Report No. 29912-PK Pakistan Devolution in Pakistan Main Report: An Assessment & Recommendations for Action September 1, 2004 Poverty Reduction and Economic Management Sector Unit South Asia Region Document of the World Bank Table of Contents Methodology ................................................................................................................................... Acknowledgments ........................................................................................................................ v11 ..v Acronyms and Abbreviations x Executive Summary ....................................................................................................................... ....................................................................................................... 1 Context ................................................................................. ..................................... This Study ... ........................................................................................... 3 Service-delivery Improvements Are Essential......................................................................................... 3 Why MightDevolutio K e y Findings................................................................................ New Incentives for Nazimeen and Senior Staff A MilestonePassed The Glass i s More t Recommendations................................................................................................................................. .15 Keyelements ofthe Suggested Strategy.......................................................................................................... 15 Strategic Goals andPractical Actions........... 16 Supporting Material .............................................................................................................................. .18 1.Service Delivery andDevolution ............................................................................................ 19 Service-delivery Improvements are Essential ......................................... ...................... 19 H o w Can Devolution Assist in Service-delivery improvements? .......................................................... 21 A Framework .......................................................................................................... Unpacking the Incentives........................................................................................ Assessing the Impact ofDevolution on Service Delivery ....................................... 2. Assessment ................................................................................................................................ 27 K e y Findings .......... New Incentivesfor Nuzimeen and Senior Staff............ Access to Justice Visible Wins When ....................................................................................................... .48 A Challenging Time for Devolution ...................................................................................................... 54 Managing Inequity Between and Within Provinces ........................................................................................ 55 Re-emergence of private goods strategies .............................. 57 The Key Challenges Facing Devolution ................................................................................................ 58 3. Recommendations .................................................................................................................... 60 A Strategy from Here............................................................................................................................. 60 Seize the Opportunity Presentedby the NFC Award ......................... 60 Enforce the Core Principles-and Accommodate Local Differences.............................................................. 63 Transmit National Policy Priorities More Effectively ..................................................................................... 64 Practical Actions toward the 7 Goals. .................................................................... 70 CompletingDevolution Accepted as e .............................................................................................. 74 Demonstrable FederalGovernment Commitment to StrengthenedRole of the Provinces.............................. 76 Demonstrable Provincial Government Commitment to Strengthening Local Government 79 Local GovernmentsChallenged to Demonstratethat they Merit Confidence ................................................. 86 Judicious Selection from a Menuof Conditional Grants.................... 90 Local GovernmentsHeld to Account ........................................................ 95 Further Clarify the Separationof Executive and Judicial Powers ................................................................. 100 11 Appendix 1:Backgroundto ADBIDfIDNVB DSDDataset2 ................................................. 102 Appendix 2: Assessment Frameworks Developedfor the Study ...................................................................................................................................... ........................................... 105 ~~ Glossary 107 References ................................................................................................................................... 108 Tables Table 1:Districts and TMAs Studied....................................................................................................... v Table 2: Progress to Date inPolitical. Administrative. and Fiscal .............................. 2 Table 3: Gender Profile o f Union Councilors by Province ............ ............................ 10 Table 4: Progress inCreating New Incentives for Nazimeen and S ............................ 12 Table 5: Progress inCreating New Incentives for the Service Providers............................................... 13 Table 7: Timetable o f Proposed Key Actions ..................... Table 6: Impact o f Changed Incentives on Sector Service Delivery............................................. ............................................ 17 Table 8: Pakistan's Social Indicators inInternational Persp ............................................ 19 Table 9: Key Indicators inSAP Focus Areas ......................................................................................... 21 Table 10: Key Incentives inthe Devolved Structures ................................. .......................... Table 11:Proportions o f Total Funding Transferred to Local Government la Basis .........24 28 Table 12: Vertical Programs as a Percentage ofthe District ADP ......................................................... 29 Table 13: Ownership ofthe Structural Reforms.......................... ................................................... 36 Table 14: Management Structure after Devolution ............................................................ 38 Table 15: SMC Structure inthe Four Provinces .... ................................................... 39 Table 16: Per Capita Resource Flows (FY 2003i04).............................................................................. 55 Table 17: Cumulative ProvincialRevenues (Normalized) ........................ ............................. 56 Table 18: Potential Increase inDistrict Budgets with New NFC Award ............................................... 63 Table 19: Characteristics and Impact o fVertical Programs and Conditional Grants............................. 66 Table 20: Principles and Better Practices inGrant Design..................................................................... Table 21:Benchmarks inMoving from Vertical InvestmentPrograms to Conditional Grants .............67 67 Table 22: Conditional Grants inMajor Federations.................................. ..................................... 69 Table 23: Actions to Bring Some ReliefNow-and to Prepare for the F ..................................... 71 Table 24: Total Outstanding Debt Against Provinces, July 2003........................................................... 78 Table 25: Devolution-Friendliness o f ADB and World Bank Projects inthe Current Lending Program. ................................................................... 93 Table 26: Possible overnments ........................................................... 97 99 Table 28: State ofthe ProvincialAccounting Records......................................................................... Table 27: Steps Toward Provincialization o f Accounts ......................................................................... 103 Table 29: Data Sources for Tables inDSD2 ....................................................................................... 104 Table 30: Assessing Institutional Changes within Local Government................................................. 105 Table 31: Assessing the Locationof the Employer Function............................................................... 106 Table 32: Assessing "Verticalness" o f Conditional Grants .......... ................................................... 106 Figures Figure 1:Key Incentives inDevolution ................................................................................................... 4 Figure 2: Growth and Female-male Literacy Gap inPakistan and Comparison Sample. 1970-2000 ...20 Figure 3: Growth and Infant Mortality inPakistan and Comparison Sample. 1970-2000 .................... 20 Figure 4: Interpreting Education Data.................................................................................................... 26 Figure 5: Interpreting Health Data ......................................................................................................... 26 Figure 6: Non-salary as a Percentage o f Total Expenditure (districts) ................................................... 31 Figure 7: DMGStaff by Grade........................................................................ ................................. 52 Figure 8: Distribution o f Positions by Grade for FUGand APUG Staff................................................ 53 Figure 9: Provincial ConsolidatedFundsunder New and OldNFC Scenarios ...................................... 61 Figure 10: The Spectrum o f Autonomy inTax Revenues...................................................................... 77 ... 111 Boxes Box 1: The Sets o f Incentives Tracked inthis Report.............................................................................. 5 Box 2: Vertical Programs-Health ........................................................................................................ 35 Box 3: Citizens Using Private Medical Care.......................................................................................... 36 ............................................................. Box 5: Vertical Programs inEducation........... Box 4: School-Based Management .. ........................................... Box 6: Vertical Programs inWater and Sanita ........................................... Box 8: Celebrate Modest Diversity........ Box 7: Provincial Interventions at the District Level............................................................................. ......................................................... ......50 64 Box 9: Incentives for Restructuring inIntergovernmentalTransfers inthe RussianFederation and Uganda .................................................................................................................................... 68 Box 10:Fine-Tuning Devolution inthe PRSP....................................................................................... 70 Box 11: Fostering Debate on Decentralization Box 12: Selling Decentralization inUganda ... Box 13: Technical Capacity Available to Othe Box 14: The PRSP Recognizes the Need to Resolve Jurisdictional Rivalries ....................................... 79 Box 15: "Debt Write-off for Performance" inPunjab........................................................................... 80 Box 16: The PRSP Recognizes the Need for Further Tax Devolution................................................... 83 Box 17: Potential Short Term Gains from AIT Devolution.................................. ........................ 84 Box 18: PotentialRevenueGains for TMAs inTransferring UIPT Collection Responsibility to the Box 19: Local Innovation: Property Tax Enhancement through Self-Assessment inHyderabad..........84 Districts ................................................................................................................................... 85 Box 20: Reducing the Throw-Forward inNWFP ................................................ ............................ 86 Box 21: The Scale o f the District Throw-Forward Problem ........... ................................................ 87 Box 22: Multiagency Planning inWest Bengal and Kerala ................................................................... 89 Box 23: Donor-Supported Transitional Vertical ProgradGrants .......................................................... 91 Box 24: Data Needs for Equalization or Conditional Transfers ............................................................. 94 Box 25: Administrative Devolution inIndonesia and Uganda ..............95 Box 26: Conflicting Lines o f Management............................................................................................ 96 Box 27: Key Features o f a Local Government Service .......................................................................... 97 Box 28: Improving Union-Level Police. Court and Local Government Relations .............................. 100 iv Methodology The Devolved Service Delivery Study (DSD) is the product o f an agreement between the World Bank, ADB, and DfID, inresponse to a request from the Government o f Pakistan that the agencies review progress toward improving service delivery through decentralization. The government has emphasizedthat it i s looking for feasible, practical recommendations that add value to a devolution process that is well advanced. It requested a range o f robust technical analyses that provide a summary o f the emerging lessons and implications for both policy and operational practices. The government made the point strongly that this should not be an assessment o f larger legal or political arrangements, rather that the focus o f the review mustbe on pragmatic actions for service delivery improvements at district (zila) and tehsil level, based on reasonable assumptions about the likely preconditions for service delivery. The authors have drawn comprehensively on the following data inthe preparation o f this three- volume report: Extensivepublished outputsfrom the NRB and a range o f informal outputs including presentations and conference notes. District Studies-new, Table 1: Districts and TMAs Studied largely qualitative, data gathered from six districts (and a selection o f tehsils within them) ontheir experiences to date gathered through the use o f a series o f sector-specific templates systematically reviewing experience to date with sectoral concerns including public expenditure, civil service, politics, water and sanitation, health, education, and access to justice issues (see Table 1).The districts identified by government for t h ~ sstudy were Faisalabad and Bahawalpur (Punjab); Karachi and Khairpur (Sindh); Bannu (NWFP); and Killa Saifullah (Balochstan). Fieldwork was conducted between June and August 2003 by a seven-member team inthese six districts. Detailed socioeconomic data for the districts are available ina separate working paper (Qadir: 2003). The team included Shehryar Sanvar, Sajjid Siddique, Ali Murtaza, Syed Ata ur Rehman, Masroor Ahmad, Nasir Jamy, Rehana H a s h , Fareeha Zafar, Mohammed Saboor and Sanvar Ban. The qualitative data from this fieldwork are cited inthis report as "ADBIDfIDIWB D S D Dataset 1." Intergovernmental Flows of Funds-a model o f fiscal flows underpinned by a comprehensive database o f revenue and expenditure for district governments and tehsils. The model allows: 0 Some analysis o f the overall fiscal architecture and inparticular the combined effect o f the Provincial and National Finance Commission Awards under altemative scenarios-illustrating the impact o f revenue reassignments or o f any changes infunding assumptions; 0 An assessment o fthe impact o f any "lumpiness" inrevenue, assessingthe cash flow problems that the provinces are flagging as the major reason for their slow movement towards the devolution o f civil service pay; and 0 A comparison to be drawnbetweenthe intendedand the actualbehavior o fprovinces in transferring funds and assessing how predictably districts receivedtheir anticipated funds. V Fieldwork and analysis was conducted between June 2003 and January 2004 by Abdul Wajid Usmani, Mirza Irfan Beg and Adeel Sattar. T h e fiscal data from this fieldwork are cited inthis report as "ADBIDflDIWBDSD Dataset 2." Background to the dataset and the methodology used to construct it are provided inthe appendix to this volume. vi Acknowledgments This report was preparedby Jackie Charlton(DfID),Musharraf Cyan (ADB), Zahid Hasnain(World Bank), Nick Manning (World Bank), Douglas Porter (ADB) and Hamid Sharif (ADB).Manuela Ferro, HanidMukhtar, Jeffrey Rinne, Anwar Shah and Salman Zaidi (World Bank) provided additional contributions on public sector institutionalissues. Raja Rehan Arshad and Shahnaz Arshad (World Bank) led the work on TMA-level reforms and on water and sanitation. InaamHaq (World Bank) led the work on health sector reforms, and Mark Poston (DfID) on education sector reforms. Further inputs were provided by Ismaila Ceesay (World Bank) on accounts and audit, by Mohammad Sarwar Khan (ADB) and Raza Ahmad (ADB) on legal andjudicial reforms, by Richard Slater (ADB consultant) on South Asian experiences, and by Louis Crouch and Dana Weist (World Bank) on vertical programs and conditional grants. Ghulam Qadir prepared the socioeconomic background data on the districts. K a i Kaiser provided particular assistance onbehalf o f the World Bank Decentralization Thematic Group inreviewing experiences inconditional grants and in ensuring consistency with the World Development Report framework. Additional helpful comments were provided by Ahmad Ahsan, Ijaz Nabi, and Tara Vishwanath. The study was conceived and planned by a Steering Committee made up o f representatives o f the Government o f Pakistan Finance and Planning departments, the NRB and three bilateral and international lending agencies: the World Bank, Asian Development Bank and the UK Department for Intemational Development (DfID).Additional financial support was provided by the Canadian Intemational DevelopmentAgency (CIDA). A study as complex and focused as t h ~ would nothave beenpossible without extensive support and s cooperation ineach o f the districts studied. The D S D team thanks the DCOs, the administrative officers and the district mayors (nazimeen) o f all six districts visited for the hospitality and the access that they kindly afforded: Haji Naimutallah o f Karachi, Ms.Nafisa Shah o f Khairpur, Mr. IqbalKhan o f Bannu, Chaudhry Zahid Nazir o f Faisalabad, Mr.Tariq B a s k Cheema o f Bahawalpur, and Nawab HassanKhan Jogezai o f Killa Saifullah. They were all generous with their time, valuable information, and insights on the administrative challenges and successes o f their districts. vii The officials interviewed during the fieldwork included: District: Provincial Government: District Nnzim Home Secretary, Govemment o f Sindh, NWFP District Nnib Nnzim Staff o f Local Govemment Board (all provinces) DCO Deputy Secretaries,Finance Department, ` EDOs for F&P, Health, Education, W&S, Revenue, Govemment o f Sindh, Punjab Community Development, as well as staff o f MD,K S W (Karachi) MD,WASA (Faisalabad) these offices Staff of PHED (all provinces) District Accounts Officer (DAO) and staff o f DAO DIG Police (Karachi, Faisalabad, Bannu, office Bahawalpur) DO F&B, D O Planning, DOIDDO HRM, DO Treasury Accountant (and other officers), Karachi Coordination (Punjab districts), ACO (NWFP), Deputy Secretary, Local Govemment Department, DO NWFP Excise and Taxation Members o f the Provincial And District Election DPO Commissions Commissioner FCR, Bannu Civil Society: District and Sessions Judge Journalists, members o f NGOs and other civil Civil Judge (and Senior Civil Judgewhere present) society organizations including CCBs Judicial Magistrates, Special Judicial Magistrates Members o f the District Bar Association (Punjab) Members o f the CPLC (Karachi, Faisalabad) PPOs, DPOs, and SHOs o f various police stations Members o flnsauf Committees Probation Officers Parole Officers Superintendents O f Jail TMA: Tehsil Nuzim in each of the TMAs visited T M A Naib Nnzim TMO o f each T M A TOs of each T M A (I&S, Planning, Finance, Regulation) Preparation o f the report was assistedby valuable comments received from government and from others on an earlier version o f this paper prepared for the "Forum on Intergovernmental Relations and Service Delivery inPakistan," hosted by the National Reconstruction Bureau and held in Bhurban inJuly 2003. Peer reviewers for the workshop paper were Bert Hofman (World Bank), Dana Weist (World Bank) and Clay Wescott (ADB).Informal comments were made by Daniyal Aziz (Chairman NRB).Other very valuable comments on the workshop paper were provided by Ahmad Ahsan (WB), Benjamin Loevinsohn (WB), InaamHaq (WB), Luis Constantino (WB), Manuela Ferro (WB), Mark Poston (DfID),Naved Hamid (ADB), Pablo Saavendra (WB), Raja RehanArshad (WB), Shahnaz Arshad (WB) and Tahseen Sayed (WB). Administrative and logistical support has beenprovided throughout by Juliet Teodosio (World Bank) and Shaista Hussain(ADB). Copyediting was undertakenby Alfred Friendly and RebeccaKary. The initial findings of the Devolved Service Delivery Study were discussed with all participating nazimeen and other senior district staff, and Ministryo f Finance Staff, at a workshop held in September inIslamabad. Comments were also received following the circulation o f a summary note on the interimfindings o f the study inOctober 2003. The conclusions were fiuther reviewed at a PFC Workshop o n "Concepts and Practices in Intergovernmental Finance" held on December 16-18,2003, organized by the Government o f the Punjab, Finance Department and the Provincial Program Support Office, DSP. Aninternalpre-review meetingheld within the World Bank onDecember 19, 2003, and chaired by Shanta Devarajan, Chief Economist (South Asia) endorsedthe logic and the structure o f the report. A summary paper was presentedat a retreat o fthe National Steering Committee (comprising representatives o f Federal finance, Social Welfare And Women's Development, the NRB and all provinces, a range o f districts, and civil society organizations) held January 10-11, 2004. Further useful comments were provided o n behalf o f the National Steering Committee. viii The study buildson an array o fprevious analytic and operational reports prepared by the World Bank and ADB-very particularly including: Asian Development Bank. 1999. "Pakistan Legal and Judicial ReformProject: Integrated Report." ADB. Manila Asian Development Bank. 2001. "Fiscal Decentralization Technical Assistance Project : Fiscal Transfers Report (Volume 11)'' ADB. Manila Asian Development Bank. 2002. "Poverty inPakistan: Issues, Causes, and Institutional Responses." ADB. Manila Shah, Anwar and others. February 19, 1996. "Fiscal Federalism inPakistan: Challenges and Opportunities." Washington D C World Bank. 1999. "Agricultural Taxation inPakistan." World Bank. Washington D C World Bank. 2000. "Pakistan-Reforming Provincial Finances inthe Context o f Devolution: An Eight Point Agenda." World Bank. Washington D C World Bank. 2001. "Reforming Punjab's Public Finances and Institutions." World Bank. World Bank. 2002. "Pakistan Development Policy Review: A New Dawn? (Report No. 23916-PAK)" World Bank. Washington D C World Bank. 2002. Pakistan Poverty Assessment: Poverty inPakistan-Vulnerabilities, Social Gaps and Rural Dynamics. Washington DC: World Bank. World Bank. 2002. "Primary EducationManagement: ProvincialCase Studies (Punjab)." UnpublishedWorldBank paper. Washington D C World Bank. 2003. "Pakistan Public Expenditure Management: Strategic Issues and Reform Agenda." World Bank. Washington DC. ix Acronyms and Abbreviations ACR Annual Confidential Reports DSD Devolved Service Delivery ADB Asian Development Bank Study ADP Annual Development Plan DSSP Devolved Social Services AEO Assistant EducationOfficer Program AG Accountant General DTCE Devolution Trust for AGP Auditor General o f Pakistan Community Empowerment AGPR Accountant General Pakistan ED0 Executive District Officer Revenues EPI Expanded Programfor AIT Agriculture Income Tax Immunization APT Appointment, Promotion and ESR Education Sector Reform Transfer FBS Federal Bureau o f Statistics BHU Basic HealthUnit FCF Federal Consolidated Fund BKN Badan Kepegawaian Negara FIR FirstInformationReport (Civil Service agency in FJA Federal Judicial Academy Indonesia) GDP Gross Domestic Product BPS Basic Pay Scale GFS IMFGovernment Finance CBR Central Board o f Revenue Statistics CCB Citizen Community Board GoWB Government o f West Bengal CDL Cash Development Loans GRAP Gender ReformAction Plan CDMD City Development and GST General Sales Tax Municipal Department HIPC Highly-Indebted Poor CED Central Excise Duty Countries CEDAW Convention o n Elimination o f H M I S HealthMonitoring Information All Forms o fDiscrimination System Against Women IMF International Monetary Fund CFAA Country Financial IT Information Technology Accountability Assessment JM Judicial Magistrate CGA Controller General o f Accounts KMA Kolkata Metropolitan Area DA Development Authority KPP Khushal Pakistan Program DAO District Account Officer K W S B Karachi Water and Sewerage DC Deputy Commissioner Board DCJCC District Criminal Justice LG Local Government Coordination Committee LGC Local Government Commission D C O District Coordination Officer LG&RDD Local Government and Rural DDC District Development Development Dept. Committee LFO Legal Framework Order DDEO Deputy District Education LGO Local Government Ordinances Officer LHC Lahore High Court DE0 Deputy District Education LHW Lady HealthWorker Officer MICS Multi-Indicator Cluster Survey DERA Drought Emergency Relief MCP Malaria Control Program Assistance MIT Member InspectionTeams (of DfID UKDepartment for the HighCourt) International Development M&R Maintenance and Repair DM District magistrate MMA Muttahida M a jlis-i-Amal DMG District Management Group (political party) DoF Department o f Finance MNA Member o fNational Assembly DO District Officer M o E Ministry o f Education DPO District police officer M o W D Ministry o f Women's DPR Development Policy Review Development, Social Welfare DPSC District Public Safety and Special Education Commission MPA Member o f Provincial DRTA District Regional Transport Assembly Authority MQM Muttahida Qaumi Movement D S Deputy Secretary (political party) X MVT Motor Vehicle Tax PSC Public Service Commission NAM N e w Accounting Model PSDP Public Sector Development NBP National Bank o f Pakistan Program NFC National Finance Commission PTA Parent Teacher Association NGO Nongovernmental organization PTSMC Parent-Teacher School NIH National Institutes o f Health Management Committee NPA National Plan o f Action ROB Rules ofBusiness NPsc National Program Steering RTA Regional Transport Authorities Committee RWSSP Rural Water Supply and NRB National Reconstruction SanitationProgram Bureau SAP Social Action Program NWFP NorthWest Frontier Province SBP State Bank o f Pakistan OPD Outpatients Department S D M Sub-Divisional Magistrate OSR Own Source Revenue SE Superintendent Engineer OZT Octroi and Zila Tax SHO Station House Officer P&D Planning and Development SJM Special Judicial Magistrate PCA Police Complaints Authority SLGO Sindh Local Government PCF Provincial Consolidated Fund Ordinance PCP Pakistan Centre for SWAP Sector Wide Approach Philanthropy SMC School Management P E N Punjab Economic Research Committee Institute TAO Tehsil Accounts Officer PFC Provincial Finance Commission TB DOT Tuberculosis Directly Observed PHED Public Health Engineering Treatment Scheme Department TMA Tehsil Municipal PIFRA Pakistan Improvement to Administration Financial Reporting and TMO Tehsil Municipal Officer AuditingProject TO Tehsil Officer PIHS Pakistan Integrated Household UA Union Administration Survey uc Union Council PLA Personal Ledger Account UIPT UrbanImmovable Property Tax PML Pakistan Muslim League W&S Works and Services (political party) WAPDA Water and Power Development PO Police Order Authority PPO Provincial Police Officer W A S A Water and Sanitation PPP PakistanPeople's Party Authorities (political party) WB World Bank PPSC Provincial Public Service WDD Women's Development Commission Department PRGF IMFPovertyReductionand WDR World Development Report Growth Facility XEN Executive Engineer PRSP Poverty Reduction Strategy Paper xi Executive Summary CONTEXT Pakistan's far-reaching devolution initiative has been designed with three broad and inter-related objectives inmind: 0 To inject new bloodinto a political system considered to be the domain o fhistorically entrenched interests; 0 To provide positive measures enablingmarginalized citizens-women, workers, peasants-to access formal politics; and 0 To introduce a measure o f stability into a turbulentpolitical scene by creating a stronger line o f accountability between new politicians and local electorates. Underpinning the political strategy were other technical objectives: improved delivery o f social services; better determination and enforcement o f property and labor rightsand regulation o f economic activities; and access to justice inthe form o f improved performance by local administrations, courts and police, with greater awareness o f basic human rights protectedunder devolution. Remarkable progress has beenachieved. New local institutions with new structures for local government, new arrangements for intergovernmentalsharing o f resources, new electoral arrangements, new rules for government formation and dismissal and new opportunities for citizens to participate inthe affairs o f government have all been created. Installing this array o f new structures and accountability arrangements i s an achievement that can hardly be overstated. The process i s all the more remarkable given the other far reaching reforms that government undertook inparallel. At the same time as the devolution initiative was being implemented, the government also implemented significant reforms intax, trade, deregulation and privatization, the banking sector, anticorruption, restructuring federal and provincial legislatures and respondingto gender concerns. 1 Table 2: Progress to Date in Political, Administrative, and Fiscal Reforms PreAugust 2001 Constitution entrenchesthe Federal and Provincial govemments only. The four provinces were divided into three deconcentrated levels for administration and service delivery. Unstable and unpredictablelocal govemment system without commitment of the higher tiers to strengthen or sustain it. Post August 2001 Cross-cutting reforms. 6,458 new, directly and indirectly elected Local Governments. Methods established for both the intemal and extemal recall of the nuzimeen inlocal government. Citizen Community Boards (CCBs) conceived with abroad govemance mandate. Responsibilities for regulatingthe activities o f civil society devolved to district govemments. Particular emphasis on access to justice New institutional framework for checking police conduct introduced. Power triad o f local politicians, district commissioner and head of district police abolished. Emerging issues. O f the 126,462 new union councilors, 25 percent of the reserved (one-third) seats for women were filled. Electedpositions have in the main been won by people with previous political backgrounds, although some new blood i s evident. Slow progress in the new oversight bodies. Progress in establishingthe District Public Safety Commissions (DPSCs) has been patchy. District, province, and federal political representatives want greater control over administration and the police. PreAugust 2001 Significant presenceo f federal civil servants in the provincial administrations. Local administrations largely led by federal staff from the District Management Group. Mismanagement o f the police by the executive authority o f the state at every level. The District Deputy Commissioner role combined magisterial and administrative powers. Post August 2001 Cross-cutting reforms District govemments divided into 11departments. District CoordinationOfficer (DCO) established as the highest-ranking civil servant in the district. Some federal staff remain in the districts, but former rural district council employees now are district employees. However, some ambiguity persists about reporting arrangementsfor the large numbers o f staff from provincial employment groups. Particular emphasis on access tojustice. Office o f deputy commissioner abolished: powers divided between District and SessionsJudge, District Nazim and the DCO. Judicialpowers exercised only by thejudiciary. Emergingissues. Shifting the employer function to coincide with the new levels o f functional responsibility i s proceedingslowly, resulting in some competing lines o f control. Determinationo f the wage envelope takes place at the provincial level when salaries are still paid from the provincial Account Irather than the district-controlled Account N. Useful preparatory steps have been taken toward transfemngemployer functions to the school level, but this remains more an ambition than a reality. Autonomy o f the police enhancedwithout establishment of the new systems of accountability. Need to review allocation of quasi-judicial powers to make local administration more effective. 2 PreAugust 2001 Constitutionspecifies the functions o f the federal govemment and the areas o f shared responsibility, primarily treated as federal responsibilities. Centralization o f taxes by agreement with the provinces. Most buoyant taxes with significant revenue potential assigned to the federal govemment or ceded by the provinces on efficiency grounds. Main source o f provincial revenues is a transfer based on a share of federal tax collections determined by the National Finance Commission. The importance o f federal transfers to provinces has significantly increased over time. Post August 2001 Cross-cutting reforms. District govemments given responsibility for delivering services related to elementary and secondary education, primary and secondary health, agriculture, and so forth. Towns and tehsils have relatively unchangedresponsibilities. Union administrations have not been assigned any significant service-delivery responsibilities. Local govemments given the rights to raise some additional revenues. Provincial Finance Commissions established which have made awards for the distribution o fprovincial resources to local govemments. NB:No shifts ofresponsibility from the federal to the provincial govemments. Particular emphasis on access to justice. Additional funding provided to subordinate courts and police services. Emerging issues. Progress increating the necessary secretariats and establishing the required analytic support for the PFCs has been slow. For the majority o f districts, the provincial finance departments support budget-making. Budget execution procedures made it difficult for the district govemments to control general development expenditures. The cash flow position o f the province has become a factor obstructing the smooth flow o f funds to the districts. THIS STUDY Based on an empirical study o f 6 districts and 12 TMAs, this paper evaluates the extent to which the new structure has succeeded increating the incentives necessary for local governments to achieve at least some o f the service delivery objectives. The study uses a framework drawn from the World Development Report 2004 to integrate the analysis o f fiscal, administrative and political dimensions o f devolution with sectoral concerns: health, education, water and sanitation and access to justice. The paper addresses three notional readers. First, the "nazim o f good intention"-the recently elected local politicians who are keen to make improvements inservice delivery despite the ambiguities intheir present situation. It invites them to comment on a set o f new incentives and challenges and opportunities that devolution appears to have introduced. Second, govemment- invitingsenior officials who have worked long andhard inintroducingthese radical reforms to agree, or not, with the views expressed, the lessons learned and the recommendations offered for the next key steps. Third, donors-who, while overwhelmingly sympathetic to devolution and frequently strongly supportive, can nevertheless by their actions occasionally undermine the process. SERVICE-DELIVERY IMPROVEMENTS AREESSENTIAL Cross-country comparisons illustratethat, inPalustan, growth i s associated with fewer improvements inkey health and education outcomes than incomparable developing countries. The linkbetween increased fundingand better education or health indicatorsi s equally weak. Thus, as devolutionbegan, Palustan faced a dilemma. Its public institutions for social service delivery were demonstrably failing, and the evidence suggested that neither economic growth nor increased public expenditures were likely to have significant remedial impact. The Social Action Program represented an earlier attempt to find a way through. Ultimately SAP failed-perhaps overwhelmedby the deterioratingmacroeconomic position that squeezed public expenditures inthe social sectors, but more likely because it did not adequately prioritize 3 institutionalreform and did not tackle the thorny issues related to the incentives for staff to improve outcomes either at the policymaking level or at the service-provider level. WHY MIGHT DEVOLUTION OFFERA WAY FORWARD? The changes introducedby devolution were summarized inTable 2 above. This report puts into perspective what these imply for service delivery. Devolution changes the accountability relationships between the actors involved in service delivery. The study focuses on the changed situation that nazimeen find themselves in- and the degree to which this provides incentives for malung services work at the district and TMA level. Whether the incentives really will inturn lead to better services inhealth, education, infrastructure, and access tojustice across local governments remains an empirical question. Figure 1 summarizes these assumptions ina simplified framework for thinkingabout the prospect for making services work under Pakistan's devolution, centered on the nazimeen and their senior staff. Figure 1: Key Incentives inDevolution The incentives to make services work under devolution can be grouped under two broad headings. Nazimeen and I First,there is the set o f incentives that mightmake nazimeen want to see service delivery improvements for citizens at reasonable cost. The report considers whether Nazimeen are influencedby citizen power, inthe form o f "voice, so that they have an " incentive to direct those providers to achieve service delivery improvements (although as Figure 1illustrates, this assumes that they are ina positionto direct the behavior o fthe staffthat provide the services-motivated nazimeen with no ability to direct their front line staff will not be able to improve services). It also considers whether voice i s accompanied byprovincial eficiency incentives, so that service-delivery improvements are not simply obtained through increased spending. Second, the report reviews the incentives on those who provide the services-for example, the staff o f schools and clinics-addressing the question o f whether they are likely to want to see improvements. More specifically, it looks at managerial power, the arrangements that ensure that the nazimeen and senior staff have authority over the front line workers and, again, citizen power (thistime inthe form o f "clientpower") that provides providing potential service recipientswith some direct means to have an impact on the conduct o f service providers. The methodology used inthis report i s to identify the components o f these four sets o f incentives and to track changes that can be attributedto devolution. Box 1summarizes these incentives. 4 Box 1: The Sets o f Incentives Tracked in this Report [ncentives o n nazimeen and senior staff: Provincial efficiency incentives derive from the structure o f the revenue base, including transfers and own source revenues available to meet the new local expenditure responsibilities. Ifwell structured, these fiscal arrangements promote the efficient use o f resources inmeeting local preferences. Since nazimeen will typically face expenditure demands that exceed available resources, budgetary certainty inthe form o fpredictable resource flows andahardbudgetconstraint, coupled withadequate autonomy inresource allocation, will encourage them to use public resources responsibly and efficiently. With these arrangements inplace, nazimeen will more likely attempt to make the best trade-offs, for example, inallocations between sectors, between recurrent wage and development expenditures, and attain value for money (for example, through effective procurement arrangements). Voice offers a channel for citizens to make their demands for better services heard-the most obvious form is elections. However, voting may be based on other criteria than service delivery, as owing to years o f bitter experience voters may not see commitments on the part o f politicians to provide broad- basedpublic goods over time as credible. Hence, voters and politicians may emphasize directly visible and narrowly targeted projects such as school infrastructure. These schools may not be staffed or maintained, but voters have leamt that they cannot holdpoliticians to their promises beyond the short term. Voice i s likely to be stronger, with councilors motivated to listen to citizens, ifthere i s public assignment o f credit or blame and ifthere are obstacles to councilors providing private goods targeted at a few influentials rather than public goods that benefit many.To influence nazimeen, citizen voice must compete with the pressures from the provinces on nazimeen to serve their political interests. That "pull" from the provinces is reducedwhen the councilors have independent political legitimacy and when provinces themselves have fewer incentives to intervene inlocal affairs. [ncentives on the providers: The local governments, as represented by the nazimeen, engage with service delivery providers (for example, schools and contractors) through managerialpower, w h c h reflects their ability to require service providers to adhere to particular standards o f behavior and service. Again, as is explored in more detail inthe second volume (Annex 1-Recent History), managerial power is strengthened ifthe staff that work inthe district are employedby the district, with hire and fire authority available to local managers, subject to suitable safeguards. Recognizingthat this i s unlikely to be the case for senior staff, nazimeen need distinctive authority over their behavior including input into their performance evaluations, and a credible threat to transfer them ifthe relationship i s not productive. Managerial power i s strengthened by the external oversight providedby bodies such as Accounts Committees, Monitoring Committees and District Public Safety Commissions. It also assumes that staff with appropriate slulls are available insufficient numbers and that those staff have the skills and sufficient technical sanction powers to undertake their tasks. A shorter route o f accountability, clientpower, directly connects citizens with service provides, such as schools. Client power includes more formal mechanisms as functioning agencies that assist in dispute resolution (Departmental Grievances Redressal and Complaint Cells and Citizen Police Liaison Committees) and bodies that provide direct citizen inputs into the management o f facilities (Citizen Community Boards and School Management Committees). If,onbalance, these changingincentives are infact havingsome impact on service delivery, then, inprinciple, improvementscouldbe measureddirectly. However, there are two difficulties. First, devolution i s recent and it would be premature to attribute any change (positive or negative) in social sector outputs to new arrangements for service delivery. Second, inPakistan there i s no consensus on which data sources to use to measure intermediate and outcome indicators and little robust district disaggregated data, and it would be dangerous to judge the utility o f devolution on thebasis o fany apparent trends inthese indicators. 5 KEYFINDINGS New Incentives for Nazimeen and Senior Staff Provincial efficiency incentives The study assumes that devolution would be providing district and tehsil nazimeen with greater incentives for efficiency infiscal management ifit achieved positive movement inthree key categories: 0 Budgetary certainty-the degree to which they must live withinhard constraints and the predictability o f the budgetary transfers that they depend on; 0 Autonomy inpreparingthe budget-specifically the development, the non-salary recurrent and the salary budgets; 0 Incentives for local revenue-raising-the buoyancy and potency o f taxes assigned, the clarity o f revenue assignments and other incentives for revenue raising. The scorecard on the creation o fprovincial incentives for fiscal efficiency i s mixed. Onbudgetary certainty, as the report shows, most funding for local government i s ineffect negotiated rather than formula-based. All negotiatedrevenues provide opportunities for gamesmanship-basing current policy decisions on their likely impact on future resource transfers, rather than on their merits. The proportions o f the Provincial Consolidated Fundthat ultimately are transferred on formula basis are small. Less than 25 per cent o f all transfers mandated by the Local Government Ordinances (LGOs) are calculated on a formula basis.' However, while these transfers are negotiated, the data also reveal that there i s considerable predictability inthe transfers at the district level, with inmost cases districts receiving only slightlyless than what was budgeted. The district studies also revealed that local governments have limited autonomy inpreparing their budgets. Vertical programs are the main constraint on district and TMA autonomy inpreparing the development budget, with often the volume o f funds channeled through vertical programs dwarfing the district's own Annual Development Plan(ADP) for particular sectors. Inthe study districts, more than halfo f the ADP i s under the effective control o f the federal and provincial agencies controlling the vertical programs. Local govemments also have limited freedom in preparing the salary budget. Neither local nor provincial govemments are able to determine pay scales, which are de facto set at the federal level. Inaddition, in Sindh and NWFP, districts must seek approval from the province for the release o f the funds from Account I(the provincial account) for their salary payments.' Local govemments are hugely dependent on intergovernmentalfiscal transfers. This degree o f vertical imbalance can leadto inefficiencies inboththe levels and quality o fpublic expenditure- specifically, it creates incentives for local govemments to overspend as well as reduces the incentives to spendthese resources efficiently. The logic inboth cases i s that intergovemmental transfers lower the costs to local residents o f getting services. Arguably, the disconnection between expenditure and taxation encourages "free-riding," as well as reducingthe incentives to monitor public expenditures, as resources are collected from distant taxpayers not likely to be usingthe services producedby the local government. Thus itis at least arguable that localrevenues contain intrinsic incentives for efficiency, as local taxpayers mightbe more vigilant than distant ones and so efforts to improve own source revenues ' It is important to note that transfer has a precise meaning and refers to funds that are non-lapsing and within the ' re-appropriation authority o f the district or TMA. InPunjab andBalochistan, the salary budget is transferred directly inDistrict Account 4. 6 can imply incentives for efficiency. Incentives for local revenue effort are linkedto the buoyancy and potency o f taxes assigned to local government, and the clarity o f such assignments. Effort will be deterred ifthere is no credible connection to likely revenueresults and ifit is not evident which level o f government i s to benefit from which tax. While districts have been given rights to raise additional revenues, their tax base i s weak, and some district taxes and user charges are very difficult to impose inpractice. Inthe six study districts, for example, for 2003/04 own source tax revenues ranged from 0 percent to 8 percent o ftotal revenues. Therefore, even a large tax effort may bring only modest results intotal revenue. Inaddition, althoughthe district andTMA taxes are clearly assigned intheory, administrative confusions inthe collection arrangements blur the assignments inpractice, further reducing incentives to increase collections. Administrative complexity i s most clearly evident inthe case o f property tax (UPT). Citizen Power (Voice) This study argues that there are two necessary conditions for the localpolitical leadership to have incentives for service delivery improvements. First, local councilors themselves must have an interest in serving the needs o f their constituents and inparticular the poor. Second, since the relationshipbetween nazimeen and councilors i s a critical intermediate step inthe linkage to citizens as nazimeen are indirectly elected by councilors, nazimeen must findthe pull from the councilors more demanding than the pull from other political powerbrokers, particularly the provincial governments. Onthis basis, devolution has, onbalance, created distinct incentives for nazimeen to listento citizens. However, jurisdictional overlap makes it difficult for politicians to seek credit for improved services. When multiple levels o f government are involved indelivering the same services, the public i s unable to assign credit or blame effectively for that particular service to a particular politician. This overlap weakens incentives to perform well on service delivery and could encourage politicians to target services to their core supporters. Jurisdictional overlap i s particularly an issue inthe education and health sectors, as federal and provincial vertical programsremainprincipal tools inpromoting nationalpolicy priorities and represent a significant proportion o f local-government expenditures on service delivery. The anecdotal evidence from the study districts also suggests the strong incentives workmg on councilors to focus on private goods, such as patronage or locality-specific projects. However, the evidence from the study districts suggests that while the pressures to focus onprivate goods are strong, devolution has at least not exacerbated this problem. For example, the data shows that there i s no obvious reduction inoperations and maintenance expenditures between the allocations in2003/04,the first year local governments prepared their ownbudgets, andthat o f earlier years. Similarly, the data suggests that there has not been a significant shift from large to small development schemes, or that councilors are abandoning existing schemes to gainpolitical glory by starting new ones. The importance o fthe political relationship between the nazim and the ruling provincial political coalition emerges strongly fi-om the study. Intwo o f the six districts visited-Bahawalpur and Faisalabad-the District Nazim is allied to the Chief Minister and the rulingprovincial political coalition, while inthe other four the District Nazim i s inopposition. This relationship has a This paper's referencesto private goods include both pure private goods (that is, individual-specific) as well as narrowpublic goods (for example, locality-specific projects). Local infrastructure projects often encompass both- that is, only the local community benefits, and the contractors for the projects reap the privaterewards. 7 significant impact on the flow o f funds between the province and local governments, on the implementation o f local development schemes and on the manipulation o f appointments and transfers o fpersonnel. Effectively, the relationshipwith the province determines the extent o f political control that the District Nazim can exercise within local government as well as the freedom local policymakers have to act to meet their constituents' demand^.^ New Incentives for the Service Providers For service delivery to improve, the service providers-and notjust their elected superiors-must want to see improvements. This motivation i s provided either by ensuring that the service providers have an incentive to listento the policymakers-managerial power-or by providing potential service recipients with some direct means to have an impact on the conduct o f service providers-client power, inthe terminology o f the World Development Report (World Bank: 2003~). Managerial power over serviceproviders Inreviewing the changed incentives for front line staff andother service providersto "listen to Nazimeen," the study examines three key dimensions: 0 The extent to which staff workmg inthe district are employedby the district. 0 The adequacy o f the nazim's authority over senior staff, including inputs into the performance evaluations for senior staff and ability to require their transfer. 0 The extent o f external oversight, including a review o f the effectiveness o fbodies, such as Accounts Committees, Monitoring Committees, and District Public Safety Commissions, which are mechanisms for the monitoring o f front line staff. Managerialpower over service providers i s weak. Administrative devolution remains incomplete and, as yet, has given district staff few reasons to respond to the authority o f the nazimeen. The transfer o f functions and the physicalrelocation o f staff have not been fully matchedby the transfer o f the employer role. While the LGO provides for a sophisticated sharing o f control over posting o f civil servants based on norms which clearly delineate responsibilities, informal provincial controls are clearly evident inthe area o fAppointment, Promotion and Transfer (APT) authorities. The result i s that althoughphysically located inthe districts, most senior district staff do not consider themselves as district employees and therefore are likely to accommodate provincial pressures to transfer subordinate staff. This attitude has considerably weakened the accountability o f senior staff to local elected representatives. Infact, politically motivated inter- and intra-district staff transfers are a main area o f conflict between.provincial and local governments and are most severe indistricts that are not politically alignedwith the province. Since TMA staff largely belong to local government cadres or are the council's own employees, the problem o f l o w institutional allegiance and weak accountability i s less acute. These problems o f accountability were clearly evident inthe relationship between the District Nazim and the senior staff inthe district, particularly the District Coordination Officer (DCO) and the District Police Officer (DPO). It was generally observed that the authority o fthe nazim to write the Annual Confidential Reports (ACRs) o f DCOswas an insufficient mechanism for enforcing accountability since it was an ineffective tool for measuringthe performance o f an officer and tended to be viewed as a routine matter to be placed on an officer's file at the time o f promotion to fulfill a legal or administrative requirement. Officers tended to be given postings See the second volume (Annex I-Recent History) for a full treatment o f the political relationships between the provinces and local govemments. 8 regardless o f their performance inprevious tenures. Nazimeen also have very little authority over the DPO. Formally, at the time o f the study, only inPunjab had the District Nazim been given the initiating authority for the ACR o f the DPO, although it has been reportedthat this authority will be extended to the nazim inall provinces. However, inpractice innone o fthe districts studied were these performance evaluation reports completed inthis manner. Some nazimeen, in fact, considered that removing the authority o fthe former district magistrate incompleting the ACR o f a former district Superintendent o f Police had contributed to deterioration inlaw and order in general.5 The LGOs provide for a number o f institutional arrangements-Monitoring Committees, and District Public Safety Commissions-for the oversight o f service providers by elected representatives.6However, it was generally observed that thesebmechanisms have either not been established, or have been introducedto only a limited degree. For example, the field visits revealed that, with some notable exceptions and for several reasons, MonitoringCommittees were not performing as the regulations envisioned. Similarly, Public Safety Commissions where formed, were generally unable to elicit compliance from the police inits investigations. Citizenpower (client power) Service delivery can also be improved by improving client power, that is, strengthening the direct linkage between citizens and service providers. Inprinciple, such an approach can be more effective than focusing on efforts to make service providers more accountable to policymakers and then making policymakers, inturn, more accountable to citizens. The LGOs and Police Orders call for the creation o f various bodies, such as Citizen Community Boards and Citizen Police Liaison Committees, which can serve to improve this client power. The overall conclusion from the field visits i s that it i s too early to tell whether or not devolution has had any impact on client power for the better or the worse. While in general participatory mechanisms, such as Citizen Community Boards and School Management Committees (a pre-devolution initiative) have yet to achieve their full potential, there are numerous examples that suggest that innovative local leadership has had a positive impact on community mobilization and inturn on service delivery. ASSESSMENT A Milestone Passed Devolution has brought a decisive break with the past intwo important respects. First, local government now has unambiguous constitutionalprotection. The agreed changes to the constitution enacted on 30 December 2003, has placed the local government system, as prescribed by the provincial LGOs, inthe sixth schedule o f the 1973 Constitution and thus they cannot be altered, repealed or amended without the sanction o f the President. Second, as a result o f the LGO provisions, the reservation o f 33 percent o f local council seats for women inall three levels o f local government has proven the most strategic step ever undertaken by government toward women's political empowerment. Most directly, despite the significant number o f vacant seats, it has enabled over 30,000 women to enter into formal politics at the local government level (see Table 3). Pressreports of a high levelmeetingheld on 14 April 2004 involving the PrimeMinister, Provincial Chief Ministers, the NRB andother federal andprovinceofficials indicatethat significant changes will be introducedto the Police Order 2002 andthat anew target date ofDecember2004 hasbeen set for complete implementationof these new arrangements.Until these are officially confirmed and clarified, this report will make only passing reference to what hasbeenreported, but will not speculateon their likely impact. It is notedthat the DPSCsinclude non-electedmembers, but that these are appointedfrom alist providedby the nazimeen. 9 Province Women Women union Total union Women union nazimeen councilors council seats councilors as a percentage of total Punjab 9 20,007 72,513 27.6 Sindh 2 5,878 22,974 25.5 NWFP - 3,963 20,097 19.7 Balochistan - 2,374 10,878 21.8 TOTAL 11 32,222 126,462 25.4 Devolution marks a decisive departure from the previous arrangements that became associated with persistent failures inservice delivery. This studyreaffirms the government's view that devolution offers the best prospect for improved service delivery and access by the poor to public entitlements, including notjust health, drinlungwater and education, but muchneeded improvements inpolicing and localjustice. While it i s too soon for the proponents o f devolution to point to an overall improvement in service delivery, the study does reveal significant achievements. For instance, anecdotal evidence suggests that doctors and teachers are now more likely to be at their post and drugs are more often available inclinics. Citizens undoubtedly appreciate the improved access they have to their political representatives. Councilors, both directly and through new dispute resolution arrangements at local level, are now seen to be respondingto people's needs and concerns. This change suggests that devolution may deliver on its larger political objectives to create a better matchbetween what governments do and the public's ability to hold them accountable for performance. This study i s therefore modestly optimistic about the future o f devolution. The transformation underway through devolutionnevertheless faces significant challenges. Detractors argue that infact devolution has made matters worse. Insome respects opponents o f devolution are inthe same position as proponents; the evidence i s indeedmixed. Yet, while morale has deteriorated for some civil servants, and law and order i s thought by many to have worsened due to a failure to reestablish oversight o fthe police, no one consulted duringthis study saw any merit ina return to the old systems. Moreover, while many are concerned about the possible re-emergence o f "private goods strategies" under which local politicians resume the time-honored practice o f rewarding their constituents with private benefits at the expense o f public goods the study suggests this i s not infact a fair representation o f what i s occurring. The key variable inpromoting successful devolution would appear to be the strength o f the provincial incentives to intervene inthe districts. The province has few incentives to intervene when: 0 The ChiefMinister andor the mainprovincial coalition and the district nazim are from the same political alliance. 0 The Provincial Assembly constituencies within the district are filled with politicians from the same political alliance as the nazim. 0 The nazim has sufficient independent political standing to resist provincial interventions. 10 When the province has few incentives to intervene, the nazim gains a fairly free hand to direct district staff and, despite the inevitable resource constraints, can achieve results. However, when the provincial incentives for interventionare high,nazimeen often finds their hands tied. The Glass i s More than Half Full Table 4 and Table 5 summarize the degree to which the incentives for the key actors are changing. Dark cells indicate little progress; lighter cells suggest more forward movement. Overall, the study finds that some appropriate new incentives have indeedbeen created and that they are undoubtedly promising as they directly address the institutional weaknesses that the failure o f SAP highlighted-weak inter-sectoral allocation o f expenditures, low emphasis on non- salary recurrent expenditures and few incentives for efficient utilization o f these expenditures (World Bank: 2002a). 11 ixr I 12 t 3 Devolution i s undoubtedly a fact o f life inPakistan. It i s implausible that there will be a major returnto a discredited set o finstitutional arrangements with aproventrack recordo f failure. Equally, however, no certainty exists that devolution will now be developed and entrenched to maximize its potential returns. The risk i s that by failing to deliver visible service delivery improvements inthe short term, the devolution project will fall short o f completion and the present hybridarrangements-part new, part old-will become permanent. RECOMMENDATIONS The government emphasized the need for a robust technical analysis o f emerginglessons and feasible, practical recommendations that add value to a devolution process that i s well advanced. With the principle o fpracticality very clearly inmind,the study sets out an agenda for furthering devolution. It suggests both a strategy and some practical actions that will offer some relief now to hard-pressed nazimeen and senior staff inlocal govemments who are attempting to make the systemwork. Key elements of the SuwestedStrategy 1.Seize the opportunitypresentedby the NFCAward to conduct a well-targeted, strategic information campaign usinga variety of media-print and electronic media Note the successes.N o w i s the time for a public information campaign, setting out both progress and challenges. Lookfor opportunities that are offered to local governments following the NFC award and the improved financial position that this gives to provinces. Ensure that citizens understand what this means for local govemment budgets and priorities. Take advantage of the opening up of the electronic media to the private sector-in particular local radio. 2. Enforce the core principles-and accommodate local differences. NRB, whilst ensuring that the core principles o f devolution are not diluted, should be willing to allow some marginal local variations to meet the needs o f different areas. Implementation requires new strategies that recognize Pakistan's diversity while keeping devolution on track. The March2000 announcement o f the devolution plan and the subsequent August 2000 Local Government Planpresented a broad strategic vision. Since then, the focus has been on the technical, fiscal and administrative aspects o f implementation, superintended by the NRB.Underlying this emphasis has been concern to ensure compliance with the local government and associated ordinances. Now mightbe the time to reassert the broad goals o f devolution. Allowing diversity wouldpromote wider support. The range o f actors has grown exponentially. The basic structures for devolution have beenput inplace. Some local diversity inimplementation i s now feasible, and accepting marginal differences will avoid needless federal-provincial conflict. It would demonstrate that the federal govemment will hold fast to the core principles, while at the same time allowing local institutions to reflect local circumstances. 3. Transmit national policy priorities more effectively. Movefrom an over-reliance on verticalprograms. A large share o f financing for health, education, water supply and sanitation i s delivered inways that systematically undermine local government accountability. Vertical programs represent attempts by higher tiers o f government to influence the policy and behavior o f lower levels o f govemments by direct intervention. Retaining some influence over sub-national policy is sensible. Federal and provincial governments need some levers to influence local performanceto support national 15 objectives. Inpractice, though, vertical programs have a poor track record and they now risk damaging the credibility and nascent accountability o flocal governments. 0 Create incentives to encourage local governments to buy into national priorities. Grants coordinatedthrough the mechanism o f NFC and PFC and conditioned on outputs, outcomes or processes are an important part o f a strategy through which to promote nationwide standards andpriorities. Conditional grants can interfere with the accountability o f local representatives to their constituents. They must not be used excessively. But they do offer a pragmatic instrument through which federal and provincial governments can maintain a legitimate interest inthe institutional reforms and service-delivery arrangements at the sub- national level. Strategic Goals and Practical Actions The study identifies 7 strategic goals and practical steps that are achievable given the impressive progress made to date. The goals, along with the immediately operational, implementing measures, include: 1. Achieve public consensus that completing devolution is inevitable Issue a White Paper on strategic options for adjusting legal, administrative, political, and fiscal arrangements insupport o f devolution. 2. Demonstrablefederal government commitment to strengthenedrole o fthe provinces 0 Timetable for the PRSP commitment on functional devolution with a review o f the relative roles and responsibilities o f Federal and Provincial levels inthe context o f service delivery. 0 Set preparations inmotion for the 7th NFC Award now. 3, Demonstrable provincial government commitment to strengthening local government 0 Complete reassignment o fprovince-administered functions to local governments. 0 Incorporatethe Lessons o f Transition inRevisionsto the LGOs. 0 Major Commitment to Capacity Development. 0 Undertakeprovincial-local tax reassignments. 4. Local governments challenged to demonstrate that they merit confidence. 0 Reduce development throw forward. 0 Increase coverage o f user charges. 0 Strengthen inter-jurisdiction planning and coordination. 0 Local governments to regulate and engage with private sector service providers. 5, Judicious Selection from a Menu o f Conditional Grants inPFCNFC Transfer Systems 0 Use conditional grants-cautiously. 0 Review existing vertical programs and redesign proposals. 0 Broadenavailable array o f social indicator data and capture district data. 6. Hold local governments to account 0 Districts employ their own staff. 0 Districts maintain their own accounts. 7. Clarify the separation of executive andjudicial powers. 0 Restructurereporting obligations betweenthe District Police Officer and the District Nazim. 0 Assignnon-adjudicatorypowers to the local government executive. 0 Trainjudicial magistrates innewjurisdictions under criminal and local laws. 0 Provide HighCourt protection for judges interdicting unlawful action by officials. Establishseparate civil and criminal courts, and appoint Senior Judicial Magistrates in each district 0 Establishindependent prosecution service at district level. 16 Table 7 suggests the sequence with which these steps mightb e initiated. I Table 7: Timetable of ProposedKey Actions I FY07 ~ FY05 FY06 * IssueaWhitePaperonstrategicoptionsfor 2 E adjustinglegal, administrative,political and I 8 3ir fiscal arrangementsin support of devolution.' EstablishHighLevel ExpertGroup on federal- provincial tax reassignment(includingpossible debt-for-performanceswaps). PFC announcementsof frameworkunder which Review of the Rules of the provincial govemmentsmight use Business, the constitutiono f conditionalgrants, selectively,within thePFC the Local Government awards, to achieveoverarchingpolicy goals. Commissionand the facilities Developandpromulgatelocalgovernment and authoritiesprovidedto the public-privatepartnershiprules (as under LGO local governmentMonitoring Schedule5, Part A). Committees. Providedistrictswith full APT authorityover all Prepareandnotify the Rulesof staff at grades 19 andbelow. Businessdetailingthe Make three majorpolicy announcements: reportingobligationsbetween a) Proposedreassignmentof province- the DistrictPoliceOfficer and , administered functionsto local governments the DistrictNuzim. complete. Might reasonablyinclude: Assignnon-adjudicatory -- TMA debt-for-performanceswaps powers and supervisionof Transfer of PHED staff andbudgetsto the pretrial detention' to the local TMAs government executive. b) Formalreviews ofprovincial-localtax Establishthe Independent reassignment,to reportwithin 2 years. Prosecution Service andensure c) Major commitmentto capacity its extension to the district development. level. Reducedevelopmentthrow-forward. Increasecoverageof user charges. Empowerthe Mushawarat Committeesto provide aplatform for district-wide,medium- term planning. Trainjudicial magistratesinnewjurisdictions EnhanceHigh Courtprotection Establishseparate under criminal and local laws for subordinatejudges criminal and civil interdicting unlawful actionby courts established officials. at district level and appoint Senior Judicial Magistrates in eachdistrict 1. Such aWhite Papermight reasonablyinclude: 17 SUPPORTING MATERIAL The appendixes to this volume set out details o f the fiscal dataset that was prepared for the study, and describe three "assessment frameworks" that were developed. These frameworks (for assessing: institutional changes within local government; the location o f the employer function; and the "verticalness" o f conditional grants) could provide the basis for extended studies and for more regional comparisons. This assessment andrecommendations is accompaniedbytwo other volumes: Annex I-Recent History has two parts. The first part sets out the detailed history o f the initiative, and the political, fiscal, and administrative changes that have been made to date. Itthen provides a detailed assessment o f the impact o f these institutional changes on the key actors, particularly the nazimeen and senior staff o fthe districts and the TMAs. Annex 2-Technical Considerations provides a more in-depthreview o f a series o f contentious topics that contribute to the intensity o f the debates on devolution. It examines: 0 Options for reducing the throw-forward o f ongoing ADP Schemes in devolution 0 Developments and limitations inmonitoring data. 0 Provincial-local fiscal transfers including a review o f the weightings used inPFC formulae. 0 An examinationo fthe vertical fiscal imbalances, andthe implications from a possible change inthe NFCaward. 0 Reforming access to justice under devolution. 0 Key considerations inadministrative decentralization. 18 1. ServiceDelivery andDevolution SERVICE-DELIVERY IMPROVEMENTS ARE ESSENTIAL The track record o f social sector service delivery inPakistan i s so poor that it i s tempting to assert that almost any reform ingovernance arrangements would produce better results. Many reports, including the Table 8: Pakistan's Social Indicators in International government's PRSP, point out that Perspective Palustan's social indicatorslag behind those o f countries with comparable Indicators Actual Predicted Difference per capita incomes. According to one values for values for between estimate, Pakistan's gross female Pakistan countries actual and primary enrollment rate i s 33 (2001/02)a with predicted percentagepoints below and its infant similar incomes mortality rate 20 per thousand above (2002)b that o f countries with the same per :ertilitv I 4.5 3.8 0.7 capita income (see Table 8). idult illiteracy` 49.5 29 20 idult illiteracy-female 62 35 27 As devolutionbegan, Palustan faced a boss primaryenrollment 72 98 -26 double dilemma o f long standing. Its koss primaryenroll- public institutions for social service emale 61 94 -33 delivery were demonstrably failing, nfant mortalityrate 82 62 20 and the evidence suggested that Jnder-5mortalityd 105 91 14 neither economic growth nor a. PIHS 2001102, unless otherwise indicated. increasedpublic expenditures were b. The predictedvaluescolumnis derivedfroma cross-country likely to have significant remedial regressionaimed at explainingthe indicatorinquestionbyper capita impact. Cross-country comparisons GDP (PPP).Dataare fromthe WorldDevelopmentIndicators(World illustratethat, inPalustan, growth i s Bank). associated with fewer improvements c. PRSP,basedon 1998 Population Census. d. Ministry ofHealth. inkeyhealthandeducation outcomes than incomparable developing countries. For example, the female-male literacy gap increasedin Palustan with growth while it declined ina sample o f comparator countries (Figure 2); infant mortality declinedby 56 percent inthe comparator counties, comparedto only 27 percent in Palustan (Figure 3).' ' The comparatorcountriesare 22 low- andmiddle-incomecountriesthat on average hadthe same growthrateas Pakistan.It shouldbenotedthat these are 30-yeargrowth averages andmay maskmodest year-to-yearprogressat the countrylevel. 19 Figure2: Growth and Female-maleLiteracyGap Figure 3: GrowthandInfant Mortality in inPakistan and Comparison Sample, 1970-2000 Pakistan and ComparisonSample, 1970-2000 1 0 - 0 8 - 0 8 - 0.4 - 0 24 1 0 0 4 "" 1970 1985 2000 1970 1985 2000 1 - t Comparator countries+Pakistan +Comparator countries +Pakistan Source: World DevelopmentIndicators. Source: WorldDevelopmentIndicators. The linkbetween increased fundingand better education or health indicators i s equally weak. The Social Action Programhas to date been the Government o f Palustan's most concerted effort at improving its pro-poor social expenditures. Duringthe two phases o f the program (SAP Project 1, 1993/94-96/97, and SAP Project 2 1997/98-99/00), a total o f approximately $9 billion was spent, with 80 percent being contributed bythe government. The program aimedat increasing the physical availability and improving the quality and efficiency o f services, especially for the poor and for women, in elementary education, basic health care, family planning andrural water supply and sanitation. Outcomes were disappointing, particularly ineducation, which received roughly two-thirds o f this allocation (Table 9) (Easterly: 2003; Operations Evaluation Department: 2002). Overall net enrollmentrates declined from 46 percent in 1991/92 to 42 percent by 2001/02, while net enrollments inrural areas-a SAP focus-fell from 41 percent in 1991-92 to 38 percent in2001/02 (PIHS data). There was however, some provincial variation around this decliningnational trend-the situation inSindh, and especially Balochistan, was particularly disturbing, while NWFP showed some moderate progress. There was similarly little progress inrural water supply and sanitation. The percentage o f households with access to indoor piped drinlungwater, in fact, declined slightly from 25 percent in 1995/96to 22percent in2001/02. There were however, some positive achievements inhealth as indicatedby a significant decline inthe infant mortality rate (mortality rate among infants under age 1) and incidence o f diarrhea (a significant cause of infant death) and modest improvement inprenatal care as measured by the immunizationrate o f children from 12 to 23 months o f age. 20 The roots o f SAP'Sfailure are o f course complex and have been Table 9: Key IndicatorsinSAP FocusAreas well explored.* Onthe one hand, it may be that SAP was overwhelmed by the deteriorating macroeconomic position that squeezed public expenditures in the social sectors-for example, public expenditures ineducation declined from 2.1 percent o f GDP in1991/92to 1.7percentby 1998/99 (World Bank: 2003a). Onthe other hand, however, the program failed to improve the composition o fpublic expenditures, and non-salary inputsfailedto keep pace with development expenditures, resulting inpoor maintenance o f a. Based on record and recall. new infrastructure. Most b.Percentage o fhouseholds with accessto indoor pipeddrinking water. importantlythough,the Source: PIHS, various years. failed to adequately prioritize institutional reformand did not tackle the thorny issuesrelated to the incentives for staff to improve outcomes either at the policymalung level or at the service- provider level.' HowCANDEVOLUTION ASSISTIN SERVICE-DELIVERY IMPROVEMENTS? A Framework As noted, the existing institutionalarrangements for social sector service delivery inPalustan deliver ineffective services and neither growth nor additional investment expenditures offer a ready cure. Devolution offers a new set o f institutional arrangements that, arguably, can produce effective solutions. It i s important however, to emphasize that government policy statements did not regard service delivery and responsiveness as ends inthemselves. Insome respects, infact, the architects o f policy saw the pitfalls o f measuringthe success or failure o f devolution by a narrow emphasis on efficient service delivery. Pakistan's devolution initiative has been designed with three broadandinter-related objectives inmind: 0 To inject new blood into a political system considered to be the domain ofhistorically entrenchedinterests; 0 To provide positive measures enabling marginalized citizens-women, workers, peasants-to access formal politics; and (World Bank: 2002a) notes that one largely ignored achievement o f the SAP was that it placed social development center stage among both govemment and citizens. Thus it undoubtedly contributed to the political prioritization o f poverty reduction approaches and undoubtedly contributed to the view among senior political leaders that it i s far more efficient inthe short term to allocate the new fiscal space to debt retirement than increasing social service spending. The magnitude o f this problemof unresponsive staff can be gauged by the fact that the number o f complaints filed against key govemment departments more than tripled between 1985 and 1998 and that the proportiono f complaints that concern arbitrary decisions has risen very significantly (Asian Development Bank: 1999). Complaints are categorized as delay, inattention, neglect, arbitrary decisions and corruption. 21 0 T o introduce a measure o f stability into a turbulent political scene by creating a stronger line o f accountability between newpoliticians and local electorates. Improvements in service delivery were one o f several technical objectives, but as the overarching aims o f devolution were political, it can not be said to stand or fall on the basis o f its impact on service delivery. The changes introducedby devolution were summarized inTable 2 above. Full details are provided inthe second volume (Annex I-Recent History). Given the multitude o fpolitical, administrative, and fiscal changes, it i s important to try to put into perspective what these might implyfor service delivery. Devolution changes the accountability relationships and the relative importance o f various actors-most notably politicians and policymakers-associated with service delivery, as cast by the 2004 World Development Report (World Bank: 2 0 0 3 ~ ) .The study focuses on the capacity ' ~ and incentives faced by nazimeen for making services work at the district and TMA level. Whether the incentives really will inturnlead to better services inhealth, education, infrastructure, and access tojustice across local governments remains an empirical question. Cross-national studies findno clear evidence that decentralization has ledto better governance and improved service delivery (Azfar: 2002; Blair: 2000; Smoke: 2001)." Figure 1summarized these assumptions ina simplified framework for thinking about the prospect for making services work under Pahstan's devolution, centered on the nazimeen and their senior staff. The incentives and capacity to make services work under devolution can be grouped under two broad headings. First,there is the set ofincentives that mightmake nazimeen want to see service delivery improvements for citizens at reasonable cost. As will be seen, ina context o f indirect elections for nazimeen and complex intergovemmentalrelationships and political career paths, it i s far from self-evident that service delivery improvements for the poor will be a priority for them. Nuzimeen mustbe influencedbycitizenpower, inthe form of "voice,"" so that they have anincentiveto direct those providers to achieve service delivery improvements (although as Figure 1illustrates, this assumesthat they are ina positionto directthe behavior o fthe staffthat provide the services-motivated nazimeen with no ability to direct their front line staff will not be able to improve services.). Voice must be accompanied byprovincial efficiency incentives, so that service-delivery improvementsare not simply obtained through increased spending. lo The 2004 World Development Report (World Bank: 2003c) presents an accountability framework focused on service delivery. At its simplest,the WDR sets out an "accountability triangle," where the nodes represent actors (citizens, politicians-policymakers, and service providers4epartments) and the sides represent accountability relationships (voice, compact-management, client-power). The approach inthis paper takes these principal-agent relationships as given, and does not seek to re-run the arguments that are well set out inthe World Development Report. The approach here essentially converts these relationships into qualitative checklists or sets of incentives. Some studies are more sanguine about the impact o f local govemments. For example, inBolivia it appears that devolution has resulted inchanges in investment patterns inthe country, puttinga greater focus on education and water and sanitation, as well as to a more geographically equitable distribution o fpublic expenditure (Faguet: 2001). "Voice i s the relationship o f accountability between citizens and politicians, the range o f measures through which citizens express their preferences and influence politicians" (World Bank 2003). 22 Second, there are the incentives on those who provide the services-for example, the staff o f schools and clinics-so that they also want to see improvements. This motivation can be provided through managerial power, ensuring that the nazimeen and senior staff have authority over the front line workers or, again, through citizen power (thistime inthe form o f "clientpower") providingpotential service recipients with some direct means to have an impact on the conduct o f service providers. Unpackingthe Incentives Examining the incentives that devolution mightprovide for nazimeen and senior staff, then the provincial efficiency incentives derive fiom the structure o f the revenue base, including transfers and own source revenues available to meet the new local expenditure responsibilities. Ifwell structured, these fiscal arrangements promote the efficient use o fresources inmeeting local preferences. Since nazimeen will typically face expenditure demands that exceed available resources, budgetary certainty in the form o f predictable resource flows and a hardbudget constraint, coupled with adequate autonomy inresource allocation, will encourage them to use public resources responsibly and efficiently. With these arrangements inplace, nazimeen will more likely attempt to make the best trade-offs, for example, inallocations between sectors, between recurrent wage and development expenditures, and attaining value for money (for example, through effective procurement arrangements). As i s explored inmore detail inthe second volume (Annex I-Recent Histoly), some argue that own source revenues automatically contain more incentives for efficient use than transfers, as local taxpayers are more likely to demand results. Citizens would be able to make their demands for better services heard through the voice channel,the most obvious form o fwhich i s elections. However, voting may be based on other criteria than service delivery, as owing to years o f bitter experience voters may not see commitments on the part o fpoliticians to provide broad-based public goods over time as credible. Hence, voters and politicians may emphasize directly visible and narrowly targeted projects such as school infrastructure. These schools may not be staffed or maintained, but voters have come to know that they cannot holdpoliticians to their promisesbeyond the short term. Voice i s likely to be stronger, with councilors motivated to listen to citizens, ifthere i s public assignment o f credit or blame and ifthere are obstacles to councilors providing private goods targeted at a few influentials rather than public ones which benefit many. As an influence on councilors, voice must compete withthe incentives that the provinces can provide for nazimeen to serve their political interests. That "pull" fiom the provinces i s likely lessened when the councilors have independent political legitimacy and when provinces themselves have fewer incentives to intervene at district level. The local governments, as representedbythe nazimeen, inturnengage with service delivery providers at the local level (for example, schools and contractors) through managerialpower, their ability to require service providers to adhere to particular standards o f behavior and service. Again, as i s explored inmore detail inthe second volume (Annex I-Recent Histoly), managerial power i s strengthened ifthe staff that work inthe district are employedby the district-with hire and fire authority available to local managers, subject to suitable safeguards. Recognizing that this is unlikely to bethe case for senior staff, nazimeenneeddistinctive authority over their behavior including input into their performance evaluations, and a credible threat to transfer them iftherelationshipisnotproductive.Managerialpower is strengthenedbythe extemaloversight provided by bodies such as Accounts Committees, Monitoring Committees and District Public Safety Commissions. It also assumes that staff with appropriate skills are available in sufficient numbers and that those staff have sufficient technical sanction powers to undertake their tasks. 23 A more direct route o f accountability, clientpower, directly connects citizens with service provides, such as schools. Client power includes more formal mechanisms as functioning agencies that assist indispute resolution concerning service delivery (Departmental Grievances Redressal and Complaint Cells and Citizen Police Liaison Committees) and bodies which provide direct citizen inputsinto the management o f facilities (Citizen Community Boards and School Management Committees). Although this i s seemingly a more direct route o f accountability, as the World Development Report underscores, there may be a variety o f reasons why client power mightbe insufficient by itself. Citizens may findparticular services hard to monitor. What constitutes good health care may be difficult to assess, and require other forms o f accountability relationships (for example, oversight from health authorities or professional associations). Table 10 summarizes the variables that are fundamental to these incentives being put inplace. Table 10: Key Incentivesinthe Devolved Structures citizens Staff working inthe district are employed by the district DCO Nazimeen authority over senior Credible threat to transfer the DCO staff Nazim 's input into the performance evaluation for the DPO Districtpolicingplans approved by the DPSC Accounts Committees Monitoring Committees Extemal oversight District Public Safety Commissions Village and NeighborhoodCouncils Insaaf committees Adequate deployment of staff Adequate numbersand skills Adequatetechnical sanction powers Departmental GrievancesRedressaland Complaint Citizen dispute resolution Cells Citizen Police Liaison Committees Community management of Citizen Community Boards facilities School Management Committees 24 As will be seen, a number of crosscutting devolution changes have transformed these various incentives that are likely to affect service delivery. However, the study findings suggest that while many o f the elements for better decentralized service delivery are inplace, significant challenges for reformremains. Local governments have been provided with additionalresources under the PFC awards and their ownrevenue bases. These resource flows are proving relatively predictable, butallocations across local governments maybe insufficient andinadequately distributedacross local governments givennew expenditure assignments, and inmany cases settingpoor incentives for own source revenue mobilizations. Some o f the biggest challenges remain inthe area o f administrative decentralization. The relationships between nazimeen and core senior staff (for example, DCOs) remain complicated, as these staff are still largely beholdento the province. Various reforms to enhance citizen power have yet to prove themselves. Citizens were able to exert voice through the election o ftheir UC representatives, but nazimeen at the district and TMA level remain indirectly accountable through the system o f local councils. The Citizen Community Boards have the potential o f enhancing voice through direct involvement inthe planning and implementation o f development funds, but are confrontingmany teethingproblems. Reforms in school management committees have the potential to enhance client power. Inthejusticial sector, anumberofchangesinaccountabilityrelationships appeartobeboth strengthening and weakening the prospects o f service delivery interms o f security and access to justice, meriting special treatment. Assessing the Impact of Devolution on Service Delivery If,onbalance, these changingincentives areinfact havingsome impactonservice delivery, then, inprinciple, improvementscouldbe measureddirectly. However, there are two difficulties. First, devolution i s recent and it would be premature to attribute any change insocial-sector outcomes to new arrangements on service delivery. Second, inPakistan there is no consensus on which data sources to use to measure intermediate and outcome indicators, and it would be dangerous to judge the utilityo f devolution on the basis o f any apparent trends inthese indicators. Furthermore, there i s no obvious relationshipinthe data between intermediate indicators that can be impactedby policies-such as percentages o f schools with facilities like toilets and electricity-and outcomes. 25 Figure4: InterpretingEducationData This problemo fdata canbe seen in the discrepancies in education-and- health intermediate and outcome 90 indicators. Ineducation, the two 80 1 major sources for data are the 70 Palustan Integrated Household 60 Survey (PIHS) and the Ministry o f 50 Education (MoE). While the P M S i s 40 30 a household survey, the Ministry o f 20 Education constructs enrollment 10 numbers from (a) the Educational 0 Management Information System P MS:1998-99 PMS 2001-02 MoE 2000-01 (EMIS), which gives data only on enrollment ingovernment &IGross primenrol N e t prim enroll educational institutions and (b) estimates o f private enrollment, Figure5: InterpretingHealthData based on a census ofprivate schools conducted by the FederalBureau o f Statistics (FBS) in 1999-2000 and estimates o f enrollment in madrassahs. Given these different modes o f collection, it i s not surprising that the two sources produce very different results. Overall, the PIHS data on enrollments are considerably below that o f the Ministry o f 1999 2000 2001 2002 2003 Education-the PIHS shows net primary enrollment to be 42 percent as compared to 66 percent for the M o E (Figure 4). These discrepancies have created such controversy that the government has called into question the methodology usedinthe PIHS.'3 Inhealth, atthenationallevel, the EPIMIS(management informationsystem) andthePIHS,the two main sources for the state of immunization coverage, both show a recent improvement, but starting from very different bases (Figure 5). Data from the EPIM I S shows that immunization coverage inchildren for a number o f diseases was declining untilvery recently For example, DPT 3 immunization declined from 74 percent nationally in2000 to 69 percent in2002, before returning to 74 percent by September o f 2003. PIHS data on the other hand, shows a much lower percentage o f children immunized but also suggests a rising trend betweenthe last two sample years. l3 As a result of these controversies, the release o fthe 2001-02 report was delayed, and when it was released it included an "Explanatory Note" warning the reader that specific sections o f the report were wrong, thus calling into question the accuracy of the entire report as well as the previous PIHS rounds. While non-sampling errors are a problemin the PIHS, the M o E data are generally believed to be more problematic as the mode o f collection o f the MIS creates many sources for errors, and the estimates o f enrollment inprivate schools and madrassahs are problematic. Ingeneral, EMIS i s best suitedfor collecting important administrative information like salaries, transfers and retention o f teachers, for example, and not for tracking outcomes and performance indicators. 26 2. Assessment KEYFIND~NGS This studyproceeds from a basic understanding that two preconditions produceimprovements in service delivery. Firsti s the requirement that the incentives motivating nazimeen make them want to see such improvements and, as a corollary, that they want to see these achieved-at least in part-by the more efficient use o f existing resources. The second precondition i s that those who provide the services- for example, the staff o f schools and clinics-must also want to see improvements. Their motivation i s a response to managerial power-ensuring that the nazimeen and senior staff have authority over the front line workers+r, again, to citizen power that gives potential service recipients some direct means to guide the conduct o f service providers. Figure 1 earlier set out the assumed chain o f connections between the new incentives and the likelihood o f improved services. This section synthesizes the discussion o f these changed incentives set out in detail inthe second volume o f this report (Annex 2-Recent Histoly). New Incentives for Nazimeen and Senior Staff Provincial efficiency incentives The study assumes that devolution would be providing District and Tehsil Nazimeen with greater incentives for efficiency infiscal management ifit achievedpositive movement inthree key categories: 0 Budgetary certainty-the degree to which they must live withinhard constraints and the predictability o fthe budgetary transfers that they depend on; 0 Autonomy inpreparing the budget-specifically the development, the non-salary recurrent and the salary budgets; 0 Incentives for local revenue-raising-the buoyancy and potency o f taxes assigned, the clarity o f revenue assignments and other incentives for revenue raising. Onbudgetary certainty, as Table 11shows, most fundingfor local government i s ineffect negotiatedrather than formula-based. All negotiated revenues provide opportunities for gamesmanship, basing current policy decisions on their likely impact on future resource transfers, rather than on their merits. The proportions o f the Provincial Consolidated Fundthat ultimately are transferred on formula basis are trivial (less than 8 percent inall provinces). However, the more significant measure i s the proportion o f transfers that are, or should be, transferred to local government that are transferred on a formula basis. It i s important to note that transfer has a precise meaning and refers to funds that are non-lapsing and within the re-appropriation authority o f the district or TMA. On this basis, the formula-based proportion o f de facto transfers (transfers made inpractice) ranges from 71 percent (NWFP) to 17percent Punjab. However, again this i s a somewhat misleading figure, as NWFP and Sindh do not, formally, transfer funds for salaries as these remain inAccount 1. Ifthe proportion i s taken from dejure transfers-the transfers mandatedby the LGOs-then all provinces come inat less than 25 percent. 27 Table 11: Proportions of Total Funding Transferred to Local Governments on a Formula Basis (Rs.million) es (excluding own Formula basedtransfers Note Formula-basedtransfersare shaded. a. For Sindh andNWFP, GSTwas paidout to districts, TMAs andUAs first to meetbaseline shares andthen additional funds were distributedto new units.Punjabpaid GSTto TMAs andUAs alone, first to meet baseline shares andthen onpopulationbasisto new units. InBalochistanGSTwas distributedon formulabasisandwas usedfor financinglocal developmentas well. No transfers for capitalbudgetwere made otherwise. b. De facto transfers are definedas paymentsfrom provinces to local government that are madeto the account of the district or the TMA-that is, they are genuine transfers as they are non-lapsingandwithin the re-appropriation authonty of the district or TMA. Incase of Sindh the de facto transfers includedevelopment funds, althoughthey were made available throughAccount I,becausebudgetary autonomywas givento LGs. c. Dejure transfers are those that are mandatedby the LGO. For NWFP and Sindh, dejure transfers are considerably larger than de facto as salariesshould betransferred via Account I V as specifiedinthe LGO, but inreality are not d Punjab transferred Rs. 50,859 inrevenuesto distncts as lump sums without specifying salary, non-salaryand developmentallocations.Non-salaryanddevelopmentallocationswere formula-based district shareswhile the salary transfers were in accordancewith histoncalbaseline, althoughinsome cases negotiatedincreaseswere given. e No disaggregation i s possible from the budgetdata, however, the DSD2 shows that Rs. 42,600 were bookedas salary and non-salaryexpenditure inFY 2003 inthe 34 districts andhence the development expenditurescanbe calculated. Source DSD2, basedon 2002103 revisedbudget estimates. While these transfers are negotiated, the data also reveals that there i s considerable predictability inthe transfers at the district level, with inmost cases districts receivingonly slightly less than what was budgeted. The district studies also revealed that local governments have limited autonomy inpreparingtheir budgets. Vertical programs are the main constraint on district and TMA autonomy inpreparing the development budget, with often the volume o f funds channeled through vertical programs dwarfing the district's own ADP for particular sectors.14Inthe study districts, more than half o f the ADP i s under the effective control o f the federal and provincial agencies controlling the vertical programs (Table 12). This practice undermines the sovereignty o f local governments in planning the development o f their districts according to their own priorities. l4 See Table 19 below for a summary o f the institutional impact ofvertical programs. 28 Table 12: Vertical Programs as a Percentage of the District ADP 2001-02 2002-03 Total VP Recurrent District Total VP Recurrent District Outlay VP as YO Budget ADP Outlay VP as YO District Budget (Rs.) ADP (Rs.) (Rs.) of ADP (Rs.) (Rs.) (Rs.) o f A D P Killa Saifullah 202.42 - - - 167.86 71.79 23.42 32.60 Karachi 7275.90 - - - 11631.19 - - - Bannu 520.30 131.90 51.12 38.80 586.94 189.73 - - Bahawalpur 1368.92 239.01 177.78 74.40 1486.22 462.89 281.67 60.90 Khairpur 1637.95 293.52 145.86 49.70 1637.95 192.00 - - Faisalabad 2630.75 164.93 150.00 90.90 2630.75 710.00 358.70 50.50 Local governments also have limited freedom inpreparingthe salary budget. Neither local nor provincial governments are, inpractice, able to determine pay scales, which are set at the federal level. Inaddition, in Sindh and NWFP, salaries are transferred through provincial Account I rather than the district-controlledAccount IV. The districts must seek approval from the province for the release o fthe funds from Account Ifor their salary payments. Thus, the salaries effectively are still paid from the provincial budget. Local governments are hugely dependent on intergovernmentalfiscal transfers as taxes assigned to local governments lack sufficient revenue potential. This degree o f vertical imbalance can lead to inefficiencies inboth the levels and quality o f public expenditure-specifically, it creates incentives for local governments to overspend as well as reduces the incentives to spendthese resources efficiently. The logic inboth cases is that intergovernmental transfers lower the costs to local residents o f getting services. Arguably, the disconnect between expenditure and taxation encourages "free-riding," as well as reducing the incentives to monitor public expenditures, as resources are collected from distant taxpayers not likely to be usingthe services produced by the local government. Thus it is at least arguable that localrevenues contain intrinsic incentives for efficiency, as local taxpayers mightbe more vigilant than distant ones and so efforts to improve own source revenues can imply incentives for efficiency. Incentives for local revenue effort are linked to the buoyancy and potency o ftaxes assigned to local government, and the clarity o f such assignments. Effort will be deterred ifthere is no credible connectionto likely revenue results and ifit is not evident which level o f government i s to benefit from which tax. While districts have been given rights to raise additionalrevenues, their tax base i s weak, and some district taxes and user charges are very difficult to impose inpractice. Inthe six study districts, for example, for 2003/04 own source tax revenues ranged from 0 percent to 8 percent o f total revenues. Therefore, a large tax effort brings only modest results intotal revenue. Inaddition, although the district andTMA taxes are clearly assignedintheory, administrative confusions inthe collection arrangements blur the assignments inpractice, further reducing incentives to increase collections. Administrative complexity i s most clearly evident inthe case o f property tax (UIPT).InSindh, for example, where the province has control over various aspects o f tax policy, collection i s meant to be the responsibility o f the district, while TMAs determine 29 rating areas. This sharing o f responsibility i s likely to reduce the potential for a revenue source that could make up a significant proportiono f the fiscal gap inthe TMA budgets. CitizenPower (Voice) This study argues that there are two necessary conditions for the local political leadershipto have incentives for service delivery improvements. First, local councilors themselves must have an interest in serving the needs o f their constituents and inparticular the poor. Second, since the relationshipbetween nuzimeen and councilors i s a critical intermediate step inthe linkage to citizens, nuzimeen must find the pull from the councilors more demanding than the pull from other political powerbrokers, particularly the provincial governments. On the first point, the study notes that while it is probably true that citizens value education and health services, it i s also likely that voters believe politicians' claims to deliver patronage and infi-astructureschemes are more credible than promises to improve health and education. Inthe context o f devolution, jurisdictional overlap makes it particularly difficult for politicians to seek credit for improved services. When multiple levels o f government are involved indelivering the same services, the public i s unable to assign credit or blame effectively for that particular service to a particular politician. The overlap thus weakens incentives to perform well on service delivery and encourages politicians to target services to their core supporters. Jurisdictional overlap i s particularly an issue inthe education and health sectors, as federal andprovincial vertical programs remain principal tools inpromoting national policy priorities and represent a significant proportion o f local-government expenditures on service delivery. The anecdotal evidence from the study districts also suggests the strong incentives working on councilors to focus on private goods. Inall the districts visited, The District and Tehsil Councils hadreserved significant portions o ftheir development budgets to be distributed equally among theirrespective councilors. The councilors inturnhaveusedthese allocations mostly for neighborhood-specific schemes, such as water and sewerage network extensions. l5This focus on union-specific development has meant that union nuzimeen are not paying adequate attention to their other important role as district council members. The important council budget sessions were largely perfunctory, with budgets beingpassed ina matter o f minutes, suggesting that I district-wide issues hadnot received the attention that they deserve. Monitoring Committees, while formed, are generally not functional, inpart because councilors lack capacity and generally deal with a hstratingly non-responsive bureaucracy, but also because councilors' priorities lie elsewhere, and their attention i s focused on their particular constituency. While these incentives to focus on private goods are strong, the study did not find that devolution had exacerbated this problem. For example, the evidence inFigure 6 shows no obvious reduction inoperations andmaintenance expenditures betweenthe allocations in2003/04, the first year local governments prepared their own budgets, and that o f earlier years. The argument in favor o f continuedearmarlung-that non-salary budgets will be raided to finance new, greenfield projects-therefore does not appear justified. Similarly, the data suggests that there has not been a significant shift from large to small development schemes, or that councilors are abandoning existing schemes to gain political glory by starting new ones. l5 As was noted earlier, this paper refers to private goods to include both pure private goods (that is, individual- specific) as well as narrow public goods (for example, locality-specific projects). Local infrastructure projects often encompassboth- that is, only the local community benefits, and the contractors for the projects reap the private rewards. 30 Figure 6: Non-salary as a Percentage of Total Expenditure (districts) 20% 18% 16% 14% 12% DBahawalpur 10% Faisalabad 8Yo HKhairpur 6% 0Karachi 4% 2% 0Yo 200112 200213 200314 Source: ADB/DfID/WB D S D Dataset 2-fiom 2001/02,2002/03 revised estimates and 2003104 budget estimates. Onthe second point, the study found that although elected by them, nazimeen have few incentives to take notice o fthe district or TMA councilors. They feel some needto maintain their support, as indicated by the equal allocation o f development funds to each councilor. However, it i s strikingthat n o district nazim faced effective opposition inany o f the districts visited, inlarge part because the electoral college i s often more dependent on the nazimeen than vice versa. Onthe other hand, a clear connection that emerges from the district visits is the importance o f the political relationship between the nazim and the rulingprovincial political coalition. Effectively, thisrelationship determines the extent ofpolitical control that the DistrictNazim can exercise within local government as well as the freedomlocalpolicymakers haveto act to meet their constituents' demands. Intwo o f the six districts visited-Bahawalpur and Faisalabad-the district nazim i s allied to the ChiefMinister and the ruling provincial political coalition, while in the other four the district nazim i s inopposition. This relationship has a significant impact on the flow o f funds between the province and local governments, on the implementation o f local development schemes and on the manipulation o f appointments and transfers o fpersonnel. New Incentives for the Service Providers For service delivery to improve, the service providers-and notjust their elected superiors-must want to see improvements. This motivation i s provided either by ensuring that the service providers have an incentive to listento the policymakers-managerial power-or byproviding potential service recipients with some direct means to have an impact on the conduct o f service providers-client power inthe terminology o f the World Development Report (World Bank: 2003~). Managerial power over serviceproviders Inreviewingthe changed incentives for front line staff and other service providers to "listen to nazimeen," the study examines three key dimensions: 0 The extent to which staff working inthe district are employedby the district 0 Adequacy o f the nazim's authority over senior staff, including inputs into the performance evaluations for senior staff and ability to require their transfer. 0 Extento f external oversight, including a review o f the effectiveness ofbodies, such as Accounts Committees, Monitoring Committees, and District Public Safety Commissions, which are mechanisms for the monitoring o f front line staff. 31 The overall conclusion from the district studies i s that administrative devolution has beenpartial and, as yet, given district staff few reasons to respond to the authority o fthe nazimeen. The transfer o f functions and the physical relocation o f staff have not been fully matched by the transfer o f the employer role. Local governments have limited ability to create fiscal space by dismissingsurplus staff, and intwo o fthe four provinces (Sindh andNWFP), salaries continue to bepaidfrom Account 1.Inother cases, while the formal authority rests with the local govemment, the province continues to exercise informal control. While the LGOprovides for a sophisticated sharing o f control over posting o f civil servants basedon norms which clearly delineate responsibilities, informal provincial controls are clearly evident inthe area o f Appointment, Promotion and Transfer (APT) authorities. Inmany cases, provinces have reduced the control o f local political executives over their civil service through unilateral orders o f postings. The result i s that although physically locatedinthe districts, most senior district staff do not consider themselves as district employees and therefore are likely to accommodate provincial pressures to transfer subordinate staff. This attitude has considerably weakened the accountability o f senior staff to local elected representatives. Infact, politically motivated inter- and intra- district staff transfers are a main area o f conflict betweenprovincial and local governments and are most severe indistricts that are not politically aligned with the province, such as Bannu, Khairpur, and Karachi. Since TMA staff largely belong to local govemment cadres or are the council's own employees, the problem o f low institutional allegiance and weak accountability i s less acute. These problems o f accountability were clearly evident inthe relationshipbetweenthe district nazim and the senior staffinthe district, particularly the DCO and the DPO. It was generally observed that the authority o f the nuzim to write the ACRs o f DCOs was an insufficient mechanismfor enforcing accountability since it was an ineffective tool for measuring the performance o f an officer and tended to be viewed as a routine matter to be placed on an officer's file at the time o fpromotion infulfilling a legal or administrative requirement. Officers tended to be given postings regardless o f their performance inprevious tenures. Nuzimeen also have very little authority over the DPO. Formally, only inPunjab has the District Nuzim been given the initiating authority for the ACR o f the DPO, although it has been reportedthat this authority will be extended to the nuzim inall provinces. However, inpractice innone o f the districts studied were these performance evaluation reports completed inthis manner. Some nuzimeen, infact, considered?hat removing the authority o fthe former DistrictMagistrate incompleting the ACR o f a former district Superintendent o f Police has contributed to deterioration inlaw and order in general. l6 The LGOprovides for anumber o f institutional arrangements-such as Monitoring Committees, and District Public Safety Commissions-for the legislative oversight o f service providers. However, it was generally observed that these mechanisms have either not been established, or have been introducedto only a limited degree. For example, the field visits revealed that, with some notable exceptions and for several reasons, Monitoring Committees were not performing as the regulations envisioned. Similarly, Public Safety Commissions where formed, were generally unable to elicit compliance from the police inits investigations. Pressreports of ahigh levelmeetingheld on 14April 2004 involving the PrimeMinister, Provincial Chief Ministers, the NRBandother federal andprovinceofficials indicatethat significantchangeswill be introducedto the PoliceOrder2002 andthat anew target date ofDecember2004hasbeenset for complete implementationof these new arrangements.Untilthese are officially confirmed and clarified, this report will makeonly passing referenceto what has beenreported, but will not speculateon their likely impact. 32 Citizenpower (client power) As the WDR argues, service delivery can also be improvedby improving client power, that is, strengthening the direct linkage between citizens and service providers. Inprinciple, such an approach can be more effective than focusing on efforts to make service providers more accountable to policymakers and then making policymakers, inturn, more accountable to citizens. The LGO and Police Order call for the creation o f various bodies, such as Citizen Community Boards and Citizen Police Liaison Committees, which can serve to improve this client power. The overall conclusion from the field visits i s that it i s too early to tell whether or not devolution has had any impact on client power for the better or the worse. While in general participatory mechanisms, such as Citizen Community Boards and School Management Committees (a pre- devolution initiative) have yet to achieve their full potential, there are numerous examples that suggest that innovative local leadership has had a positive impact on community mobilization and inturnon service delivery. THEIMPACTOFTHE CHANGEDINCENTIVESON SERVICEDELIVERY: A PRELIMINARY ASSESSMENT As discussed above, it is too early to determine the effect o f devolution on service delivery outcomes. This section therefore focuses on the effect o f the changed incentives for nazimeen and service providers on the four key sectors impactedby devolution-health, education, water and sanitation, and access to justice. Health Thestructural changes The overall picture for the health sector under devolution i s encouraging. The transition was relatively smooth and the transfer of staff and facilities was accomplished without major disruption to services. Some districts took advantage o f the changes to restructure more comprehensively, and devolved procurement seems to be workmg well. The post o f executive District Officer Health (EDO-H)has been created and i s filled inall districts. All the health facilities, including district headquarters hospitals butnot tertiary (teaching) hospitals, have been placed under this office. Training and staff development institutions that were earlier under the divisional headquarters are now largely under the EDO- H." However, only inPunjabare training institutionsfully devolved (except medical colleges which remain with the provincial government). Insome provinces, like NWFP and Punjab, this transfer provided an opportunity to carry out an extensive rationalizationo f the health facilities and to obtain data on utilization, citizen access and missingstaff. Different provincial requirements have produced a number o f structural models for health district departments. Some predate devolution and were part o f the previous district decentralization plan. InNWFP,for example, there are three district types identified onthe basis o f size, and structures are tailored to these different types. Thus the number o fposts and the specialties within a district depend on the type o f district. Type A, Peshawar, i s the largest district with a wide range o f posts, departments and specialties. InType B, medium-size districts, the number o f specialties and l7 There are exceptions, for example, inKilla Saifullah and Bannu, where inthe earlier district there were no such institutions and the district health development center and nursing schools remain under the direct control of the provincial health services academy. 33 departments decline, and smaller Type C districts that do not cover the full range o f health care rely more on provincial expertise. The procurement o f medicines and purchase o f biomedical equipment have been devolvedto the districts, and purchase committees headed by the District Coordination Officer (DCO) have been constituted. The committees are responsible for technical scrutiny, standardization and purchase o f medicines, capital goods and medical equipment. For Balochistan, NWFP and Punjab, a provincial committee performsprequalification, and a system o frate contract eliminates the role o f intermediaries. InSindh, the absence o f a system for rate contract or prequalificationby the provincial government has ledto complaints relating to the purchase o f medicines. Additionally, inKarachi Tehsilresponsibilityfor healthhasbeenrelinquishedto the city district, anddifferent purchase procedures are ineffect for health facilities that hadbeen municipalrather than provincial units. Some o f the outstanding issues inharmonizingprocurementpractices and rules are still not resolved as the Rules o f Business are not yet inplace. As argued earlier, while the anecdotal evidence suggests that there are strong incentives on local politicians to focus on private goods, the study did not find that devolution had exacerbated the problem, as evidenced by the protection o f operations and maintenance allocations inthe district budgets. Specifically inthe health sector, the picture that i s emerging suggests that the new powers to procure are being used well, and citizens are increasingly finding that health facilities have medicines available for them. Not only i s the supply picture undoubtedly improved, but it i s also noteworthy that the press has publicized few reports o f corruption incontracting for medicines. Citizenpower (voice) While ingeneral Monitoring Committees have failed to achieve their potential, insome districts, notably Karachi, Monitoring Committees were functional, inlarge part due to the more politically aware and educatedbody o f councilors. Reports o f these committees were submitted regularly, and the health committee claimed to have increased the attendance o f staff by 20 percent, improved the quality o f maintenance, and repair, a claim that was generally supported by NGOs andjournalists. Similarly, the city health committee stated that it had visited 25 hospitals inthe last year to ensure quality o f service and reduce absenteeism. The committee members believe that this improvement has been achieved because o f their consistent monitoring and not because o fthe respective EDO's interest inthe improvement o f services. Evenhere though, there i s considerable frustration with the bureaucracy, as exemplified by the fact that the District Council has passed over 300 resolutions, many o f them calling for disciplinary actions against particular staff, but without muchresult. Provincial efficiency incentives Devolution has not yet produced district ownership and the consequential district oversight over vertical programs (Box 2). As outlined inmore detail inthe second volume (Annex I-Recent Histoly), vertical programs are ingeneral problematic because they limit local government's autonomy inpreparing the budget, and limit public assignment o f credit and blame, thus distorting relations betweenlocal service providers and councilors by reinforcing vertical relations with higher governments. This lack o f local ownership i s most marked inrelation to the Lady Health Workers, but similar problems o f delays inthe release o f funds from the province and weak supervision and monitoring by the districts beset the EPI, MCP, and TB DOTS programs. While lack o f capacity and training are clearly important factors, lack o f full district ownership i s a bindingconstraint. Districts do not allocate funds or other resources for monitoring and supervision, and without this involvement program targets are unlikely to be met. 34 Box 2: Vertical Programs-Health The National Program for Family Planning and Primary Health Care, better known as the Lady Health Worker (LHW) program i s the mainfederal program, w h l e others, such as the Extended Program for Immunization (EPI), the Malaria Control Program, and TB DOTS,are primarilyprovincial programs with varying degrees o f district involvement. The LHW program i s the government's flagship health program as well as a major focus o f donor interventions, primarilybecause it has developed a reputation for overall effectiveness. The program i s highly vertical, inthe sense that the most important policy decisions, such as the number o f LHWs to be employed, as well as financing, rest with the federal government. Onthe other hand, key operational responsibilities, such as decisions on the allocation o f LHWs within a district, as well as overall recruitment and supervision o f the LHWs, are district functions. Currently, the program employs 70,000 LHWs, with an I-PRSP target o f 100,000 by 2004105. A DflD-funded evaluation is one o fthe few data sources available on the program (Oxford Policy Management: 2002) although this was undertakenprior to devolution. The results o f the study are mixed. Interms o foutcomes, the studyrevealsthat use ofprenatal care anddelivery services, use o f contraceptives, immunization coverage and prevention and treatment o f common illnesses inchildren under five are all better inLHW areas than incontrol groups. For example, it found 56 percent full immunization coverage inLHW areas as compared to 38 percent incontrol areas, and 20 percent modem CPR inLHW areas vs. 14 percent incontrol areas. For the most part, this difference i s due to the fact that LHWs tend to serve more economically developed areas.' Once socioeconomic variables are statistically controlled for, L H W s impact i s positive but muchmore moderate.2 Interms o fprocesses, the studyfinds that ingeneralLHWsare recruited according to the criteria establishedby the federal government and are also adequately trained. However, there were major problems with the timely payment o f LHW salaries and serious undersupplies o f medicines and contraceptives. For example, only a third o f the LHWs surveyed hadbeenpaidinthe month preceding the survey; a t h d hadnot beenpaidfor four months or more; and more than 20 percent o f L H W s hadrunout o f essential medical supplies inthe last three months. Despite their extensive training, the knowledge o f the L H W s was limited, especially incurative care. Fewer than 25 percent knew the correct dose o f antibiotic neededto treat pneumonia. Supervision was weak, primarilybecause supervisors lacked adequate access to vehicles, but this weak supervision had more to do with quality than frequency, as 70 percent o f the LHWs hadbeen supervised inthe last month, and 82 percent inthe last two months. Some o fthe mainproblems with the program-delayed release o f salaries, poor supervision-have primarilyto do withits vertical nature. Release procedures from the federal level to the LHWs inthe case o f salaries and to the districts inthe case o f non-salary inputs are cumbersome and need streamlining. As the study revealed, LHW performance is closely linked to the quality o f supervision, which inturn implies that the role o f district governments is vitally important. However, since political leadership inthe districts is largely unaware o f the program, a crucial incentive for improving the quality o f supervision is lackmg. 1.This is inpart due to the federally-determined LHW selection criteria a) stipulating that LHWs be relatively educated women and b) requiring a hnctioning health facility inthe area from which the L H W s are to be recruited. 2. Specifically, the presence o f L H W s increases the probability o f a child's being vaccinated and o f the use o f prenatal care during pregnancy. However, LHWs appear to have no impact o n contraceptive use once socioeconomic factors are controlled for. 35 Managerialpower Local ownership o f the reforms remains very limited. Table 13 documents the absence o f organizational structure and some keyjob descriptions for district health offices infour o f the six study districts. Implementationplans, though, were available infour o fthe districts. Table 13: Ownershipof the StructuralReforms Source; ADBIDflDIWB DSD Dataset 1. Inaddition, personnel anomalies have developed inthe processofallocating staffto thenew district structures. Inparticular, former hierarchicalrelationships have become distorted. For example, the doctor incharge o f the district headquarters hospital has been placedunder the supervision o f the ED0 (H),whereas previously the District Health Officer and the DHQdoctors would all report to the Divisional Director Health Services. Also, the Medical Superintendent o f the hospital, a grade 18 employee, supervises doctors ingrades 19 and 20. Box 3: CitizensUsingPrivateMedicalCare An emergingpicture inthe 6 districts, one borne out by the CIET social audit reports, shows more Killa Saifullah 29 percent and more citizens migrating from the public to the Karachi 75 percent private sector for health care (Box 3). Predating Khairpur 81 percent Bahawalpur 55 percent devolution, this movement demonstrates Faisalabad 37 percent dissatisfaction with existing public sector health Bannu 21percent provision at the local level and presents a challenge for district health departments. As the Source: (CIET: 2003). private provision o f health grows, district health departments must provide not only government services but oversight and regulation o f the private sector as well. Currently, most such departments maintain no record o f private sector services, and no legal registration requirements exist except inone district.** Providing regulatory oversight-a key future task for district health departments-will require new legal frameworks and effective monitoring by trained district health teams. The issue, however, does not yet seem to be on district planning agendas. In sum The key impacts of the changed incentives inthe health sector are as follows: Citizen power (voice) o While ingeneral MonitoringCommitteeshave failed to achieve their potential, there are isolatedexamples o f these committees reducing staff absenteeismand improving operations and maintenance Provincial efficiency incentives Bannuwas the only district where all clinics and hospitalswere registeredandpaid fees that, while collectedby the district, were submittedto the provincial govemment. 36 o There is complete lack o f district ownership and corresponding lack of disttlct oversight over vertical programs. 0 Managerial power o Local staff ownership o fthe reforms is, as yet, very limited o Reporting lines and managerialrelationships are not quite finalized o Districts are unprepared for regulation o fthe growingprivate sector Education Thestructural changes Ineducation, thenew district structures have evolved hrther andare more settledthaninthe health sector. T h i s i s probably because, prior to devolution, provincial education departments were already a long way down the road o f decentralization and appropriate management structures hadbeen established at the sub-provincial level. The sector has seen considerable experimentationat the district level-thus, although inall districts the basic structure o f the District EducationOffice i s inplace, below that level a variety o f education department structures are emerging to meet district-specific circumstances (see Table 14). Important variations among provinces and districts characterize the devolution o f education management. The structure of the Departments o f Education inthe Punjab, Sindh and NWFP include three directorates-Primary, Secondary and High-while Balochistaninitially merged them into one but subsequently created a separate directorate for colleges. At the district level, Punjab has created separate departments for education and literacy, while the other provinces have opted for single departments for both sectors. InKarachi, the district govemment has not yet achieved uniformity inmanaging its own schools and those that were formerly under the tehsils. Procedures for tehsil schools-such as procurement rules-are still inplace and so although schools are all now nominally under the Karachi EducationDepartment, different rules and procedures allow former tehsil schools latitude denied to district schools. These parallel service delivery arrangements have resulted injurisdictional conflicts and overlapping and confusion among the concerned stakeholders inlocal govemment. 37 Table 14: Management Structure after Devolution Balochistan Sindh NWFP Punjab Provincial Departmentfor level Educationanda Education Directoratefor and Literacy Literary and Departments Non-formal Education Provincial 1.Primary 1.Primary 1.Primary 1.Primary level 2.Secondary 2.Secondarv ~~econdary Directorates 3.High 3.High 3.High District Killa Saifullah- Karachi- Bannu-joint Faisalabad- Bahawalpur- Education joint department separate department separate separate and Literacy departments iepartments departments Departments District DEO-Primary DEO-EE (M) DO-EE (M) DEO-EE(M) DEO-EE(M) Education DEO-Sec. DEO-EE (F) DO-EE (F) DEO-EE(F) DEO-EE (F) 0fficers- DE0 Admin. DEO-Sec. DO-Sec. DEO-Sec.(M) DEO-Sec. DE0 DEO-HS DEO-College DEO-Sec.(F) DEO-College DEO-Acad. DO-Special DEO-College DEO-Special DEO-Tech. Education DEO-Special Education DEO-Special DO-Sports Education DEO-Sports Education DEO-Sports DEO-Sports DO-SEMIS and Planning Deputy DDO DDO (M) DDEOs at DDEOsattehsil District DDO (F) tehsil level level Education 0fficers- DDEO Assistant ADO ADDO-E AEOs at AEOs at markaz District ADDO-S markaz level level Office- AEO/ADO Source: ADBiC D i m DSDDataset 1. Instaffing the newstructures, PunjabandNWFP districts are typical inthat District Education Officers have been appointed for Secondary Education, Male Elementary Education, Female Elementary Education, Colleges, Special Education and Sports under the ED0 (Education). A gender consequence o f devolution, most clearly observable inNWFP, i s that the district structure reduced the management posts available for women education officers. Prior to devolution, there were both women and men district education officers. Following devolution this post has been merged and, inthe majority o f cases, filled by men. 38 Structural Punjab Sindh NWFP Balochistan features Membership 11 14: parents to be 8: parents to be 7 elected elected Chairperson Head teacher Head teacher and Parent Head teacher parent Political No union Yes: union No representation I Yes: councilor 1 councilor representation Authority to hire I No I Yes I No I No and fire teachers Authority for IINo 1 Yes I No I N o school construction FY04 budgetary Rs.450 million Rs.247 million None allocation Provinces also differ intheir arrangements for school-based management (Table 15). School Management Committees predate devolution but have come to be associated with it, and inrecent years have beenrevived, largely because o f donor engagement at the provincial level (Box 4)." Sindh has gone the furthest, with SMCs beinggiven powers, inaddition to those for maintenance and repair and procuremento f school materials, to recruit primary teachers on contract and to undertake construction o f shelterless schools.'' Sindh i s also the only province that has formally incorporated the local political leadership within SMCs and therefore given councilors a direct stake inthe affairs o fthe local schools. I Box 4: School-Based Management I School-based management was a key feature o f the Social Action Program, and School Management Committees and Parent Teacher Associations (PTAs) were established in 1994 inall provinces as part o f this project. These committees were to performcertain school-based functions, such as procurement o f teaching materials, repair and maintenance o f the school facility, increasing enrollment, reducing teacher absenteeism and raising funds or generating other resources for school improvement. Overall, studies reveal that these committees failed to meet these objectives.' Ofthe many reasons for this failure, the most important was a general lack o f understanding about the complexity o f achieving true community-based management inrural societies marked by considerable socioeconomic inequality. Most PTAs were formed by a process o f selection rather than election, with the head teacher determining membershlp and controlling the bodies. Parents ingeneral lacked interest or were unaware o f their roles and responsibilities. Evenmotivated ones hadno formal mechanism for disciplining teachers, and were limited to sending complaints to line-department officials who inturnhad little incentive to act on these complaints. The PTAs also lacked adequate funding. One survey revealed that inone district about a quarter o f SMCs had received funds since their formation, while inthe other district none hadreceived any money since they Names vary with the provinces-SMCs inSindh, School Councils inPunjab, ParentTeacherAssociations (PTAs) inNWFP, ParentTeacher School Management Committees (PTSMCs) inBalochistan. 2o SindhDepartment of Education, notificationno. PAIASIMISC-III2002, dated October 13,2002. This notification states that only female teachers will be appointed to primary schools and that teachers' salarieswill be fixed at a rate of Rs. 40 per monthper student, but may not exceed Rs. 2,000 per month. 39 were formed (Department o f Education and Multi-Donor Support Unit: 1998). Moreover, the funds were delayed, primarily due to complicated release procedures that released funds first into the PLAs o f the then DeputyCommissioners and then onwardto the SMC accounts. The survey also found that ina number o f cases committee members, particularly co-signatories (who were to be nonteaching members o f the SMCs) on SMC accounts, were relatives o f the school teachers or head teacher, ties which raised questions about corruption and greatly reduced the SMC's ability to monitor school staff. As a result o f these problems, the SMCs/PTAs became defunct. In 1999 all government fundingwas halted. 1. For example, according to one survey o f SMCs intwo districts o f Punjab, inmany cases these committees existed only inname, with roughly two-thirds holding either no meeting or fewer than 5 meetings inthe past year (Department o f Educationand Multi-Donor Support Unit: 1998).Other surveys reveal similar results. For example, the Pakistan Rural Household Survey, 2001 (conducted by the World Bank) showed that while 52 percent o f the sampled schools possessed a PTNSMC, only 68 percent o f these had received government funds, and a quarter had not met withthe head teacher inthe past month. Citizenpower (client power) SMCs operate with varying degrees o f effectiveness. Inmost cases, head teachers still largely control SMCs, continue to select members, and school management remains de facto with the staff. Most SMC members, moreover, know little ifanything about their roles and responsibilities. For example, a majority o f SMC members inSindh hadnever receivedthe notifications delineatingtheir powers, and inPunjab it was observed that School Council members hadnever seen the proceedings register designed for supervision and management by the members (Indus Resource Center: 2003; World Bank: 2002b). Incases inPunjab where teachers were not from the same village, there was little interactionbetween them and the committee members, a fact that i s not surprising since SMCs are not involved inhiring and firing (World Bank: 2002b). Fear o f audit objections i s also resulting inSMC funds being underspent. This is primarily because school managershavenot been issuedwritten instructions andrules o f business inrelation to the use o f funds. InBalochistan, for example, over 1,000 audit objections are currently pending against school heads. Itis too early to say whether devolution has had any impact one way or the other on SMCs. On the one hand, there are many anecdotal reports o f SMCs being subjected to harmful interference bydistrict governments ingeneral, andunion nazimeen inparticular, particularly withregards to the use o f SMC funds (IndusResource Center: 2003; World Bank: 2002b). It appears that in many districts o f Sindh, locally elected officials were pushingfor union nazimeen to be chairpersons o f SMCs, or were pressuring school headmasters to make purchases from certain dealers.21 Inother cases, activeencouragementofSMCsbydistrictandunionnazimeenhasresultedin significant improvements in school quality. The example o f the Khairpur Shelterless school i s a case inpoint: innovative district and union leadership has resultedinthe constructiono f cheaper and higher quality schools. Similarly, indistrict Rahim Yar Khan, the National Rural Support Program, in collaborationwith the district government, has attempted to revitalize SMCs inone particular union council, and preliminary results show a positive impact on enrollments. The district government has ensured that where this experiment i s talung place, the teachers postedin the relevant areas are all from the same village, an expedient that increases the communities' ability to discipline nonperforming staff. Inlarge part, the district government has been able to achieve this reallocation o f the teaching staff because o f its close alliance with the provincial government. 21 The notification clearly specifies that chairpersonshave to be membersof the community. Union councilors are to be members ofthe SMC. 40 Provincial efficiency incentives Vertical programs in education also pose a challenge for district education planning (Box 5). Districts have no freedom or flexibility to use these resources according to their own priorities, and no additional fundingto support the recurrent cost implications o f vertical programs i s available. The literacy department inKarachi, for instance, receives its funding through a pipeline directly from the provincial government. This support, however, effectively frees the literacy department from the control o f the district government under which it i s meant to work and puts it under provincial control. The ED0 Literacy establishes schools, which are often runby NGOs. The education and literacy departments do not work together with the result that schools are opened by one agency without taking account o f what the other i s doing. Some areas have 5 or 6 schools within a small radius, while other areas have none. Box 5: Vertical Programs in Education The EducationSector Reformprogram (ESR) (2001-2005) i s the lynchpin o f the federal government's reform strategy ineducation. It carries a projected cost o f Rs. 100billion over and above current annual national expenditures o n education. I t s largest component aims at rehabilitation o fprimary schools; other priorities include technical education, higher education, literacy, quality assurance, promoting public- private partnerships and mainstreaming madrassahs. The federal government determines not only the size o f the program, but also the specific allocation to be made for each area o f emphasis. Funds go to the provinces as development grants and then onwards to district Account IV.Inthe past, inan apparent attempt to prevent re-appropriation, these funds were then immediately shified from Account IV into the PLAs o f DCOs. The largest component o f ESR i s the rehabilitationo f facilities inexistingprimary schools. While districts are to identifythe beneficiaries, it is possible that political pressure will cause MPAs to be involved as well. Since the operational and maintenance responsibilities for these schemes rest with the districts, ifthe districts do not fully own the projects, it i s not clear that they will accord the operations a highpriority. These problems o f district ownership worsen when some ESR programs bypass the districts altogether. For example, the literacy stream and the early childhood education program are being managed at the district level by the ED0 Literacy through post-devolution, separate Literacy Departments rather than, as before, withinthe EducationDepartment. For these programs, Federal funds go to the provincial government, which channels them onward-without any involvement o f the district government-to the ED0Literacy, an office supposedly subordinate to district officials. Managerial power A key challenge for the districts i s to put inplace adequate monitoring and oversight mechanisms inrespect ofthe private sector, whilst at the same time not discouraging the growtho fthis increasingly important profit and nonprofit sector. The private sector i s o f increasingimportance inthe urbanareas ofPakistan, butis insignificantinruralSindhandBalochistanwhere private schools account for only 3 percent and 1percent o f overall enrollment, respectively. InPunjab and NWFP the private sector accounts for 42 percent and 59 percent, respectively, o f schools in rural areas (Federal Bureau o f Statistics: 2000). The growth o freligious schools i s particularly pronounced inBalochistanand NWFP. Inall provinces, inlower-income groups enrollment in NGOschools has increased-between 1991 and 1995-96, enrollment inprivate schools increased by 70 percent as opposed to 4 percent inpublic schools.** 22 PIHS RoundIand Round 11. 41 I n sum The key impacts o f the changedincentives inthe education sector are as follows: 0 Citizen power (client power) o While it is too early to tell whether devolution has had any impact one way or the other on SMCs, there are cases o f innovative leadership resultingin significant improvementsin school quality 0 Provincial efficiency incentives o Vertical programs are creating problemsindistrict educationplanning 0 Managerial power o Effectiveregulation ofthe increasingnumbers o fprivate schools and madrassahs remains a challenge Water and Sanitation Thestructural changes Before devolution, the provincial Public Health EngineeringDepartment (PHED) had the main responsibility for the development and maintenance o f water and sanitation services, particularly for large-scale projects and particularly inrural areas. Inaddition, Development Authorities (DAs)andWater and Sanitation Authorities (WASAs) were providing similar services inlarge, urban centers. Although inlegislativeterms water and sanitation are now clearly assigned to tehsils or towns (except inthe case o f city districts, where they are district responsibilities), the emerging implementation arrangements tell a messy story o f unevenretention o fprovincial control. Eachprovince has devolved PHEDina different way, even though the pre-devolution structure o fthe department was the same inall four provinces and the same clause o f the LGO governed devolution inall provinces. InBalochistan, thePHEDdepartmentremains attheprovinciallevel, withthe Secretary and Chief Engineer operating from Quetta. The office o f Superintending Engineers (SE), located before devolution at the divisional level, has beenreplaced by that o f the Executive Engineer (XEN).Not district offices inany formal sense, they are instead deconcentrated arms o fthe provincial department. As a result, TMAs are only providing solid waste management, street lightingand street soling projects.23 Water supply development, operations and maintenance (only inurban areas) remain with the district PHED offices. Sindh oversaw the most faithful interpretationo f the LGO. The province PHED office was abolished and merged into the Local Government and RuralDevelopment Department (LG&RDD), andthe staffwas devolved at the TMA level. However, inKarachi, the KWSB has continuedas an autonomous body leaving little responsibility for water and sewerage with the 18 towns. XENs work with the TMA staff but report to the K W S B management and are not accountable to the TMA. This lack o f devolution o f the K W S B i s a highlypolitical issue given the demands from the population. As a result o fnonperformance, TMAs and Unions are investing indistributionnetworks andother small-scale projectsinapiecemeal style. Punjab has also devolved PHEDto the TMAs but has preserved the PHED structure with SE circles being maintained infour areas o f the province. Additionally, DAs and WASAs inlarge cities, such as Faisalabad, have been devolved to one tehsil, without appropriate reorganization. Therefore, while the WASAs' service network spreads inneighboring TMAs as well, those TMAs are denied arole indecisions on the quality or coverage o f service provision. 23 Solingrefers to the pavingof streets and open drains with bricks. 42 Finally, NWFP provides an example o f the recentralization o f the PHED.Initially, PHED had been devolved to the district, not TMA, level. However, inM a y 2003, the provincial government divided the province into six circles, each headedby a Superintending Engineer (SE) o f W&S department, a newly establishedpost withjurisdiction over 4-5 districts.24These divisions, or circles, now have responsibility for water and sanitation services. Administratively, TMAs have acquired a much-strengthened structure under devolution. There are separate posts for Planning, Infrastructure and Services, Finance and Regulation, inaddition to a Chief Officer, who i s a remnant o f the former Municipal Committee. However, the failure to complete the devolution o f PHED has deprived tehsils o f the quantity and quality o f technical expertise they need. The burden on TMAs has been increased with more UCs added to the jurisdiction o f each TMA. The Municipal Committees (MCs) indistricts and tehsils that used to provide municipal services to the urban populations are now required, with little increase in staffing or resources, to deliver these services to a much wider jurisdiction, includes sizable rural areas not previouslypart o f their mandate. Provincial efficiency incentives Former municipal authorities have continuedas before intheir delivery o f municipal services and, insome post-devolutioncases, havebeeninnovativeintheir approaches. The new, primarily rural, tehsils,however, lacking an existing resource base from which to operate services or make new investments to expand coverage, have faced major problems. Furthermore, the independent and freestanding nature o f each tehsil does not allow any equalizationacross the tehsils o f any one district. There are no mechanisms for prioritizing resources within a district to ensure that the relatively deprivedrural areas begin to buildservices equal to those o f the former municipalities. Apart from Karachi, which as a city district i s able to plan for its whole area, other districts have no legal mechanismunder the legislation for undertakmg cross-TMA water and sanitation schemes. Without that option, they have difficulty planning urban infrastructure. The continuation o f vertical programs inwater and sanitation inall districts, funded either by the federal or provincial government and often supported by donor agencies, represents a challenge to district ownership (Box 6). NWFP presents an extreme example o f the distortions these programs can cause, as the move to established SE circle offices was made at the behest o f MPAs and some senior W&S department officials inorder to win control o fthe DERA and DIMRCprograms that are funded by the World Bank and ADB, respectively. InBannu district, these programs were consideredby the district government to be redundant or politically motivated, and indicative o f the provincial govemment's interference in district matters, as project sites are identified by MPAs or MNAsbut not by local councilors and there i s no local inputindetermining the programs' priorities. Box 6: VerticalProgramsinWater and Sanitation The Khushal Pakistan Program (KPP) andthe ProvincialTameer Programs are two ofthe main vertical programs inthe Water and Sanitation sector. While neither specific to this sector and are broadly aimed at generating employment through public works, such as construction o f roads, electricity and schools and basic healthunits, water, sewerage and sanitation schemes form a significant proportiono fthese programs. The two programs are, however, quite different inthe degree o f district control over planning and 24 A Circle is an administrative division (for technical departments such as PHED) that clusters several districts together. As an example, SE Bannuhasjurisdiction (scheme identification and Technical Sanction power) over four districts: Bannu, Lakki Marwat, D IKhan, and Tank. 43 implementation. InKPP, the federal government provides the broad criteria, while local governments, in particular union councils, identify the specific schemes that are thenapproved by the DDC. By contrast, the Tameer Programs inthe four provinces are highly vertical; districts have very little say inplanning, budgeting, or expenditure control. Inthese programs, members o fthe provincial assemblies (MPAs) can each identify development schemes o f up to Rs. 10 million for any sector inthe provincial domain. In Punjab, the District Development Committee has to give the technical sanction to these schemes, thereby providing district governments with a modicumof influence. InSindh, the MPAs have the choice o f seeking technical approval from the district or provincial P&D departments. The pronounced political visibility o f water and sanitation schemes has made them a serious cause o f conflict between provincial and local governments. With regards to the Tameer programs, findings from the district visit suggest that districts were opposed to many o f these schemes, an attitude that calls into question their long-term sustainability. Insome cases even the KPP projects have become a tool inconflicts between provincial and local governments, with informal provincial controls over the DCO and EDOs being used to approve schemes over the objections o f the district political leaders. Managerial power Overall, the survival o f PHED and related structures inall provinces hampers the delivery o f services by TMAsby restricting the resources and personnel available to them to plan and execute large schemes. InBannu, for example, the acting ED0 has a technical sanction limit o f Rs.2 million, whereas the superintending engineer-still within a provincial circle-can sanction schemes o f up to Rs. 10 million. It i s feared that TMAs may become graveyards o f small, nonfunctional infi-astructureprojects instead o fbeing able to address health and hygiene issues in water projects, or drainage insoling projects-work for which integration i s essential. Some o f the teething problems could be alleviatedif the provincial governments disseminated the Rules o f Business (ROBS) more actively. Their delays, however, appear to have prompted local bodies to evolve their own division o fresponsibilitiesand their own processes and procedures. Of greater concern, provincial departments have been able, ineffect, to re-appropriate the responsibilities o f local governments, as observed inthe case o frural water and sanitation Citizenpower (client power) While ingeneral CCBs have yet to become fully functional, as explained indetail inthe second volume (Annex I-Recent History), there are some success stories where active involvement o f councilors and community organizations has had a positive impact on service delivery. In Bahawalpur, for example, the work o f an active CCB greatly improved the sewerage system o f a neighborhood-a new sewerage line was built by the members, which was o f a higher quality and builtat a lower cost than as compared to those constructedby the Works and Services department. The key ingredients o f this success story were support from the union government, as exemplified by the regular release o f funds, and an educated and proactive membership. Another success story comes form district Jhelum inPunjab, where the district government, inpartnership with a union government and the National Rural Support Program (NRSP), has launcheda pilot project under which CCBs have been created and are being provided on-the-job training through the implementationo fphysical infrastructure schemes, including sewer-drain and reservoir construction. In sum The key impacts o f the changed incentives inthe water and sanitation sector are as follows: 0 Provincial efficiency incentives 44 o Limitedresourcebases are posingfinancial constraints on the new, primarily rural tehsils o The independent and freestanding nature o f each tehsil i s constraining common planning or equalization across the tehsils o Vertical programs are contributingto the pressure from the provincesto undermine devolution. Managerial power o The continuationo f the provincial PHED andrelatedstructures is restrictingthe resources and personnel available to the TMAs o The Ruleso fBusiness have not been disseminated, andprovincial departments have been able to effectively re-appropriate the responsibilities o f local governments. 0 Citizen power (client power) o While ingeneral Citizen Community Boards have failed to meet their potential, there are some examples o f community management resulting inimproved service delivery Access to Justicez5 The structural changes The key reforms inaccess to justice are the abolition o f the executive magistracy and the changes introduced inthe Police Order 2002 aimed at establishing an institutional framework that replaced extra-departmental checks (the district magistrate) with arrangements to make police answerable to citizens and the rule o f law, and increasing the efficiency and competence o f the police. Devolution inthe justice sector rests on complete separation o f the judiciary from the executive and began in 1996 following a Supreme Courtjudgment. This entailedreassigning criminal trials under major criminal statutes tojudicial magistrates. The executive magistrates were, however, still left with somejudicial functions under the Police Act 1861, Police Rules, and local and special laws. Insome cases the Executive Magistrates could still pass sentences o f up to three years imprisonment. With the abolition o f the office o f the District Commissioner, who was also the District Magistrate, alljudicial powers o f the executive magistracy are now vested inthe judiciary. Inaddition, the executive functions relating to law and order and local regulation have been either transferred to the police, or the judiciary, or have not been reassigned. Each Civil Judge now acts also as a Judicial Magistrate, while the District and Session Judge exercises the judicial powers o fthe erstwhile District Magistrate. The centerpiece o f the police reforms i s the creation of District Public Safety Commissions (DPSC) to institutionalize community monitoring o f the police. Together with the judiciary, DPSCs are the bodies to prevent and punishpolice excesses and ensure that the police function inaccordance with law. Managerial power: nazimeen authority overpolice Executive officers inall the districts associate the end o f the executive magistracy with weak implementation, regrettingthat they must now prosecute and prove every allegedviolation before an independentjudicial magistrate. More profoundly, the vesting o f the former District Magistrate's powers inthe district and session judge has fundamentally changed the relationship between the district executive and the police. While overall responsibility for law and order i s given to the District Nazim, nazimeen generally do not feel that the law adequately defines their relationshipwith the district or city police officer, nor do they have access to the legal and regulatoryknowledge within their executives, such as was characteristic o f the executive 25 Detailsofthepolice andjudicial reforms are provided inthe third volume (Annex 2-Technical Considerations). 45 magistrates. Where there i s functional and constructive relationshipbetween the police and nazim, it i s generally because o f a personalrelationshiprather than defined institutional roles and responsibilities.The immediate impact o f friction between the two has been a decline inthe rate o f prosecutions under local laws. There are four major reasons for this decline. First,thejudiciary, which is now requiredto enforce municipalregulations that executive magistrates previously oversaw, has not been orientedin its newjurisdiction under the local laws, and Judicial Magistrates do not effectively use the summary procedure available to themunder the law. The increase inthe number o f cases handledby the courts has further exacerbated delays. Second, the executive generally lacks the capacity and awareness to enforce local laws before independent courts. Improvements indisposal o f cases can arise from more specialization within the judiciary by creating separate civil and criminal courts. All four highcourts have agreed in principle to put such arrangements inplace, but the pace o f implementation i s not uniform. Weak enforcement by the executive i s also tiedto old patterns o fpatronage that are particularly endemic where violation o f local laws can bring substantial economic benefits, as in encroachments on local government property. Third,no municipal authority cannow direct police to take action and enforce local laws at the district level. Formerly the police were obliged to obey the summons o f the executive magistrate who usually accompanied the district officers to enforce local and special laws which tended, typically, to be `politically difficult' to implement. The executive magistracy was seenby the executive as being responsible for implementation o f these laws ina way that allowed politicians to remain away fkom the fkay. Abolition o f the executive magistrate i s now perceivedto have left the various departments, concerned with breeches inthese laws, indisarray.. N o w it i s the police who decide when to come and inwhat strength. When assistance was provided promptly it was usually on request o f the districts' nazimeen. More often, assistance was provided with considerable delay. District staff also explained that the police didnot provide adequate security, and-without executive magistrates as buffers-they had to face threats o f violence that reduced their zeal for enforcement. Fourth, and most wonyingly, it seems that political interference contributes to weak enforcement o f local laws. Several DCOs for instance, noted that direct interference by the nazimeen or simply lack o f interest by them in enforcement o f local laws impeded enforcement. Yet where nazimeen took an interest inenforcement, as inthe case o fprice control laws inBahawalpur, enforcement rose above pre-devolution levels. Inprinciple allfourprovinceshaveorderedthe separationofinvestigationfkomthewatch-and- ward function. However, the situation inthe six districts was far from uniform. Investigative activity hadnot been separated at all inKilla Saifullah, while such separation heldonly at the thana level inBahawalpur and hadbeen implementedinonly half o f Faisalabad. Inmost districts the investigatingofficers reportedto the StationHouseOfficer (SHO) who heads the police station and i s incharge o f its watch-and-ward policing. This dual responsibility has blurred the distinction between the two ostensibly separate police functions. Furthermore, the police station watch-and-ward staff was still investigating local and special laws.26 Where it has occurred, the separation has not had any demonstrable, positive effect such as an increase inconviction rates nor led to improved citizen perceptions ofpolice performance. 26 The Police Order (Article 18(4)) allows such investigations by police station staff where the punishment for an offence does not exceed three years imprisonment. 46 Instead the general belief i s that since citizens now have to pay bribes to two police forces and that corruption has increased. Any efficiency resulting from this internalreorganizationi s unlikely until an independent prosecution service i s established. Itis also generally perceived that thejudiciary has beenunable to effectively control police excesses. Even before the abolition o f the district commissioner/magistrate and the executive magistracy, there was growing disquiet about the ability o f the executive and the courts to control the police. However, before devolution, citizens could approach the executive magistrate to redress gnevances against the police. Now, the process i s more complicated. The citizen must either approach the DPSC, which i s not presently regarded as a viable option, or seekjudicial intercession, which i s a lengthy and expensive process. The separation presents an opportunity for a more independent judiciary to put a stop to unlawful police behavior. However, judicial magistrates do not appear to be sufficiently conversant with the relevant provisions o f the Criminal Procedure Code under which they can hold the police accountable. Moreover, the lower judiciary lacks protection against reprisals by police or other officials whose conduct i s penalized and many feel they are insufficiently supported by the highcourts when this occurs. Manyjudges reported a worrying correlationbetween criminal acts against them and disciplining o f unlawful police behavior. Traditionally, illegal detentions by the police have been a major cause for citizen complaints against the police. The Police Order seeks to curb this abuse by empowering the nazim to inspect police stations and by criminalizing the failure to produce arrestedpersons before magistrates as requiredby the Criminal Procedure Code. However, supervision o f the police through regular inspections o f police stations was almost nonexistent. Under the previous system, the District Magistrate was empowered to visit police stations under the Police Rules issued under the now defunct Police Act. Although by law, the District and Session Judge i s empoweredto inspect police diaries and to order the release o fpersons illegally detained, such officials were not prepared to inspect police stations for this purpose. There i s clearly a need for better inspectionby a third party independent o f the police. DPSCs could play this role but, as discussed above, are not fully functional. Managerial power: external oversight As outlined indetail inthe secondvolume (Annex I-Recent Histoly), ingeneral institutional mechanisms for external oversight of service providers have either not been established, or have been introducedto only a limited degree. This problem i s especially evident inthe case o f DPSCs. Effective DPSCs that enjoy public confidence do not exist inany o f the six districts studied. Only inthe two districts o f Punjab and Khairpur inSindh have DPSCs been notifiedand some preliminary meetings held. Inno district has the DPSC approved a police plan as required by law.*' The DPSC inKhairpur hadno budget, while those inBahawalpur and Faisalabaddid.In all districts there i s a widespread perceptionthat DPSCs will be ineffective incontrolling the police. A success ofthe new system is the establishment and functioning o fthe District CriminalJustice Coordination Committees (DCJCC) designed to enhance interaction among various agencies o f the district criminaljustice system. These committees, made up o frepresentatives o f the prisons, judiciary and police services, are supposed to deal with the irritants inthe working o f that system. Despite its role inenforcingmunicipal offenses, the committee surprisingly lacks a representative 27 More recently, policing plans havebeencompletedand approvedby DSPCs inmany districts of Balochistan. 47 o f the district DCJCC also acts as a horizontal accountability mechanism for the institutions involved. The DCJCC has been established inKhairpur, Bahawalpur, Faisalabad and Bannu, but not in Killa Saifullah because the officers who are supposed to be its members are not posted inthe area. Similarly inKarachi it was not clear who would sit on the committee, as Karachi i s still divided into five Judicial Districts. Inthe districts where the committee hadbeen established, it was reportedthat communication and coordinationamong the various actors inthe criminal justice system had improved. I n sum The key impacts o fthe changedincentives inthe access tojustice sector are as follows: 0 Managerial power: nazimeen authority over police o The end o fthe executive magistracy, andthe lack o fnazimeen authority over police, has resulted inthe decline inrate o fprosecutions under local and special laws o The nazimeen andthejudiciary are unable to check police abuses 0 Managerial power: external oversight o DPSCshave ingeneral failed to carry out their mission, andthere is growing public perceptionthat the police are out o f control. The lack o f improvement in the public perception o f police i s exacerbated by the continuation o f police abuses such as non-registration o f First Information Reports (FIRS)and illegal detentions. o No arrangement exists to replace the Executive Magistrateineachpolice station jurisdiction. The power to inspect places o f pretrial detention has been assigned to the n ~ z i m . ~ ~ o District Criminal Justice Coordination Committees (DCJCC) have succeeded in enhancing the interactionamong various agencies o f the district criminaljustice system. VISIBLE WINS WHENPOLITICS ALLOW Judgments about the success or failure o f devolution do not depend solely on its impact on service delivery, but anecdotal evidence suggests some progress. Reports are common, for instance, that doctors are more often present inhealth facilities and teachers in schools than was the case before devolution. This increasing availability o f staff may also be relatedto the increasing likelihood that local facilities, particularly for health, will actually have medicines available for treating patients. Where procurement i s fully devolved, as inPunjab, evidence suggests that public medical facilities are now well-stocked and therefore more used. Some provinces and districts have rationalizedfacilities. Punjab rationalized Basic HealthUnits,many o f which didnot meet local needs. Other achievements that mightbe attributable at least inpart to devolution include community participation fostering reduction inschool-building costs and some increases inenrollment and retention. Significantly, at the union level, there i s considerable evidence o f councilor involvement inlocal dispute resolution and redressal o f gnevances against service providers. 28 Insome districts (for example, Sargodha and Attock) representatives ofthe district government, inbothcases EDOsfor law, have been invitedto attend. 29 The executive magistrate's jurisdiction was coterminous with the territorial jurisdiction o fthe police, under Cr.PC Section 1641364. The nazim's weak control over the DPO, and his lack of support by an executive body acquainted with the law, has diluted the pre-trial activities of the police. 48 The study found particular areas o fpromising progress: Khairpur district: Khairpur i s among the most innovative districts interms o f community involvement in service delivery. School Management Committees, with the help o f union councils, have been given the responsibility to construct shelterless schools o f which 35 have beencompleted. With community monitoring, their construction costs have come down considerably without compromising quality. These schools have been constructed at around Rs. 200,000-250,000, while similar ones builtby the Education department inthe past cost approximately Rs. 780,000. This initiative i s a good example o f developingpartnerships with local communities and fostering community ownership o f local schools. It should be widely disseminated to other districts for replication. The Khairpur pilot confirms international findings o f similar programs that: the schools are well constructed; costs are remarkably lower; and completion time i s much shorter. Bahawalpur District: Ahmadpur East hospital has a capacity o f 60 indoor beds with two air- conditioned wards. The district has collaborated with the hospital to achieve significant improvements. The operating theater i s fully equipped, and a qualified surgeon conducts major and minor operations. The Medical Superintendent o fthe hospital i s an anesthesiologist. An orthopedic surgeon working inthe post o f medical officer looks after trauma and bone diseases. A child specialist and obstetrician are also posted there. Medicines are in good supply, and three casualty medical officers provide emergency treatment free o f cost. Inaddition to procuring the new equipment, the district government has also managed to clear Rs. 2.6 million o f previous liabilities for buildingconstruction. District Criminal Justice Coordination Committees have been established and meet regularly in all 6 districts except Karachi and Killa Saifullah to carry out investigations, raise issues o f juvenile justice, insist on timely challans (indictments), and on the production o f witnesses in accordance with the law. The delegation to the Sessions Court o fthe power to issue writs o f habeas corpus has been greeted positively as providing a relatively inexpensive remedy against the continuing practice o f illegal detentions by the police and-in some areas-by powerful elites. Gulshan-e-Iqbaltown, Karachi: Incollaboration with Transparency International, the town administrationhas established a Customer Service Center to register citizen complaints. An elaborate procedure has been developed to track the action taken on these complaints and to gauge the level of customer satisfaction. The town government i s also developing an interactive and informative web site while inthe process o f electronically connectingall 13 union councils with the town administration, as well as providing the union andtown offices with computers through which citizens can register their complaints on-line. The town administration i s also consolidatinga variety o f existing data on the municipal infrastructure, and linlungit to a GIS system, and has launched a property survey inorder to increase property tax collections. TehsilJaranwalla, Faisalabad: The town administration has taken a number o f innovative steps to improve service delivery. With private sector and NGO involvement, the tehsil i s constructing a comprehensive database o f the existing infrastructure, including, for example, the number o f water outlets, pipes, drains, connections, joints and other details o f sewerage, gas and telephone connections. This database will then be integratedinto a GIS system and usedto prepare a master plan o f the town. With baclung from the TehsilNazim and based on their own consultations, the 19 women councilors inthe Tehsil Council have pooled their individual allocations for development schemes to develop a Women's Resource Center. This center will have a shelter for 49 abused women and embroidery, computer and other industrial training facilities. Even more encouragingly, the woman MPA from the tehsil has contributedher individual development allocation to the construction o f this scheme. The same nazim, responding to citizen complaints, has undertaken activities that do not come under his directjurisdiction. When Palustan Railways didnot respond to requests to improve the road at two o f the city's main railway crossings, he ordered his staff to use their own resourcesto improve it. Similarly, as a measure to support the local traffic police, his TMA has hired eight people from a private security agency to helpmanage traffic and lessenthe impact o f traffic jams. Union councils, Karachi: Councilors are demonstrating innovation inresponse to increased contact with citizens. One union council has purchasedmanhole covers and street lights from its own funds even though such maintenance falls outside their purview. The naib nazim claimed that the public does not have the patience to wait while the KWSBtakes months to respondto requests to supply manhole covers.The unionhas also hiredsanitation workers (kundimen)on contract, despite the objections o f the provincial auditor that such an initiative goes beyond their scope o fresponsibility. Another union council has mobilized the local community to restore a local park. Neighboring union councils are also cooperating to improve service delivery. For example, one council that didnot have adequate resourcesto purchase a tractor and a trolley (sanitation vehicles) has developed an arrangement with its neighboring union council whereby one has purchasedthe tractor and the other has purchased the trolley, and the vehicles are shared between the two. Nuzimeen and Senior Staff Can M a k e a Difference The new incentives have seemingly allowed determined and dynamic nazimeen to achieve significant results. The key variable i s the strength of the provincial incentives to intervene inthe districts. The province has few incentives to intervene when: The Chief Minister and/or the main provincial coalition and the District Nazim are from the same political alliance The Provincial Assembly constituencies within the district are filled withpoliticians from the same political alliance The nazim has sufficient independent political standing to resist provincial interventions. When the province has few incentives to intervene, the nazim gains a fairly free hand to direct district staff and, despite the inevitable resource constraints, can achieve results. However, when the provincial incentives for intervention are high, nazimeen often finds their hands tied (Box 7). Box 7: Provincial Interventions at the District Level I n Khairpur district, the nazim i s a social activist tumed political leader. She brings to herposition a vast experience ininvestigativejournalism, activism inwomen's rights, environmental conservation and social uplift.The daughter o f former ChiefMinister andPPP stalwart Qaim Ali Shah, she has effectively usedher political background despite its having also been a major impediment, since her father's opponents-chief among them the powerful Pir o f Pagar+have become her opponents as well. As one o ftwo female nazimeen inthe country, she has some significant achievements to her credit. She has encouraged the formation o f over 100 CCBs inthe district, with the assistance o f UC nazimeen and the ED0 Community Development. Hers is one o fthe few districts that has been active inthis sector. A joint committee o f the treasury and opposition members o f the district council has approved a total o f 98 schemes, while the DDWC under the DCO's chair has approved 59 schemes. Forty CCBs have deposited 50 their contributions to these schemes. Unfortunately these CCBs have become hostage to the political polarization that typifies the administration o f this district and have not yet received funding themselves. The successful initiative to fund the construction o f shelterless schools, using district funds and the contributions o f beneficiary communities managed with the assistance o fUnion Councils, has also stalled due to conflict with the province. Apparently on instruction from the province, the DCO didnot approve the new structures and have not beenregularized by the EducationDepartment for the same reasons. I n District Bannu, the provincial government mounts a great deal o f interference inthe affairs o f the district govemment. The provincial government often overrides the role o f the district government, even, insome instances, sending instructions directly to the ED0 (Education) instead o f following the proper procedure. In southern districts of NWFP,provincial and federal politics revolve around the development o f water schemes. Because many o f the schemes-mostly donor funded-are provincial schemes, nazimeen were given no say inthe development decision-making process. The nazimeen therefore prevented their Works and Services EDOs from attending meetings inPeshawar on these vertical programs. The response o f the province has been to establishnew provincial posts at the circle level (group o f districts) to enable the province to go forward with provincially approved projects without involving local governments- effectively cutting them out o f major water and sanitation decision-making and ensuring that MPAs influence these decisions. For example, inBannu the DERA schemes for 2003-04 were finalizedby the MNA andMPAs from the district ina meetingwith the provincial Works and Services Department. A Decisive Break from the Past Devolution has brought a decisive break with the past intwo important respects. First, local government now has unambiguous constitutional protection. The agreed changes to the constitution were enacted on 30 December 2003, following a vote inthe Senate, and National Assembly (29 December). The local government system, as prescribedby the Provincial Local Government Ordinances, i s now includedinthe sixth schedule o f the 1973 Constitution. Under Article 268 (2) o f the Constitution, laws specified inthe sixth schedule cannot be altered, repealed or amended without the previous sanction o fthe President. Further, it i s evident that there i s growing support for devolution across all political parties. Second, as a result o fthe LGO provisions, the reservation o f 33 percent o f local council seats for women inall three levels o flocal government has proven the most strategic step ever undertaken by government toward women's political empowerment. Most directly, despite the significant number o f vacant seats, it has enabled over 30,000 women to enter into formal politics at the local government level. The consequences have already been felt inprovince and federal politics-not just interms o f increasedwomen's representation, butalso inthe fiscal and administrative domains as well. The Federal cabinet and the cabinets o f Punjab, Sindh and Balochistanlook set to approve multiyear Gender ReformAction Plans which will commit them to wide-ranging reforms consistent with devolution in electoral procedures, women's representation inthe civil service and interms o f introducingnew measures in statistics, budget and expenditure management to track the impact o fresource allocation decisions on women. Thus, devolution marks a decisive breakwith the previous arrangements-arrangements that, as the earlier discussion highlights,were associated unambiguously with the persistent failures in service delivery. VISIBLE LOSSES As with any dramatic policy change, there are costs as well as gains. Two concerns emerged prominently duringthe course o fthis study: civil servant morale and law and order and public safety. 51 Civil Servant Morale Discussions with senior staff in all districts emphasized that career paths for the provincial and federal staff postedinthe provinces following devolution, causing a serious loss inmorale. There are two ways o f looking at this contention. These senior staff saw devolution as, at best, a confusing and demanding institutional change and, at worst, as a sudden and unfair blow to their career prospects. Others regardedthe move as a long overdue restraint on privileged groups who hadprotected their fast-track career options for too long and dismissed the protests as the complaints o f those who feared losing rents from district postings that were often thought to be particularly l~crative.~'Figure 7 shows the profile o f the DMGgroup. Figure 7: DMG Staff by Grade I 300 1 250 200 150 Women 100 50 0 17 18 19 20 21 22 BPS grades Source: FederalGovemment Civil ServiceCensus, 2000. Although few innumber, District Management Group staff (DMG) and the Provincial Civil Service (Executive) were the most affected and the most vocal about their concerns. The DMG group felt the loss o f their career prospects most acutely. This occupational group had, historically, been posted to the districts and divisions, at the level o f Commissioner, Deputy Commissioner, Additional Deputy Commissioner and Assistant Commissioner. Also, as one o f three groups inthe All Palustan Unified Grades (MUG), statistically they had a better than average chance o fpromotion to senior grades. At the highest level of the civil service, o f the total o f 47 BS-22 officers, 29 belong to the MUG, more than 60 percent o fthe total (Figure 8). 30 For the reasonswhy cadre interests are particularly deeply entrenchedinPakistan, see (Kennedy: 1987). 52 Figure8: Distributionof Positionsby Grade for FUG and APUG Staff 1,000 800 600 400 200 Source: FederalEstablishment Division. Other provincial occupational groups also felt that devolution had diminished their career prospects and their status. Provincial Public Health Engineering staff inparticular regarded their proposed transfer to the tehsils as particularly demeaning. This concern lies behind much o f the determined resistance to the transfers, opposition so effective that, inthe main, water and sanitation continues to operate as a provincial function with technical and administrative sanction powers continuing to lie inthe province. Inconsidering morale, the operational question isnotwhether the loss ofcareer prospects and privileges wasjustified. The question i s how to re-establish a sense o fpurpose and enthusiasm in those who must provide the administrative lead for the new local governments. Law and Order The reforms dealingwith access to justice were undoubtedly among the most complex, ambitious and controversial elements of devolution. They entailed three major legal and institutional changes: 0 Although devolution i s often seen as predicated on the 1973 Constitutional principle o f the separation o fpowers, it i s more accurately understood interms o f Article 37(i) o f the C~nstitution,~~which refers to subsidiarity. Nonetheless, devolution didresult infurther separation o f the executive from thejudiciary by abolishing the post o f deputy commissioner and dividingthe power o f that office among the District and Sessions Judge, district nazim, the DCO, and DPO. This change also entailed scrapping the office o f district magstrate, one o fmany hats o f the Deputy Commissioner, and the cadre o f Executive magistrate^.^^ 0 With the abolition o f the deputy commissioner, the Police Order 2002 was originally craftedto insulate police from political interference and provide a "unity o f command" bymahngthe district police officers(DPOs) responsible for allpolice professional 3' "The state shalldecentralize govemment administrationso as to facilitate expeditiousdisposal of its businessto meet the convenience andrequirements of the public." 32 It shouldbenotedthat the ED0 Revenueremains with trial jurisdiction incases relating to landrecords and land revenue, continuingan age-oldtradition extendingfrom the office of the ziladar, to collector,to the deputy commissioner. 53 matters including postings and transfers o f 0ffice1-s.~~Unlike the previous police head o f a district, the DPO i s no longer supervised by or accountable to any part o f the local civil service, but accountable to new structures-District Public Safety Commissions and Police Complaints A~thorities~~-andto thejudiciary that has now assumed all o f the judicial and quasi-judicial powers formerly held by district commissioners. Police reforms also include creating a separate independent prosecution service. The federal government has proposed a model law to set up independent prosecution agencies, and this draft i s under active consideration by provincial governments. Apart from Balochistan, however, promulgation o f the new laws has been miredinturf disputes between the province departments o f law and home. As noted earlier, the key concerns are: the power o fthe District Executiveto enforce regulations has been weakened; nazimeen have little control over the police; there has been some decline in prosecutions; and there i s a widespread perception that corruption has risen. Of most concern, it seems that weaknesses in external oversight o f the police have made the police less governable. The consequence i s that, inpopular perception, the incidence o f police excess and torture during pretrial detentionhas increased markedly, along with a perceptionthat police are unresponsive to increasing (or better reported) violence against women (Human Rights Commission o fPakistan: 2004). As a result police reforms are inserious danger o f beingperceivedas a one-way street on which the police has simply secured more resources while becoming less accountable. Risingcorruption inthe police service exacerbates corruptioninother public offices, motivated by the sense that corruption and unresponsiveness can occur with impunity.The failure o fthe police to provide basic public safety i s undoubtedly associatedinthe public perception with a decline inaccountability inother service areas (for example, revenue and tax, land, irrigation) which impact on the poor. A CHALLENGINGTIME FORDEVOLUTION As notedearlier, devolution hadbothbroad governance objectives and some narrower technical goals. Although this study i s modestly optimistic about the future o f devolved governance in Palustan, inthe near term, devolution will achieve few if any o f these objectives. It i s certainly unrealistic to anticipate that substantial improvementsinservice delivery could be detected after such a short period. Thus while the evidence does not allow the proponents o f devolution to point to many specific improvements, as systems bed down, detractors mightbe able to point to some areas inwhich devolution has seemingly made matters worse without compensating systemic improvements. Even the gains have been less substantial for the districts than for the TMAs, as the pervasive and often undermining intrusion o f the province into district affairs i s a particular problem. Inreality, devolution is undoubtedlya fact oflife. Itisnot plausibleto expect any substantial return to a discredited set o f institutional arrangements with a proven track record o f failure. However, undoubtedly opposition to it (from the senior bureaucracy and fromprovincial politicians inparticular) still remains. There i s no certainty that, inthe face o f this opposition, devolution will now be developed and entrenched to maximize the returns that it mightoffer. The 33 It is reportedthat powers of postings andtransfers ofDPOshavebeenvestedinthe provincial ChiefMinisterwho will appoint on the basis of a list preparedby the ProvincePolice Officer. Further, Members of Provincial Assemblies inthe districtjurisdiction will becomemembersofthe DPSCs.See footnote 16. 34 It is reportedthat the 14April 2004 decisionhas mergedthe functions o fthe PoliceComplaintsAuthority with the ProvincePublic Safety Commission.See footnote 16. 54 riskis that infailing to deliver visible service delivery improvements inthe short term, the process will freeze inplace and the present hybridarrangements will be accepted as permanent. There are other potential challenges for devolution on the horizon: 0 Managing inequity between and within provinces 0 Reemergence o fprivate-goods strategies. Managing Inequity Between and Within Provinces The likely significant increase inthe fiscal transfers to the provinces under the 7' NFC Award herald a major advance in devolution. However, as shown below, giventhat the current award i s generally fiscally equalizing, the likely accompanying shift away from a population-based formula underpinning the divisible pool presents some potential risk o f increased fiscal inequity if these transfers are designed inappropriately. Any evaluationo ffiscal equalization inPalustanrequires somejudgment about the provinces' current own fiscal capacity,35expenditure needs, and the fiscal gap betweenthe To put current federal-provincial fiscal equalization inperspective, Table 16 sets out different revenue categories, expressed on a per capita basis-with those most associated with local tax effort or local tax base on the left and those best thought o f as equalization transfers arranged on the right hand side.37While the Own Source Revenues (OSR), Straight Transfers, and OZT replacement grants (which together can be broadly categorized as provincial "own" revenues) vary significantly across provinces, the divisible pool allocations are by design the same across all provinces. Table 16: Per Capita Resource Flows (FY 2003/04) Source: Federal and provincial budget estimates, and W o r l d Bank staff calculations. Fiscal capacity o f a sub-national government equals the revenues that couldbe raised by that government if it applied national average tax effort to all its tax bases. Two major caveats are in order on the expenditure needs side. First, there is typically at best a limited correspondence between fiscal resources, spending, and public sector outcomes. Efficient use o f resources rather than absolute levels often matter more for ultimate performance. Secondly, the expenditure levels o f the "richest" province may not necessarily be what i s realistic or desired for all other localities. The four comuonents o f the mainstream federal-wovincial transfers are as follows: The divisible pool provides revenue sharing; Straight transfers represent the return o fresource royalties, charges and excises after deducting a federal collection fee to the province o f origin; Special lump sumtransfers are provided to NWFP and Balochistan provinces for backwardness; and In 1999,a fourth component was effectively added to the Award, the "Octroi andZilu Replacement Grant," although inpractice the transfers o f the 2.5 percent GSTbegan only from fiscal 2002-03. 55 To gauge the current equalization impact o f the transfers, Table 17presents the cumulative per capita revenues as a percentage o f the national average, moving from left to right. Table 17: CumulativeProvincialRevenues(Normalized) Source: Federal andprovincial budget estimates, and World Bank staff calculations. Overall, comparing column 3 (own source revenues + straighttransfers + OZT replacement) to column 5, it can be seen that per capita revenues do converge with, by design, the divisible pool making the greatest contributiontoward this c~nvergence.~'Sindh and Balochistanhave the highest levels o f locally derived revenues (due primarily to royalties on natural gas and oil), which appear to reflect the higher fiscal capacity given the assigned tax basis. Their per capita "own" revenues are more than one and a halftimes the national per capita average. The subsequent transfers reduce these per capita disparities somewhat. These figures suggest that the current system o f divisible pool and other grants provides some degree o f fiscal equalization, although to be sure it would be necessary to assess whether the remaining variance reflects differences inexpenditure needs, and these data are not available. The results are intuitively reasonable, however, ifexpenditure needs inBalochistan are driven by the provinces size and remoteness. It should however, be noted that while on per capita terms the current NFC formula does result insome fiscal equalization, population i s only one factor determining fiscal needs, and other factors like poverty and backwardness are equally important. Thus the current divisible pool arrangements, when coupled with the subventiongrants, mightbe providing a reasonable degree o f fiscal equalization while beingbased on populationmeasures which offer bothtransparency and simplicity (Shah: 1994). The proposals currently under discussion to include backwardness, tax effort and other indicators inthe NFCformula, if designedpoorly, could make the system complex without achieving fiscal equalization. The sensitivity o fNFC awards to any move away from population as the overarchingcriterionin the divisible pool allocations-whether toward area, poverty, or the Human Development Index (HD1)-suggests the need for creative solutions grounded intechnical 38 It is frequentlyobservedthat the vertical fiscal imbalancescoupledwith the correspondinglyhigh levels of transfers maybe settingdisincentivesfor the provincesto raise their own revenues. Historical evidence, for example, suggests that the provincesmight be under performingon OSR, somethingwhich i s aggravatedby the transfers (Ahmadand Wasti: 2003). 39 The choice of fiscal-capacityand expenditure-needsindicatorproxiesusedintemationally varies significantly (Ma: 1997; Shah: 1994). Especiallyonthe expenditure-needsside, simpler indicatorsare typically preferred, as comprehensiveandreliable sub-nationaldata are typically in short supply. Populationis typically the default expenditureneeds. Indonesiarecentlyintroducedthe use of area, regionalprice levels andpoverty inadditionto populationas measuresof expenditureneeds(Hofman, Kadjadmiko andKai Kaiser: 2002). 56 Devolution certainly need not exacerbate problems o f inequity (Smoke and Schroeder: 2003). The few cross-national, empirical analyses conducted o f this propositionreveal no clear-cut pattern. For example, one study o fregional inequality inboth developed and developing countries (Shanka and Shah: 2003) found that unitary countries were on average more regionally unequal than federal countries. The study hypothesized that this negative relationshipbetween decentralization and inequality results from the greater political risks these disparities pose for federal states and therefore the greater incentive to restrainthese inequalities. Most transition economies have equalization components intheir grant programs to sub-national government^.^' Even without changes inthe NFC formula, there are risks inthe present federal-provincial arrangements. Although inprinciple, the formula-based distribution o f the divisible pool combined with the grants has potential for fiscal equalization, since the former relies on revenue- sharing, any changes inthe overall tax situation can result inunforeseen consequences for equalization(Shah: 2002). The PFC awards offer a further risk o fincreased fiscal inequity. Inprinciple, the present generation o f PFC Awards i s likely to be fiscally equalizing to some degree, on the simplistic but not unreasonable assumption that formula-based allocations to districts are more likely to be equalizing than discretionary allocations based on political clout. Inaddition, the formulae have, to varying degrees, factored underdevelopment as an indicator o f fiscal need. However, current data do not allow any determination o f the changes introducedby the first round o f PFC awards, and it i s clear that the new, primarily rural, tehsils with no existing resource base from which to operate services have faced major problems. It i s evident that the delivery o f municipal services indensely populatedurbanandperi-urbanareaspresents a different type ofchallenge thantheir delivery inthe often dispersed, poor, rural areas. Developingrural water and sanitation requires significant investment. At the same time, rural tehsils are poor, and their scattered populations increase the cost o f service provision. The independent and freestanding nature o f each tehsil does not allow any equalization among tehsils o f any one district. N o mechanism for prioritizing resources within a district ensures that the relatively deprived rural areas begin to build services equal to those o f the former municipalities. Re-emergence of private goods strategies The pull on councilors to provide private rather than public goods will remain very real for reasons o f history, patronage and political visibility. The oversight mechanisms are well designed but still intransition. As yet, District Councils andtheir MonitoringCommitteesare not providing any strategic check on these tendencies. CCBs have yet to take offto any significant degree. Inmany districts the Mushawarat Committees, designated as the main institutional vehicles for district-wide planning, are not functioning properly primarily because o fpolitical tensions between the district and tehsil nazimeen. While checks and balances are introduced and institutionalized, the well-trodden path o fpatronage-based hiringand distribution o f small schemes undoubtedly beckons local councilors. The second volume (Annex I-Recent History) sets out the limited evidence available so far. Reassuringly there i s little sign that expenditures are being moved from operations and 40 Latvia, Lithuania, Poland, Romania, Russia and Ukraine have adopted transfer formulae that explicitly incorporate either fiscal capacity and/or expenditure need equalization concems. In developing countries, programs usingan explicit standard o f equalization are untried, although equalization objectives are implicitly attempted inthe general revenue-sharing mechanisms usedin Argentina, Brazil, Colombia, India, Nigeria, Mexico, Pakistan and South Africa. These mechanisms typically combine diverse and conflicting objectives into the same formula and fall significantly short on individual objectives. Because the formulae lack explicit equalization standards, they fail to address regional equity concems satisfactorily. 57 maintenance expenditures toward new investments, from large to small schemes, or from completing ongoing schemes toward more new schemes. However, the fiscal evidence i s limited and nuanced. The well-established pattem might come to dominate again intime. THEKEYCHALLENGESFACING DEVOLUTION Insum, the keychallengesfacing devolution canbe summarizedas follows: e Completing devolution is not seen to be inevitable. Although the abandonment o f the reforms i s unlikely, especially now that local governments are protected under the Sixth Schedule o f the Constitution, there i s some hope, insome quarters, that it mightbe frozen at its current stage and there i s a widely heldview that its completion i s tied to the sustainability o f the Musharraf government. Verticalprograms undermineplanning and budgeting in local government and reduce the commitment of senior sta8 Vertical programs have largely butnot entirely outlived their usefulness. Givenpressure to improve service delivery and the perceptionthat central control i s required to overcome local governance problems, federal governments and donors alike have favored vertical programs. That approach, however, represents an outdated mindset that held central control to be feasible, a view refutedby very strong evidence that such top-down management causes as many problems as it solves. Senior stafl andpoliticiansfrom theprovincial governmentsfeel devalued. Devolution was seen as a zero-sum game inwhich any gains for local government were inevitably losses for the provinces. This problem i s clearly evident inthe jurisdictional conflicts outlined indetail inthe second volume (Annex 1-Recent History), particularly the problem o f politicians inhigher levels o f government claiming credit for the efforts o f local governments (Box 12,Annex 1-Recent History). This verdict will not change without some credible signals being sent about the future role o fthe provinces and, very particularly, without the transfer o f functions from federal to provincial governments along with at least the mediumterm prospect o freducingthe chronic vertical fiscal imbalances. e Senior staflandpoliticians from the local governments consider themselves to be under threatfrom theprovinces. Many feel that the provincial governments will seek ways to undermine their performancehowever determined their efforts. This feeling i s certainly justified indistricts that are politically opposed to the province (for example, Killa Saifullah and Khairpur), where the province can use its administrative control over senior district staff, such as the DCO, to influence the flow o f funds and the implementation o f local development schemes. Mirroring the federal-provincialrelationships, some credible signals o f future intentions are necessary; they needto include further transfer o f some functions such as PHEDto local authorities and the prospect o f some tax reassignments and debt-for-performance swaps. Certain weaknesses that have emerged inthe implementation o f the LGO concerning sanctioning powers inlocal government, the monitoring committee and the weak oversight offered by the Local Govemment Commission also need to be addressed. Senior staff andpoliticiansfrom local governments have not yet realized the degree to which they are in charge of their own affairs. For example, as discussed inthe second volume (Annex 1-Recent History), local governments remain heavily reliant on provincial revenue transfers and, while notingthat the taxes assigned to them lack buoyancy and potency, they have not fully utilized user charges as potential sources o f own source revenue increases. The past has, o f course, shaped the present, and despite the large numbers o f new councilors, a legacy o f dependency and o f waiting for instruction hangs on. Local governments are now ina position to reduce the development throw- 58 forward, to increase the coverage o f user charges and to improve inter-jurisdiction planning and coordination Federal andprovincial governments do notperceive many alternatives to vertical programs. The debate has been cast interms o f either seeking, ineffectively, to over- determine the behavior o f sub-national governments through vertical programs, or abandoning both the programs and any prospects for communicating the PRSP goals from the federal to the provincial and then to the local level. The potential use o f conditional grants as an interim measure to achieve service-delivery and poverty- reduction goals has not yet won the full accord o f federal and provincial governments. The range o f such grants available to federal and provincial governments can be increased through negotiations with the donors and by improving the range and quality o f social-indicator data available. e District governments do not employ their own stas As noted earlier, and elaborated in the second volume (Annex I-Recent Histoly), local governments have little de facto control over the APT authorities o f local staff, and especially little control over the DCO and DPO.The ambiguities inthe present arrangements and the reality that, for many district staff, there i s no pressing reason to follow the instructions or the policy direction set by the nazim or the council will only be overcome through the creation o f a distinct, local-government, career service. Accountability also requires the provincializationo f accounts as a first step toward complete preparation by districts and the maintenance o f their own accounts. 0 Theseparation ofjudicial and executivepowers is incomplete. The reporting arrangements between the District Police Officer and the nazim are uncertain and the transfer o f all non-adjudicatory powers to the courts have combined to lower confidence inthe police andjudiciary even further. Nazimeen havelittle control over the police and as a result the enforcement o fregulations has suffered, there has been some decline in prosecutions, and there i s a widespread perceptionthat corruption has risen. 59 3. Recommendations A STRATEGY FROMHERE To bolster devolution as a process and, inparticular, its impact on service delivery, its architects and supporters need to take a strategic view o f the challenges they face and the responses within their grasp. This chapter offers a range o fpractical actions that can be taken immediately to add momentumto the transfer o f authority, resources and responsibilities. It opens, though, by examining three key considerations shaping an overall strategy that would 1) Raise awareness of devolution's gains andpotential, 2) Incorporate flexibility within the framework o fbasic principles and 3) Make local governments allies inpursuingnational priorities. Seize the Opportunity Presentedby the NFC Award Note the successes The successes o f devolution are not widely known. Three are clear. First,it has changed the incentives for nazimeen, senior staff and service providerswithin districts and TMAs: Provincial efficiency incentives. o Budgets are predictable o Significant autonomy has been achieved inpreparing the non-salary recurrent budget o Despiteproblems with the tax bases and administrative collection arrangements, clear revenue assignments have beenprovided along with some incentives for local revenue-raising. Citizen power over nazimeen. o Incentives have been created for nazimeen to respondmore directly to the concerns o f citizens without resorting to the time-honored approach o f providing patronagejobs and small schemes. o Councilors have become public figures, earning significant political legitimacy following good voter turnout andpolitical contestation and evidence o f increased contact between citizens and policymakers. Citizenpower over service providers. o There are some sporadic but encouraging signs inrelation to the Citizen Community Boards and the School Management Committees. Second, it has already led to some visible service-delivery and fiscal gains: Anecdotal reports find doctors more often present inhealth facilities and teachers in schools than was the case before devolution. Reportedly the likelihood has increased that local facilities, particularly inhealth, will actually have on hand medicines available for treatingpatients. Determined and dynamic nazimeen can already achieve significant results as long as they have a free hand and no undue interference from the provinces. Contract-based employment i s being usedto prevent permanent liabilities arising. Finally, it has marked a decisive break from the past: The 1973 Constitutionalprovision regardingthe separation o fpowers has finally been achieved inlocal governance. Local government now has unambiguous constitutional protection. 60 0 33 percent o f local council seats are reserved for women in all three levels o f local government. A public information campaign setting out bothprogress andproblemswould assist in entrenching devolution and in spurringa more open debate about the options from here. N o w i s the time for such a campaign. The NFC decision i s likely to be announced soon and offers a significant opportunity to kick-start this debate. TheSigniJicanceof the likely NFCAward Figure 9 depicts the increase inprovincial consolidated funds that might emerge as a result o f the new NFC award. Figure 9: ProvincialConsolidated Funds under New and Old NFC Scenarios Current transfers 200,000 - 180,000 - 160,000 - 140,000 - 120,000 - Capital Rev. Reciepts 100,000 - 80,000 - 0Dev. Rev Reciepts 60,000 - 40,000 - 0SpecialTransfers Straieht Transfers Scenario 1-NFC Award at 42% Scenario 2-NFC Award at 50% -I 250,000 7 250,0001 200,000 200.000 Capital Rev. Reciepts 150,000 =Capital Rev. Reciepts ODev. Rev. Reciepts 150,000i O D e v . Rev Reciepts 100,000 0Special Transfers 100,000 O S p e c i a l Transfers Straight Transfers .Straight Transfers 50,000 E4 42% Provincial Share 50,000 L 1 I 1 I Assumptions: The projections with increasing provincial shares have been simulated on the total revenue collected inFY 2003. The inter-provincial sharing formula remains the same (population based and the population proportions o f 1998 census are used). There is no change inthe amount o f straight transfers, special transfers or provincial OSR. Note; These assumptions model the transfers as still including the existing grants-despite the probability that these would be incorporated in the poolLon the basis that ifthey are in fact incorporated into the pool then a formula will be sought that roughly replicates the current pattern, and so the effect would be the same. Source: ADB/DfID/WB DSD Dataset 2-and staff calculations from the flow o f funds model. Inturn,the federal government appears to be advocatingthe abolition ofthe OZT offset and special grants, as well as a revision o fthe Divisible Pool allocation formula. The current 61 allocation formula mightbe extended beyond population to include area or the Human Development Index.Moving to a full formula-based allocation from the Divisible Pool carries significant political challenges. Political reality may require that relative changes inoverall provincial allocation amounts not differ significantly between provinces. Inthat case, the provinces would be likely to receive on the order o f 10percent more federal transfers each, a not insignificant windfall. Intum, Sindh may argue for additional revenue sharing (for example, o f customs revenues), as formula-driven transfers tend to work against it. The provincial-local implications o fa revisedNFC are unclear. Not only is there much uncertaintyabout what the changes intransfers to the provinces will be, individual provinces' decisions on how to pass these resources on to their local governments may be even more uncertain. Any simulation o f the impact o f a revisedNFC award on local governments i s therefore highly speculative. However, higher federal divisible pool transfers to the provinces enhance the significance o fthe provincial finance commissions (PFCs). Overall increases in provincial-local transfers will likely exceed current allocations on existing recurrent expenditures (mainly district wages). These additionalresources will therefore likely be distributedinaccord withthe evolving PFC formulas. The abolition of the OZT offset, implicit inthe full merging of the additional GST revenues inthe Provincial Allocable, will also end the implicit earmarlung o f these resources to TMAs based on their historical levels o f collection. Assumingthat the resultant increases inprovincial resources are made available to local govemments under the same sharing arrangements, including the application o f the current Provincial Retainedand Allocable ratios as well as maintaining salary and development proportions, then a 50 percent NFC award would impact on DSDDistrict spending as shown in Table 18. Itmay, however, be argued that the Provincial Allocable should enjoy a proportionately greater share o f the increase, given the PRSP spending priorities. 62 Table 18: Potential Increase in District Budgets with New NFC Award I Baseline NFC at 37.5 percent NFCat 50 percent allocation (Rs. million) Develop- Percent of District Development Salary ment Salary previousbudget Source: ADB/DfID/WB DSD Dataset 2-and staff calculations from the flow o f funds model. Assumptions: The simulation compares the projections with the baseline revised budget estimates o f FY 2003. The inter-provincial sharing formula remains the same (population-based, using the population proportions o f the 1998 census). There i s no change inthe amount o f straight transfers, special transfers or provincial OSR. Provincial Allocable i s calculated using the current ratio but applied to the entire Provincial Consolidated Fund. The share o f the district is calculated usingthe formula for sharing development funds prescribed by PFC inthat province. The additional revenues are divided among development and salary allocations in accordance with their FY 2003 proportions. Assuming, not unrealistically, that the major share o f additionalresources is infact available for development spending, the impact inthe six districts studied would o f course be even more dramatic, ranging inthe study sample districts from 270 percent for Killa Saifullah to 7,400 percent for Faisalabad (see the thirdvolume, Annex 2-Technical Considerations, for details). The magnitude o f increase i s so significant as to warrant a rapid consideration by PFCs o f the lunds o f grant systems to be usedto transfer these resources inways that also pursue PRSP priorities. A public-information campaign launched intandemwith the announcement o fthe NFC Award mightset out the limited evidence on devolution's successes and challenges-but more importantly, it could highlightthe goals that devolution i s seeking to reach and invite debate on the best ways to achieve them. Enforce the Core Principles-and Accommodate Local Differences The implementation challenges have evolved intwo ways. First, the message of devolutionhas gone from the broad and visionary to the narrow and technical. The March2000 announcement o f the devolution plan and the subsequent August 2000 Local Govemment Planpresented a broad strategic vision (National Reconstruction Bureau: 2000). Subsequent debate has focused on the technical issues o f implementation-with a strong focus on the technical fiscal and administrative aspects and an underlying concem about the degree o f compliance with the Local Govemment Ordinances. 63 Box 8: Celebrate Modest Diversity The ED0 Education inRawalpindi announced that he will decentralize hispowers to enable teachers to seek resolution o f human-resource management issues at the tehsil level. The ED0 said, "It was physically impossible for me to handle all issues. The district education department has to handle 22,000 personnel. H o w is it possible for a single person to deal with such a large number o fpeople, especially when there are heaps ofproblems?" The DDO, the DEOs, and the DDEOshave been empowered to deal with all cases o fteachers intheir respective areas. Office bearers o f Ittehad-e-Assataza Pakistan expressed their great appreciation o f the EDO's step, which they said, had solved the long-pending problems o f the teachmg community. Source: Dawn, January 19,2004. Second, the range o f actors has grown exponentially. While the NRB alone supervised the pre- October 2002 phase o f devolution-defining policy, assigning new functions, powers and responsibilities to existing and newly created organizations and enacting laws and regulations to make these arrangements durable, inthe present phase this i s neither possible nor desirable. Unlike the first stage o f devolution, the process o f adapting the new institutional arrangements to local circumstances will now depend on the sustained efforts o f elected and administrative officials at federal, province and local levels, not least by devising ways to productively engage private sector and civil society inthisprocess. The increasing complexity o f the actors that must implement the next stages o f devolution argues for some relaxation o f the one-size-fits-all approach, with some recognition that provinces are likely to follow slightly divergent paths. Since local governments across the country face very diverse situations (Qadir: 2003), some variance inimplementation i s not only inevitable but also welcome (Box 8). Turninga constructive blindeye to some o f the more marginal differences has preventedneedless federal-provincial conflict, and has started devolution along a path that will be o f strategic value inthe future-where basic principles are recognizedand enforced, but where, in the detail, local institutions are tailored to local circumstances. N o w that the basics are inplace, the message from the federal government mightperhaps return to the level o fbroader principles rather than detailedimplementation instruction^.^^ T o achieve this, ina context o fmany veto-players, strong leadership from the Prime Minister's Office will be crucial. Transmit NationalPolicv Priorities MoreEffectivelv Movefrom an over-reliance on vertical programs At the federal level, specific purpose or conditional grants inthe context o fPakistan's intergovernmental fiscal system remain small compared to other types o f transfers, although a variety o f programs have existed inthe past (Ahmad and Wasti: 2003).42Provincial governments rely largely on block transfers from the divisible pool (revenue sharing), "straight transfers," and "special lump sumtransfers" for backwardness. "Other transfers and development grants," similar to conditional or special purpose grants, total 2.3 percent and 3 percent o f federal recurrent transfers inFY 2002/03 and 2003/04. However, these figures do not include on-lent donor projects, which continue to play a significant role inprovincial finances. These are 4' Tight-loose tactics, that is, holding fast to core principles,while allowing diversity in implementation, are well known. See particularly (Bellamy and Taylor: 1998; Peters and Waterman: 1982; Witkin and Berry: 1975). 42 Internationalexperience with conditional grants varies significantly, althoughthey appear to find some role alongsideunconditionalrevenue sharing via block grants inmost federations (Shah: 1994, Annex F). The third volume (Annex 2-Technical Considerations)presents the evidence conceming their usage. 64 undoubtedly significant; the federal and provincial governments continue to operate a wide range o f vertical programs, many o f them effectively earmarked, conditional grants (see the third volume, Annex 2-Technical Considerations). Provincial governments use earmarlung to a very large degree. Fromthe perspective o f the districts, between 30 percent and 60 percent o f the total resource flows for development projects within their geographical boundaries are earmarked. Inthe study TMAs,between 19percent and 59 percent o f the transfers are within the appropriation authority o f the TMAs. Insum, Pakistanuses conditionalgrantsto avery limited degree-but when it does, it tends to favor earmarkingthrough vertical programs. The result i s that a large share o f financing for health, education, water supply and sanitation i s deliveredinways that systematically undermine local government accountability. Vertical programsrepresent attempts by higher tiers o f government to influence the policy and behavior o f lower-level governments by direct intervention. Retaining some influence over sub- national policy i s sensible. Federaland provincial governments need some levers to influence local priorities inorder to achieve national objectives such as minimumstandards ineducation and health, to foster internal economic union and to meet spillover effects (see Shah: 1994). In practice, though, vertical programshave a poor track record, and they are now damaging the nascent accountability, credibility and authority o f local governments. Pragmatism not purity Table 19 shows the relationship o fvertical programs to other forms o f conditional grants and to unconditional block grants. One goal i s to seek movements, where possible from left to right in order to reduce the damage that the transfers cause to sub-national accountability. This entails replacing vertical programs inthe medium term as the primary instruments for expressing federal and province government policies and, inparticular, PRSP service delivery commitments. 65 - Table 19: Characteristics and Impactof Vertical Programsand ConditionalGrants I Conditionalgrants Unconditionalblock I grants Vertical programs Most Least vertical vertical Shown inthe Higher level Sub-national Sub-national government budgetof: government govemment (but might require pre- authorization byhigher level gove-Fmen!).......... ............................................................................... ............... Conditioned Not conditional- Inputs Processes, i Not conditional on: based on project (earmarking), outputs or agreement 0'Sub-national............ outc!?.Ees PE9ceSs.!3 ................................i.......................................... .................. ......... Expenditure Parallel project Rely on existing provincial or local govemment control and facility managed govemment- accounts control accountability by higher level but with arrangements: government special arrangements (local imprest with either ex ante or ex post .C!?!??O!S) ....................... ......................................................................................................................................................................... Awarded Individual mnistries or departments : NFCPFC through: Degree that Parallel systems Parallel systems plus quasi legal Integrated with statutory systems are compacts arrangements by design made part of normal functioning of sub-national governments: ................................................. .......................... ................................................................................................................................ ~ Transaction Veryhigh- Moderate-proliferation o f grants j ................................................................................................................. i Low-few, strategic ~ costs: special purpose I creates demands on local conditional grants. unitsand administration establishments 'I . Likely damage to High 4 Low sub-national accountability The approach to that goal, however, needs to be realistic. A simplistic conversiono f all transfers into formula-based block grants mightpresent some risks. Inparticular, although it i s changing rapidly under the new devolved arrangements, the political economy still prevalent at provincial and local-government levels might provide incentives for provincial and local governments to use additional decentralized resources for other purposes than meeting PRSP goals.43Thus where the challenge i s to provide incentives to meet higher-level standards or otherwise to influence local priorities, then grants conditioned on outputs, outcomes or processes can be tools of a reasonable 43 This risk has also been noted in other settings. See (Andrews and Schroeder: 2003). 66 strategy. Grants should not be conditionedon inputs(that is, not earmarked on inputs)see Table 20, and should not attempt to cover multiple objectives within the same grant. Table 20: PrinciplesandBetter PracticesinGrantDesign Grant Objective Grant Design Better Practices Practices to avoid To bridge fiscal gap Reassignresponsibilities Tax abatement and tax base Deficit grants Tax abatement sharing inCanada Tax by tax shanng Tax base sharing To reduce regional General Non-matchingFiscal Fiscal equalizationprograms of General revenue fiscal disparities capacity equalization transfers Australia, Canada and Germany sharingwith multiple factors To compensate for Open-endedmatching transfers Republic of South Africa grant benefit spillovers with matching rate consistent for teachinghospitals with spill-out ofbenefits Settingnational Conditionalnon-matching Indonesiapre-2000 roads and Conditionaltransfers minimumstandards block transfers with conditions primary educationgrants with conditions on on standardsof service and Colombia and Chile education spending alone access transfers Ad hoc grants Influencing local Open-endedmatching transfers Matching transfers for social Ad hoc grants priorities inareas of (with preferably matching rate assistance as inCanada highnational but low to vary inversely with fiscal local priority cauacitv) Stabilization Capital grants provided Limituse ofcapital grants and Stabilization grants maintenance possible encourage private sector with no future participationby providing upkeep requirements political and policy risk guarantee Source; (Shah: 1994, 1998). Also in (World Bank: 1997). While inferior as a means o f transferring full accountability to sub-national governments, conditional grants represent a pragmatic approach one which recognizes that the federal government maintains a legitimate interest inprovincial, district and local institutionalreforms and service delivery arrangements. Table 21: BenchmarksinMovingfrom VerticalInvestmentProgramsto ConditionalGrants To the extent that they are necessary, conditional grants should aimto incorporatethe features listedinthe right-hand column o f Table 21. However, inthe first instance, data problems might require that these conditional grants take the form o f the transitional mechanisms also set out in Table 21, to bereplaced as quickly as data systems improve by conditional grants coordinated through the mechanism o f the NFC and PFC. Box 9 provides some illustrations. 67 Table 19 had highlighted the significance of coordinating such grants through the mechanisms of the NFC andPFCs, but also notes that realistically, this i s not attainable inthe short term. Box 9: Incentives for RestructuringinIntergovernmentalTransfers inthe RussianFederationand Uganda Incentives for restructuring in the Russian Federation A federal Regional FiscalReformFund(RFRF)has been established with World Bank assistance. Disbursements from the fund are made to regions that have chosen to enter a competition-conditional on their success inimplementing two-phase reforms o f public finance. Disbursement o f the first phase i s conditional on (inter alia): Comprehensive budgets Introduction o f 3 year budget framework Reduction inwage arrears Reductions innon-monetary receipts Publication inthe mass media o fbudget, debt and procurement details. Disbursement o f the second installment i s conditional on: Compliance with minimumquality standards inbudget management Establishment o f single treasury account Reductions inoff-budget resources. Conditional and Unconditional Transfers in Uganda The Local Government Development Program (LGDP) has beenpivotal inUganda's fiscal decentralization. Since the beginning o f fiscal decentralization inFY 1994, Uganda faced a familiar quandary. H o w could development resources be devolved to local governments to allow elected leaders to plan and allocate them according to local preferences, w h l e at the same time insuring the PRSP objectives for health, education, water, rural roads, and so forth would be prioritizedby local councils? By 1998 the need to address this question was becoming more acute: as a result o f HIPC, budgets for health, education, water, roads and other PRSP priorities were increasing rapidly. Rather than address the question, donors and central government hadpreferred to create additional vertical programs and highly restrictive conditional grants. Evaluations o f local government performance demonstrated that this was having adverse impacts on local government governance and undermining devolution. The LGDP was created to provide a common set of standards for judging performance and at the same time to channel resources to local governments inways that allowed local councilors to decide how they would be allocated. All local governments have access to a formula-based local government grant, however their access and level of hnding i s dependent o n the annual local-government assessment. Localgovernments are assessed on their corporate performance, against minimumrequirements and benchmarks related to areas such as planning, budgeting, financial management, engineering capacity-all o f w h c h are derived from existing laws. Those local governments that do not meet minimumrequirements get no local development grant. However they do remain eligible for a capacity-building grant enabling themto upgrade their performance. The hghest-scoring local governments inthe assessment receive 10percent greater allocations under the Local Development Grant. The LGDP framework has provided a strong incentive for local governments to upgrade their corporate performance. It shows that benchmarkmg can be carried out ina developing-country context o f limited administrative capacity. It also has become a common accountability platform for a range o f donor- financed, sector-specific grant systems. 68 Risks of Overuse Over-reliance on conditional grants can prove detrimental to local government performance and accountability. Disaggregated data are very difficult to obtain, but Table 22 suggests that typically, in federaljurisdictions, no more than 10percent o f federal-provincial transfers take this Table 22: ConditionalGrants inMajor Federations Federation Specific PurposeGrants Approximate Share Argentina Variety o f special funds 43 percent in 1997 (Haggard and Webb: 2001, table 5) Brazil Health and education earmarking Less than 10 percent as per 1988 Constitution (FundeflSUS), and guidelines on (Haggard and Webb: 2001, table 5) minimumexpenditures inthese Discretionarygrants 2 percent o f total state and sectors. municipal revenues in 1999 (World Bank: 2002)' LimitedDiscretionary Funding Mexico Aportaciones restricted (Webb and Less than half in 1998, and declining (Haggard and Gonzales: 2003) Webb: 2001, table 5) Nigeria EducationTax Fund(ETF) (World Bank: 2001) Russia Inthe late 199Os, various special or conditional grants accounted for over a third o f transfers to the regions (Martinez-Vazouez and Boex: 2001). South By 1998, adoptedpolicy to shift Africa toward unconditional grant programs ("equitable share"), and freeze levels o f existing discretionary grants (Bahl: 2001) United Wide array o f federal sectoral Over 90 percent o f U S federal grants to states and States programs local governments inthe 1990s were conditional or categorical. However, sub-national governments were only dependent on the federal government for about a fifth o f their revenues (Ma: 1997). Consequently, the share o f conditional grants in sub-national revenues totaled only about 18 percent. Moreover the extent o f deconcentratedvertical programs inquite limited. Inadditionto the generalgrant program, the federal government makes grants to the states and municipalities (which have their own constitutional status) for a variety o f specific purposes, and states also conduct projects o n behalf o f the national government. Moreover, there has been a significant use o f "voluntary" or "negotiated" transfers, often for politicalpurposes. The Source of funding matters Over and above these design features, any federal-provincial conditional grants must, o f course, be funded from the federally retainedportion o f the federal divisible pool. It will be difficult to argue for their funding from the provincial share o f revenue sharing to which the provinces can argue they have a constitutionalright. Province-local conditional grants are less constrained as local governments are legislative and not constitutional creations. 44 See (Craig andPorter: 2003) for an account of the negative consequences o f over-reliance on conditional grants in Uganda. 69 PRACTICAL ACTIONS TOWARD THE 7 GOALS The three-point strategy discussed above can provide apushto make devolution hlly operative. It supports the PRSP(Box 10) and serves to hurryhistory toward the following goals: 1. Devolved governance accepted as irreversible; 2. The federal government credibly committed to strengthen the future role o fthe provinces; 3. The provincial governments credibly committed to strengthen the future role o f local govemment; 4. Local governments challengedto demonstrate that they merit confidence; 5. A menu o f conditional grants enabled; 6. Local governments held to account; and 7. Fullseparation o f executive andjudicial powers. Box 10: Fine-Tuning Devolutionin the Some steps can bringimmediate relief. Others can PRSP prepare the ground for the medium term. Setting a series o f clear goals i s the foundation o f a tight- "Devolution i s now a ground reality. The challenge now is the imp1ementationOf loose strategy that maintains the core principles but all its components and consolidation that allows local diversity. To entrench the gains to include (a) some modifications such as the date, it i s important to move forward inthose areas direct election o fzila and tehsil nazimeen; that offer opportunities for visible wins, at the same (b) some additions such as federalto time avoiding needless battles for small gains. provincial devolution to strengthen provincial autonomy; and (c) federal and The actions proposedto meet these goals are not provincial restructuring." original but have been derived fi-om wide-ranging studies and consultations, supported by other recent Source: (Government of Palustan: 2003, p. reports (see Table 23). 591 70 - m s . -0 9 N r- 0 e o d r-: I Completing DevolutionAccepted as Inevitable Box 11: Fostering Debate on Decentralization "Decentralization i s a dynamic process. No country ever gets it right on its first try. Circumstances change, and the nature anddesignof intergovernmentalfiscal relations should change also. An important aspect of establishing an adequate institutional framework for decentralizationis thus to buildinsome `error-correction' mechanism,that is, to permitand encourage the adaptive development and evolution of the system inresponse to changes inneeds and capacities. "Muchmore informedandopendiscussionofthese mattersthannowprevails inmost countriesis needed. Regularpublication ofrelevant data would help, but one cannot rely solely on an interestedparty (the central government) to carry out, let alone publish, all the informationneededfor public discussion.In countriesinwhich intergovernmental fiscal issues are important, considerationmight thereforebe given to establishinga small nongovernmentalresearch institute focusing on local governmentproblems." Source: (Bird:2001). The goal i s to achieve a consensus among Palustan's governments, citizens and donor partners that stable and significant assignment o f taxing powers and functional responsibilitiesto elected sub-national governments i s the only credible strategy for lasting improvements in service delivery and achievement o f PRSP commitments. Achieving the goal requires engagement with a muchbroader range o f stakeholders (Box 11). Palustan's devolution was well served by publication o f the Local Government Plan 2000. This creative document set the agenda and established a common framework for the host o fparties to devolution. Since then, a slew o f ordinances, amendments, codes, regulations and guidelines, not surprisingly, have filled inthe space createdby the Devolution Plan2000. Other countries at comparable stages o f devolution have found it vital to revisit the original statements, like that published by the NRl3 March2000, and prepare a white paper to communicate clearly the nature o f the steps government intends to take to complete the devolution project. Such a strategic paper would revisit the original Devolution Plan, identify elements that may be reconsidered or configured differently (for example, the role of Public Safety Commissions at union, thana, and tehsil levels, or Village and Neighborhood councils), modified (for example, how to implement the LGO inthe Cantonments or Federally Administered Tribal Agencies), or newly introduced (for example, the proposed five-part independent service group). It would outline how remaining steps are to be introducedover time and present a renewed appeal to the political parties and donors. Most importantly, such a paper could assist inconsolidatingthe currently broad citizen support for devolution (Pattan Development Organization: 2003). Although the main political parties were initially opposed to devolution, it now enjoys near bipartisansupport. A white paper that sets out the potentialbenefits interms o f legitimacy and accountability that mightbe conferredby direct elections for district and tehsil leaders, as recently proposed by the government (Government o f Pakistan: 2003, p. 54), might also provide an opportunity for Palustan's political parties to make official, organizational commitments to devolution through their party manifestoes. Such a response becomes more likely ifthe policy statement explores possibilities for local governments to operate on an explicitly party political basis. This study has confirmed the key role that party affiliations continue to play inlocal government politics and the all-important relationship with province governments. 74 Box 12: Selling Decentralization inUganda After 6 years o f implementation, the Government o f Uganda in2001 designed and published a Fiscal Decentralization Strategy to streamline and communicate a process for moving the next stage o f decentralization forward ina way that enhanced local government budget autonomy, increased transparency-not just about fiscal transfers, but also importantpolicy decisions andperformance-and ensured that strong linkages were created withthe government's Poverty Reduction Strategy process. The Fiscal Decentralization Strategy, approved by Cabinet in2002, has beenpivotal inthree important respects. First it has clearly delineated to sector ministries and local governments the way forward on completing fiscal decentralization. Second, it has challenged donors to harmonize their procedures ina more devolution-friendly way. And third, the Fiscal Decentralization Strategy has helped gamer support for a broadly representative debate about the reforms involving civil society, parliamentarians, private sector and the civil service. Source: (Government o f Uganda: 2003, p. 2). For donors, the paper might reinforce the fact that for countries inthe process o f devolution and where a significant share o f spending for social services i s dependent on donor support, harmonizing donor procedures with devolution i s critically important (Box 12).45 Insuring that donors have sufficient, accurate, and timely information about progress and government intentions i s one part o f the equation. The other part i s insuringthat donors are reassuredthat the LGO, relatedordinances and the hard-wonbudget, accounts andprocurement codes form a common accountability framework for donor support. The goal: persuade donors that insisting on vertical programs to protect special-purpose project delivery arrangements i s less and less necessary. This study has documentedmany commendable efforts to performwell inthe face o fthe obvious constraints posed by a new and incompletely implemented Devolution Plan. But little i s known o f these efforts. Devolved governance i s seen as the only credible approach to a problem that has eluded Pakistan since Partition, namely to establish a set o f institutional arrangements capable o f efficiently convertingpublic resources into the delivery o f services. Nevertheless, remarkably few o f the many initiatives being taken by local governments receive the recognition they deserve. In country, the debate on devolutionremains somewhat polarized and superficial. Internationally, the contrast could not be greater between the depth and intensity o f academic andpractitioner debate on the future o f the big-bang devolution inIndonesia and the paucity o f evaluation and critique o f a reforminPalustan, which has, inmany ways, been more fundamental. One possible mechanism for highlightingthe progress that some determinednazimeen have made i s a series o f federal and provincial award ceremonies, celebrating particular local successes. Other efforts shouldbe made to increase the level o f research and balanced media coverage o f devolution. Action recommended: Issue a White Paper on strategic options for adjusting legal, administrative, political and fiscal arrangements in support o f devolution. 45 (Annan: 1999; Brown, Foster, NortonandNaschold: 2001; Development Assistance Committee: 2003; Foster and Mackintosh-Walker:2001;Teskey and Hooper: 1989) provide particular insights into the potentialuse of conditional grants within SWAPS. 75 DemonstrableFederalGovernmentCommitment to StrengthenedRole of the Provinces Timetablefor the PRSP Commitment on Functional Devolution The PRSP makes the point that deconcentration o f functions will not be complete untilfederal to provincial devolution takes place to strengthen provincial autonomy (Government of Palustan: 2003). The government hadpreviously announced its intention at the Pakistan Development Forum inM a y 2003 to examine the roles of federal government inrelation to those o f the Provinces and reallocate responsibilities more clearly to provinces on the basis o f subsidiarity (assigning responsibilities to the level o f government best placedto deal with the associated responsibilities, costs andbenefits). Candidates for disengagement by the federal government include: populationplanning, agriculture, local govemment, tourism development, transport and communications, delivery o f health, social welfare and women's development and education services.46This will entail the restructuringo f the federal level to eliminate unitsthat are unnecessarily duplicating provincial functions, such as education, health, and agnculture. It will also involve restructuringthe provincial level to strengthen capacity for policy formulation as well as winding down direct execution o f functions being assumed by the local governments. The proportionate cuts intotal employment would be small-around 3 percent o f the existing workforce-representing some Rs.27.1 billion inwage bill savings. The proportionateprogram savings would be more significant. However, the point i s not primarily fiscal although some fiscal sharing via the NFC, to matchnew functional responsibilities, including through a special federal grant out o f federal share of the divisible pool may be necessary. The significance o f the transfer o f functions i s that it demonstrates that the federal government recognizes that provinces are not just a problemto be bypassed inthe marchtoward devolved governance. Some degree o f hostility from provincial governments toward local governments and district governments, inparticular, was noticeable throughout the study. Untilrelations between the federal and provincial governments are adjusted so as to complete devolution, such tensions can be expected. Provinces unambiguously considered that they had lost power and influence as a result o f devolution. Devolution o f these functions would ease provincial governments' concern about becomingredundant inthe wake o f the major devolution to the district governments, would liftthe morale of the provincial civil service andenable the federal government to focus on its key strategic functions. Action recommended: 0 Federal government to confirm its planned timing for functional reassignment to provinces. Preparations in handfor the 7`j NFCAward now The problems o f inappropriate federal-provincial tax assignments and burdensome and expensive provincial debt are well known and have been debated for some time. 46 Similar arguments aremadein(Shah and others: 1996;World Bank: 2003a, p. 87). 76 Figure 10: The Spectrumof Autonomy inTax Existingprovincial taxes and user charges Revenues are inefficient and inequitable, and are Il incapable o f generating sufficient Own Subnationalgovemment sets tax rate and decentralized Tax revenues to meet a significant share o f Revenue Subnationalgovemment sets tax rate only provincial expenditure needs (Malik: Subnationalgovernment sets tax base onl? 2003; Shah and others: 1996; Revenue split may be changed \\ith World Bank: 2000). Admittedly, consent of Subnational government provincial tax effort has, inaddition, been Tax Revenue split fixed in Iegijlanon (ma! be Revenue changed unilaterallv by the central . . lacklusterbut insignificant part this has Sharing govemment) been due to reliance on numerous, Revenue split determinedby the central govemment frequently changing, low-yield tax Central govemment sets rate andbase of centralized instruments, and poor tax records that Subnationalgovemment tax have created strong incentives for tax Source: Developedfrom (Ebel andYilmaz: 2002). evasion. Two measures have been consistently recommended. First, while collectionremaineda federal responsibility, the provincial revenue base could be expanded by allowing sharing o f selected federal tax bases (income tax and possibly GST) through provincial add-ons over and above the federal rates. Second, some taxes and charges can be devolved to provincial governments (for example, the specific-purpose excise duties such as those on tobacco products) (see Figure 10). The rationale for these moves was most recently developed in (World Bank: 2000,2001) and emphasized in (Malik: 2003). A further andmore complex question concerns an expansion inthe base o fthe GST on services. This apparently attractive propositionis fraught with technicalproblems, as while this is (by default) a provincially assigned revenue, most o f the services on which it should be charged are federal functions (telecoms, electricity transmission and distribution, gas distribution, a large proportion of transport, including air travel, railways, and sea travel). The provincial governments have an interest inexpanding the GST base to these services, and a committee has been established to finalize recommendations on this matter. Most o f these services are taxed by imposing a federal excise duty or GST while treating the relevant outputs as goods rather than services. This i s not the only problem. At the moment, the revenue from the tax i s distributed as per the NFC formula. As a result, the provinces disagree on which services to expand the GST to. For example, since the bulk o f financial services are located in Sindh and Punjab, NWFP and Balochistan would like the tax to be expanded on these services inorder to obtain more revenue without either significant impact on economic activity inthe province or major political fall-out. Despite these difficulties, on balance this i s a sensible policy option. The challenges faced by the provincesas they labor under a highburden o f expensive Cash Development Loans (CDLs) are also well identified inrecent studies (World Bank: 2000).47In addition to explicit liabilities, the provinces have substantial contingent ones, mostly inthe form o funfunded pension liabilities (pensions are pay-as-you-go) and debt guarantees and workers' compensation obligations o f provincial autonomous bodies (Table 24). Due to the rapid growth in provincial employment inthe early 1970s,pensionpayments have increased sharply. 47 CDLs constitute the overwhelmingmajority of all provincial debt. The other debts are marketloans, foreign exchange loans on-lentby the federal govemment and floating debt-the bulk o f it acquired inthe mid-1990sas a resultof the 1997NFCprovision permittingprovincesto borrow fromthe State Bank of Pakistanto meet short- termcash flow. That leeway inadvertentlyencouragedfiscal mismanagementinthe provincesby allowing their governmentsto finance their revenuedeficitsthroughborrowing. 77 Tab1e24: Urgedby the federal g~vernment,~'provinces Outstanding DebtAgainst Provinces, July 2003 have usedthe opportunity presentedby recent structural adiustment and uublic-resource Province Debt Interest (2003- World Bank and ADB to seek federal government (2003) 04) dispensation for early retirement o f the CDLs. Punjab 82,983.28 12,237.44 Their action has created a more sustained fiscal Sindh 40,121.60 5,936.27 space at the provincial level, since the interest NWFP 35,427.3 1 4,298.16 payments include a significant subsidy from the Balochistan 15,386.22 Inpreparing for the nextNFC Award, the government mightalso wishtopropose the size and scope o f a standing secretariat with a significant technical capacity to evaluate options for future Commissions and prepare detailed policy alternatives. Box 13 provides some comparisons. Box 13: Technical Capacity Available to Other Intergovernmental Finance Commissions The Finance and Fiscal Commission (FFC) inSouth Africa was established by Section 220 o f the Constitutiono f the Republic o f South Africa. The Financial and Fiscal Commission was institutedto provide advice withrespect to the equitable sharing o f revenue and other intergovernmental fiscal issues, and to limit subjective political decisions regardingthe allocation o f public resources inthe spheres o f national, provincial and local government. The FFC i s a permanent expert commission. Acting as an advisory body, it i s mandated to make recommendations on financial and fiscal matters to parliament, the provincial legislatures and any other institutions o f government when necessary. The FFC i s separate from government and therefore is able to perform impartial checks and balances betweenthe three levels o f government. It facilitates cooperation on intergovernmental fiscal matters. Excluding the outsourced legal advice and humanresource functions, the FFC has 23 permanent professional staff, divided between a finance administration and operations department and a research program. Staff include a librarian, information specialists and specialist researchers. 48 The FinanceMinister urgedprovincesto developprovincial debt-repaymentstrategies so as to reduce their debt servicingcosts when presidingover the first meetingofthe reconstitutedNationalFinance Commission(NFC) to discussthe 6thNFC Award onNovember 19,2003. 49 The FederalMinister of Financehas announcedthe government's willingnessto revise the rate of intereston loans extendedto theprovincesinnextyear's budget.(Dawn, February22,2004). 78 Action recommended: Establish HighLevel ExpertGroup on federal-provincial tax reassignment (including possible debt-for-performance swaps), to report within 2 years Ensure adequate technical capacity and resourcing for standing NFC secretariat DemonstrableProvincialGovernmentCommitmentto StrengtheningLocal Government Reassignment of province administeredfunctions to local governments completed Mirroringthe need for federal-to-province transfers o f functional responsibilities, functions mandated to local government but currently retained by provinces must also be shifted to the next level. The evident mismatch between responsibility and accountability for water and sanitation i s a clear example o f incomplete devolution. The incomplete transfer o f authority over service providers both inspecialist agencies (for example, WASA, Karachi Water and Sewerage Board, and the Sindh Katchi Abadis Authority) and province departments (for example, PHED/LGRD) has resulted inthe unevenretention o f provincial control. To the extent that it i s not completely cynical about local government performance (CIET: 2003), the public wants to see action and i s unlikely to be impressedby nuanced points concerning the inability o f councilors to respond because they lack control over the administrative machinery. The current situationhas led to some significant problems for the credibility o f newly elected district and TMA leaders who are unable to deliver on improved water and sanitation services. Box 14: The PRSP Recognizesthe Needto Progress has been made inPunjab and Sindh. ResolveJurisdictionalRivalries InNWFP andBalochistancompletingthe transfer o f authoiity will require significant "At this point, there is some degree of overlap and fiscal transfers also, not least to cover salary mismatch between responsibilities for design o f the program, implementation and delivery. Ths is recurrent expenditureof R ~3.0 ~ million ~ 4 in partly due to the ongoing devolutionprocess NWFP and Rs. 134.1 millioninBalochistan. devolving from upper level to lower levels o f However, inaddition to transferringthe staff government, and thus, the government's effort in to the TMA, Some effort must be made to deal the next 2-3 years will need to better align with the large accrued debts o fthe TMAs. accountabilities and expenditures, resources, and Otherwise, those obligations will preventthem administrative and policy functions." from developing new schemes or maintaining large existing schemes and thus prevent the Source: (Govemment of Pakistan: 2003, p. 68). effective implementation o fpublic-health Inreducingthe debt burden, however, itwillbe importantnotto signal any softeningo fthe budgetconstraints for TMAs. Recurrent non-salary liabilities, which historically have been largely comprised of electricitjr charges, could be subsidized through cost sharing with the district and with no impliedcommitment for future accumulation o f liabilities. Alternatively, TMAs could be expected to bear the costs through the mediumo f their shares inbuoyant, GST-based transfers. Provinces can, for instance, adjust TMA non-salary liabilities through a debt-for-performance swap (Box 15). The provinces can prescribe process and fiscal targets including cost recovery and linkprogress to drawing downa conditional grant. Major amounts o fthe grant should go toward financing (or adjusting) liabilities with a significant amount left for local, discretionary allocation. Inaddition, incentive grants shouldbeprovidedto TMAs once they meet the targets and get write-off for local discretionary allocation. 79 Box 15: "Debt Write-off for Performance" in Punjab Punjab's 122 TMAs are burdenedby unpaid debts, electricity bills andpensiondues o f the staff transferred to local governments. Recently, the Department o f Finance, Government o f Punjab, estimated these liabilities at Rs. 3.7 billion-more than ten times Punjab's total UIPT collection inFY 2003. And for the 15 TMAs inPunjab for which data was compiled under the DSD, liabilities amount to 292 percent o f their own-source collection for FY 2003. The Punjab government announced a one-time waiver o f these liabilities, but has decided to provide this as a conditional grant to be sure that a hardbudget constraint and incentives for improved local tax collection as well as important non-tax revenues like water and sanitation charges are reinforced. The grant i s conditioned on TMAs' undertaking an analysis o f collection inefficiencies and preparing and securing council endorsement o f a plan for raising own-source revenues over 3 years. The Department o f Finance, with t e c h c a l assistance from the Decentralization Support Program, will help TMAs to meet their obligations andbe eligible for the grant. It i s possible that matching grants may come from the Province under PFC to helD offset the Political costs o f local action. TMAs have varying technical capacity. Groups ofTMAs canmeet the requirementbyjointly financing senior engineers, thus allowing them to bringthe technical and financial sanction powers within their administrative scope and ending their reliance on province offices. Action recommended: 0 TMA debt-for-performance swaps. 0 Transfer o f PHED staff andbudgets to the TMAs Incorporate the Lessons of Transition in Revisions to LGOs, Rules of Business and Delegation of Power. The transition to devolved governance has highlightedthree key areas o f the LGOs and secondary legislation that have not resulted ineffective local structures. Although this failure i s due inpart to weak implementation, the NRE3 and other key agencies have signaled the time has now come to revisit the legislationrather thanjust reiterate what it originally intended. The Delegationo fPower instruments havemeant that many district staffnow have less financial authority than before devolution. Inmany districts, EDOs are category I1officers, whereas the equivalent position previously had category Ifinancial powers. As the Principal Accounting Officer i s a Category Iofficer and therefore has higher financial sanction and procurement powers, the budget preparation process and expenditure authorizationwithin the districts i s undulycentralized. There is a similar needto enhance the technical sanction powers o ftechnical staff inTMAs. Among other problems, the low level o f sanctioning authority provides the pretext for the recentralization o f Public Health Engineering to the Province. Monitoring Committees are not performing as envisioned. Among the reasons: improper constitution o f the committees without a formal vote by the Council and denying the opposition any opportunity to participate; lack o f training o f the councilors; inadequate traveling allowances and transport facilities; and inability to enforce decisions. The Local Government Commissionshave no authority to impose a settlement and are seen, largely, to represent the interests o fthe province rather than as unbiased facilitators o f the devolution process. Refinements to the LGO, the Delegation o f Powers and the Rules o f Business can resolve these problems. A review could provide the opportunity to use these instruments as further incentives 80 to local governments to perform. For example, the LGO could allow district governments, subject to a satisfactory business plan, to amend their structures and merge or bifurcate some departments-allowing for the possibility that large city districts do not needthe same structure as smaller rural districts. Action recommended: 0 Province-by-province review o f the LGO, Rules o f Business and the Delegation o f Powers inorder to improve the functioning o fthe Local Government Commission and the facilities and authorities providedto the local government Monitoring Committees and to recognize well-performing local governments by allowing themto move to more relaxedregulatory regimes. Major commitment to capacity development Capacity i s primarily a function o f institutions rather than the reverse-and so training and traditional capacity-buildinghas little lasting impact ifthe basic incentives are weak. However, while there are many shortcomings inimplementation o f the array o f new local institutions introducedunder the Police Order and Local Government Ordinances, there i s evidently great need for province by province capacity development: includingbasic orientation and training o f elected and administrative officials, and the provision o f adequate facilities and equipment. Although the Devolution Transition Grants provided by federal government inthe last two fiscal years addressed this need for many local governments, a large number still require basic facilities. This i s most markedinBalochistanwhere local governments, particularly at the tehsil levels, lack basic structures and equipment necessary for maintenance o f even the most basic records and files essential for local government functioning. Inthe initial stages ofdevolution, the NRBshouldered theresponsibility for capacity development, through their advocacy o f the Devolution Transition Grant, support for Devolution Transition Teams, public awareness and special purpose training activities-such as inrespect o f Chapter XI1and XIIA o f the LGO-this i s no longer feasible or appropriate. It i s essential that provincial governments, together with local governments address this requirement directly since they have a particular interest and responsibility to see that this occurs ina way that i s tailored to needs inboththe administrative and legislative organs o f government. Ongoing province-focused public sector reformprograms5' are providing targeted support to departments responsible for finance and planning, only sporadic efforts are being made to reorient the key line departments51most affectedby devolution. Staff o fprovincial departments are aware that devolution has substantially changed their role with respect to direct service delivery, but few ineducation, health, water and sanitation, or inthe productive sectors havebegunto develop capacity for their newroles in sector policy and standards or technical support and mentoring o f their corresponding local government departments. More concerted efforts are required inrespect o f new institutions. The Province Finance Commissions, through support from the DecentralizationSupport Program, are actively engaged inresearch, training and other preparations for the upcoming Awards to local governments, but there i s a glaring absence o f similar efforts with respect to the Local Government Commissions. Similarly, the lengthy disputes regardingthe police complaints authorities and Public Safety Commissions has been a convenient excuse for the lackluster support provided to these institutions.N o w that agreement Supportedby the World Bank inSindhandNWFP andthe ADB inPunjab andBalochistan. Exceptions are the World Bank supportedEducationSector Reformprogramin Punjaband the SindhDevolved Social Services programsupportedby the ADB. 81 appears to have been reached,s2there i s pressingneedto prepare and notify the Police Rules and embark on a major capacity development effort for these critically important bodies. Similarly, training and other support i s also necessary for the members o f provincial assemblies. Although UNDP andUSAID are providing some capacity-building support there remains a critical needto ensure province policymakers are oriented inhow local government i s intendedto work and in the steps necessary to activate critically important province bodies-such as the Accounts Committees and other assembly committees. At local government level, this report argues that the core principles o f devolution shouldnot only be reconfirmed, but that diversity should be encouraged inhow they are applied. This principle might apply notjust to how local government structures are created-bearing inmind the obvious differences inrequirements from rural Balochistanto central Karachi-but also in approaches to local capacity development. Inrural Balochistan, for instance, NGOs have considerable experience with provision o f water supply facilities, inPunjab, private sector agencies, includingNGOs, play a large role inprimary education. Responsive capacity development requires arrangements through which local governments can contract-in such services from a range o f sources, includingprovincial agencies, NGOs and the private sector. This demand-led approach to capacity development inturnwill require that local governments are also encouraged to develop their capacity to become `clients' for these soft sector services, and that NGOs and private sector firms-in planning, accounts, audit, procurement, evaluation and training-learn to respondto local governments as consultants inaddition to their emerging role inthe implementation o f province programs. Several existing modalities could be utilized for capacity development. First, as noted, there i s a plethora o f vertical programs with a primary focus on delivery o f services. Typically, these include efforts to buildlocal capacities inprogram execution. As argued earlier inthis section, a strategy i s needed to convert these programs and transit toward greater use o f formula based transfers. This presents an opportunity to similarly redefine the focus and purposes o f support for capacity development under these programs so as the progressively integrate `executive' functions-planning, budgeting, expenditure controls and accountabilities-within local government operations. Second, the multi-donor supported Devolution Trust for Community Empowerment i s presently contractingNGOs to support CCB-local government relations. Additional support i s required, to extend this effort and to ensure it i s well attuned to local variations inneeds. Third, to promote better matching o f capacity development to local needs, there i s scope to expand arrangements being made under the Decentralization Support Program whereby local governments and provinces reach agreement on specific training and relatedneeds through a participatory planningprocess, jointly appraise proposals and monitor the quality of services provided by private contractors. DFID, USAID and the ADB are consideringproviding local government capacity development grants linked to development grants within the PFC system. Under this arrangement, local governments that perform well, that i s effectively utilize capacity development grants and demonstrate performance against common eligibility criteria, would qualify to access further budget support which may be utilized through their development plan and budget process. Action recommended: Training for provincial staff ineducation, health, water and sanitation departments concerning their new roles insector policy and standards or technical support and mentoring o f their corresponding local government departments. ~ 52 See footnote 16. 82 0 Targeted support for Local Government Commissions, Police Complaints.Authoritiesand Public Safety Commissions. Training for the members o f provincial assemblies concerning how local government i s intendedto work, and inthe steps necessary to support the Accounts Committees and other assembly committees. 0 Training for local government staff inmanaging the contracting-in o f services from a range o f sources, including provincial agencies, NGOs and the private sector. PRSP (Box 16) and elsewhere (World Bank: 1999,2000,2001). Box 16: The PRSP Recognizes the Need for Further Tax Devolution For thedistricts,muchofthedebate "Presently the local governments depend primarily upon fiscal revolves around the possible transfers from province governments for meeting their Of A~icultural Income expenditure requirements. Localrevenue mobilization i s an Tax to the districts. Currently a important activity for the sustainability o f local govemments. Future strategy will look a further devolution of taxes to the Provincial tax inall four Provinces, local levels, widening o ftax base, rationalization of local AIT is a contentious issuein taxes and improvement o f collection and recovery Palustan, inpart because the mechanisms." Constitution i s silent about this tax. H ~since ~generates agriculture ~ ~ Source: (Government of Pakistan: 2003, p. 56). ~ ~ , 83 Box 17: Potential Short Term Gains f r o m AIT Devolution o f AIT i s technically Devolution complex. On the one hand, converting the tax into a land-based farm tax on Assuming that a 20 percent increase inAIT collection small(withnoexceptions)and farms were to result from the decision to share AIT collection withprovinces on a 50-50 basis, the added revenue would transferring assessment and collection represent about 5 percent o f Bahalwalpur's development responsibilities to the districts4r just budget and 4 percent o f Faisalabad's. However, an added sharing the tax directly with the incentive could be created by the provinces' agreeing to districts-ou1d create a and pass its 50 percent share back to districts as a matching much-needed additional resource base grant (that is, collection at district level, half given on the for them. On the other, the tax itself i s basis o f the point o f collection and the other half as a anomalous as it differentiates between matching grant), with the province remaining responsible agricultural and other forms o f for setting rates and controlling policy. Onthis basis, it i s income+pening the door to reasonable to assume a far greater increase incollection. A 50 percent increase incollection, along with the manipulationandevasion. ~ M ~it ~ ~ revenue from the matching grant, would amount to 12 i s administratively problematic to percent o f Bahawalpur's development budget and 10 devolve as land registries are held at the percent o f Faisalabad's inFY 2003. InFaisalabad's case, provincial level (World Bank: 1999, this is equivalent to half the total spent on agriculture in 2ooo). FY 2003. For Bahawalpur, the increase amounts to twice its outlays inFY 2003 on Social Welfare. For the TMAs, the Urban Immovable Property Tax holds significant revenue Source: ADBIDfIDIWBDSDDataset2. potential, and there i s scope bothto Box 18: PotentialRevenue Gains for TMAs in Transferring UIPT Collection Responsibility to the Districts Itis difficult to assess the potentialrevenue gains for the TMAs o ftransferring collection responsibilityto districts. A 20 percent increase inUIPT collection would mean that Rs. 118.6 million in additional resources would be available inPunjab for allocation. A 20 percent increase is not an unreasonable expectation, indeed for many TMAs where collection is particularly low, t h s rise would be within easy reach. As a percentage of existing development spending, the increase would be significant, even for predominantly rural TMAs. Although for TMA Jaranwalla the projected increase represents only 4.8 percent o f development spending, this TMA has a very highproportion o f development spending inthe total budget. For other rural TMAs the gain would be considerable. For Yazman inBahawalpur, the add- o n represents a 48.8 percent increase inFY 2003 development spending. ISource; ADBiDfIDIWB DSD Dataset 2. Itis often notedthat collection could be rationalizedbymalungthe districtsresponsiblefor collection in all circumstances, broadly maintaining the current revenue-sharing arrangements between them and the TMAs (World Bank: 2000,2001). Given the lack o f buoyancy inproperty tax inrecent years, despite tremendous growth inresidential and commercialproperties, there i s a strong case to argue that the transfer o f collection duties would give local councils a much greater incentive to utilize this tax. Inaddition, other administrative constraints must be removed and a general survey will be required to update valuation tables and capture new properties. The 84 assessment formula will also needrevision along with a review o f the administrative structure and computerization o f records (Malik: 2003).53 Box 19: LocalInnovation: Property Tax Enhancementthrough Self-Assessment inHyderabad The Indian cities o f Hyderabad and Bangalore have introduced self-assessment, property-tax systems to good effect. Bothmunicipal corporations decided to place the responsibility for tax assessment on the individual ratepayer rather than municipal employees. The aim was to increase revenue and reduce the scope for collusion, rent-seelung and corruption. When the process began, there hadbeen n o revision o f assessments inHyderabad for at least 20 years. T o tackle the likely deadlock o fblocks, appeals and litigation associated with tax revision the Corporationintroduced self-assessment in 1999 based o n existing legalprovisions to make a return true to the best o f "your knowledge." Owners and occupiers submitted returns guidedby residents' welfare associations that hadnegotiated on plausible rate guidelines. Sample verification occurs to track untruthful filing, which-where identified-results in higher assessmentpenalties or criminal charges with publicity. Revenues more than doubledinthe first three years. Source: ADB staff files. Further devolution o f collection to the districts has already occurred on a pilot basis. However, the situation inNWFP highlightsthe complexity o f the choices involved. NWFP has recently transferred collection o f UIPTto local govemments inNowshera and Swabi, while keeping the assessment at the provincial level. The result has been a drop incollection attributable, according to the provincial Department o f Finance, to capacity problems at the district Box 19 illustrates other innovations. Again mirroring the situation inrelation to the federal-provincial assignment o ftaxation, awareness o fthese issues has produced little consensus on remedies. As a result, the imminent PFC awards will not reflect any proposals for significant changes intax assignments. An opportunity has been missed. To avoid missingthe next one in 3 years, provincial governments should establish formal reviews o f options for reassigning these taxes, with a reporting date at least 1year before the anticipated end o f the forthcoming PFC Awards. There i s however a significant difference betweenthe provincial govemments' preparing options ahead o f the next round o f PFC Awards and the federal govemment's preparing options ahead o f the nextNFC Award. First,this advance move requires some acceptance on the part o f the federal government that provinces can and should follow separate paths. There i s no intrinsic reason why the devolution o f AIT and UIPT should look the same inNWFP as inPunjab. Second, realistically, some provincial governments have less capacity than the federal government to undertake the long-termperspective analysis required and will need additional incentives to stimulate timely, effective action. Inpreparingfor the next PFC Awards, the provincial governments mightwishto propose the size and scope o f a standing secretariat, with a significant technical capacity, which can evaluate options for future Commissions and prepare detailedpolicy alternatives. 53 Other innovations are also possible. Suggestions are also advanced from time to time to give the property owner the option to assess the rental value o fhis or herhouse and paythe property tax accordingly. If,on revaluation by the ETD, ifit is found lessthan the market value or rental value, the owner would berequiredeither to offer the house at self-assessedvalue to the govemment or pay the differential with high penalty rate. 54 The point that there are no certainties in reforming land and property taxation is well made in (Birdand Slack: 2002). 85 Action recommended: Provinces to announce formal reviews o fprovincial-local tax reassignment, to report within 2 years. changed? Has council consideredthe long- term, recurrent-cost implications o fnew Box 20: Reducing the Throw-Forward inNWFP development spending? The NWFP Government has been trying to rationalize Improving the efficiency o f resource its ADP to avoid the practice o f allocating nominal amounts to ongoing schemes and to reduce throw- allocationwill entail three responses by forward liabilities that had accumulated to such an local governments, as illustrated inBox 20 extent that they equaled 3.24 years of the current for NWFP. First, a clearer link i s needed ADP. Recently the province went through successive between investment decisions and a rounds o f reprioritizing projects (breaking them into 4 medium-term planning process. This priority groups, o f which they only fimded the top 3) requires, at heart, that investments be better and consequently reduced the ADP throw-forward 8 connected to reliable indicators and percent below the level o f FY 2003. As a result o f the knowledge o f local needs and o f likely reprioritization, the number o f new projects for FY resource availability beyond the annual 2004 was reduced from 980 to 723, and the number o f completed projects i s expected to increase from the budget-planning process. It should be 342 originally targeted to 4 13 for FY 2004. possible now for districts and TMAs to work out likely revenue over a three-year period I Source: Govt. of NWFP calculations. (the PFC award and OSR projections) and indicate how income will be distributed across different sectoral priorities that may inturn shift over time according to council's priorities. 55 Throw forward canbe defined as the claim on future developmentresourcesby the existingportfolio of development projects 86 Box 21: The Scale of the District Throw-Forward Problem For four o f the districts inthe study (Karachi, Faisalabad, Bahawalpur and Khairpur) information was compiled concerning throw-forward o f district ADPs on June 30, 2001-immediately before the elected local governments took charge. The results were contrasted with snapshot data concerning the throw- forward in2003104 to assess the impact o f the inherited throw-forward, its claim on districts' development resources and whether the district governments have been ina position to start new schemes. The main conclusions are: The four district govenunents combined inherited a throw-forward o f about Rs. 3 billion, which would have taken district governments 2.6 years to complete at the 2001102 level o f allocation and if no new schemes were started inthe interim. Incomparisonthe throw-forward in2003104 is 28percent lower, mainly becauseKarachi and Faisalabad cut their throw-forward significantly, whereas throw-forward o f ongoing schemes increasedmore than threefold inBahawalpur and 19percent inKhairpur, indicating that the smaller districts initiated a large number o fnew projects in2002103. The overall, inherited and self-generated throw-forward in2003104 still represents a significant constraint on the ability o f district governments to start more new projects. See the thirdvolume (Annex 2-Technical Considerations) for further details. Second, medium-term plans and annual budgets needto reflect a clear connection between development proposals and recurrent cost requirements. Third, andmost immediately, local governments needto rationalize andreduce their development throw-forward. Most have a large portfolio o f incomplete projects that shouldbe reviewed to cull redundant or unviable obligations and to prioritize the revised list for completion ina phased manner. NWFP (Box 21) has recently achieved some progress at the province level; most district governments are capable o f doing the same. Action recommended: Local governments to rationalize and reduce their development throw-forward, culling redundant or unviable obligations prioritizing the revised list for completion ina phased manner. Coverage of user charges increased Inthe six study districts, there are encouragingtrends inown-source revenue collection. Appropriate levying o f user charges is, again, frequently discussed and has been extensively reviewed (Malik: 2003; World Bank: 2000).s6However, user charges are politically difficult. Although districts mightrecognize significant potential for increased user charges inthe health sector (outpatient services, inparticular) and insecondary education, they are likely to judge that the political costs outweigh the likely fiscal gains. A further disincentive i s that, ingeneral, effort at the facility level i s currently not rewarded. For example, where modest user charges have been levied ineducation, they have been collected at the level o f the school and then sent to the district for reallocation. Inhealth, outpatient department fees are more often than not deposited inthe provincial account. Determination and imaginationmightrealize some short-term progress. For example, at the TMA level, Punjab has provided some incentives to TMA nazimeen for cost recovery inwater and 56 The broader utility of user charges is well set out in (Bahl: 1999; World Bank: 2003c), and a usefulassessment of the challenges involved is provided in(Bird: 2003). 87 sanitation by waiving electricity arrears, loans liabilities and pension liabilities ifcost recovery matches the TMA's own expenditures on water-supply systems. There are early but impressive results (Box 15).57 Selective application o f user charges could establish a visible linkage betweenthe paying public and the quality o f services that they value and for which they are willing to pay. However, this move requires a careful political strategy; province politicians would, inmany cases, be only too happy to portray districts as rapacious and unconcernedwith the poor. Since their policies determine what services can and cannot be charged for, the provinces would have to feel some benefit from the change. Action recommended: Provinces to ensure that user charges are returnedto local governments. Districts to identify opportunities for increasing user charges inhealth and secondary education, providing some incentives at the facility level for collection. Strong interjurisdictionplanning and coordination Crafting workable arrangements for resolving disputes and making consistent plans and decisions is a challenge for any multi-representative electoral system. Palustan is attempting an innovative approach to inter-local government coordination and integration. CalledMushawarat Committees, they aim to achieve through a nonhierarchicalpolitical process o fnegotiation and compromise that which i s frequently achieved elsewhere only through an administrative h i e r a r ~ h y . ~ ~ The public credibility and fiscal efficiency o f local governments depend on concerted efforts to clarify jurisdictional responsibilities and demonstrate that workable arrangements exist for coordinating planning and implementation o f vertical programs both among tiers o f local government and with province governments (Box 22). The Mushawarat Committees and DDCs provide highly visible platforms to nurture this capability. The former could provide a platform for district-wide, medium-term planning as well as resolve common disputes aboutjurisdictions and responsibilities. DDCslikewise present a forum for district, province and federal political representatives to ensure resources are usedmore inwiser, more publicly credible ways. Both may be usedto encourage civil-society engagement indebates about local service delivery policy, plans and budget processes. Although it i s true that Mushawarat Committees might offer a rationalbasis for the effective coordination o f development programs at the district and TMA levels, it i s not inthe individual interest o f the district or any single TMAto ensure their success. The political incentives discussed earlier make councilors value schemes that publicize their individual contributions to their constituents more than they value coordination. To break the impasse and tilt the balance in favor o f coordination, additional incentives mightbe needed from the province. 57 See also (Sidat Hyder Morshed Associates (Pvt) Ltd.: 2003). 58 Since October 2002, the challenge of institutionalizing this new approach has been increased by the resistance o f MNAs and MPAs who can see successful local government planning, investmentand disputeresolution as undermining their position as coordinators. 88 Box 22: Multi-agency Planning in West Bengal and Kerala The District Development Plan (DDP) and the Metropolitan Development Plan (MDP) process has been established inWest Bengal as a planning system to integrate the activities o f local governments and other spending departments at district and metropolitan levels. District Planning Committees have been constituted inall districts inaccordance with the District Planning Committee Act 1994. The DDP i s intended as both a comprehensive multi-sectoral plan and as an instrument for optimizing and equalizing resources. An infrastructure plan is an essential derivative o f the comprehensive plan. At the same time GoWB i s the first State inIndia to have made provision for a Metropolitan Planning Committee (MPC). One was established inOct 2001 for the Kolkata Metropolitan Area (KMA) to converge and guide the various local DDPs within a wider strategic framework o f economic, social and environmental development. The M P C has 60 members, 40 o f whom are elected by and from the elected representatives o f urban and rural local bodies inKolkata Metropolitan Area. A similar system o fintegrateddistrict planning has beenestablished inKerala where 40 percent o fthe outlay o f the State Five Year Plan has been allocated for decentralized planning and development mediated through local government. The first stage o f local planning i s based on a participatory planning process known as the Peoples Plan Campaign. Based on a strategic perspective inevery gram panchayat, these strategies and needs assessments are then formulated into the Annual Plan by specialpanchayat and municipal task forces. The Annual Plans are then consolidated into block and district plans to be spatially and functionally integrated into a district Perspective Plan. Source: ADB staff files. Action recommended: Districts to empower the Mushawarat Committees to provide a platformfor district-wide, medium-term planning. Local governments regulate and engage withprivate sector serviceproviders As the private provision o fhealthand education grows, the role o fthe responsible district departments must become overseer andregulator o f the private sector as well as provider o f government services. Currently, most district health or education departments maintain no record of services providedby the private sector, and are poorly equipped to promote or regulate the sector. A first, near term step toward rectifying this weakness would be for provinces to promulgate local government public-private partnership rules (as under LGO Schedule 5, Part A). While the role o f local government i s not necessarily limited inthis way, these rules should, in the interim, focus on two types o frelationship: 1) where the local government i s outsourcing services, which potentially includes the entire local government mandate; and 2) where local governments are regulating the market, for example, inrapidly expandingprivate-sector delivery o f health and education services. Action recommended: 0 Provincial governments to develop and promulgate local govemment public-private partnership rules (as under LGO Schedule 5, Part A). 89 Judicious Selection from a Menu of Conditional Grants Cautious use of conditional grants As noted earlier inthe section on Transmit National Policy Priorities More Effectively, at the federal level, Pakistanuses few specific purpose or conditional grants compared to other types o f transfers-and when used, the form o f conditionality preferred for provincial-local transfers i s earmarking. One goal o f reform i s to reduce the damage that such transfers cause to sub-national accountability. However, an overly rapid conversion o f all transfers into formula-based block grants i s a hostage to fortune, as it would leave the federal and provincial governments without levers through which to communicate policy preferences that mightbe at odds with local politically determined concerns. Recognizing that to some degree they runcounter to the logic o f transferring full accountability to sub-national governments, conditional grants represent a pragmatic approach which recognizes that the federal and provincial governments, respectively, maintain a legitimate interest in provincial, district and local institutionalreforms and service delivery arrangements. As set out above, these are nuanced instruments and there are three concerns to be borne inmind: 0 They can only be employed at the margins-ver-reliance on conditional grants i s likely to be detrimental to local government performance and accountability. 0 Itis important to coordinate such grants through the mechanisms o fthe NFC andPFCs, although notingthat this i s not attainable inthe short term. 0 Any federal-provincial conditional grants must, o f course, be fundedfrom the federally retainedportion o f the federal divisible pool as provinces can argue they have a constitutional rightto the provincial proportionso fthe shared revenues. This constitutionalconstraint does not apply at the provincial level. Withthese constraints inmind, there are some likely candidates for conditional grants at boththe federal and provincial level. At the federal level, inassistingprovincial-local tax reassignments, the federal government could, inprinciple, propose special conditional grantsto provincial governments, necessarily made outside o fthe NFC, conditioning access to the grant on the existence of a publishedplan for reviewing tax assignments and amounts payable that reflects the coverage and success o fpilot projects. Provincial governments could propose conditional grants to districts, initially outside the PFC ifit i s too late for their inclusion inthe imminent awards, which: 0 Provide incentives for the reduction indevelopment throw-forward 0 Provide matching incentives for the increased application o f user fees within a range acceptable to the province 0 Provide incentives for districts and the associatedT M A s to develop integrated district plans through empowered Mushawarat Committees and to fund schemes consistent with the plans. Bearing inmindthe significance o f using conditional grants sparingly, then the effect o f such conditional grants could also be achieved by government requestingthat such conditionalities are included inprovincial budget support credits or loans provided by donors and the multilateral lending agencies. 90 A c t i o n recommended: Federal and provincial governments to consider using conditional grants, selectively andifnecessary outside ofthe NFCPFCs initially, to achieve overarchingpolicy goals. Existing verticalprograms reviewed andproposals redesigned Vertical programs remain significant features of intergovernmental relationships. Inadditionto the government-funded vertical programs, donors provide extensive support for federal and provincial vertical programs. For local governments, highlyvertical programsnot only undermine accountability and operational efficiency. Additionally, they create an environment in w h i c h fknding inthe key sectors is largely determinedelsewhere. The result impels even well- intentioned local councilors either to shift funds about rapidly as they try to cover the priority sectors or to second-guess next year's amount of vertical-program allocations. Insome cases, the programs can create incentives for recentralization. Donors have begun to recognize these problems and have started initiating fewer vertical and more devolution-friendly projects. The World-Bank-supported Punjab Education Sector Reforms, the proposed Punjab Municipal Services Improvement Project andthe ADB-supported Southern PunjabBasic UrbanServices Project and the SindhDevolved Social Services Programexplicitly recognize local government ownership (Box 23). These programs provide examples of how far, andhowrapidly, donors are prepared to realign their operations inaccordance with the changed reality presented by devolution. Box 23: Donor-Supported Transitional Vertical ProgradGrants The ADB-financed Southern Punjab Basic Urban Services Project, approved inlate 2003, is designed to encourage good performance by TMAs and to shift from a traditional vertical program toward modalities more consistent with the PLGO 2001. During2004, the first year, funds will be allocated through the Local Government and Rural DevelopmentDepartment under a special PFC grant o n the basis o f TMA- specific investment plans for small-scale water supply schemes. Insubsequent years, funds will be allocated under a formula-based grant. Access to funds inany year will depend o n TMA's compliance with "Basic Eligibility Criteria" that are applicable to all TMAs. The criteria include: Compliance with PLGO 2001 requirements o n Monitoring Committees. Strengthening o f Council secretariat and budget provisions made for continued t e c h c a l capacity. Complaints cells and procedures establishedand linked with intemal audit procedures. Action plans prepared, detailing how Tehsils will transfer functions to Union Administrations in accordance with PLGO 2001. Requirements allocating 25 percent o f development budget for CCB projects observed by TMA, and efforts made to disseminate CCB guidelines and Rules o f Business widely. Water Supply and Sewerage User Charges action planprepared, for detection, disconnection and regularization o f illegal connections, and detailing time-bound strategy for rationalizing water tariffs, improving collection efficiency. Other revenue-enhancement measures, including initiation o fbaseline survey to determine revenue potential, incorporating licensing, encroachment, temporary tehbazari, parlung and transport stands. Inthe following years, the funds available to a TMA will be determinedby formula, and performance. Two sets o fperformance measures will be applied: Investment Project-Spec@ Performance Measures (IPPM). These include simple process measures to be observed for the effective use o f resources allocated to individual investment projects. Examples include 91 operations (staffing, equipping levels); community participation; compliance with physical progress targets; quality o f construction; observance o f PLGO 2001 requirements on maintenance o f accounts, tender and contracting rules; monitoring and evaluation. A minimumscore for IPPMwill be mandatory. The shares o f T M A s failing to meet the minimumperformance standard will be credited to an investment pool andmade available for reallocation to other TMAs based on their perfonnance. Reform Implementation PerformanceMeasures (RIPM). These measures willjudge performance in carrying forward the reforms under the eligibility criteria. They include therefore, scoring systems that measure the functioning o f complaints cells, the implementation o f commitments for transfer o f hnctions to UCs; at least 80 percent o f utilization o f budget reserved for CCB projects, and so on. Supported by the World Bank, the Punjab Education Sector Reform Program focuses on improving access through the rehabilitation o fprimary schools, and improving quality through teacher training and free textbook provision. The centerpiece o f this program i s a Terms o f Partnership Agreement between the provincial government and participating districts. Based on these accords, the province will provide conditional grants to districts inAccount IV outside the PFC framework. These grants will be conditional on a number o f items such as increased education expenditure allocations, meeting o f reform and other education-indicator targets. Governments at all three levels needto find ways to bringvertical programs under local- government planning and budgeting systems and to lay the ground now for a shift to the kinds of formula-based grant systems envisaged inthe Local Government Plan 2000. Since donor-funded projects represent the bulk o f the province-local government vertical programs, reviewing the pipelines o f donor projects i s the logical place to start inidentifyingpragmatic opportunities to make this transition. Inthe various programs to providebudget support to federal andprovince governments, donors should work with governments to ensure attention to the intergovernmentalaspects o f the linkages betweentheir budget support-at federal or province levels-and the local governments. More generally, inproject aid, donors must ensure that project modalities do not weaken but rather support the linkages between elected leaders and citizens and with service providers. Frequently, the project aid modalities sidestep local policymakers through parallel institutional arrangements. As the study has revealed, a large share o f donor project aid i s effectively outside the influence and, sometimes, the ken o f the elected leaders responsible for planning and delivering services intheir jurisdiction. This separation creates great difficulties inplanning and budget processes so that a fair and equitable distribution o f resources i s ensured and exacerbates the problems o f local ownership and long-termrecurrent financing o f the facilities and services created. 92 Table 25: Devolution-Friendlinessof ADB and World Bank Projectsin the CurrentLending Program DimensionsAssessed olicy priorities, site Average Scores (YO)48.5% 16.8% 34.7% Medium Term Impact 8. Will the project or program envisage a graduated increase inlocal 14 8 6 government responsibility inthe implementation o f the project? 9. Will it encourage diversity with provincial devolution models to 3 12 13 accommodate local needs and innovation? 10.Will it envisage the use of conditions grants for major policy 7 5 16 initiatives? Average Scores (YO)128.6% 1 29.8% 141.7% Notes: ADB Proiects: SME SectorDeveloDmentPromam; DecentralizedSocial Services, Sindh; Balochistan Road Sector DeGelopmentProject;PunjabResourceManagement Program, SouthernPunjabBasic Urban Services; FATA RuralDevelopmentProject; Sustainable Livelihoods in BaraniAreas Project(Punjab); BalochistanPublic ResourceManagementProgram; DecentralizedSocial ServicesProgram(Punjab); Restructuringo fTechnicalEducationandVocationalTraining System; RawalpindiEnvironmentalImprovement Project;NWFP Road Sector and Sub-regional Connectivity; Industrial Efficiency andEnvironmental ImprovementProject;DecentralizedSocialServicesProgram(Sindh11). WorldBank Projects: Highway Rehabilitation;PovertyAlleviation Project11; Punjab EducationSAC; Public Sector CapacityBuilding Project;Community GovernmentInfra. (CIP 2); SindhOFWM; PovertyReduction Support Credit; PIFRA; Tax Adm. ReformProject;PunjabOFWM; PunjabMunicipal ServiceImprovement Project; Communitybased MaternalandChild HealthProject;NWFP SAC 11; Sindh SAC 11. Source: Responsesfrom surveys of task teamleadersinthe World Bank and ADB undertakeninNovemberand December2003. As apreliminary step, this studyhas assisted the World Bank andADB to begina review o f upcoming assistance to Pakistan. It i s apparent that whereas some assistance may be assessed as devolution-neutral, that i s supporting upstream institutional reforms, capacity-building or urgent service delivery improvements, the larger share o f assistance would benefit from commitments to be devolution-encouraging. Table 25 summarizes an analysis conducted simultaneously on ADB and World Bank lending for 2004-05. Action recommended: More comprehensive review o f donor project portfolios and pipelinesto ensure that all projects are assisting ina phased migration away from vertical programs. Array of social indicator data available broadened and district data captured The quality o f data available about service-delivery performance, from expenditure data all the way through to measurements o f human-development outcomes, i s weak at the district level. 93 Unless quality improves, it will not be possible for federal government to make reliable judgments about the impact o f macroeconomic or poverty policy, for provinces to establish performance-based grant systems to buy better performance from local governments (Box 24), for local governments to plan, or for the citizens o f Palustanto gauge how well their elected leaders are doing and hold them accountable. Box 24: Data Needs for Equalization or Conditional Transfers "In selecting formulas for equalizationtransfer, a major consideration is the availability o fdata. Inmany developing countries, data constraints force the central government to adopt relatively simple models with few variables. The central government agency incharge equalization transfer can use a variety o f data sources. Inmost countries, the easiest source is the statistics provided by the central government's statistical agency. Inaddition to the central statistical agency, line ministries can oftenprovide more detailed statistics on need indicators such as demographic composition, land areas, student enrollment rates, health indicators, length and quality o f roads, electricity consumption, number o fpolicemen, etc.. .Data on local tax bases are often supplied by local tax authorities. Incases where the central and local governments share the same tax bases and the center i s responsible for tax collection, tax base data can be easily obtained from the central tax authorities. Incases where local tax bases differ from central tax bases, local tax authorities shouldbe required to provide annual tax base figures to the central government agency in charge o f transfer. It is necessary to enact a law or issue a central government ordinance o n fiscal transfer that obliges local authorities to submit accurate data o n a timely manner. Naturally, incentives exist for local governments to under-report.. . inorder to receive more transfers. To prevent such practices, it i s necessary to include penalty clauses on fraud inthe law or ordinance. Equally important i s that the central government audits the statistical reports submitted by local governments (either by directly sending officials to local governments or hiring independent auditors)." Source; (Ma: 1997). A range o f data sources provides input,intermediate, and outcome indicatorsdisaggregated at the district level (see the thirdvolume, Annex 2-Technical Considerations, for a survey o f available data sources). Ingeneral, the administrative data sources, such as the MIS, are not as reliable as household survey data from PIHS. Presently, two survey data sources are representative at the district level: the CIET Social Audit and the MICS, and one inprocess, the CWIQ. Ofthese, the CWIQ and the CIET appear to be the most promising, although itmustbe recognizedthat each provides different information to support different processes. CWIQ may be used to support the definition o f indicators for design o f conditional grants and then to track expenditure and performance against process and output. CIET data, which cannot be used inthe same way, provides a more qualitative gauge-what do citizens think-about particular service-delivery or cross-cutting issues. This additional source o f information has great potential for elected leaders, inparticular, when itcomesto budget-planningdecisions or gauging the performanceo f particular sector service-delivery staff. Provided that these methodologies can be adjusted to ensure local ownership, and that the requisite capacity is created for the surveys to be conducted regularly and properly, then they shouldbe able to provide a range o f indicators for local government planning and budgeting, as well as a basis for a conditional transfer mechanism. The success o f a monitoring system and its impact on policy critically depend on the institutional arrangements and the capacity withinthe monitoring institutions. Progressing from the I-PRSP, the PRSP publishedin December 2003 lays out an institutional setup inmore concrete terms, specifying a general mechanism for tracking and reporting outcomes, with the National PRSP Secretariat as the coordinatinginterface. Clarity i s still lacking, however, on such key specifics as the mechanisms through which the data and reports generatedby the monitoring institutions would provide timely feedbacks into the policy process. On the issue o f capacity, while the 94 document identifies four vital areas where capacity has to be developed at the Secretariat and other levels o f government, no clear plan has yet been advanced to address this need. Specifically, the sustainability o f the monitoring system, o f which the CWIQ is a part, will depend critically on the capacity o f the Federal Bureau o f Statistics (FBS) to conduct complex surveys effectively and on its ability to function as an autonomous organization. The PRSP does not include a plan for enhancing capacity at FBS or for insuringthe autonomy that would enhance the credibility o f its data. Action recommended: 0 Set benchmarks and specifications for the Federal Bureau o f Statistics to report on the data that will be available to for the 7th Award, at least 3 years before the anticipated end o f the 6th Award. Local GovernmentsHeld to Account Districts employ their own staff The challenge to managerial authority inthe districts arises fiom the transfer o f large numbers o f staff from provincial employment groups, particularly Rural Development, Local Government, Health and Education.'' These civil servants have a strong historical connection with the provinces and provide a convenient point o f entry for provincial involvement indistrict affairs. In addition, some see the province as a logical next step intheir career. Box 25: Administrative DevolutioninIndonesiaand Uganda InIndonesia, the 2001 decentralization saw a massive transfer o fcentralgovernment civil servants, government assets, and documentation to regional governments. The transfer o f civil servants proceeded in two phases. Personnel of 5 departments, some 150,000 people, were already reassigned to the regional governments in2000, when the departments they worked for were either abolished or converted into a state department without an apparatus inthe regions. Law 22/1999 gives the regions the right to hire and fire its civil service, but there are conflicting laws and other forms of implementing regulations that appear to circumscribe this autonomy. The local rightto manage the civil service i s limited inseveral ways. Incontrast to Law 22/1999, the civil service law (Law 43/1999) retains much control at the central level. The central government still determines to a large extent civil service wages: base wage, position allowances and family and rice allowances are still set by Presidential Instruction. Inaddition, the Ministry o f Finance sets limits to honoraria to be received by civil servants for project management, and the like. And BKN, the civil service agency, i s developingjob classifications and accompanying qualification standards, which it wants to see applied to local civil service as well as central civil service. Government regulations, which require nationwide advertising for certain posts, and local advertising for others, further limithiringprocedures. Source; (World Bank:2003b). InUganda: "Before decentralization, line Ministrieshaddirect control over staffworking inDistricts. Line Ministries initiated the appointment, confirmation, promotion, training, discipline and the transfer o f such staff. With decentralization, the above staff were transferred to Districts and so were the functions previously exercisedby the Line Ministries. Staff became employees o f the respective District and Urban Councils and movement between Districts and betweenDistricts and Ministries was minimized. s9 TMAs do not face the sameproblemas they inheritedstafffrom the former urbancouncils, and from rural district councilsbut other than some PHEDstaffwhere thesehavebeendevolved, they inherited few provincial staff. 95 withdecentralization,anofficercanleaveoneDistrictorUrbanCouncilforanother,andfromaDistrictor Urban Council to a Ministry, if: (i) There i s an advertisedpost to which he or she applies and i s appointed: (ii) H e or she learns o f a vacant post and applies to (the respective council) to fill it." Source; (Republic ofUganda: 1999). Administrative devolution i s always limited.Few sub-national governments have untrammeled authority to act autonomously as employers (see Box 25 and the thirdvolume, Annex 2- Technical Considerations). However, as Box 26 illustrates, there i s an urgent needto move significantly inthis direction (World Bank: 2001, 2002a). Realistically, establishing the districts as the primary employers o f their staff may not be a feasible single step for Pakistan District Governments. Staff from the federal occupational groups and senior staff from provincial groups would strongly resist such a move, fearing that their career paths hadbeen unreasonably limited,restricting them to a few management positions in a small government. Infact, as local government develops, there i s every reason to assume that a dynamic market in senior local government positions will develop, with pay and career options at least as attractive as those inthe federal and provincial governments. However, inthe short term, a two-part local government service could provide the senior staff inthe district with the reassurance that they have both security and career prospects, while insuringthat all staff inthe districts, senior andjunior, are managed only from within the district. I Box 26: ConflictingLines of Management "The heated debate inthe Peshawar District Council on Tuesday when the elected members resentedthe non-implementation o f the resolutions passed by the council, is not particular to Peshawar. This i s happening all over Pakistan innearly all the District Councils. The Peshawar District Council hadpassed seventy-eight resolutions duringthe year 2003; the concerned departments responded to only fifteen. According to the members o f the council the District Coordinating Officer failed to do his duty by not sending the resolutionto the concerned departments." "The crux o f the whole matter i s the plurality o f authority under which the different cadres o f employees work inthe districts. Most officials on senior posts inthe districts are federal government appointees and their service rules and promotions are regulatedby that government. They, however, work inthe provincial government departments. The provincial governments can influence themby writing adverse reports on themandtransferring them to unattractive locations. These officers when placedindistrict departments are supposed to do the biddingo f the district administration. But since the districts have no influence with them whatsoever, these senior officers do not feel the urgency o f following the directions o f the district nazimeen or members." 1Source: "Governance Issues" electronic newsgroup-22 January 2004. One method for achieving this i s to create a distinct Local Government Career service comprising those staff at grade 16 or below, inwhich the district government assumes responsibility for preparingthe salary budget, recruitment, career management and performancemanagement for junior staff. They would thus be employed unambiguously by the district. Inparallel, provincial staff could apply to join aprovincial Local Government Service. Districts wishing to fill senior vacancies would interview and accept staff from the pool o f eligible provincial Local Government Service members, who would serve inthe district for fixed-term appointments while retaining their substantive appointment with the province.60Table 26 summarizes this option. 6o The Services Departments o f the provincial govemments might need some assistance or capacity-building in order to manage the pool andto ensure that district govemments had access to career historiesof staff in the Provincial Local Govemment Service pool. 96 Table 26: PossibleFuture for the Staff inDistrict Governments Staff groups Appointment Control Establishment Pay policy overthe control salary budget Staff from federal employment groups (primarily District Management Group (DMG) and Provincial Federalor With the the Audits andAccounts Group) Local Provincial Province With the Federal Formerlyprovincial Government Government province Govemment or employment groups- Grades Service province particularly Public 17+ HealthEngineering, Rural Development, Grades To be passed Local Govemment, 1-16 Local Appointed by from the With the and Education Government the District District province to the province Employeeswho had earlier Career Government district within indefinitely belonged to the rural district Service one year councils Ideally control of the establishment, and ultimately authority over pay policy, would be devolved for the Local Government Career Service, but realistically these are longer term options. Box 27 sets out the more immediate steps. Box 27: Key Features of a Local Government Service Key features o f the Provincial Local Government Service could be: 0 The Service i s a voluntary pool that senior provincial staff can apply to join. 0 Staff who had applied and been found suitable for district employment would remain intheir provincial posts pending assignment. Pool staff could be accepted into the district o n clearly specified terms, with explicit perfonnance agreements. 0 Staff members who wished to resign from the district service could, with the permission o f the provincial government, retumto the province. The district could terminate the appointment to the district, with the staff member automatically retumed to the province, with the right o f appeal to the Provincial Public Service Commission in the event o f unfair dismissal Key features o f the Local Government Career Service would be: The District Service to be established separately ineachprovince by provincial legislation. Legislationto ensure that all districts within a province retainthe same pay and grade structure. T o cover grades 1-16 only. Recruitment to the districts to be at departmental level within the district, with the Provincial Service Commissions to oversee the quality o f the recruitment processes, but not to be involved directly inthe recruitment. Transfer between districts would entail resignation from one and appointment by another. Transfer from districts to provincial or federal government employment would entail resignation from the district. The enabling legislation could allow for the possibility o f small provincial quotas o f new appointees to be drawn from the staff o f districts. The legislation could also allow for the Local Government Commissions to establish small secretariats to ensure consistency inpersonnel practices between districts within the province. Steps intransferring responsibility for managing the Local Government Career service to the districts: 97 Fullcontrol over the salarybudget for the LocalGovernment Career Service would bepassedto the district immediately-allowing for local determination o f the size and composition o f the district workforce. PFC transfers to the districts would initially be based on 2003104 staffing figures, recognizing that disparities indistrict employment have arisen historically and are beyond the control o f districts in the short term. PFC transfers to the districts for salaries would move incrementally toward a formula-based transfer within 5 years that allowed for employment levels based on similar per capita ratios, with some allowance for the diseconomies faced by small districts and by remote rural districts. If,within 5 years, any comparatively overstaffeddistrict couldnotreduceits staffing norms through attrition to those allowed inthe NFC awards, then the province must offer the option o f transferring unwanted staff to a provincial surplus pool. Inthe short term, the province would maintainrestrictions onthe total headcount, while allowing districts to determine staffing composition. Within one year, provinces would relinquish control over the district establishments completely, with discipline exercised o n district employment practices through PFC provisions that penalize districts for staffing beyond provincial norms. Ifthis or a similar arrangementisnotinstitutedinthe short term, there is stillanurgentneedto reduce the opportunity for provincial micromanagement o f district staff. Inprinciple, APT authority for the devolved staff inthe provinceshas largely been passed to the districts. However, for staff inSindh, Balochistan (health), and NWFP (health), the province has authority over APT down to grade 16. At the very least, some rationalization o f the APT authorities i s necessary to make sure that the district has full authority over all staff at grades 19 and below. Enforcing this rule inpractice would remain challenging. One possibility for enforcement would be for intra- district and inter-district transfers o f TMA staff at any grade to be subject to veto by the Mushawarat Committee, and for inter-district transfers o f district staff at any grade to be subject to veto by the district council. Action recommended: Provincial governments to establish that the districts have full APT authority over all staff at grades 19 and below. Options for the creation o f a Provincial Local Government Service (comprising provincial employees wishing to work inlocal government), and a Local Government Career Service (an autonomous local government cadre) to be set out inthe White Paper. Districts maintain their own accounts The AGP and CGA are indispute over their respective authorities over staff. Inreality, the cadre histories have resultedinthe CGA having very little managerial or professional authority over the accounting staff. That authority i s hrther reducedby the dispute between the federal and provincial accounts cadres. The consequence i s that district accounts are prepared by a combination o fprovincial accounts staff (who have few incentivesto follow the professional advice o f the CGA) and federal accounts staff (who also have few incentives to follow the advice o f the CGA, but inaddition feel no need to cooperate with their lower-status provincial colleagues). Compoundingthe problem, the new accounting responsibilitieso f the District Accounts Officer coupled with the minimal capacity available has created a series o f technical accounting problems. The consequence i s that district-level accounts are poorly prepared and maintained, largely by staff with no loyalty to the district. 98 Itis necessary to address three key issues to resolve the transitional problems. First,the various technical problems associated with the development o fNew Accounting Model and the transition to the new codes must be resolved. Second, ambiguities inthe staffing responsibilities and reporting arrangements, includingresponsibility for preparation and maintenance o f accounts and oversight and enforcement o f compliance with accounting standards, must be addressed. Finally, there i s a significant problem of capacity. The federal cabinet decision on 6 November 2002, requiring the provincializationo f accounts by 2006, sets a direction toward local responsibility for account preparation. However, responsibility for administration o f district accounts should eventually be fully transferred to a district cadre- as oversight and enforcement o f compliance with accounting standards i s transferred from the CGA to the Provincial Accountants General. Establishingbenchmarks as outlined inTable 27 would allow the federal govemment to create incentives, through conditional grants, for provinces that are complying. Table 27: Steps toward Provincialization of Accounts Issues to be resolved Near rm Mediumteri :tomid-2006) District TMA District TMA Technical issues Development o f Refine, based on Refine,based No action N o action NewAccounting piloting through on piloting needed needed Model PIFRA through PIFRA Managing Modify LG Modify Review and Pilot test, then transition to new Accounts Code to debug roll-out NAM codes NAM systems to all TMAs compliance. following Roll-out NAMto roll-out all districts Staffing Responsibility for Federal- TMA (Tehsil District TMA responsibilities preparation and provincial Accounts and reporting maintenance o f (Federal Pakistan Officer) arrangements accounts Audit and Accounts Service, the Provincial Treasury cadre) Oversight and CGA designs Province CGA to CGA to enforcement o f system o f designs system impose impose compliance with sanctions for o f sanctions for sanctions for sanctions for accounting noncompliance. noncompliance. violations violations (but standards (but handing handing over over to to Provincial Provincial Accountant Accountant General post General post 2006) 2006) Capacity-building Buildcapacity o f Train on Local Retrain and existing a) federal existing system govemment induct accounts and b) cadre qualified province treasury responsible accountants for accounts into service 99 Action recommended: 0 Federal government should be providing incentives for meetingbenchmarks toward the provincialization o f accounts as a first step toward complete district preparation and maintenance o f their own accounts. Further Clarify the SeDaration of Executive and Judicial Powers Devolution has abolished the executive magistracy and its functions vested inthe judiciary, executive (local governments) or police. Promulgation o fthe Police Order 2002 has left many uncertainties inthe respective responsibilities o fthe DPO and nazimeen, includingboth district and tehsil levels, for the broad definition given to law and order. There i s a real risk that failure to clarify these relationships, and ensure they are,adequately dealt with inthe forthcoming Police Rulesmayjeopardize the entire reform effort. As noted, there i s widespreadpublic concernthat police corruption and public safety have deteriorated since promulgationof the Police Order 2002. Evidently, national leaders are acutely aware o fpublic concern.61Box 28 suggests that some local action can be helpful. I Box 28: Improving Union-Level Police, Court and Local Government Relations The Local Government Plan envisagedPublic Safety Committees at union level, and a role for CCBs in police monitoring. By and large, this has not beenpart of the roll-out o f the Police Order 2002. However, Tehreek, a Quetta-based NGO that works inpreventive-health service delivery inUC Gawalmandi has taken this idea forward. Recognizingthe highpriority placed o n public safety andjustice issues, Tehreek's dialogue with the UC nazim and local police-station and community leaders has ledto formation o f an Insaaf Committee and a Musalihat Committee. An office has been established to facilitate police- community liaison, the functions and responsibilities o f the committees, as detailed inthe LGO are displayed outside the office, and the initiative has been greatly appreciatedby citizens. This initiative is currentlv being redicated inother unions inOuetta. The goal i s to conclude the separation o f powers with the result that external oversight and accountability o f the police improve and workable arrangements are put inplace for the enforcement o f local and special laws, including the assignment o fnon-adjudicatory powers and supervision o f pretrial detention to the executive.62 Action recommended: 0 Provincial governments to prepare and notify the rules of business detailing the reporting obligations betweenthe District Police Officer and the District N ~ z i m . ~ ~ 0 Provincial governments to assign non-adjudicatorypowers and supervision ofpre-trial detention to the local government executive, reducingthe burdenon the subordinate courts, and updating fines and fees for violation o f local and special laws, and municipal regulations. 0 Federal Judicial Academy (FJA) to train all Judicial Magistrates intheir newjurisdictions under criminal and local laws especially to protect citizens from abuse o fpower and harassment. Inparticular, where writ o f habeas corpus i s issued, the court should also order action against the officials concerned and report the same to the HighCourt for On 27 January 2004, the President and Prime Ministerjointly presided over a further meetingwith senior cabinet officials andthe Province ChiefMinisters to set arenewed deadline for "full implementationofthe Police Order 2002 in letter and spirit." Dawn 28 January 2004, "August 14Deadlinefor Police Reforms Fixed." This refers to the nazim's power to visit police stations and demand the release of personsillegally detained and not thejudicial powers pertainingto the grant of custody or remand of the accusedto police. 63 The 14 April 2004 meeting (see footnote 16.) reportedly reaffirmed the nazim's responsibility to initiate the DPO's ACR inrespect of law and order. Police rules have not yet beenprepared. 100 follow-up by the Registrar with those actions reported inthe annual report o f the High court. 0 HighCourtspublicly reaffirm their commitments to monitor any adverse actiontaken against subordinatejudges who bar unlawful behavior o f officials, particularly the police. The declaration should include a commitment to exercise its powers o f contempt against the highest officials concerned. 0 HighCourts shouldexpedite the establishment of separate civil andcriminal courts and lighten the administrative and supervisory burden o f the district and sessionjudge by appointing a Senior Judicial Magistrate ineach district to assist with administrative matters relating to the exercise o f criminaljurisdiction by the subordinate courts. 0 EachProvincial govemment to establish an independent prosecution service with offices at district level, and strengthen arrangements for supervision byjudiciary and local government officials o f pre-trial detention and supervision o f investigation. 101 Appendix 1:Backgroundto ADB/DfID/WB DSDDataset 2 This dataset comprises: provincial receipts and expenditures; fiscal transfers to the districts; and district government receipts and expenditures-as known at the end o f FY2003. This dataset has been constructed from a range o f sources: 0 Annual budget statements o f eachprovince for FY 2004 and revisedestimates. 0 Budget notifications o f the Departments o f Finance ineachprovince showing the budgeted share o findividual district government o f eachprovince were used. 0 Monthly actual transfers to each district government from the Department o f Finance o f each Province were used, drawing on notifications issued to the State Bank o f Pakistan for release into account IV. 0 Monthly Detail Books o f account IV from the Provincial Accountant General's office. These are available centrally at the office o f accountant general o f eachprovince. Table 29 provides details o f the data sources. Data drawn from district budget documents alone are unreliable for two principal reasons: 1. Budget estimates are only roughproxies of actuals. Duringthis transitionperiodthe estimates have been preparedby adding a percentage onto the previous years' budgets. These are therefore inexact pictures o f actual expenditures. 2. Presentation o fthe district budgets does not follow a set format and can vary from individual district to district, creating problems for comparability. Aggregate data o f DAOs are also unreliable, as reports o f Provincial Consolidated Fundrevenues include all revenues collected ina district without differentiating between Account Iprovincial levies and dejure district revenues deposited inerror or temporarily inAccount I(see Table 28). Dependingupon the level of aggregation at times they can include federal revenues as well. As such they provide a false picture o f district government revenues and often are not even a close proxy. Obvious errors resulting from this approach are inclusion o f the following revenue items inthe districtrevenues: agriculture incometax, collectionfees onarrears oflandrevenue, the sale o fnon-judicial stamps and papers, motor vehicle tax, entertainment tax, and so forth. 102 Table 28: State of the Provincial Accounting Records Punjab NWFP Balochistan Sin& The accounting records in InAG NWFP, the records Accounting records inAG DAOs in Sindh were not AG Punjab are- were not yet computerized. Balochistai are able to record the computerized and monthly DAOs from each district computerized and monthly transactions o f Account IV Detail Books were accounts office submit Detail Books were separately. As with NWFP, available showing the monthly cash accounts and prepared monthly showing DAOs from each district details o f expenditures and civil accounts o f the the expenditures and accounts office submit receipts o f the Account IV transactions incurred receipts of the districts. monthly cash accounts and and Account Iof each during the period. civil accounts of the district. transactions incurred during the period, but these contain many errors. Expenditures from Account IV and Account I, are often recorded inthe same document. As a result, a reliable database o freceipts and expenditures for province o f Sindh couldnot be constructed. The accounting records used to construct the database are the records o f transactions at the point at which they are incurred. These records are reconciled with each authority responsible for incuning expenditure or recording receipts. As standard accounting formats are used, the accounting records are comparable. The process o fbuildingthe dataset took the following into account: 1. Different provincesare at different stages o ftransitionto preparationo f district accounts. The transition from provincial accounting formats to district level accounts i s still not complete. This means that the actual decentralized functions and revenue items have been changing byjurisdiction, by province, and over time. This requirednumerous revisions in data collection and datasets. New codes for district accounts were only notifiedin January 2004 (with implementation o f NAMthese will change once again). N o district accounts have beenprepared to date. District-level data remains embedded inprovincial accounts. 2. Many o f the revenues which shouldbe bookedunder Account IV are still being recorded inAccount I. expenditures appear inthe aggregate Province ConsolidatedFund District boolungs inthe DAO and later AG's records. To adjust for these difficulties a mapping o f statutory assignments over the codes o f classificationwas carried out. Minor heads in case of revenue items and function codes incase o f expenditures were selected from the Chart o f Classificationand used to map onto corresponding assignment o f functions and revenues in accordance with the four LGOs.These were then each tracked month by month inthe accounting records. 3. Appropriate amendments were made for shared taxes. For instance urban property tax collected ina district i s now a tax to be sharedbetween district government and TMAs in the ratio o f 15:85 with the latter returning to point o f collection. For this purpose the district level data used inthe paper depicts the total collection and then counts only 15 percent o f the tax for district governments. This process ledto some delays as the staff inthe offices o fthe Accountants General were initially reluctant to hand over data. It was only toward the end o f the fiscal year, after repeated attempts, that data for all the districts inPunjab, Balochistan and Sindh were available with the 103 respective AGs. Data were flowing in from districts in fits and starts at times with errors. Data from the AG office o f Punjab and Balochistanwas easy to collect. It was difficult to obtain data from AG NWFPas the records inthis office were not computerized and cash accounts and civil accounts were obtained and photocopies were made from these accounts. Provincial AGs will be finalizing appropriation and finance accounts by the end o f third quarter o f the current year for fiscal year 03. The data will be updated and any errors that are found at this stage will be corrected. Table 29: Data Sources for the Tables in DSD2 Datasource Punjab INWFP IBalochistan Provincial Receipt and Annual budget statements o f each province for FY 2004. Revised estimates have Expenditures been used Fiscal Transfer Data Tables For the budgeted share o f each district govemment, the budget notifications o f the Departments o f Finance ineach province showing the budgeted share o f individual District Govemment o f each province were used. For actual transfers, the statements depicting monthly actual transfer to each district govemment from the Department o f Finance o f each Province were used, drawing- on notifications issuedto the : ite Bank o f Pakistan for re ise into Account IV. District Govemment receipts Monthly Detail Books o f All receipts are being All receipts are being Account IV were used deposited inAccount I deposited inAccount I from the Accountant therefore, monthly cash therefore, the monthly General's office. Where accounts o f Accountant Detail Books o f Accountant neededthe required General's office were General's office were used codes have been usedto construct the to construct the receipts disaggregatedfrom receipts tables. The tables. The required codes Account Irecords required codes have been have been disaggregated disaggregated from from Account Irecords Account Irecords District Government Monthly Detail Books o f Monthly civil accounts The expenditure o f each expenditures Account IV from the o f Accountant General's district govemment has Accountant General ofice were used to been collected from office were usedto construct the expenditure monthly Detail Books construct the expenditure tables collected from the office o f table. the Accountant General. The photocopiedrecords were used to construct district by district expenditures 104 Appendix 2: AssessmentFrameworksDevelopedfor the Study Despite the voluminous researchand practitioner literature, enormous conceptual and practical difficulties remain inassessing the pace and depth o f devolution. This study has developed three key frameworks that could, with adaptation, be used for extending an assessment o f devolution to other settings. First,the study uses a schema for focusing on a key set o fincentives at local government level, for which there i s ample justification inthe theoretical and research literature, as key to the success o f devolution initiatives, and which can provide a test o f the degree to which devolution i s beginning to "bite" by creating appropriate incentives for efficient and responsive services within local government. These are summarized inTable 30, butwere set out inmore detail in Table 10 above. The second volume (Annex 1-Recent History) illustrates how these tests were applied inpractice. Table 30: AssessingInstitutionalChangeswithinLocalGovernment Incentives for Provincial efficiency Budgetary certainty Nazimeen and senior incentives Enhanced autonomy inpreparing the budget staff Incentives for local revenue-raising Citizen power or Councilors motivated to listen to citizens voice The pull from councilors i s stronger than that from the Incentives for service I Managerial power I Provinces Staff working inthe district are employedby the district providers I Nazimeen authority over senior staff External oversight Adequate deployment o f staff Client power Citizen dispute resolution I I Communitv management o f facilities I Second, within this overall schema, the study also has employed a distinctive approach to measuringthe degree to which staff are formally employedby local government4ffering some precision to the nebulous assessment o f administrative devolution. This i s set out indetail inthe thirdvolume (Annex 2-Technical Considerations), butis summarized inTable 31. 105 Table 31: Assessingthe Locationof the Employer Function Dimensionsof "Employership" I s the local government responsiblefor: Budget Control Paying staff from its own budget Determine the wage envelope Authority to dismiss surplus staff Establishment Control Control overall staffing numbers Control staffing numbers in local offices and facilities Recruitment Formal employer Authority to hire Independent merit-based recruitment mechanism(for example, PSC) Career Management Promotion Transfers within local government Horizontalmobility Performance Management Direct and supervise activities and tasks Conduct evaluations Financial rewards or promotion Ability to discipline or fire Pay Policy Set overall wage rates Set local hardship or remoteness allowances Third, the study sets out a new approachfor assessing the degree to which conditional grants represent deconcentrated expenditures, with damaging consequences for local ownership and accountability-and corresponding progress toward institutingdecentralized expenditures inthe form ofunconditionalblock grants. This i s set out indetail inTable 19 above, and summarized in Table 32. Table 32: Assessing "Verticalness"of Conditional Grants Deconcentrated Delegated Decentralizedexpenditures expenditures expenditures Conditional grants I Unconditional I Verticalprograms blockgrants Shown inthe Higher level Sub-national Sub-national government ~ budget of: govemment i government : (butmight j require pre- i authorization i byhigher level .........................................................................!i ovemment) ............................................................................................................................................................... , Inputs .....g.................................................. Conditionedon: Not conditional- Processes, outputs Not conditional ~ based on project ..agreement:............................... i........................................................ ..:or I (earmarking), or outcomes processes Expenditure Parallel project Sub-national Rely on existing provincial or local control and facility managed govemment- govemment accounts control accountability by higher level but with arrangements: government special arrangements (local imprest with either ex ante or ex post controls) Awarded through: Individual ministries or deDartments NFCPFC 106 Glossary Anjuman Society or organization Challan Indictment Gully koocha Street and neighborhood Grampanchayat Local council (India) Insaaf Justice Insaaf Committee Local committee whose function i s to help citizens pursue complaints about the functioning o f subordinate courts Jirga Traditional or tribal council Kotwal Police officer inMoghul times Kutchey Localmagistrate courts Majlis-i-Shoora Consultative council Madrassah Islamic religious school Markaz Headquarters or center Mohtasib Ombudsman Muk mu-ka) An arrangement; reference to corrupt deals madebypolice Musalihat Reconciliation Musalihat Anjuman (or Society, organization, or committee for reconciliation Anjuman-i-Musalihat) Naib Nazim Deputy mayor Nazim Mayor Nazimeen Mayors Panchayat Council Qazi Judge under Islamic law Shariat Islamic law Sifarish Recommendation o f a patron Taluka Town (same meaning as tehsil, but used solely in Sindh) Tehbazari Use o fpedestrian walkways for the selling o f merchandise Tehsil Town Thana Police station Zakat Alms Zila District Ziladar District officer (historical term) 107 References Ahmad, Nuzhat, and SyedAshraf Wasti. 2003. "Pakistan" (http://www.adb.org/DocumentsiBooks/Intergovernmental_Fiscal-Transfers/). InY. Kimand P. Smoke (eds.) IntergovernmentalFiscal Transfers inAsia: Current Practice and Challenges for the Future.Manila: AsianDevelopmentBank. Andrews, Matthew and Larry Schroeder. 2003. "Sectoral DecentralisationandIntergovernmental Arrangements inAfrica." Public Administration and Development. 23. 2940. Annan, J. 1999. "Ghana Health Sector-Wide Programme: A Case Study" (http://www.capacity.orgn?ieb~CapacityrWeb/UKContentlConten~Content.nsE/vwSearchIntemet/984 1AA088 E269B5DC1256D90003971DC?OpenDocument&Highlight=0,swaps).JSA Consultants (prepared for D A C I/CD Network andthe Policy Branch o f the Canadian International Development Agency (CIDA). Accra. Asian Development Bank. 1999. "Pakistan Legal and Judicial ReformProject: IntegratedReport." ADB. Manila. Azfar, Omar. 2002. "Conditions for Effective Decentralized Govemance: A Synthesis o f Research Findings" (http://www.inform.umd.edulIRIS/IRIS/PEG/synthesisqaper.pdf).Paper presentedto the Regional University Consortium Conference: Regional Economic Development ina Decentralizing Indonesia. Center for Institutional Reform and the Informal Sector. College Park MD. Bahl, Roy. 1999. "Implementation Rules for Fiscal Decentralization (Workmg Paper 99-1)" (http://isp- aysps.gsu.edulpapers/ispwp9901.pdf).Georgia State University.Atlanta, Georgia. Bahl, Roy. 2001. "Equitable Vertical Sharing andDecentralizing Government Finance inSouth Africa." Georgia State University. Atlanta. Bellamy, Christine and John A. Taylor. 1998. Governing inthe InformationAge. Buckingham: Open University Press. Bird,Richard. 2001. "Intergovernmental Fiscal Relations inLatinAmerica: Policy DesignandPolicy Outcomes" (http://www.iadb.orgisds/doc/Intergovernmentaltotal.pdf).Inter-American Development Bank. Washington DC. Bird,Richard M.2003. "Local andRegionalRevenues: Realities andProspects" (http://www.worldbank.orgiwbiipublicfinance/documents/fiscalfederalism-Russia~ird_revenues.pdf). World Bank. Washington DC. Bird,Richard M.andEnidSlack. 2002. "Land andProperty Taxation :A Review (Land Workshop Material, Budapest, Hungary, April 3-6, 2002)'' (http://www 1.worldbank.org/publicsector/DecentralizationSub- nationalEconomics/June2003Seminar/LandPropertyTaxation.pdf).World Bank. Washington DC. Blair, Harry. 2000. "Participation and Accountability at the Periphery: Democratic Local Govemance in Six Countries." World Development. 28 (1). 21-29. Brown, Adrienne, Mick Foster, Andy Nortonand Felix Naschold. 2001. "Working Paper 142: The Status o f Sector Wide Approaches" (http://~.odi.org.uMpppg/publications/workingqapers/142.pdf). Centre for Aid and Public Expenditure, Overseas Development Institute. London. CIET. 2003. "Community Social Audit Report." National Reconstruction Bureau. Islamabad. Craig, D. and D.Porter. 2003. "Poverty Reduction Strategy Papers: A New Convergence." World Development. 31(1). 53-69. Department o f Education, Government o fPunjab and Multi-Donor Support Unit. 1998. "Social Action Program: Survey of School Management Committees." Government o f Punjab and the SAP Multi- Donor Support Unit.Lahore. Development Assistance Committee. 2003. Harmonising Donor Practices for Effective Aid Delivery (http://www.oecd.org/dataoecd/0148/20896122.pdf).Paris: OECD. Easterly, William. 2003. "The Cartel o f Good Intentions: The Problemo f Bureaucracy inForeign Aid." Policy Reform. 1-28. Ebel, Robert D. and Serdar Yilmaz. November, 2002. "On the Measurement and Impact o f Fiscal Decentralization: Part 2." Working Paper No. 2809. Evans, Anne with NickManning. March 11, 2003. "Decentralization: A Review o f Staffing Practices in Eight Countries-Draft Paper Prepared for the World Bank." Washington DC. 108 Faguet, Jean-Paul. 2001. Does Decentralization Increase Responsiveness to Local Needs? Evidence from Bolivia (World Bank Policy Research Working Paper 2516). Washington DC: World Bank. Federal Bureau o f Statistics. 2000. "1999-2000: Census o fPrivate Educational Institutions inPakistan." Statistics Division, Government o f Pakistan. Islamabad. Foster, Mick and Sadie Mackintosh-Walker. 2001. "Working Paper 157: Sector Wide Programmes and Poverty Reduction" (http://www.odi.org.uWpppgipublications/workingqapers/157.pdf). Centre for Aid andPublic Expenditure, OverseasDevelopment Institute. London. Government o fPakistan. 2003. "Accelerating Economic Growth and Reducing Poverty: The Road Ahead (Poverty Reduction Strategy Paper)" (http://www.finance.gov.pWpoverty/prsp~03 .pdf).Government of Pakistan. Islamabad. Government o fUganda. 2003. "Fiscal Decentralization: The Way Forward." Ministry o fFinance. Kampala. Haggard, Stephen and Steven B.Webb. 2001. "Political Incentives and Intergovernmental Fiscal Relations: Argentina, Brazil and Mexico Compared." World Bank. Washington DC. Hofman, Bert, Kadjadmiko and Kai Kaiser. 2002. "Evaluating Indonesia's Fiscal Equalization." World Bank. Jakarta. Human Rights Commission o f Pakistan. 2004. "Police Torture and IllegalDetention." HRCP. http://www.hrcp-web.orgipo1icetorture.cfm.Access Date: Feb 1, 2004. Last Update: dk. IndusResource Center. 2003. "Institutional Strengthening of Smcs." Karachi, Pakistan. Kennedy,Charles H. 1987. BureaucracyinPalustan. Karachi: Oxford University Press. Ma, Jun. 1997. "Intergovernmental FiscalTransfer: A Comparison o fNine Countries" (http://www.worldbank.org/htmvdec/Publications/WorEcpapers/WPS 1800seriesiwps1822iwps1822.pdf ). Economic Development Institute o fthe World Bank. Malik, Sohail J. 2003. "Study on Tax Potential inNWFP: An InterimReport." Islamabad. Martinez-Vazquez, Jorge and Jameson Boex. 2001. "Russia's Transition to a New Federalism." World Bank Institute. Washington DC. Mirza, Naeem. 2002. "Women's Participation inLocal Government Elections." Aurat Publication and Information Service Foundation. National ReconstructionBureau. 2000. "Local Government Plan." Government of Pakistan. Islamabad. Operations Evaluation Department. 2002. "Annual Review o f Evaluation Activities in2001" (http://www.adb.orgiDocuments/perslannualrev2OO.pdf).Asian Development Bank. Manila. 1 OxfordPolicy Management. 2002. "Extemal Evaluation o fthe NationalProgram for Family Planning and Primary HealthCare." OPM. Oxford, UK. PattanDevelopment Organization. 2003. "Local Government Elections 2001-Phase 111,IV andV." Pattan Development Organization. Islamabad. Peters, Thomas J. and Robert H.Waterman. 1982. InSearch o f Excellence: Lessons from America's Best- RunCompanies. New York: Harper andRow. Qadir, Ghulam. 2003. "Socio-Economic Variances inthe Devolution and Service Delivery Study Districts-Working Paper." World BankiDfID. Islamabad. Republic o f Uganda. 1999. "Human Resources Management Manual for Local Governments inUganda." Kampala. Shah, Anwar. 1994. The Reform o f Intergovernmental Fiscal Relations inDeveloping and Emerging Market Economies. Washmgton DC: World Bank. Shah, Anwar. 1998. Balance, Accountability, and Responsiveness: Lessons About Decentralization. Working Paper No. 2021. Washington DC: World Bank. Shah, Anwar. 2002. "Fiscal DecentralizationinDeveloping and Transition Economies." InR. Blindenbacher and A. Koller (eds.) Federalism ina Changing World-Learning from Each Other. Montreal and Kingston, Canada: McGill-Queen's University Press. Shah, Anwar and others. February 19, 1996. "Fiscal Federalism inPakistan: Challenges and Opportunities." Washington DC. Shanka, Raja andAnwar Shah. 2003. "Bridging the Economic Divide within Nations: A Scorecard on the Performance o f Regional Development Policies inReducing Regional Income Disparities." World Development. 31 (8). 142141. Sidat Hyder MorshedAssociates (Pvt) Ltd. 2003. "Punjab Local Government Finances-Financial Profiles o f Selected Tehsil Municipal Administrations (TMAs)." Karachi. 109 Smoke, Paul. 2001. FiscalDecentralizationinDevelopingCountries: A Reviewof Current Concepts and Practice.New York: UnitedNationalResearchInstitute for SocialDevelopment. Smoke, Pauland Larry Schroeder. 2003. "Intergovernmental FiscalTransfers: Concepts, International Practice andPolicy Issues." InY.-H. KimandP. Smoke (eds.) IntergovemmentalFiscalTransfers in Asia: Current Practiceand Challengesfor the Future. Manila: AsianDevelopmentBank. Teskey, GrahamandRichardHooper.1989."Uganda EducationStrategic InvestmentProgramme: Case Study" (http:llwww.capacity.orglWeb~CapacityiWebiLTK_Content/Content.ns~vwSearc~ntemetl6S53B9298 DB23DBCC1256DF80059E96B?OpenDocument&Highlight=O,swaps).Department for International Development.Nairobi. Webb, StevenB. and ChristianT. Gonzales.2003. "Bargaining for aNew FiscalPactinMexico." World Bank.Washmgton DC. Witkin, H.A. andJ. W. Berry. 1975. "Psychological Differentiation inCross-CulturalPerspective." Journal o f Cross-CulturalPsychology. 1.4-87. World Bank. 2001. "Nigeria FederalPublic ExpenditureReview." World Bank.WashngtonDC. World Bank. 2002. "Brazil: Issues inFiscalFederalism(Report No. 22525-Br)." World Bank. Washington DC. World Bank. 1997.WorldDevelopmentReport 1997-the State ina ChangingWorld. New York: Oxford University Pressfor the World Bank. WorldBank. 1999. "Agricultural TaxationinPakistan." World Bank.WashingtonDC. World Bank. 2000. "Pakistan-Reforming ProvincialFinancesinthe Context of Devolution: An Eight PointAgenda." World Bank.WashingtonDC. WorldBank.2001. "Reforming Punjab's Public FinancesandInstitutions." World Bank. WorldBank. 2002a. "Pakistan DevelopmentPolicy Review: A New Dawn?(Report No. 23916-PAK)." World Bank. Washington DC. WorldBank. 2002. "Primary EducationManagement:ProvincialCase Studies (Punjab)." Unpublished World Bankpaper. WashingtonDC. World Bank.2003a. "Pakistan Public Expenditure Management: Strategic Issues andReformAgenda." World Bank.Washington DC. WorldBank.2003b. "The World BankRegionalPublic ExpenditureReviewfor Indonesia: Decentralizing Indonesia" (http:lllnweb18.wor1dbank.orgleapleap.nsf/0i7a59aff0cdda926047256dc800138fa7?~enDocument). WorldBank. Jakarta. World Bank.2003c. World DevelopmentReport2004: Makmg Services Work for Poor People. WashmgtonDC. 110