools. wage expenditures Health: 19 districts, 100 health facilities Gauthier, Bernard. PETS-QSDS in Sub-Saharan Africa: A Stocktaking Study. 2006. Uganda QSDS 2003 Health Financing, drugs, 10 District administrations vaccines, and supplies (out of 45), 155 (public, private for profit and non profit) health facilities, 1617 patients Gauthier, Bernard. PETS-QSDS in Sub-Saharan Africa: A Stocktaking Study. 2006. Yemen QSDS 2006 Education Three complementary 16 schools in 12 districts studies examine from three governorates and budget practices, another in 240 schools in four teacher and school governorates A Guidebook resource management issues and teacher absenteeism. http://www-wbweb.worldbank.org/prem/prmps/publicsector/PETS%20Studies/Yemen/Yemen%20Final%20Report%20Edited.pdf Continued on next page 55 56 Annex A. World Bank PETS (Continued) Using PETS to Monitor Projects and Small-Scale Programs Survey Year of Resource Flow Country Type Report Sector(s) Monitored Sample Size Yemen PETS 2005 Education Ghabayen, Rafat. "Public Expenditure Tracking Survey: Education Sector ­ Republic of Yemen" Dec 2005. Zambia PETS, QSDS 2004 Education Non-wage funding 33 districts, 182 primary for basic education schools (grades 1-9) (fixed-school grant, discretionary non-wage grant program) Gauthier, Bernard. PETS-QSDS in Sub-Saharan Africa: A Stocktaking Study. 2006. Zambia PETS 2009 Health Forthcoming Annex B. CSO-led PETS** Survey Year of Resource Flow Country Organization Type Report Sector(s) Monitored Sample Size Albania* 2A Consortium PETS 2008 Health Budgeting system that 47 health care facilities (31 funds Albanian health primary health care centers, care 6 hospitals, and 10 local government units) http://tap.resultsfordevelopment.org/resources/improving-public-expenditure-effectiveness-health-sector Bangladesh* Concern for QSDS 2010 Health Expenditures of 26 Drop-In-Centers (DIC) in Environmental government HIV/AIDS Bangladesh Development and prevention program in Research Bangladesh Expected June 2010 Ghana* Center for PETS 2010 Education Disbursement and 30 public basic primary Democratic usage of the Capitation schools in Ghana Development Grant in public primary schools Expected June 2010 Continued on next page A Guidebook 57 58 Annex B. CSO-led PETS** (Continued) Using PETS to Monitor Projects and Small-Scale Programs Survey Year of Resource Flow Country Organization Type Report Sector(s) Monitored Sample Size Guatemala* National Center PETS 2008 Education Expenditures for 41 schools--20 urban and 21 for Economic government programs rural--around Guatemala Research aimed at getting City poorer students to go to and stay in school, specifically six programs intended to lower the cost and increase the quality of education--by providing scholarships, free school meals, milk, textbooks, and supplies, and teaching kits http://tap.resultsfordevelopment.org/resources/expenditure-tracking-improve-effectiveness-public-education-guatemala India* Indo-Dutch PETS 2008 Health Flow of resources 30 rural Primary Health Project to selected Primary Centers in two Karnatakan Management Health Centres districts, Chamarajanagar Society (CR Nagar) and Bellary http://tap.resultsfordevelopment.org/resources/following-health-delivery-trail India* Public Health PETS 2010 Health Flow of funding 60 facilities in 32 districts Foundation of for drug financing, (two states) India procurement and distribution Continued on next page Annex B. CSO-led PETS** (Continued) Survey Year of Resource Flow Country Organization Type Report Sector(s) Monitored Sample Size Expected June 2010 Indonesia* Centre for PETS 2008, Education 2008: Seven financing 2008: 38 elementary and Regional 2010 schemes that secondary primary schools Information and fund Indonesia's in two Indonesian regions: Studies decentralized Serang Regency, in Central education system-- Java, and Gresik Regency, three that fund in East Java. 2010: 30 junior operating expenses high schools (public school) and four that fund in Tangerang Regency, investments. 2010: Tangerang City, and Tracking schools' real Magelang Regency expenditures of BOS Funding http://tap.resultsfordevelopment.org/resources/education-budget-efficiently-spent Indonesia* Bandung PETS 2010 Education School Operational Aid 30 schools in Bandung City Institute of (BOS) Program (urban) and Kabupaten Governance Bandung Barat (rural) Studies Expected June 2010 Continued on next page A Guidebook 59 60 Annex B. CSO-led PETS** (Continued) Using PETS to Monitor Projects and Small-Scale Programs Survey Year of Resource Flow Country Organization Type Report Sector(s) Monitored Sample Size Kenya* Institute for PETS 2008, Education Administration of The 2008 study included Policy Analysis 2010 Kenya's Secondary 48 schools in all eight and Research Education Bursary constituencies (districts) Scheme in Nairobi: Makadara, Embakasi, Kamukunji, Starehe, Westlands, Lang'ata, Dagoretti and Kasarani. The 2010 study surveyed 202 schools, in the remaining 202 constituencies in Kenya. http://tap.resultsfordevelopment.org/resources/expenditure-tracking-secondary-education-bursary-scheme-kenya Malawi Civil Society PETS Ongoing Education Education Funding 500 schools (10% of the Coalition for total number of schools), as Quality Basic well as district assemblies, Education district education offices, education supplies units, and teacher training colleges. Information on this study available at http://www.internationalbudget.org/resources/expenditure/index.htm Malawi* Community PETS 2010 Education Teacher salary 300 teachers--100 teachers Based Monitoring administration--flow from each of the regions of Project of funds from Treasury Malawi to Primary Schools Continued on next page Annex B. CSO-led PETS** (Continued) Survey Year of Resource Flow Country Organization Type Report Sector(s) Monitored Sample Size Expected June 2010 Moldova* Institute for PETS 2008 Education Public budgetary 30 schools, selected on Development and spending for the pre- the basis of: geographic Social Initiatives university schools location, size, type of institution (primary, gymnasia, lyceum) http://tap.resultsfordevelopment.org/resources/targeting-education-policy-review-decentralized-public-spending Mongolia Open Society PETS 2004 Education Mongolia general 2005 study with WB and Forum education funding, OSF: 118 schools from divided into two 21 aimags and UB--1416 separate funds- respondents--were sampled operational expenses (salaries) based on number of students, and heating, electricity, water... based on past expenditures at the school level. Information on this study available from Ramkumar, "Civil Society Experiences in Monitoring Government Budgets" and http://www.opensocietyforum.mn/en/index.php?sel=news&menu_id=229&obj_id=690 A Guidebook Continued on next page 61 62 Annex B. CSO-led PETS** (Continued) Using PETS to Monitor Projects and Small-Scale Programs Survey Year of Resource Flow Country Organization Type Report Sector(s) Monitored Sample Size Nigeria* ABANTU for PETS 2010 Education Tracking the flow 30 schools were selected Development of education funds from the three senatorial from states to local Districts--one urban, two governments (LGAs) rural. and from LGAs to schools. Expected June 2010 Pakistan* Institute of PETS 2010 Education Tracking budgetary 40 primary and secondary Social and Policy allocations and schools in two districts, as Sciences disbursements at well as a beneficiary survey district and school of 600 households. level Expected June 2010 Paraguay* Centro de PETS 2008 Education Public funding for 30 primary schools in the Análisis y primary schools Central region of Paraguay Difusión de la Economia Paraguaya http://tap.resultsfordevelopment.org/resources/education-expenditures-budget-tracking-analysis-thirty-paraguayan- educational-institutions Continued on next page Annex B. CSO-led PETS** (Continued) Survey Year of Resource Flow Country Organization Type Report Sector(s) Monitored Sample Size Peru* Research Center PETS 2008 Health Budgeting for health 14 health facilities, with of the University sector in Metropolitan surveys administered to 100 of the Pacific Lima--From Ministry health personnel. of Economy, to the Ministry of Health and intermediate level agencies and then to frontline providers. Specifically focuses on National Strategies for Vaccination and the Tuberculosis Prevention and Control http://tap.resultsfordevelopment.org/resources/route-expenditure-and-decision-making-health-sector Romania* Romanian PETS 2008 Education Public funding 32 schools in four counties Academic Society for education in of Romania (randomly decentralized counties selected from eight counties participating in a pilot decentralization strategy by the government of Romania) http://tap.resultsfordevelopment.org/resources/effectiveness-and-efficiency-public-expenditure-romanian-education-system A Guidebook Continued on next page 63 64 Annex B. CSO-led PETS** (Continued) Using PETS to Monitor Projects and Small-Scale Programs Survey Year of Resource Flow Country Organization Type Report Sector(s) Monitored Sample Size South Public Service QSDS 2010 Education Tracking funds for 30 public primary schools Africa* Accountability the School Nutrition within the Grahamstown Monitor Programme (SNP) Education District in the Eastern Cape Province of South Africa in the Grahamstown education district Expected June 2010 Tanzania Research PETS 2001, Health and 2001 Study monitored Unknown on Poverty 2004 Education discretionary funds Alleviation of local councils. 2004 study monitored funds earmarked for primary schools, comprised of a capitation grant for recurring costs and development grant for other costs such as textbooks, school maintenance and teacher training Information on this study available from Ramkumar, "Civil Society Experiences in Monitoring Government Budgets" Continued on next page Annex B. CSO-led PETS** (Continued) Survey Year of Resource Flow Country Organization Type Report Sector(s) Monitored Sample Size Uganda* Coalition for PETS 2010 Health Tracking funds for 41 health centers were Health Promotion essential medicines randomly sampled from the and Social from the Ministry of two districts--one urban, Development Finance to front line one rural service providers Expected June 2010 Zambia Zambia National PETS 2005 Education Education Funding to Unknown Education see if funds allocated Coalition for specific programs were reaching schools and whether guidelines for fund utilization were being met. Information on this study available from Ramkumar, "Civil Society Experiences in Monitoring Government Budgets" * TAP grantees ** There may be other CSOs that have completed PETS exercises; however, this table encompasses all of those known by the authors. A Guidebook 65 66 Using PETS to Monitor Projects and Small-Scale Programs Annex C. Resources for Power Calculations Lenth, R. V. (2006-9). Java Applets for Power and Sample Size [Com- puter software]. Retrieved April 12, 2010, from http://www.stat.uiowa. edu/~rlenth/Power http://statpages.org/#Power http://www.statsoft.com/textbook/power-analysis/ http://www.raosoft.com/samplesize.html http://www.dssresearch.com/toolkit/spcalc/power_p1.asp Annex D. Communications Tools from Past TAP-Supported Organizations Audience Organization Research Topic Communications Tool/Strategy Goals Policymakers · National Center · Expenditure · Individual meetings with targeted To inform policymakers for Economic tracking of high-level officials to share their of findings and Research primary school findings and recommendations potential solutions, (Guatemala) programs to give them an opportunity (and a · Romanian · Education PETS · Policy briefs directed towards need) to respond to the Academic Society policymakers at different levels of public government · Research Center · Expenditure · Seminar highlighting study of the University tracking of TB and findings and asking policymakers of the Pacific Immunization to speak on these issues (Peru) Programs Public/Service · National Center · Expenditure · Cartoon strip explaining the To inform the public of Users for Economic tracking of results of the PETS to students, findings, to empower Research primary school teachers, and parents them and give them (Guatemala) programs · Posters (displayed in schools) tools to better monitor · Centro de · Education PETS outlining the findings of the PETS spending and service Análisis y and what students can do to delivery Difusión de improve transparency of school la Economia funds Paraguaya · Meetings with parent associations (Paraguay) to develop tools to monitor facility- level spending A Guidebook Continued on next page 67 68 Annex D. Communications Tools from Past TAP-Supported Organizations (Continued) Using PETS to Monitor Projects and Small-Scale Programs Audience Organization Research Topic Communications Tool/Strategy Goals NGOs and · Indo-Dutch · Primary Health · Developing networks of like- To provide NGOs with CSOs Project Expenditure minded NGOs to share findings information that they Management Tracking and strategies on these issues can use in their own Society (India) grassroots advocacy or lobbying work · Centre for Budget · Budget analysis · Documentary film screenings and Policy for health and presenting the budget process Studies (India) education sectors in state of Karnataka and examples of knowledgeable and effective government officials and incompetent ones Media · Center for · Teacher · Toolkit and press packet for While the media itself Democratic Absenteeism journalists to use to interview is not generally a Development Tracking presidential and parliamentary target audience, CSOs (Ghana) candidates about the problems can provide tools to and potential solutions to teacher the media to help absenteeism disseminate their findings to a wider audience 69 References Ablo, Emmanuel and Ritva Reinikka. 1998. "Do Budgets Really Matter? Ev- idence from Public Spending on Education and Health in Uganda." Policy Research Working Paper 1926. The World Bank: Washington, DC. Alvarado, Betty, and Eduardo Moron. 2008. "The Route of Expenditures and Decision Making in the Health Sector in Peru." Centro de Investigacion de la Universidad del Pacifico: Lima. [http://tap.resultsfordevelopment.org/ resources/route-expenditure-and-decision-making-health-sector]. Amin, Sarnia, Jishnu Das, and Markus Goldstein, Eds. 2009. "Are You Be- ing Served? New Tools for Measuring Service Delivery." The World Bank: Washington, DC. Cuevas, Mario, and Jorge Lavarreda. 2008. "Expenditure Tracking to Improve the Effectiveness of Public Education in Guatemala." Centro de Investiga- ciones Economicas. Das, Jishnu. 2004. "Equity in Educational Expenditures: Can Government Subsidies Help?" Policy Research Working Paper 3249. The World Bank: Washington, DC. Das, Jishnu, Stefan Dercon, James Habyarimana, and Pramila Krishnan. 2004a. "Public and Public Funding Basic Education in Zambia: Impli- cations of Budgetary Allocations for Service Delivery." Africa Region Human Development Working Paper Series No. 62. The World Bank: Washington, DC. Das Gupta, Monica, Varun Gauri, and Stuti Khemani. 2004. "Decentral- ized Delivery of Primary Health Services in Nigeria: Survey Evidence from the States of Lagos and Kogi." Africa Region Human Development Working Paper No. 70. 70 Using PETS to Monitor Projects and Small-Scale Programs Gauthier, Bernard and Waly Wane. 2006. "Leakage of Public Resources in the Health Sector: An Empirical Investigation of Chad." Working Paper. The World Bank: Washington, DC. Gauthier, Bernard. 2006. PETS-QSDS in Sub-Saharan Africa: A Stocktak- ing Study. Gauthier, Bernard. 2009. "Some Elements of Guidance for the Design and Implementation of PETS/QSDS." Presentation ­ TAP Workshop. Kampala, Uganda. Habyarimana, James. 2008. "Public Expenditure Tracking Surveys: An Ap- plication to Education." Presentation. The World Bank: Washington, DC. Oyugi, Lineth, Andrew Riechi and Evelyn Anupi. 2008. "Public Expenditure Tracking of Secondary Education Bursary Fund in Nairobi Province, Kenya." IPAR. Kenya. Ramkumar, V. 2008. "Obstacles Faced by IMOs in Implementing PETS," Transparency and Accountability Project June IMO Conference. Brookings Insti- tution: Washington DC.. Reinikka, Ritva and Jakob Svensson. 2001. "Explaining Leakage of Public Funds." Policy Research Working Paper Series. The World Bank: Washing- ton, DC. Reinikka, Ritva and Nathanael Smith. 2004. "Public Expenditure Tracking in Education." International Institute for Education. UNESCO. Sundet, Geir. "Public Expenditure Tracking Surveys: Lessons from Tanza- nia". Bergen: Chr. Michelsen Institute (U4 Brief 2007:14). Sandilya, Indira. 2009. "Communication for Impact." Presentation ­ TAP Workshop. Delhi, India. World Bank. 2003. World Development Report 2004: Making Services Work for Poor People. The World Bank: Washington, DC. Ye, Xiao and Sudharshan Canagarajah. 2002. "Efficiency of Public Expendi- ture Distribution and Beyond: A Report on Ghana's 2000 Public Expendi- ture Tracking Survey in the Sectors of Primary Health and Education". Africa Region Working Paper Series No. 31. The World Bank: Washington, DC. ECO-AUDIT Environmental Benefits Statement The World Bank is committed to preserving Saved: endangered forests and natural resources. The · 3 trees Office of the Publisher has chosen to print · 1 million British Using Public Expenditure Tracking Surveys to thermal units of Monitor Projects and Small-Scale Programs total energy on recycled paper with 30 percent postconsumer · 295 pounds of net fiber in accordance with the recommended greenhouse gases standards for paper usage set by the Green Press (CO2 equivalent) Initiative, a nonprofit program supporting pub- · 1,422 gallons of lishers in using fiber that is not sourced from waste water endangered forests. For more information, visit · 86 pounds of solid waste www.greenpressinitiative.org. Public expenditure tracking surveys (PETS) are a tried and tested meth- odology to identify delays in financial and in-kind transfers, leakages, and other inefficiencies in government programs. This guidebook aims to provide a starting point for civil society organi- zations and other organizations interested in conducting public expen- diture tracking surveys, both on a small and a larger scale. It is designed to lead users from conceptualization to dissemination of findings, with an emphasis on using evidence effectively to influence policy decisions at any level. Based on the experience of the World Bank in measuring and improving the effectiveness of service delivery, this approach helps empower citizens to keep service providers accountable through better information, communication, and engagement. ISBN 978-0-8213-8519-7 SKU 18519