Tanzania - Evaluation of the Poverty Reduction Strategy Paper (PRSP) process and arrangements under the Poverty Reduction and Growth Facility (PRGF)

82133 INDEPENDENT EVALUATION OFFICE INTERNATIONAL MONETARY FUND OPERATIONS EVALUATION DEPARTMENT WORLD BANK The United Republic of Tanzania Evaluation of the Poverty Reduction Strategy Paper (PRSP) Process and Arrangements Under the Poverty Reduction and Growth Facility (PRGF) July 6, 2004 This case study was prepared jointly by the independent evaluation units of the World Bank and IMF, as background to separate but complementary evaluation reports available at www.worldbank.org/oed/prsp and www.imf.org/ieo. The other country case studies conducted in support of these evaluations are also available on-line at the above addresses.  2 Contents Page Abbreviations and Acronyms ....................................................................................................4 Executive Summary ...................................................................................................................6 I. Introduction.....................................................................................................................11 II. Background and Context.................................................................................................12 A. Poverty ......................................................................................................................12 B. Economic and Political Context................................................................................13 C. National Policy and Strategy Formulation................................................................16 III. The Poverty Reduction Strategy Paper (PRSP) Process.................................................19 A. Relevance..................................................................................................................19 B. Preparation ................................................................................................................22 C. Application of the Underlying Principles and Assessment of Change .....................24 D. Preliminary Results ....................................................................................................40 IV. IMF Role and the PRGF .................................................................................................42 A. IMF Inputs ................................................................................................................42 B. Aligning the PRGF with the PRSP ...........................................................................46 C. Movement Towards “Key Features of the PRGF”. ..................................................53 V. World Bank Effectiveness ..............................................................................................59 A. Bank Support During Formulation ...........................................................................59 B. Bank Support and Alignment for Implementation....................................................62 VI. World Bank-IMF Collaboration in Tanzania..................................................................68 A. Coordination and Priority Setting .............................................................................68 B. Joint Staff Assessments.............................................................................................69 C. BWI Structural Conditionality..................................................................................71 VII. Conclusions and Lessons ................................................................................................73 Boxes 1. Priority Areas of the National Poverty Eradication Strategy (1997) ..............................17 2. Definition of Poverty in the PRSP and NPES ................................................................30 3. Lack of Pro-Poor Targeting of Expenditures..................................................................34 4. The Poverty Monitoring Master Plan (PMMP) ..............................................................34 5. Survey Results on the PRSP: The Role of the IMF ........................................................54 6. Poverty Reduction Support Credit (PRSC).....................................................................66  3 Figures 1. PRSP Priority Sector Expenditures, 1998/99–2002/03 ..................................................31 2. Overall Budget Deficit (after grants) ..............................................................................48 3. New Commitments by Sector (FY98-00 and FY 01–03) ...............................................65 Tables 1. Growth and Investment Indicators (Period Averages), 1991–2002................................15 2. Chronology of the PRSP/PRGF Process.........................................................................23 3. Issues Around which the PRGF-Supported Program of 2001/01–2002/03 was Designed ..................................................................................43 4. Financing the Domestic Deficit Under the PRGF-Supported Program..........................48 5. Number of Structural Measures in ESAF and PRGF-Supported Programs for Tanzania by Type and Sector.................................................................54 Annexes I. Survey Results ................................................................................................................79 II. Table 1: IMF Structural Conditionality .........................................................................84 Table 2: World Bank Adjustment Lending Conditionality.............................................85 Table 3: An Assessment of the Adaptation of the IMF’s Internal Policy Process Based on Briefing Papers and Review Departments’ Comments Associated with Tanzania’s ESAF (1996–99) and PRGF (2000–03) .........................86 Table 4: Key Features of PRGF-Supported Programs with Respect to Tanzania’s PRGF, 2000–03 .........................................................................................89 III. List of Interviewees.........................................................................................................92 References................................................................................................................................96  4 Abbreviations and Acronyms ASDP Agricultural Sector Development Program BEST Business Environment Strengthening in Tanzania BWIs Bretton Woods’ Institutions CAS Country Assistance Strategy CCM Chama Cha Mapinduzi CG Consultative Group CSD Civil Service Department DAC Development Assistance Committee Danida Danish International Development Agency DAWASA Dar-es-Salaam Water and Sewerage Agency DFID Department for International Development EPZ Export promotion zone ESAF Enhanced Structural Adjustment Facility ESRF Economic and Social Research Foundation FAD Fiscal Affairs Department (IMF) FDI Foreign Direct Investment GDP Gross Domestic Product I-PRSP Interim Poverty Reduction Strategy Paper HBS Household Budget Survey HIPC Heavily Indebted Poor Countries HQ Headquarters HRDS Human Resource Development Survey HSDP Health Sector Development Program IDA International Development Association IEO Independent Evaluation Office IFMS Integrated Financial Management System ILFS Integrated Labor Force Survey IMF International Monetary Fund IMG Independent Monitoring Group JSA Joint Staff Assessment LOI Letter of Intent M&E Monitoring and Evaluation MDAs Ministries and development agencies MDF Multilateral Debt Fund MTEF Medium Term Expenditure Framework NGO Nongovernmental Organization NPES National Poverty Eradication Strategy ODI Overseas Development Institute OED Operations Evaluation Department PAF Performance Assessment Framework PADEP Participatory Agricultural Development and Empowerment Program PDR Policy Development and Review Department (IMF) PEDP Primary Education Development Program  5 PEM Public expenditure management PER Public Expenditure Review PFP Policy Framework Paper PHDR Poverty and Human Development Report PMMP Poverty Monitoring Master Plan PoMoS Poverty Monitoring System PPA Participatory Poverty Assessment PRBS Poverty Reduction Budget Support PREM Poverty Reduction and Economic Management network (World Bank) PRGF Poverty Reduction Growth Facility PRS Poverty Reduction Strategy PRSP Poverty Reduction Strategy Paper PSIA Poverty and Social Impact Assessment PWMI Poverty and Welfare Monitoring Indicators REPOA Research on Poverty Alleviation Sida Swedish Agency for International Development Cooperation SOE State-owned Enterprise SSA Sub-Saharan Africa SWAP Sector-wide approach TA Technical assistance TAS Tanzania Assistance Strategy TRA Tanzania Revenue Authority UNDP United Nations Development Program VAT Value-Added Tax  6 EXECUTIVE SUMMARY The PRSP process The initial formulation of the PRSP took place within an extremely compressed timetable, under pressure to reach the HIPC completion point. Civil society inputs in the preparation of the PRSP were limited and hence had little influence on the initial policy content of the PRSP. The Ministry of Finance officials indicated that at the beginning they were frustrated over the lack of clarity of what was expected from a PRSP. A country-led and comprehensive poverty reduction strategy, set in the context of the long-term Vision 2025, already existed in the form of the National Poverty Eradication Strategy (NPES). With respect to partnerships, the Helleiner report had initiated a fruitful dialogue with the country’s external partners and the Tanzania Assistance Strategy (TAS)—aimed at coordinating donor assistance—was under formulation. So what potential did the PRSP have when it was introduced? A review of the existing strategy documents and the evolution of the process, as evidenced from the preparatory phase and the subsequent three years of implementation, suggests that the PRSP process has indeed added value in some important areas, although challenges remain. A key contribution of the process has been the opening up of a dialogue between the government and civil society. Civil society is now more actively engaged in national policy- making through channels such as participation in the public expenditure review (PER). The PER has broadened to become the main avenue for public policy debate and strategy formulation in Tanzania. The institutionalization of the policy dialogue, through the PER process, where the World Bank played a critical supporting role, has been crucial—although it is not possible to attribute this development to the PRSP approach per se. The PER process provided a forum for follow up of the policy debate which was very important given initial gaps in the poverty reduction strategy. It is broadly participatory, government-led and encourages domestic and external partnership. It also facilitates a multi-sectoral results- orientation of the PRSP via links to the budget and assessments of the poverty focus of expenditure. One area where stakeholder involvement has been limited is on macro policy issues— involving mainly the government, donors, and the World Bank and the IMF. While this is partly due to the weak technical capacity among civil society groups, it also reflects limited success by the IMF in framing key macro policy issues in a manner digestible to a wider audience and in ensuring that all major macro issues were included in that debate. IMF staff were not clear as to what their role should be if some key policy issues were not emerging and were reluctant to intervene too actively in what was meant to be a government-led domestic debate. The PRSP has enhanced the mobilization of resources and focused them on priority sectors— with public expenditure reviews as major inputs. During the PRSP process the shares of priority sectors in expenditures have risen steadily, although PER analyses have shown that spending within priority sector needs to be better targeted to the poor. The PRS process has substantively enhanced national processes for poverty monitoring and a master plan now  7 guides all monitoring activities. But the feedback between monitoring of results and policy actions remains in need of improvements. In retrospect, the issue of poor domestic capacity has not been adequately considered given the broad agenda of reform—both within the process of implementing the PRSP and more broadly as part of ongoing government reforms—particularly management and coordination at the sub-national level. Lack of capacities has led to serious delays, and may ultimately lead to inadequate outcomes. The PRSP process has been partnership-oriented. External partners view the PRSP as the guiding framework for development assistance and have attempted to streamline their processes accordingly. The BWIs have also aligned their processes and their behavior in the field, along the lines of the PRSP approach, although in the view of domestic stakeholders more remains to be done. Two issues of concern on partnerships are the still low levels of involvement in the process of the business community and the continuing high transactions costs for the government of numerous donor missions, some of which are still not well coordinated. In the initial phase of its implementation, the PRSP has been mainly focused on improving service delivery, particularly in education and health, by increasing social sector expenditures. Several stakeholders also noted that the poverty diagnosis of the PRSP was not comprehensive enough, with a bias towards nonincome aspects. The income-oriented elements of a broader growth and poverty reduction strategy were delayed by the late completion of sector specific strategies, notably for agriculture and rural development. This also reflects the legacy of the initial focus on social expenditures of the HIPC process. The completion of sector strategies and action plans in later years and the more elaborate PRSP progress reports have helped to expand on and balance the implementation of the strategy. The ownership of the PRSP is strongest at the level of top leadership and the Ministry of Finance, followed by priority sector ministries, but is found to be much lower in nonpriority central ministries. Among civil society and local government, ownership has been mixed, being a function of their involvement in the process, with Dar-based groups more favorable to the process than those outside. Parliament has been largely outside the process, and its ownership is low. The main conclusions from Tanzania’s PRS experience are fourfold. First, it is the process, not the document itself that matters. Despite the compromised quality of the initial participatory process, the insufficient poverty diagnostics and lack of a broader growth strategy in the initial PRSP, three years after the launch of the PRSP policy-makers have been able to flesh out the strategy and improve poverty monitoring mechanisms. Second, while there has been commitment to the strategy at the top leadership level, there is by no means a full consensus on many key issues among all stakeholders involved in the exercise. The sustainability of the PRS process depends upon a broader level of participation by Parliament, nonpriority sector ministries, and civil society, including in the macroeconomic policy framework. Third, capacity constraints and lack of clarity about the role of the BWIs  8 if major policy issues are not emerging in the debate are currently hindering such participation and need to be addressed. It is also clear that wide differences of view about the scope and goals of participation remain. Fourth, the PRSP process requires closer World Bank and IMF collaboration, including a clearer delineation, as part of the PRSP framework, of what analytical inputs should be delivered. IMF role and the PRGF The consultation processes under the PRSP did, over time, have a significant impact on the content of macroeconomic policies, and the PRGF-supported framework did show considerable flexibility. In this sense, there was eventually an “alignment” with the key objectives of the PRSP although the process was not driven by preparation of the original PRSP. In assessing how the PRSP process affected the scope of policy choices available to the government and how the PRGF adapted to these policy choices, we examined in depth four policies in the macroeconomic area. They included two cases, the PRGF programming framework and cash budgeting, where a broader debate in the PER working groups led to a more flexible fiscal stance and increased priority expenditure better insulated from temporary budget shocks; one case, trade reform, where the political nature of the problem and the lack of a domestic debate on the lines of the PER evaded easy resolution; and a last case, fiscal federalism, which, though considered a priority by the government and with clear relevance for the PRSP, was approached very cautiously by IMF staff, in light of the fairly limited local capacities and the IMF’s own experience from other countries, and thus the alignment of the PRGF remained incomplete. Lack of capacities within government led to delays in policy formulation and implementation, although the process of interaction and debate finally led to the convergence of views on key macroeconomic issues. However, effective networking by key government officials, reinforced by donor concerns that the initial macroeconomic framework did not sufficiently accommodate increased aid flows, helped bring onboard technical expertise from local consultants, academics engaged in local research-oriented NGOs, and others from abroad. Their participation in the PER macro working group and workshops organized by the government, as well as technical contributions to policy discussions helped move the process forward. In this sense, the substantive revisions that were made in the original PRGF program design reflected an effective contribution of the institutional framework set up for consultations between stakeholders and the use of outside expertise to strengthen capacity. However, this participatory process was incomplete since it largely involved government, the BWIs, and bilateral donors, with civil society stakeholders largely on the sidelines. With regard to internal IMF policy formulation and review processes, the overall conclusion is that the IMF does seem to have evolved towards better discussion of the PRGF-PRSP linkages, as well as poverty issues more broadly, and did show flexibility in modifying the program design. However, while the program was able to accommodate higher aid inflows and did modify the targeted fiscal path, in light of the slower-than-expected-response of private sector investment and demand for credit, the assumptions on growth remained largely  9 exogenous to program design. Moreover, PSIA still did not play a significant role as a guide to policy formulation. World Bank effectiveness The Bank played a facilitative role, providing technical advice to the government during the short formulation period of the PRS as well as in identifying key process issues, which influenced the sectoral contents of the strategy and helped promote a more participatory process. Stakeholders were critical of the Bank’s “intrusive” influence in the initial stages of the formulation of the strategy, but acknowledged that this keenness was tempered by the country office. Overall, the principal drafters of the PRSP described the formulation as a government-led process with the BWIs not influencing the substance of the final strategy. Government officials, however, did complain about having to adhere to “formatting requirements” for the BWIs. Civil society members, on the other hand, were critical that during formulation the Bank did not exert more influence on the government to increase their participation. But they were cognizant of the dilemma faced by the Bank in urging the government to expand the consultative process while respecting the “country-led” principle. Further, the Bank as well as the civil society was reluctant to hold up Completion Point under HIPC to demand a more extensive consultative process. During implementation, the Bank has promoted the participation of civil society within the national PRS processes with encouraging results. Civil society participation has been a feature of the PER process and is being further strengthened year by year. An area where the Bank could have been better prepared for the PRSP was in assisting Tanzania to conduct a Household Budget Survey prior to 2000. During implementation of the PRS, however, Bank staff have provided substantial technical advice in analyzing the results of the Household Budget Survey and in establishing the Poverty Monitoring System. The Bank’s technical expertise and sectoral inputs in education and health and via the PER process have supported the 1st and 2nd Annual Progress Reports and have been greatly appreciated. The PER process in particular has been well recognized for its support of the PRSP and its principles, and the Bank has played a leading role in promoting and expanding the PER from an external technical assessment to a country-led participatory process involving a wide range of stakeholders. Besides providing secretariat services to the bi-weekly PER meetings until the end of 2003, Bank staff have fostered the momentum of the process and have continued to provide high-quality technical expertise. In other areas of the Bank’s analytical support, some key shortcomings came to light in the formulation of the PRSP, and some of these remain. There are continuing knowledge gaps between the micro-macro linkages, specifically the micro-economic social and poverty impacts of macro-economic and structural policies, analysis on the sources of growth and the poverty impact of alternative policy and strategy options. At the sectoral level, the depth and breadth of the Bank’s analytical work was uneven prior to the formulation of the PRSP.  10 Work in agriculture, for example, should have started much sooner and the Bank could have been more assertive in promoting the growth agenda in the PER/MTEF processes. The Bank’s lending portfolio in the last three years has been well aligned with the PRSP, and the recently introduced PRSC signals a further shift towards alignment with country processes. The PRSC1 is programmatic, aligned with the PRS policy matrix and has much less conditionality than previous policy-based lending instruments. The Bank’s overall behavior vis a vis respecting the “country-led” principle is widely recognized as having improved significantly. The Bank was already moving in this direction but the PRSP has helped moved it much further and faster. Interviewees described a “sea change” in the Bank’s behavior, and have cited the strong role of the country office in promoting this change. But there is still room for improvement, with some national stakeholders feeling that the “Washington consensus” has become a “Washington-Dar” consensus. Government officials and NGOs also feel that some sector teams are more resistant to the principle of country ownership and often fail to adequately consider alternative viewpoints. The Bank has continued to promote the “partnership” principle. The Bank’s proactive role in the participatory PER process, the alignment of the PRSC within the framework of the PRBS, and its engagement with the DAC, are important steps that have enhanced the collaboration of donors in Tanzania. There are still complaints about the Bank’s coordination at the sector level which relate to the large numbers of missions that regularly arrive from the Bank. At the same time stakeholders have acknowledged that the complaints at the sectoral level are strongly correlated with capacity problems within ministries, an area where the Bank’s and donors' technical assistance has not been successful in addressing. World Bank-IMF collaboration The PRSP process has fostered greater coordination amongst the World Bank and IMF teams working on Tanzania, although there remain important gaps—including the setting of priorities for timely delivery of PSIA. On conditionality, the BWIs have narrowed their focus to areas of their specialty, but this has not resulted in substantial reductions in aggregate conditionality. The Joint Staff Assessments of the PRSP and the Progress Reports have been important in providing a forward-looking perspective on the process but have been fairly uncritical, especially with respect to implementation capacity at the local level. As commentaries on a multi-stakeholder process, however, questions have been raised on their limited inclusiveness.  11 I. INTRODUCTION 1. This evaluation of the Poverty Reduction Strategy Paper (PRSP) and Poverty Reduction Growth Facility (PRGF) processes in Tanzania was jointly undertaken by the Operations Evaluation Department (OED) of the World Bank and the Independent Evaluation Office (IEO) of the IMF. The goal of the evaluation, as emphasized in the joint terms of reference, is to inform the policies and internal processes of the World Bank and the IMF, but not to evaluate the performance of the government.1 2. The evaluation is based on interviews and documents collected in the course of two visits to Tanzania in February and April, 2003, during which a number of stakeholders, including government officials, Members of Parliament, external development partners, business groups, academics, NGOs, as well as other civil society groups were met.2 The evaluation also draws upon in-depth interviews of staffs of the World Bank and the IMF, including current and former members of the Tanzania country teams in Washington and Dar es Salaam; published and internal documents from the two institutions; and recently published and nonpublished studies on PRSP-related processes in Tanzania.3 A survey covering 100 stakeholders, familiar with the PRSP process, was undertaken in May, 2003.4 3. The report is laid out as follows: Section II provides a background to the poverty situation in Tanzania as well as the economic and political context. The section also looks at the national strategies and processes for policy formulation that preceded the introduction of the PRSP and the PRGF. Section III discusses the Poverty Reduction Strategy Paper process in Tanzania, starting with a look at the relevance of the process to the key development challenges facing the country and then turns to the issue of how the underlying principles of the PRSP have been applied. It concludes with an analysis of the preliminary results. 1 The report covers Tanzania’s experience with the PRSP/PRGF process up to mid-2003. A draft of the report was transmitted for comments to the Government of Tanzania. Subsequent developments, including the PRS review process for the second PRSP and the Third Annual Progress Report, have not been discussed explicitly in the assessment. 2 The February 2003 visit comprised the scoping mission, whose outcomes also helped shape approaches in the other country cases. Members of the scoping mission included from IEO, David Goldsbrough (Deputy Director) and Steve Kayizzi-Mugerwa, and from OED, Bill Battaile (Task Manager) and Madhur Gautam. Members of the main evaluation mission, April 2003, were Madhur Gautam and Shonar Lala (OED) and Steve Kayizzi-Mugerwa (IEO). 3 A full list of interviewees is given in Annex II, Table 4. 4 See Annex I  12 4. Section IV evaluates the PRGF/PRSP process in Tanzania from the point of view of the IMF—specifically at how the PRGF-supported program adapted following the launch of the PRSP, including the changes that have taken place in the IMF’s internal policy formulation processes. The section concludes with a look at the problems encountered in aligning the PRGF-supported program with the PRSP, using an approach that focuses on how policies responded to four specific issues, and an evaluation of the extent to which the IMF- supported program evolved towards the key features of PRGF-supported programs. 5. Section V discusses the role and effectiveness of the World Bank in supporting the PRSP process in Tanzania. The analysis focuses on two areas: i) support during PRSP formulation, with particular attention paid to the way the Bank carried out its support with respect to the “country-driven” principle; and ii) alignment during implementation of the Bank’s financial and nonfinancial assistance to PRSP-defined priorities as well as its role in supporting the “country-driven” and “partnership” principles. 6. Section VI looks at World-Bank-IMF collaboration during the PRSP process, dwelling on the issues of coordination and priority setting, the joint staff assessment of the PRSPs, and aggregate Bank-Fund conditionality. Section VII provides the main conclusions and lessons of the evaluation. II. BACKGROUND AND CONTEXT A. Poverty 7. By any measure, poverty in Tanzania is deep and pervasive. The country ranked 160th on the 2001 Human Development Index, below the average for Sub-Saharan Africa. Per capita income was US$280 in 2002, which is well below Sub-Saharan Africa’s average of US$450. According to the Household Budget Survey for 2001/02, 12.6 million Tanzanians, 36 percent of the population, live under the basic needs poverty line, with about one in five living in abject poverty. The majority of the poor live in rural areas and are largely subsistence farmers. Inequality is high and regional and urban/rural income disparities are marked. 8. Over the past decade, Tanzania has not seen any significant downward trend in poverty. Only Dar es Salaam saw a substantial decrease in the population under the basic needs poverty line, from 28.1 percent in 1991/92 to 17.6 percent in 2001/02. The lack of national household survey data between these two points renders it difficult to assess whether poverty has been correlated with trends in real GDP per capita, which declined during the first half of the decade but has increased since 1995. But inequality has increased slightly over the last decade in both urban and rural Tanzania.  13 9. Nonincome or social indicators also portray a picture of stark poverty. 5 In 2002, life expectancy at birth was only 43 years, while a third of under-fives suffered from malnutrition. Infant mortality rates were over 100 for every 1000 live births. In recent years, this situation has been exacerbated by HIV/AIDS which is now the leading social concern, with serious demographic and economic implications. It is estimated that 10 percent of the population was infected by 2000. Stunting is widespread, with 44 percent of under-fives being moderately stunted in 1999. An important exception to the above numbers is the level of adult literacy. With 76 percent of the adults classified as literate, Tanzania’s population is more literate than many of its neighbors due to high primary school enrollment rates in the 1970s, which have not, however, been sustained.6 10. As in the case of income poverty, trends in social indicators in the past decade are disappointing. Net enrollment rates for primary education increased marginally in the 1990s, from 54.2 percent in 1990 to 58.8 percent in 2000, but were still lower than rates in the early 1980s. Since 2000, rates have increased substantially to 65.5 percent as a result of the cancellation of user fees but the quality of education has declined markedly during the period. In 2001, only one-third of boys passed the Primary School Leaving Examination and a fifth of girls. The 1990s showed little progress in improving health markers with stagnation in infant and under-five mortality rates and a steady increase in the prevalence of HIV. B. Economic and Political Context 11. The hopes of rapid economic growth and poverty reduction that have existed since independence in 1961 have yet to be realized. While external shocks and increasing foreign debt have been very disruptive in the past three decades, the fragility of the economy is also reflective of poor initial conditions at independence, including lack of trained manpower and paucity of investments in social infrastructure and production capacities. Analyses have also cited lack of flexibility in the policies pursued by the government since the late 1960s as a major factor (see Bigsten et. al. 2001). 12. Beginning in 1967, the government adopted the Arusha Declaration as a blueprint for the country’s development strategy. It involved the nationalization of most of the economy, as well as the resettlement of villagers in larger communities in the countryside in order to improve their proximity to services. This policy stance, based on Ujamaa (or African socialism) was supported by many donor countries as well as multilateral institutions, 5 The last Participatory Poverty Assessment (PPA), carried out in 1997, identified landlessness, lack of agricultural inputs, inability to feed oneself, lack of access to health and education services, dependency and powerlessness, and physical disability as defining characteristics of the poor (Narayan 1997). 6 Average adult literacy rates for Sub-Saharan Africa were 62.4 percent in 2001 (Government of Tanzania, 2003a).  14 resulting in considerable inflows of aid. However, the Ujamaa experiment failed to deliver the promised rapid social advancement—for example following the onset of the villagization project in 1973 agricultural output began to decline. Although the goal of Ujamaa was a self- reliant economy, Tanzania was to become more dependent on donors than before (Collier, 1991). 13. The late 1970s and early 1980s were marked by a deepening economic and social crisis. The war with Uganda at the end of the 1970s and the second oil shock led to a severe resource shortage. The economic policies pursued in the previous decade were questioned, with policymakers divided between those supporting and opposing market reforms. Importantly Julius Nyerere, the leader of the country from independence, retired in 1985 but continued to be the leader of the ruling party (CCM). 14. The collapse of the economic system in the 1980s led to a decline in growth, weakening of the state and to the disengagement of the public at large from any meaningful political participation. From the mid-1980s Tanzanian leadership came under increasing pressure to undertake democratic reforms, both from within the country and from its external development partners. Change—combining economic reforms with political liberalization and decentralization—has not been easy. The government has had to address a number of competing demands under serious financial and human resource constraints. Moreover, the control regime in Tanzania had generated a number of vested interests and its removal was bound to be resisted by politically powerful groups. Thus, earlier economic reforms were inconsistently implemented and were subject to reversal (see Wangwe, 1997). 15. A confluence of factors—poor economic performance, inappropriate policies and deteriorating governance—led to a breakdown in the relations between Tanzania and its external partners in the mid-1990s. In mid-1994, a group of “Independent Advisors on Development Cooperation Issues Between Tanzania and its Aid Donors” was formed to assess whether development cooperation could be strengthened. The report of the group (known as the Helleiner report, after its chairman) contributed to a major revamping of relations with donors. With rapprochement in the wake of the Helleiner report and with a new government at the helm, Tanzania has undertaken far-reaching economic reforms since the mid-1990s. Multiparty democracy has also been strengthened with the first multiparty election held in 1995. The ruling Chama Cha Mapinduzi (CCM) remains the dominant political force in government and in other important areas of political life in Tanzania. Although political liberalization has been difficult to implement in its entirety, the evolution to multi-partism has been spared the serious political disturbances that have been witnessed in neighboring countries. Indeed, in combining economic and political reforms with peaceful conditions at home, Tanzania has distinguished itself in a region otherwise dominated by serious domestic disturbances. 16. Tanzania finally embarked on sustained economic reforms in 1996 because the government was willing to implement reforms, even politically costly ones, and the donors insisted on greater accountability. First, the elections of October 1995 were won by Benjamin Mkapa, with his victory seen as an endorsement for continued economic reform. Second, the  15 donor community indicated more emphatically than before that in the absence of tangible results in implementation and accountability it would not continue funding the government’s programs. 17. Table 1 shows that compared to the previous period, economic performance improved during 1996–99, with increased growth and lower inflation. This trend was strengthened during 2000–02. 18. The most important budgetary challenge has been the government’s poor capacity to generate revenue. In spite of the creation of the Tanzania Revenue Authority (TRA) in 1996 and the introduction of VAT in 1997, government revenue has not risen much above 13 percent of GDP (compared to an average of close to 20 percent for Sub-Saharan Africa). The main causes of poor revenue generation are tax evasion, discretionary and statutory tax exemptions, a very limited tax base, and weak tax administration. Towards the end of the 1990s, the government adopted a more holistic approach to tax reform with the removal of tax disincentives and the streamlining of the legislation and licensing procedures for domestic and international trade. 19. Trade expanded by less than expected during the 1990s, despite trade liberalization and the privatization drive aimed at encouraging foreign investment and inflows of capital goods. This was partly to blame on the negative impact of commodity price shocks. Increased aid inflows, following HIPC, helped to reduce the current account deficit. The HIPC initiative also led to a decline in the large external public debt, from 130 percent of GDP in the early 1990s to 75 percent of GDP in 2000 (Table 1). Table 1. Growth and Investment Indicators (Period Averages), 1991–2002 (In percent of GDP; unless otherwise indicated) Economic Indicator 1991–95 1996–99 2000–02 1991–2002 Gross fixed capital formation 24.4 15.8 16.6 19.6 Real GDP growth (in percent) 1.8 3.8 5.5 3.4 Real GDP per capita, in 1990 U.S. dollars 181.7 183.9 196.2 186.1 Real GDP per capita growth (in percent) -1.0 1.1 2.7 0.6 Inflation, average (in percent) 27.4 14.0 5.3 17.4 Government balance including grants -2.8 -1.1 -2.0 -2.0 Grants 2.9 3.8 5.0 3.7 Government revenue excluding grants 12.4 12.5 12.8 12.5 External current account balance (including grants) -8.7 -9.8 -2.7 -7.6 Real exports of goods and services growth (in percent) 4.4 1.8 12.5 5.5 External public debt 106.5 131.4 98.5 75.7 Sources: The Tanzanian Authorities; and IMF database (WETA)  16 20. Agricultural output has grown at 3.7 percent per annum since 1990 although characterized by swings due to both weather and price shocks. Performance in the industry sector has been closely patterned on the speed of economic reform. In the early 1990s, performance was uneven and sometimes negative, but picked up strongly after 1995. Mining has accounted for a large part of the recent expansion as the investment undertaken by multinationals began to generate higher output. In contrast, manufacturing output has been below expectations. The business community has blamed this on poor incentives, including high taxes, inadequate tariff protection and smuggling. The service sector has expanded rapidly since the late 1990s, a reflection of the higher priority that education and health have received in government spending in recent years and the expansion into these areas by the private sector. 21. In the new millennium, however, the question of sustainability remains. In a retrospective look at the post-independence economic performance, the government’s Vision 2025 statement from 1998 lists four factors which “have impeded the realization of the goals of the development plans and programs” in Tanzania. • A donor dependency syndrome and a dependent developmental mindset. • A weak capacity for economic development. • Failures in good governance and in the organization of production. • Ineffective implementation syndrome. 22. As the current regime reaches the end of its political mandate (presidential elections will be held in 2005) these challenges have come to the fore once again. How should the government ensure that the current reform process is maintained? The advice given to the government on this score has sometimes been conflicting. Some domestic activists have argued, for example, that tying reform processes more closely to the political process by including parliamentarians, leaders of local governments, and even the opposition, would ensure the survival of the process. Others have suggested, however, that robust lock-in measures, notably legislation and the opening up of the economy to market competition, are the only viable avenues to ensuring the longevity of the reform process. C. National Policy and Strategy Formulation 23. In the second half of the 1990s, the government introduced a number of national strategies for development and poverty reduction as well as changes in its budget and policy formulation processes: a) Development Vision 2025. b) National Poverty Eradication Strategy (NPES). c) Tanzanian Assistance Strategy (TAS).  17 d) The Public Expenditure Review/Medium Term Expenditure Framework (PER/MTEF). 7 24. The Vision 2025 document was an attempt at updating Tanzania’s national vision, first expressed in the Arusha Declaration. It projects a future free of poverty and with high levels of human development, with Tanzania graduating to a middle income country by 2025. Importantly, it discusses the imperatives of integrating Tanzania into the new global economy—competitive, diversified, and economically stable (growing at 8 percent and above) and characterized by good governance and the rule of law. However, beyond these broad development goals, the document does not provide a strategy of how to achieve them nor does it discuss the policies required in any detail. 25. The National Poverty Eradication Strategy of 1997 was more specific in its goals and strategies. Within the broader Vision 2025, policymakers saw NPES as the means to coordinate policies and programs in the medium to long-run (up to 2010) (see Box 1). The NPES defined goals in the areas of economic growth, income levels, primary education, literacy, access to water and sanitation, disease burden, mortality and health, employment, housing and infrastructure—some with indicators and targets specified, and others without.8 Although the NPES was formulated under a participatory process, it failed to specify priorities among its many planned interventions, making it inappropriate as an action plan. Crucially, the macroeconomic framework, costing strategies for the program, and its targets were not explicit. Box 1. Tanzania: Priority Areas of the National Poverty Eradication Strategy (1997) A. Policies and strategies for creating and enabling environment for poverty eradication: governance, coordination, participation. B. Policies and strategies for creating capacity for poverty eradication: economic growth, agricultural, industrial sector and mining sector development, promotion of cooperatives, infrastructure, resource mobilization. C. Sectoral policies and strategies for poverty eradication: education, health, water supply, employment, environment and settlement. 7 In addition, CG meetings were moved to Dar es Salaam to reinforce domestic-policy making processes. An Anti-Corruption Strategy which emphasized increased transparency, accountability and governance, and an External Debt Strategy aimed at preventing the country from reverting to serious external debt problems post-HIPC, were also introduced. 8 For example, one of the specific goals was to increase the average income to between US$ 300-500 per annum by 2010, while one without a specified target was to reduce the rate of unemployment among women and youth.  18 26. The Tanzania Assistance Strategy was the government’s attempt to “restore local ownership and leadership by promoting partnership in the design and execution of development programs” after the deterioration of relations with donors in 1995. However, the multi-stakeholder Working Group drafting the TAS was interrupted by the work required for reaching completion point under the HIPC initiative—that is preparation of the PRSP. The interruption of the TAS process implied that, when completed in 2002, the document’s focus had narrowed to achieving best practices in development cooperation—promoting local ownership and leadership, improving predictability of aid and harmonizing domestic policies and procedures. Monitoring aid effectiveness in terms of progress made in poverty reduction was to be undertaken under a PRSP-led process. 27. The public expenditure review (PER) process had become the main avenue for policy debate and strategy formulation among domestic stakeholders and with donors by 1998— before the PRSP process was launched. At the apex is a PER working group, which is responsible for guiding all aspects of the PER process, including approving the annual work program. A PER macro sub-group debates aspects of the macroeconomic framework. There are also sector working groups focusing on areas such as health and education. The PER groups include government officials, civil society, external partners and the private sector. The PER process is credited with raising a number of important issues, including the need to adjust the application of the cash budget to reduce its negative impact on priority expenditure. PER working group bi-weekly meetings, drawing on the wide network that the PER process has created in the last few years, have provided key inputs into the government’s policy analysis. 28. The PER process revolves around the government’s budget cycle, and thus feeds directly into it. The PER’s work program is prepared in July, at the beginning of each financial year. In May of the following year (a month before budget presentation), the PER working group organizes a PER consultative meeting, attended by representatives from all stakeholder groups, including government, the donor community and civil society organizations. Three hundred people attended the PER consultative meeting in May 2002. 29. The PRSP process picked up both positive and negative features of earlier strategy- making exercises. The government's experience with the formulation and implementation of national strategies was useful for the PRSP process, including a quicker process of internalization and ability to map out participatory strategies. It is quite likely that some of the eventual shortcomings of the PRSP, including lack of concrete operational guidelines and costing of interventions, were spillovers from the earlier strategies. The proliferation of poverty eradication strategies also made it harder for officials and other stakeholders to view the PRSP as different from previous attempts. Indeed there was a real risk of seeing the  19 PRSP as only one of many strategies rather than the strategy to encapsulate all poverty eradication efforts.9 III. THE POVERTY REDUCTION STRATEGY PAPER (PRSP) PROCESS A. Relevance 30. Any assessment of the relevance of the PRSP process in Tanzania should bear in mind two factors : first, is the issue of the potential value added of the PRSP, given that a number of national strategies were already in place or in the process of being developed when the “requirement” of developing a PRSP was “imposed” by the BWIs, as perceived by many stakeholders in Tanzania. The second factor refers to the distinction between the actual PRSP document and the strategy as subsequently elaborated—important since the document was not intended to be an end in itself but a “live” document with periodic updates. 31. In the view of many stakeholders, the PRS is fully consistent with the existing approaches and strategies of the government for poverty reduction but, in principle, adds value in some important respects, building on earlier efforts. Survey respondents ranked the relevance of the PRSP approach highly.10 As noted earlier, the Vision 2025 and the NPES were too broad in scope and lacked sufficient prioritization of actions to make them fully operational. Some specific areas in which the PRSP process was particularly relevant are discussed below; how well the process has actually delivered on these potential value additions is addressed in subsequent sections. 32. The results-oriented principle of the PRS was intended to help government mobilize and focus resources on the priority sectors. While in the 1990s, the Rolling Plan and Forward Budget’s priority sectors were broadly defined: social services, infrastructure, law and order and the environment, by the first year of the MTEF for FY00–02, the priority focus had been narrowed (to roads, water, health, education, agriculture and the judiciary) and priority items within these priority sectors were defined for the first time. So a trend towards prioritization and protection of expenditures did take place particularly after the introduction of the MTEF. 9 A paper commissioned by Danida (2001) to look at Tanzania’s experience with the PRSP claims that since most of the interventions presented in the Tanzanian PRSP were already contained in the NPES, Vision 2025 and TAS, the PRSP added little. 10 Overall, relevance received the highest mean score among all the categories in the survey. The questions—“The PRSP process has added value to the country’s response to the challenges of poverty reduction, compared to previous strategies/policy frameworks” and “The PRSP framework improves on past modalities for donor assistance”—were among the five most positive responses in the survey, with mean scores of 3.4 on a scale of 1 (Completely Disagree) to 5 (Completely Agree).  20 But there was still much progress to be made in the strategic prioritization of expenditures in Tanzania, and the PRS process provided a framework for achieving this.11 33. The PRS has also helped to improve the monitoring of poverty reduction targets and bring a results-orientated approach to the poverty reduction efforts. Prior national poverty strategies and monitoring initiatives had been characterized by lack of baseline data and sporadic use of targets. Indeed, the Country Economic Memorandum for 2000 identified the “lack of analytical capacity and good data at the country level” as among the key challenges in the design and execution of poverty reduction programs in the country.12 The NPES, for example, refers to 21 goals, of which only 9 had specific targets. The 1999 Poverty and Welfare Monitoring Indicators (PWMI) compiled a list of 75 indicators in 14 sectors, but here too few targets were included, and recent baseline data were not available. Moreover, a coordinated consistent approach to data collection and poverty monitoring along with a feedback mechanism into policy formulation and expenditure allocation, critical for fine- tuning and modifying a “living” poverty reduction strategy document to the emerging needs and trends, was missing. 34. The PRS enabled the participation of a broader swathe of civil society in national policy-making than had been the case thus far. Although NGOs proliferated in the 1990s, their policy dialogue with the government was limited to political or governance issues.13 Vision 2025, NPES and TAS had been prepared in a consultative manner, but the emphasis of participation was on hearing the views of line ministries and government officials and the participation of NGOs and other civil society organizations was minimal.14 Encouraging “a more meaningful participation in the development process”15 by all citizens was central to efforts to enhance accountability and governance and was consistent with contemporary research on the benefits of broadening decision-making power via greater inclusion.16 11 See Government of Tanzania and World Bank (2001), p. 108. 12 Ibid. Chapter 4 13 Mogella (1999) 14 From interviews undertaken for the evaluation. Government officials indicated, however, that CSOs had, at this point in time, serious capacity constraints which prevented them from participating effectively. 15 See Government of Tanzania and World Bank (2001), p. 43. 16 See Government of Tanzania and the World Bank (2001). See also Isham, Narayan and Pritchett (1995), and Narayan and Pritchett (1997).  21 35. In other areas as well, the relevance of the PRSP is high in terms of consistency with the country’s own on-going processes and agenda. In Tanzania, a country-led process was already in practice, with the government taking leadership in key strategic areas following the reforms initiated in 1995. While the ownership of development projects and programs prior to the PRSP was characterized as “unsatisfactory,”17 the formulation of a wholly-owned and country-led strategy, the TAS, to look at issues of donor-government cooperation was underway at the time of the PRSP’s introduction—with working groups set up and background papers commissioned.18 The Policy Framework Paper (PFP) had also evolved into an increasingly collaborative process with the last PFP of 1998–99 being considerably more country-led than the previous one in 1994.19 36. Similarly, on the principle of enhancing partnership, by the time that the PRSP was initiated, Tanzania’s relationship with external partners had improved significantly following the crisis in 1994–95 (Helleiner, 1999). The TAS, which was prepared in consultation with the donor community, was meant to provide a common framework for development assistance.20 The consultative and expanding PER process that was in place since 1998 was already opening the lines of communication between the government and its external partners, as indeed the decision to relocate the annual CG meetings to Dar es Salaam in 1997. At the sector level, sector-wide approaches (SWAPs) were nascent but did exist in the health sector. A system of program aid via the Multilateral Debt Relief Fund (MDF) already existed in 1998 (albeit for servicing external debt obligations to the multilateral institutions). Thus, while the PRS principle of partnership was fully consistent with these ongoing efforts, the changes were already under implementation and the PRSP helped reinforce these processes. 37. Finally, Tanzania had already adopted a comprehensive and long-term strategic approach to poverty reduction. The NPES was prepared using inputs from a rural participatory poverty assessment (PPA), and it encompassed a broad definition of poverty and priority areas, including many income and nonincome measures. Governance, economic 17 Helleiner (1995) 18 TAS work on priority areas was retained for the preparation of the PRSP. 19 OED (2000), Helleiner (1999) 20 With the goal of reviewing the original Helleiner recommendations and to provide an updated analysis of the partnership process, an Independent Monitoring Group (IMG), including Tanzanians and others, was set up in 2002 to monitor the evolution of partnership and the extent to which the TAS principles were being adhered to. The work of the group provides a good example of the institutionalization of the review processes in Tanzania (see Alonso and Utz, 2003). The monitoring exercise was to be repeated on the basis of 2-3 years. Among the recommendations of the Independent Monitoring Group (2002)was the need for the government to adopt reduced aid dependence as a policy objective.  22 growth and poverty eradication feature as prominent priority areas. The long term Vision 2025 provides the context for the NPES and lists achievements to be attained by 2025.21 Thus while not all targets in the NPES and Vision 2025 were quantified and neither explicitly acknowledges the limited sustainability of previous programs, a comprehensive and long- term outlook and strategy were clearly evolving. B. Preparation 38. The formulation of the PRSP took place within an extremely compressed timetable, under pressure to reach HIPC completion point (see Table 2). The government began the preparation of the PRSP in March 2000, once the Interim Poverty Reduction Strategy Paper (I-PRSP) was completed. The I-PRSP made no pretence to being participatory, its main purpose being to provide a road map to the HIPC completion point, including steps to be taken in the preparation of the full PRSP. 39. For the consultations for the PRSP, the government organized zonal workshops during May–June 2000 in all regions of Tanzania, with the goal of soliciting views from the whole country. The PRSP document was prepared by a technical committee of Tanzanians experienced in working on development issues (Government of Tanzania, 2000a). Members of Parliament reviewed the document in July 2000, and one month later, a draft of the PRSP was presented for discussion at a national workshop, attended by representatives from all stakeholder groups (including the World Bank and the IMF). The draft was subsequently cleared by the cabinet in September. Thus, six months after its initiation, the PRSP was ready.22 It highlighted four broad sets of measures: sound macroeconomic policies; expenditure prioritization to ensure that support to key programs and social services was maintained, in spite of budgetary constraints;23 enabling the poor to share increasingly in the benefits of globalization, by promoting export-oriented policies in agriculture and industry; and raising investment. 21 Not all these achievements have specific quantifiable targets assigned to them, however. One goal is a “reduction in infant and maternal mortality rates by three-quarters of current levels.” Another less specific goal indicates that “Tanzanian society should show strong adherence to and respect for the law.” 22 As of September 2003, the average time between completion of the I-PRSP and PRSP for all countries was 20 months (World Bank and IMF, 2003). 23 In other words since revenue was only expected to rise by only 0.7 percent of GDP, no strong revenue effort was anticipated.  23 Table 2. Chronology of the PRSP/PRGF Process Event Date Tanzania declared eligible for debt relief under enhanced HIPC initiative November 1999 Interim-PRSP endorsed/HIPC decision point reached March 2000 PRGF-supported program 2000-2003 launched March 2000 Draft PRSP debated in zonal workshops in Tanzania May-June 2000 Draft PRSP debated by Parliament July 2000 Draft PRSP debated at National Workshop August 2000 Draft PRSP reviewed and approved by Tanzanian Cabinet September 2000 PRSP approved by IMF/World Bank Boards December 2000 First progress report of the PRSP September 2001 HIPC completion point reached November 2001 Second progress report of the PRSP April 2003 Sources: IMF and World Bank Reports cited in the references. 40. The joint staff assessment (JSA) of the PRSP, completed in November 2000, found the document weak in several areas—including poor macro-micro linkage and inadequate costing of the proposed interventions. The JSA also noted that the PRSP did not discuss adequately the fiscal, monetary and structural policies required to ensure sustained levels of growth, nor did it justify how the sources of growth were identified. The issue of unambitious revenue targets was also noted. Nevertheless, the mood in government and among donors was in favor of bringing the process to closure. Many Bank and IMF country staff interviewed as part of this evaluation expressed the view that Tanzania’s good macroeconomic performance, rather than the speed at which the PRSP was being drafted and any “gaps” in the document, should have been the basis for reaching the HIPC completion point. They argued that focusing on the document rather than the process belied the claim that the ultimate goal was domestic ownership and a country-driven process.24 41. Ministry of Finance officials indicated to the joint evaluation team that they had received imprecise guidance from the Bank and the IMF at the beginning of the PRSP process. They noted that after responding to the concerns of one institution, similar concerns would be raised by the other. Given their limited capacities, this complicated the preparation of the PRSP. The authorities’ complaints raise the issue of insufficient clarity on the tradeoff between ensuring a country-driven process and providing an operational policy framework with guidance on the advice that staffs were expected to give during the PRSP process. 24 Donors had sent a “joint bilateral note” in support of Tanzania to the Boards of the IMF and World Bank in Washington in early 2000.  24 Although the joint staff assessment argued that the PRSP could have taken a broader perspective on poverty alleviation issues, the Executive Boards of the Bank and Fund found that the document provided an adequate basis for Fund and Bank support. In making this decision, the BWIs took into account the progress the country had achieved in creating an environment for sustainable growth, and the government’s commitment to the implementation of a comprehensive set of pro-poor policies which they hoped would be further elaborated in the subsequent PRSP progress report.25 C. Application of the Underlying Principles and Assessment of Change Country-driven with broad-based participation Ownership of the PRSP process 42. The stakeholders interviewed for this evaluation were divided in their views on the issue of ownership of the PRSP. One view held that conceptually, since the PRSP was an externally imposed condition, a requirement for qualifying for HIPC debt relief, there could not be genuine ownership of the document or the strategy. The other view is that the process itself has been beneficial—it has added value to official thinking and has helped move government policy towards more effective poverty reduction. These contrasting views are reflected in the survey findings, in which responses to questions on whether the process was country-driven were among the most polarized in the survey.26 Going beyond these arguments, perhaps a truer measure of ownership is the commitment to the strategy and its implementation. 43. A second, more universal, criticism of the process in Tanzania is that the need for approval by the Boards of the BWIs contradicts one of the key underlying principles of the PRSP process. 27 This remains a thorny issue, as there is an inherent tension between the goal 25 In the event, the draft of the first PRSP progress report was also found to be weak, having failed to respond to many queries raised by the JSA on the full PRSP, including presentation of a poverty profile on the basis of a household budget survey, better costing of PRSP programs, and linkages of the macroeconomic policies to poverty reduction. The IMF’s main review department (PDR) had even wanted to have the decision on HIPC completion point delayed until a better report was prepared, although area department staff argued that viewed as a process, and taking account of constraints, the PRSP approach was proceeding well. The progress report (Government of Tanzania 2001a) was eventually published in November 2001, enabling the country to reach HIPC completion point. 26 See Annex I for more information. 27 The Boards do not “endorse” the PRSPs in and of themselves. Rather, the Joint Staff Assessment accompanying the PRSP “recommend endorsement (or rejection) of the strategy as a basis for Bank and Fund concessional lending to the country concerned” (World (continued)  25 of aligning aid programs to the PRSP and the fiduciary responsibility of external partners to their own taxpayers and what they perceive to be high priority issues. The tension thus boils down to differences in priority setting between external and domestic stakeholders, differences in the quality of analysis and strategy, and how to set acceptable standards of performance, with implications on domestic managerial and technical capacity. 44. While at the level of top leadership there is complete ownership of the process, that of the Parliament is more circumscribed. For instance, MPs feel that their current role in approving budgets is more of rubber stamping, which is ironic given that they are ultimately accountable to their constituencies for the performance of the government. The central administration, however, feels that the function of the Parliament is to provide oversight and to hold government accountable—there is thus a potential conflict of interest in involving MPs in the technical aspects of the process, which could compromise their oversight function. 45. Within government, the ownership of the PRS is strongest in the Ministry of Finance, followed by priority sector ministries, and is much lower in nonpriority sector ministries and local government. Another source of friction is the current focus on priority areas and shifting of the budget to specific activities at the cost of others. Not being a priority sector has naturally generated a decline in interest in and commitment to the PRSP process.28 However, even among the priority sectors, the buy-in into the principles of the strategy, primarily shifting from an input to a results-oriented focus, is limited. Several ministries and agencies remain more committed to their sector programs, and the associated donor funding, than to the PER/MTEF, nor are they willing to subject themselves to the budget discipline that the overall PRSP process seeks to enforce. Here the role of donors in financing off- budget projects is important—many have yet to provide details on the full extent of their commitment. 46. Ownership among the civil society groups is a function of their involvement in the process. Many of the major Dar-based NGOs are more positive about the PRSP and are more actively involved than some of the more remote rural based NGOs (although most are members of networks with representatives in Dar es Salaam). This does not imply, however, that the CSOs endorse all of the elements and the processes associated with the PRSP and its implementation—indeed a number are highly critical of parts of the strategy and aspects of its implementation. More important, most of the NGOs are interested in discussing options and issues in a constructive manner and the government has, overtime, become more open to this dialogue as discussed below. Bank/IMF 1999a&b). Domestic stakeholders were not aware of this distinction or regarded it as unimportant, not affecting the central role of the BWIs in the process. 28 See also KK Consulting Associates (2001).  26 Broad-based participation Participation in the poverty reduction strategy debate 47. There have been two distinct phases to the consultation process: the PRSP preparatory phase and the implementation and monitoring phase. As noted, the timeline for the preparation of the PRSP was very condensed because of the rush to get to the HIPC completion point and precluded adequate consultations with both national and international Dar-based NGOs and grass-roots organizations, a shortcoming which has been widely criticized. The World Bank unsuccessfully tried to convince the government to carry out a more inclusive participatory process which would involve civil society. But civil society inputs had limited influence on the policy content of the PRSP.29 Subsequently, however, a dialogue has opened up between NGOs and the government. Civil society is now much more actively engaged in national policy-making through channels such as working groups for the public expenditure review, poverty monitoring and the Consultative Group meetings. 48. This evolution has been part of a learning curve for both the government and the NGOs. Prior to the PRSP, the involvement of NGOs in national policy making was very limited. NGOs lacked capacity to effectively participate in policy debates. In some instances, this lack of capacity still limits constructive engagement in policy discussions, especially macroeconomic policy. McGee et. al. note that “Civil society itself recognizes that it is better prepared now than at the outset for such processes, and is occupying the spaces that the PRS has opened to demand of government the institutionalization of pluralist, participatory and deliberative policy processes.”30 In light of the grassroots linkages and the comparative advantage of NGOs in reaching the poor and in service delivery, for example some NGOs are undertaking their own monitoring activities, the government has realized that it stands to benefit from the feedback and transparency that an open participatory process brings. 49. While the process is thus widely acknowledged to have become much more participatory and inclusive, there remain several areas for improvement. The NGOs are in general highly critical of the manner in which the initial consultative process was carried out especially the selection of NGOs for invitation to the zonal and other workshops as this had implications for setting the agenda.31 With better organization of NGOs and other CSO umbrella groups, there is now more effective participation in national workshops. But the 29 See McGee with Levene and Hughes (2002). 30 Ibid. 31 For example, representatives of the NGO Policy Forum explained that had NGOs participated more actively in the PRSP formulation process, they would have put governance issues on top of the agenda. In their view, lots of money was being thrown at poverty but without ensuring that it reached the poor. Trade unions added that they would have emphasized sustainable livelihoods and employment.  27 composition of the various PRSP working groups remains nontransparent and the criteria used to determine participation is viewed with suspicion. For example, for the poverty monitoring working groups, NGOs are mostly confined to the Dissemination working group, which, though consistent with the advocacy nature of many of the NGOs, implies little NGO presence in the more influential working groups.32 There are also complaints from some of the underprivileged and vulnerable groups (such as the disabled persons) about not having adequate access to the national processes in order to have their voice heard. 50. During the PRSP process, many academics have, in individual capacities and as members of research-oriented NGOs played an important role in policy formulation.33 These institutions have been key participants in the PER process from inception and have also held workshops and training seminars on their premises for the benefit of civil society. They thus play an important intermediating role, while also helping to enhance analytical skills among civil society groups. 51. As noted, the role of the Parliament was limited. This is partly a result of weak institutional capacity and lack of a more proactive involvement at the grassroots level, notably during the zonal workshops, but is also because the government has done little to more actively engage parliamentarians. The PRSP document was presented to the Parliament during a brief seminar to obtain feedback and political endorsement. Moreover, MPs enter very late in the planning and budgeting process, when it is too late to make any major changes to priorities or budget allocations. A Tanzanian academic has argued, in a background paper commissioned for this case study, that since Parliament, the ruling party, CCM, other political parties and CSO groups are still relatively weak in comparison to the government, the process has been monopolized by the central bureaucracies, leading to their expansion. He is concerned about the implications of this recentralization for the sustainability of the process (Mukandala, 2003). Participation in the formulation of the macroeconomic framework 52. The Ministry of Finance and Bank of Tanzania were most involved in the formulation of macroeconomic policies, while knowledge and participation elsewhere in government were very rudimentary. Indeed the Independent Monitoring Group (2002, p.18), found that a number of line ministries were entirely “out of the loop”, especially with respect to their involvement in macro discussions. More than two-fifths of the survey respondents disagreed 32 There is one NGO representative in the Research and Analysis working group, and none in the Surveys and Census and Routine Data working groups. 33 The most notable of such groups are the Economic and Social Research Foundation (ESRF) and Research on Policy Alleviation (REPOA). The latter also acts as the secretariat for the poverty monitoring working group for the PRSP. See also Government of Tanzania (2001c).  28 that the formulation of the macroeconomic framework of the PRSP had been sufficiently participatory.34 Some respondents blamed their limited involvement on lack of incentives, since keeping abreast of the changes implied little change in their professional situation. A number of nonpriority sector ministry officials argued that since their activities were not on the priority list for expenditure allocation (in contrast with health and education, for example) their views were not taken seriously by government.35 53. Many NGOs and other civil society groups, such as trade unions, business groups and employers’ associations argued that while macroeconomic policies were now more frequently and openly discussed by the government and in broadly inclusive forums, there was nevertheless an important difference between attending meetings with government and donors, which they now did routinely, and being able to have an impact on policy (see for example Evans, 2001). Further, they said that although macroeconomic policies affected their members in a variety of ways, the rationale for and importance attached to the policy choices of the government were poorly explained in the gatherings. Some respondents argued that while NGOs were looked upon as allies by the government during its attempts to reach HIPC completion point, their demands for direct involvement in all major policy processes thereafter, including macroeconomic policies, irritated the government. Moreover, lack of resources constrained NGOs and other groups from effective participation. A number of NGOs noted in interviews that although donor funding was available during the PRSP formulation process to facilitate their participation, funds became quite scarce after its completion. 54. Although a good number of civil society groups had heard of the PER working group and some were members of the sector PERs or of those for cross-cutting issues, many NGO representatives felt that membership of the key macro sub-group was too restricted. Ministry of Finance officials as well as the World Bank country team indicated that membership of the working groups was not restricted but that frequent attendance was a requirement if the process was to continue to be useful. However, one group cited the exclusion of an individual they had nominated who was making statements in newspapers about corruption and transparency. The few from civil society that attended the meetings of the macro sub-group 34 44 percent of respondents disagreed or completely disagreed with the statement “The formulation of the macroeconomic framework of the PRSP has been sufficiently participatory.” The mean for this question had one of the five lowest standard deviations (indicating high level of consensus) among the questions in the survey. See Annex I for details. 35 For example, the Ministry of Lands was of the view that since it was responsible for mapping land and physical structures it was crucial to activities in agriculture, tourism and mining and that it was necessary to be put on the priority list for expenditures, and to have access to donor funding (which had dried up when it became a nonpriority sector). The Ministry of Lands has since become a priority sector.  29 thought that discussions were often difficult to follow, partly for lack of adequate training in economics. Thus capacity constraints were also important in limiting CSO participation in the macro policy debate. 55. A number of IMF staff interviewed were of the view that if the macroeconomic framework and related policies had emanated from a more robust domestic debate, some of the subsequent disagreements on policy might have been avoided. However, the macroeconomic debate had to be initiated by the government, either through Parliament or via the participatory processes it was encouraging. Although there was certainly a debate over expenditure priorities, some of the other issues, including the overall macroeconomic framework and the tradeoffs it involved, did not surface during the initial participatory process. It was also hoped that trade issues, on which there had been some disagreements with the IMF, could have been addressed more comprehensively as part of a broad-based PRSP process and public debate. The most notable disagreement was over the pace of tariff reform, which exposed domestic industries to foreign competition. Interviews also suggested that there was some ambiguity about the role that IMF staff were expected to play, especially if some major policy issues were not emerging in the broader policy debate. Nevertheless, as will be discussed further in Section IV, a more comprehensive discussion of macroeconomic program design and the accommodation of aid flows did subsequently take place, albeit with little input from civil society. Comprehensive in recognizing the multi-dimensional nature of poverty 56. The definition of poverty in the PRSP includes both “income” and “nonincome” poverty. The definition relied on feedback from the zonal workshops conducted for the PRSP and on a variety of (somewhat dated) sources—the 1991/92 Household Budget Survey, the 1993 Human Resource Development Survey (HRDS), the 1995 PPA, and a 1998 survey done by REPOA. A comparison of the facets of poverty in the PRSP with those in the NPES shows that the two definitions overlap in many areas, although there are differences—with the PRSP also including social well-being (mainly governance issues) and vulnerability as important dimensions of poverty (see Box 2). Thus, the PRS’s poverty definition was wider than the poverty profile of the NPES. 57. A majority of stakeholders, across all groups including government, indicated that they considered the current balance in the PRS between social sectors (education and health) and other sectors important for poverty reduction to be inappropriate since it did not pay enough attention to issues of income generation. Stakeholders noted that the poverty diagnosis in the PRSP was not comprehensive enough because the conceptualization of vulnerability, particularly with regard to disabled persons, was shallow, and feedback from the poor at zonal workshops on how governance severely affects their well-being was not given sufficient weight in the PRSP document. They blamed this on inadequate consideration of their inputs during the rushed consultative process.  30 Box 2: Definition of Poverty in the PRSP and NPES PRSP NPES Income Poverty Low Incomes Education Illiteracy Survival High Mortality Rate/Poor Health Services Nutrition Malnutrition Clean & Safe Drinking Water Inadequate Clean and Safe Water Supply Social Well-Being/Governance Vulnerability Environmental Degradation Unemployment Homelessness and Poor Housing 58. Aside from the definition of poverty, the focus in implementation of the PRS has been skewed towards improving human capabilities and survival (mainly primary education and basic health)—at the expense of both income poverty and other aspects of nonincome poverty (such as nutrition, vulnerability and social well-being/governance). In monitoring, for example, the Second Annual Progress Report does not include any social well- being/governance (although progress in governance is discussed36) or vulnerability indicators, acknowledging only that defining vulnerability indicators is challenging and that an on-going PPA exercise will help address this in the next progress report.37 59. Resource allocation has shown similar patterns with two-thirds of total priority sector expenditures (30 percent of total expenditure) in 2001/02 targeted at the human capabilities and survival dimensions of poverty i.e. education and health. On the other hand, the agriculture and roads sectors received less than a third (28 percent) of priority sector expenditures in the same period. It is, however, difficult to extrapolate from this, as in some sectors, such as agriculture, financial inputs may be less important than policy reforms. However, the draft PER on the development budget for 2003 notes that “the strategic weakness in implementing the PRS suggested by the 2002/2003 budget estimates relates to areas of spending that might be expected to directly generate sustainable increases in the 36 See Ronsholt et. al. (2003). The study indicates that the Local Government Reform Program emphasizes governance indicators more than the PRSP. However, baselines are not yet available. 37 The Third Progress Report (2002/03) which was completed in April 2004 (Government of Tanzania, 2004), that is after the Tanzanian case study was finalized, acknowledges that challenges exist in the translation of macro level achievements to the micro level, in commitment to good governance and in ensuring quality and equity in social service delivery.  31 income opportunities (livelihoods) of the poor, particularly through investment in rural infrastructure and support for improvements in productivity of poor farmers.” 60. In part, the lop-sided implementation is a result of delays in completing sector specific strategies. While health and primary education are most advanced in strategy formulation, with programs for each developed in 1999 and 2001, respectively, the secondary and tertiary sub-sectors are still developing their strategies and programs. The agricultural sector strategy was not finalized until 2001, while the corresponding Agricultural Sector Development Programme (ASDP) was not ready at the time of the evaluation mission. Similarly, while a Rural Development Strategy has been developed, its program of implementation still needs to be costed. In environment and gender, work is still on-going to integrate and mainstream these cross-cutting areas into the PRS process. 61. These delays also reflect the interests and attention of external partners. As discussed below, donors continue to play a major role in setting priorities in sector working groups. Thus, the more active working groups such as health and primary education, where SWAPs have been in place for some time, Figure 1. Tanzania: PRSP Priority Sector Expenditures, 1998/99-2002/03 also have the most advanced strategies and programs, and are able to bid for a bigger 1200 1000 60% 50% share of the budgetary resources. 800 40% 600 30% 400 20% 62. In addition, observers have 200 0 10% 0% commented that minimal attention has been 98/99 99/00 00/01 01/02 02/03 paid to the composition of growth and its impact on poverty in order to make necessary In Billions of Tz Shillings As % of Total Expenditures changes and to fine-tune the PRS strategy.38 The income poverty indicators Source: The Tanzanian Authorities; and figures from 1998/99 to 2001/02 are actual; 2002/03 are as budgeted. in the Second Progress Report measure rural and national poverty lines but do not contain intermediate indicators to track the effect of macroeconomic and structural reforms on poverty. The lack of attention to distributional issues is compounded by the weakness previously mentioned in monitoring vulnerability in the PRS. Results-orientated and focused on outcomes that benefit the poor Prioritization and mobilization of public expenditures in priority sectors 63. The PRS has enhanced the prioritization and the mobilization of financial resources in the priority sectors. Although efforts were ongoing in this area (as discussed in Section IIA on Relevance), particularly with respect to the MTEF for FY00–03 and the analytical work deriving from the PER process, the PRS has provided a poverty-focused policy framework 38 An Agriculture Survey was planned in 2003 to explore the linkages with poverty.  32 for budgeting that was lacking in the earlier Rolling Plan/ Forward Budget.39 A DFID/ODI assessment of the MTEF in Tanzania also concludes, on the basis of preliminary findings, that one of the positive factors influencing the success of the MTEF exercise has been the role of the PRSP in identifying policies and priority programs that form the basis for prioritizing the allocation of public expenditure resources.40 64. The share of PRS priority sector expenditure increased from an average of just over half of recurrent expenditures in FY99 and FY00, the two years prior to the PRSP, to 68 percent of recurrent expenditure in FY02. In nominal terms, total priority sector expenditures rose from approximately T Sh 500 billion in FY01 to T Sh 760 billion in FY02, an increase of over 50 percent (substantial in light of the low inflation) (Figure 1). As a percent of GDP, priority sector expenditures have grown from 6.5 percent in FY01 to 8.9 percent in FY02. Recurrent expenditures within priority sectors have also increased from 36 percent in FY01 to 45 percent in FY02 and are anticipated to comprise 50 percent of total recurrent expenditures in the next budget year. Priority sector expenditures are budgeted to further expand to 10.1 percent of GDP in FY03. 65. Perhaps as an indicator of the initially very optimistic projections of the PRSP, a 2002 study for the EU finds that budgeted allocations to the priority sectors in FY03 were below original projections. Moreover, the report also notes stagnation in the projected expenditure shares of the priority sectors between FY03 and FY05.41 However, differences between the original projections and the budgeted amounts reflect the view of the government that the PRSP is a “live document.” Moreover, they noted that since the original costing of sectoral programs was weak, adjustments were required. 66. The MTEF and budget guidelines for FY02 and FY03 have seen an increase in the number of sectors designated as priority sectors (called “key areas”). They now include 8 items additional to those mentioned in the PRSP: Police and Prisons, Good Governance, Census, Energy, Core Statistics, Lands, Controller and Audit General, and the Poverty Monitoring and Master Plan. This has contributed to an increase of 18 percent in the allocation to priority sector expenditures (Ronsholt, 2002), which should be taken into account in judging the expansion in “priority spending”.42 However, the expansion of priority 39 See Government of Tanzania (2001d) and Naschold and Fozzard (2002). 40 Bird (2002). 41 See Ronsholt (2002), cited in Ronsholt et. al. (op. cit). 42 Although “justice” had been declared a priority sector, respondents at the Ministry of Justice and Constitutional affairs during interviews for this evaluation noted that the emphasis was on lower level courts, while the prisons and the police were not considered priority sectors. The expansion was thus a clear attempt to correct the anomaly, but also indicating that a learning process was taking place.  33 sectors has not resulted, strictly speaking, from a broad consultative process, although the PER framework, where such discussions take place attracts a diverse cross section of stakeholders. There is also evidence that in FY02, the government protected priority sectors (where expenditure increased by 2.4 percent of GDP) in spite of donor budget support falling below estimates, indicating a high degree of commitment. Poverty focus of public expenditures within priority sectors 67. Improvements in strategic resource allocation via the MTEF process have been supported, and indeed facilitated, by the participatory Public Expenditure Review (PER) process, active in Tanzania since FY98. The scope of the PER process has variously included tracking of expenditure allocation and its impact, support to the government’s budget management process and systems, fiscal management and support to improve the coordination of donor aid into the budget. Chaired by the Ministry of Finance, with its secretariat originally housed at the World Bank, the PER process has involved a progressively larger group of stakeholders in its working groups including line ministries, external partners and civil society. 68. One of the objectives of the annual PER is to provide analytical work on the effectiveness of expenditures in the priority sectors. The FY03 PER in fact specifically set out “support to the implementation and monitoring of the Tanzania PRS” in its Terms of Reference, reflecting the greater institutionalization of the PRS in broader government processes. 69. However, spending within sectors is not well targeted to the poor. Box 3 demonstrates the lack of pro-poor targeting of expenditures within many priority sectors. This is in spite of the fact that the PER uses PRS targets as the benchmark for assessing overall performance in the sector. While this demonstrates the alignment of the PER with the PRS, it raises concerns that the PRS targets are not adequately defined to serve as the appropriate guide for assessing sectoral performance in the PER. Box 3. Tanzania: Lack of Pro-Poor Targeting of Expenditures The targeting of expenditures within priority sectors is not necessarily pro-poor according to findings of the 2003 reviews of sectoral expenditure. In health, allocations to local government are not correlated with poverty, taken as a proxy for health/health service need. The geographical distribution of development expenditure increases in the water sector does not suggest an effective poverty focus. Interviews with government officials in the Ministry of Health and Water and NGOs working in these sectors confirmed these findings. In general, geographical resource allocation is mainly determined by historical trends and funding of pre-existing facilities which leads to a bias in allocation of resources against the poorest regions and districts of the country. _____________________ Source: Government of Tanzania (2003c).  34 Poverty monitoring 70. The PRS process has substantively enhanced national processes for poverty monitoring although there are some inconsistencies in the follow through of actual PRS targets. 71. As described earlier (in the section on relevance) the NPES, and later, the PWMI, were both inconsistent in their elaboration of targets, and lacked robust and recent baselines. The baseline for indicators in the PRSP, selected from the PWMI, was based on a Household Budget Survey from 1990/91, and thus was outdated. 72. The PRSP described the planned surveys from which baselines and targets would be established, and referred to the development of a more elaborate monitoring system, which was underway. In December 2001, the Poverty Monitoring Master Plan (PMMP) (see Box 4) was launched which sets up the institutional framework for monitoring and evaluating the PRSP, the Poverty Monitoring System (PoMoS). Box 4. Tanzania: The Poverty Monitoring Master Plan (PMMP) The 2001 PMMP situates the institutional home for poverty monitoring in the National Poverty Monitoring Steering Committee, with a broad membership of different stakeholder groups. Four Thematic Working Groups (TWGs), the Surveys and Census Group, the Routine Data Group, the Research and Analysis Group and the Dissemination, Sensitization and Advocacy Group report to the Steering Committee. The Poverty Monitoring Secretariat provides administrative and coordination support to the Working Groups and the Steering Committee. ______________ Source: Government of Tanzania (2001b). 73. The government completed the 2000/01 HBS and the Integrated Labor Force Survey (ILFS), and published in 2002 the Poverty and Human Development Report (PHDR) which analyzes the results from the HBS. The Second Progress Report has updated the income poverty, human capabilities and survival targets of the PRS using the new baselines from the 2000/01 HBS. The results of the August 2002 Population Census, an Agricultural Survey and a Participatory Poverty Appraisal (PPA) in the pipeline for 2003 are anticipated to further enhance the PRS baseline and targets. 74. The Local Government Reform Program has attempted to create further sub- indicators for use at the local level. As mentioned, however, baselines are not readily available for the LGRP indicators (some of which would be quite problematic to obtain). 75. One area of confusion, however, is the government’s follow through of targets from the original PRSP. Besides the human capabilities and survival targets which have been  35 enhanced, revised and explicitly monitored in the Second Progress Report, targets in other areas have been omitted (for example, “development of a private sector strategy by 2003,” “expansion of investment productivity”), mentioned in passing in the text (“average time taken to settle commercial disputes”), or have never been defined explicitly (“strengthened professional and cost effectiveness of public service reform.”) Feedback of monitoring into performance 76. While the PER/MTEF and poverty monitoring systems processes have contributed to better tracking of poverty expenditures, prioritization and protection of priority sectors, and clearer tracking of targets with baselines, there is still a major gap vis-à-vis the feedback of PRS monitoring to actual policy formulation and performance. Information across the results chain (inputs, outputs, outcomes and impact) has not been integrated by policy-makers in a meaningful way. Some of the strengths and weaknesses are: i) Within sectors, allocations and re-allocations are generally informed by and follow the recommendations of the PER. For example, the 2002 Education Sector PER raised the issue of inadequate financing of operating costs in secondary education. The government therefore increased the allocation of funds to the department of secondary education. ii) Findings from the PER rarely inform resource reprioritization across sectors as sectors are bound to work within the budget ceiling given to them by the Ministry of Finance. There is some element of performance budgeting which requires that ministries and development agencies (MDAs) have to meet certain action plans before they get their budget—but that only covers “other charges” or about 20 percent of the priority sector budget. The fact that there is no inter-sectoral reallocation is not surprising since donors require that allocations to some sectors remain protected or keep growing. iii) Performance assessments of public service personnel are done by a separate agency, the Civil Service Department (CSD), which determines salaries and hiring. Although, recent documents and actions plans of CSD refer to the PRSP strategy, actual assessments are not based on PRS targets. iv) The Local Government Reform Program has also developed an M&E system for local government (from where 70 percent of PRSP funds will be disbursed). However, it is different and separate from the PRSP monitoring system. v) Each central ministry has its own method for data collection, but which eventually feeds into aggregate national statistics. The information flow is unidirectional (flows upwards) and does not feed back into decision making at the local level. The Routine Data Thematic Working Group has thus encountered a number of problems related to the parallel systems of data collection in the different agencies.  36 77. Overall, the PRS is significantly more results-oriented than previous strategies. Much progress has been made in funding priority sectors, and in improving national processes for poverty monitoring both in terms of availability of data and setting up an institutional mechanism. However priority sector expenditures are not necessarily pro-poor and the feedback from monitoring into implementation is insufficient and likely to be hindered by capacity constraints at all levels, particularly at the sub-national level. Partnership-orientated involving coordinated participation of development partners 78. Before the advent of the PRSP, the TAS was being developed to guide government- donor relations. The government views the PRSP and TAS, eventually completed in 2002, as complementary: the “PRSP provides a medium-term strategy for poverty reduction in the context of the HIPC Initiative. Whereas the PRSP details the characteristics of poverty and monitorable benchmarks for poverty reduction in the medium and long term, TAS provides a broad strategic national framework for managing external resources to achieve Tanzania’s development objectives” (Government of Tanzania, 2002). 79. The external partners view the PRSP as the guiding framework for development assistance, and indeed the PER/MTEF and budget processes are now fully aligned with the PRSP. However, the objective of the PRSP is narrower in the government’s view—focused on short to medium-term poverty focused activities. To the extent that the priority sectors identified are getting preferential treatment in budget allocations, this could potentially pose severe long-run development challenges in meeting other expenditures necessary for promoting long-term growth and sustaining poverty reduction. 80. From the beginning of the PRSP process, external partners have been consulted regularly and at key points in the process leading up to the finalization of the PRSP. This participation was, on the whole, at a healthy arms length so as not to interfere with the indigenous development of the PRS. Units of the World Bank and the IMF charged with the task of providing technical support for PRSP development in member countries were initially very eager to be involved in a “hands on” manner in the PRSP process but eventually stepped back. However, it is also acknowledged that at the beginning the World Bank and the IMF were not sure what their individual or joint roles in the process would be. The World Bank, with a stronger presence in the country than the IMF, provided technical advice and played a facilitation role which was widely appreciated. During the implementation phase, the process has continued to actively engage development partners in shaping national policies and strategies (e.g., CG meetings are now country based, and much more open and inclusive than before; external and internal development partners are invited to the annual PER/MTEF workshops, as well as to the poverty week, etc.) 81. Some of the NGOs have been critical of a perhaps too influential World Bank role in the process—sentiments partly stemming from the suspicions arising out of the PRSP being a condition for access to HIPC relief and the need for BWIs’ approval of the PRSP as a suitable basis for their concessional lending. But in fact, it was the World Bank’s insistence on a greater role for the NGOs that made possible a level of consultation that, despite its  37 shortcomings, was wider than experienced before and has opened the door for the broader consultative process that followed. Although the IMF was less involved in the participatory processes of the PRSP, the IMF’s approach has been increasingly influenced by the PRSP process. The domestic debate, especially that under the PER, has led to changes in program design, especially with respect to a more flexible fiscal stance and increased priority expenditures. However, dialogue on these issues seems to have been largely confined to that between the government, the BWIs and donors. Civil society gave little input (and many are unaware that the program did adapt). This appears to represent a combination of limited technical capacity by CSOs on these issues and the fact that the IMF did not frame them in a way that would contribute to a broader debate. These issues are discussed further in Section IVB. 82. However, the acceptance of NGOs as fully-fledged partners in the PRSP process has not been friction free. First has been the issue of legitimacy. In a country with a strong central government and where previously most civic activities were linked to the ruling party, the rise of the NGOs sector during politically turbulent times, demanding participation in all policy processes, and the fact that the first set of NGOs were funded by donors, who also emphasized the need for their participation in all processes, led to suspicion in government circles that NGOs were being used by external partners to put pressure on the government— not least in the political arena. Second, the right of NGOs to speak on behalf of the poor and to demand transparency and accountability from the government was questioned. Some parliamentarians blamed NGOs for supplanting them from their policy-setting role. In an effort to bring “order” to the NGO sector, but interpreted by other stakeholders as an attempt by the government to restrain its activities, an NGO bill was passed by Parliament in November 2002. Indicating a departure from the broad-based approaches that are being encouraged, the bill was not discussed outside official circles, notably by the NGO community on which it was targeted. 83. Beyond process, external partners are also substantively involved through various forums, but particularly through the sector and other working groups. They continue to wield substantial influence on strategy and policy formulation at the sectoral level, which eventually feeds through to the national level. The majority of the regular attendants at group meetings are external partner representatives and NGOs, with key government counterparts often missing. The discussions are often dominated by external partner representatives and specialists and it is far from evident whether the priorities are fully domestically owned and supported. There are also several complaints about some of the roles played by external participants, such as setting the meeting agenda or writing consultants’ terms of reference. These actions in part reflect the lack of time and capacity in participating government agencies, but also indicate the continuing legacy of external partner influence on domestic strategy formulation and the setting of priorities, raising concerns about the ownership of strategies and programs, even in the high priority sectors. 84. An important outcome of donor efforts at aid harmonization and policy alignment was the establishment of the Poverty Reduction Budget Support (PRBS) group in early 2000. This group replaced the Multilateral Debt Fund which was created to assist the country to  38 remain current in debt repayments to the multilateral agencies but had become redundant with the introduction of the HIPC initiative. The PRBS group, comprising eleven like- minded donors, participates in the PER process, holding regular meetings with the government, and is seen as an important player in the poverty reduction strategy.43 Support from the PRBS group is directly channeled to the budget, with its conditionality based on a commonly negotiated performance assessment framework (PAF). The PAF focuses on macroeconomic management, public expenditure management and governance.44 Generally, external partners are keen to participate in these harmonized processes, and recently the government held a joint portfolio review of the World Bank and the UN system. However, although donor harmonization is seen as desirable, and a move in the right direction, a recent review (Sida, 2002) has argued that it could have a downside, as when the most influential members of the group use it as a cartel, putting the government in a poor negotiating position.45 85. Through the PER process, the government has been able to establish a data base on donor commitments to the country over the medium term, enabling it to design its MTEF. Since this information is freely available, peer pressure is thus generated on the donor side. 86. Two other issues on partnerships are the still relatively low level of involvement in the process of the business community and the still high transactions costs for the government. Although private-sector led development has been emphasized during the PRSP process, the participation of the business sector has been marginal. Interviews with the umbrella organizations of the business community revealed that a number of issues which they considered important, including taxation, electricity tariffs, water supply, infrastructure development and the regulatory environment were not given adequate consideration.46 43 Three meetings are held with the government per year: an annual review in November; a mid-term review in March; and a review after the budget in August. 44 The PRBS is now linked to the World Bank’s Poverty Reduction Support Credit negotiated in early 2003, with conditionality outlined in the PAF (see PRBS Donors and the Government of Tanzania, 2002). Participants in the process are tied together by a memorandum of understanding. However, there is also a local Development Assistance Committee (DAC) which promotes dialogue among the donor community. Conditionality in the IMF-supported program is not directly part of this coordination mechanism. 45 The Sida (2002) report argues that the government’s purchase of an expensive air–defense system had led important donor agencies to threaten an aid embargo, urging other members of the PRBS group to do the same. A number of donor respondents also indicated the fear of being turned into “silent partners,” while others cited concerns over governance issues. 46 The newly started Business Environment Strengthening in Tanzania (BEST) program is yet to have an impact.  39 87. Although as noted, donor processes as well as those of the BWIs have been streamlined and aligned with the government’s budget cycle, the transaction costs of the government have not been reduced. The number of missions remains high, while donors are still reaching bilateral agreements with the government. Moreover, external partners have their own time constraints and prefer to deal with high-ranking officials in key ministries, notably the Ministry of Finance, during negotiations in order to expedite processes. As a result, policymakers have very little time left over for “thinking.” There is thus a serious case of dispersion of capacities, authority and resources to manage key functions. Government officials indicated that while donor alignment will eventually reduce these costs, setting up the mechanisms for monitoring and the follow-up required to ensure accountability and transparency will likely increase them in the medium term. A key recommendation of the IMG, yet to be implemented, was that the annual PER and the CG meetings should be merged as a way of reducing transaction costs. Based on long-term perspective 88. The narrow focus of the PRS implementation on human capabilities and survival (discussed earlier) raises concerns on whether the priorities of the PRS are consistent with the objectives of the long-term Vision 2025—namely achieving a high quality of life for all Tanzanians, a society characterized by good governance and the rule of law, and an economy which is competitive, diversified, semi-industrialized and macro-economically stable. The PRS, as it is currently being implemented, is not heading in the direction of achieving the holistic goals of the Vision 2025. 89. An example of a less than comprehensive focus of the PRS is in the education sector, which absorbs the bulk of PRS funding. PRS targets focus exclusively on primary education, which has also seen the highest budget increases in recent years. The implementation of the Primary Education Development Program (PEDP) has favored access and enrollment rates over issues of quality. There is a need to include PRS targets from, and gradually expand resource commitment to, the secondary and tertiary sectors in a holistic manner in order to “create a critical mass of a well-qualified human resource to face the challenges of globalization” if the objectives of Vision 2025 are to be met.47 90. The financial sustainability of the PRS in the long-term is an additional concern. Investments in primary education and health care, without significant accompanying efforts to boost domestic resources through income generation and savings, are not tenable in the long run. The government has recognized that it must broaden its own revenue base in order to ensure the sustainability of programs in education.48 Local governments are unlikely to be 47 Government of Tanzania (2003c). 48 Ibid.  40 able to meet the future maintenance costs of schools funded by the central government.49 Similarly, although the health sector has seen an increased share of central government expenditures it is largely from donor basket funding. 91. The response of the government to these relatively new concerns, which are being expressed by a broad group of stakeholders and in its own public expenditure reviews will dictate the long-term sustainability of the PRS priorities and its commitment to the Vision 2025. D. Preliminary Results 92. Despite the compromised quality of the participatory process and insufficient poverty diagnostics, for lack of adequate data, the PRS process is now well entrenched in the policy making and strategy formulation processes in Tanzania. However, there are a number of risks associated with the multiple initiatives currently being implemented, in particular, fiscal and administrative decentralization, civil service reform and public sector reform. The implementation of the PRS down to the grassroots level requires that the initiatives be better sequenced in order to enable capacity to be developed at the local and regional levels. Establishing a realistic mechanism for implementation at the sub-national levels remains the biggest challenge that the government faces in delivering the results promised by the PRSP. 93. Although it is too early to give a definitive evaluation of the impact of the PRSP approach, at the time of the evaluation mission, the process had been under implementation for some three years and CSOs and other observers had reported a noticeable improvement in service delivery, including at the village level. The Second PRSP Progress Report records a sharp increase in gross enrolment at the primary school level from 77.6 percent in 1990 to 100.4 percent in 2002.50 There were also increases in the number of teachers as well as classrooms at that level. The Report sees the tackling of the low quality of education as an important remaining challenge, however. In the health sector, the supply of drugs has greatly improved, including at local health units, attributed to an innovative cost-sharing approach drawing on the views of local communities. However, the high prevalence of HIV/AIDS and malaria have implied a rather marginal impact on the overall disease burden and other health indicators. 94. However, there have been no strong impacts on rural incomes and consumption levels. Although the Second PRSP Progress Report notes that a measure of policy coherence has returned to agriculture with the completion of its sector strategy, the government remains keenly aware of the inadequate real sector response and has already expanded the list of priority sectors. It is also planning to revisit the strategy in efforts at striking a better balance 49 Government of Tanzania (2003b). 50 The Third Progress Report (2002/03) was completed in April 2004 (Government of Tanzania, 2004) and was not available at the time of writing this report in late 2003.  41 between income and nonincome aspects of poverty. The Report indicates that the next steps will include paying more attention to cross-cutting issues, including the environment, gender, HIV/AIDS and employment, as well as the implications of the restructuring of local government. 95. A level of institutionalization, in key central ministries and for key policy formulation and budget processes, has been achieved under the PRS. The government has established a good monitoring and evaluation system, capable of providing necessary feedback into the policy and strategy formulation, although weaknesses remain. Capacity constraints and the political economy of changing policies and budget allocations imply that the penetration of the process and principles at the lower, and arguably more critical levels for achieving results, is slow. 96. There is ample evidence that the policy and strategy formulation process, at least at the national level, is much more participatory and transparent than before, and the focus on poverty reduction priorities has increased.51 IV. IMF ROLE AND THE PRGF A. IMF Inputs IMF inputs into the macroeconomic framework 97. The PRGF-supported program was launched in April 2000, following the completion of the interim-Poverty Reduction Strategy Paper (I-PRSP) in March 2000, which was a prerequisite for reaching decision point under the HIPC initiative. Thus, the PRGF preceded the full PRSP by six months. It became the IMF’s main vehicle for providing inputs into the macroeconomic framework of the poverty reduction strategy in Tanzania and for participation in domestic debates and processes and collaboration with the World Bank other donors. IMF inputs into the PRSP 98. The IMF was involved in the formulation of the PRSP’s macroeconomic policy framework both via its PRGF-supported program, which, as noted, was in place before the full PRSP, and directly via its policy advice during staff missions, correspondence between staff and the authorities, and participation by resident mission staff in the policy debates in Tanzania during the period that the PRSP was being formulated. However, while the draft PRSP discussed extensively some aspects of poverty reduction, such as service delivery and human capital development, strategies for growth as well as issues related to revenue 51 This general indication of positive learning by doing effects of the PRS process is strengthened by the results of the recent Third Progress Report (Government of Tanzania, 2004), with prioritization and effectiveness seen as crucial to further progress.  42 generation and public expenditure management, highlighted in the IMF-supported program as requiring focus, were addressed in only a limited manner. The conclusion reached was that the PRSP preparation required more time to meet these requirements, although a joint IMF- World Bank mission did not propose more specific interventions to support the process. 99. This experience suggests that IMF staff faced a dilemma. On the one hand, the PRSP was expected to become the government’s overall policy guide, including in the macro area, which required that it address all major issues in a manner that would provide broad operational guidance for other parts of the policy framework, including the PRGF-supported program. On the other hand, it was important that the PRSP emerge from a country-driven process. Interviews with staff suggest that they had insufficient guidance on how they should intervene in the domestic debate if important macro policy issues were not being addressed. Thus, while an arms-length approach to the PRSP preparation process seemed expedient, a number of procedural and technical difficulties discussed in this evaluation could have been avoided if a means of interaction with the authorities during PRSP preparation had been devised so as so ensure that all major policy issues were addressed in the process while not impairing domestic ownership. Design and content of the PRGF-supported program 100. The medium–term objective of the PRGF was to create conditions favorable for higher growth, with a particular focus on poverty reduction. Macroeconomic objectives included increasing growth from 5.2 to 6 percent annually by the end of the program, and reducing inflation from 7 percent to rates comparable to those among Tanzania’s main trading partners, that is about 4 percent. Monetary expansion was to be kept at about 10 percent per year. Surprisingly for a program focusing on poverty reduction and the expansion of service delivery, total government expenditure was expected to remain relatively unchanged at about 16 percent of GDP. Grants as a percent of GDP were also projected to decline by a full percentage point during the program. The implication was that increasing priority expenditure would require reduction in nonpriority spending relative to GDP. At this stage, the higher resource inflows expected from the HIPC initiative and grants were not yet sufficiently factored into the projections. The original PRGF thus projected a fairly restrictive fiscal stance. 101. However, comparison of the PRGF-supported program to the ESAF arrangement of 1996–1999 indicates differences in three important aspects: (i) public expenditure reviews came to play an important role in the design and implementation of the PRGF-supported program, eventually increasing resource flows to priority sectors; (ii) the program coincided with an unprecedented inflow of donor support to Tanzania, which meant that a broader range of expenditure priorities could be addressed within a wider resource envelope; and (iii) the two programs also differed in terms of conditionality (see below). 102. In retrospect, the design of the PRGF-supported program did show flexibility to allow changes in focus and targets, as dictated by the requirements of the PRSP (see Table 3). In the first year of the program the emphasis was on the consolidation of macroeconomic  43 stability; commitment and expenditure controls to ensure transparency and accountability; revenue generation; and the expansion of credit to the private sector. In the second and third years of the program—that is after the introduction of the PRSP—the focus was extended to ensuring higher outlays on priority sectors; increased focus on the IMF’s core areas of responsibility; and modifications to the design of the financial program, especially toward increased fiscal flexibility. Table 3. Issues Around which the PRGF-Supported Program of 2000/01–2002/03 was Designed Emphasized from More pronounced the beginning of after first year of the program the program 1. Consolidation of macroeconomic stability X 2. Focus on structural reforms in the IMF’s core areas of X responsibility 3. Reduced conditionality X 4. Commitment and expenditure controls to ensure transparency X and accountability 5. Higher outlays on priority sectors X 6. Increasing fiscal flexibility X 7. Revenue generation X 8. Expansion of credit to the private sector X Source: IEO Staff assessment, based on program documents. Changes in internal fund procedures Changes in approach in the field 103. The PRSP/PRSP process has led to a number of changes in the IMF’s procedures in the field. Interviews at the Ministry of Finance, Bank of Tanzania and the President’s Office, Planning and Privatization, indicated that since 2000 the key economic projections were reached by consensus. This was considered by government officials to be an important departure from the past, when missions routinely arrived in Tanzania with the bulk of their analysis largely completed. An equally important sign of change, from the point of view of government, is the fact that Tanzanian officials now write the government’s letter of intent (LOI), with IMF staff commenting afterwards as appropriate. In the past the first draft of the LOI was written in Washington, with the government commenting when it was completed. Officials also noted that IMF staff had shown greater flexibility in light of domestic institutional constraints. For example, while the IMF had wished to have the Accountant General report directly to the Minister of Finance and not to the Pay Master General as was the tradition, the government argued successfully that this would rob the latter of his authority. The IMF conceded the point, even though staff had concerns with the approach adopted. 104. Government officials and the donor community also indicated that the timing of IMF missions was now better aligned with the budget cycle of the government. The first (review)  44 mission comes earlier in the year before the government’s budget plans are firmed up, enabling the IMF to make initial pronouncements on where it sees the process moving. The second mission arrives later in the year, enabling the IMF to undertake a review of the implementation of the new budget. By implication, the IMF missions are also better aligned with other donor processes, notably the Poverty Reduction Budget Support (PRBS) group which provides budget support to the government, and where the missions’ assessments are crucial to the group’s work. However, many donors, including UN agencies, felt that IMF staff still preferred to remain in the background, even when they could help to explain to stakeholders why the IMF was pressing for certain macroeconomic measures. They also argued that the organizational model of relying on relatively short-term missions from Washington and a small resident mission staff in Dar es Salaam tended to limit what can be achieved by the IMF in terms of participating in a broader policy debate. 105. Most of those interviewed agreed that the IMF has made it a point to listen to a broader range of views, and not just those from the Ministry of Finance and the Bank of Tanzania. However, respondents also noted that mission interactions with officials outside these institutions were in the nature of public relations, providing little opportunity for exchange of views on policies. 106. The role of the IMF Resident Representative in the domestic policy debate in Tanzania has clearly been enhanced during the PRSP/PRGF process. The Representative is not only invited to all major meetings on the economy, often with a major role assigned, but is also routinely invited to workshops organized by NGOs and groups with which the IMF would have had little contact only a few years ago. The IMF Representative is also considered a key member of the donor community, in the case of the PRBS group expected to attend meetings and to provide advice on macroeconomic issues on a regular basis. However, the IMF Representative’s office is quite small (the only other technical member of the staff being an economist recruited from the Bank of Tanzania) and can only take on a limited number of assignments and tasks. 107. As will be discussed further below, the existence of an established mechanism—the public expenditure review (PER) process—which channeled ongoing discussions on economic policy among stakeholders beyond the initial PRSP discussions, facilitated the opening up of the policy debate in the macro area even though many stakeholders do not yet feel fully included.52 52 Initially, the IMF was not engaged in the PER process, while the bulk of the donor community was on board. However, interviews with officials and the donor community suggest that active involvement on the part of the IMF Resident Representative from the late 1990s, including regular attendance of the meetings of the PER macro sub-group, helped raise the IMF’s profile in the domestic policy debate.  45 Changes in internal IMF policy formulation procedures 108. To assess the extent to which internal IMF processes have adapted to the PRSP and PRGF, we assessed internal briefing papers and review departments’ comments against a series of specific criteria derived from the objectives and key features of the PRSP and PRGF (see Appendix 1).53 The overall conclusion of this exercise is that the IMF does seem to have changed its internal processes with respect to how briefing papers and internal review comments address the key policy issues to be discussed with the authorities – although the changes have been more incremental than expected. Internal policy formulation and review processes have evolved towards better discussion of the PRGF-PRSP linkages, as well as poverty issues more broadly, while a similar improvement in focus is evident with respect to mobilizing external resources for domestic financing of priority expenditures as well as encouraging domestic participatory processes in resolving policy issues. However, internal briefs and associated review comments rarely discuss policy alternatives, nor do they suggest ways of protecting program objectives in the event of negative shocks. A key shortcoming, given the emphasis on it in policy statements of the IMF, is lack of discussion of PSIA, including how to make it operative. Although mentioned in later briefs, PSIA has still not been a substantive input into the process. Change in Technical Assistance (TA) 109. The Fiscal Affairs Department (FAD) has been the main source of IMF TA to Tanzania in the period since the mid-1990s, and is used here to illustrate how IMF TA aligned with the PRSP.54 In the area of fiscal policy, the PRSP set the following goals: Allocate budget expenditure according to the medium expenditure framework and the PER guidelines; expand the use of the integrated financial management system; and strengthen expenditure management capacity at the local level. 53 For each criteria, the processes were ranked according to a four-scale range (highly inconsistent, inconsistent, consistent, highly consistent). While such judgments inevitably involve a subjective element, the division of the broad issue of the extent of adaptation of internal briefing paper processes into a series of sub-criteria, with a broad indication of what we would expect to see to warrant a particular classification, should provide some degree of uniformity to the judgments. Further details of the methodology underlying the classifications will be provided in an Annex to the main evaluation report. It should be stressed that this particular assessment is focused on the contents of the briefing papers, not on final outcomes of negotiations. Moreover, we include a ranking of “ESAF” briefing papers only to obtain an indication of what changes have taken place over time. There is no suggestion that briefing papers for this period could or should have been “consistent” with approaches that were only fully developed later (see Annex II, Table 3). 54 The other main areas of technical assistance were with regard to monetary policy, banking reforms and statistics.  46 110. The main areas of IMF fiscal TA to Tanzania in 2000–2002 focused on public expenditure management, fiscal decentralization and tax administration, respectively, and therefore do appear to have addressed the PRSP objectives. The TA reports underlined the need to raise institutional capacities at the central and local levels in order to ensure better budget execution and to raise public sector accountability and transparency. IMF staff also urged a slowdown in financial decentralization to allow for the development of technical capacities at the local government level. 111. More generally, interviews undertaken in the Ministry of Finance during the evaluation indicated that the government was not always satisfied with the procedures used in procuring IMF TA and how it was administered. For example, a number of officials expressed concerns that the government was not allowed to identify its own consultants, even though, in its view, there was a sufficient number of local consultants that could undertake some of the tasks. Officials argued that their limited involvement in the choice of consultants tended to reduce their ability to monitor TA activities. Moreover, in their view, IMF TA was often too closely linked to IMF conditionality—for example targeted at helping the government prepare legislation related to key reforms, and not to capacity building. However, IMF staff argued in interviews that while there might have been such incidents in the past, IMF TA is offered on a mutually agreed basis, with the government allowed to specify its nature and duration. They also pointed at the crucial role played by the IMF in strengthening the policy analysis department of the Ministry of Finance. They also noted that in many cases the IMF is, thanks to its expertise, only the implementer of the TA, while financial assistance is from a bilateral donor, with the scope of the activities already agreed with the government. B. Aligning the PRGF with the PRSP 112. Interviews undertaken for this evaluation revealed that, at least in the first year of the PRSP, the PRGF-supported program provided the benchmarks and guidance required by the government and donor community to plan their activities, making it the key macroeconomic framework behind the process. Was this a transitional issue and are there signs that the situation is changing over time as the PRSP process becomes more embedded? 55 To assess how the PRSP process has affected the scope of policy choices available to the government and how the PRGF has adapted to these policy choices, we have examined in depth four policy issues in the macroeconomic area. The examples include two cases where a broader 55 Another way of posing this question would be in terms of temporal alignment (i.e., of whether the timing and phasing of the PRGF program formulation process was aligned with the PRSP cycle) and of policy alignment (i.e. of whether the program, especially its macroeconomic framework incorporates and is consistent with the strategy set out in the PRSP). See Box 1 “Defining the Concept of Alignment” in Aligning the Poverty Reduction and Growth Facility (PRGF) and the Poverty Reduction Strategy Paper (PRSP) Approach— Issues and Options (see IMF, 2003).  47 debate in the PER working groups led to a convergence of views that in turn led to important adjustments to the macroeconomic measures; one case where the political economy nature of the issues and lack of a domestic debate along the lines of the PER have evaded easy resolution; and a last case where, in spite of the government seeing it as a priority and discussion under the PER working group, and with clear relevance for the PRSP, IMF staff remained skeptical and alignment of the PRGF remains incomplete. The examples are: (i) how the financial programming framework was adapted to make it more accommodating to increased aid flows and domestic financing; (ii) how the cash budgeting process was relaxed—along with greater realism in overall budgetary projections; (iii) trade reform; and (iv) fiscal federalism. Adapting the financial programming framework 113. The macroeconomic framework underlying the original PRGF-supported program as well as the PRSP was designed along the lines of the earlier ESAF-supported program, reflecting the earlier focus on restoration and maintenance of macroeconomic stability. Notably, the government continued to maintain a tight fiscal stance, involving net repayments to the banking system, in a bid to boost credit to the private sector. The private sector was, however, not able to use the expanded liquidity to any large extent, while, thanks to HIPC-related inflows and grants, international reserves were expanding rapidly, reaching seven months of imports of goods and services in 2002, as opposed to the four months targeted in the PRSP and PRGF. The combination of higher concessional financing, a sharp accumulation of reserves, negative net bank credit to the government, and continued fiscal arrears suggested that the fiscal stance was too tight. Moreover, there were signs that the wrong things were being squeezed by the cash budget controls. This was of real concern to the donor community which commissioned papers under the auspices of the PER macro sub- group to look at issues of fiscal flexibility and budget sustainability (see Bevan, 2000; 2001). 114. Ultimately, the debate was shaped by two issues: First, the medium-term expenditure framework (MTEF) required modification in order to accommodate increased aid flows. The PER macro sub-group argued that it was possible to finance a more ambitious set of targets and recommended that the MTEF projections be based on the most likely scenario—which was that aid inflows would remain at a high level in the medium to long run. Second, that the envisaged “crowding in of the private sector” via fiscal restraint that characterized much of the design of the ESAF-supported program and continued during the first annual program of the PRGF-supported program had in practice overestimated the speed and response of private sector credit demand. Although macroeconomic stability had been restored, the government was reluctant to undertake net domestic financing to expand priority expenditures or to compensate delays in donor disbursements. 115. The PER macro sub-group’s analysis of the problem found immediate resonance within the government and the donor community. Moreover, the IMF did show flexibility in adapting the financial program to the revised circumstances (see Table 4 and Figure 1 (where original ESAF and PRGF projections are inserted)). The overall deficit (before grants) was allowed to increase almost twofold between 2000/01 and 2002/03 from 5.4 to 9.8 percent of  48 GDP. Priority expenditures were projected to increase to 9.6 percent of GDP in 2002/03 from 4.7 percent in 1999/2000. The bulk of the expenditure expansion was thanks to increased grants and concessionary inflows. The government has not increased net domestic financing to any large extent (see Table 4). However, the government had undertaken net domestic repayments on the order of over 1 percent of GDP per year since 1996, and the fiscal flexibility implied by their removal was substantial. Table 4: Financing the Domestic Deficit Under the PRGF-supported Program: Actual Outturns (In percent of GDP) 1999/00 2000/01 2001/02 2002/03 Overall deficit (before grants) 7.8 5.4 5.6 9.8 Grants 4.5 3.8 4.5 5.7 Overall deficit (after grants) 3.3 1.6 1.1 4.1 Financing Foreign (concessionary loans) 1.5 1.2 1.4 3.3 Domestic 1.8 0.4 -0.3 0.8 Sources: See IMF, 2000a; 2001a; 2002a,b. Figure 2. Tanzania: Overall Budget Deficit (after grants) * 5.0 4.0 Actual outcome 5th Review 4th Review estimates, March estimate,November 3.0 2002 2002 Percent of GDP 2.0 1.0 Original PRGF 0.0 targets, 2000 Actual outcome -1.0 Original ESAF targets, 1996 -2.0 Actual data targets/estimates -3.0 96/97 97/98 98/99 99/00 00/01 01/02 02/03 * Positive number signifies deficit and negative number a surplus. Source: IMF Staff reports; WETA database. 116. The experience described above suggests two important lessons. First, the consultation processes under the PRSP did, over time, have a significant impact on the content of macroeconomic policies, and the PRGF-supported framework did show considerable flexibility. In this sense, there was eventually an “alignment” with the key objectives of the PRSP although (for reasons discussed in the preceding section) the process was not driven by preparation of the original PRSP. Second, the process of interaction and debate that led to revisions in the framework involved some, but not all, stakeholders:  49 government, donors, and the IFIs were involved, but civil society did not participate in any major way on this specific issue. This appears to have reflected in part their technical capacity constraints. But the example suggests that even this should not have been an insuperable obstacle, because one of the key inputs was provided by an outside academic advisor (funded by a donor). However, this relative lack of civil society participation was unfortunate, since it has influenced their perceptions of the process. Indeed, evidence from interviews and the stakeholder survey indicate a striking lack of recognition within civil society that the macroeconomic framework did adapt. The cash budget 117. The cash budget was introduced in early 1996 by the new government via the creation of a central payments office at the Ministry of Finance and the institution of a single treasury account at the Bank of Tanzania. It works on the simple principle that expenditures in a given period (one month when introduced) should not exceed available resources (i.e. domestic revenue plus donor support). Though meant to be a temporary instrument for correcting fiscal indiscipline, it has proved resilient, helping the government to achieve macroeconomic stability. However, from very early on in the process, the cash budget was seen as a crude tool not able to meet the resource allocation and prioritization required under the poverty reduction strategy. First, it invariably led to expenditure cuts whenever revenue interruption was detected, even if of a temporary duration. Despite attempts to ring fence them, priority sectors also suffered from the effects of the cash budget, directly through reduced outlays, and indirectly via impacts on the nonpriority sectors, such as infrastructure, to which they were linked. Moreover, although priority sectors were guaranteed a certain amount of money on a monthly basis, this was too short a period to enable forward planning. During field interviews conducted for the evaluation, officials in sector ministries argued that the cash budget gave too much power to the Ministry of Finance, as the ultimate allocation to a ministry during a serious cash shortage was arbitrary. Since priority expenditures included wages to civil servants and debt servicing, the item most hit by the cash budget was “other charges” or working capital. Civil servants continued to earn their wages but without basic supplies and equipment. Thus the system impacted negatively on the allocative efficiency of the budget. Moreover, given poor commitment and expenditure controls, government agencies and ministries procured goods on the basis of promissory notes, and incurred subsequent arrears.56 118. All these factors contradicted the goals of medium-term expenditure budgeting and prioritization, which required a degree of predictability of cash flow in order to enable planning, as well as the focus on poverty reduction. To protect social expenditures it was necessary to make the cash budget system more flexible and its rationing mechanism more transparent. Reviews of the PRGF-supported program highlighted the need to replace the 56 IMF staff estimated that the stock of arrears amounted to one percent of GDP at the end of 2000.  50 cash budget with a more transparent system for resource allocation. On its part, the PER working group focused on resolving the negative effects of the cash budget. As part of its work program in 2001, the working group commissioned a study on fiscal policy management. In the light of HIPC debt relief and higher inflows of grants, the study recommended a fiscal expansion, along with the adoption of a more flexible process for allocating resources to priority sectors (Rutasitara, 2001). This had a direct impact on budget execution and led to a switch from monthly to quarterly allocations for priority sectors, although nonpriority sectors were still allocated money on a monthly basis. Thus although the cash budget is still required for expenditure discipline, the cash budget system did adapt to demands for increased expenditure flexibility. Once again, the consultative process was crucial, albeit again with government, donors and the multilateral agencies taking the lead, and with only limited civil society participation. Trade reform 119. A large part of Tanzania’s trade reform was undertaken during the ESAF. At the launch of the PRGF in 2000, the government’s trade reforms had been ongoing for the previous four years. They had virtually eliminated all nontariff barriers on imports, while the general export duty was also removed. The simple average of import duties was reduced from 23 to 16 percent. Further reform during the PRGF-supported program has been extremely difficult, with the government coming under increasing pressure from local producers to limit the pace and extent of trade liberalization. The reduction in customs duty rates proposed in the 1999/2000 budget was opposed by local producers. 120. Trade issues were mentioned with very little specificity in the PRSP. The document refers, for example, to the “importance of promoting export-oriented expansion” as well as to the “need to enable the poor to share increasingly in the benefits of globalization” without specifying the policy measures required. There was no mention of further tariff reforms or discussion of the implications for the poverty reduction strategy of the fairly advanced regional integration project on which Tanzanian has embarked with Kenya and Uganda. During negotiations of the PRGF-supported program in late 1999 and early 2000, IMF staff advised the government to undertake further trade liberalization, including a reduction of the top tariff rate from 25 percent and lowering the number of nonzero tariff bands from four to three. However, the government found it difficult to lower tariffs any further, fearing a decline in revenue and a political backlash.57 In 2001, under pressure from local producers the authorities imposed suspended duties ranging from 10 percent to 50 percent on 12 categories of imports to counter “dumping”. Sugar imports also now required licensing and registration (a direct concession to the local sugar mills). IMF staff were concerned about this escalation and advised government to remove the surcharges and controls on sugar imports, 57 Although review departments had advised that a 15 percent top tariff rate be included in the HIPC completion point conditions, the government convinced the mission team that this would damage its revenue base, which was still heavily based on trade taxes.  51 arguing that the policies were distortionary, with adverse impacts on the poor groups that the government sought to protect under the PRSP. A completion of a timetable for the elimination of the former was to constitute a performance criterion under the 2002 program. 121. Thus, unlike fiscal management issues under the PER—on which agreements were reached under processes that allowed for significant, albeit imperfect, participation—trade reform was not addressed in any meaningful way by the PRS process and has subsequently proved difficult to move forward. Moreover, there was little effort by the government to analyze the impact of the strategy on poverty or to identify potential winners and losers among the poor. Invariably all efforts at further reduction in customs duties and other forms of protection have been counteracted by new measures to protect domestic producers.58 Recently, a National Business Council, chaired by the President, was created as a forum for policy discussions between the government and the business community and trade policies will be high on its agenda. It seeks to establish a partnership with government that will ensure increased investment, extension of business activities to the countryside and greater compliance on issues such as taxation. However, the approach does not envision a participatory process on the lines of the PER. Fiscal federalism 122. From the point of view of poverty reduction, fiscal decentralization is very attractive. It is seen as not only bringing resources closer to where the people affected can monitor their use but also as a means of encouraging local ownership. There are many references in the PRSP to the government’s intention to devolve fiscal responsibilities for key services to local authorities, including primary education, security and justice, as the most effective way of empowering local communities and reducing poverty. However, IMF staff have consistently opposed a rapid pace of implementation. The report of a (2001b) mission on fiscal decentralization advised that the process be slowed down considerably. It identified a number of problems that needed to be addressed before fiscal decentralization could be expanded. These included lack of budgetary transparency at the local level, lack of clarity on procedures for analyzing local annual budgets, and a general lack of technical capacity at the sub- national level. 123. However, while conceding the problems highlighted by the IMF, the government has not been willing to abandon what it sees as one of its important legacies. Already in 1996, it began ceding control over the provision of public services to selected local governments, 58 Legislation on export promotion zones (EPZs) was presented to Parliament in 2002 but was not ratified. The government has also raised the issue of manufacturing under bond (see the Second PRSP Progress Report, Government of Tanzania (2003b)). IMF staff have argued that although these are responses to competition for investors in the region, they contradict the thrust of earlier policies, which was to create an even playing field for private sector participation in the economy.  52 with further expansion planned after 2004. In light of its economic and political importance, the question of fiscal decentralization has become a key focus of the work of the PER working group. A number of training seminars have been organized for local level officers under PER auspices. In 2002/2003 the government increased efforts at extending the integrated financial management system (IFMS) to local governments, first embarked on in the late 1990s. 124. In retrospect, how should the IMF have responded to fiscal decentralization, a process that was strongly supported by government as well as other domestic stakeholders, but one where the IMF’s cross-country experience suggested a well-founded need for caution on pace and sequencing? There are no easy answers, but clearly there was an argument for mobilizing strong TA (not necessarily from the IMF) to try and overcome the technical obstacles.59 Merely slowing down the process is not a sufficient response to a legitimate domestically- owned objective of stronger local communities that are able to prioritize their own expenditures. 125. In conclusion, while establishing causality is not straightforward, the latter years of the PRGF-supported program seem to indicate increasing influence of the PRSP process on macroeconomic policy. The examples shown above indicate that thanks to a domestic debate on important aspects of the macroeconomic framework, in this case on financial programming and the cash budget, the PRGF-supported program became better aligned to the PRSP. Although there were practical limitations to the nature of the debate, with civil society not able to participate fully due to lack of capacities and an evidently more dominant role played by donors as financiers and providers of technical expertise, the results from this debate, in terms of a more flexible fiscal stance and increased priority expenditures, better insulated from temporary budgetary shocks, were fully in line with the spirit of the PRSP. On the other hand, in areas where a broader debate had not taken place, such as in the area of trade liberalization and tariff reform, progress was slow. Last, fiscal federalism falls in a different category. Though clearly a political project, it is one with broad domestic support. The IMF supports it in principle, but is concerned about the speed of implementation, when capacities, especially at the local level are still fairly limited. 59 An IMF technical assistance mission in March 2001 on fiscal decentralization outlined the actions required, including strengthening capacities at the local level in the areas of budget preparation and execution as well as fiscal and financial reporting. Given Tanzania’s more than 120 districts the capacity-building requirements suggested by these measures are enormous (see Harrison 2001).  53 C. Movement Towards “Key Features of the PRGF” 126. This section looks at the extent to which the PRGF has moved towards the seven key features of PRGF-supported programs (see Annex II, Table 4). Broad participation and greater ownership 127. The first annual program (2000/01) of the PRGF was not based on the full PRSP, but drew on the earlier ESAF and on the I-PRSP. Though there was limited stakeholder participation in the macroeconomic framework formulated during this period, the processes on which the poverty reduction strategy was based, such as the National Poverty Eradication Strategy (NPES), Vision 2025 and the Tanzania Assistance Strategy (TAS), had been debated widely within government, civil society and by Parliament and the donor community. 128. The next two years of the PRGF-supported program, were more closely related to the PRSP. Macroeconomic measures, especially fiscal policy, were subjected to greater scrutiny and domestic debate under the PER process. A major outcome of the debate was the reaching of a domestic consensus on the need to increase priority expenditures, with the wider deficit financed by grants and concessional loans. Priority expenditure doubled during the course of the PRGF. 129. However, not all aspects of macroeconomic policy, notably trade liberalization and fiscal federalism, opened up to such debate while the participation of civil society in the macroeconomic policy debate was relatively limited. Embedding the PRGF in the overall strategy for growth and poverty reduction 130. The PRGF-supported program in Tanzania was embedded in the overall strategy for growth and poverty reduction in three respects. First, public expenditures were increasingly prioritized, with emphasis on accountability and ability of local level recipients to provide feedback to the center. Second, the program encouraged further reforms in key areas of economic activity such as taxation, land reform, and the financial sector, notably microfinance, with a view to expanding private sector activities, not only in the modern sector but also in the informal sector and the rural areas where most of the poor derive their livelihood. Third, with respect to growth, the program focused on the consolidation of macroeconomic stability and creation of incentives for investment. There have been two important shortcomings. First, the government has not been able to increase its revenue by much over the past decade: the revenue effort failed to go much beyond 12 percent of GDP, even as total expenditure as a percent of GDP rose to over 20 percent by 2002. A weak revenue effort puts the long-term viability of the poverty reduction strategy into question. Second, as indicated below, the ability to assess the impacts of these interventions on poverty has been limited by lack of PSIA.  54 Budgets that are more pro-poor and pro-growth 131. Tanzanian budgets have become more pro-poor during the PRGF, at least in the specific sense that sectors thought to have a huge impact on the welfare of the poor such as health, education, justice and other areas, such as HIV/AIDS, identified in the Tanzanian PRSP receive an increasing share of total budgetary resources. Budgeting is now more firmly based on the mid-term expenditure framework (MTEF), with priorities set under the participatory framework of the PER. The government has also increased the share of funding going to the purchase of goods and services (other charges) as opposed to the wage bill and statutory expenditures (see Box 5). 132. The extent to which budgets have been pro-growth is less clear. First, the development budget remains small. Moreover, there is concern that despite the alignment implied by the PER process, data on total donor-supported public investment projects remain partial and incomplete, leading to variations in the estimated size of the overall budget deficit. Second, while increases in social outlays in current expenditure are bound to raise the level of human capital in the long term, with positive growth effects, a lack of a rigorous method of assessment implies that the efficiency of that spending and its growth impacts are not known with confidence. Although there is an ongoing debate in Tanzania on how best to incorporate key sectors crucial to growth, such as transport and energy, into national priorities without distorting the emphasis on poverty reduction, what the government can do within its current budget constraints—and without greater revenue mobilization—is fairly limited. Thus, while the program design was changed to accommodate higher aid flows and the scope for a higher fiscal deficit was recognized, as before, growth assumptions remained largely exogenous to program design. Box 5: Tanzania: Survey Results on the PRSP: The Role of the IMF The section of the stakeholder survey concerning the role of the IMF in the process reveals that while government respondents had a favorable view of the Fund’s role, the view of other respondents was often the opposite. However, in each case, close to 40 percent of the respondents did not know or were not sure how to respond—with the largest number of those without an opinion being from civil society. • With respect to whether budgets linked to the PRGF are more pro-poor and pro-growth than before, government and donor respondents agreed, without ambiguity (scores of 3.5 and above), while international NGOs and civil society generally disagreed (scores of 2.5 and below).1/ • With respect to whether the PRGF indicates greater flexibility in policy design, government respondents again agreed, while donor respondents were neutral. However, international NGOs and civil society did not see the PRGF as enhancing flexibility. • Finally, on whether the role of the IMF in the PRSP process was helpful, government respondents were only moderately positive, while other stakeholders generally disagreed. _______________ 1/ The survey used a 5-point scale, with higher scores indicating a higher degree of agreement with the indicated proposition. See Annex I for more details.  55 Ensuring appropriate flexibility in fiscal targets 133. In the face of the government’s low capacity for revenue generation and its ambition to maintain low inflation, there was not much fiscal flexibility under the ESAF. The government depended on a cash budget to ration resources and undertook net repayments to the banking system in order to consolidate macroeconomic stability and encourage the expansion of credit to the private sector. As described earlier, a debate undertaken under the auspices of the PER, with the participation of donors and civil society and the government, made it clear that the fiscal stance in the initial PRGF-supported program was too tight as evidenced by a combination of higher concessional financing, a sharp accumulation of reserves, negative net bank credit to the government, and continued fiscal arrears. The debate led to the abandonment of the government’s policy of maintaining negative net domestic financing and to the restructuring of public expenditures towards increased shares of priority sectors. Thus the PRGF now accommodates higher total spending as well as a higher deficit than the ESAF-supported program. In this sense, the fiscal “policy space” under the IMF- supported program was broadened. More selective structural conditionality 134. Structural conditionality under the PRGF has been progressively streamlined, though not necessarily reduced in terms of number of conditions (See Table 5 and Appendix 2). The first year of the program had a number of structural carryovers from the ESAF-supported program, with performance criteria and structural benchmarks imposed on a total of eight areas including trade liberalization, public finance, financial sector development, financial transparency and good governance, petroleum, parastatals, poverty and statistics, in all totaling twenty-five structural measures, with seven focusing on privatization and SOE reform. The last two years of the PRGF saw a sharp reduction in both the numbers of conditions and the areas on which they were imposed, focusing essentially on public finance, financial sector development and governance and financial transparency. Within public finance, the number of structural measures has not been reduced, but there has been a shift of focus towards institutional reforms, including legislation, and public debt management. However, in discussing conditionality, it is important to note that the process is partly endogenous. For example, structural measures on state-owned enterprise reforms were reduced because privatization had matured—similarly for financial sector development. Still, in terms of numbers, conditionality in the last two years of the PRGF was not very different from the ESAF levels (Aggregate IMF/World Bank conditionality is discussed in section VI). Emphasis on measures to improve public resource management/accountability 135. The government has been able to increase the level of accountability in central government. The Integrated Financial Management System was introduced at the central level to provide a government-wide system for improving accounting practices and  56 accountability, and subsequently rolled out to the districts.60 The various reviews of the PRGF have included items that focused on streamlining procedures and on closing loopholes that facilitate abuse. In this regard, important legislation and anti-corruption strategies were introduced in 2001, notably the Public Finance Act. It established requirements for internal and external audits for every government ministry and agency. With a view to raising accountability at the local level the amounts of money sent to the districts from the center are published in local newspapers and on notice boards in local administrative centers. Local authorities are also required to present quarterly reports of their revenue sources. With regard to monitoring whether targets under the poverty reduction strategy are being met, the government launched an elaborate poverty monitoring master plan. However, to be effective all these innovations require adequate capacities for data collection and analysis, which are scarce in Tanzania. Monitoring and feedback thus tend to be slow, often relying on ad hoc monitoring and evaluation efforts by donors (see for example Ronsholt et. al., 2003). It is thus too early to tell with confidence what the impact on the ground has been. Social impact analysis of major macroeconomic adjustments and structural reforms 136. Poverty and social impact assessments analyses (PSIAs) have not been undertaken in Tanzania, so far. This has not been for lack of suitable areas for their application. IMF staff had as late as mid-2002 identified increased electricity tariffs, retrenchment in the parastatal sector, civil service reform and VAT exemptions as possible areas for undertaking PSIAs. The reason for this delay is not quite clear. During a May 2002 IMF mission, the authorities expressed doubt at their capacity to effectively undertake PSIAs, although IMF staff were of the view that some of the analysis could be done by domestic consultants. Funding was also a constraint, with the World Bank indicating that its budget for economic and sector work for the 2003 fiscal year did not include any support for PSIAs—still, the World Bank is currently undertaking a PSIA of the reform of crop boards. In retrospect, the practical difficulties related to implementing PSIAs, including costs, were underestimated as well as the willingness of policymakers to give the new instrument a high priority. 60 A recent IMF Board paper (see IMF 2002c) on improving poverty expenditure tracking in HIPCs, presents a total of 15 criteria that define the quality of a country’s public expenditure management system (PEM). Tanzania had satisfied eight of these by the end of 2001, placing it in a group of nine countries that required “some upgrading” of their PEM systems, as opposed to “substantial upgrading” for some fifteen countries (of the sample of 24 countries none had reached the top level of “little upgrading” required).  Table 5. Tanzania: Number of Structural Measures in ESAF and PRGF-Supported Programs by Type and Sector Breakdown 1/ Prior Performance Bench- Total formal Sectoral breakdown of structural conditionality (numbers) Program actions criteria marks conditionality Public Finance Financial Good SOE Other 5/ (1) (2) (3) (1+2+3) i. Revenue sector governance 3/ reform 4/ ii. Expenditure reforms iii. Other 2/ ESAF 1996-99 (i) (ii) (iii) First Annual 3 3 12 18 7 1 1 2 2 5 Arrangement Oct. 1996 Second Annual 3 12 15 4 1 2 6 2 Arrangement Nov. 1999 Third Annual 4 15 19 5 2 2 8 2 Arrangement Jan. 1999 Total for ESAF 3 10 39 52 16 4 1 6 16 9 PRGF 2000–03 - 57 - First Year (including 3 22 25 5 2 3 2 1 7 5 First Review) Mar. 2000; July 2000 Second and Third 5 11 16 3 8 5 Reviews Feb. 2001; Sept. 2001 Fourth and Fifth 5 12 17 3 2 7 2 2 1 Reviews Mar. 2002; Nov. 2002 Total for PRGF 13 45 58 11 4 18 4 3 7 11 Sources: IMF Staff reports for various arrangements (ESAF) and reviews (PRGF). 1/ The PRSP was endorsed by the Boards of the IMF and the World Bank in December. 2000, while the PRGF program began in March 2000. 2/ Includes institutional reforms and legislation in the fiscal area. 3/ Includes financial transparency 4/ Includes privatization 5/ Includes civil service reform, trade reform and public sector reforms  - 58 - V. WORLD BANK EFFECTIVENESS 137. Section A describes the key analytical inputs from the Bank in the process of formulation of the PRSP and assesses Bank behavior during PRSP preparation. It concludes by addressing whether the Bank’s support was conducive towards country ownership. Section B looks at alignment of the institution’s support following the PRSP. The Country Assistance Strategy covering FY01–03 provides the basis for the Bank’s planned lending and nonlending program. As the CAS was prepared before the PRSP, the overlap between PRSP priorities and the Bank’s analytical and lending activities, as they were implemented, is examined. The section concludes with an assessment of the Bank’s role in respecting country ownership, promoting donor coordination and working with national development partners since the PRSP was introduced. 138. In Tanzania, the process of formulating the PRSP took six months and the country has been implementing the PRS for more than three years. The findings in the first section relate therefore to a very short phase of formulation, relative to the much longer experience of implementation, to which adequate weight must be given. A. Bank Support During Formulation 139. As noted earlier, the government’s primary concern during formulation was to satisfactorily complete a PRSP in order to receive debt relief under HIPC. The overriding political expediency of achieving debt relief set the tone for most of the government’s dealings with the Bank. In general, the government was quite keen to receive guidance and advice in order to complete the PRSP and receive HIPC relief. Bank’s role 140. The Bank’s assistance for the PRSP was spearheaded by the Country Economist based in Washington DC. The Country Director and his team in the country office also played an important role in supporting the team responsible for drafting the PRSP. The network anchor in PREM provided direction and support during formulation. During the six month period of formulation, Bank staff assessed Tanzania’s progress towards a draft PRSP, proffered advice on areas that would be covered by the JSA and gave extensive comments on drafts. Washington-based staff carried out one joint Bank/IMF mission at the end of June 2000.61 61 OED did not have access to Bank correspondence files for this period. Information on this mission was drawn from the IMF’s “Briefing Memorandum for Staff Visit for Discussions on the PRSP,” June 22, 2000, which states that this mission went out with the objective of reviewing the nature and scope of the participatory process, encouraging the authorities to continue involving donors, discussing the data to be used to establish monitoring indicators for the PRSP and following up on progress with HIPC Completion Point conditions.  - 59 - 141. The role played by Bank staff during the preparation of the PRSP was described variously by stakeholders as “facilitative” and as one of “active involvement.” As reported in earlier sections, HQ staff were initially very eager in providing assistance to the government in preparing the PRSP. Other external partners were very critical of these “intrusive” actions by HQ staff which they believed did not respect the government’s ownership of the PRSP (since most other donors were not very involved at the start of the PRSP process). However most stakeholders agreed that Bank staff at the country office responded quickly to these objections and were able to effectively contain Washington’s influence on the process. Key members of the team responsible for drafting the PRSP concluded that the process was “government-led”. 142. An area that senior government officials criticized was the “formatting requirements of BWIs” for the submission of the PRSP to the Bank’s Board of Directors.62 Government officials objected to the fact that they had to adhere to BWI formatting requirements for a nationally owned document. In addition, the editorial changes the Bank made to the draft PRSP have been viewed by some stakeholders as indicative of the Bank’s intrusiveness in the process, resulting in the final draft of the PRSP losing some of its local “flavor” (Mukandala 2003). The principal drafters of the report, however, were adamant that no substantive changes were introduced into the draft. 143. During formulation, the Bank encouraged the government to deepen the consultative process for the PRSP but was unsuccessful in this area. Yet, the JSA did not reflect the widespread opinion of stakeholders that the consultative process was rushed and mechanistic. Civil society groups are critical that the Bank did not exert more influence to facilitate their participation, particularly in discussion of the macroeconomic framework for the PRSP. They acknowledge, however, the Bank’s dilemma between respecting country ownership and “enforcing” participation as a conditionality. The Bank, for its part, did not want to hold up Tanzania’s Completion Point under HIPC to allow a more effective consultative process. For that matter, neither did civil society. Thus, the design of the PRSP with its links to HIPC put the Bank and other stakeholders in a difficult position vis a vis the consultative process. Analytical inputs 144. Prior analytical work conducted by the Bank was useful in supporting the preparation of the PRSP. Previous studies included a comprehensive Country Economic Memorandum (2000) and from 1997 onwards, a series of Public Expenditure Reviews.63 The absence of a Poverty Assessment, given that a Household Budget Survey had not been carried out since 62 The policy documents issued by the BWIs at the time that Tanzania was formulating its PRSP do not stipulate formatting requirements for submission of the PRSP to the Boards of the two institutions (World Bank/IMF 1999a and b). 63 The role of the PERs is discussed in the next section.  - 60 - 1990/91 and the ability of the Bank to contribute in this area, is noteworthy. This was flagged in OED’s Country Assistance Evaluation, completed in early 2000, which recommended that the Bank could assist Tanzania in “helping to improve its social and economic database”64 and the June 2000 CAS had responded by planning a FY02 Poverty Assessment upon release of the results of the 2000 Household Budget Survey. It is unclear why a survey and Poverty Assessment had not been carried out earlier. The PRSP was significantly handicapped without recent survey data from which the government could draw a baseline for its targets. 145. In its advisory role, the Bank provided technical inputs which eventually influenced the sectoral contents of the PRSP. First, the Bank stressed the importance of tackling corruption, which had been raised as an issue by many participants during the consultative process. As a result, the government included the judiciary as a priority sector in the PRSP (although civil society felt that this did not constitute sufficient attention to governance). Second, cross-cutting issues such as gender and the environment were revised based on Bank feedback. Finally, Bank staff strongly influenced the policy decision made in the PRSP to eliminate user fees for primary education.65 National and international NGOs had been lobbying intensely on this issue during the formulation of the PRSP. The government (and the Bank) had resisted canceling user fees on the basis that cost-sharing was an essential component of their strategy, and the draft of the PRSP shared with stakeholders in August 2000 reflected this view. But the final draft of the PRSP that was submitted to the Cabinet abolished user fees for primary education. 146. Many civil society organizations expressed the concern that the strong influence of the Bank on the contents of the PRSP, while appreciated in the case of eliminating user fees in primary education, was a potential risk to country ownership. Government officials interviewed did not have any complaints on or reactions to the Bank’s influence on the policy contents of the PRSP as the desire to reach HIPC Completion Point was foremost on their mind. 147. The Bank played a substantial advisory and facilitation role in the formulation of the PRSP. Civil society was critical of this role but appreciated the Bank’s influence on some of the policy content of the PRSP. Yet they regretted that the Bank had not exerted more influence on the government in order to open up greater space for their inclusion. This highlights the tensions for the Bank between respecting a country-led process and influencing the government to ensure broad-based participation of all stakeholders. Government officials for their part were not very concerned that the Bank had wielded considerable influence on the process beyond their serious objections to “formatting requirements” and heavy editing of the PRSP by the Bank. Given the uncertainty of what was expected of a PRSP, and under 64 OED (2000), p.28. 65 For further information, see the following accounts: Alonso (2002) and Holtom (2002).  - 61 - pressure to reach HIPC Completion Point, the team responsible for drafting the document was generally appreciative of the Bank’s guidance and support. B. Bank Support and Alignment For Implementation Bank’s role 148. The Bank continues to play a strong coordination and facilitation role following endorsement of the PRSP by the BWIs in November 2000. It has continuously engaged in a dialogue with the government on the formulation of the 1st and 2nd Annual PRSP Progress Reports (dated August 2001 and March 2003) as well as on overall implementation of the PRSP. Support is formally coordinated through the PRSP Task Manager based in Washington, although the country office continues to play a strong role in PRSP discussions, with the cluster leader for all tasks related to the PRS process, public expenditure management and programmatic lending located in the country office. 149. Based on the information available to OED, the Bank conducted 4 PRSP related missions between October 2000 and March 2003. An October 2000 mission, conducted jointly with other donors, had the objective of clarifying the role of Bank assistance in developing the Rural Development Strategy and exploring linkages with the government’s ongoing work on agricultural sector strategy. A May 2001 joint Bank/IMF mission discussed progress with the implementation of the PRSP and provided initial comments to the government on the PRSP draft Progress Report. A December 2001 mission followed up on issues related to the outstanding poverty analysis and implementation of the Poverty Monitoring System. The mission had intended to initiate a discussion of the impact of growth on poverty, but this discussion was not advanced much further. Finally, a mission conducted in March 2003 reviewed the draft JSA of the 2nd Progress Report with the government. Bank behavior in promoting underlying principles 150. Subsequent to the formulation phase, the Bank’s overall behavior in terms of promoting country ownership is widely perceived to have improved significantly. There has been a “sea change” in the Bank’s behavior, and this was reported across the board from representatives in government, civil society and donor agencies. Using 1995 as a benchmark year, the Bank’s record in supporting country ownership was very poor. The Independent Monitoring Group reported in that year that “in practice the Bank (along with other donors) ignores the first conclusion of its own recent report on aid effectiveness that aid is most effective when it supports programs and projects that are “owned” by the recipient country.” The Bank was also criticized by other donors for its “unwillingness to engage in serious substantive discussions” with them.66 From this dismal point, the Bank progressed quite rapidly. The collaborative 1998-99 PFP and 1998 and 1999 PER processes had indeed 66 See Helleiner et. al. (1995).  - 62 - already been noted as a best practice by OED for stimulating “support for a country-driven formulation of the development strategy.”67 151. With the introduction of the PRSP and its principles, the Bank’s respect for country ownership has gone much further. A senior Ministry of Finance official noted that “bank behavior has changed dramatically in supporting country-driven processes”. There has been a reduction in the Bank’s adjustment lending conditionality which is now derived from the PRSP (discussed in detail below). One sector ministry noted that the World Bank has played a key role in improving the government’s ability to decide and the donor’s ability to give advice and assistance; an NGO in that same sector was in full agreement with this characterization. Two mechanisms have added to the impact of the PRSP in supporting this change. First, the Bank’s continuous involvement in working groups around the PER/MTEF process has provided a collaborative framework for much of its support (described in next section). And second, the relocation of the Country Director to Tanzania in 2000 supported by a team of Tanzanians at the country office has been critical in shifting ownership from the Bank to the country. 152. But there is always more room for country ownership—as one government official quipped, “The consensus has moved from Washington to Washington-Dar consensus.” Staff in the country office do a better job of fostering country ownership than sector missions from Washington. The latter often fail to adequately consider alternative viewpoints and are not willing to relinquish control to the ministry. While ministry officials were widely appreciative of the Bank’s technical skills, they complained that sector staff like to “come into the kitchen too often while you are cooking.” Weak technical capacity at all levels within government is part of the problem and the technical assistance provided by the Bank (and other donors) has not been successful in addressing this problem. At the same time, Bank sector staff “hold hands too easily” which inhibits learning amongst ministry staff. Project implementation units run separately from ministry processes create parallel structures that have also discouraged capacity-building within the ministry. Severe disbursement pressures on sector staff work further aggravate the problem. 153. The Bank has continued to promote the “partnership” principle of the PRSP. By 2000, the Bank had already made significant progress in enhancing “country-led donor coordination, encouraging transparency in aid flows and promoting harmonization of donor procedures” (OED, 2000). Moving the 1997 Consultative Group meeting to Dar es Salaam was a major step in this direction undertaken by the Bank. Subsequently, the Bank has gone further. The government and other donors pointed out the Bank’s proactive role in involving donors in the PER process, its alignment of the PRSC within the framework of the PRBS, and its constructive engagement with and support of the DAC as important measures that have contributed to greater donor coordination. At the sector level, two sector ministry officials noted that the Bank had exercised its power in a useful way by bringing donors 67 OED (2000), p.26.  - 63 - together and taking the lead in supporting government to do business. Here too, progressive leadership within the Bank office enhanced by the relocation of the Country Director to Tanzania in 2000 has played a critical role. Results from the OED/IEO survey in Tanzania provide support for the Bank’s positive role in donor coordination – almost 40 percent of respondents strongly agreed that the World Bank is promoting coordination of donor assistance.68 154. There are mixed reviews of the Bank’s coordination with donors at the sector working level. While the situation continues to improve, a strong criticism from several donors is that missions from Washington often arrive without adequate notice to or coordination with other partners working in the same sector. Government officials were unanimous in their view that there were too many missions from the Bank. Here too lack of capacity in the government is part of the problem. Senior Bank staff are aware of the “mission overload” and are working on improving coordination. They acknowledge that the PRSP has added new demands on the government’s time and that this has not been reciprocated by observing “quiet times” or asking less of the government in other areas. “The other shoe has not dropped.” Bank staff also noted that transaction costs are likely to be higher in a transitional period as the existing portfolio and projects are disbursed. 155. Although the Bank was criticized for not exerting greater influence on the government to involve civil society during the formulation of the PRSP, during the implementation phase the Bank has been responsible for their involvement in PER working groups. The Bank also attempted to encourage the government to carry out a broad-based participatory process during the preparation of the first Progress Report but was not successful. Civil society would like the Bank to insist upon greater representation of NGOs in the 4 working groups of the Poverty Monitoring System . But this presents a dilemma for the Bank—in fostering broader country ownership, it risks harming the country-led principle. Alignment of analytical and advisory assistance 156. During the implementation phase of the PRSP, the Bank provided technical guidance in key areas of the PRSP, including analysis of the Household Budget Survey, advice on the Poverty Monitoring System, sectoral inputs in education and health, support for institutional reform and investment climate promotion, and for the PER process through specific areas of expertise. This is acknowledged both by government officials and other external partners, who continue to look to the Bank for technical inputs and reviews. 157. Stakeholders unanimously acknowledged the positive role of the Bank-supported Public Expenditure Review process in PRSP implementation. As discussed in previous 68 Of 100 persons surveyed, 29 strongly or completely agreed with the following statement— “The World Bank is promoting coordination of donor assistance.” (23 answered “Didn’t Know” or “Unsure” and were withheld from the analysis.)  - 64 - sections, the Bank had conducted PERs prior to 1997/98 which were primarily external analytical exercises. From 1997/98 onwards, the PER process changed in two distinct ways. First, the process was aligned to directly support the preparation of budget guidelines and the MTEF while also monitoring the composition and effectiveness of the past year’s expenditure allocation and supporting the development of budget management systems. Second, the broad engagement of various stakeholders was encouraged. The PER Working Group now actively includes central and line ministries, external partners and representatives from civil society. In addition, Macro and Sectoral Working Groups were set up with similar membership. All groups have an ongoing work program and meet year round on a bi-weekly basis. The PER process today is broadly participatory, government-led, encourages domestic and external partnership and facilitates a multi-sectoral results-orientation of the PRSP via links to the budget and assessments of the poverty focus of expenditure. 158. The leadership role of the Bank in stimulating a participatory PER process has been widely acknowledged. Until recently, the Bank was providing secretariat services for the process which included offering its offices as the physical venue for meetings and publishing the PER report. In end 2003, they moved this function to the Ministry of Finance. Throughout, Bank staff have provided excellent technical inputs. In the words of a Bank staff member, the PER process supports “the right mix of capacity-building and criticism”. 159. There have been shortcomings in other areas of the Bank’s analytical support for the PRSP. There is a continuing knowledge gap about the micro-economic social and poverty impacts of macro-economic and structural policies, the sources of growth (particularly in the informal sector), and the poverty impact of alternative policy and strategy options. The Bank acknowledges that this remains an area for much work to be done, but notes that past efforts have been hamstrung for lack of appropriate and adequate data – even to undertake a proper diagnostic of poverty in the country.69 160. At the sectoral level, the depth and breadth of the Bank’s analytical work has been uneven across sectors. This has resulted in different sectors being at different stages of development of their sector strategies, which in turn has resulted in uneven implementation of the PRS priorities. For example, in agriculture the Bank staff acknowledge that analytical work should have started much earlier than it did. In maintaining a balance between growth and poverty reduction efforts, the Bank, via its lending program, has continued to focus on the growth agenda through its support for business climate issues to promote investment and for institutional reform (e.g. in capacity building for public expenditure management). But it has not been assertive in promoting the growth agenda to maintain the needed balance in policy formulation in the PER/MTEF and budget processes. 69 IMF staff have noted that the Bank has stated that it did not have the resources to carry out PSIAs in FY03. The Bank has agreed with the government to conduct a PSIA on agricultural tax reforms in FY04.  - 65 - Alignment of lending 161. The Bank’s CAS for FY01–03 was approved in June 2000, a few months prior to the endorsement of the PRSP. It is not known why the timing of the 2000 CAS was not delayed in order to align with the PRSP, but its focus—private sector and infrastructure development, sustainable rural development, improved social infrastructure, public sector reform and institution building—is in line with the priorities of the PRSP. New commitments across the three year CAS period reflect a similar picture of alignment: projects that fall into PRSP priority sectors comprised 75 percent of new commitments. (The Songo Songo Gas Development and Power Generation project accounts for the majority of the commitments that were outside priority areas of the PRSP.) Compared with the previous three year period, Bank commitments have broadened from public sector governance and private sector development into priority areas such as education, social protection and water supply (see Figure 3, which compares new commitments by sector during FY01–03 and FY98–00). Figure 3: New Commitments by Sector (FY98–00 & FY01–03) Water Supply and Sanitation Social Protection Rural Sector FY01–03 Private Sector Development Public Sector Governance FY98–00 Health, Nutrition and Financial Sector Economic Policy Environment Energy and Mining Education 0% 10% 20% 30% 40% 50% 60% 70% Source: World Bank Business Warehouse. 162. Bank support of sector budgets increased significantly after the PRSP. Commitments in SWAPs before the PRSP were US$22 million—since the PRSP, an additional US$207 million of IDA credits in Tanzania support sector-wide programs. This represents an increase from 5.2 percent of commitments in the 3-year period before the PRSP to 28 percent of commitments in the three-year period after the PRSP. (To the extent that SWAPs are based upon a country-owned program and improve donor coordination, this would demonstrate alignment with PRSP principles.) Before the PRSP, the Bank was participating in the Health Sector Development Program (HSDP) which it had joined in FY00. In FY02, the Bank became the largest financier of the Education Sector Development Program (ESDP) through the US$150 million Primary Education Development Program (PEDP). In addition, the Bank’s US$57 million Participatory Agricultural Development and Empowerment  - 66 - Project (PADEP) which became effective in July 2003 supports the government’s Agricultural Sector Development Program (ASDP) and the Rural Development Strategy. 163. The Bank’s total budget support through adjustment lending (including sector adjustment loans) has actually declined as a proportion of new commitments since the PRSP was introduced. Between FY98 and FY00, 47 percent of new loans were adjustment loans.70 The percentage of adjustment lending from FY01 to FY03 decreased to 39 percent of new loans. In absolute terms, new adjustment loans have increased from US$195 million in FY98-00 to US$284 million since the PRSP. 164. The May 2003 Poverty Reduction Support Credit 1 (PRSC1) represents almost half of the US$284 million in adjustment lending commitments since the PRSP was approved. The PRSC1 is a programmatic lending instrument specifically designed to support the implementation of the PRSP. It supports government efforts in three areas: (i) reforms in the areas of agricultural and rural development, (ii) institutionalizing Tanzania’s monitoring and evaluation systems, and (iii) strengthening public sector performance and the efficiency and effectiveness in the use of resources (see Box 6). Box 6. Tanzania: Poverty Reduction Support Credit (PRSC) Approved by the World Bank’s Board of Directors in May 2003, the PRSC is designed as a series of three annual single tranche operations of US$132, US$150 and US$175 million each. The PRSC selectively supports the implementation of priority areas of Tanzania’s Poverty Reduction Strategy. It emphasizes the “effectiveness of expenditure programs and holding sectors accountable for results and outcomes of programs through a strong M&E system”. Second, it “supports greater focus on pro-poor growth policy reforms, especially those affecting rural incomes”. The first tranche, PRSC1, became effective in July 2003 and included a grant component of US$32 million in addition to a US$100 million credit. The US$132 million disbursement was based upon the completion of seven prior actions. Two of the prior actions were contingent upon publishing the 2nd Annual Progress Report for the PRS and its approval by IDA. The other five were actions in debt management, financial management, rural development, institutional reforms, and private sector development. The PRSC1 also specifies seven triggers that would make Tanzania eligible for proceeding with PRSC2. ___________________ Source: World Bank (2003a). 165. The PRSC1 breaks new ground in aligning the Bank’s policy-based lending with the underlying “country-led” and “partnership” principles of the PRSP. First, conditionality has been reduced under the PRSC1 as compared to previous adjustment lending operations in Tanzania. The 1997 Structural Adjustment Credit, the 2000 Programmatic Structural Adjustment Credit and the 2001 Primary Education Development Program had 24, 28, and 23 conditions respectively. The PRSC1 has 14 conditions—7 prior actions and 7 triggers for PRSC2—but it is only the first in a series of three credits. Relative to the size of the 70 This does not include the HSDP which is an investment Adaptable Program Loan.  - 67 - programs, however, conditionality has declined much more. The US$132 million PRSC1 is two-thirds the size of the US$190 million PSAC but has half the conditions. The reduction in conditionality is largely due to a decrease in conditions in financial and private sector development, mainly related to privatization of SOEs. Under the PSAC, 40 percent of conditions fell into financial and private sector development. Of the 14 conditions in PRSC1, 14 percent of conditions are in that category. 166. Second, the PRSC1 allows the government greater control over its aid flows. Its ex- post conditionality permits the government to know with certainty what disbursements to expect within the year. It also allows flexibility in the triggers for future tranches so that disbursement is not held up by factors outside the government’s control.71 167. Third, the Bank has coordinated with other donors in adding the PRSC policy measures to the existing PRBS policy matrix, the PAF, and aligning the timing of the PRSC1 with the preparation and review of the PRBS. The PAF had already been developed by the government and PRBS donors. Its policy matrix uses the PRS as its guiding framework, is broadly aligned with the objectives of the PRS and is synchronized with the PRSP progress reports. The PAF was extended to include a greater focus on rural incomes to provide an adequate policy matrix for the PRSC1. 168. Finally, the PRSC1 conditions were determined in a participatory process with the government and PRBS donors. The OED/IEO evaluation team was invited to observe a meeting in April 2003 in which the Bank, PRBS donors and the government negotiated prior actions and triggers. This session demonstrated how the Bank is working in partnership with other donors to highlight joint priorities and a common framework of action for the PRSC. 169. Overall, government officials and donors were of the opinion that the PRSC1 supports the PRSP and were appreciative that the Bank had aligned its instrument with the PAF policy matrix. Some donors were keen to see a larger share of the Bank’s portfolio dedicated to this instrument. But the PRSC1 credit is almost a fifth of new loans committed in the past 3 years, and in terms of disbursements, represents 75 percent of average annual disbursements in the past three years. 170. Although the 2003 CAS had not yet been finalized at the time of writing this case study, the country team has indicated that it will focus on the private sector development pillar of the PRSP and on building government capacity for its implementation, in a framework of gradual transition towards budget support. Most interviewees agree that capacity building will be key in supporting implementation of the PRSP. Bank assistance for private sector development will put greater emphasis on the neglected “income” poverty side of the PRSP. Until the CAS is actually finalized and implemented, it will not be possible to 71 The PRSC program document specifies the delay in DAWASA’s privatization, where buyers were not available, as an example of this.  - 68 - evaluate its alignment with the PRSP. But its timing has been delayed to align with the next round of the PRSP. Alignment of other activities 171. The World Bank Institute has carried out various PRSP related workshops and seminars both at a national and regional level in which government officials, civil society and donors from Tanzania participated. The content of these courses is aligned with the PRSP, and has focused extensively on capacity building in monitoring and evaluation. The impact of these courses is difficult to assess as no feedback on these activities was available from interviewees. One area where support to learning activities was found lacking was the Bank’s role in building capacity amongst the Parliament in order to empower it to oversee government. VI. WORLD BANK-IMF COLLABORATION IN TANZANIA 172. There are three aspects of Bank-Fund collaboration that are especially relevant to the PRSP/PRGF process: the nature of coordination and priority setting in the two institutions’ analytical inputs to help foster a broad-based, country-driven policy debate; the Joint Staff Assessments (JSA) of the PRSP; and developments with regard to the “aggregate” level of conditionality and division of labor between the two institutions. A. Coordination and Priority Setting 173. Most of those interviewed, including IMF and World Bank staff, indicated that day- to-day cooperation was good. A number of joint missions took place (e.g. for monitoring progress on the PRSP), and on issues of mutual interest, including reform of the electricity sector or fiscal issues in local government reform, World Bank-IMF teams had joint meetings with government officials. Several government officials said that an even greater coordination of missions would help reduce the burden on limited local capacity. At the beginning, however, the World Bank and IMF teams raised issues in an uncoordinated manner and government officials expressed frustration over the lack of clarity from the BWIs on what was expected from a PRSP. But, country teams in both institutions eventually reached fairly similar conclusions regarding the pace of reforms and the process of reaching HIPC completion point—i.e., that Tanzanian policymakers had shown a willingness to stay the course and that quicker access to HIPC relief was justified. However, the one-year rule for reaching HIPC completion point after the full PRSP was introduced and Tanzania’s HIPC completion point was delayed up to the end of 2001. This delay was received with a certain amount of irritation and anger within government, who had expected to reach HIPC completion point soon after completion of the PRSP. Although country teams argued against this delay, the completion point was still postponed. In retrospect, confusion over how to interpret the Boards’ wish to maintain flexibility in the timing of Tanzania’s completion point was a major cause of the misunderstanding. 174. Two aspects of Bank-Fund inputs to the broader policy debate are worth noting. First, as already noted above, the IMF was less involved than the World Bank in the  - 69 - “participatory” elements of the debate.72 Second, in spite of the prominence given to PSIA in policy documents related to the PRSP, none has been undertaken thus far in Tanzania. The long lags involved in producing such assessments appear to reflect their technical complexity and high information requirements that demand considerable preparation. As noted earlier, the World Bank was not able to undertake the analysis within its 2003 work program and the issue was postponed to 2004. In contrast, the IMF appears to have regarded such assessments as essentially a World Bank responsibility. 175. The experience with both these aspects suggests that a greater integration into the PRSP process of what the BWIs themselves should deliver is needed. For example, a clearer delineation, as part of the PRSP framework, of what analytical inputs should be delivered, when, and by which institution would be consistent with the “results based” emphasis of the PRSP approach. It would also give the authorities a greater input into what the priorities should be. B. Joint Staff Assessments 176. From early 2000 to September 2003, the World Bank and the IMF have written a total of four joint staff assessments (JSAs) on Tanzania, including on an interim-PRSP, full PRSP and two PRSP progress reports—partly indicating the fact that the country was one of the first to complete a full PRSP as well as to reach HIPC completion point. The period covered is also long enough to enable an intermediate evaluation of the impact of the JSAs on the PRSP process. In looking at Tanzania’s JSAs, four aspects will be focused on: the clarity and candidness with which the poverty reduction strategy and related policies are described; the extent to which the risks to successful implementation of the strategy are discussed; the impact of JSAs on progress reports, and thereby the PRSP process, and the reaction of other stakeholders to the JSAs as a pronouncement on domestic policies. 177. The JSAs are generally accurate in their presentation of the PRSP process in Tanzania. Both the JSA for the interim-PRSP and that for the full PRSP acknowledge the importance of previous national processes in paving the way to a rapid PRSP process—while also acknowledging that the rush to reach HIPC completion point was also an important factor. The JSA for the full PRSP provides a good discussion of the consultation process, although following the JSA guidelines, it does not formally assess the participatory element.73 However, while acknowledging the shortcomings of the PRSP and the progress 72 Some respondents argued that this reflected the asymmetry of presence in the field, with the World Bank having a fully-fledged country office that gave it daily access to policymakers and other stakeholders, while the IMF had a small office, with missions from headquarters forming the primary basis of its interaction. 73 The Boards of the World Bank and the IMF have made clear that any such assessment of participation in the JSAs should be left to the Boards themselves, to avoid any suggestion (continued)  - 70 - reports, often in concluding remarks, JSAs are fairly uncritical. In contrast internal IMF reviews were much more critical of the quality of the first PRSP progress report, which failed to respond to the many queries raised by the JSA on the full PRSP—including the need to present a poverty profile for the country on the basis of a household budget survey, better costing of PRSP programs and linkages of the macroeconomic policies to poverty reduction. On the other hand, the JSA is candid in noting that in some areas, such as health and education, resource constraints will make it difficult to arrive at the medium-term goals set in the PRSP. 178. All JSAs discuss in some detail the nature of the risks to the successful implementation of the poverty reduction strategy. The JSA for the full PRSP acknowledged structural obstacles including imprecise data and the fact that though the need was greatest in the rural areas, policies, owing to weak institutional capacity, were untested there and the level of preparation was inadequate at all levels. The economic risks discussed include external shocks and low revenues, which put priority spending at risk. Though assuming continued donor inflows, they also indicated that these were not beyond reversal. 179. The PRSP progress reports have helped to fill the gaps remaining in the poverty reduction strategy. The quality of the progress reports has improved over the past two years, and many of the gaps left over from a hastily drafted initial PRSP have been closed. The second progress report includes for the first time a good macroeconomic analysis, including a linking of macro policies with growth and poverty reduction, first requested in the JSA for the full PRSP, two years earlier. It also takes up issues on poverty monitoring and compares social economic and poverty-related outcomes for 2002 with PRSP targets. 180. During an era when partnership, transparency and ownership are seen to be important aspects of reform, a document such as the JSA, jointly and exclusively drafted by the World Bank and the IMF, but pronouncing on processes and outcomes in the poverty area in which a number of stakeholders have keen interest was bound to generate controversy. However, it is mainly among international NGOs and bilateral donors in Tanzania, some of the latter being members of the PRBS group, where concerns over the JSA have been raised. They argue that the “J” must be expanded in order to make it more inclusive of other development partners. In their view, issues of importance to the JSA, such as donor coordination and domestic political feasibility, are much too important to be left to the BWIs alone and that more systematic mechanisms for incorporating broader inputs need to be developed. that the JSA is imposing a ‘political test’. In practice, the criteria by which the participatory process might be judged by the Boards are unclear.  - 71 - C. BWI Structural Conditionality 181. This section assesses changes in Bank-Fund conditionality in light of the PRSP process, by addressing three questions: i) what has happened to the level and focus of aggregate Bank-Fund conditionality; ii) in areas of shared interest, to what extent have there been changes in conditionality and the associated division of labor between the Bank and the IMF; and iii) is Bank and Fund conditionality (post-PRSP) in line with the policy priorities outlined in the PRSP? In analyzing conditions from both institutions, it is important to note that there is no simple method of aggregation, making comparison at specific points in time difficult. The nature, timing, and triggers associated with each institution’s conditions vary by program and year, depending in part on the architecture of the programs being implemented. Thus, this analysis looks at the changes in the level and focus of conditionality (including in areas of mutual interest) over a broad time period. 182. The analysis covers Bank and Fund structural conditions for lending operations from late 1996 to 2003.74 Drawing on the original lending and program documents, conditionality is broken down into 21 categories (see Annex II, Tables 1 and 2). For the IMF, conditions are extracted from annual arrangements under the ESAF and requests and reviews under the PRGF. World Bank conditionality is drawn from lending documents for adjustment operations, both structural and sectoral. For Tanzania, these adjustment operations are the 1997 Structural Adjustment Credit 1 (SAC1), the 2000 Programmatic Structural Adjustment Credit 1 (PSAC1), the 2001 Primary Education Development Program (PEDP) and the 2003 Poverty Reduction Support Credit (PRSC1).75 For purposes of this analysis, the results can be categorized according to the following three broad questions: What has happened to the level and focus of aggregate Bank-Fund conditionality? 183. In aggregate, BWI conditionality in Tanzania has decreased, but not dramatically, since the endorsement of the PRSP at the end of 2000. The Bank’s operations prior to the PRSP, the SAC1 and the PSAC1, had 24 and 28 conditions respectively. The PEDP had marginally less conditionality than the two previous adjustment credits (23 conditions). The PRSC1, with 14 conditions, shows a greater decline in conditionality, although the PRSC1 is 74 The lending period was chosen to reflect a sufficient length of time before and after the PRSP during which Bank and IMF operations were active and—where possible—to incorporate a full ESAF- or PRGF-supported program and Bank-approved adjustment operation. 75 World Bank conditionality is recorded in the fiscal year in which the associated adjustment operation was approved. As such, the government may have actually felt the burden of conditionality associated with a second or higher tranche at a time different from the approval of the credit. For these—and other reasons—simply adding up the number of IMF and World Bank conditions at a particular point in time can be misleading.  - 72 - the first operation in a series of three expected credits. However, the number of conditions per dollar of Bank commitments is significantly less under the PRSC1. The decline in IMF conditionality has been modest and occurred only after the first year of the PRGF arrangement. 184. Two points related to this overall trend can be noted. First, reforms in the areas in which conditionality saw a dramatic decline, notably SOE reform and privatization, were largely complete. For similar reasons, conditionality in pricing and marketing, notably for petroleum products, was no longer required as reforms there had been concluded. Second, the IMF seems to have increased the level of conditionality in 2000 in a bid to “round-off” reforms in anticipation of the completion of the PRSP. The subsequent reduction in the number of IMF conditions tended to return the total to ESAF levels. 185. The distribution of conditions between policy areas has shifted significantly following the introduction of the PRSP. Debt management, both domestic and foreign, has received increased focus after HIPC completion point from both the IMF and the World Bank, with a number of conditions introduced to prevent debt escalation. In its 2003 program, the Bank introduced conditions related to progress in PRSP implementation and monitoring, and others in the agricultural sector. In areas of shared interest, to what extent have there been changes in conditionality and the associated division of labor between the Bank and the Fund? 186. Before the completion of the PRSP, the World Bank and the IMF tended to impose conditions on a number of broadly similar areas. Post-PRSP, the overlap has been reduced. The areas of mutual interest now are fiscal treasury systems, debt management, and institution building, regulation and transparency. 187. In fiscal treasury systems, the World Bank has focused on public financial management—notably the development and implementation of the Public Financial Management Reform Program. The IMF has focused on the role of the IFMS in increasing financial transparency at the sub-national level. In debt management, the Bank is concerned with strategies and regulations for prudential debt management—that is with the institutional environment—while the IMF focuses on macro-critical aspects of the debt problem, including domestic arrears. In the area of institution building, regulation and transparency, both institutions have focused on strengthening the business environment and financial transparency. The Bank and Fund also share an interest in land issues, notably the reform of the Land Act—but with the Bank focusing on implications for agricultural development and the IMF on the impact of land transactions on the financial sector. 188. The Bank has moved out of tax/revenue reform, expenditure reform and the financial sector, and the IMF out of the social sectors and civil service reform. Streamlining, in the case of the IMF, is illustrated by increased focus on fiscal policy, where conditions comprised more than half of the PRGF total. This increase in overall fiscal conditionality,  - 73 - even as IMF conditions were being reduced elsewhere, perhaps explains the feeling at the Ministry of Finance that conditionality had increased. Is Bank and Fund conditionality (at least post-PRSP) in line with the policy priorities outlined in the PRSP and designed to make a meaningful contribution to the major objectives articulated in the PRSP? 189. The Bank’s conditionality post-PRSP is aligned with priorities of the PRSP. The Primary Education Development Program, which accounts for the majority of its conditions post PRSP, has been accorded the highest priority in the PRSP. About a third of the conditions in the PRSC are explicitly tied to capacity building for and progress in monitoring and implementing the PRS. The remaining conditions, including conditions in agriculture and public financial management, are drawn from a program that expands upon the policy matrix of the PRSP. 190. IMF conditionality has mainly been geared towards ensuring the soundness of the domestic financial system and the transparency of public accounts. Although the PRSP was far from specific on the requirements in these areas, domestic debates in Tanzania have indicated the importance of effective fiscal policy, debt management and treasury systems at the sub-national level to the PRSP’s goals of priority expenditure protection and poverty reduction. VII. CONCLUSIONS AND LESSONS The PRSP process 191. One way of summarizing the conclusions about the PRSP process in Tanzania would be to ask what evidence the evaluation suggests about two fundamental questions. First, has the process affected the “way of doing business” in Tanzania, by the government, BWIs and donors especially, as encapsulated in the underlying principles of the PRSP? Second, have these intended changes in process led to changes in actual results? While it is too early to analyze the final impact on the ultimate objectives of growth and poverty reduction, we can ask whether the process led to significant differences in the types of policies implemented, in the direction that give greater emphasis to growth and poverty reduction. In this regard, key messages to emerge from the evaluation would be as follows: • The answer to both sets of questions would be yes, albeit with significant qualifications. There have been important positive changes in the way of doing business in Tanzania in recent years, although some of the most important (e.g. the PER system and the shift toward a more partnership role with donors following the Helleiner Report) predated the PRSP. Policies have been affected in some important areas—specific examples would include the composition of expenditures, the modification of the macroeconomic program to accommodate higher aid flows, and the elimination of fees at the primary school level—but the approach has had little impact so far on other important policy areas (e.g. trade policy).  - 74 - • There are undoubtedly tensions between different aspects of the PRSP process, most notably between the emphasis on a country-driven approach and one where key steps in the process have been specified by the BWIs. While these tensions are perhaps inevitable, the most useful elements of the approach are those which build upon domestic institutions and initiatives. These should be emphasized over “Washington- driven” elements of the process. In this respect, the close link with the externally- driven HIPC debt relief timetable affected the pace and quality of the domestic policy debate. • Tanzania is a good illustration that it is the PRSP process, not the document itself that matters. The initial PRSP was weak in many respects. However, three years after the launch of the PRSP, policy-makers have been able to flesh out the strategy and improve poverty monitoring mechanisms. This also raises the issue of why the national strategies that preceded the PRSP were not used more concretely as a basis for the PRS process. 192. The following discussion, organized around the key principles of the PRSP process, elaborates on these messages. 193. A homegrown country-led comprehensive poverty reduction strategy, set in the context of the long-term Vision 2025, already existed in the form of the National Poverty Eradication Strategy. With respect to partnerships, the Helleiner report had initiated a fruitful dialogue with the country’s external partners and the TAS was under formulation. So what potential did the PRSP have when it was introduced, besides adding to the proliferation of national strategies in Tanzania? 194. The process provided an opportunity to mobilize and focus resources in the priority sectors; enhance the monitoring of poverty reduction targets; and encourage broad-based participation of civil society in national processes. These were areas that were lacking up to that point in national strategy formulation. Has it delivered in these areas? Evidence from three years of implementation suggests that it has indeed added value in some important areas, although challenges remain. 195. Ownership is ultimately judged by commitment to the strategy, which has been high, especially at the top leadership level. However, a thorny issue that remains to be resolved is the need for endorsement by the BWIs. Among civil society and local government, ownership has been mixed. It is a function of involvement in the process, with Dar-based groups more favorable to the process than those outside and the priority ministries more committed to the PRSP than others. An important limitation is that Parliament has been only marginally involved in the process and its ownership is equally low. In order to extend the political legitimacy of the process and ensure sustainability, bringing Parliament more formally into the process is imperative. 196. The process has, with time, become more participatory and inclusive. Civil society participation in this process, while significant, was incomplete and the source of some  - 75 - frustration to many NGOs. This partly reflected different expectations/ambiguity about what participation was meant to achieve—consultation versus participation in policy decisions. The involvement in macro policy issues was more limited—involving mainly the government, donors and BWIs—than other sectors. While partly an issue of technical capacity, it also reflected limited success of the IMF in framing broader policy issues in a manner digestible to a wider audience and in ensuring that all major policy issues were included in that debate. This was linked in turn to the lack of clarity, in the PRSP approach itself, about what precisely should be the IMF staff’s role in such a debate. 197. However, the institutionalization of the policy dialogue, through the PER process, where the World Bank played a critical initiating role, was crucial. It provided a forum for follow up of the policy debate which was very important given initial gaps in the strategy. The PER process today is broadly participatory, government-led and encourages domestic and external partnership. It also facilitates a multi-sectoral results-orientation of the PRSP via links to the budget and assessments of the poverty focus of expenditure. 198. While the PRSP recognizes the multidimensional nature of poverty, it has not been sufficiently comprehensive in its implementation. The conceptualization of vulnerability in the PRS was weak, nor was much consideration given to the relationship between governance and poverty. During implementation, the focus has been on nonincome poverty at the expense of income poverty, hence the emphasis on social sectors and less on investment in productive sectors of the economy to generate sustainable increases in the income opportunities of the poor. On the whole, the PRS has been results-oriented and focused on outcomes that benefit the poor. It has enhanced the mobilization of resources and the focus on priority sectors—with public expenditure reviews as major inputs. During the PRSP process the shares of priority sectors in expenditures have risen steadily, although PER analyses have shown that spending within priority sector needs to be better targeted to the poor. The PRS process has substantively enhanced national processes for poverty monitoring and a poverty monitoring and master plan now guides all monitoring activities. However, the feedback between monitoring of results and policy actions requires improvements. 199. In retrospect, issues of poor domestic capacity were not adequately considered. The success of the PRSP process was premised on effective management and coordination at the sub-national level, particularly in rural areas. However, both central and local government structures lacked adequate technical skills to implement and follow-up key aspects of the process. Lack of capacities led to serious delays, and ultimately to inadequate outcomes. That some progress has been made in spite of capacity constraints is partly thanks to effective networking by key government officials, bringing onboard technical expertise from local consultants and academics engaged in local research-oriented NGOs, and others from abroad. 200. Although the private sector, including small scale enterprises, is mentioned in PRSP documents as the future engine of growth and source of employment, its participation in the process has been limited thus far. Its relative lack of involvement is of concern for the sustainability of the poverty reduction strategy as a whole.  - 76 - IMF inputs into the PRSP process and their impact 201. The IMF’s main input into the PRSP process was the macroeconomic framework which was based on the PRGF. Though initially taking a conservative approach to assuming higher aid flows, and incorporating a rapid increase in private sector credit demand that proved too optimistic, the PRGF became increasingly aligned with the PRSP process— allowing, in the light of sustained donor support, higher spending on priority sectors. Its conditionality was streamlined, while its focus reverted to the IMF’s core areas of competence: monetary and fiscal policy and public expenditure management. 202. The PRSP process has had impacts on internal IMF procedures, with the timing of IMF missions becoming better aligned to the budget cycle of the government. IMF staff now interact in the country with a more diversified group of stakeholders. Still, with a limited presence in the field, the IMF is yet to meet demands for broader interactions with CSO groups and to devote more time to explaining its macroeconomic policies. One illustration of this gap is that, while the fiscal program design did adapt significantly over time—largely in response to government and donors—there was little civil society involvement in this debate, or recognition that anything has changed. The IMF’s technical assistance is focused on the key areas of expenditure management, tax administration and fiscal decentralization, and therefore appears to have been aligned with PRSP objectives. However, government officials are demanding more say in the use and mode of delivery of the IMF’s TA, which they still see as too closely linked to program targets rather than capacity building. It is also of concern that fiscal TA has only had a modest impact on increasing tax revenue. 203. The internal policy formulation and review process of the IMF has also changed during the PRSP/PRGF process. In internal documents, there are more discussions of poverty related issues, as well as PRSP-PRGF links, while there has been a similar improvement in considering the mobilization of external resources for domestic financing of priority expenditures. However, discussion of policy alternatives or how to protect program objectives in the event of negative shocks, is still limited. Moreover, even three years into the implementation of the new approach, PSIA is yet to provide an input in the IMF’s analysis or as a contribution to the broader policy debate. Thus, while there have been changes in internal processes, they have been more incremental than expected. 204. The latter years of the PRGF-supported program indicate that, thanks to a domestic debate on important aspects of the macroeconomic framework, in this case on financial programming and the cash budget, the PRGF-supported program became better aligned to the PRSP. The results, in terms of a more flexible fiscal stance and increased priority expenditures, better insulated from temporary budgetary shocks, were in line with the spirit of the PRSP. However, where a broader debate had not taken place, such as in the area of trade liberalization and tariff reform, progress was slow.  - 77 - World Bank effectiveness 205. The Bank played a facilitative role and provided technical advice to the government during the short formulation period of the PRS, which influenced the sectoral contents of the strategy. Stakeholders were critical of the Bank’s “intrusive” influence in the initial formulation of the strategy which they perceived as overstepping the boundaries of country ownership but acknowledged that this keenness was tempered by the country office. Overall, the principal drafters of the PRSP described the formulation as a government-led process. but government officials did complain about having to adhere to “formatting requirements” for the BWIs. 206. Civil society were critical that, during formulation, the Bank did not exert more influence on the government to increase their participation. However, they were cognizant of the dilemma faced by the Bank in urging the government to expand the consultative process but yet respecting the “country-led” principle. Further, the Bank was reluctant to hold up Completion Point under HIPC to demand a more extensive consultative process. During implementation, the Bank has done its best to promote the participation of civil society within the national PRS processes with encouraging results. Efforts to further strengthen capacity of civil society to participate in the PER process are also underway. 207. An area where the Bank could have been better prepared for the PRSP was in assisting Tanzania to conduct a Household Budget Survey prior to 2000. During implementation of the PRS, however, Bank staff have provided substantial technical advice in analyzing the results of the Household Budget Survey and in establishing the Poverty Monitoring System. The Bank’s technical expertise and sectoral inputs in education and health and via the PER process have supported the 1st and 2nd Annual Progress Reports and have been greatly appreciated. The PER process in particular has been well recognized for its support of the PRSP and its principles, and the Bank has played a leading role in promoting and expanding the PER from an external technical assessment to a country-led participatory process involving a wide range of stakeholders. Until recently, Bank staff provided secretariat services for the bi-weekly PER meetings which have now been moved to the Ministry of Finance Bank staff have been highly committed to and wholly engaged in the PER process and continue to provide high-quality technical expertise. 208. In other areas of the Bank’s analytical support for the PRSP, some major shortcomings are widely recognized. There are continuing knowledge gaps between the micro-macro linkages, specifically the micro-economic social and poverty impacts of macro- economic and structural policies, analysis on the sources of growth and the poverty impact of alternative policy and strategy options. At the sectoral level, the depth and breadth of the Bank’s analytical work has been uneven in supporting PRS priorities. Work in agriculture, for example, should have started much sooner and the Bank should have been more assertive in promoting the growth agenda in the PER/MTEF processes. 209. The Bank’s lending portfolio in the last three years has been well aligned with the PRSP, and the recently introduced PRSC signals a further shift towards alignment with  - 78 - country processes. The PRSC1 is programmatic, aligned with the PRS policy matrix and has much less conditionality than previous policy-based lending instruments. 210. Overall, the Bank’s behavior vis a vis respecting the “country-led” principle has improved significantly. The Bank was already moving in this direction but since the PRSP, it has moved much further. Interviewees described a “sea change” in the Bank’s behavior, and have cited the strong role of the country office in promoting this change. From the “Washington consensus,” national stakeholders feel the consensus is now a “Washington- Dar” consensus, indicating that there is still room for improvement, however. Government officials and NGOs feel that sector teams in particular are more resistant to the principle of country ownership and often fail to adequately consider alternative viewpoints. 211. The Bank has continued to promote the “partnership” principle. The Bank’s proactive role in the participatory PER process; the alignment of the PRSC within the framework of the PRBS, and its engagement with the DAC are important steps that have enhanced the collaboration of donors in Tanzania. There are still complaints about the Bank’s coordination at the sector level which relate to the large numbers of missions that regularly arrive from the Bank. At the same time stakeholders have acknowledged that the complaints at the sectoral level are strongly correlated with capacity problems within ministries, which the Bank’s technical assistance has not been successful in addressing. World Bank-IMF collaboration 212. The process has intensified the interaction of the country teams of the BWIs both at headquarters and in the field. Generally, the PRSP process has reinforced the need for increased focus on areas in which each institution has a comparative advantage and to collaborate in those areas where a common approach is required. However, there is also need for a uniform approach to the streamlining of conditionality. In resolving the tension, in the PRSP approach, between ensuring a country driven process and the need for the PRSP to provide operational guidance, the Joint Staff Assessments of the PRSP and the progress reports have been important inputs, providing a forward-looking perspective on the process. However, as commentaries on a multi-stakeholder process, questions are bound to be raised on their limited inclusiveness. Clearer “rules of the game” on the BWIs role in the process— along with a more results-based emphasis on what the institutions are expected to deliver as analytical inputs into the ex ante policy debate are needed.  - 79 - ANNEX I SURVEY RESULTS ANNEX Tanzania 1. As part of the OED and IEO evaluations of the PRSP Process and the PRGF, a survey of PRSP stakeholders was administered in each of the ten countries where a case study was undertaken. The objective of the survey was to obtain perceptions of the PRSP process and the role of the World Bank and IMF in supporting the initiative. 2. A standard survey of 39 questions was administered in each country. The full questionnaire can be found on both of the evaluation websites www.worldbank.org/oed/prsp and http://www.imf.org/external/np/ieo/2002/prsp/index.htm. The survey consists of four main components: information on respondents; the PRSP process (covering ownership, results orientation, comprehensiveness, partnership-orientation and long term perspective); World Bank performance; and the role of the IMF. In most cases, respondents were asked to indicate the extent of their agreement with statements on a five point scale76. The survey was translated, into local languages, where necessary, and pre-tested. A local consultant with survey experience was engaged in each country to assist with administration of the survey. Survey results were coded by the local consultant and sent back to Washington and an outside contractor, Fusion Analytics, was hired to analyze the data. 3. The survey was targeted at key groups within the three main categories of PRSP stakeholders: Government, Civil Society, and International Partners.77 Within each group, the survey sought to obtain an institutional view and was targeted at the most knowledgeable individuals. Respondents were asked to define the nature of their involvement in the PRSP process, and their level of familiarity with the PRSP document, the Bank, and the IMF. Given the targeted nature of the survey, respondents who were “Not Aware” of the PRSP process were excluded from the results. The specific samples were selected using three main inputs: information gained through the country case study mission; participants listed in the PRSP document; and input from the local consultant. In some cases, samples were circulated to obtain broader input on their composition. The study teams also identified a set of highly relevant respondents in each country for whom a survey response was required. These included core ministries and agencies (Finance, Economy, Central Bank…), key PRSP-related ministries (Health, Education, Agriculture…), and major donors. Survey 76 The five point scales used in most questions offered a range from 1: Completely Disagree to 5: Completely Agree. Respondents could also mark 0 for Don’t Know or Unsure. 77 Fourteen stakeholder groups were identified: Government—central government, line ministries and sector agencies, local government, Parliament—Civil Society—local NGOs, business sector, labor unions, academia, media, religious organization, political party, other—International Partner—donor, international NGO. Results at the stakeholder group level will be presented in the aggregate analysis across all countries.  - 80 - ANNEX I questionnaires were tracked in order to ensure responses were obtained from key groups, however, individual respondents could choose to remain anonymous. 4. The following section presents findings from the survey applied in Tanzania. Section A provides an overview of the survey respondents, including the nature of involvement and familiarity with the process. Section B provides an aggregated snapshot of stakeholder perceptions of the PRSP Process across each of five main sub-categories. Section C provides the mean results for all questions concerning the role and effectiveness of Bank and Fund support. Section D presents results for questions with the most positive and negative responses and questions where there was the greatest consensus or disagreement on issues.  A. Respondent Information 1. Composition of respondents (n = 100) 2. Nature of involvement (%) Involved in Implementation / Monitoring Only 36% Civil Society Involved in both Strategy and Implementation / 0% 49% Monitoring Governm ent 30% 4% Direct Contribution to Strategy Consulted During Strategy Only 7% International Donor NGO 8% Not Involved but Aware 51% 13% - 81 - Not Aware 2% 3. Level of Familiarity PRSP document 46% Bank 40% IMF 32% ANNEX I  B. THE PRSP PROCESS Relevance 3.27 Partnership- 3.05 Oriented Country-Driven 2.93 Results-Oriented 2.93 Comprehensive / 2.70 Long-Term - 82 - Based on a five point scale, where 1: Completely Disagree to 5: Completely Agree. Respondents could also mark 0 for Don’t Know or Unsure. C. BANK AND IMF World Bank IMF Q34: World Bank assistance Q38: Gov't budgets linked to 3.16 supports PRSP priorities the PRGF are more pro-poor 2.94 and pro-growth than before Q36: World Bank promoting coordination of donor 3.08 assistance Q39: Design of the PRGF Q33: World Bank strategy is 2.97 aligned with PRSP program indicates more 2.92 flexibility Q32: World Bank involvement has been very 2.82 helpful ANNEX I Q37: IMF involvement has Q35: World Bank activities 2.69 2.81 been very helpful provide relevant inputs Based on a five point scale, where 1: Completely Disagree to 5: Completely Agree. Respondents could also mark 0 for Don’t Know or Unsure.  D. Composite Results Table Means with Means with highest lowest Percent Percent Percent standard standard don’t know Question agree disagree deviations deviations or unsure Most positive responses Q27 Partnership-oriented: Donor coordination improved 53.0 Q10 Relevance: PRSP improves on past modalities 46.0 Q9 Relevance: PRSP adds value 45.0 Q16 Results-oriented: Outcomes benefit poor 44.0 Q15 Country-driven: Government continues to engage stakeholders 43.0 Most negative responses Q21 Comprehensive: Alternatives fully explored 45.0 Q22 Comprehensive: Macroeconomic framework participatory 44.0 Q23 Comprehensive: Balance between growth and poverty reduction 41.0 Q13 Country-driven: Your stakeholders were consulted 40.0 Q17 Results-oriented: Realistic targets and plans 38.0 - 83 - Most polarized responses Q29 Partnership-oriented: Coordination b/t World Bank and IMF improved 2.94 Q30 Partnership-oriented: Quality of Bank/Fund collaboration 2.91 Q13 Country-driven: Your stakeholders were consulted 2.90 Q15 Country-driven: Government continues to engage stakeholders 3.19 Q14 Country-driven: Final document was modified to accommodate viewpoints 2.92 Areas of Greatest Consensus Q28 Partnership-oriented: Current donor coordination 3.08 Q10 Relevance: PRSP improves on past modalities 3.38 Q17 Results-oriented: Realistic targets and plans 2.81 Q23 Comprehensive: Balance between growth and poverty reduction 2.54 Q8 Relevance: PRSP is a good model 3.37 Most unfamiliar areas Q29 Partnership-oriented: Coordination b/t World Bank and IMF improved 34.0 Q30 Partnership-oriented: Quality of Bank/Fund collaboration 33.0 Q19 Results-oriented: Results feed back 16.0 Q25 Partnership-oriented: Donors supported formulation 16.0 Q18 Results-oriented: Structure to monitor results 13.0 Note: The percentage of those who “agree” or “disagree” relates to the total that responded to each question. The percentage of “don’t know or unsure” is calculated on the basis of the respondents in the sample (100). To determine areas of low level of agreement among the sample (polarization), and high level of agreement (consensus), the ANNEX I means for all the questions were sorted by their standard deviation. The five highest (most polarized) and lowest (greatest consensus) ranking means are given in the table.  Table 1. IMF Structural Conditionality in Tanzania: ESAF (1996–99) and PRGF (2000–03)76 ESAF 1996–1999 PRGF 2000–2003 First Year Second (including first Second and Fourth and First annual annual Third annual review) third reviews fifth reviews arrangement arrangement arrangement Total Mar. 2000; Feb. 2001; Mar. 2000; Oct. 1996 Nov. 1997 Jan. 1999 ESAF July, 2000 Sept. 2001 Nov. 2002 Total PRGF Exchange System Central Bank Reform Trade Regime 1 1 1 3 Capital Account Pricing and Marketing 1 2 3 Public Enterprises: Reform/Restructuring/ Privatization 2 5 5 12 6 6 Fiscal—General Government Tax/Revenue Reform 6 4 5 15 5 3 5 13 Fiscal—General Government Expenditure Reform 1 2 3 2 1 2 5 Fiscal – Treasury Systems 1 1 1 4 5 - 84 - Fiscal – Debt Management 2 5 7 Fiscal Other 3 3 Civil Service Reform 1 1 Social Sector/Social Security System/ Social Safety Net 2 2 Financial Sector 2 2 2 6 3 1 2 6 Agricultural Sector Labour Market Economic Statistics 3 2 2 7 1 2 3 Corporate Restructuring and Governance Institution Building, Legal and Regulatory Framework, Transparency 1 1 1 2 2 5 PRSP Development and Implementation Other Total 15 15 19 49 25 16 17 58 1/ Note that the PRSP was approved by the IMF and World Bank Boards in December 2000. During the second and third annual arrangements of the ESAF, three conditions (not met earlier) were re-introduced, but with slightly different wording. The condition related to the privatization of the Tanzania Telecommunications Company mentioned in the second annual arrangement (and not met) was repeated in the third annual arrangement. Under the PRGF, conditions were not repeated in subsequent program years (see Table 5 in the text for sources). ANNEX II  Table 2: World Bank Adjustment Lending Conditionality in Tanzania (1997–2003) June 1997 June 2000 October 2001 May 2003 Structural Adjustment Programmatic Primary Education Poverty Reduction Credit I (SAC1) Structural Adjustment Development Program Support Credit 1 Credit I (PSAC1) (PEDP) (PRSC1) Exchange System Central Bank Reform Trade Regime Capital Account Pricing and Marketing 1 1 Public Enterprises: Reform/Restructuring/ Privatization 8 8 Fiscal—General Government Tax/Revenue Reform 1 4 Fisca—General Government Expenditure Reform 3 1 Fiscal—Treasury Systems 1 2 Fiscal—Debt Management 2 Fiscal Other 1 Civil Service Reform 1 1 - 85 - Social Sector/ Social Security System/ Social Safety Net 7 Financial Sector 3 2 23 Agricultural Sector 2 Labour Market Economic Statistics Corporate Restructuring and Governance Institution Building, Legal and Regulatory Framework, Transparency 10 4 PRSP Development and Implementation 3 Other Total 24 28 23 14 NOTE: a) The PRSP, approved by the Boards of the World Bank and IMF in December 2000, is represented by a bold line in the table. b) Data excludes one waiver that was granted for the 1997 Structural Adjustment Loan. c) Conditions for PRSC1 include prior actions for board effectiveness and triggers for PRSC2. ANNEX II  Table 3. An Assessment of the Adaptation of the IMF’s Internal Policy Process Based on Briefing Papers and Review Department Comments Associated with Tanzania’s ESAF (1996–99) and PRGF (2000–03) Assessment criteria Summary of change in policy approach as indicated by ESAF—and Ranking of briefs and comments linked to PRGF-related briefs/review department comments programs on a scale of 1-4 ESAF, third ESAF/PRGF PRGF, fourth annual (12/1999) review, arrangement 12/2001 (8/1998) A. Briefing Papers 1. Does the brief discuss links The December 2001 brief (for the fourth review of the PRGF) made N.A. 2 2 between the PRSP and the PRGF frequent reference to the poverty reduction strategy. Although the objectives? poverty discussion itself is sparse, it is an improvement on the ESAF. 2. In resolving the key issues Overall, the change from the ESAF was not drastic and the result must identified by the mission does the be deemed unsatisfactory, albeit with some evidence of continued brief leave room for discussions of improvement over time. The PRGF brief in early 2000 was fairly alternative policy options to achieve prescriptive in tone, while the specific actions required in fiscal and - 86 - major objectives? monetary policy were reduced in the December 2001 brief. Cash budget -- 3 -- Capital account liberalization 2 -- -- Civil service reform 2 2 -- Credit policy 2 -- -- Debt and contingent liabilities -- -- 2 Fiscal federalism -- 2 -- Financial sector reform -- -- 1 Governance -- 3 -- Privatization 1 2 -- Public expenditure 3 2 2 Revenue performance 2 2 2 Stabilization and growth -- 2 2 Trade liberalization -- 3 -- ANNEX II  Table 3. An Assessment of the Adaptation of the IMF’s Internal Policy Process Based on Briefing Papers and Review Department Comments Associated with Tanzania’s ESAF (1996–99) and PRGF (2000–03) (continued) Assessment criteria Summary of change in policy approach as indicated by ESAF- and Ranking of briefs and comments linked to PRGF-related briefs/review department comments programs on a scale of 1-4 ESAF, third ESAF/PRGF PRGF, fourth annual (12/1999) review, arrangement 12/2001 (8/1998) 3. Are alternative macroeconomic Proposals to intervene in agricultural marketing were discouraged, but 1 2 2 frameworks and the tradeoffs the Bank of Tanzania was allowed to pursue more flexible monetary between them considered? policies. 4. Does the program discuss how it As an indicator of change, PRGF briefs discussed contingent liability 1 1 2 will protect key objectives in the issues, but in a fairly limited manner and not linked to poverty issues. event of unanticipated negative However, in the face of a 30 percent expenditure reduction in 2001, the shocks? December 2001 brief did not discuss how to protect government expenditure from a similar shock in the future, noting only that priority - 87 - expenditures were sufficiently protected. 5. Does the brief allow for flexibility There was evidence of improvement over time—the ESAF brief noted 2 2 3 to use additional concessional the need to switch aid from debt repayments to priority expenditure, external financing if available? which was picked up by the PRGF-related briefs, with the December 2001 one expressing concern over the contractionary impact of the program. It planned discussions with the government and donors, with a key suggestion of relaxing the fiscal stance by increasing donor financing. 6. Is a participatory process in The public expenditure review process launched under the ESAF had 3 2 3 resolving key issues considered? some influence on targets. The PRGF briefs continued to underline the importance of a participatory approach with respect to public expenditure issues. 7. Does the brief identify policy There is very little discussion of the need to evaluate the poverty and N.A. N.A. 2 issues where poverty and social social impacts of policies. Since PSIAs were highlighted in IMF impact assessment (PSIA) inputs statements as key to the whole PRGF/PRSP process, lack of would be necessary/useful? discussion/guidance in the briefs points to a potential problem. ANNEX II  Table 3. An Assessment of the Adaptation of the IMF’s Internal Policy Process Based on Briefing Papers and Review Department Comments Associated with Tanzania’s ESAF (1996–99) and PRGF (2000–03) (concluded) Assessment criteria Summary of change in policy approach as indicated by ESAF- and Ranking of briefs and comments linked to PRGF-related briefs/review department comments programs on a scale of 1-4 ESAF, third ESAF/PRGF, PRGF, annual (12/1999) fourth arrangement review, (8/1998) 12/2001 B. Review departments’ comments 8. Is the need for more “policy The comments on the PRGF briefs indicate that review departments were 1 2 3 space” for home grown options more willing than before to defer to the mission’s superior knowledge of recognized in the review process? conditions in the field (for example regarding the appropriateness of reducing tariffs in an election year). However, there are no concrete examples of the government being encouraged to pursue alternative policies of its own. 9. Are poverty issues discussed or There is evidence of increased discussion of poverty issues in comments 1 3 2 - 88 - their absence highlighted in review on the PRGF briefs, including the need for a more relaxed fiscal stance in departments’ comments? order to enable the government to meet its priority expenditure targets. As the PRGF-supported program neared maturity, however, comments on poverty issues receded markedly. Presumably, departments felt that poverty issues had by then become sufficiently internalized. 10. Did review departments press Under the ESAF-supported program, review departments demanded, as a 1 3 4 for more or less conditionality rule, more specificity in both language and targets. In contrast, comments (prior actions, PCs, structural on PRGF-related briefs indicated a willingness to reduce conditionality. benchmarks) in the policy areas In review comments on the December 2001 brief, there was very little specified by the mission? discussion of conditionality, indicating movement towards reduced policy prescription. Source: Internal (briefing papers) IMF documents. 1/ The ranking scheme is based on the degree of consistency with the PRSP/PRGF approach : 1=Highly inconsistent ; 2=Inconsistent ; 3=Consistent ; 4=Highly consistent. ANNEX II  Table 4. Key Features of PRGF-Supported Programs with Respect to Tanzania’s PRGF, 2000–03 Key Feature Comments on specific items from Tanzania’s Overall remarks PRGF program and related staff reports 1. Broad participation and greater a. First annual program of the PRGF was • Participation in the formulation of ownership a. Do main elements of PRGF based on I-PRSP. Subsequent annual the macroeconomic framework was draw from country’s PRSP? programs based on full PRSP. limited until the PER process was b. Was the PRSP produced in transparent b. Yes (but see remarks) expanded. Civil society participation process with broad participation? c. Yes limited by capacity constraints. c. Was PRSP produced by country d. Yes: For example, the Bank of Tanzania authorities? was allowed to expand money supply, even d. Do JSAs/staff reports highlight flexibility in absence of concrete evidence on money in accepting country choices? demand behavior 2. Embedding the PRGF in the overall a. Expenditures have been prioritized, and • The integrated financial strategy for growth and poverty reduction emphasis on accountability raised. The management system (IFMS) was - 89 - a. How have macroeconomic and other financial programming framework was rolled out to the whole country. Low policies been influenced by growth and modified. technical capacity has reduced poverty objectives? b. Tax reforms, reduction of arrears in public effectiveness. b. Highlight aspects of the PRGF program expenditure, microfinance, legislation to that promote private sector development enable use of land as collateral, anti- c. Does the PRGF’s contribution to the corruption and establishment of a national strategy focus on areas within the IMF’s area debt management strategy. of expertise and responsibility c. Yes: public expenditure management, financial sector and tax reform. 3. Budgets that are more pro-poor and pro-growth a. Re-orient government spending towards a. Budget more firmly based on the MTEF, • Clear shift of resources towards activities that benefit the poor with prioritization of expenditure social sectors, with priority allocations emphasized. doubling under the PRGF. ANNEX II  Table 4. Key Features of PRGF-Supported Programs with Respect to Tanzania’s PRGF, 2000–03 (continued) Key Feature Comments on specific items from Tanzania’s Overall remarks PRGF program and related staff reports b. Poverty-reducing programs facilitated by b. Improve efficiency and targeting of spending in introduction of identification codes for priority • Less clear evidence of resource key sectors relevant to growth and poverty transfers to agriculture. expenditure as well as the rolling out of the IMFS. reduction c. stress tax reforms that simultaneously improve c. VAT introduced on petrol products and • Tax system not used for income efficiency and equity surcharges on imports removed. But at the same distribution goals. d. Improve data and monitoring to track time, VAT exemptions and other import expenditures surcharges introduced. • Notable emphasis on data quality d. A new statistics law was submitted to improvements. Household budget Parliament. survey done in 2001 and used to establish a poverty profile for Tanzania. 4. Ensuring appropriate flexibility in fiscal targets a. Yes, undertaken in the PER/MTEF exercises • The PER/MTEF process has been a. Present more normative macro-projections to and in the PAF of the PRBS group. crucial in establishing a credible - 90 - signal financing needs b. Done within PER/MTEF processes. Program b. Where warranted, seek commitments of higher modified in second year to accommodate higher structure for projecting needs and aid flows and build into the program resource inflows. negotiating with the donor c. Use PRSP to identify contingent expenditures c. Not explicitly undertaken community. that could be added if more aid were forthcoming d. Not explicitly done • Some flexibility achieved with the d. Indicate how fiscal targets would be modified in the event of key shocks abandonment of the net domestic repayments. But so far the government has eschewed domestic financing for fear of the resumption of inflationary pressures. 5. More selective structural conditionality a. Conditionality has been streamlined and areas a. Limit conditionality to key measures, central to on which imposed reduced. the success of the strategy. b. Conditionality was confined to core IMF areas b. Confine IMF conditionality to measures in the of responsibility. In third annual program, IMF’s domain; exceptions must be justified conditionality was extended to public debt management and trade policy (both with ANNEX II important implications for fiscal policy).  Table 4. Key Features of PRGF-Supported Programs with Respect to Tanzania’s PRGF, 2000–03 (concluded) Key Feature Comments on specific items from Tanzania’s Overall remarks PRGF program and related staff reports 6. Emphasis on measures to improve public a. The public expenditure review process has • The cash budget is still the most resource management/accountability allowed some of this to be done, although civil effective instrument for budget control, a. Fiscal policies and objectives should be open to society participation was greater on expenditure but with high efficiency and equity public debate composition issues than on the broader fiscal implications. strategy. b. Develop transparent monitoring systems to b. An Assessment and Action Plan for tracking • Pro-poor expenditure tracking has improve efficient delivery of public services poverty expenditure was introduced. An helped reduce resource leakage in the integrated financial management system (IFMS) countryside. c. For HIPCs, include specific mechanisms for is meant to bring transparency to public sector monitoring use of debt relief management. The Public Finance Act and Public d. Consider selective conditionality on fiscal Procurement Act were enacted in 2001. governance measures c. PER process has been instrumental in this regard and an expenditure tracking system was instituted. d. Yes, “financial transparency and - 91 - good governance” was a key conditionality 7. Social impact analysis of major a. Not done • PSIAs have not been undertaken so far, macroeconomic adjustments and structural b. Not explicit although IMF departments have indicated reforms c. Not done. that they should have preceded the a. Demonstrate that distributional effects of introduction of levies on water supply, substantial macro-adjustments or structural hike of electricity tariffs and increases in reforms have been considered petroleum prices. b. Highlight countervailing measures to offset temporary adverse effects on the poor c. Bank should lead if technical impact analysis is needed, but PRGF documents should indicate what work was done and how it influenced policies Source: See IMF documents on Tanzania: IMF 2002a, b and c; IMF 2001a and b; IMF 2000a and b (issues under “key features” are exactly as outlined in this document). ANNEX II  - 92 - ANNEX III List of Interviewees Government President’s Office Mr. O.M.S. Bendera, Deputy Permanent Secretary, Planning and Privatization Mr. Arthur Mwakapugi, Director, Macro Economy Division, Planning and Privatization Mr. Sadikiel Kimaro, Personal Assistant to the President Vice President’s Office Mr. Raphael O. S. Mollel, Permanent Secretary Dr. Servacius B. Likwelile, Director, Poverty Eradication Division Mrs. Anna Mwasha, Poverty Reduction Portfolio Mr. Pascal B. Assey, Poverty Eradication Division Mr. Servus A. Sagday, Poverty Eradication Division Mr. Amon Manyama, Poverty Eradication Division Mr. Latson Msongole, Poverty Eradication Division Ministry of Finance Mr. Peniel M. Lyimo, Permanent Secretary Mr. Ramadhan Khijjah, Deputy Permanent Secretary Mrs. Joyce Mapunjo, Acting External Commissioner, External Finance Mr. Jerome J.R. Buretta, Senior Management Officer Bank of Tanzania Mr. Daudi Ballali, Governor Mr. Peter Noni, Director, Economic Policy Mr. Alemu Aberra, Advisor, Bank of Tanzania Ms Kessi Sia Mbatia, Director, Financial Markets Dr. Joseph Leina Masawe, Deputy Director, Monetary Affairs Mr. N. K. Mboje, Principal Economist Ministry of Justice and Constitutional Affairs Mr. Johnson Mwanyika, Permanent Secretary Mrs. Caroline Mdundo, Planning Section Ministry of Water and Livestock Mr. Charles Nyamrunda, Deputy Permanent Secretary Mr. A. Nyenza, Director of Planning Ministry of Lands and Human Settlement Mr. E.E. Mshanga, Director of Planning Mr. A. Msangi, Commissioner Prof. A. Hayuma, Assistant to Permanent Secretary  - 93 - ANNEX III Ministry of Education and Culture Ms. M. Malale, Permanent Secretary Mr. R. Mpama, Acting Commissioner for Education Mr. A. G. Mwakalinga, Director of Planning Ministry of Health Dr. E. W. Manumbu, Director of Planning Dr. Faustin Njau, Project Coordinator – Health Sector Reform Ministry of Agriculture Ms. Janet Simkanga, Senior Economist Ministry of Energy and Minerals Mr. Gray Mwakalukwa, Commissioner for Minerals Ministry of Transport and Communications Mr. Salim Msoma, Permanent Secretary Mr. A.N. Temba, Policy and Planning World Bank Ms. Judy O’Connor, Country Director Mr. Jim Adams, Former Country Director Ms. Melanie Marlett, Country Program Coordinator Ms. Sumana Dhar, Economist Mr. Benno J. Ndulu, Former Lead Specialist and Acting Country Manager Mr. Benson Ateng, Senior Operations Officer Mr. Robert Utz, Senior Economist Mr. Phillip Mpango, Economist Mr. Jim Smith, PRSP Task Manager Ms. Miria Pigato, Former PRSP Task Manager Ms. Jeni Klugman, Lead Advisor Ms. Denyse Morin, Senior Public Sector Specialist Mr. Ladisy Chengula, Rural Development Specialist Mr. Hamisi Munyimvua, Consultant IMF Mr. Ali I. Abdi, Senior Resident Representative Mr. Tsidi Tsikata, Former Resident Representative Jurgen Reitmaier, Assistant Director, Former Mission Chief David Andrews, Former Mission Chief Mumtaz Hussain, Economist, Policy and Review Department Harry Snoek, Former Senior Economist, African Department Treichel Volker, Senior Economist, African Department Meshack Tunee Tjirongo, Senior Economist, African Department Andrew John Tiffin, Economist, Economist Program  - 94 - ANNEX III Donor Community Mr. Mamoru Endo, Economic Cooperation Adviser, Embassy of Japan Mr. Okuyama Takashi, Advisor on Government Finance, Japan International Cooperation Agency Mr. Jackson, M. Biswaro, Chief Programme Officer, Japan International Cooperation Agency Mr. Bertil Oden, Counsellor, Embassy of Sweden Ms. Tone Tinnes, NORAD Ms. Lisbeth Jespersen, Royal Danish Embassy Ms. Caroline Sergeant, Head of DFID, East Africa Ms. Fiona Shera, Economic Adviser, DFID East Africa Mr. Axel Pougin de la Maisonneuve, First Secretary, Delegation of the EU Mr. Detlef Mey, Director, GTZ Mr. Wolfang Ahrens, GTZ Mr. Onesmo N. Shuma, Project Management Specialist – Private Sector, USAID Mr. John Hendra, Resident Coordinator, United Nations & Resident Representative, UNDP Ms. Valerie Leach, UNICEF Mr. Isiye Ndombi, UNICEF Mr. Harold Randall, UNICEF - Zanzibar Mr. Teferi Seyoum, Resident Representative, UNFPA Mr. Thor Oftedal, Deputy Representative, UNFPA Ms. Miia Hanninen, UNAIDS Mr. Anthony Rutabanzibwa, Program Officer, ILO Mr. David Howlett, Poverty Environment Adviser, UNDP Mr. Pim Van Der Male, Programme Specialist, UNDP Business Community Mr. Elvis A. Musiba, President, Tanzania Chamber of Commerce, Industry and Agriculture Mr. Julius M.M. Matiko, Vice President – Agriculture, Tanzania Chamber of Commerce, Industry and Agriculture Mr. Aloys J. Mwamanga, Vice President – Industry, Tanzania Chamber of Commerce, Industry and Agriculture Mr. Dunstan G. Mrutu, Executive Secretary, Tanzania National Business Council Trade Unions Mr. Nestory Kilala Ngulla, Secretary General, Trade Unions’ Congress of Tanzania Mr. Hassan M. Raha, Deputy Secretary General, Trade Unions’ Congress of Tanzania Mr. John Gonza, Director of Economics and Research, Trade Unions’ Congress of Tanzania Mr. Ezekiah T. Oluoch, Deputy General Secretary, Teacher’s Trade Union Nongovernmental Organizations Prof. Haidari K.R. Amani, Executive Director, Economic and Social Research Foundation Prof. S. Wangwe, Senior Fellow, Economic and Social Research Foundation Ms. Betty Masanja, Workshop Organizer, Economic and Social Research Foundation Prof. Joseph Semboja, Executive Director, Research on Poverty Alleviation (REPOA)  - 95 - ANNEX III Mr. Rakesh R. Rajani, Executive Director, Hakielimu Ms. Josephine Marealle-Ulimwengu, Deputy Program Director, CARE International Mr. Geoffrey Chege Gathuka, Country Director, CARE International Ms. Mary Mwingira, TANGO Prof. Marjorie Mbilinyi, Member, Tanzania NGO Forum Rev. Christopher Mwakasege, Executive Director, TASOET, Arusha Mr. Dominick de Waal, Water Aid Dr. Frederick Kigadye, Executive Director, Christian Social Service Commission Mr. Basili Liheta, Member, GENDO, Dodoma Mr. Joel Joram, Chief of Adminstration and Finance, INADES Formation Mr. Alphonce Katunzi, Managing Director, INADES Formation Mr. Kaganzi Rutachwamagyo, Tanzania Federation of Disabled Peoples’ Organizations Mrs. Eva Biswaro, Treasurer, Tanzania Association for Mentally Handicapped Mr. Alphonce Hume, National Vice Chairman, Tanzania Association for Mentally Handicapped & Director, Amnesty International University of Dar es Salaam Prof. Rwekaza S. Mukandala, Professor of Political Science and Public Administration, Dean Faculty of Arts and Social Sciences Dr. Longinus Rutasitara, Head, Department of Economics National Parliament, Dodoma Chairpersons of Committees: Finance and Economic Affairs, Hon. Njelu E.M. Kasaka Public Accounts Committee, Hon. Leonard N. Derefa Local Authorities Accounts Committee, Hon. Mgana I. Msindai  - 96 - References Alonso, T. Rosa Maria (2002), “The Elimination of the Enrollment Fee for Primary Education in Tanzania. A Case Study in the Political Economy of Pro-Poor Policies,” Draft Paper, World Bank Institute (Washington: The World Bank). Alonso, T. Rosa Maria, and Robert J. Utz (2003), “Fiscal Aspects of PRSP Implementation in Tanzania,” Draft Paper, World Bank (Washington: The World Bank). Bevan, David (2000), “Public Expenditure Review: Tanzania 1999/2000 – Fiscal Policy and the Macroeconomic Context,” Department of Economics (Oxford: University of Oxford). ___________(2001), “Tanzania Public Expenditure Review: 2000/2001 – The Fiscal Deficit and the Sustainability of Fiscal Policy,” Department of Economics, University of Oxford. Bigsten, Arne, Deogratias Mutalemwa, Yvonne Tsikata, Samuel Wangwe (2001), “Tanzania,” in Aid and Reform in Africa, ed. by Shantayanan Devarajan, David Dollar and Torny Holmgren (Washington: The World Bank). Bird, Andrew (2002), Assessment of the MTEF in Tanzania: Initial Conclusions and Recommendations,” Unpublished report under the project Design and Implementation Features of MTEFs and Their Links to Poverty Reduction (London: Department for International development/Overseas Development Institute). Collier, Paul (1991), “Aid and Economic Performance in Tanzania,” Transitions in Development: The Role of Aid and Commercial Flows, ed. by Uma Lele and Ijaz Nabi (San Francisco: ICS Press). Danida (2001), “Review of the PRSP Processes in Tanzania,” A Contribution to the International Review of the PRSP Process (Copenhagen: Danish Ministry of Foreign Affairs). Evans, Alison (2001), “Institutionalizing the PRSP approach in Tanzania,” Chapter 9 in PRSP Institutionalization Study: Final Report, Strategic Partnership With Africa (Washington: The World Bank). Evans, Alison and Erasto Ngalwea (2003), “Tanzania,” Development Policy Review, 21(2), p. 271- 287. Government of Tanzania (2004), Poverty Reduction Strategy paper, Third Progress Report 2002/03, Dar es Salaam. ____________________ (2003a), Poverty and Human Development Report 2003, Dar es Salaam. ____________________ (2003b), Poverty Reduction Strategy Paper, Second Progress Report 2001/02, Dar es Salaam.  - 97 - ___________________ (2003c), Public Expenditure Review FY03- Managing Public Expenditures for Poverty Reduction. Report on Fiscal Developments and Public Expenditure Management Issues, Sectoral Reviews, Dar es Salaam. ___________________ (2002), Tanzania Assistance Strategy—A Medium Term Strategy for Promoting Local Ownership and Development Partnerships, Dar es Salaam. ___________________ (2001a), Poverty Reduction Strategy Paper, First Progress Report 2000/01, Dar es Salaam. ___________________ (2001b), The Poverty Monitoring Master Plan, Dar es Salaam. ____________________(2001c), “An Assessment of the Impact of Macro and Sector Policies on Poverty Reduction Initiatives,” Final Report (Dar es Salaam: Economic and Social Research Foundation). ____________________(2001d), Public Expenditure Review FY01, Dar es Salaam. ____________________(2000a), Poverty Reduction Strategy Paper, Dar es Salaam, May. ____________________(2000b), Report of the National Consultative Workshop on the Preparation of the Poverty Reduction Strategy Paper, August, Dar es Salaam. ____________________ (2000c), Interim Poverty Reduction Strategy Paper, Dar es Salaam. Government of Tanzania and the World Bank (2001), Tanzania at the Turn of the Century—From Reforms to Sustained Growth and Poverty Reduction, Dar es Salaam and Washington D.C. Harrison, Graham. (2001), “Post-Conditionality Politics and Administrative Reform: Reflections on the Cases of Uganda and Tanzania,” Development and Change, Vol.32: 657-697. Helleiner, Gerald, (1999), “Tanzania Assistance Strategy: Critical Issues,” prepared for the Tanzania Consultative Group and Related Meetings, Dar es Salaam. Helleiner, Gerald, Tony Killick, Ibrahim, N. Lipumba,Benno Ndulu and Knud Erik Svendsen (1995), Report of the Group of Independent Advisors on Development Cooperation Between Tanzania and its Aid Donors (Copenhagen: Danish Ministry of Foreign Affairs). Holtom, Duncan, (2002), “Tanzania’s Poverty Reduction Strategy Paper—Everyone wants a success story,” Feedback Report for Interviewees, Centre for Development Studies (Swansea: University of Wales). Independent Monitoring Group (IMG) (2002), Enhancing Aid Relationships in Tanzania: Report of the Independent Monitoring Group (Dar es Salaam: Economic and Social Research Foundation).  - 98 - International Monetary Fund (2003), “Aligning the Poverty Reduction and Growth Facility (PRGF) and the Poverty Reduction Strategy Paper (PRSP) Approach—Issues and Options,” SM/03/94, March (Washington: International Monetary Fund). _______________________ (2002a), “Tanzania— Staff Report for the 2002 Article IV Consultation, Fifth Review Under the Three-Year Arrangement Under the Poverty Reduction and Growth Facility,” EBS/02/187, November (Washington: International Monetary Fund). ________________________(2002b), “Tanzania—Fourth Review Under the Three-Year Arrangement Under the Poverty Reduction and Growth Facility,” EBS/02/56, March (Washington: International Monetary Fund). _________________________(2002c), “Actions to Strengthen the Tracking of Poverty-Reducing Public Spending in Heavily Indebted Countries,” SM/02/30, March (Washington: International Monetary Fund). __________________________(2001a), “Tanzania—Staff Report for the 2001 Article IV Consultation and Third Review Under the Three-Year Arrangement Under the Poverty Reduction and Growth Facility,” EBS/01/153, September (Washington: International Monetary Fund). ___________________________(2001b), “Tanzania – Fiscal Decentralization Issues,” Fiscal Affairs Department, May (Washington: International Monetary Fund). ___________________________(2000a), “Tanzania –Request foe a Three Year Arrangement Under the Poverty Reduction and Growth Facility,” EBS/00/44, March (Washington: International Monetary Fund). ____________________________(2000b), “Features of IMF PRGF-Supported Programs,” SM/00//193, August (Washington: International Monetary Fund). Isham, Jonathan, Deepa Narayan, and Lant Pritchett (1995), “Does Participation Improve Performance?” The World Bank Economic Review 9(2): 175-200. KK Consulting Associates (2001), “Tanzania—Poverty Reduction Strategy Paper – An Assessment of the Depth of Understanding of the PRSP in the Government,” Draft Report, Nairobi, Kenya. Mogella, Cosmas (1999), “The State and Civil Society Relations in Tanzania: The Case of the National NGOs Policy,” Civil Society and Governance Programme (Brighton: IDS, University of Sussex). Mukandala, S. Rwekaza (2003), “The Political Economy of the Poverty Reduction Strategy in Tanzania,” Background Paper for OED/IEO Evaluation of the PRSP/PRGF – Tanzania Case Study, Department of Political Science (Dar es Salaam: University of Dar es Salaam).  - 99 - Mwakasege, Christopher (2002), “Making PRSP Work in a Globalized Economy, The Tanzanian Experience of Implementing PRSP for the first two years—An NGO Perspective,” (Arusha: Tanzania Social and Economic Trust). Narayan, Deepa (1997), Voices of the Poor:Poverty and Social Capital in Tanzania (Washington: The World Bank). Narayan, Deepa and Lant Pritchett (1997), “Cents and Sociability: Household Income and Social Capital in Rural Tanzania,” Social Development and Development Research Group, Policy ResearchWorking Paper No. 1796 (Washington: The World Bank). Also, in Journal of Economic Development and Cultural Change (July, 1999). Naschold, Felix and Adrian Fozzard (2002), “How, When and Why Does Poverty Get Budget Priority? Poverty Reduction Strategy and Public Expenditure in Tanzania,” Working Paper 165 (London: Overseas Development Institute). OED (2000), “Tanzania—Country Assistance Evaluation,” Operations Evaluation Office, Report No. 20902-TA (Washington: The World Bank). PRBS Donors and the Government of Tanzania (2002), “Tanzania—Joint PRBS Annual Review,” (with support of Dag Aarnes, consultant), Dar es Salaam. Ronsholt, Frans (2002), “Analysis of Government Expenditure Allocations to Priorities of the Poverty Reduction Strategy in Tanzania 2002,” Draft, September, Copenhagen, Denmark, Ronsholt, Frans with Richard Mushi, Bedason Shallanda and Paschal Assey (2003), “Results Oriented Expenditure Management: Country Study—Tanzania,” Working Paper 204, (London: Overseas Development Institute). McGee, Rosemary, Josh Levene, and Alexandra Hughes (2002), “Assessing Participation in Poverty Reduction Strategy Programs, a Desk-Based Synthesis of Experience in Sub-Saharan Africa,” Research Report 52 (Brighton: IDS, University of Sussex). Rutasitara, Longinus (2002), “Highlights of Public Expenditure Review FY02 (PER02),” Draft Report to Public Expenditure Review Working Group, (Dar es Salaam: Ministry of Finance). Sida (2002), “Part III: Tanzania” in Supporting Ownership, Swedish Development Cooperation with Kenya, Tanzania and Uganda (Stockholm: Swedish International Development Cooperation Agency). Wangwe, Samuel (1997), “Economic reforms and Poverty alleviation in Tanzania,” Discussion Paper No. 16 (Dar es Salaam: Economic and Social Research Foundation). World Bank (2003a), “Tanzania—Poverty Reduction Support Credit and Grant Project,” Program Document (Washington: The World Bank).  - 100 - World Bank and IMF (2003), “Poverty Reduction Strategy Papers—Progress in Implementation and Detailed Analysis of Progress in Implementation,” September (Washington: The World Bank and International Monetary Fund). ___________________(1999a) “Building Poverty Reduction Strategies in Developing Countries,” September (Washington: The World Bank and International Monetary Fund). ___________________ (1999b) “Poverty Reduction Strategy Papers—Operational Issues,” December (Washington: The World Bank and International Monetary Fund).