Report No. 22870-BR Brazil Planning for Performance in the Federal Government Review of Pluriannual Planning (In Two Volumes) Volume 1: Main Report December 12, 2002 Brazil Country Management Unit PREM Sector Management Unit Latin America and the Caribbean Region Document of the World Bank CURRENCY AND EQUIVALEI E PTS Currency Unit - Real (R$) December 1997: R$1.12/US$ December 1999: R$1.21/US$ January 2001: R$1.96/US$ Weihtsandmensuires Metric System FiscaR Year January 1 to December 31 ABRIEViATRONS AND ACRONH7.YWS NDIES Banco Nacional de Desenvolvimento Econ6mico e Social CGP Commissariat G6n6ral de Plan IlCMS Imposto sobre Circulagdo de Mercadorits e Servigos (Value- added Tax) IlPEA Instituto de Pesquisa Econ6mica Aplicada LDO Lei de Diretrizes Orqamentdrias (Budge. Directives Law) LOA Lei do Orgamento Annual (Annual Bud-et Law) MARE Ministry for Federal Administration a. Leform of the State MOH Ministry of Health MOT Ministry of Transport MTEIF Medium-Term Expenditure Framework OECD Organization for Economic Co-operatio and Development PAB Piso de Asistencia Bdsica PPA Plano Plurianual de Agao SEDES Secretaria de Desenvolvimento SIlG Sistema de Informaqdo Gerencial SOF Secretaria de Orqamento Federal SPHI Secretaria de Planejamento e Investimeizos Estrat6gicos Vice President: David de Fer:a ti Country Director: Vinod Thom as Sector Director: Ernesto May Sector Manager: Ronald Myers Lead Economist: Joachim von Amsberg Task Manager: Yasuhiko Matsida TABLE OF CONTENTS EXECUTIVE SUM AR Y ................................................................................................. I UNDERSTANDING THE PPA ............................................... .......... I ASSESSING THE PPA.........................II.. ...................... Ii IMPROVING AND CONSOLIDATING THE PPA ........................................... V RECENT ADVANCES AND OTHER DIMENSIONS OF THE PPA 2000-2003......IX INTRODUCTION ..................................................................IX THE PROGRAM MANAGEMENT MODEL IN THE ORGANIZATIONAL ENVIRONMENT....... ................ X COSTS ........................................................................... XI ANNUAL EVALUATION OF THE PROGRAMS AND OF THE MULTIYEAR PLAN...............................XII SIGPLAN: NETWORK MONITORING ...............................................XIII TERRITORIAL PLANNING .........................................................XIV THE PROGRAM AS A FACILITATOR OF INTERNAL AND EXTERNAL PARTNERSHIPS ............XV 1. UNDERSTANDING THE PPA .................................................................................. 1 INTRODUCTION ...................................................................1 WHAT IS THE PPA? CONSTITUTIONAL DEFINITION AND HISTORICAL BACKGROUND ..................2 REVIVING THE CAPACITY TO PLAN: THE CURRENT PPA INITIATIVE ........ ...............8 SUMMARY ......................................................................14 2. ASSESSING THE PPA .......................................................................................... 16 PPA AND MANAGEMENT REFORM ..................................................16 SECTORAL DIMENSIONS OF THE PPA PROGRAM MANAGEMENT..........................16 CRITICAL IMPORTANCE OF IMPROVING ALLOCATIVE EFFICIENCY ........................25 KEY CHARACTERISTICS OF BRAZIL'S PUBLIC EXPENDITURES .......... .................26 PPA AS A TOOL TO IMPROVE ALLOCATIVE EFFICIENCY ....................................28 INSTITUTIONAL CONSTRAINTS To BETTER PUBLIC EXPENDITURE MANAGEMENT.................35 3. IMPROVING AND CONSOLIDATING THE PPA..........................................36 PLANNING FOR PERFORMANCE......................................................36 IMPROVING THE TECHNICAL DESIGN OF THE PPA MODEL..............................37 ADDITIONAL REFORM MEASURES TO SUPPORT THE PPA CONSOLIDATION....... ...........42 TACKLING THE EXTERNAL CONSTRAINTS TO THE PPA .................................49 SEQUENCING THE REFORM MEASURES ..............................................51 TABLES TABLE 1.1: MAIN STEPS IN PREPARATION OF PPA 2000-2003 AVANCA BRASIL...............14 TABLE 2.1: PROGRAM EXECUTION IN MINISTRIES OF HEALTH AND TRANSPORT, FISCAL YEAR 2000 ........ ...........................................................19 TABLE 2.2: SUMMARY OF THE EXPERIENCE OF FOUR PPA PROGRAMS IN TWO SECTORS .............21 TABLE 2.3: BRAZIL - FEDERAL BUDGET RIGIDITY ...............................................27 TABLE 2.4: CONGRESSIONAL BUDGET ALLOCATIONS To NON-STRATEGIC AND STRATEGIC PROGRAMS, 2000 .......................................................30 TABLE 2.5: ALLOCATIONS OF DISCRETIONARY EXPENDITURES IN LOA AND BUDGET EXECUTION DECREE, 2001 (SELECTED MINISTRIES, MILLION OF RS.) ..................................31 TABLE 2.6: SECTORAL DISTRIBUTION OF Low EXECUTION PROGRAMS .....................33 TABLE 3.1: RECOMMENDED MEASURES FOR CONSOLIDATING THE PPA PERFORMANCE FRAMEWORK ......................................................................54 FRGURES FIGURE 3.1: ILLUSTRATIVE DECISION TREE FOR PPA PROGRAlv EVALUATION . ...............41 BOXES Box 1. 1: THE "Exos" STUDY . .............................................. ........12 Box 2. 1: MONITORING IMPLEMENTATION: THE FAMILY HEALTH PROGRAM ...........20 Box 3. 1: BUDGETING IN BRAZIL ..............................................39 Box 3.2: PROGRAM REVIEWS IN CANADA ................................. .....40 Box 3.3: BUDGET REPORTING IN AusTRIA ...............................44 Box 3.4: SWEDEN'S PUBLIC EXPENDITURE AREAS .............. .................46 ACKNOWLEDGEMENT This report was prepared based on the World Bank missions in September and December 2000 and March 2001. The World Bank team was composed of Yasuhiko Matsuda, Geoffrey Shepherd, William Dorotinsky, and German Rios. Jos6 Orihuela and Henry Gonzalez provided research assistance. Shafiq Islam (consultant) prepared a background paper on PPA 2000-2003 prior to the commencement of the field work by the Bank team. In two videoconferences in July and August of 2000, a discussion of the PPA was held between selected Bank experts in public expenditure management and representatives of the Secretariat for Planning and Strategic Investments (SPI) of the Brazilian Ministry of Planning, Budget, and Management. 'he team also benefited from advice and comments from various World Bank staff. The peer reviewers were Allen Schick and Barbara Nunberg. Throughout the preparation of the report, the team was supervised by Gobind T. Nankani and Vinod Thomas (Country Director), Suman Bery and Joachim von Amsberg (Lead Economist), and Claudia Costin and Ronald Myers (Sector Manager). In carrying out the mission, the team received logistical support from Patricia Bacelar and Julia Conter at the World Bank Brasilia Office, and Patricia Mendez at headquarters. Finally, the team acknowledges strong support from and professional collaboration by the staff of the SPI, especially the Secretary Josd Paulo Silveira, Ariel Garces Pares (Director of Planning), Andrd Cavalcanti, Mariana Meirelles, and Alofsio Pddua Pinto. The team is also grateful to staff of the Ministries of Transport, Health, Environment, and Social Assistance, who generously offered their time to respond to our requests for interviews, sometimes more than once, during the three missions. A COM2PLEMENTARY NOTE Y THE COVERNMENT Because of the lapse of time since the finalization of the field work (spring of 2001) and the final publication of this report (January 2003), the findings of this report are now somewhat outdated. The Government has since made additional progress on many fronts. There are also a number of important aspects of the PPA 2000-2003 that the report deliberately or inadvertently left out of its analysis. To fill these gaps and to provide an update of the more recent progress, the Government has prepared a complementary note, which is included in this report after the Executive Summary. EXECUTIVE SUMMARY UNDERSTANDING THE PPA Introduction 1. National development planning was once seen as an important tool of economic development. However, today few countries continue with systematic planning exercises. In those countries that still carry out comprehensive development plans, they are often little more than expressions of the government's and society's wishes and intentions, divorced from the reality of day-to-day public policy-making. In this context, Brazil stands out for its commitment to formal government planning, as enshrined in the country's Constitution, and more importantly, for its serious ongoing efforts to revive planning as a tool for improving the effectiveness of the public sector to deliver public goods and services and thus promote the country's broader socioeconomic development. 2. In Brazil's 1988 Constitution, the PPA, together with the Law of Budget Directives (LDO) and the annual budget laws (LOA), forms part of a set of legal instruments for fiscal and public expenditure management. Its main role, as a tool to improve the planning and budgeting process, is to provide the government with strategic guidelines for the allocation of public resources to improve efficiency and ultimately to achieve a higher level of development. Brazil also has a long history of state planning and administrative reforms. But during the macroeconomic crises of the 1980s, the Brazilian state partially lost its capacity to plan. The current PPA initiative, starting with the PPA 1996-99 and continuing with the PPA 2000-2003, is a conscious and major effort on the part of the government to revamp the eroded planning capacity and to use planning as a strategic tool of public management. 3. Brazil's Multi-year Plan (PPA) is a unique instrument. It is primarily an instrument to direct strategic allocation of the federal government's budgetary resources. With the PPA 2000-2003, the government has attempted to link the plan to the annual budget to ensure its effective implementation. The current PPA is also a vehicle with which the Federal Government is attempting to re-build the state's institutional capacity for directing the country's development primarily through better use of budgetary resources, as well as by mobilizing private sector and external financing. This report, prepared during the first and second years of the PPA implementation, offers preliminary assessments of its potential and accomplishments, and discusses some of the key challenges it faces in fulfilling its stated as well as implicit objectives, focusing on: (i) the introduction of a new management model and its impact on operational efficiency and results orientation of the public sector; and (ii) the integration of planning and budgeting and its impact on improving allocative efficiency. More recent advances in the implementation and the evolution of the PPA framework as well as aspects of the model that this report does not explore are briefly reported in the following complementary note prepared by the Ministry of Planning, Budget and Management. i The Current IPPA Initiative 4. It was with the PPA 1996-1999 and especially with its Brasil em AVdo Program (as a key thematic component of the PPA) that the Brazilian government began its current efforts to revamp its planning capacity. The primary goal was to create adequate conditions for embarking on a program of strategic investments, essential for integrated development of the economy (especially but not exclusively in infrastructure development). A particularly innovative aspect of the Brasil em AVdo Program is the new program management approach: " All government actions organized as Projects: Clear objectives, as well as time, cost and quality requirements for each project were specified. o Each project assigned to a Manager: A manager was appointed for each project to organize actions from various areas into the project, pull together resources and overcome obstacles. o Management Information System (MIS): A computer network interlinked all managers and partners to allow immediate access to information concerning all projects, with the aim of enabling these lead actors to share information, promote partnerships and identify impediments to achieving goals. o Access to resources linked to progress of the project: The MIS also permitted another innovation of the management model - a manager's access to financial resources depended on the progress of the project, at least in some cases. 5. Given the successful implementation experience of the Brasil em Agdo Program, the government decided to expand its management model to the entire federal government programs. Thus, the PPA 2000-2003 AvanVa Brasil was designed based on a completely overhauled new program classification (now common with the budget classification). The entire federal government budget was broken down into 388 Programs, each with clear objectives, indicators for financial and physical execution, and a Manager. The government also maintained consistency between the Plan and the medium-term macroeconomic framework to avoid a typical weakness of government plans - lack of fiscal realism. Also, for the first time, the government carried out a systematic and comprehensive evaluation of the PPA Programs, based on Program Managers' self- assessments. In 2001 and 2002, the evaluation reports were presented to Congress and made public, which is another demonstration of the government's commitment to transparency and managerial improvement. 6. Besides being a government plan for guiding public resource allocations, the PPA, as a broad development plan, is also intended to accomplish a wider set of objectives, including gaining political legitimacy for the President's governing proposal, communicating it to society, and translating it into concrete government actions. In addition, both the Brasil em AVdo Program within the PPA 1996-1999 and the PPA 2000- 2003 Avanga Brasil have been used explicitly as vehicles for promoting public-private and federal-sub-national partnerships, as well as a driver of a fairly ambitious public sector management reform agenda. ii Summary 7. Brazil's PPA is a unique and serious attempt to use techniques of planning to transform the Brazilian Federal bureaucracy into a modem, results-oriented entity that effectively provides public goods and services that are demanded by citizens. Its goals of revamping the state's capacity to plan and instilling a culture of entrepreneurial management in the public sector are ambitious, yet its approach to implementation is advisedly cautious. Its conceptual design is built explicitly on experiences of the previous PPA, with which a series of management innovations were successfully piloted. The development of the Plan followed a careful preparation process that lasted more than two years. The "model" is multi-faceted and highly complex, as is the Brazilian Federal bureaucracy. It is also an evolving process, in which the Government is constantly introducing adjustments to the model's design as it encounters practical implementation challenges. ASSESSING THE PPA PPA and Management Reform 8. A primary objective of the PPA 2000-2003 is to improve impact of government programs through a more results-oriented, "managerial" public administration. To achieve this objective, the management "model" of the PPA appears to be based on the following principles: (a) The division of public business into Programs with clear objectives and performance indicators is a catalyst for advancing a culture of performance orientation. (b) These Programs are linked to budgetary resources in order to ensure that these final needs can be effectively served. (c) Program Managers will be able to contribute to meeting these final needs by working across sectoral and bureaucratic boundaries. (d) Monitoring and evaluation can contribute to better Program management. 9. An examination of four selected PPA Programs in the health and the transport sectors suggest that the PPA is proving a catalyst in advancing a culture of performance orientation: budget and budget execution information is becoming more useful to people inside and outside government; the area for political discretion is being reduced by better definition of spending items; and public servants are getting more into the habit of thinking about the final needs they are supposed to serve. But these observed effects are modest (and impossible to measure); they are stronger in some Programs than others; and the PPA is no doubt only one of a broader set of factors (including public demand) impelling the changes. 10. The sectoral context - the nature of the tasks and the organizations - seems to be an important factor in determining managerial behavior and Program performance. The degree of commitment to the PPA varies across ministries, and so do their planning iii capacity. In some ministries, organizational changes are being considered (or have already been implemented) in ways that are logically consistent with the PPA's program logic. 11. Besides the sectoral characteristics, program characteristics may also matter. Although evidence is not clear and definitive, the mini case studies suggest that impacts of the PPA managerial model, which relies heavily on individual Program Managers' own initiatives, may depend on whether or not the program itself contains strong reform/developmental content. In spite of the expectation that better specifications of objectives and outputs, combined with greater transparency (through more intuitive program classifications and public information dissemination) would lead to more results orientation, the PPA model itself may not provide sufficiently strong incentives for managers to perform and pursue better outcomes. Motivating managers to perform may be more difficult with routine programs than with "reform" programs. 12. Our overall impression of the PPA management model in the light of the evidence we have reviewed is that the initiative is well-timed and that it is well-proportioned. But, the ultimate success of the managerial reform dimension of the PPA initiative will depend on effectively tackling additional challenges that have so far not been incorporated explicitly in the PPA model, including: o Addressing the issue of incentives to line managers; o Strengthening sectors' and ministries' capacities for strategic planning; and o Developing better human resources. Criticall Importance of Improving AUocaglve Efficiency 13. Effective integration of planning (a mechanism for clarifying the government's policy priorities) and budgeting (a mechanism for resource allocation) is essential for the government's goal of improving performance-oriented public management and for good program performance. Without adequate resources and predictability in their availability within a credible hard budget constraint, program implementation will likely suffer. Further, clarity in strategic priorities is essential for efficient allocation of scarce resources (especially in conditions of constant fiscal adjustments as in Brazil). 14: Based on the implementation record for 2000 and part of 2001, our conclusions are that the PPA initiative has not yet attained tangible improvements in allocative efficiency in spite of the government's concerted efforts to integrate the plan to budgets. First, despite its transparency that has brought much clarity to objectives of each of the 388 Programs, the PPA does not identify relative priorities among these Programs or among groups of Programs. A list of strategic programs is presented in an annex to the LDO each year, but no explanation is provided, and the PPA itself does not offer explicit criteria by which these strategic programs are selected each year.' 1 According to the Ministry of Planning, an exercise was conducted to pre-allocate resources strategically to priority programs, but these signals of prioritization were not made explicit in the PPA document itself. A possible interpretation is conflict avoidance. Budgeting as a resource allocation exercise can be inherently conflictive, pitting sectoral and regional interests against each other in a zero-sum struggle over finite iv 15. There is also evidence that the actual resource allocations do not always reflect the strategic designation in the LDO annex both during budget approval (as Congress could introduce amendments without respecting the strategic designation in the LDO, although the situations improved in 2001 and 2002) and during budget execution (as the Ministry of Planning and the Ministry of Finance rationed commitment and cash ceilings to discretionary programs throughout the fiscal year, thus implicitly favoring many of the non-strategic, obligatory spending programs). In many cases, being designated as strategic does not guarantee better-than-average execution of strategic programs. 16. Finally, the comprehensive program evaluation carried out in 2000 was designed primarily to identify obstacles to program management. Although the results of these evaluations were reportedly reflected in the annual PPA revision exercise, the evaluation method can be diversified to enhance its linkage to strategic resource allocation decisions more effectively. The Ministry of Planning is working on developing a more diversified and comprehensive evaluation system that includes deeper evaluations of selected programs. IMPROVING AND CONSOLIDATING THE PPA Planning for Performance 17. The PPA is a unique planning instrument that is intended to enhance the federal bureaucracy's performance orientation. The basic tenets of its performance framework consists of (i) clear identification of government objectives and priorities; (ii) integration of planning and budgeting for improved allocative efficiency; and (iii) promotion of entrepreneurial management through transparency and partnership and through flexible and results-oriented program management. 18. Achievements over the past few years have been considerable. With SPI's strong technical leadership and flexible approach to implementation that permits learning by doing, a good institutional infrastructure for performance-oriented planning and management has been put in place. The directions that the government aims at with the PPA model and related reform measures are broadly in line with the recent OECD trend in public management reforms. Yet specific ways in which these goals are pursued show Brazilian touches that reflect the country's different historical and institutional contexts. In short, this is a serious reform attempt with a large dose of innovation and strong government commitment to its success. The PPA holds a potential to catalyze a range of reform measures, which, if pursued as complements to specific measures for improving the PPA's own design features, could contribute to developing a more performance-oriented institutional environment in Brazil's federal bureaucracy. 19. As impressive as the efforts so far have been, the PPA model and the way in which it has been implemented is not flawless. There are weaknesses within the model itself that need to be addressed. There are also several external constraints that, while they lie outside the purview of the PPA initiative as such, should still be removed, or at least alleviated, for the PPA to have lasting impact on improving government performance. To resources. By obscuring the relative gains/losses among programs, the government might have opted to dampen potential conflict especially at the crucial moment of launching the new PPA framework. v address these internal weaknesses and external constraints, the government might consider three sets of actions: (i) those aimed to improve some aspects of the technical design of the PPA model; (ii) additional public management reform that complements the PPA's performance framework; and (iii) broader structural and institutional reforms to remove external constraints to the PPA's consolidation. fimproving the Technical Design of the PIPA Model 20. Progress so far has been uneven in achieving the several parallel objectives of the PPA. There has been more progress on clarifying government objectives with respect to each of the 388 programs than on signaling relative priorities among these objectives that compete for limited public resources. Important initial steps have been taken to permit effective integration of planning and budgeting, for example, by developing a common programmatic classification system and a common inventory (cadastre) of programs and actions. But, further changes in program costing and evaluation are necessary to aid budgetary and management decision-making. The appointment of the Program Managers and use of the management information system and performance indicators are expected to boost the performance orientation. But the lack of explicit attention to the relations between the Programs and the formal organization of the government bureaucracy limits the PPA's potential to be a catalyst for performance improvements among ministry staff other than the Program Managers and a few who work directly with them. 21. Much of this unevenness arises not so much from obstacles faced during implementation as from the relative emphases given to different objectives in the design of the PPA model. It is our impression that by a conscious choice the government did not stress the PPA's contribution to better resource allocation or to organizational incentives, instead focusing its initial attention on clarifications of program objectives and designs, and on establishing the Program Managers as key actors in the new framework. To build on the strengths and address the weaknesses in the PPA design, specific measure should be taken to: o sharpen PPA priorities (as a necessary condition for more effective allocation of resources to clearly identified strategic priorities); o introduce full program costing (as a basis for better estimates of relative efficiency of the Programs and more explicit trade-off among them in budget decision-making); o introduce in-depth program evaluation (as a tool to identify potential for savings by eliminating or rationalizing inefficient/ineffective Programs/Actions); and o monitor actual roles of Program Managers (as a prelude to developing more systematic strategies of organizational reforms). Additional Reform Measures to Support the PIPA Consolidation 22. Ultimately, the effectiveness of the PPA's performance management framework will depend on effective and performance-oriented bureaucratic apparatus. The PPA provides the federal bureaucracy with some tools for enhancing its performance vi orientation. But, the PPA initiative itself does not address a host of related issues, such as a framework for better accountability for results within sector ministries, or a much needed rationalization of the federal budget process (e.g., a draft law to regulate the budget process and financial management is pending in Congress). Additional reform measures, especially in the areas of organization and management of sector ministries and government-wide budget process, are necessary to complement the core of the PPA's performance framework. 23. An exhaustive needs assessment of organizational and budget management reforms is beyond the scope of this report. However, several measures appear particularly relevant for supplementing the PPA initiative. One of the limitations of the PPA initiative is the lack of explicit strategy to incorporate ministries' organizational structures and incentives. As all Programs are implemented by the ministry hierarchy (including some dependent entities), it is imperative that the PPA's entrepreneurial and performance-oriented management approach be adopted by those other than the PPA Program Managers. Therefore, some of the priority measures include: * aligning organizational structures and incentives with the PPA's Program management logic, through human resource management reform and possible organizational restructuring of ministries; * fostering greater ministry ownership of the PPA by, for example, involving sector ministries much more closely in the PPA program development; and2 * linking the PPA to more robust sectoral strategies, as a basis for designing PPA strategic programs and also for ministries' own performance management framework. 24. Effective integration of planning and budgeting would require additional measures to improve the annual budget process. Taking advantage of the PPA's medium-term strategic focus, the government might consider: * adopting a procedure whereby budgetary resources are allocated to broader strategic categories, such as the 28 macro objectives, on a medium-term basis before detailed allocation program-by-program - the LDO could possibly be turned into an instrument for establishing such a strategic expenditure framework, but this would require, among other actions, passage of a new budget and financial management law; * simplifying the planning-budgeting process, either by collapsing the three legal instruments (PPA, LDO, and LOA) effectively into a single process, or by streamlining each - The procedural simplification is also expected to facilitate shifts in SPI's and SOF's roles from the current ones with heavy emphasis on "rowing" functions (e.g., through' detailed control of commitment ceilings), to more strategic "steering" functions; 2 In the initial PPA program development process, the ministries were invited to participate, but their capacities to articulate strategies and programs varied significantly across ministries, and in general, they tended to under-exploit the opportunities offered. Strengthening sector ministries' strategic policy-making capacities should be part of a pending and ongoing agenda of reform. vii o reducing, gradually, the observed divergence between the approved and the executed budgets and thereby improving the credibility of the budget as a binding guide for resource allocations and as a reliable management tool - this would require more effective constraint on budget inflation during the budget approval process. Tackling the External Constraints to the PPA 25. Among the most challenging obstacles to the ultimate success of the PPA initiative is the set of external constraints that reduce the PPA's intended effect on performance management. We have identified three such constraints: (i) the constant "threat" of fiscal adjustment and the considerable expenditure rigidities that arise from constitutional earmarking and other past fiscal decisions (e.g., debt payment); (ii) budget inflation, primarily by Congress, which renders the approved budget impractical as an expenditure management tool; and (iii) difficulties of inter-organizational coordination in the Federal bureaucracy (not to mention across levels of government, which we have not touched upon). There are no quick and easy solutions to these constraints, as these have political as well as technical origins. But measures need to be taken consistently and continuously to reduce their effects over time, including structural reforms to reduce expenditure rigidities, efforts to improve the technical basis of congressional budgeting without usurping its legitimate role in the budget process, and development of a framework for greater ministerial ownership of PPA model and Programs so as to reduce scope for inter- ministerial differences in preferences and priorities. 26. Finally, the PPA's transparency can be a potentially very powerful instrument to push the necessary reform agenda on the above-mentioned institutional constraints. For example, by making explicit the extent to which the expenditure rigidities limit the government's ability to allocate public resources to strategic programs, the PPA might raise public awareness of the problems caused by earmarking of expenditures, as well as deficit spending that reduces the discretionary spending in the future. The current presentation of the PPA does not always convey these overarching constraints on efficient expenditure allocations and difficult policy trade-offs that they force upon the government, in spite of its very transparent format. Instead, it might, in some contexts, give precisely the opposite impression that the government has much room for flexibly pursuing policy objectives through a relatively large number of programs. The PPA's transparency as a catalyst for broader reforms in related areas can and should be exploited further. viii RECENT ADVANCES AND OTHER DIMENSIONS OF THE PPA 2000-20033 INTRODUCTION 1. The integrated model of planning, budgeting and management has been a reality in Brazil since January 2000, with the implementation of the 2000-2003 PPA.4 The implementation of the model involves deep changes in the federal public administration. It produces direct consequences for the culture of organizations and the performance of civil servants. It changes the criteria for expenditure allocation, execution and control, and in particular it changes the relations between the State and society, as it directs all of its actions to meet the demands of the "citizen-customer." 2. The nature of the changes implies a long process, over several multi-annual plans, including those of state and local governments. Its ultimate destination is undefined, nor its final form. Nonetheless, one thing is certain in this process - nothing will be as prior to the 2000-2003 Multiyear Plan, with the creation of programs as a reference for the allocation of resources and the constitution of such programs into management units. 3. This is why this World Bank study, which is now being published, is so important. It represents the testimony of public sector management experts who followed the implementation of this new model from the very beginning. The study was started at the same time and at the same pace the plan was introducing a new concept of planning, budgeting and program management in the Brazilian public administration. It thus also shows the courage and intellectual engagement of its authors when they disregarded the lack of a historical antecedent of the object under analysis, a pre-condition of any academic study, and opted for being part of the implementation process of the model, accepting the challenge of seeking to understand the logic and the extent of the changes, rigorously investigating the weaknesses and risks, as well as defining in many cases the conditions required for the consolidation of the proposed improvements as an effective alternative to entrepreneurial management. 4. As the model resulted from a learning process of the authors and staff of the Ministry of Planning, it seemed important to update briefly some aspects of the model that were introduced after the study was practically concluded, as well as other aspects which apparently could not be approached in such depth as the authors themselves might have wished during the preparation of the study. Hence, these issues are being introduced after the preliminary version of the study was concluded. This update is, thus, an additional contribution to the study, made with the aim of contributing to the discussion the Bank study has generated for its quality and factual rigor. Furthermore, this note reflects the strength of the Bank study to serve as a reference for discussions among and reflections by the Ministry staff about the adoption of the Model by the Federal Public Administration even nearly three years after its implementation. Below are the subjects related to the 3 This is a complementary note prepared by the Secretariat of Planning and Strategic Investment to update recent progress in areas covered by the Bank report and to comment on aspects of the PPA model which the report did not explore. 4 PPA Plano Plurianual = Multiyear Plan ix improvements that have been implemented or are being implemented after the preparation of the study. At the same time, it is the opportunity to present some issues in more details, which were not addressed in depth in the Study but which would require the World Bank's attention in a future report. THE PROGRAM MANAGEMENT MODEL IN THE ORGANIZATIONAL ENVIRONMENT 5. The progress of the management by programs, particularly in 2001, revealed the importance of taking into account the organizational dimension to better solve the insertion of program management in the organizations of the Public Sector and to define the attributions of the manager and the responsibilities of the organization and of the manager in program implementation. 6. The 2001 annual evaluation of the Multiyear Plan incorporated, for the first time, an evaluation of the organizations (of the public administration), asking each sector about its organizational practices, besides the evaluation by programs, based on a model developed by the Secretariat of Management (SEGES). The intention was to understand the stage of each of the organizations in terms of good management practices and their interaction with the management by programs. By introducing this new form of evaluation, it is expected to explore over the coming years the existing complementarities between the Reform proposed by the 1995 Master Plan, anchored in the organizations and the management by programs. This means, in other words, to explore the complementarities between productivity enhancement the Master Plan offered and the incorporation of the focus on results given by the management by programs. Another important step was the consolidation of a partnership between the Secretariat of Planning and Strategic Investment (SPI) and the SEGES for the annual evaluation process and the internalization of the management by programs in the organizations, with the definition of clearer rules for the management of government action in the new model proposed by the PPA. 7. The results of the annual evaluation of the 2001 PPA, compared to the evaluation results of 2000, show that the organizations of the Federal Government are adapting themselves gradually to the management by program model. For example, in 2000, the evaluation of program performance showed that only 72 managers considered the organizational structure adequate for the management by program. In 2001, the number increased to 150. Further 141 programs considered that the organization was in the process of becoming adequate, totaling 88% of the programs in those two brackets. The numbers showed greater insertion of the managers and their teams, as well as the final program delivery units, in the management cycle. Although 2ess significant, it is important to stress that the administrative structures have been fitted to the management by programs in the Federal Public Administration. It has been noted that the participation of managers is significant in events that were previously restricted to the specialized and centralized areas of sectoral entities, permitting greater insertion of the managers in the process. 8. The efforts of the government entities to create, in most cases, coordination structures directed towards the insertion of the managing teams to the management cycle and towards the optimization of the complementarities amongst programs have been x equally significant. Amongst the 22 sectoral entities evaluated, 17 showed that they have a structure and/or authority responsible for the coordination of managers. 9. One of the largest challenges of the model is the creation of organizations that reconcile management by function with management by objectives. The management by program introduces a conflict between the departmental structure and performance by objectives. This conflict produces positive effects up to a certain extent, as it contributes towards the equilibrium between the efficiency and efficacy targets and the results of the organization vis-A-vis society. The challenge, in these circumstances, is to speed up the transformation process of the organization in its various aspects, including; structure, values, leadership, strategy, staff development, and development information and control systems. COSTS 10. The study evidences, rightly, the need of appropriating important elements of costs, especially personnel costs, to the programs. This deficiency is progressively being solved with the reintroduction of the Program of Expenditure Management Monitoring and Institutional Evaluation (PAGG). In early 2002, a task force, coordinated by the Executive Secretary of the Planning Ministry and involving the areas of Planning, Budgeting, Control and Personnel of the Government, was created to implement in 2003 a system of personnel costs with programs as management units. The purpose of the system under development, which will involve the program manager in its operationalization, is to permit a more complete budgeting process and a cost-benefit assessment of each program. 11. The adoption of the program as a cost center requires the creation of the role of an action coordinator, who is responsible for the administration of those costs at the action level, for the programs with actions in more than one department within the ministry and for the multi-sector actions. 12. Another aspect, less known about the benefits expected to be achieved with the introduction of the system, refers to the opportunity given to the manager to identify, at the various levels and organizations carrying out part of the program, the human resources involved in the program, thereby enabling an effective administration of human resources and responsibilities, having as a reference the demands of the implementation of the program and the expected results. 13. To guarantee the development of the cost appropriation system, some procedures are being developed: V Development of a module in the SIAPENET,5 which will be linked to the SIAFI- Gerencial;6 / Preparation of a manual; / Training of civil servants; 5 SIAPENET - Sistema Integrado de Administragio de Recursos Humanos por intermddio da INTERNET = Integrated System of Human Resources Administration by means of the INTERNET 6 SIAFI - Sistema Integrado de Administragdo Financeira do Governo Federal = Integrated System of the Financial Administration of the Federal Government xi / Issue of an Order of the Planning Ministry legally establishing the methodology and the updating of the cost system. ANNUAL EVALUATION OF THE PROGRAMS AND OF THE MULTIYEAR PLAN 14. Although evaluation was mentioned as one of the bases of the new model in the study, no detailed analysis of the process was done at that time. The evaluation, however, presents methodological innovations, which were introduced to take into account a comprehensive annual evaluation of a set of programs and of the Plan, as well as of the need to disseminate the management principles based on performance and achievement of results. 15. Thus, the evaluation of programs and of the Multiyear Plan became one of the stages of the management cycle of the Federal Government expenditures. Furthermore, the results of the annual evaluation generate practical consequences for the improvement of programs and their administration. 16. In this context, the annual evaluation process is carried out fully in a computerized environment, in three stages by level of implementation of the Multiyear Plan - program managers, sectoral ministries and the Ministry of Planning, Budgeting and Management. It adopts the principle of self-evaluation by the management team of the program and develops so as to form a process of interaction of managers, teams and sectoral supervisory bodies for the critical analysis of results and restrictions, as well as agreement on possible improvements. It values, before anything else, the examination of results to better understand the performance registered and the design of the programs. More than controlling costs or measuring the implementation capacity of the government, to evaluate means, in the model, to compare the effort made with the demand and the satisfaction of the beneficiary, fitting the programs increasingly to the expectations of society. The evaluation process, carried out in a participative and transparent manner, aims at providing the continuous learning of the management teams, strengthening their capacity to promote and coordinate with partners the achievement of results. 17. The final product of this work is the Annual Evaluation Report of the Multiyear Plan, which is presented to the National Congress by April 15 of each year. The intention is to make the parliament members aware of the importance of using the Report as a reference for the analysis of the revision of the Plan and of the annual budgets. 18. In its second version in 2001, the Evaluation Report of the PPA incorporated improvements in both its methodology, as well as in the processes developed for its elaboration. The first step for the introduction of changes was the survey carried out between July and August 2001, of 30 people selected amongst managers, sectoral managers, officers of the TCU7 and managers of the central entities of the management cycle. The research evidenced the need for significant changes in the process of mobilization and training of the parties involved, in the evaluation guides and questions, in the recommendations to the evaluators and in the form of participation of the government entities during the evaluation process. 7TCU - Tribunal de Contas da Uni5o - The Federal Court of Audit. xii 19. Based on these results, changes were made in the Evaluation Manual to improve the quality of the work carried out by the managers and sectoral government entities. At the same time, an effort was made of awareness and mobilization of the management of sectoral government bodies, the managers of the programs and their teams. Furthermore, management and sectoral teams were trained to structure a network of evaluators at the management and sectoral teams. 20. To have evaluation results reflected in the allocation process and in the public administration is a great challenge of the model of management by programs. To achieve this, training of managers and administrators is indispensable, as well as learning of new methodologies and collective construction of a management culture of performance evaluation oriented towards results that is marked by participation and by transparency. The process was developed in order to integrate the organizations responsible for planning, budgeting and management, to create the conditions required for the implementation of the improvements. SIGPLAN: NETWORK MONITORING 21. Although the study shows the importance of the existence of a management information system that gives transparency to and enables the control of the actions of the Federal Government, the potential of this innovation was not duly ex lored. Before being a physical and financial follow-up system of programs, the SIGPLAN is an instrument for introducing and practicing the concepts of management of restrictions and for network monitoring, that is, an instrument for sharing solutions by program managers with the network of managers and the Planning Ministry. The prospective focus is another important feature of the system, as it enables a joint and preventive action of the different agents for solving potential restrictions. The intention is to by-pass possible future discontinuities, improving the performance of the programs. Restrictions mean any problem that cannot be solved by the manager. 22. The SIGPLAN is still in the process of consolidation, for the complementary sectoral systems of program management monitoring have not yet been fully developed as foreseen in Decree number 2829/98. 23. For that purpose, a study to be financed with World Bank resources is under bid, containing: / A diagnosis and critical analysis of the current situation of all Ministries as to the management information systems that have been or are being developed for the programs under the responsibility of the Ministry, with recommendations as to their adequacy for the administration of the programs of the Ministries and the needs of improvement; / A project for the development of a management information system that takes into account its main functions of gathering information for the physical execution of the 8 SIGPLAN - Sistema Integrado de Informag6es Gerenciais e de Planejamento = Integrated System of Management Information and Planning xiii program actions, management of restrictions and evaluation. Additionally, the main services or links with other systems of the Federal Government have to be identified, to help the manager in the functions of program administration. v A diagnosis and critical analysis of the conditions for implementing the services of the current SIGPLAN, with a view not only to improve such services, but also to extend them to other services which may better meet the demand of managers in the administration of their programs. 24. Nonetheless, during the 3 years of implementation of the Plan, significant advances have been registered in the monitoring of government action. In the first year, control was achieved over only 37% of the actions by being able to correlate physical performance to financial performance. In 2001, this index rose to 49%. For 2002, the target is to control 65% of the actions carried out with budgetary resources. Another important improvement is the institutionalization of the SIGPLAN, as it has become a reference for the preparation of the General Balance Sheet of the Union, the Annual Evaluation Report of the PPA and for the Address of the President to the National Congress with respect to the physical performance related to the budgetary expenditure of the Federal Government. Finally, it is worth mentioning the role of the SIGPLAN as a system of intensive follow-up of strategic programs, to guide the flow of resources and the administration of restrictions. TERRITORIAL PLANNING 25. The features of the new planning model, introduced by the 2000-2003 Multiyear Plan, required the rethinking of the concept and of the procedures of traditional spatial planning, which was always concerned with the development of the depressed regions. Under this understanding, spatial planning never got beyond being another sectoral policy, which disputed resources with other policies, such as education, health or transport. The study of the National Zones of Development and Integration, prepared for the 2000-2003 Multiyear Plan, readdressed the subject, treating the territory as a planning unit capable of serving at the same time as a reference to local development and to the allocation process of resources. 26. Although the World Bank Study mentioned the National Zones of Development as an element of the model, spatial dimension was not given the emphasis that it deserved due to its importance in the context of the Plan. 27. In the new model, spatial planning was re-emphasized as an instrument capable of promoting rationalization and improvement of the quality of the expenditures of sectoral policies, as these policies were no longer defined only from the point of view of sectoral demands, but also from the point of view of territorial demands. The results of this initiative are numerous. In the first place, it values a multi-sectoral and integrated government action, which has the program as its reference. The programs in the areas of energy and transport are good examples of this. In the first case, the energy programs of the zone integrate the investments in generation, transmission and distribution, as well as express the changes in the energy matrix, including the construction of gas power stations throughout the territory. In the second area, there are the transport corridor programs, xiv which group the transport investments related to the need to modernize the logistics of the country in a multi-modal form. 28. Secondly, it renders a more participative planning viable, by allowing the States to be involved in the elaboration of the Plan and of the programs. Two meetings with local leaders were held per State to discuss the investment portfolio for 2000-2007, which influenced in many cases the definition of the investments in the States. The project portfolio in the area of water resources, for example, was totally remade following the request of the Northeastern region. 29. Thirdly, it promotes a decentralization of the actions of the Federal Government based on three large priorities - to ensure that the process takes into account an integrated and balanced view of territorial development, to permit the regions to have conditions, in terms of services and infrastructure, to compete for productive investments and to improve the systemic competitiveness of the country, and lastly, to coordinate government action in the perspective of preservation of natural resources and guarantee of quality of life. 30. In the fourth place, it stimulates a larger focus on expenditures, by means of a more precise delimitation of the target population, mapping of its problems, as well as of the local factors for development. 31. In the fifth place, it introduces the concept of spatial planning as a reference for the private sector decision making process on investments, besides creating a longer-term horizon - 8 years - to plan development, taking into account 100% of the Brazilian territory. Evidence of this are the evaluation results of the degree of realization of the portfolio's investment opportunities relating to the National Zones of Integration and Development, carried out in early 2002. This shows that over 52% of the 492 projects of economic infrastructure for the 2000-2007 period have been finished, initiated, concessioned or bid. THE PROGRAM AS A FACILITATOR OF INTERNAL AND EXTERNAL PARTNERSHIPS 32. Another aspect of the model, less explored in the study, is the possibility it opens to combining resources and rationalizing expenditures through the incorporation of partnerships, which became explicit in the management instruments of the Government since the 2000-2003 PPA, by organizing partnerships, both internal and external. 33. The changes resulting from this new role of the State open up opportunities to combine public and private actions to achieve the purposes of the Plan. Hence the programs created by the 2000-2003 PPA, besides defining the actions financed by the treasury budget, the social security budget, and the state-owned companies' budgets, all following the need for a fiscal equilibrium, incorporate also external partnerships, resources of the private initiative, the Third Sector, and sub-national units - States and Municipalities. The Brazilian experience has shown the importance of partnerships, both in the areas of infrastructure and in social development, in an environment of significant fiscal restrictions. Two examples may be quoted: the programs aimed at the reduction of infant mortality, which enabled a decline from 41 infant deaths in each 1000 infants born in 1995 to 33 to each 1000 infants born in 2001, thanks to the joint action of non- xv government organizations, the central government and local governments; and the construction of the railroad that links the grain producing areas, in the Center-West of the country, to the Port of Santos, in the State of Sdo Paulo - FERRONORTE, a private investment, which was made possible in view of the combination of public and private actions. 34. The fact of the programs being designed from the perspective of the demands of society gave rise, naturally, to multi-sectoral programs. These programs are a combination of actions of different entities of the administration, linked by a program, whose management is placed in the hands of one sole ministry and of one sole manager. Government action obtains the possibility of solving problems of the society with minimum resources, provided the corporative barriers of the traditional administration are broken. However, this is not an easy task and experience, even though quite recent, shows that the public-public partnership can be viable when the management is directed to results in the society and that sectoral actions that are carried out on their own generate little efficacy for development. 35. The improvement in the quality of public expenditures depends on the most difficult transformation to be conquered: the entrepreneurial behavior of public managers, inserted in a traditionally bureaucratic organizational environment. In this sense, the partnership becomes an innovative form of implementing public policies, less by the possibility of contributing with resources and more by the quality of the relationship that it imposes between the State and society, practicing the concept of co-responsibility in place of a paternalistic State. However, the environment of fiscal discipline and restriction does not always value conveniently this principle in the public policy. And it often deviates the manager from this practice, as its probable success may be interpreted as an alternative to reduce expenditures. xvi 1. UNDERSTANDING THE PPA INTRODUCTION 1.1 National development planning was once seen as an important tool for guiding and catalyzing economic development in both developing countries and a number of developed countries. However, by the early 1980s, if not earlier, the general lack of evidence that comprehensive planning had generated growth and development led to the widespread view that planning had failed to live up to its expectations. A 1983 World Bank working paper concluded that: First, planning failed to live up to expectations in most developing countries. There is no clear association between a high degree of planning efforts in these countries and their performance in terms of growth. Countries that performed best were those that combined effective planning for the public sector with the avoidance of price distortions for the economy as a whole. Second, the technical, administrative, and political causes of the failure of comprehensive planning in mixed economies are inherent in the process and are unlikely to be remedied merely by more strenuous efforts to strengthen the planning machinery.9 1.2 The loss of faith in planning arose not necessarily because of the technical quality of the plans that had been prepared in these countries but because of the perception of the inherent limitation of a planning approach in a dynamic economic context where the market was firmly recognized as the main driver of economic transactions. Today, relatively few countries still carry out comprehensive planning exercises, and fewer still use their plans as a real governing tool. Most plans appear to be not much more than symbolic expressions of the government's and society's wishes and intentions, while actual economic and social policies are made and implemented through other mechanisms. 1.3 In this context, Brazil stands out for its commitment to formal government planning, as enshrined in the country's Constitution, and more importantly, for its serious ongoing efforts to revive planning as a tool for improving the effectiveness of the public sector to deliver public goods and services and thus promote the country's broader socioeconomic development. Especially since 1996, when the first Cardoso administration announced its first Multiyear Plan (Plano Plurianual, or PPA 1996-99) with the core of its public investment program entitled Brasil em Aqdo (Brazil in Action), the Brazilian government has stepped up its efforts to use its national plan as an instrument for guiding resource allocations for better development impact.'0 1.4 With the PPA 2000-2003 Avanqa Brasil (Brazil Forward), the Brazilian government took additional steps in further developing its PPA, not only as a plan as such but also as a vehicle for driving an ambitious attempt to introduce performance orientation in the federal public sector. In preparing the PPA 2000-2003, the federal government overhauled the budget classification system, in use since the 1970s, into a new programmatic system that more accurately describes concrete government actions, tied to a set of broader strategic 9 Ramgopal Agarwala, Planning in Developing Countries: Lessons of Experience. World Bank Staff Working Paper. No. 576. 1983. While the majority of the projects in the Brasil em AVdo were capital investment projects, some activities composed mostly of current expenditures were also included. 1 objectives ("macro objectives"). Now the universe of federal government activities is grouped into 388 programs, each with a clear statement of objectives, a set of performance indicators, indicative financial allocations for the four-year PPA period (including in some cases, the required funding from outside the federal government - the private sector, sub- national governments), and a manager in charge of ensuring effective program implementation. 1.5 Brazil's PPA is a unique instrument. It is not only a comprehensive development plan that, for the first time in Brazil, attempts to link the plan to the annual budget to ensure its effective implementation," but it is also a vehicle with which the Federal Government is attempting to re-build the state's institutional capacity for directing the country's development. Because of its multifaceted objectives and characteristics, the current PPA initiative does not conveniently fit existing notions of government plans or any other similar concepts and practices such as a medium-term expenditure framework. In his presidential message in the Final Report on the PPA 1996-1999, President Cardoso proclaimed that: It (the PPA) is not central planning a la ex Soviet Union. It has little to do with French indicative planning. Certainly it is not something to expect from a 14 12 neoliberal government." The specialists will have to discuss... 1.6 The PPA's rather unique features make it a challenge to assess its strengths and weaknesses from an appropriate perspective.'3 Given these and other peculiarities, the PPA needs to be evaluated on its own terms rather than mechanically against benchmarks of similar, better known international models such as the medium-term expenditure framework of OECD countries. This report offers preliminary assessments of the potential and accomplishments so far of the PPA, and discusses some of the key challenges it faces in fulfilling its stated as well as implicit objectives. The report specifically focuses on two main aspects of the PPA initiative: (i) the introduction of a new management model and its impact on operational efficiency and results orientation of the public sector; and (ii) the integration of planning and budgeting and its impact on improving allocative efficiency. It concludes by offering some suggestions for further improving and consolidating the PPA. WHAT is THE PPA? CONSTrTUTIONAL DEFINMON AND HISTORICAL BACKGROUND 1.7 An obvious place to start in attempting to understand what the PPA is and what it is meant to accomplish is the 1988 Constitution, which specifies the role of the PPA within the broader public expenditure management system of the country. It is also important to understand the historical background to the current PPA initiative. This is an ongoing process, and a fuller appreciation of its potentials and challenges can only be gained by understanding longer historical processes within which the initiative is embedded. " Efforts to establish such direct links between a development plan of any kind and annual budgets are virtually unheard of outside of Brazil. 1 Brasil em AVdo, 1996-1999: Relatorio Final. Ministry of Planning. 1 An aspect of the PPA resembles the trend in OECD countries to move toward a medium-term expenditure framework (MTEF), or budgeting in a multi-year framework. But the current government also expects the PPA to be a mechanism for fostering public-private partnership in carrying out much needed infrastructure development, or promoting effective intergovernmental coordination in social policy implementation. 2 PPA and the Budget Process: Constitutional Definition14 1.8 Article. 21 §9 of the 1988 Constitution mandated that the government should "prepare and carry out national and regional plans for ordaining the territory and for economic and social development." It further specified that "the law that institutes the multi-annual plan shall establish, by region, the directives, objectives, and targets of the Federal Government for the capital expenses and other expenses resulting therefrom and for those regarding continuous programs" (Art. 165 §1), and that budgets, budget directives, and all national, regional and sectoral plans and programs be prepared in accordance with the multi-annual plan (Arts. 165, 174 §1). A newly elected President must produce a four- year plan and send it to the National Congress by August 31s of the first year of his/her mandate. The Congress deliberates the plan and returns its approved version to the Government by the closing of the legislative session (December 15 ). 1.9 The annual budget cycle starts with the proposal for the Budget Directives Law (LDO), which is normally sent by the President to the Congress by April 15 of each year. According to the Constitution, LDO "shall contain the targets and priorities of the Federal Government, including the capital expenses for the following fiscal year, shall guide the preparation of the annual budget law, shall deal with changes in tax legislation and shall establish the investment policy for official promotion financing agencies" and "amendments to the bill of the budget directives law may not be approved if they are incompatible with the multi-annual plan" (Art. 165). In the LDO, the Government defines its priorities for the year and provides guidelines for the preparation of budgets by line ministries. Congress typically approves the proposed LDO with some changes by the end of June.5 In the following two months, the Government prepares its draft Annual Budget Law (LOA) and submits it to Congress by August 31 for its debate and approval. 1.10 Thus the PPA, together with the LDO and the LOA, forms part of this set of legal instruments for macroeconomic and fiscal management in Brazil. Its main role, as a tool to improve the planning and budgeting process, is to provide the government with strategic guidelines for the allocation of public resources to improve efficiency and ultimately to achieve a higher level of development.16 It appears, at least from the Constitutional text, that the PPA is meant to be primarily a public investment plan with the flexibility that the plan might also cover non-capital expenditures. 14 This section is partially based on "Review of Brazilian Fiscal Institutions," in Brazil.: Structural Reform for Fiscal Sustainability, (Draft) World Bank, Report No. 19593 BR (June 7, 2000). 15 The government's proposal is reviewed by the Joint Committee that represents both houses of the Congress, where various amendments will be introduced by its sub-committees. The Joint Committee votes on the recommendations of these subcommittees and passes on the amendments to the Plenary Session of the two Houses for final voting. The President can veto any line item approved by the Congress within 15 days, but the Congress has the opportunity to reject such a veto within 30 days by an absolute majority in a joint session of the houses using a secret ballot (Art. 66). In case negotiations in the Congress or between the Congress and the Government are prolonged, the previous year's budget remains in effect on a pro rata basis until a new LOA is enacted. 16 In this sense, Brazil's PPA is a rather different kind of instrument from indicative plans (e.g., in France in the post WWII period) whose objectives are to influence the private as well as the public sector's economic behavior by forecasting economic trends and signaling the government's intention as to where efforts are directed on a priority basis. 3 Brief History of Planning in Brazil 1.11 Brazil has a long tradition of active state involvement in the economy. It was also rather slow in abandoning the state-led model of economic development in the 1980s and 1990s. Even after initiating policies of market liberalization in the early 1990s and consolidating them by the end of the decade, Brazil still looks to the state for a guiding role in the economy more so than most of its neighbors. 1.12 Part and parcel of this statist tradition in Brazil is a history of state planning. One of the first systematic national plans in Brazil was the 1956-61 Targets Plan (Plano de Metas) implemented by the government of Juscelino Kubitscheck. Covering five key sectors (energy, transportation, food production, basic industries, and education), the Targets Plan identified "thirty targets/goals for specific bottlenecks in the economy" and their achievements "required the cooperation of different government agencies as well as different private sector actors."1 The Targets Plan is usually described as having been a relative success as a means of promoting import-substitution industrialization in Brazil.18 1.13 After the military take-over in 1964, successive governments continued relying on various national development plans to stimulate industrial development of the country. In the 1960s, IPEA was created as an independent government planning institute to prepare a ten-year plan (Plano Decenal). However, as the state's direct and indirect role in guiding or executing investments expanded, its capacity to effectuate needed structural adjustments of its national industries grew increasingly weak. By the early 1980s: The Brazilian State neither commanded nor effectively cooperated with the private sector. It had no secure control over real resources. Rather the Brazilian public sector experienced progressive fiscal difficulty over this period, diminishing its capacity to guide needed structural adjustment. . . . State enterprises were a direct instrument, but they foundered on the need to increasingly secure their finances abroad. Increasingly, parastatals became means of obtaining resources for other activities rather than agents for priority real investment. Weakness was self- fulfilling. Larger and larger efforts were necessary to persuade the private sector in the validity of signals emanating from policymakers. As subsidies multiplied in response to private demands, the relative inducements required to undertake the reallocations of resources became blurred, while deficits increased."9 1.14 Having expanded so much without increasing its effectiveness, the Brazilian state was close to collapse under its own weight even before the onset of the macroeconomic crises of the 1980s. The state's capacity for internal coordination was seriously debilitated. "Until the early 1980s, the Planning Ministry did not even know the number of public agencies that existed."20 1 Barbara Geddes, Politician's Dilemma: Building State Capacity in Latin America, (University of California Press, 1994, p. 67). 1o Werner Baer, The Brazilian Economy: Growth and Development, (Praeger, 1995, pp. 61-62). 19 Albert Fishlow, "A Tale of Two Presidents: The political Economy of Crisis Management," in Democratizing Brazil: Problems of Transition and Consolidation (Oxford University Press, 1989, p. 110). 2 Kurt Weyland, "From Leviathan to Gullivar? The Decline of the Developmental State in Brazil." Governance: An International Journal of Policy and Administration, Vol. 11, No. 1, Jan. 1998, p. 55. 4 1.15 The difficulty faced by Brazil's development planning apparatus toward the 1980s is more likely to be due to broad structural constraints on the Brazilian economy (e.g., the oil shocks) and the "exhaustion" of the import-substitution model of economic development. However, the state's weakened institutional capacity and the fact that the development plans and budgets were not fully integrated with each other would have contributed to the eventual ineffectiveness of the plans as a government instrument.21 1.16 Since the 1980s, two sets of events have added to the list of built-in obstacles to effective planning in Brazil. First, and probably the most important, is the series of economic crises that plagued the country for much of the 1980s and until mid 1990s, when the Real Plan finally stabilized the economy. In this high-inflation context, in which budget execution inevitably diverged significantly from approved budgets, multi-year plans (or even annual budgets) were made basically unusable as a tool for economic management. Thus the capacity and the tradition of planning that the Brazilian state possessed since the 1950s (in spite of the growing ineffectiveness of the more recent plans) were lost to a large extent in the 1980s. 1.17 Changes in the institutional structure of the state that have been taking place since the democratic transition and the 1988 Constitutional reform have also affected the Brazilian state's capacity to plan for the whole economy. One such change is the decentralization process, which gave sub-national governments substantial autonomy, 22 power, and resources without corresponding expenditure responsibilities. The Federal Government cannot dictate to sub-national governments its will on policy priorities and resource allocations. In the democratic context, the structure and the actual functioning of certain political institutions also limit the scope of rational, technocratic planning in Brazil. It is well-established among students of Brazilian political institutions that the current electoral and party systems encourage particularistic interests, often regionally or locally based, to influence political processes including public finance management (e.g., through the pork barrel politics of investment resource allocations). The Executive branch's capacity to contain such particularistic pressures emanating from the Congress and state governors is weakened because of the President's need to form coalitions among weakly disciplined political parties.23 1.18 Finally, the successful implementation of trade liberalization and privatization in the early 1990s also limited the state's role in the economy and thus its real capacity to plan. Before the 1980s, the state claimed the "commanding heights" of the economy by, for example, controlling 44 percent of the fifty largest banks and providing 70 percent of industrial credits through public development banks such as BNDES.24 In today's more liberalized market, however, the state's capacity to direct financial resources and influence private-sector behavior is far more limited. This makes the government's task of obtaining 21 Alfredo H. Cost-Filho, "0 --jamento no Brasil: A Experi8ncia Recente." Pensamiento Iberoamericano, No. 2, 1982. pp. 57-76. For a more detailed discussion on this and related issues, see Brazil: Issues in Brazilian Fiscal Federalism, (draft) World Bank, Report No. 22523-BR, June 27, 2001. 23 See for example, Barry Ames, " Electoral Rules, Constituency Pressures, and Pork Barrel: Bases of Voting in the Brazilian Congress." The Journal of Politics, Vol. 57, Issue 2 (May, 1995), 324-343. etc. 24Werner Baer, The Brazilian Economy: Growth and Development, 4' edition (Praeger, 1995, p. 251). 5 the cooperation from the private sector implement the plan simultaneously more important and challenging. 1.19 There has also been a set of countervailing factors, both historical and more recent in origin, that have mitigated the weakening of the Brazilian state's institutional capacity. Historically, parts of the state, such as some of the decentralized entities like Petrobras, and BNDES, have been more or less insulated from the perverse impacts of patronage and clientelism, and occupied by well-qualified professionals. These "pockets of efficiency" served, to some extent, as sources of technocratic capacities within the state, and in many cases, contributed to development of bureaucratic career structures.25 Although professional capacities in different parts of the Brazilian state vary, bureaucratic competence at the federal level compares favorably to its counterparts in most of the neighboring countries. Civil servants are relatively well remunerated and professionalized (especially those in special career streams such as planning, budgeting, finance, tax administration, and law), and there appears to be a reasonable degree of stability in the high level civil service. These developments are products of the long history of administrative reform in Brazil, which forms another important dimension of the historical background to the current PPA initiative with its clear attention to managerial reforms of the federal public administration. Administrative Reform in Brazil since the 1930s 1.20 The managerial reform intended by the PPA occurs in the context of a country whose Federal government has developed a system of "bureaucratic" (or Weberian) administration, but which has tried on some occasions, most recently under the first Cardoso administration, to move towards greater performance orientation. 1.21 Serious administrative reform has been a feature of Brazilian public life since the 1930s, though the direction and type of reform have changed, sometimes reversed, over time. The basis for a modernized public administration was laid in the 1930s. Systems for a meritocratic civil service, public procurement, and budget control were established. From the start, the rules of what Bresser Pereira (1999) calls a "bureaucratic reform" emphasized a rigid control of processes, and there has been a more or less permanent effort to further reform the administrative process. Success in this has been limited, until recently at least. 1.22 During Brazil's "miracle" years, administrative reform was colored by the "developmentalist reform" of 1967 and took a different turn. Planning, decentralization, and delegation led to the growth of state-owned enterprises and autonomous agencies ("Autarchies" and Foundations). Meritocratic principles in the central administration were de-emphasized, while the new agencies outside the central administration came to be a source of qualified senior public servants who were "borrowed" by the central administration. The strictly Weberian approach of the 1930s was replaced by one that was more informal and one that emphasized "islands," rather than system-wide performance. Some efforts were made from the 1980s to recuperate a meritocratic civil service by developing specialized careers, but the dominant perception was that increasing patronage and clientilism were undermining the quality of the administration. ' For an account of the roles of these technocrats during the military dictatorship, see Ben Ross Schneider, Politics within the State: Elite Bureaucrats and Industrial Policy in Authoritarian Brazil, (University of Pittsburgh Press, 1991). 6 1.23 In response to this, the 1988 Constitution went in a different direction and re- established merit as the criterion for new civil-servant entrants. However, the measures introduced for building meritocracy actually reinforced the inflexibility of the system by giving job stability and generous benefits to public servants at all levels of government, and thus severely limited the power of the Executive to restructure the administrative apparatus or make it more performance-oriented. 1.24 The Collor Government in 1990-92 was concerned with macroeconomic crisis more than the quality of the public service. It cut public-servant jobs (though there was a lot of successful resistance to this) and it froze salaries. The Government also severely reduced the number of ministries, only to have the next Government (of Itamar Franco) reverse the course. Meanwhile, the prestige of civil servants was damaged and administrative capacity was substantially weakened as a result of the economic crisis that began in the 1980s. 1.25 Administrative modernization came back on to the agenda in a new way in 1995, with the creation of the Ministry for Federal Administration and Reform of the State (MARE), under the first Cardoso Government. MARE developed an ambitious approach to putting the house back in order, enshrined in its "Plano Diretor." The Government achieved important constitutional reforms to reverse some of the inflexibilities of the 1988 Constitution. It also did much to re-establish discipline and merit in the civil service. On another front, the Government made some notable advances in transparency through various information technology initiatives. 1.26 But MARE made less progress in its most ambitious objective of moving from a "bureaucratic" to a "managerial" administration, principally by radically altering the structure of the federal executive. The strategy took its cue from the leading reforms then going on in some OECD countries (especially the United Kingdom), but the design of the reform was substantially adapted to the Brazilian situation. New organizational entities would be created to hive off large amounts of government activity: "Executive Agencies" would be a key service arm of the government, enjoying considerable managerial autonomy and linked to ministries through performance contracts; "Social Organizations" would be created under private law to provide contestable services of a public nature with partial public subsidy. In the event, few such new entities have been created, partly because the government was not prepared to provide priority in the flow of budget funds to such agencies. But it may also be the case that changing the management culture through changing the organizations was too frontal an approach to reform, hence too threatening to existing interests in the bureaucracy. While it did not move as far as it had hoped towards creating a results-oriented culture, the government's reform efforts of 1995-99 did succeed in a recuperating the effectiveness of at least parts of the public administration (as well as of public financial administration). They also succeeded in putting further performance-based managerial reform firmly on the agenda. 1.27 Brazil today has a Federal administration that is still far from perfect, but which, compared to many of the countries of the region and to middle-income countries more generally, is in many ways well endowed. 7 o The government has a cadre of qualified and professional middle- and higher-level staff, thanks to a merit-based system of careers and reasonable pay levels.26 Yet numerous imperfections remain in career rules and salary systems. For instance, horizontal (inter- ministry) and vertical (promotion) mobility are impeded. While some federal public servants were found to be receiving comparatively generous compensations vis-h-vis their private-sector counterparts, the competitiveness of salaries with the private sector has been eroded since the mid-1990s due to the salary freeze. This has affected some careers more than others. One result of these imperfections is to accentuate the unevenness of levels of capability and professionalism across ministries. o The improvement in fiscal stability since the Real Plan has allowed a recovery in the budget and financial management system as an instrument of managerial control. o Brazil's federal ministries are generally effective, relative to their homologues in many other countries of the region. For better or worse, the ministries are marked by different cultures, perhaps reflecting their different objectives and challenges, different institutional histories, and the often distinct characteristics of their professional corps. o The public administration is, of course, subject to particularistic political pressures. Fragmented parties, coalition government, and the strong constitutional role of Congress and state governors are all elements that contribute to these pressures. Yet Brazil's administration does not appear to face the intense level of day-to-day clientilism and politicization that so strongly undermine the administration of some other countries in the region. o In spite of the overall capacity of the administration, in a comparative-regional context at least, its financial and other procedures, ministerial structures, and working cultures have been designed (prior to the PPA at least) to concentrate on form rather than results. This is reflected in almost seven decades of effective application of strongly hierarchical, rule-based bureaucracy and the persistent failure of reform efforts to change that paradigm toward results orientation. Effective hierarchy has probably served Brazil better than the ineffective systems undermined by patronage and clientilism, which typify many countries of the region. But there now appears to be a strong consensus in the leading parts of the government that it is time to transform a bureaucracy organized along Weberian lines into a more managerial format. REVIVING THE CAPACITY TO PLAN: THE CuRRENT PPA INrflATIVE 1.28 It was against this historical backdrop of the tradition of state planning and administrative reform that the Cardoso administration launched its concerted efforts to revive the federal government's planning capacity. The initiative has been driven by the same concern for public sector efficiency and a desire to instill results orientation in the federal bureaucracy that motivated the ambitious administrative reform agenda of the MARE's Plano Diretor. The PPA's approach to pursue these objectives, however, is dramatically different from the Plano Diretor's. Unlike the Plano Diretor, the PPA 26 The Federal government also has a reasonably effective system for making senior political appointments: The DAS system incorporates both senior career officials and political appointments into a limited number of posts, with a uniform structure of remuneration and appropriate qualification requirements. DAS thereby creates an effective corps of senior officials, as well as provides for transparent and meritocratic rules for political appointments. 8 initiative places far less emphasis on organizational structures than on clarification of goals and objectives, linked explicitly to an effort to improve the public expenditure management apparatus. At the same time, many of the PPA's implications for broader managerial reform agenda remain implicit. PPA 1996-1999 and the Brasil em Aio Program27 1.29 It was with the PPA 1996-1999 and especially with its Brasil em Agdo Program (as a key thematic component of the PPA) that the Brazilian government began its current efforts to revamp its planning capacity. The primary goal was to create adequate conditions for embarking on a program of strategic investments essential for the integrated development of the economy. During its four-year period, this plan executed about US$ 40 billion on 42 investment projects in economic infrastructure and social development. These projects were selected "for their capacity to reduce production and trade costs, complete infrastructure links to improve competitiveness in the economy, attract new investments from the private sector and promote partnerships, as well as reducing regional and social inequalities." 1.30 The Brasil em Agdo Program also introduced a new management model which has been further developed in the current PPA. The main features of this new model were the following: * All government actions organized as Projects: Clear objectives, as well as time, cost and quality requirements for each project were specified. * Each project assigned to a Manager: To promote accountability, a manager was appointed for each project to organize actions from various areas into the project, pull together resources and overcome obstacles. * Management Information System (MIS): A computer network interlinked all managers and partners, and allowed immediate access to information concerning all projects, enabling these lead actors to share information, promote partnerships and identify impediments to achieving goals. * Access to resources linked to progress of the project: The MIS also permitted another innovation of the management model - a manager's access to financial resources depended on the progress of the project. 1.31 According to the Government's own assessment, twenty-five of the original 42 projects were completed while either meeting or surpassing the original goals. Given the successful implementation of many of these programs, the Government expanded the Brasil em AVdo to 58 investment projects in 1999. The Ministry of Planning claims that the Brasil em Agdo proved "itself to be the most successful experience ever to be accomplished by the federal public administration."29 In the Government's own view, three factors account for the success of Brasil em A!do; (i) selectivity of the priority projects; (ii) use of "advanced methods" of management; and (iii) public-private partnership. 27 This section is based on a background paper for this report by Shafiq Islam (2000), Brazil - Multiyear Plan: PPA 2000-2003 - a Background Paper. 2 See Ministry of Planning, Budget and Management (2000), p.7. 29 Ibid., p.9. 9 PPA 2000-2003: Avana Brasil 1.32 For the 2000-2003 period, the Government decided to design a new PPA building on key elements of Brasil em AVio, and expanding the model to the entire universe of government actions. While Brasil em AVdo was primarily a core public investment plan (even though, strictly speaking, not all the projects were capital investment projects), the new PPA was to apply its planning and management approaches as a comprehensive development plan that included not only public investment projects but all the recurrent activities as well as those actions that did not involve public expenditures. Innovative features of the PPA 2000-2003 1.33 As a tool for improving the federal government's public expenditure management, the PPA 2000-2003 contains several innovative features, which, if fully implemented, would greatly improve transparency and accountability of government actions. 1.34 Program budgeting and objectives orientation as a means to management reform: In AvanVa Brasil, all Federal Government expenditures are grouped in 388 Programs, each with a clear objective and a set of actions deemed relevant for achieving the objective and performance indicators for monitoring progress in implementation. Budgetary resources are intended to be allocated not to organizational units but to Programs, based on the philosophy that resource allocations should be linked to pursuit of objectives rather than maintenance of bureaucracy. For similar reasons, the architects of the PPA 2000-2003 originally downplayed the organizational aspect, and only recently have incorporated an explicit vision of organizational reforms as an element of its management reform agenda. AvanVa Brasil also integrates capital and current expenditures under the same Program categories, thus reinforcing the objective-based programmatic approach to public expenditure management. By appointing a manager for each program, Avanga Brasil also establishes links between programs and their objectives on the one hand and government personnel responsible for their delivery 0on the other. This arrangement is expected to promote better accountability for results., 1.35 By making government objectives and actions transparent, Avanva Brasil is expected to facilitate partnership between the public and private sectors and between levels of government in the Federation. The same transparency is also expected to improve internal coordination/partnership within the Federal Government itself. Because of the interdependent nature of many of the programs, managers are forced to seek cooperation of other programs/agencies to achieve the program goals. The presence of managers in turn makes day-to-day contact among related programs easier and smoother since they would not necessarily have to let communications go through the ministries' organizational hierarchy. Finally, the transparency aids better social control by civil society and the Congress. 1.36 Along with the new program budget format, the Federal Government has introduced a strong element of transparency by widely announcing the initiative and making 3 However, the current transitional arrangement still leaves room for evading accountability because the Program Managers are not given resources and actual authority to make managerial decisions: their roles are limited to monitoring and coordination/facilitation of actions by those who are directly involved in and responsible for program implementation. This point will be developed below. 10 information available by internet to the entire Federal bureaucracy (through its Management Information System, SIG-Plan) and the public. While the design of the programs is still far from perfect (for example, the technical basis for their costing, definition of performance indicators and so on still leave room for further improvement), the new classification system has made it easier for both public administrators and external agents (e.g., Congress, citizens) to understand what the government does with what resources for what end. Also for the first time, the government carried out a comprehensive evaluation of the PPA programs after the first year of implementation, and reported the results to the Congress. The Ministry of Planning is continuing its efforts to improve its monitoring and evaluation systems. The evaluation was repeated in 2001 with some methodological improvements. 1.37 Explicit link between the plan and annual budgets and consistency with the macroeconomic framework: One of the innovative features is the explicit effort to link the plan to the annual budgets. Many development plans worldwide have failed to achieve their intended results because of typical institutional separation between plans and budgets. For the first time in the history of planning in Brazil, the PPA 2000-2003 has been linked to the annual budget documents (so far in FY 2000 and 2001) and provides them with a multi-year perspective to priority-setting and resource allocations. 1.38 Similarly, the current PPA was developed with a commitment to ensure consistency between its four-year resource envelope and the macro fiscal framework. Developing a plan that is explicitly linked to, and bound by, the macroeconomic projection is an important advance toward effectively linking the plan to budgets and thus assuring the plan's credibility. The importance of this aspect of the new plan is particularly high in Brazil's current fiscal conditions that require constant adjustments to maintain macroeconomic stability. PPA as an indicative planning tool? 1.39 The PPA, as a broad development plan, is also intended to accomplish a wider set of objectives, including gaining political legitimacy for the President's governing proposal, communicating it to society, and translating it into concrete government actions. In addition, both the Brasil em AVdo Program within the PPA 1996-1999 and the PPA 2000- 2003 AvanVa Brasil have been used explicitly as vehicles for promoting public-private and federal-sub-national partnerships, as well as a driver of a fairly ambitious public sector management reform agenda. 1.40 An aspect of the PPA 2000-2003 that gives it an element of indicative planning is the fact that its preparation was based partly on a large-scale technical study (Study of the "National Integration and Development Zones," known familiarly as the "Eixos" study because of its use of the concept of regional development axes).3' This study identified investment needs and opportunities not only for the public but also for the private sector, and contributed to the development of a portfolio of investment opportunities for both sectors covering eight years (2000-2007). 31 This "indicative" aspect of the whole planning exercise is intended to signal investment opportunities to the private sector and guide the government's public-private partnership strategy. 11 1.41 It was also around this Eixos study that the government engaged in a broad consultation process with representatives of ministries, state and municipal governments, the Congress, private businesses, and NGOs. Requests and suggestions submitted at these consultation sessions were reviewed by the technical team that prepared the study, and some of them were incorporated in the final version of the study "within the technical parameter of the study."32 A rigorous assessment of the technical quality of the Eixos approach and its preparation process is beyond the scope of this report. However, it does appear that the study's preparation process is reminiscent of an aspect of French indicative planning: The main practical role of the CGP (Commissariat G6ndral de Plan) has been to orchestrate a complex operation of consultation and participation in discussions exclusively concerned with plan elaboration. The process constitutes the core of what is peculiar to French planning, and many thousands of people are involved in an exercise that was once a major effort to produce a broad social consensus o the nation's medium term priorities. Box 1.1: The "Eixos" Study The Consortium of Consultants produced a R$317 billion portfolio of 822 potential public and private investment opportunities for the period of 2000-2007, which served as one of the principal inventories of ingredients for developing the set of 365 Programs for the PPA 2000-2003. The study estimates that about a third of these investments 'will be financed by the private sector, while the government hopes that the level of private participation will reach 50 percent over the following eight years. Eixos (Zones): An Eixo is defined as a strict geographic and mutually exclusive areas that may cut across political boundaries of Brazil's states. The study divided the country into nine zones, with each municipality belonging to only one zone, but some states split into more than one zone. The preliminary classification of zones was produced by putting together a set of variables such as transportation network, transportation flows, spatial distribution of production and potential for dynamic expansion in a multi-criteria evaluation matrix. This classification was then revised by incorporating, among other things, sector and spatial integration analyses. The nine zones are: Arco Norte, Madeira-Amazonas, Araguaia-Tocantins, Transnordestino, Sio Francisco, Oeste (West), Sudoeste (Southeast), Rede Sudeste and Sul (South). Guided by a set of common and complementary variables such as geographical characteristics, resource endowments, production and transport linkages, these zones are in turn combined into sets of two (except for Rede Sudeste) to create five broader zones: Amazon (Arco Norte and Madeira-Amazonas), Northeast (Transordestino and Sdo Francisco), Midwest or Center-West (Araguaia-Tocantins and Oeste), Southeast (Rede Sudeste) and South (Sudoeste and Sul). Portfolio of Investment Opportunities: The raison d'Atre for each investment project in the portfolio is to eliminate an obstacle to the development and integration of the primary goals identified in earlier stages of the study. The projects were identified and conceived to have multiplier effects in order to generate other investment projects. Integration of related projects was a basic criterion for the conception of other new projects. For example, the construction of a railway may lead to . derivative projects for highways, water-ways, ports, power plants, and 32 Ministry of Planning, Budget and Management (January 2000), Recent Development in Federal Government Planning. 3 Saul Estrin and Peter Holmes (1982), "Recent Developments in French Economic Planning," Economics of Planning, Vol. 18, No. 1, p. 2. 12 telecommunications which, in time, may lead to new investment projects in social development, science and technology and' environmental protection. This synergistic process is also expected to catalyze participation of the private sector in new profitable investment opportunities. on its own or in partnerships with the public sector. Source: Shafiq,Islam (August, 2000), "Brazil Multiyear Plan: PPA 2000-2003 A,Background Paper." PPA 2000-2003's preparation process 1.42 The PPA 2000-2003 was developed over more than two years of careful preparation, building on the lessons learned from the Brasil em Agdo Program in the PPA 1996-1999. The process consisted of intensive staff training and technical/analytical preparation, combined with a series of consultations (mainly with stakeholders within the broader public sector, but less with those outside the public sector). The programs were defined on the basis of certain strategic guidelines from the Ministry of Planning based on President's Strategic Orientation. This was discussed and defined on the basis of the President Cardoso's governing plan as a presidential candidate during the electoral campaign for his reelection, and composed of the following elements: four "strategic directives," 28 "macro objectives," and five "agenda." 1.43 Strategic Directives (Diretrizes Estratdgicas): One key component of the final formulation of Strategic Orientation of the President took the shape of a set of four Strategic Directives:35 * Consolidate economic stability and sustained growth * Promote sustainable development aimed at generation of employment and income opportunities * Combat poverty and promote citizenship and social inclusion * Consolidate democracy and defense of human rights 1.44 Macro-objectives: These strategic directives were further disaggregated into a set of 28 Macro-objectives. These objectives range from topics as diverse as reform of public finances to improvement of health, education and housing to increasing exports. Each macro-objective contains a set of three to five broadly formulated goals and actions of the government. For example, the actions under the macro-objectives of "Combat Hunger" are: V Eradicate malnutrition among children under two by means of provision of foodstuffs together with basic heath services. V Distribute basic food baskets where hunger problems are most acute, in partnership with states and municipalities, and preferentially using local products. V Promote nutrition guidance in the government health and education programs. 3 Further details are available in "Recent Developments in Federal Government Planning" Ministry of Planning, Budget and Management, January 2000 35 The National Congress added two more Strategic Directives which were included in the Budget Directives Law that shaped the preparation of the PPA 2000-2003 and the Federal Budget for the year 2000: Reduce interregional disparities; and Promote rights of minorities who are victims of prejudice and discrimination. 13 These 28 macro objectives became the closest thing the PPA 2000-2003 has to substantive expression of the government's strategic priorities. 1.45 Agenda: The other main part of the Strategic Orientation is five "agenda" or challenges. These are cross-cutting themes that government actions should address in pursuing specific, substantive objectives identified as macro-objectives. , National Integration and Development V Reform of the State / The Environment / Employment and Income Opportunities / Information and Knowledge Simultaneously, each ministry prepared its own Strategic Orientation to guide the PPA program elaboration. It is not clear from the available documentation, however, how these agenda and ministries' strategic orientations influenced and are reflected in the actual program design. Table 1.1 summarizes the PPA 2000-2003 preparation process that spanned about 3 years. Table 1.1: Main Steps in Preparation of PPA 2000-2003 Avanfa Brasil Date Steps Jun 1997 1. Study of national integration and development zones commissioned Mar 1998 2. Training of planning and budget officers 3. The Planning Secretariat internal seminar Oct 1998 4. Decree 2829 as a legal basis for reorganizing all government actions into programs. 5. Series of manuals prepared to guide the PPA preparation process Nov 1998 6. A new budget classification system introduced Dec 1998 - Mar 1999 7. Inventory of actions Jan. 1999 8. Creation of the Ministry of Planning, Budget and Management Feb - May 1999 9. Mobilization for elaboration of the PPA 2000-2003 (including definition of the President's and ministries' Strategic Orientations) Mar.- Aug 1999 10. Discussion of the results of the "Eixos" Study Apr 1999 11. Training of more than 500 technical staff for program elaboration Apr - Jul 1999 12. Elaboration of PPA programs 13. Allocation of resources to PPA programs within the projected fiscal resource envelope Aug 1999 14. Consolidation of the PPA and the presentation of the PPA Law to the Congress together with the 2000 annual budget law 15. Launching of the Avanga Brasil web site Sep - Dec 1999 16. Appointment and training of managers 17. Reorganization of the SPI Jul 2000 18. Congressional approval of the PPA 2000-2003 Law SUMMARY 1.46 The PPA 2000-2003 is a unique and serious attempt to use techniques of planning to transform the Brazilian Federal bureaucracy into a modern, results-oriented entity that 14 effectively provides public goods and services that are demanded by citizens. Its goals of revamping the state's capacity to plan and instilling a culture of entrepreneurial management in the public sector are ambitious, yet its approach to implementation is advisedly cautious. Its conceptual design is built explicitly on successful experiences of the previous PPA, with which a series of management innovations were successfully piloted. Its implementation strategy does not explicitly seek to rationalize or re-organize the existing government structure, which may be seen as threatening by those affected (i.e., line ministries). Instead, it attempts to introduce greater performance orientation through clarification of government actions, their objectives and performance indicators, and through various measures to enhance transparency and coordination. Possible organizational reforms in line with the PPA model are left to each ministry's initiative. 1.47 The "model" is multi-faceted and highly complex, as is the Brazilian Federal bureaucracy. It is also an evolving process, in which the Government is constantly introducing adjustments to the model's design as it encounters practical implementation challenges. For example, learning from the relatively poor performance in protecting strategic programs in 2000 (the first year of PPA implementation), the Government reduced the number of strategic programs/actions from 80 to 52, and devised a meticulous mechanism to ensure that available resources flow to well-performing strategic programs (discussed below). While there are weaknesses in the model, as discussed below, this flexibility may prove to be one the greatest strengths of the current PPA initiative. 15 2. ASSESSING THE PPA PPA AND MANAGEMENT REFORM 2.1 A primary objective of the PPA 2000-2003 is to improve impact of government programs through a more results-oriented, "managerial" public administration. Explaining why the government launched the new planning initiative, the Ministry of Planning web site emphasizes that "the society wants public administration with results orientation, which translates into more food, employment, security, schools, hospitals, housing, water and sewage."3 Thus, we have first examined the extent to which the PPA 2000-2003 is catalyzing changes in program management in the federal government. 2.2 The management "model" of the PPA is by no means fully spelled out. But the principle building blocks appear to be as follows. First, the division of public business into Programs is a catalyst for advancing a culture of performance orientation because the level of information to the public about government spending and its objectives is raised and because business comes to be organized more around the final needs that government serves than the inputs that go into them. Second, these Programs are linked to budgetary resources in order to ensure that these final needs can be effectively served. Third, Program Managers will be able to contribute to meeting these final needs because they will be able, with the Planning Ministry's support, to work across sectoral and bureaucratic boundaries. Fourth (as already discussed), monitoring and evaluation can contribute to better Program management. This "model" so far lacks a particular vision of the organizational dimensions of management reform, which contrasts starkly with the ambitious management reform agenda pursued under the Plano Diretor during the previous administration. 2.3 To gauge the impact of the PPA on the Federal administration, we examined in some detail four Programs in two sectors, health and transport. Four cases are too few to be representative of the 388 programs, but they do give a clear flavor of the issues raised by the management model across a broader canvas. Our cases suggest that the PPA is proving a catalyst in advancing a culture of performance orientation: budget and budget execution information is becoming more useful to people inside and outside government; the area for political discretion is being reduced by better definition of spending items; and public servants are getting more into the habit of thinking about the final needs they are supposed to serve. But these observed effects are modest (and impossible to measure); they are stronger in some Programs than others; and the PPA is no doubt only one of a broader set of factors (including public demand) impelling the changes. SECTORAL DIMENSIONS OF THE PPA PROGRAM MANAGEMENT Synopsis of PPA Programs in Health and Transport37 2.4 This section provides a summary comparative assessment of the four programs that we examined to gain insight into concrete managerial challenges of the PPA model. The Programs examined are: the Family Health Program and the Outpatients and Hospital 3 "Por que o Avanga Brasil?" at the Ministry of Planning Avanga Brasil web site, http://www.abrasil.gov.br/anexos/anexos2/index.htm. 3 See the Annex Chapter 1 for more detailed descriptions of these Programs. 16 Attendance Program in Health and the Araguaia-Tocantins Corridor Program and the Maintenance of Federal Road Network Program in Transport. The mini case studies reveal interesting differences and similarities among them along several dimensions of program management. Annex I provides a more detailed description of these Programs and assessments of their management. Program design and characteristics 2.5 First, the cases represent different types of programs, though by no means all the possible variations are captured in four cases, and thus different managerial challenges. The Family Health and the Araguaia-Tocantins Corridor Programs have been designed as vehicles for a major reform or developmental agenda in each sector, while the Outpatients and Hospital Attendance and the National Road Network Maintenance Programs are more routine in nature, though they, especially the latter, also contain some reform elements in improving how these routine activities are to be carried out. 2.6 With the possible exception of the Corridor Program which was born partly on the basis of the Eixos study, the origins of these Programs predate the PPA. The Family Health Program was born as an experimental community health agent program in 1994, and, on the basis of its own success, progressively grew into the MOH's core reform program that supports decentralization of primary health care and policy shift from curative to preventive medicine. The Program Manager perceives that the PPA has given the Program greater visibility and better guarantee of resources. Similarly, the Road Maintenance Program and the Outpatients Program did not owe their existence to any strategic planning exercise linked to the PPA initiative. These had existed (albeit not necessarily in the same names) as routine government activities that needed to be carried out irrespective of the government's 38 particular policy orientation in the respective sectors. 2.7 Architects of the PPA intended to introduce greater transparency and programmatic coherence to the internal structure of these Programs through better specifications of program objectives, physical execution targets, and "actions" to pursue them. But in a number of cases, the advances appear to be limited, leaving scope for additional tightening of program design. To the extent the PPA is inducing strategic and managerial improvements at the program level, it is likely to be through individual efforts of Program Managers, aided by the greater transparency. In those cases where major changes are taking place, it appears that the impetus for change comes not primarily from the PPA but from the broader dynamics of sectoral reform, such as decentralization in the health sector, or greater private-sector participation in the transport sector. Program execution 2.8 One aspect of the PPA program management that is clearly affected by sectoral and program characteristics is the financial execution of the Programs. Budget execution is a useful first-cut indicator of program management performance, as low budget execution may be a result of poor program management (which could only be ascertained by a 3 Again, specific ways in which road maintenance activities were to be carried out have been re-examined in the Ministry of Transport, which has resulted in several innovative approaches to the task. Nonetheless, the task of maintaining the federal road network itself is not a new one that originated in the PPA. 17 program evaluation), or it may be a cause of the program's failure to achieve its implementation goals (which could be identified by carefully examining budget execution performance of different sectors and types of programs). As discussed in the section on expenditure management, fiscal adjustments, manifested in terms of in-year downward adjustments of expenditures, are facts of life in today's Brazil. The two sectors under review here illustrate contrasting expenditure dynamics that are likely to affect program execution differently. Invariably, social sectors are better protected because of the explicit prioritization given to them in the 1988 Constitution. Transport programs, on the other hand, tend to suffer from expenditure adjustments far more than health programs or an average strategic program, even though 12 out of the 30 MOT programs, representing 90 percent of the MOT budget, were designated as strategic. Some of the key findings are summarized below and in Table 2.1. * In the budget execution decree, the level of spending released to the MOH in 2000 was 98 percent of the level authorized by the budget, compared to 51 percent for the MOT programs.4 * The combination of increases in authorized spending and the low level of funds blocked by the budget execution decree meant that by the end of the year the health ministry had executed 16 percent more than the initial discretionary budget. The transport ministry executed 18 percent less, despite significant increase in year-end spending. * Because the Health Ministry's Programs are dominated by recurrent expenditures such as personnel costs, pensions, and formula-driven transfers to sub-national governments or SUS facilities, the flow of expenditure across the year is reasonably regular. As a result this Ministry did not show the same acceleration in end-of-year commitments and accruals as did the Ministry of Transport. * In Transport, in contrast, there was a significant unblocking of funds towards the end of the year, when the Finance Ministry was more certain of their availability. This led to a rapid increase in commitments (42 percent) and accruals (95 percent) at the end of the year, compared to 28 percent and 59 percent, respectively in Health. For instance, accrued spending in Transport almost doubled from the level reported in mid-October to that of mid-February, 2001. 3 In any government, capital expenditures tend to be more volatile than recurrent expenditures simply because of the nature of capital investment projects. In the case of Brazil, the constant need for fiscal adjustment (discussed below) has become an additional ingredient that has an effect on resource flow to programs. 4 The figure for all social sectors was 96 percent, infrastructure sectors 63% and other sectors 58%. 18 Table 2.1: Program Execution in Ministries of Health and Transport, Fiscal Year 2000 (as reported in mid-February 2001) Health Transport AIL Strategic All other Direct Progriams Programs* 1. The budget execution decree - spending allowed as percent 98% 51% 90% 97% of budget law (discretionary spending only):41 2. Spending allowed as percent of budget law (discretionary 116% 82% 68% n.a. spending only):42 3. Year-end increase in level of spending commitments (as 28% 42% n.a. n.a. between reports of early October 2000 and mid-February 2001: 43 4. Year-end increase in accruals (as between reports of early 59% 95% n.a. n.a. October 2000 and mid-February 2001:" * Except Special Operations Source: World Bank estimates based on Senado budget data base and Decree No. 3473, 18 May 2000 Organizational arrangements, accountability and the role of managers 2.9 There is also an interesting difference between the two Ministries in terms of the extent to which changes in sectoral organization and accountability are moving hand-in- hand with the PPA's programmatic logic. With the PPA 2000-2003, the MOT has adopted a "matrix" approach, whereby the new Secretariat of Development houses most of the "development-oriented" Programs (i.e., all of the nine regional corridor programs), whose implementation requires coordination among various "input" units (e.g., Secretariats of Road and Water Transport). The MOT is currently going through a wholesale re- organization to split policy-making/coordination functions of the Ministry more clearly from regulatory and policy/program execution functions of the semi-autonomous regulatory and executive agencies. While this initiative apparently has an origin different from the PPA 2000-2003, its philosophy is consistent with the way in which the MOT has adopted the "matrix" approach to managing the PPA Programs.45 2.10 In the health sector, in contrast, the PPA Programs have apparently been defined in such a way as to fit the existing activities and organizational structure of the Ministry. Both Programs examined in this report are managed by Directors with line management responsibility. Although the Manager of the Family Health Program effectively acts as a "super" manager who supervises and coordinates four other, much smaller but closely related, basic-health-care Programs (covering nutrition, Hansen's Disease, chronic degenerative diseases, and tuberculosis), she does so in her capacity as Director of the unit and not as PPA Program Manager. The Manager of the Outpatients and Hospital Attendance Program is also the Manager of a Ministry Department. But the Program and the Department appear to work more in isolation from the rest of the Ministry than Family Health. The Manager of this Program did not perceive much to have changed as a result of the PPA. 4'Limite Austorizado/lUi. 42 42 (Autorizado - Bloqueado)/Inicial (NB percent can exceed 100% because authorized spending rose during year) 3 Increase in Empenhado. 4Increase in Liquidado. 4 At the same time, it is interesting and ironic that the kind of organizational changes proposed in the MOT are akin to the organizational change strategy pursued by Piano Diretor of the previous government. 19 2.11 The differences in the program characteristics and the ministries' approaches to incorporating the PPA's programmatic logic into their organizational arrangements are also reflected in different roles PPA Program Managers are playing. On the one hand, Managers in the transport sector see their roles primarily as facilitators, or "fixers," whose main responsibility is to facilitate intra-organizational coordination (among different "input" units) or clear bureaucratic hurdles that impede program implementation (e.g., obtaining environmental licenses, or securing necessary additional funding). The two Program Managers in the health sector, in contrast, have the potential advantage of being line managers with authority over line staff. But, the extent to which they actually capitalize on this advantage in managing their respective Program seems to differ, with greater reform orientation visible in the Family Health Program. 2.12 One indication of the way in which the PPA's approach is changing the managerial cultures at the program level is the use of monitoring and evaluation. In this regard, the two ministries face starkly different conditions. In general, outputs in the transport sector (e.g., a bridge built) are far more readily observable and measurable than in the health sector. Nevertheless, there was evidence that at least the Department of Basic Health, where the Family Health Program is housed, is developing a culture of information and evaluation. It makes good use of the information system for managing the PPA: the Manager of the Family Health Program stresses the importance of systematizing program-level information. Last year the Ministry of Health evaluated three Programs, including Family Health, and it has now commissioned a study to evaluate PAB. It was not clear to what extent the other Programs were also strengthening their systematic use of monitoring and evaluation. Box 2.1: Monitoring Implementation: The Family Health Program In the Family Health Program, the biggest Action, the transfer of resources (PAB-fixo) to municipalities with full responsibility for basic health care, is inherently difficult to monitor. The physical target, number of people benefited, is difficult to verify since the service is a per-capita transfer to municipalities and it is difficult to establish whether these resources get to finance basic health care without leakage or diversion and are spent efficiently. The Ministry is, however, making progress: an integrated information system for basic health care, which provides information on health teams and family-health levels, works well, though the Ministry can still make further progress in establishing the accuracy of the statistics that the municipalities feed into the system. The Ministry has some power to withhold the incentives-based transfers it makes (PAB-variavel, the second largest Action), but can do little to suspend the PAB-fixo transfer (and has to rely on public auditing bodies to call for action to be taken). 20 Table 2.2: Summary of the Experience of Four PPA Programs in Two Sectors Transport Program: Transport - Health Program: 1 Health Program: Regional corridor Program: Saude da Familia Atendimento (ag. Araguia- maintenance Ambulatorial,- Tocantins) (e.g. ManutenVdo da Malha Rodoviaria) _ Program structure_ Nature of Program Time-bound Permanent (routine) Core of major long- Routine activity: (investment activity; specifiable term reform automatic pass- projects); intra- product; direct. initiative; product through of money; sectoral difficult to monitor; little monitoring of coordination; requires coordination effectiveness. specifiable product; with local direct. governments. Origin of the Program strategy Reaffirmation of Program strategy Reaffirmation of Program evolved as fruit of existing routine evolved over time as existing routine long-standing activity, with some consequence of activity. planning tradition. PPA attempts to decentralization SEDES as key PPA improve policies. process and strategic interlocutor. learning. Better accountability PPA moderately PPA moderately PPA improves PPA has little effect. through limits discretion. limits discretion. visibility of Program transparency? somewhat. Greater productivity Improves service Objectives clarified, Program apparently Program less reform- through better provision by but impact so far already well oriented and PPA did definition of improving intra- limited. conceived and PPA not change this. business? sector coordination. did not change this. Program management Changes in sectoral Better coordination of activities through a Key Programs were defined so as to fit into organization and matrix approach related to PPA Programs, but existing organizational structure, i.e. the accountability "input" units continue to control resources. sector forewent more fundamental change. Sector restructuring is proceeding independently of, but in spirit of, PPA. The role and value Managers are Managers are . Line manager is Line manager is added of the effective facilitators, facilitators/fixers, Program Manager Program Manager. Program Manager especially in intra- perhaps facing and uses Program as The Program sectoral greater bureaucratic umbrella for related remains routine, coordination. hurdles than other smaller Programs. hence little value parts of the sector. added from PPA. The Manager's link Program Managers value relationship with Program Manager Link is weak. to SPI SPI and line to President as channel of values moral support problem solving. (We could not confirm the and information from effectiveness of this channel.) SPI Budget management and prioritization Did sector budgeting Budget & Planning Office did not adopt Incremental budgeting still predominant in change to a Program Program as basis for resource allocation. Budget & Planning Office's actual practice. basis? Were Program Volatile/poor Volatile/poor High level of High level of priorities respected? execution as a result execution as a result execution because of execution because of (N.B all four of large cuts and of large cuts. mandatory nature. mandatory nature. Programs were Congressional Strategic in 2000) amendments. Monitoring & Evaluation The use ofM&E by We are not aware that this has been an Ministry's own MIS Little M&E. Manager, sector, and important tool in the sector (including and M&E are being SPI SEDES). developed (pre- dating PPA). 21 Program outcome 2.13 After only 1.5 years of implementation, it is clearly too soon to evaluate the PPA's impact on each program's final outcomes. But, the first PPA Evaluation that was conducted by the SPI provides a first look at program performance across the board. This evaluation was based on self reporting by Program Managers themselves and ministry representatives. The subjective method used might limit the reliability and validity of the findings as self- reporting generally tends to bias the evaluation in the positive direction on the one hand, and point to exogenous factors as causes of management difficulty, on the other. But the comprehensive evaluation still offers good approximations on how the Programs have performed so far and what may be some of the factors influencing their performance. 2.14 According to the Evaluation report, fully 75 percent of the 343 Programs that were covered in the exercise achieved their targeted results. Forty-three percent of the Program Managers have "high" expectations that their performance goals will be met by the end of the four-year PPA period, whereas another 41 percent have a "medium" level of expectation that they will. The Managers are also quite positive about their achievements in seeking partnerships - 60 percent reported their partnerships with other entities such as sub-national governments and the private sector to be satisfactory. Besides 32 percent reported that their Programs did not involve partnership, thus leaving only 8 percent of the total surveyed with unsatisfactory partnership performance. 2.15 In spite of this strong optimism among the Program Managers, only 27 percent found that the design of their Program was adequate to resolve the problem or address the needs in society, while 10 percent thought it was not adequate. Forty percent thought the availability of financial resources were insufficient, and among them, 12.5 percent of the total thought the resource insufficiency "seriously affected" program execution. Thirty-six percent complained of interruption in resource flow as affecting program implementation. Evidence not captured in these summary figures indicates that there have also been instances of managerial failures (e.g., a road maintenance program that did not execute at all due to mismanagement of the procurement process). However, inadequate program design (i.e., weak planning) and insufficient funding (i.e., possibility of weak budgeting, though it may also be a reflection of inefficiency at the program level) are two of the factors mentioned by an important minority of Program Managers. The PPA and Managerial Outcomes 2.16 What were the outcomes in the four Programs we have looked at? In terms of the four managerial instruments of the PPA, we might venture to note two areas of achievement, one problem area (so far), and one area where the results are not yet clear. * The PPA has been a catalyst in advancing a culture of performance orientation. First, the designation of Programs and Actions is leading to greater public information about what the government is using its money for and reducing the areas of political discretion. Second, the very definition of many of these Programs - transport linkages rather than road building, for instance - emphasizes outputs, not inputs. The performance-orientation effect was more noticeable in the Programs of the transport sector, though not absent in health. These are of course subtle, perhaps modest, 22 advances and they are not the product of the PPA alone: they are occurring within a more general shift of the Federal public sector to performance orientation. * Program Managers have, in two of the four Programs (both of them in the transport sector and typical of the other Programs in this sector), added value in their role of coordinator and facilitator. * The PPA was relatively powerless in 2000, its first year, to guarantee financial resources for Programs. There were significant differences between the programs in the Health and the Transport Ministries in terms of resource predictability (as inferred from the budget execution data). Designations as "strategic" did little to protect flow of funding, if the programs were in the transport sector. 4 * Given the newness of the PPA, it has been difficult for us to fully judge the effectiveness so far of Program monitoring and evaluation. 2.17 One of the most striking impressions from the comparison of the four Programs has been the apparent importance of the sectoral context - the nature of the tasks and the organizations - in determining managerial behavior and Program performance. The Ministry of Transport, or at least parts of it (notably SEDES), seems to have embraced the PPA as an opportunity to pursue a particular performance-oriented reform agenda, that had its root in pre-PPA efforts. Brasil em Agdo gave the opportunity for the Ministry to test the new management model. The Ministry was also well-placed to benefit from the PPA because its "products" are relatively simple to produce, especially investment projects, which benefit greatly from the model of Manager-as-facilitator. Moreover, the Ministry's relatively strong planning capacity - deriving from the predominance of engineering professionals - made it a "natural" partner for the PPA. The MOT is also embracing organizational change. It has proposed a new model separating policy making and planning (to be retained in the core Ministry) from regulation and service delivery (to be handled by attached agencies).47 2.18 On the other hand, the Ministry of Health, like other social-sector ministries, has a weaker planning capacity. One of the two Programs we looked at, Family Health, was certainly as reformist as any of the transport Programs, but it did not benefit as much from the PPA as a strategic planning and management tool. An obvious reason for this is that the nature of the problem to be solved was different, requiring political decisions and management rather than facilitation. But it also seems to be the case that the MOH has been more resistant to organizational change than Transport. Had it been more open to change, it is conceivable that a more radical redesign of Programs would have required different types of managerial expertise. 2.19 Besides the sectoral characteristics, program characteristics may also matter. Although evidence is not clear and definitive, the mini case studies suggest that impacts of the PPA managerial model, which relies heavily on individual Program Managers' own initiatives, may depend on whether or not the program itself contains strong 4 To be accurate, it should be noted that the "strategic" designation was done in the LDO for 2001, issued in 2000. This was after the PPA was already prepared, and was midway into the 2000 budget execution. 47 The Ministries of Environment and of Science & Technology have similarly used the PPA as a catalyst for adopting a more strategic approach and for reorganizing. 23 reform/developmental content. In spite of the expectation that better specifications of objectives and outputs, combined with greater transparency (through more intuitive program classifications and public information dissemination) would lead to more results orientation, the PPA model itself may not provide sufficiently strong incentives for managers to perform and pursue better outcomes. Motivating managers to perform may be more difficult with routine programs than with "reform" programs. 2.20 Our overall impression of the PPA management model in the light of the evidence we have reviewed is that the initiative is well-timed and that it is well-proportioned. It is well-timed because Brazil's Federal government has developed a professional cadre capable of greater performance orientation, the recovery of economic stability in the country now makes performance-oriented management of public expenditures more feasible, and the progress of democracy is creating a demand for better public services. It is well- proportioned in the sense that it is incrementalist and not over-ambitious in its approach to implementation, and it allows different parts of the government to proceed at different speeds. 2.21 By the very token that the PPA model has been cautious, it will also have to respond to several future challenges. We wish to emphasize three such challenges. * First, continuing progress in performance-oriented management will require the government to address the issue of incentives to line managers. The current "model" is of a Program Manager who facilitates across the matrix, but does not take responsibility for managing resources and producing the requisite outputs with them. But performance management is strongly promoted when line managers have their accountability for outputs matched with their authority to produce them. The government has since recognized the importance of the contributions line managers and staff, in addition to the Program Managers, can make to program performance. However, how to integrate the PPA programs and Program Managers within each ministry's organizational structure and management cycles is a pending agenda that needs to be confronted mostly during the next PPA cycle. * Second, responsiveness to the new management model from within individual Programs has depended in large parts on sectoral characteristics and the nature of organizations. The SPI can improve Program design at the margin, but it is not always able to make bad Programs into good Programs, for instance to make them more transparent or to make them promote greater policy reform. Changing what ministries do may ultimately require political decisions, but the evidence suggests that sectors' and ministries' capacities for strategic planning can affect the quality of Programs and the quality of reforms. This capacity cannot be ignored if performance orientation is to progress more evenly across ministries. * Finally, human resources are a problem in many parts of the Federal public service. This topic has not been directly addressed in this report, if only because transport and health did not provide strong contrasts in this respect - both sectors are reasonably staffed with professionals, but compared to some of the central ministries, such as Planning and Finance with strong professional career structures, may suffer from the lack of a cadre of permanent senior civil servants. In health at least, better human 24 resources are part of the solution to the problem of obtaining a better capacity for strategic planning, and instill greater performance orientation. CRITICAL IMPORTANCE OF IMPROVING ALLOCATIVE EFFICIENCY 2.22 One of the promises of the PPA in improving effectiveness of government actions and results on the ground is its potential to provide clearer sense of government priorities and more predictable funding to these priority programs. Resource insufficiency and unpredictability are not the only, or sometimes not even the most important, reasons for poor program execution. Only systematic and deep program evaluation would reveal a mix of factors affecting performance of a particular program. But it is fair to say that without adequate resources and predictability in their availability within a credible hard budget constraint, program implementation will likely suffer. Further, clarity in strategic priorities is essential for efficient allocation of scarce resources (especially in conditions of constant fiscal adjustments as in Brazil). In short, operational efficiency is hard to achieve without allocative efficiency. 2.23 It is clear from the letter of the 1988 Constitution that the PPA is expected to play this role of guiding resource allocations. Thus we have assessed the extent of integration between the PPA and the budget beyond the improvements of the budget classification (including the important fact that the first year of the PPA is identical to the 2000 budget) by examining: (i) the extent to which the PPA clarifies the government's strategic priorities; (ii) the extent to which budgetary resources are allocated in accordance with the stated priorities of the government; (iii) the extent to which budget execution respects the ex ante prioritization, and (iv) the extent to which the evaluation process is likely to contribute to improving allocative efficiency. 2.24 In some cases, specific ways in which programs were designed contributed to important improvements in allocative efficiency as duplications of efforts were eliminated or reduced. For example, before the PPA 2000-2003, there were multiple entities involved in basic sanitation without clear demarcation of the division of responsibilities among them. The explicit program design in the PPA 2000-2003 introduced much improved rationality in assignment of responsibilities and resources across these entities. 2.25 These improvements notwithstanding, based on the implementation record for 2000 and part of 2001, our conclusion is that the PPA initiative for the 2000-2003 cycle has not fully exploited the instrument's potential for attaining tangible improvements in allocative efficiency in spite of the government's concerted efforts to integrate the plan to budgets. There are measures that can be taken - and some have already been taken by the Government (SPI) - to address some of these observed weaknesses within the current PPA framework. There are, however, broader measures outside the scope of the PPA as such, that will need to be tackled in the medium term, including continued efforts to reduce expenditure rigidities. These obstacles are identified at the end of the chapter. Suggested actions are discussed in the concluding chapter of this report. 25 KEY CHARACTERISTICS OF BRAZIL'S PUBLIC EXPENDITURES4 2.26 Before entering item-by-item assessment of the PPA's effectiveness as a resource prioritization mechanism, it is necessary to lay out some of the most salient characteristics of Brazil's federal public expenditures as background against which the PPA has to perform. Two characteristics in particular - (i) the continued need for fiscal adjustments; and (ii) expenditure rigidities - and their implications for the PPA are discussed below. Continued Need for Fiscal Adjustments 2.27 Although Brazil has successfully controlled the historical pattern of high inflation with the Real Plan in the mid 1990s and has sustained its effect into the early 2000s, its economy remains fragile and vulnerable to external shocks. Besides, the effective control of inflation has ironicaly had some negative effects on the fiscal balance through several mechanisms, including:4 * The rise in real interest rates, which have imposed heavy fiscal burdens on the government through the accumulated interest obligations on the very large domestic debt; and * The inability of the government to offset real increases in salary and pension liabilities through inflation. 2.28 Fully aware of these vulnerabilities and the structural constraints to fiscal management (as discussed below), the government has been pursuing a tight fiscal policy, aiming (and succeeding) to achieve primary surpluses of 3 percent or more in the last few years. Thus fiscal adjustments have been the utmost priority of the Brazilian authorities in charge of economic policy management over the last several years. Expenditure Rigidities 2.29 A significant obstacle to the government's efforts at fiscal adjustment arises from the well-known problem of expenditure rigidities. These rigidities have their roots in the constitutional mandates to fund social spending (e.g., health, education, unemployment), transfers to sub-national governments, and wage and pension benefits, as well as from the interest payment for the ballooning public debts. Table 2.3 reports the distribution of federal expenditures by economic classifications of spending. Throughout the period covered, close to 80 percent of the total budget has been rigid for one reason or another, though the relative weights of different items have shifted over time.50 There is evidence that the government's efforts to reduce the personnel expenditures (especially the wage expenditure) and "de-earmark" a portion of the constitutional transfers to states and 48 This section draws from Chapter 3 "Reviewing Expenditures in Brazil," in Brazil: Structural Reform for Fiscal Sustainability, Report No. 19593 BR, World Bank, June 2000. 49 See Eliana Cardoso (2001) "From Monetary and Fiscal Reforms in the 1990s to Sustained Growth," (Annex to the Implementation Completion Report for the Fiscal and Administrative Reform Loan) for a discussion on this so-called Patinkin effect, where real spending decreases under inflationary conditions, as well as for a broader discussion on Brazil's macroeconomic reforms in the 1990s. 5 The figures likely underreport the extent of rigidities as there are expected to be some spending items under the category "other capital and current expenditures" that are not subject to re-allocation, at least in the short run. 26 municipalities have had a positive impact. Between 1995 and 2000, the combined share of the constitutional transfers and wages decreased from 40.6 percent to 34.7 percent. But in the meantime, the persistence in the relatively high interest payment has offset the fiscal gain from the adjustments in the constitutional transfers and wages. Table 2.3: Brazil - Federal Budget Rigidity 1995 1996 1997 1998 1999 2000 Constitutional Transfers 13.5% 13.4% 12.1% 12.5% 11.8% 12.3% Wages 27.1% 25.7% 24.9% 23.3% 21.9% 22.4% Social Security Benefits 25.6% 26.2% 25.7% 25.3% 24.7% 25.4% Interest 12.0% 12.2% 11.8% 15.0%. 19.3% 14.9% Total Rigid Expenditures 78.2% 77.5% 74.5% 76.1% 77.7% 75.1% Other Current and Capital Expenditures 21.6% 22.4% 25.4% 23.8% 22.3% 24.9% Total Expenditures 100% 100% 100% 100% 100% 100% Source: SOF Note: There are additional rigid components within the "Other current and capital expenditures." According to an estimate by World Bank staff, this seems to ranges between 3 to 5 percent. 2.30 Some of the components of the rigidities (e.g., expenditure earmarking to social sectors) are defined at a fairly broad level that would permit a degree of flexibility within them (i.e., thus the Constitutional provision does not prevent the Ministry of Health from shifting allocations from one program to another within its own budget envelope). But apart from these exceptions, the overall effect of the rigidities is to limit the effective reach of the PPA as an instrument for guiding strategic allocation of public resources. This is clearly manifested in the data that show that the programmatic core of the PPA expressed as the 28 macro objectives, including the large social security payments (part of the rigid spending items), jointly account for less than 35 percent of the total budget. Besides, there is no clear indication that intra-sectoral budget allocations are approached in a strategic manner (e.g., following an explicit sector strategy and/or program evaluations). Implications for the PPA 2.31 The peculiarities of the policy mixes in Brazil place disproportionate burden of fiscal adjustment on expenditure control. The government has so far maintained the aggregate fiscal discipline by pursuing and obtaining large primary surpluses. But, the persistent need for expenditure control, in the face of the expenditure rigidities, forces the government to target adjustments in the relatively small discretionary budget, and strains its ability to protect resource allocations to PPA priorities. A number of the PPA programs are unlikely to receive the amount of funding indicated in the original annual budget in a given year. This situation accentuates the PPA's need to be a flexible tool that provides strategic guidance to the policy makers on relative priorities of spending, especially when the expected levels of expenditures available for key programs do not materialize. 27 2.32 While the PPA design could be fine-tuned to allow it to better cope with the resource uncertainty imposed by the fiscal constraints in the short run, in the medium to long run, more permanent solutions to better fiscal management must be sought. Without a system that guarantees a minimum level of predictability in resource flows, performance-oriented public expenditure management is difficult to achieve. The government has taken several measures, including constitutional amendments, to alleviate the expenditure rigidities. The series of fiscal reform measures and the Law of Fiscal Responsibility is expected to exert effective control over fiscal management at all three levels of governments in the federation. But, for the time being, Brazil has no option but to continue to pursue tight fiscal policy and pursue structural reforms such as administrative reform and pension reform. Results of these measures are critical for the consolidation of the PPA. PPA AS A TOOL TO IMPROVE ALLOCATIVE EFFICIENCY PPA as a Priority Signaling Mechanism 2.33 Greater transparency is one of the most important advances of the PPA 2000-2003. The traditional opacity of the budget has been greatly diminished with the introduction of the innovative budget system that follows self-explanatory program structure. Both public managers and outside stakeholders are now able to make much better sense of what the government is doing with its budget, and what it intends to do over the four-year period covered by the Plan. This move toward, and the strong commitment to, transparency is fully in line with the current international practices of good public expenditure management, and in many ways puts Brazil ahead of the other countries in the region. 2.34 There is, nonetheless, scope for improving the PPA model in the area of transparency, which could be addressed both during the implementation of the current PPA (partially) and improved upon when the next PPA will be prepared. First, we find it would be desirable to improve the PPA's clarity in signaling government priorities. Although the PPA document lays out the 28 "macro objectives" and maps the 388 programs under them, this in itself does not clarify what the relative priorities of this government are among the 28 "macro objectives" and the 388 programs. From an outsider's point of view, it is difficult to understand how different criteria used to define PPA programs, such as the "macro objectives," "strategic orientations" of the President and of the Ministries, and the President's five "agenda" are reflected in program design. 2.35 PPA Programs are often presented, for example, on the Avanga Brasil web site, without following a common classification format, such as sectoral, or functional classifications. We appreciate the government's decision to move away from the original functional-programmatic classification system, which was seen as too rigid to represent accurately the reality of the government's expenditure programs. But the traditional functional classification is still useful for international comparison, and at any rate, it should have been replaced with another classification that aggregates expenditures beyond the program level. In this sense, the Evaluation report, for example, lists all the Programs by "macro objectives." This is potentially a useful exercise, but the government has not followed this practice consistently, and has not taken full advantage of the potential that it offers in presenting broader patterns of expenditure allocations beyond the program level for the benefit of outside observers. For example. there does not seem to be any aggregation of 28 expenditure data by "macro objective" even though one could construct such a table using the government and the Senate's databases which are quite useful. The latest version of the Avanga Brasil web site utilizes a yet another set of substantive categories (e.g., "Life in Cities," "A Better Future"), which are rather vague in their content. 2.36 In addition, the PPA is unclear as to how different programs relate to specific sectoral strategies. The general weakness of the ministries' sectoral strategies as basis for the PPA programs has implications for both prioritization during budget execution and program implementation, as discussed below. The PPA Program design was based on a comprehensive cadastre of existing government actions. The Programs were designed by technical staff from SPI, SOF and the relevant sector ministry, but the an adequate strategic program review at the ministerial level did not precede this exercise. Although each ministry's Strategic Orientations were supposed to be one of the various criteria used to define the Programs, the analytical base of these Strategic Orientations is unclear, and its quality is likely to be uneven at best. But, it appears there is further scope for improving the mix of programs and actions by applying more rigorous ex ante prioritization criteria in the next round of PPA elaboration.51 2.37 Finally, the actual indication of strategic programs and actions is made not in the PPA itself but in the annual Budget Directives Law (LDO). This itself is not a problem. In fact, the Constitution itself assigns the LDO, rather than the PPA, the task of identifying the government's strategic priorities on an annual basis. But the way these priorities are defined in the LDO leaves room for improvement. The LDO simply lists a set of strategic programs/actions with no explanation as to why these are strategic programs, how they are linked to specific sectoral strategies of relevant ministries, and how they relate to each other in addressing common problems (e.g., macro objectives), if at all. Neither does the LDO provide estimates of budget allocations to these priority programs, nor (at least in the LDO 2001) an account of how those programs performed in the previous year. Finally, just as the PPA's linkage to sectoral strategies is unclear, sector ministries' participation in priority- setting in the LDO appears to be limited. This could affect the degree of sector ministries' "ownership" of the identified priorities. It would make more sense to let the PPA play the role of articulating what the government's broad priorities are for the four-year period, why they are the priorities, and how the government intends to address them through specific programs. PPA Priorities and Budget Management PPA priorities and budgetformulation 2.38 The Government made great progress toward integrating the PPA with the annual budget in 2000 by presenting to Congress an annual budget proposal for 2000 that was identical to the first-year portfolio of the PPA2000-2003. Despite this important accounting measure, the real linkage between the PPA (and the set of strategic priorities as identified in the LDO) and the allocations in the Annual Budget Law (LOA) is still weak. 51 The Government has, under the technical leadership of SPI, undertaken efforts to adjust the PPA's program mix, identifying a number of actions and some programs to be removed from the Plan after the Plan was approved. The Government has also taken the decision to revise the PPA each year in order to adjust it to the changing fiscal conditions of the country. 29 2.39 Our argument that the PPA itself is rather ineffective as a priority signaling mechanism is reflected in the way in which the Congress approves the budget. In 2000, when the LDO issued a year earlier had not identified a set of strategic programs, a sizable portion of those programs that would later be designated strategic received actually less allocations than the original government proposal. Using the same database from which these figures were calculated, the average funding increase to the non-strategic programs was 10 percent, compared to only one percent to the strategic programs. Although Congress has legitimate powers to shape the budget, including by changing relative priorities among expenditure items, the fact that priorities were apparently shifted between the PPA, a law sanctioned by the same Congress, and the annual budget suggests the failure of the PPA to serve as a tool for arriving at binding decisions on prioritization and resource allocations. 2.40 The situation improved, at least from a narrow and specific point of view of assuring adequate funding to the strategic programs, in 2001 and 2002. By this time, the LDO annex had identified the strategic programs. The portions of the strategic programs receiving less allocations than the government proposal were reduced from 30% of all the strategic programs in 2000 to 20% in 2001 and 2002. At the same time, the portions of the strategic programs receiving more than the government proposal increased from less than 20% in 2000 to around 70% in 2001 and 2002. Although a definitive answer would require a more careful analysis of the congressional budget process in these years, there is prima facie evidence that priority signaling makes a difference in actual resource allocation at the budget approval stage. Table 2.4: Congressional Budget Allocations to Non-strategic and Strategic Programs, 2000-2002 2000 Non-strategic programs Strategic Programs Number Percent Number Percent Received more than Government request 42 16 15 19 Received Government Request 180 70 41 51 Received less than Government request 36 14 24 30 Total 258 100 80 100 2001 Received more than Government request 99 33 39 72 Received Government Request 45 15 4 7 Received less than Government request 157 52 11 20 Total 301 100 54 100 2002 Received more than Government request 93 32 41 69 Received Government Request 106 36 6 10 Received less than Government request 92 32 12 20 Total 291 100 59 100 Source: World Bank staff calculation based on the PPA program database supplied by SPI. PPA priorities and budget execution 2.41 The government deals with the problem of budget inflation due to the revenue overestimation by Congress to accommodate its members' budget amendments by issuing 30 budget execution decrees which adjust ministries' budget allocations downward, depending on the actual projected resource envelope. With this decree, the execution process becomes a second actual mechanism for setting spending ceilings by ministry (at the levels of both expenditure commitment and cash payment). 2.42 Since non-discretionary expenditures have de jure and/or de facto guarantee of resource allocations, the budget execution decree (e.g., No. 3746 for the 2001 Budget) is essentially concerned with programming execution of discretionary expenditures (Article 1 explicitly excludes from the Decree's purview the "non-discretionary" spending priorities of the budget). This gives the Ministries of Planning and Finance the ability to apply implicit decisions about inter-ministerial prioritization by adjusting ministry budgets downward at different rates.52 To take the extremes, in 2001, the Ministry of Education had its budget allocation (permissible level of expenditure commitment) only reduced to 97 percent of the originally budgeted amount, while the Ministry of Sports & Tourism saw its authorized amount decline to 36 percent of the original budget (Table 2.5). The commitment and cash ceilings set in these decrees are revised periodically through the fiscal year depending on the changes in the revenue performance. In 2000, for example, some of the ceilings were re- adjusted at least five times since the issuance of the original decree. Table 2.5: Allocations of Discretionary Expenditures in LOA and Budget Execution Decree, 2001 (Selected ministries, million of Rs.) Decree/ Agri Educ Social Health Trans Culture Environ Sport & Nat'l TOTAL LOA I Assist -port -ment Tourism Integ. Total 0.59 0.97 0.91 0.96 0.70 0.71 0.77 0.36 0.48 0.87 Strategic 0.84 1.00 1.00 1.00 0.85 0.76 0.87 0.37 0.98 0.90 Rest 0.56 0.97 0.91 0.96 0.53 0.70 0.74 0.35 0.25 0.87 Source: World Bank staff estimates. 2.43 As discussed earlier, resource prioritization is compromised by the expenditure rigidities arising from the extensive earmarking and policy decisions to protect certain expenditures such as debt payment (commonly mandatory), and personnel and health service costs (treatment differs by country). Because of the presence of these mandatory spending items, the amount of the budget available to fund discretionary programs is severely limited. Therefore, it can be concluded that any impact the PPAILDO and related measures has on resource prioritization are rather limited. Given this institutional reality, the most reliable method of protecting resource allocation to a given area is to have it earmarked in the Constitution, precisely as accomplished by the Ministry of Health in the 1998 constitutional amendment, or in specific laws. From one point of view, the constitutional earmarking of health expenditures in a country with known deficiencies in social indicators indicates a strong government commitment to prioritizing health sector programs. But from the point of view of good budget management, this is exactly how the 52 It is the Ministry of Planning that is more interested in allocation aspects of the budget execution decree. Treasury officials are more concerned about keeping the total spending within the deficit target, and are not so concerned about how they achieve this spending control, though at least one former government official believed that at least in the past Treasury used the power to ration cash to make allocative decisions during budget execution. 31 perverse incentives built into the system work against rational allocation of public resources according to government priorities.53 The smaller the discretionary budget and the weaker the budget system's capacity to provide predictable funding to priorities, the greater the incentives faced by individual claimants of the budget to have "their" portions of their budget protected by earmarking. This incentive leads to excess and proliferation of earmarking as a preferred means of resource allocations. 2.44 A third weakness is spending ministries' flexibility to transfer resources between activities during budget execution, and thereby effectively invert strategic PPA priorities. Given the high degree of centralization in budget management in Brazil, whether sector ministries have "excessive" abilities to re-allocate budgeted funds is perhaps a debatable assessment. But, several interviewees referred to the problem of opportunistic budget execution whereby sector ministries commit their funds to "their own" priorities first and force the central ministries, Planning and Finance, to make additional funds available to cover the latter's priorities. We were not able to gather any direct evidence to corroborate these observations, but it appears that one of the significant challenges of government management in Brazil is inter-ministerial coordination, including in the area of public expenditure management. The fact that SPI is placing so much emphasis on the supra- organizational nature of the PPA Programs, and the PPA as a vehicle for improving partnership within the public sector (albeit not only among ministries) is one of the indications of this problem. 2.45 In 2001, the Ministry of Planning is seeking to remedy this problem by introducing a new method for protecting resource allocation to priority programs. With the new method, Cash Flow Control, the Ministry of Planning monitors program execution on a monthly basis and re-allocates the available budgetary resources among the discretionary strategic programs depending on their execution performance. The effectiveness of this new method in protecting allocations to strategic programs and actions remains to be seen. However, at the level of spending-ceiling allocations in the budget execution decree for 2001, the strategic programs as such do not appear to be much better protected than the non-strategic programs. Whereas the decree allocated to the strategic programs on average 90 percent of the originally approved level of funding, it provided 87 percent to the non-strategic programs. 2.46 At least in 2000, the government's performance in executing the budget according to its strategic priorities was also not particularly good. The budget execution data show that on average non-strategic programs had higher execution levels than the strategic programs. A number of the strategic programs were capital investment programs. Some of them had their spending ceilings increased near the end of the fiscal year, which would have affected efficiency of program execution. s3 The Minister of Health revealingly commented at a seminar at the World Bank that while he, as an economist, used to be opposed to budget earmarking, as Minister of Health, he considered as one of his greatest achievements the constitutional amendment to effectively earmark the health ministry budget. This is a telling example of how a behavior that is necessary to adapt to the constrained institutional reality in order to protect a particular interest (in this case of the Ministry of Health's) produces a decision that may be sub- optimal from a collective point of view. 32 2.47 Among the strategic programs, 25 out of the 80 programs/actions designated as strategic in 2000 had the budget execution ratio lower than 60 percent of the authorized amount.54 Fifteen of the 25 programs claimed to have had their implementation affected by problems of resource flow (insufficient, blocked, delayed, etc.). Predictably, programs that are primarily capital investment operations appear to be affected disproportionately by the problems of resource unpredictability and insufficiency. Of the 15 programs whose low execution is attributed to resource problems, five each are in the Ministries of Environment and Transport, and one each is in the Ministry of Mines and Energy (Energia das Pequenas Comunidades), Special Secretariat of Urban Development (Saneamento 6 Vida), and the Ministry of Health (Saneamento Bdsico). The sectoral breakdown of the 25 programs is as follows. Table 2.6: Sectoral Distribution of Low Execution Programs Ministry Poor execution due to Avg. execution Avg. execution resource problems/Poor ratio ratio for the whole execution ministry (No. of programs) (%) (%) Environment 5/7 38 73.8 Transport 5/6 38 66.8 Urban Development/Presidency 1/3 42 68.6 Dev. Industry, External Trade 1/3 43 79.1 Mines & Energy 1/1 49 72.1 Agriculture 1/1 27 80.1 Health 1/1 59 96.4 Culture 0/1 15 83.3 Fiscal Tribunal 0/1 47 99.8 Communications 0/1 57 83.7 Source: http://www.senado.gov.br/orcamento/ PPA evaluation and allocative efficiency 2.48 Among the particularly commendable features of the PPA initiative is the serious effort to introduce some measure of evaluation in the planning-budgeting cycle. While program evaluation is not new in Brazil, to the best of our knowledge, this is the first time a comprehensive evaluation of entire federal government programs was conducted to assess the extent to which government programs achieved their targets, consider whether the programs were well conceived, and identify obstacles to effective program implementation. The evaluation was based on self-administered surveys of PPA Program Managers and ministry representatives, which were reviewed and verified by the Ministry of Planning sector analysts. This is a valuable exercise that should be continued. 2.49 But, the way the whole evaluation exercise was approached leaves room for improvement, especially in exploiting the potential of program evaluation methods as a means to improve allocative efficiency. From the way the PPA evaluation was designed, it is evident that the government's primary interest has been in evaluating how effectively ' Sixty percent is an arbitrarily chosen cut-off point to identify those programs that had substantially lower execution ratio than the average, which was 80 percent in 2000. The 80-percent execution ratio at the aggregate level is itself a very low figure which reflects, partially at least, the problem of budget inflation as discussed above. 33 each program is being managed and identifying obstacles to achieving program goals. Relatively less emphasis seems to be placed on measuring actual impacts of the programs, establishing their actual contribution by analyzing the cause-effect relations between the observed impacts and the program outputs, and assessing efficiency in program execution and sharpening expenditure prioritization by identifying those programs and/or actions that could be eliminated, merged, or otherwise modified to better fit overall strategic priorities of the government or the sectors. 2.50 From a perspective of improved public expenditure management, it would be valuable to have both performance monitoring of programs and organizations, as well as deeper program evaluations. Performance monitoring of achievement of outcomes or objectives can indicate which programs or organizations are performing poorly, but cannot get at the root cause of the problem. For this, deeper program evaluation is necessary. Given the way some PPA programs were designed as supra-organizational, several different evaluation techniques may be necessary. Separate management or performance audits may be needed to assess the organizations that carry out the activities that underlie PPA programs. Performance of the implementing organizations may be a core source of execution problems, but this aspect is not fully addressed through the current PPA program structure or evaluation system. 2.51 Perhaps for Brazil, a relevant question is not so much how to carry out a deeper program evaluation, since the Institituto de Pesquisa Economica Aplicada (IPEA) is well equipped to perform such evaluations, as when to do it (since it is relatively costly and time consuming and thus cannot be done for all programs),5s and how to use the evaluation results to inform the planning and budget process. At the very least, the results of the evaluations must be public, and accessible to Program Managers, sector ministries, SPI, and SOF. Within the planning and budget process, SPI and SOF analysts responsible for ministries (and other government entities) should be the primary users of the evaluation results to scrutinize budget proposals of the entities for which they are responsible. A second point of integration is the Secretaries of Management and Human Resources in the Planning Ministry, which are responsible for staffing decisions. Sector ministries are perhaps the only place where an integration of programs and organizations can occur, where resource decisions (staffing and funding) can be holistically integrated with objectives, agencies, and programs. However, from international experience, sector ministries generally do not use program evaluations to modify spending on programs as much as to improve program or organization management. 2.52 From a perspective of the central planning and budget offices, it is desirable to know not only which programs are working well/poorly but also which organizations are working well/poorly. In most developed countries, programs are designed within organizations, linking program performance to a range of program-specific as well as broader organizational variables (e.g., staffing, management skills, accountability arrangements). The budget process can then serve to pressure improvements in program management within each organization, and certainly minimize additional investment in under-performing programs. The current Brazilian budget process, however, does not promote a holistic view 55 See Figure A2.1 for a suggested scheme for deciding which programs should be evaluated more deeply. 34 of organizations or programs, and thus are likely to under-exploit full potential of program evaluations in both improving resource prioritization and program management. INSTITUTIONAL CONSTRAINTS TO BETTER PUBLIC EXPENDITURE MANAGEMENT 2.53 While the PPA model can and should be improved upon by fine-tuning some of its technical aspects, overcoming the main constraints to fulfilling the PPA's potential and thus to fomenting performance orientation in Brazil's federal bureaucracy will be a major challenge. We have identified three such constraints: fiscally unconstrained congressional budgeting; the structural rigidity of the federal budget; and the difficulty of inter-ministerial coordination. 2.54 The first two constraints, the congress's budgetary behavior and the structural rigidity of the budget, are beyond the PPA's potential to address. Both have essentially the same root in the nature of Brazilian electoral and party politics, and are quite difficult to correct. However, it appears difficult to make significant improvements in planning- budgeting integration without addressing these institutional constraints. 2.55 The third constraint, the difficulty of inter-ministerial coordination, arises from two sources, corporativismo (entrenched bureaucratic interests and rivalry that often results in ministries' lack of responsiveness to central direction for change), and the dynamics of coalition politics, which introduces political dimensions to the inherently complex inter- ministerial policy coordination. Coalition politics, obviously, is difficult to manage. Even though the PPA's architects have placed strong emphasis on promoting synergy, partnership and inter-institutional coordination, and the PPA model does provide a number of facilitating conditions and incentives for coordination, it is not clear to what extent it will be able to overcome those aspects of inter-agency discord arising primarily from partisan political reasons. 2.56 Corporativismo is a reflection of the relatively strong organizational identity of ministries and other government agencies. This is a positive attribute when compared to the common pattern of extremely weak institutionalization of public bureaucracy in most developing countries. Brazil's challenges are clearly different from those faced by other countries like Bolivia, Peru and Venezuela. Brazil's challenges are to seek ways of making various government entities responsive to the central direction for change, while promoting their own policy and management initiatives, in such a way as to improve performance of the entire state apparatus. The PPA model offers promises for introducing such changes, but the degree of ministry ownership of the PPA model and the centrally determined priorities appear to be uneven. The fact that the Ministry of Planning is forced to institute a mechanism like Cash Flow Control to keep ministries from allocating the available budgetary resources to non-priority programs is evidence of such problems. The somewhat awkward relationship between the Ministry of Planning, which is in charge of planning and budget formulation, and the Ministry of Finance, which is in charge of fiscal policy management and budget execution, is a variant of this problem of corporativismo that affects efforts at linking plans and budgets effectively. It is also possible, as is often the case in systems with strong bureaucracy, that the ministries' responsiveness even to their own ministers' policy directions may at times be limited. Reforming strong organizations is always a difficult task, and the PPA model may not be an exception. 35 3. IMPROVING AND CONSOLIDATING THE PPA PLANNING FOR PERFORMANCE 3.1 The PPA is an ambitious attempt at introducing performance orientation in Brazil's federal public sector through a well-considered, cautious approach to implementation. Thanks to the relatively advanced state of the federal public sector, Brazil is as ripe as any developing country, and perhaps even some of the OECD countries, to embark on a path toward performance-oriented public management. The PPA offers one promising path toward that goal. 3.2 The essence of the performance-enhancing framework of the PPA 2000-2003 is built on the following three key features: (i) clear identification of government objectives and priorities; (ii) integration of planning and budgeting for improved allocative efficiency; and (iii) promotion of entrepreneurial management through transparency and partnership and through flexible and results-oriented program management (i.e., not bound by conventional organizational boundaries, at least in theory). These features are interdependent in that achievement of the PPA's ultimate objective - better public services to attend to citizen demands - would require advance in all three aspects. 3.3 Achievements over the past few years have been considerable. With SPI's strong technical leadership, a good institutional infrastructure for performance-oriented planning and management has been put in place composed of the new program structure, Program Managers, the SIG, and the incipient and evolving efforts to strengthen integration of planning, budgeting and management under the principle of transparency and partnership. The ways in which past lessons have been reflected in designs of new initiatives and present difficulties have been addressed with constant adjustments are impressive. The directions that the government aims at with the PPA model and related reform measures are broadly in line with the recent OECD trend in public management reforms. Yet specific ways in which these goals are pursued show Brazilian touches that reflect the country's different historical and institutional contexts. In short, this is a serious reform attempt with a large dose of innovations and strong government commitment to its success. More importantly, the PPA holds a potential to catalyze a range of reform measures, which, if pursued as complement to specific measures to improve the PPA's own design features, could contribute to developing a more performance-oriented institutional environment in Brazil's federal bureaucracy. 3.4 As impressive as the efforts so far are, the PPA model and the way in which it has been implemented is not perfect. There are weaknesses within the model itself that need to be addressed. There are also several external constraints that, while they lie outside the purview of the PPA initiative as such, should still be removed, or at least alleviated, for the PPA to have lasting impact on improving government performance. This chapter discusses both actions for addressing weaknesses in the internal design of the PPA and those for tackling external constraints to consolidating the PPA. The main argument is that the government should consider three sets of actions: (i) those aimed to improve some aspects of the technical design of the PPA model; (ii) additional public management reform that complements the PPA's performance framework; and (iii) broader structural and institutional reforms to remove external constraints to the PPA's consolidation. 36 IMPROVING THE TECHNICAL DESIGN OF THE PPA MODEL 3.5 Over the last few years, the government has made substantial progress in each of the three inter-related areas of the PPA's performance framework: (i) clarifying government objectives and priorities; (ii) linking planning to budgeting; and (iii) promoting entrepreneurial management. But these objectives have not been pursued with equal emphases. There has been more progress on clarifying government objectives with respect to each of the 388 programs than on signaling relative priorities among these objectives that compete for limited public resources. Important initial steps have been taken to permit effective integration of planning and budgeting, for example, by developing a common programmatic classification system and a common inventory (cadastre) of programs and actions. But, further changes in program costing and evaluation are necessary to aid budgetary and management decision-making. The appointment of the Program Managers and use of the Management Information System (SIG) and performance indicators are expected to boost the performance orientation. But the lack of explicit attention to the relations between the Programs and the formal organization of the government bureaucracy limits the PPA's potential to be a catalyst for performance improvements among ministry staff other than the Program Managers and a few who work directly with them. To build on the strengths and address the weaknesses in the PPA design, this section discusses four priority actions: (i) sharpening PPA priorities; (ii) introducing full program costing; (iii) introducing in-depth program evaluation; and (iv) monitoring actual roles of Program Managers. Sharpening the PPA priorities 3.6 Our assessment of the PPA implementation so far has found that the PPA 2000- 2003 is not particularly effective in signaling government priorities. As discussed below, one way of addressing this weakness is to take advantage of the LDO in the annual budget cycle. But, there are also adjustments to the PPA itself that could sharpen prioritization. For example, the PPA could become a document that introduces the government's broad policy priorities with relatively detailed and concrete explanation of a set of strategic programs only. Given the Constitutional requirement that all expenditure programs must be included in the PPA, the plan document could contain an annex that lists all government programs and actions, just as the PPA 2000-2003 does. But it may not be necessary for the PPA to provide detailed cost estimates for all the programs for the entire four-year period. The technical basis of these estimates is suspect, and because of the constant fiscal adjustments and other exogenous shocks to the government's expenditure programs, these estimates are likely to be altered during implementation in any case. At worst, this could become a false sense of precision that the PPA is not capable of delivering in reality. 3.7 Instead of providing what some see as excessive details, the core of the plan document could be a sharper description of the government's rationale for selecting a particular set of programs as strategic, what outcomes it intends to achieve, and how it intends to implement those programs. In this sense, it appears to us that Brasil em Agdo and Projeto Alvorada are more effective in presenting coherent sets of priority programs that together pursue key developmental goals (mostly infrastructure development in Brasil em ACdo and poverty reduction in Projeto Alvorada). As a devise for signaling government priorities, these two programs seem to be more effective than Avanga Brasil. 37 3.8 This is not, however, a proposal for a complete reversion to the PPA 1996-1999. Now that all government actions have been organized by program with performance monitoring measures, future PPAs should preserve this structure. But a vast majority of these programs and actions will always be ongoing, routine ones with limited change content. These programs and actions should be managed with the same transparency and results-orientation as the strategic programs. But unless there are significant changes in program objectives or in the way in which they are managed, it may be unnecessary to highlight them in the PPA plan document, which should, instead, focus primarily on introducing policy changes that each new government intends to bring about. Full program costing 3.9 The PPA's potential for improving program efficiency and strengthening program performance is limited by the incomplete move toward program budgeting because of the exclusion of important cost items, especially the personnel cost, from the program costing This is reflected in the current Brazilian annual budget formulation process, in which resources are allocated to programs, in theory, on the basis of the estimated costs to achieve particular outcomes, rather than to administrative units (Ministries) on the basis of inputs (personnel, etc.) required to run them (Box 3.1). Although the PPA 2000-2003 integrates capital and current expenditures under the same Program categories, thus reinforcing a holistic, outcome-based programmatic approach to public expenditure management and avoiding a typical disconnect between the investment and recurrent budgets, for any given program, personnel and overhead costs are not included. The programs thus represent, roughly, the marginal cost of the program or attaining the 56 objective. 3.10 We are fully aware that the constitutional and other practical restrictions limit the government's ability to move toward full program costing in actual resource allocation decisions (i.e., even if staffing is found to be excessive, downsizing is not easy, if not impossible). But a practice of full program costing would allow re-allocation of resources, including human resources, among programs within the same administrative unit, be it a department, a secretariat, or a ministry. It might also facilitate larger-scale re-allocation across ministries within the same overall staffing level. And if the staffing level is found to be excessive vis-A-vis the total program requirements, it could open up a way toward necessary rationalization of the staffing level and structure in the federal bureaucracy - at least the government would have better information to judge cost-effectiveness of delivering a particular program for considering an appropriate staffing level. 3.11 The PPA's program budgeting holds the potential to catalyze such changes, but without full cost programming, that potential is currently under-exploited. It would therefore be advisable for the government to develop a methodology to reflect personnel cost in a program cost structure on a pilot basis. Further, it would be interesting if some programs within a given ministry could be exempted from the existing personnel rules on an experimental basis to pilot more flexible program cost management. 56 Marginal program funding and performance objectives based on the full-effort of all resources devoted to the program creates a conceptual misalignment. This model seems to assume that there is little relation between inputs other than funding, such as staffing, and outputs or outcomes. 38 Box 3.1: Budgeting in Brazil The broad functional steps in the Brazilian budget formulation process are:, 1. Definition of the macro-economic projections, estimates of public revenues, and plan/set expenditure and debt levels. 2. Development of indicative sectoral limits based on analysis of current operating costs, programs, and activities, The ceilings include several subsets of estimates: a) Personnel costs: A detailed personnel registry provides information on wages, new hires, and early retirements. This is a known number., b)' Debt: Part of debt service for loans related to each ministry's programs is allocated to the Ministry and part,to the Ministry of Finance. The budget office keeps records of all contracts. c) Activities (continuing): The cost of continuing the current programs is estimated by the budget office, and refined by the line Ministry. This includes administrative costs in support of personnel (utilities, etc.), plus mandatory activities (health, pensions, etc.). d) Activities (new or expansions): Line ministries develop these proposals and submit to the Ministry of Planning for review. e) Projects (continuing): Line ministries estimate the cost of maintaining the economic viability of continuing on-going projects. f) Projects (new or expansions): Line ministries estimate the cost of expansion of projects or of new projects. Within the available resource estimates, personnel costs, debt payments, and continuing activities are treated as priorities in the annual budget process, and new activities, continuing projects and new projects are treated as residual. If additional resources are available, they first go to new or expanded activities, then to continuation of projects, and finally to new or expanded projects. However, these items are only the marginal cost of the activity or project. The personnel costs and associated overhead costs (running costs) are separately funded (a and c). This process assures the basic obligations of the state - personnel and fixed administrative costs, as well as transfer payments, receive funding,first. Systematic use of program reviews and evaluation 3.12 One of the laudable attempts as part of the PPA initiative is the introduction of a comprehensive annual program evaluation. As discussed briefly in Chapter 2, the first annual evaluation was carried out primarily to gather self-assessments of program performance and obstacles to achieving results from the point of view of the Program Managers themselves (verified by a ministry official and SPI). In this instance, the government chose to emphasize the breadth of coverage in the evaluation exercise rather than the depth of individual program evaluations. The choice of breadth over depth is an appropriate approach to gain insight into how the PPA as a whole is performing at least at this early stage. 3.13 Similar to the issues raised above with respect to program costing, the way in which the PPA program evaluation was designed and conducted could be expanded to include various differentiated approaches to evaluation in order to take full advantage of the PPA's performance management framework. Most notably, the current approach to program evaluation under-exploits the potential of the PPA's program budget format in improving efficiency in resource allocation among programs. The 388 Programs in the PPA 2000-2003 were developed "bottom-up" on the basis of a cadastre of existing government actions, instead of following a "top-down" method whereby each sector 39 ministry defined its own sector strategy and identified programs that fit its strategic framework. Much of the effort was likely devoted to identifying clear social objectives for each of the ongoing government actions, rather than on eliminating those actions deemed inconsistent with the government's and the sector's strategic priorities. 3.14 It would therefore be useful if the PPA evaluation included more in-depth look at some of the Programs with a view to identifying those programs and actions that might still be rationalized, by either elimination or merger with other programs and actions. The same evaluation could be used to improve the quality and relevance of the performance indicators chosen for each program and action. This does not need to be a one-off exercise of axing a large number of programs and actions, but instead can be an integral part of the PPA cycle whereby each year incremental adjustments are made to gradually improve overall efficiency of resource allocation. Several OECD countries incorporate comprehensive expenditure reviews as part of their annual budget cycle (e.g., UK's spending reviews). Box 3.2: Program Reviews in Canada In the mid-1990's, Canada undertook a comprehensive program review to reduce spending and reallocate between low and high priority programs. The program reviews were built around six criteria: " Does the program serve a public interest? o Is this an appropriate role for the government? o Could this be better done by another level of government - provincial? - municipal? o Could this be left to the private sector or volunteer sector? o Could the program be delivered more efficiently? o Is it affordable? The reviews were organized under the auspices of the Cabinet, headed by a Minister for reform, and involved the participation of other Cabinet Members. While a one-off exercise, Canada has tried to integrate this overall review and prioritization process into their annual budget cycle. Serious economic conditions necessitated this approach, but the results were positive enough to warrant trying to integrate elements of the exercise on an annual basis. Based on Managing Strategic Change: Learning from Program Review. Ottawa: Canadian Centre for Management Development, June 1998. 3.15 Since it will not be possible to evaluate all programs in great depth, the government should develop a set of criteria for determining which programs should be subjected to deeper program evaluation, which would allow a better understanding of causes of unsatisfactory program outcomes. Figure 3.1 offers an illustrative decision tree. 40 Figure 3.1: Illustrative Decision Tree for PPA Program Evaluation Design flaw? Redesign Retrain/hire Correct finance problem Yes Weak HR, preparation, etc.? eYes No Exogenous to Achieved financial execution Program? targets? - No Acieve program objectives?NoIthprga Yes strategic? YeLN Ye Success Evaluate Other remedies 41 Monitoring actual roles of the Program Managers 3.16 Among the most innovative measures to improve accountability and transparency in the PPA 2000-2003 is the appointment of Program Managers and the introduction of the flexible program format in organizing government actions. The PPA model itself does not impose a particular prescribed role on these Managers, and a variety of roles have sprung up depending on the personalities of the Managers themselves, types of program they manage, the ministry's approach to incorporating the PPA into their organizational structure, and so on. Given the importance of sectoral context, as discussed in Chapter 2, we believe it would be unwise to try to define a singular set of roles to the PPA Program Managers. In that sense, the Government's current approach to let sector ministries define their own PPA Manager-ministry relations is a sound one. 3.17 However, a downside of this laissez-faire approach, combined with the apparently uneven commitment to the PPA management model among ministries, is that in some ministries, the PPA Managers may end up being symbolic creations with no real contribution to improving management in those parts of the government. As discussed in the next section, effective implementation of PPA Programs will require more than competent and motivated Program Managers. The whole federal bureaucracy will have to transform itself into a more performance-oriented entity. The PPA Program Managers may be able to play a catalytic role in inducing that transformation. It would then be ideal if the Ministry of Planning developed a better understanding of how the relationships between the PPA Program Managers and the ministry organizations are unfolding on the ground. Periodic surveys of PPA Program Managers combined with independent reviews of these relationships in selected sectors might prove useful. 3.18 In the medium term, this knowledge could contribute to better aligning the PPA's programmatic logic and the existing bureaucratic organizations since seeing these Programs as de facto organizational units for implementing government actions creates a range of practical problems that we do not think can be resolved without incorporating the organizational dimension of performance management more explicitly. In our view, however, the PPA's program logic and the organizational reality of government bureaucracy can be made compatible without major adjustments to the model itself, and a few additional measures to introduce an organizational reform component into the PPA initiative. ADDITIONAL REFORM MEASURES TO SUPPORT TH PPA CONSOLIDATION 3.19 Ultimately, the effectiveness of the PPA's performance management framework will depend on effective and performance-oriented bureaucratic apparatus. The PPA provides the federal bureaucracy with some tools for enhancing its performance orientation. But, the PPA initiative itself does not address a host of related issues, such as a framework for better accountability for results within sector ministries, or a much needed rationalization of the federal budget process (e.g., a draft law to regulate the budget process is pending in Congress). In this section, we address two broad sets of additional reform measures, those related to organization and management of sector ministries, and those related to government-wide budget process, which we see as necessary complements to the core of the PPA's performance framework. 42 Organizational and Management Reform Aligning organizational structures and incentives with the PPA 3.20 So far the government has deliberately left the organizational dimensions of performance-oriented reform out of the PPA initiative. From a tactical point of view, this "stealth" approach to reform may be a sensible one that generates limited perceived "threat" to the existing bureaucratic organizations, which often see reform initiatives from the center of the government with suspicion. But, budgetary resources flow through ministries, and all Programs must be implemented by them. Therefore, the PPA's potential for improving performance in program delivery will depend, ultimately, on how the model will incorporate the organizational aspects of the federal bureaucracy. This will necessitate, sooner or later, addressing the incentives of the line managers and staff to perform and deliver program objectives. 3.21 Such a reform would require, at least two lines of measures. On the one hand, there is a need to reform the entire human resource management system to allow, gradually over time, more performance oriented management of federal staff. For example, there may be a need to re-examine the existing professional career system and the civil service tenure. The careers and the tenure in Brazil's federal bureaucracy have so far played a critical role in professionalizing different streams of public administrators in key government functions. But at the same time, there is a concern that these career streams create internal segmentation of the bureaucracy that works against horizontal coordination. The public management career could prove to be a useful counter measure, but its impact should be monitored and carefully evaluated in due time. 3.22 On the other hand, each ministry should re-examine its organizational structure, management process and accountability arrangement to see if these are in line with the PPA's Program management logic and the philosophy of entrepreneurial management. The development of PPA Programs may mean, in some cases at least, that the existing organizational structure is not the most adequate to coordinate implementation of related Programs and Actions. Each ministry will have to confront the question of how to incorporate the PPA Program Managers more formally within its operational process, and how to assign responsibility and clarify accountability among the Program Managers and other ministry staff. The Ministry of Planning might consider developing a pilot program of technical assistance to a few ministries to learn how to align the PPA's program logic with ministries' organizational structures incentives can be achieved. Fostering ministry ownership of the PPA 3.23 Given the diversity and the complexity of Brazil's federal bureaucracy, and the government's sensible laissez-faire approach to its implementation, it will probably be more productive to let each sector ministry devise specific details of its own organizational and management reform. It was apparent in our review that different sectoral dynamics might necessitate different ways of organizing the ministry to achieve better alignment with the PPA's program logic. Key to a ministry-led organizational and management reform is each ministry's commitment to improving performance and ownership of the PPA's philosophy. 43 3.24 From our limited review of program management, we note that degrees of ownership by sector ministries are uneven, with a few ministries clearly showing strong buy-in, while others seem indifferent. It is conceivable that some of the ministries with relatively weak ownership would develop a stronger and more proactive commitment to the PPA if the program development and selection process were guided by the ministries' own strategic priorities. There would obviously be a need to ensure that the ministries' strategic priorities are consistent with those of the President and that the programs are designed in a way that would satisfy the given technical standard established by the Ministry of Planning. 3.25 While the initial implementation of the PPA may have required a more centralized approach with the SPI exercising strong technical leadership, subsequent iterations may benefit more form increased line ministry own initiatives in early phases of PPA development. As a performance culture develops, providing responsibility and opportunities for self-direction are important corollaries of accountability and performance management. Thus, consistent with the trend in OECD countries, the Ministry of Planning might gradually devolve greater managerial flexibility to sector ministries (especially in human resource management), as they develop demonstrated commitment to and capacity for performance management. For example, reforms adopted in the Austrian budget process in the 1990's focused on reporting requirements, and relied heavily on self- management by ministries themselves (Box 3.3). The Austrian approach is one of establishing a more transparent and accountable framework for Government, building on mutual trust between line ministries and the center of the government and elected officials. Box 3.3: Budget Reporting in Austria "In order to establish compliance between the government programme and the budget frame as well as to guarantee the success of the budget consolidation programme, the federal government decided in the Council of Ministers on 4 December 1996 to establish a task and ministry specific system of budget reporting, which should provide data on different hierarchic levels for different purposes (planning, execution and review), identify weak spots, mention their cause and find solutions. It has to provide: * documentation of the work of each ministry and the implementation of political decisions; * data for performance review for each ministry * tools for strategic management and decision making for each ministry. These obligations are laid down in the latest revision of the Federal Budget Act. For personnel administration, the Ministry of Finance has been providing a monthly staff report since Spring 1996. This report shows the development of personnel expenses (employment and retirement expefises) for federal employees and-regional teachers, the extra-charges, the number of full time equivalents employed, personnel fluctuation and retirements for each ministry, including agencies. It contains target-performance comparisons as well as divergence analysis. Since January 1997, every ministry has to submit abidget report on the development of budgetary expenditures and receipts and important indicators, as well as on progress of consolidation s7 In addition to the Transport and Health Ministries which are discussed in some detail in this report, and in the Annex I, we also interviewed officials from the Ministry of Environment, the Ministry of Pension and Social Assistance, the Ministry of National Integration. 44 measures in their respective fields of competence., This report contains an analysis of the detected deviations and propositions for effective measures to correct these deviations." Source: Public Management Developments in Austria. OECD. 1998 Linking the PPA to sectoral strategies 3.26 One way to encourage greater and more substantive involvement of sector ministries in the PPA process is to use sector strategies as a basis for defining the strategic programs. If successful, such an exercise may also have a beneficial effect in strengthening the ministries' commitment to pursuing the stated PPA program objectives, and to better utilizing the management model that revolves around the PPA Program Managers. For this, sector ministries will have to develop strong strategic planning capacity so as to be able to design their sectoral programs with technical rigor. Definition of strategic programs by sector ministries themselves (in close coordination and negotiation with the Ministry of Planning to ensure consistency with the President's priorities as well) is likely to reduce the possible problem of priority inversion by sector ministries during budget execution, whereby they allocate available resources to their own priorities, which may or may not necessarily coincide with the PPA strategic programs. Budget Process Reform 3.27 As we showed in Chapter 2, the intended integration of planning and budgeting is not yet complete, and the government's institutional capacity to translate its priorities into budget and program execution is still weak. Without effective protection of priority programs during the budget execution process, the PPA's ability to produce better program outcomes is limited. There are weaknesses in the PPA's priority signaling, as well as severe constraints on allocative efficiency imposed by the problems of expenditure rigidities and the persistent need for expenditure adjustments. Some of these obstacles (e.g., structural reforms that require constitutional amendments) are beyond the PPA initiative itself to resolve. But since it is through the annual budget process that resources are assigned to the identified priorities, additional improvements in plan-budget linkage may be possible by further rationalizing the annual budget process. We mention three issues for the government's consideration: (i) introduction of greater strategic focus in the budget allocation criteria; (ii) simplifying the planning-budgeting process and changing SPI/SOF's roles; and (iii) the credibility of the approved budget. Strategic resource allocation at a broad expenditure level 3.28 Once the PPA's priority signaling is improved, a corresponding change in the budget process will be necessary to allow allocation of resources to these broad strategic priorities. Key to using the PPA as a strategic plan of this nature and as a tool to enhance the annual budget process are early decisions on main expenditure envelopes, starting with the overall spending level for the medium term and broad allocations to strategic expenditure categories. What is important is that these resource envelopes be binding on the annual budgetary decisions rather than merely being projections. Adjustments to these medium-term spending envelopes may be necessary at times, but any deviation from them should be explained and discussed in the annual budget process. 45 3.29 To move the PPA more explicitly in this direction, one option is to adopt something along the Swedish model of two-step budgeting in the first phase of which the total spending envelope and allocations to a fixed set of broad expenditure areas are determined before the government and the parliament enter into discussing specific spending proposals (Box 3.4). Decisions on specific programs under each resource area would be left to the second phase of the annual budget process. Within the broad expenditure areas, ministries' strategic plans could be used to set priorities. In Brazil, the current macro objectives may serve as equivalents of the Swedish expenditure areas except that the coverage of the 28 macro objectives is not comprehensive as they cover only about 35 percent of the total budget (excluding debt payments). Comprehensiveness of coverage is one of the basic requirements for a good budget, including for allocative efficiency. Thus in the long run, it will be imperative for the government to increase the portion of the budget that could be meaningfully covered by the macro objectives (or their future equivalents). By the same token, the PPA's usefulness as a strategic plan will depend critically on successes in reducing expenditure rigidities, and thus increasing the effective coverage of strategic allocation criteria to larger portions of the public expenditures. Box 3.4: Sweden's Public Expenditure Areas The Swedish budget comprising about 500 separate appropriations is grouped in 27 Expenditure Areas. This is a method for clearly presenting expenditures in individual policy' areas In Sweden's multi-year budget framework, indicative funding levels for these 27 Expenditure Areas are determined before details of annual budgets are worked out. Furthermore, there- are 13 ministries in Sweden, and thus naturally some Expenditure Areas are shared by more than one ministry, which is assigned an expenditure ceiling for' the specific appropriations that. correspond to it. The Expenditure Areas are: 1. The Swedish political system 2. Economy and fiscal administration 3. Tax administration and collection 4. Justice 5. Foreign policy administration and international co-operation 6. Defense 7. International development assistance 8. Immigration and refugees 9. Health care, medical care, social services 10. Sickness and disability benefits 11. Old-age benefits 12. Family and children's benefits 13. Unemployment benefits 14. Labor market 15. Study support 16. Education and university research 17. Ciltiure, the media, religious organizations, leisure 18. Planning, housing supply, construction 19. 'Regional development 20. General environment and development 21. Energy 22. Communications 23.' Agriculture, forestry, fisheries 24. Business sector 46 25. General grants to municipalities 26. - Interest on-debt 27. Contribution to European Union bidget Source: Jon Blondel (2001), "Budgeting in Sweden", OECD Journal ofBudgeting, Vol. 1 3.30 In fact, a common trend in performance-oriented budget reform in OECD countries is to extend the time horizon of budget process to a medium term (typically two to three years), and use it as a forward expenditure planning process. While the PPA starts from the other end (i.e., starting from a multi-year plan and deriving annual budgets from it), it can also serve as a basis for a Brazilian equivalent of a medium-term expenditure framework (MTEF), whose objective is to provide the government with a medium-term strategic plan within a pre-established macro-fiscal and strategic allocation frameworks. 3.31 The Law of Budget Directives (LDO) could possibly be an instrument with which a broad expenditure allocations can be reviewed, discussed and established with their medium-term estimates in each annual budget cycle. Indeed, the original purpose of the LDO appears to be to set broad fiscal and strategic parameters for the given year's annual budget proposal. However, the absence of an up-to-date budget and financial management law has forced the government to use the LDO as a substitute for such a law. This has made the LDOs rather complex laws in large volumes with multiple purposes. This may have distracted both the preparation process and the congressional debate on the law from the fiscal and strategic parameters themselves. Simplifing the planning-budgeting process and changing the roles of SPI and SOF 3.32 In a performance-oriented budget process, the roles of the central planning and budget offices (i.e. SPI and SOF) should shift from preparing detailed annual budgets and ensuring their execution through close monitoring of its implementation (so-called "rowing" role) to establishing broad strategic directives, reviewing and evaluating program performance, and challenging sector ministries' spending proposals to force difficult but necessary policy and expenditure trade-offs (so-called "steering" role). Since the PPA initiative itself will need continuous adjustments and improvements, as has been done so far, both SPI and SOF will also have to devote important amounts of their attention to these additional tasks besides the annual planning and budgeting exercises. 3.33 But Brazil's planning-budgeting process is very complex. Besides the initial preparation of the Plan, the full PPA process involves annual evaluations of the PPA Programs, annual revisions of the PPA Law, annual preparations of the LDO and the LOA, constant monitoring of program execution, and periodic training and other complementary activities throughout a year. Introducing additional reform measures, including improvements to the technical design of the PPA, while managing the annual planning- budgeting process will be a daunting task. The administrative costs of managing the PPA process appear to be quite high already. The Ministry of Planning so far seems to have been able to manage the heavy work load. But there is a danger that the quantity of the work will eventually overwhelm the Ministry and affect the quality of its work. Such a situation would be somewhat analogous to the fate of the Planning, Programming and 5 Schick, Allen (1997), The Changing Role of the Central Budget Office, OECD, Paris. 47 Budgeting System (PPBS) introduced in the U.S. in the 1960s only to be abandoned later on, though elements of it remained in some departments. A main reason for the PPBS' failure was "excessive bureaucratic paperwork" required to manage the process.59 3.34 A risk of such an effort-intensive process is that it is likely to draw the Ministry of Planning into micromanaging the process (e.g., Cash Flow Control) and away from the strategic "steering" role. While it is not clear to what extent the government itself intends to move the roles of the central planning and budget offices toward this "steering" role, the recent emphasis on PPA program evaluation suggests SPI's recognition of the importance of this role, which would be a step that advances the PPA's performance framework in line with its ultimate objective of creating a more effective government. Thus simplification of the planning-budgeting process is critical both to shift it to greater performance orientation and to make the PPA process itself sustainable over time (i.e., by lessening administrative costs of managing the process). 3.35 There are at least two ways of thinking about how to simplify the planning- budgeting process. From a certain point of view, predominant in most OECD countries which no longer conduct formal planning exercises, the current Brazilian system with separate planning and budget processes can be questioned (though the former feeds into the latter). Some government officials at a workshop on budget and financial management held in October 2000 reportedly wondered whether the PPA, with nearly as many allocative details as annual budgets, was replacing the latter. As one radical solution, the government might consider options for effectively combining the PPA, LDO, and LOA into a single process and document. The Constitution does require each of these documents, but does not specify the form they must take. Perhaps the best means of integrating PPA with the budget is in fact to collapse these exercises into one. 3.36 Alternatively, the government might consider streamlining the role of each legal instrument so as to simplify the whole planning-budgeting process and avoid apparent overlaps among them. This might mean, for example, discontinuing the annual PPA revisions and instead using the LDO as a document to assess why program performance has deviated from the original PPA projections (on the basis of annual program evaluations) and explain how the deviations will be reconciled or corrected in the coming fiscal years through adjustments in budget allocations and other measures. A key premise for this proposal is that a new budget management law will be passed to alleviate the LDO from the burden of defining (and re-defining each year) budgeting and financial management rules. This way, the LDO could become a more focused document that facilitates prior debates on the overall medium-term fiscal parameter and expenditure priorities, including discussion on new policy initiatives (e.g., it could be the document in which broad strategic allocations a la Sweden, perhaps, to the macro objectives would be agreed upon) before entering more detailed cost assignment to each program. 5 "Program Budgeting in the Developing Countries," Annex II to Ramgopal Agarwala (1983), Planning in Developing Countries: Lessons ofExperience. World Bank Staff Working Paper No. 576. 48 Credibility of the approved annual budget 3.37 Given the repeated practices of budget adjustments during the fiscal year, achieving allocative efficiency at the level of approved budget is not a guarantee that priority programs will actually be executed as priorities. Improved predictability of resource flows from plan through budget execution enables managers to manage, and gain the fullest benefits of planning. SPI has tried to address this at the budget execution stage through flow control, which is probably necessary and perhaps one of few measures that is feasible given Brazil's constant need for expenditure adjustments. But such a measure is a sub- optimal solution to strategic allocation of resources. Ideally, the approved budget should serve as a binding and thus credible instrument to guide resource allocations throughout the fiscal year. 3.38 Establishing budget credibility is one of the most difficult challenges faced by most developing and middle-income countries. No easy solution has been found, yet, although some of the causes are well understood. Sometimes divergence between approved and executed budgets arises from unexpected exogenous shocks to the economy (and thus to the government's actual revenue collection) that is beyond the government's control. Other times, it occurs because of the government's weak capacity to control its cash flows or because the unit in charge of budget execution and cash management uses its discretionary power to re-make the budget during execution for political purposes.60 Also in countries with limited institutional capacity, executed budgets regularly diverge from the approved budgets because of the government's inability to formulate a good budget that anticipates all the spending needs throughout the year. 3.39 Most of the reasons that have to do with weak institutional capacity for good budgeting and financial management are not applicable to Brazil. But one of the reasons for the frequent deviation between the approved and executed budgets in Brazil is the exogenous shock. A way to remedy this problem is to prepare the budget on the basis of a conservative revenue estimate and create a large-enough contingency reserve. When the Ministry of Finance feels the need to restrict actual spending levels during the fiscal year, it could first reduce the contingency reserve before beginning to affect programs, especially the direct programs (and especially the strategic programs). But this method is unlikely to be effective against another main source of the deviation between the approved and executed budgets in Brazil, which is the lack of sufficient institutional constraint on budgetary behavior by Congress (as discussed briefly below). TACKLING THE EXTERNAL CONSTRAINTS TO THE PPA 3.40 Among the most challenging obstacles to the ultimate success of the PPA initiative is the set of external constraints that reduce the PPA's intended effect on performance management. We have identified three such constraints: (i) the constant "threat" of fiscal adjustment and the considerable expenditure rigidities that arise from constitutional earmarking and other past fiscal decisions (e.g., debt payment); (ii) budget inflation, primarily by Congress, which renders the approved budget impractical as an expenditure management tool; and (iii) difficulties of inter-organizational coordination in the Federal 6 It has been pointed out to the Bank team that in the past this was partly the case in Brazil as well. 49 bureaucracy (not to mention across levels of government, which we have not touched upon). There are no quick and easy solutions to these constraints. But measures need to be taken consistently and continuously to reduce their effects over time. 3.41 To reduce expenditure rigidities, Brazil has no option but to maintain the tight fiscal policy of the past several years, combined with continued efforts at structural reforms, especially administrative and social security reforms (to reduce the fiscal cost of personnel expenditures and also to make the federal civil service more agile and performance oriented),6' and a fundamental review of the earmarking arrangements. There is also scope for reforming the revenue and expenditure assignment across levels of government - establishing more tight correspondence between functions and financing across levels of government might result in reduction in fiscal transfers to sub-national governments in some cases (e.g., to municipalities).62 As the share of debt payment in the expenditure rigidities has been increasing over the last years (Table A2.2 in Annex II), it will also be important to continue with the policy of targeting large primary surpluses and strengthen the institutional capacity for debt management. 3.42 Introducing greater discipline into congressional budgeting may be a difficult task because of its political ramifications. Also, as a fundamentally political phenomenon, its resolution will probably defy any purely technical solution. Nevertheless, a creative institutional solution that binds both the Executive and the Congress to a pre-agreed revenue estimate prior to considering individual spending proposals may be possible and necessary. Although international experiences in this area are scarce and few concrete examples exist to guide Brazil's search for a solution, a worthwhile idea would include establishment of an entity, either a fully independent fiscal council63 or a collegial body (e.g., commission) that serves as a negotiating table, where a binding agreement can be reached among key budget stakeholders representing both the Executive and the Congress, and perhaps even non-governmental actors, prior to the annual budget appropriations process. 3.43 The third constraint, the difficulty and the importance of inter-organizational coordination, is partly being addressed within the current conceptualization of the PPA model. On the one hand, designing Programs as "supra-organizational" (at least conceptually, if not in reality in all cases) seems to reflect the architects' desire to overcome the traditional ministerial boundaries as obstacles to coordinated actions on the ground. In some cases, programs are designed explicitly as multi-sectoral programs involving more than one ministry as an executing agency, precisely to overcome the existing ministerial boundary as an obstacle to inter-sectoral coordination.64 This is reinforced by the appointments of Program Managers, whose main functions include 61 See Brazil: Critical Issues in Social Security, Report No. 19641-BR, June 19, 2001. 62 See Brazil: Issues in Brazilian Fiscal Federalism, Report No. 22523-BR, June 27, 2001. 63 Such an independent council was suggested by Ricardo Hausmann et al. "Reforming budgetary institutions in Latin America: The case for a national fiscal council," in Evaluation and Development: The Institutional Dimension. Robert Picciotto and Eduardo Wiesner, eds. The World Bank (1998). 6 The Ministry of Planning generally appears quite positive about the performance of these multi-sectoral programs. It would be quite welcome to conduct a systematic evaluation of the set of multi-sectoral programs in terms of their effectiveness in coordinating actions across ministries and achieving the program objectives. 50 coordination of different implementation units within the Program as well as coordination with other related Programs. The extent to which these methods will prove successful is yet to be seen. 3.44 The Government might consider additional measures that could also facilitate inter- ministerial coordination, or at least, minimize inter-ministerial conflict arising from conflicting objectives pursued by different ministries, especially between each of the sectoral ministries and the Ministry of Planning, as reflected de facto in the problem of "priority inversions." Once again, this could be achieved by giving sector ministries greater roles in the PPA formulation process, including program definition and design, and priority setting. The problem of "priority inversion" during budget execution, whereby line ministries may sometimes try to spend the available funds on programs that are not officially considered priorities, could partly be alleviated if the "official" priorities more fully coincided with sectoral priorities. That, in turn, might lessen the need for the Planning Ministry to try to protect priority programs through the budget execution decree, which adds to the transaction cost of managing the PPA process. Instead, SPI and SOF could dedicate more of their resources and attention to monitoring program performance rather than ensuring adequate financial execution, which is only a minimum necessary condition for good program performance. 3.45 It is understandable that when the new PPA model was first being developed, SPI and SOF had to play a strong central coordination and leadership roles as sector ministries were justifiably unfamiliar with the new "rules of the game." But as sector ministries leam and accept the new framework, the need for close supervision by SPI and SOF on program development should lessen. In return, greater ministerial ownerships are expected to strengthen accountability for program management within each ministry, since presumably the minister and senior ministry officials would be more committed to ensuring satisfactory delivery of "their own" program objectives. 3.46 Finally, the PPA's transparency can be a potentially powerful instrument to push the necessary reform agenda on the above-mentioned institutional constraints. For example, by making explicit the extent to which the expenditure rigidities limit the government's ability to allocate public resources to strategic programs, the PPA might raise public awareness of the problems caused by earmarking of expenditures, as well as deficit spending that reduces the discretionary spending in the future. The current presentation of the PPA does not always convey these overarching constraints on efficient expenditure allocations and difficult policy trade-offs that they force upon the government, in spite of its very transparent format. Instead, it might, in some contexts, give precisely the opposite impression that the government has much room for flexibly pursuing policy objectives through a relatively large number of programs. The PPA's transparency as a catalyst for broader reforms in related can and should be exploited further. SEQUENCING THE REFORM MEASURES 3.47 Much of 'the attention to the PPA process thus far has been on developing the technical aspects of the system, including the automated information system, Program Managers, etc. This is appropriate at the early stages of the reform. But with these in place, the focus should shift more to the strategy for institutionalizing the PPA and further 51 integrating this process with the budget and general management processes. Much of the gaps found in this review are precisely in this area of integration. In the long run, there is a need to address broader institutional constraints that arise from the nature of politics in Brazil. These include, as discussed above, the way in which the Congress gets involved in federal budgeting, and the extensive constitutional and legal earmarking which has created strong structural rigidity in the federal budget. We realize these are difficult issues, but the available evidence indicates that the scope for effectively integrating planning and budgeting, and thus improving program management, depends critically on clearing these institutional obstacles. 3.48 Sequencing of reform measures will be important for the eventual success of the whole initiative. It is unlikely that the government will be able to tackle the multitude of additional challenges that this Report has identified (and others we may have failed to diagnose) all at once. But sequencing is more an art than science. We therefore suggest the following ideas only to provoke some thinking on the part of the government reformers. First Step - fimprove Technical Content of the PPA Programs 3.49 A first set of additional measures to improve the technical content of the PPA Programs could be concentrated on refining the program definition. Programs that are sharply defined in terms of their objectives, performance indicators and costs can more easily and systematically be evaluated than those that are fuzzily defined. Better evaluation, in turn, is an important contributor to improving allocative efficiency. To this end, the government could further refine the methodology to cost PPA Programs (so that their efficiency can be evaluated) and introduce deeper evaluation of selected programs within the annual evaluation cycle. The continued development and full deployment of the Management Information System, and its effective linkage with other government information systems (e.g., SIDOR, SIAFI) is also critical. Second Step - Foster Greater Ministerial Ownership and Accountability 3.50 Once inefficient programs are identified, sectors will need a strategic framework within which merits and demerits of different programs can be assessed. Thus, as a precursor to the next round of PPA elaboration, it would be useful to encourage sector ministries to develop explicit sectoral strategies (within a notional hard-budget constraint to force these strategies to be fiscally realistic). SPI could perhaps assess the adequacy of the sectoral strategies, and offer those ministries with adequate strategies greater leeway in defining their PPA Strategic Programs. A set of PPA Strategic Programs could then be derived from these sectoral priorities, supplemented by the President's own priorities. The PPA could then be a more effective vehicle for clarifying to society what the government's priorities are and why. 3.51 There should also be an effort to align sector ministries' incentives to implement PPA Programs with the PPA's overall goals. Critical questions are what kinds of roles the PPA Managers are actually playing in different ministries, how they relate to staff in different parts of the ministry, and how the organizational arrangements of the ministries should be adjusted, if at all, to accommodate the PPA's Program management logic. A 52 survey of PPA Managers and ministries probing these questions, complemented by expert organizational assessments would be a useful approach to clarify these questions and devise, ministry-by-ministry, organizational reforms to accommodate the ministerial structure to the PPA. Third Step -iReform the Budget Process 3.52 In order to strengthen the integration between planning and budgeting, and to let the budget process reflect programmatic prioritization of the PPA more accurately and reliably, it might be necessary to reform the budget decision-making process. A few ideas were suggested above in the section on "Managing the PPA Process More Efficiently." An important additional element would be some measures to discipline congressional budgeting without denying Congress the very legitimate role in influencing the government's budget decisions. An independent review of congressional budgeting might be a useful starting point on this reform agenda. This third step should culminate in approval of a new budget and financial management law, and its effective implementation. 3.53 The broader structural reform agenda, which may require constitutional amendments, should be pursued continuously and consistently. These reforms with high political content cannot be planned ahead. The approach will necessarily have to be opportunistic. But, the need for reforms in these areas (e.g., administrative reform, social security reform) should always be emphasized. 53 Table 3.1: Recommended Measures for Consolidating the PPA Performance Framework Three Pillars of the PPA Measures to improve internal Additional public management Broader structural and institutional Performance Framework design reforms reforms 1. Clarify government objectives o Sharpen PPA presentation by re- Use of annual budget reporting (e.g., and priorities focusing the PPA document to Austria) and program reviews (e.g., explanation of policy rationale Canada) as part of the annual budget and identification of priority process programs rather than detailed o Development of sectoral planning costing and specification of each capacity program 2. Integrate planning and o Full program costing to a Medium-term expenditure o Administrative and social security budgeting for better allocative incorporate personnel costs in framework consistent with the LFR reforms efficiency program costs targets and using broad expenditure o Reform of the congressional o Deeper evaluation of selected categories (e.g., Sweden) budgetary procedures programs prior to budget a Completion of the new budget preparation management/process law 3. Promote entrepreneurial o Monitor PPA Program Managers o Organizational restructuring to align o Administrative reform management with transparency as a step to classify different roles organizational incentives and and partnership and relationships with the accountability more closely with the respective ministry organizations, PPA logic - ministry by ministry and to design a PPA-consistent o Development of sectoral planning organizational reform strategy capacity and sector strategies as bases for PPA program design ministry by ministry and as a means to foster greater ministerial ownership of the PPA o Broader human resource management reform (e.g., possible change to the tenure and the I professional career system). 54