Report No. 26654-IN India Fiscal Decentralization to Rural Governments January 7, 2004 Rural Development Unit South Asia Region Document of the World Bank 1 CURRENCYEOUIVALENTS Indian Rupee (Rs) US$l.OO = Rs. 47.37 (May, 2003) FISCAL YEAR (FY) April 1-March 31 ACRONYMS ACR Annual ConfidentialReport MP Member of Parliament;Madhya Pradesh AG AccountantGeneral NGO Non-governmental Organization AGR Annual Growth Rate NLGORR Non-loan Gross Own Revenue Receipts AP AndhraPradesh NTR Non-taxrevenue BP Block Panchayat OBB Operation Black Board CEO Chief Executive Officer OBS Other Backward Caste CFC Central FinanceCommission OFEs OwnFund Expenditures css CentrallySponsoredSchemes OSR Own SourceRevenues DDP Desert Development Program PAA PerformanceAudit Authority DPC DistrictPlanningCommittee PESA Provisionsfor Extensionto ScheduledAreas DPAP Drought Prone Area Program PRI Panchayat Raj Institution DRDA DistrictRural Development Agency PS Panchayat Samitis EAS Employment Assurance Scheme RIDF Rural InfrastructureDevelopmentFund ECB EffectiveClosingBalance SAS StateAdministrativeService EFC EleventhFinanceCommission sc ScheduledCaste EGS Education Guarantee Scheme SEB StateElectricityBoards EO ExecutiveOfficer SFC StateFinance Commissions GDDP GrossDomesticDistrictProduct SGSY Swarnjanyanti Gram Swarozgar Yojana Go1 Governmentof India SPAF StatePerformanceAudit officer GoMP Governmentof Madhya Pradesh SSFC Second State Finance Commission GP GramPanchayat ST ScheduledTribes IAS Indian Administrative Service TAD Tribal Area Development IAY IndiraAwas Yojana TP TalukPanchayat IRDP IntegratedRural Development Program UP Uttar Pradesh JGSY Jawahar GramSamridhi Yojana VEC Village EducationCommittee JRY Jawahar Rozgar Yojana VPF VillagePanchayat Fund MI Minor Irrigation VRM Village Road Maintenance MLA Member of LegislativeAssembly VTC VehiclesTax Compensation MLC Member of LegislativeCouncil ZP Zilla Panchayat Vice President: Praful Pate1 CountryDirector: Michael Carter SectorDirector: ConstanceA. Bernard SectorManager : AdolfoBrizzi Task Leader: Geeta Sethi I I .. 11 Fiscal Decentralizationto RuralGovernments VOLUME I-Main Report' Table of Contents I INTRODUCTION 1 I1 OBJECTIVE OF THE STUDY .. ............................................................................................................. ........................................................................................ 2 I11 SUMMARY 2 IV ISSUES INTHE CURRENT SYSTEM .. ....................................................................................................................... .......................................................................... 5 V. RECOMMENDATIONS ................................................................................................ 15 A. ClarifL Expenditure Assignments and Local Government Hierarchy......................... 15 B. Strengthen Local Government Budgetary Autonomy .................................................. 17 C. Augment Panchayat Resources................................................................................... 17 D. Redesign Transfers ..................................................................................................... 20 E. Improve Accounts andAccountability......................................................................... 24 F. Createan Information System on Local Finances....................................................... 25 V I CONCLUSION . ................................................................................................................ 25 REFERENCES ....................................................................................................................... 27 Listof Tables Table 1:Population of Lowest Government Tier inSelected Countries ................................................ 16 List ofFigures Figure 1:Control over Expenditures by Grampanchayats for FunctionsDevolved to themis Figure 2: Local Revenuesas a Percentage of Total Revenues for Selected Countries (1999) ................9 increasinginKerala (Case Study on Tholicode GramPanchayat) ......................................... 7 Figure3: GrowthinPer Capita Own Source Revenues and StatePlan Grants (Transfers) inKerala ...11 Listof Text Boxes Box 1: PracticalPrinciples from InternationalExperience for FiscalDecentralization.......................... 4 Box 2: LocalProperty Taxation ............................................................................................................ Box 3: Allocating Transfers By Formula.............................................................................................. 19 21 Box 4: Systems for Vertical Sharing OfIntergovernmentalTransfers ................................................. 22 Box 5: Problems inRedesigningthe Transfer System: the Kerala Example ........................................ 23 'FiscalDecentralization to Rural Governments is a four volume study; Volume I-Main Report (Volume in hand); Volume 11-Technical Analysis; Volume 111-Karnataka Case Study; Volume IV-Kerala Volumes 11.111. and IV are available uponrequest. Case Study. ... 111 Acknowledgments This report was prepared by a team led by Geeta Sethi (SASRD) and comprising Roy Bahl (Dean, Andrew Young School of Policy Studies), Luis Constantino (SASRD), and Fitz Ford (TUDUR). The peer reviewers were Junaid Ahmad (WSP) and RichardBird (University of Toronto). We especially wish to thank them for their contributions and comments, Mr.Vijay Anand, (Secretary Planning, government of Kerala), Mr. Ranjit Banerjee (Joint Secretary Expenditures, Ministry of Finance, Government o f India), Ms. Sudha Pillai (ex-Joint Secretary Panchayats, Government of India), and Mr. Raghunandan (Secretary Rural Development and Panchayats, government of Karnataka). The team i s also extremely grateful to Jennie Litvack (MNSED), Sanjay Pradhan and Dana Weist (PRMPS) for their support. We would also like to acknowledge Bayar Tumennasan and Pablo Saavedra (University o f Atlanta) for helpingwith the analysis and research; P.K. Subramanian (SARFM), Madhavan Balachandran (SASRD), and Rama Krishanan (Consultant) for helping us understand the financial management and accounting aspects inbothKarnatakaand Kerala, and Anand Mathew (Consultant) for helping us tie all the loose ends on information onpanchayats inIndia. The KarnatakaTeam was ledby Dr.GovindaRao (Ex-Director, ISEC) and comprisedMr. Amar Nath, andMs.B.P. Vani. The Kerala Team was ledby Professor. M.A.Oommen (Professor, University o fTrivandrum) and comprised Mr.Joy and Ms.Nirmala. Productionand organizational assistancewas providedby Grace Domingo and Yoshiko Masuyama. We also acknowledge the support of Sarita Rana, DeborahRicks and Lilac Thomas for production. 1 Fiscal Decentralization to Rural Governments I. INTRODUCTION 1. The Government of India after several decades of deliberation introduced two Constitutional Amendments - the 73`d (rural decentralization) and 74'h Amendments (urban decentralization) to - give effect to its objectives of decentralization. The 73rd Constitutional Amendment established mandatory provisions for decentralization to local governments in rural India. To understand the new institutional landscape and its impact on development, the Bank, in agreement with the Government o f India and certain state governments, prepared a phased approach for providing policy assistance on rural decentralization in June 2000. It would (i) start with a broad overview of rural decentralization in India which has now been completed; (ii)investigate the status of fiscal decentralization to rural local governments which i s the focus of this study; and (iii) finally, analyze the institutional framework for rural service delivery in priority sectors. These would be complemented by parallel studies and knowledge dissemination inspecific areas. At the same time the Bank would begin exploring possibilities for lending, takmg into account the reality, opportunities and needs o f local governments in rural India. Hence, the work program was designed inthree phases: 0 Phase 1:The Overview ofRuralDecentralizationinIndia (FYOO-FYO1, completed). The objective was to increasethe Bank's understanding of what i s happening. The output was an "Overview of Rural Decentralization inIndia" givinga broadperspective o f the key issues. Because this was an internal report, the study and specific results received only limited dissemination. 0 Phase2: Cross-sectoral issues (FYO1-FY04, ongoing). The objective i s to selectively assist states inimproving the design and monitoring of decentralization as well as developing a decentralization strategy. The approach is to highlightthe mismatch between policy objectives and current practice andmake recommendations on the way forward. The outputs would be: (i) Fiscal DecentralizationStudy (study at hand), (ii) Impacts o f Decentralization Study (ongoing), (iii) ofLocalOrganizationsStudy(ongoing), (iv)MunicipalFinancesandAccountability Roles (ongoing), and (iv) State DecentralizationDiscussions (KarnatakaDecentralization Forum [May 20011, Kerala Decentralization Forum [September 20011,Workshop onFiscal Decentralization to Local Governments [December 20021; Workshop onRoles of Associations of Local Governments [June 20031). 0 Phase 3: Rural Service Delivery (FY04-FY05). The main objective would be to evaluate service delivery from a sectoral perspective, inparticular for those sectors characterized by externalities and meritorious goods (for example, water and sanitation, health, education, natural resources,judiciary) andpropose improvements that are consistent with the decentralization policies. 2 11. OBJECTIVE OFTHE STUDY 2. The objective of this study is to review panchayat finances, identify issues and make recommendations to improve the fiscal decentralization system. UsingKamataka and Kerala as case studies, this report describes the context for rural decentralization in India, summarizes the current fiscal situation of district, block and village panchayats, identifies the main problems arising from the current practice, and outlines a set of reform options that might better satisfy the objectives of the central and state governments. Kamataka and Kerala were examined because these two states are generally thought to be more advanced than others in implementing rural fiscal decentralization and are therefore relevant for drawing lessons for the rest o f India. 3. The study is presented in four volumes. The volume at hand (Volume I) i s the Main Report. Volume I1summarizes and compares the status of fiscal decentralization inKarnataka and Kerala and i s intendedfor policy makers bothat the central level and inother states. Volumes I11and IV are the more detailed case studies o f Karnataka and Kerala respectively, and are more relevant to policy makers in these states. This main report identifies problems associated with the current system of devolution and recommends ways to improve it to meet the objectives of the decentralization program. The next section encapsulatesthe key messageso f the study. 111. SUMMARY 4. Inter-governmental relations are mostly hierarchical. There is a hierarchical arrangement of center-state-local governance. The 731d Amendment to the Constitution made India one of the most politically decentralized countries in the developing world. By the Constitution and in practice, local government has an urban and a rural stream. Rural local governments are divided into three levels: district (zilla), block (taluk), and village (gram) panchayats. The design and implementation o f the decentralization program are a state government responsibility. However implementation of key aspects of the program i s lagging. Districts and blocks have no taxing powers and little expenditure autonomy. Both inKarnataka and Kerala, they more or less function as spending agents of higher-level governments. Grampanchayats are closest to the people, have some independent taxing power, and some discretion in expenditure decisions. Hence they are the primary subjects o f this study. 5. Karnataka, with a population o f 52 million people, has 5,870 rural local governments with 27 zilla parishads (districts), 175 block panchayats, and 5,659 Gram panchayats. The average gram panchayat inKarnataka has a population of 5,000. The total number of elected politicians inrural governments is 84,886. Of these, 44 percent are women, more than the 33 percent mandated by the Constitution in the reservation system. Since the 73rdAmendment, there have been two local elections and two State Finance Commissions (SFCs). Panchayats in Kamataka have been nominally assigned all the 29 expenditure subjects prescribed in the Constitutional Amendment, including rural drinking water, primary and secondary education, primary health, as well as assigned funds and functionaries (state employees at the panchuyut level). About 30 percent of state public expenditures are channeled throughpanchayats. 6. Kerala has 31 millionpeople representedin 1,157 local governments, o f which 53 are municipalities, 14 zilla panchayats (ZP), 152 block and 991 grampanchayats. The average gram panchayat in Kerala has a population of 25,000. The total number of elected ruralpoliticians i s 12,117. Kerala comes closer to having effectively devolved funds, functions and functionaries. Of the 29 expenditure subjects assignedto panchayats, Kerala has devolved functionaries and funds for 15 o f these functions. About 18 percent of state expenditures (and 30 percent of plan expenditures) are channeled throughpanchayats. 7. The stories of Karnataka and Kerala illustrate the variety of challenges that India faces in implementing decentralization and achieving its benefits in terms of improved delivery of services to rural communities. The two states came to their current stage o f the decentralization process from 3 different directions. Karnataka has had a long history o f initiatives in support of decentralization, beginninginthe 19' century, while Kerala came belatedly to its support. What distinguishes Kerala from the other states in India is the "big-bang" approach it has taken. The strategy was to implement a significant fiscal decentralization programand to "learn by doing". As a result Kerala has now developed a decentralized fiscal structure in terms of grant design and the empowerment of the gram panchayats. Karnataka's i s the more cautious approach where the state government still holds the balance o f fiscal decision-making power. While there are important differences betweenthe two states in some aspects of the systemic weaknesses observed, there are also similar flaws. Neither approach has achieved effective empowerment o f local governments and neither appears to have substantially ratcheted up local government revenue raising, although the Kerala model i s more developed. Both states have weak, outdated and poorly functioning financial management systems that debilitate the policy making and planning process, as well as the management and accountability o f the decentralized system. In the absenceofreliable information on the revenues and expenditures o f local bodies, neither the states nor the center can lead a reasonable fiscal decentralization program. This also raises questions regarding the validity of central and state finance commissionrecommendations. 8. Perhaps the main weakness in Karnataka's experience has been its limited devolution of meaningful authority and capacity to local decision makers, while Kerala's disappointments are traceable to the strain on state finances in the years after the implementationof its decentralization program. The consequences of these weaknesses have been similar despite their different origins. Local govemments have had relatively few resources and relatively little discretion with which to make a significant impact on service delivery. Perhapsas a result, local communities have not engagedfully with the processes of planning and accountability to the extent hoped for by policy makers. 9. An important lesson learned in these case studies is that one necessary condition for a well- functioning system of fiscal decentralization is a healthy state financial position. Because the Constitutional Amendment defined decentralization to be a state subject, until state governments can improve their deficit position, local governments can expect continued underfundingof their present grant entitlements, resistance to new program development, and hesitationto assign more own source revenues to local governments. Before India can fix the fiscal position o f its local governments (rural and urban), it must fix the fiscal problems of its state governments2. This study is limited to a review of the fiscal relations between the state and local governments and thus does not comment on solutions to state finance problems. 10. International experience offers some practical lessons for the design and implementation of a system of inter-governmental fiscal relations. As decentralization is often pursued with different objectives in mind and is a highly political process that needs to reflect the particular circumstances of a country, there i s no single `best' system. However, there i s a set of 11 general principles that many analysts believe are necessaryfor a fiscal decentralization system to work. These are summarized inBox 1below. For example in Karnataka, as a result of fiscal compression, expenditure decentralization has been declining during the last decade. See Volume I11for more details. 4 Box 1: PracticalPrinciplesfrom InternationalExperiencefor FiscalDecentralization (1) All o fthe key components o f a decentralized fiscal systemneedto be addressed: significant expenditure responsibility and discretion, independent revenue raisingpowers, elected local council, and local officers and employees who are accountable to the elected local council. (2) Finance follows function-first should come the assignment o f expenditure responsibility to local governments and then the assignment o f revenue responsibility should be determined, (3) There must be a strong state ability to monitor and evaluate the intergovernmental fiscal system. (4) One intergovernmental system does not fit the urban and the rural sector--these governments have very different capabilities to deliver and finance services, and certainly different capabilities to borrow. These differences need to be recognized, that is,. where different local governments are given different financing powers and expenditure responsibilities. (5) The assignment o f revenue powers to local governments should linkrevenue and expenditure decisions at the margin. (6) Higher-level governments must follow the rules o f the intergovernmentalfiscal system that they themselves create. But the higher-level government does not always keep the rules that it makes. For example the imposition o f unfunded expenditure mandates o n local governments, the under funding o f transfer programs and so on. (7) The intergovernmental fiscal system shouldbe kept simple. (8) Intergovernmental transfers should be designed to match a set o f clearly specified objectives. (9) The intergovernmental fiscal system should be builto n policies that consider impacts o n all levels o f government - central, state and local. (10)Ahardbudget constraint should be imposed on local governments. (11)The intergovernmentalfiscal system shouldbe flexible to accommodate changes. Sources: Bahl, 2001. Details inVolume 11, Chapter IV. 11. The first phase of the program of reforms recommended by the study for Karnataka and Keralawould require substantialefforts and commitment. Step one i s the development by the states of a blueprint for their system of fiscal decentralization. Step two i s to recognize that correcting the current problems and moving decentralization closer to the Government of India's (and the states') objectives requires addressing a number o f detailed issues. In Karnataka, it i s necessary to revisit the detailed assignment of functions, the fiscal transfer system and the administrative arrangement for supporting decentralization, and capacity buildingat the state and local level. InKerala, it i s necessaryto fix the financial management system, to addressthe issue of the administrative arrangements and also to buildthe capacity to support decentralization. For boththis phasealso envisagesapackage ofmeasuresto revise and update the financial management systems. Both states have problems with the productivity of their local revenue systems, and particularly with the property tax. Finally, although Kerala has put into place appropriate downward accountability measures, many are not working properly because o f design, problems o f execution, or because the potential impact of local government program expenditures i s inadequateto provide incentives to engage. Insummary: First, expenditure assignments need to be clarified between the differenttiers o f the government. The expenditure assignments should be set out by the state, spelled out insome detail at the level o f sub-functions and should largely replace the `concurrent' approach that i s now inplace. 0 Second, decentralization should result in local governments having autonomy and sufficient 5 resources to provide meaningful services to their communities. Earmarked transfers (schemes) designed by higher levels o f government cannot dominate local government finances in a successful fiscal decentralization program. Under the present arrangement,panchayats make no contribution to the design o f the schemes and are given little discretion in implementation. They also have limitedautonomy over their staff. It i s important to begin developing these capacities at the local government level to effectively deliver localpublic goods. Third,panchayat resources need to be augmented through increased discretionary funds. There are at least two possible options- (i) consolidate some o f the schemes into untiedfunds for GPs; and (ii)improve the property tax systeminthe states. Fourth, the transfer system needs to be redesigned by making it transparent through a formula driven allocation. The critical element of reform is to unbundlethe transfer system from scheme based transfers into broad programs and impart flexibility and autonomy to local governments to make allocations according to their priorities. Fifth, an appropriate and properly functioning financial management system is necessary to provide meaningful information to policy makers and systems managers who need to steer decentralization on an ongoing basis. An important contribution of this study emerged from perhaps the primary difficulty encountered during its execution. The initial intent o f this work was to rely on data made available by the State Finance Commissions. However, initial tests on this data demonstrated that the fiscal information system in the two states does not provide an adequate base for designing, implementing and evaluating the decentralization policy. The development of a proper fiscal information systemi s an importantrecommendation of this study. Sixth, downward accountability depends not only on the design and effective execution of appropriate mechanisms for informing and involving communities in decision-making and performance oversight; but it also depends on local governments having enough money and enough expenditure responsibility to attract and holdthe interest of the community. Seventh, rural local government finance in India is part o f an intergovernmental system that includes the fiscal relations between the center and the states, the finances of state governments, and the finances o f the urban governments within the state. The system must be designed as a whole if it i s to succeed. Inparticular, an issue in India i s that the fiscal health and development o f local government finances is very much dependent on the fiscal health o f the state governments. IV. ISSUESINTHE CURRENT SYSTEM 12. The systems of intergovernmental transfers in the two states are different. InKarnataka, the total expenditure ofrural local governments i s equivalent to about 20 percent of the total state expenditure, with 20 percent spent at Zilla Panchayats (ZP), 18 percent at Taluk Panchayats (TP) and 62 percent at Gram Panchayats (GP). Eightypercent of the budget of the local governments is through a complicated system o f 428 state and central government conditional grants to rural bodies. The conditions on these schemes are stringent enough to leave these local governments with little expenditure discretion. The Gram Panchayatsreceive only eight of these schemes (grants), which, together, finance about 80 percent of GP expenditures. InKerala, on the other hand, 18 percent o f the state expenditures take place at the local government level with 22 percent spent at ZP level, 18 percent at TP level and 66 percent at GP level. About 80 percent of grants are for planpurposes and are allocated as general-purpose grants. Thus while both states would appear to have devolved similar shares of public expenditures to the third tier, Kerala has devolved more discretion. The decentralized planning process in Kerala has a built-in mechanism for bottom-up planning for the use of these funds. It seems clear that Kerala has decided that 6 the gram panchayat will be the principal unit of local self-government. There appears to be more ambivalence about this in Karnataka, where the gram panchayats are the only units o f rural local government with independent taxing powers, but account for only about 6 percent of total rural local government spending. 13. In both states the fiscal framework for rural local government finance has flaws, but the shortcomings in the two states are different. One might expect that the goals the government wants to achieve with fiscal decentralization are better local public services, increased revenue mobilization, a more efficient use of public monies, and more involvement of people in the process o f their governance. There are several reasons why the present system falls short of meeting these goals. The case studies of Kerala and Karnataka point to several problem areas, including expenditure assignments, local autonomy over budgets, fiscal role of the Grampanchayats, the transfer system, fiscal pressures and revenue efforts by grampanchayats3. These are discussedbelow. ExpenditureAssignment 14. In Karnataka the assignment of expenditure responsibilities i s not clear. Kamaka has nominally assignedall 29 subjects to panchayats. Butperhapsbecause some of these functions were thought to have important externalities, or perhaps because state politicians and bureaucrats were unwilling to relinquish power, Karnataka has retained the responsibility for a large number of key decisions regarding these services and left local governments with limited discretion. Further, relatively little has been done to unbundle the general subjectsprescribed inthe Constitution, and make expenditure assignmentsat a more detailed level of activity and sub-activity. As a result, there i s an unclear assignment of expenditures between the three tiers that has led to duplication of efforts, some failure to assume leadership in the delivery o f services, coordination problems and some uncertainty on the part of the local population as to where accountability lies. 15. In Kerala expenditure assignments are much clearer. Assignments took place at the level of "activities" and complementary legislation was issued to change the role of key line agencies. Functions have been divided according to the various tiers and in most cases broken into activities to avoid overlapping. Responsibilities are therefore clearer. Grampanchayat functions are broken into: (i) three mandatory functions with 27 items, (ii) general functions with 14 items, and (iii) sectoral functions 19 subdividedinto 74 items (see Volume IV). This clarity o f assignments made the state government more accountable in its implementation o f decentralization, which has been progressing since 1995. The change inexpenditure patterns for functions allocated to GPs i s illustrated inFigure 1. In 1995 only 32 percent of the expenditures for functions assigned to gram panchayats took place at the discretion o f the Gram Panchayat. Over a five-year period, this allocation has changed whereas the gram panchayat now has control over 68 percent of the expenditures. Details inVolume I11for Kamataka and Volume IV for Kerala. 7 Figure 1:Control over expenditures bygrampanchayats for functions devolved to them is increasing inKerala (case study on Tholicode gramp a ~ ~ ~ a ~ a ~ 1995-1996 GPBudget 68% Source: Fiscal datacollected for tbe study, 2003. (Details inVolume IV). LocalAutonomy 16, One major constraint to true fiscal decentralization is the lidted expenditure discretion that has been given to the local governments, though the situation varies in both the states. In Karnataka, there is almost no discretion in the expenditure of plan funds, Almost 50 percent are earmarked for salaries, 20 percent for transfers to individuals, and most of the reminder for input purchases where there is little local discretion. The situation is similar for non-plan funds (89 percent for salaries and transfers). b o , and others, 2003, inVolume In,estimate that inKamataka, even though a significant portionof public expenditures are being channeled through~ uineffect expenditures~ ~ ~ actually controlled by local governments are well less than 10 percent of total rural spending by local bodies, The lack o f independent sources of revenue and total dependence of ZFs and TPs on transfers makes the entire fiscal decentralization process for these levels hostage to the transfer system. There are almost no general purpose transfers at the district and block levels, The scheme-based devolution and micro allocation o f resources even within the schemes do not provide any leeway to the districts and in making allocation decisions according to their priorities. In fact, the entire fiscal ~~~~~ decentralization process has ensured that the district and block~ ~ are mere~disbursing~agents of~ a the higher level govements. However, gram~ ~because they have own~source revenues, and u ~ ~ ~ because they receive their tTansfers more inthe form of unconditional transfers, do have more discretion to allocate expenditures to meet local needs. It is estimated that they have complete authority to allocate a about one-third o f their budgets, and limited control over expenditures o f another one-third. InKerala, Gram panchayats have more expenditure discretion because grant funds tend to be unconditional, and there i s more locally raisedrevenue. However, inboth states, the higher-level governments retained most powers to hire, fire and compensate local government employees, hence Panchayat Raj Institutions (PRIs) have relatively little control over their personnel budgets. In Kerala, there is some administrative autonomy. For professionals such as doctors, engineers, agricultural officers, and so on, there i s a sort o f dual control with line departments exercising limited control over local employees. Although the state government determines compensation, the panchayat committee does have some powers to initiate disciplinary action. Fiscal Role of Gram Panchayats. 17. The fiscal role of Gram Panchayats in Karnataka is negligible and in Kerala, small. The role o f the Gram panchayats in rural local finance is quite small. Yet, the gram panchayat i s arguably the best choice for local autonomy given that this tier o f govemment i s closest to the people and the fact that it has some taxing power. The analysis for Kamataka (1998 data) shows that the level o f expenditures incurred at the GP level i s extremely l o w both in per capita terms (Rs. 68) and as a percent o f gross domestic product (0.38 percent). Even as a ratio o f total expenditure o f rural local governments, the role o f GPs is quite limited (6 percent o f total expenditures). Inaddition, overwhelming proportions o f expenditures are incurred for specific purposes such as JGSY4, welfare o f scheduled castes, and salaries. Of course, these are necessary, but it i s worth noting that they are directed to benefit a small section o f the population withinthe GPs. The activities o f GPs do not evoke appreciable interest and response from the majority o f people. Many have commented on the apparent lack o f local interest inthe business o f thepanchayats as evidenced by the poor attendance at the gram sabhas5.The level o f GP spending is significantly higher in Kerala. In 1999, the average GP spent Rs. 328, and total spending by GPs was equivalent to about 1.4 percent o f state domestic product. As a share o f total state govemment spending, gram panchayat expenditures rose from 4 percent in 1994to 8 percent in 1999. Revenue Efforts 18. Revenue effort by Gram Panchayats i s weak in both states. Kerala and Kamataka have given the gram panchayats independent revenue raising power, but in neither state has it been effectively used. Thus overall, in both states, own source revenues are negligible-less than one percent o f Gross State Domestic Product from their own sources6. Local revenues as percentage of total revenues are small by international standards (Figure 2). A number o f reasons have been suggested for this weak revenue perfonnance: 0 An unwillingness ofthe local officials to enforce the tax laws. 0 The limitedcapacity o f local government officials to effectively administer a revenue system. 0 Weak administration procedures, for example, tax rolls are not maintained, assessments are not done in a scientific way and are not updated, and collection procedures are not efficient. 0 The property tax base i s not defined ina way to be amenable to capturing the increase in property values for purposes o f taxation. 0 The need for more productive bases to tax. An employment generation scheme through labor intensive infrastructure. Village meetings for all voting population. Holding o f these meetings i s mandatory as per the State Acts. The ratio o f own source revenue to State Domestic Product i s 0.1percent inKerala, and 0.08 inKamataka. 9 0 Taxpayers see little benefit from paying taxes to local governments. 0 Local officials have little incentiveto raise additional taxes, that is, face littlepenalty if they do not, and little reward ifthey do. Relatively few of the Grampanchayats have opted to levy the highestproperty tax rate allowed. Figure 2: LocalRevenues as a Percentageof Total Revenues for Selected Countries (1999) 30.00% 25.00% 20.00% Localrevenues / Total Rev. 'I5 00% 10 00% 5.00% 0.00% Austria Brazil Canada India(*) Malaysia Mexico South Spain United Africa States =Owncalculations basedonthe report ofthe EleventhFinance Commission 2000-2005, India Source: Own calculations based on Government Finance Statistics Yearbook, 2000, IMF. Total revenues include all transfers (tied, untied, current and capital). 19. Collection rates on local taxes in Karnataka are very low. The estimated average amount o f own source revenues for GPs was Rs. 16 per capita and only 0.08 percentage of GDDP in 2001. That even this small amount accounted for more than 20 percent of total revenues reinforces the point made above that the overall role o f the GP inproviding public services i s negligible. Collection rates on the property tax are very low. On average, per capita collections were only Rs. 8 in2001, and showed a decline from the previous year. Almost 80 percent of GPs in the four sample districts that were studied in detail had property collections less than Rs. 10. The collection rate, on average, was less than 70 percent of the demand in2000-01. Inmany GPs, the level of collections i s so low that the viability of the property tax i s inquestion. 20. Accessibility to markets and public perception of service delivery are important determinants of revenues. There are lessons to be learned from the variation in per capita property tax revenues and per capita total own source revenues across GPs. Why do some Gram panchayats collect more than others? Per capita own revenues and per capita revenue from property taxes were regressedon a number o f variables representingrevenue capacity and the demand for services in the GPs in the four sample districts in Kamataka. These equations bring about some interesting insights. There seems to be no relationship between per capita revenues (including per capita property tax) and the level of development ina GP'. This is not surprisingina systemwhere the property tax is leviedas a lump sumtax onhousing units. The lack of relationship between per capita own source revenues and the proxies of agricultural incomes is surprising. Interestingly, per capita property taxes and per capita own revenues are significantly and positively related to the variables representingaccessibility to markets - the roadlength per square kilometer area and proportion of metalled roads. We also find that total own source revenues For more details refer to Volume 111-Karnataka Case Study. 10 tend to be higher inGPs that are closer to a district or block headquarters, primarily because non-property tax revenues (fee and charges) are higher. The absence o f a relationship with the property tax again points out the inability o f the property tax system to take advantage o f the higher values o f property in GPs proximate to major towns and cities. The findings about public perception o f service delivery are also interesting. For example, the availability o f water supply in a GP does not tend to be associated with increased revenues for that GP. 21. In Kerala own revenues have been increasing even with larger transfers (Figure 3). InKerala, there are some parallels with the Karnataka experience, but also some significant differences. In 1999, own source revenues accounted for about 18 percent o f total GP revenues, down from 31percent in 1994. The decline was a result o f the increase in state transfers and not due to a decline in local own source revenues. Infact, the level o f per capita own source revenues increased from Rs. 24 in 1994 to Rs. 57 in 1999 (Rs. 33 in real terms). Therefore, while the level o f collections is still weak, it i s well above the average level o f Karnataka, and there has been some growth. Within own source revenues, non tax revenues, however, have been much more buoyant than tax revenues. Per capita amounts have increased from Rs. 7 in 1994 to Rs. 29 in 1999. The tax base does not appear to have responded to a booming construction sector in the state. Nor do they reflect the relatively large inflow o f foreign remittances. This is likely due to a failure o fthe valuation systemto capture increases inproperty values. There is also evidence that panchayats may be unwilling to push aggressively for higher property taxes: o f 50 GPs sampled for detailed analysis, only three have chosen to use the maximum tax rate permitted by the statutes and 23 have chosen the minimumrate. It i s estimated that only about 40 percent o f the revenue potential has been captured. 11 Figure 3: Growth in Per Capita Own Source Revenues and State Plan Grants (Transfers) in Kerala Own Source Revenues of GPs 60 -4-Per Capita (Nominal) 50 da 40 3c4 ''I30 20 10 0 1993-94 1994-95 1995-96 1996-97 1997-98 1998-99 State Plan Grants to GPs 250 1993-94 1994-95 1995-96 1996-97 1997-98 1998-99 Source: Data collected for the study, 2003. (Details inVolume IV). 22. There are wide variations in per capita own source revenues among GPs in Kerala, and much can be learned from understanding the determinants of these variations'. Per capita taxes range from Rs. 4 to Rs. 327 around a mean of Rs. 27, and per capita non taxes fkom Rs. 9 to Rs. 586 around a One might ask why the localrevenue system has performed so much more poorly than the state system. There are a number o f answers to this. First, the state has access to broadbased taxes on income and consumption whereas the local govemments may tax only property and levy user charges. The built-in growth for these local taxes i s less strong, and require discretionary adjustments (e.g., revaluation or changes inuser prices) to realize the increase. Secondthe growth inthe tax base i s largely inthe urbanrather than the rural areas. The growth inrural economies i s usually located insectors where tax assessmentand collection i s most difficult. The problem i s compounded by the weaker tax administration skills at the rural local government level, and the inherently greater difficulty that local officials have with collected from the local population. Finally, rural local governments inIndia have a difficult time showing local voters wheat heir money buys. 12 mean of Rs. 57. Usinga multipleregression analysis, the Kerala study i s able to explain about 25 percent o f the variation among the 615 Grampanchayats for which data i s available. Ingeneral, per capita own source revenues are significantly higher in GPs where the level of economic activity is stronger, where population i s larger, and where there i s a smaller land area. The percent of SC-ST population is not a significant determinant. We find evidence o f "district effects", that is, after all else i s accounted for, GPs in some districts raise significantly more than GPs in other districts. This suggests the possibility of revenue effort beingrelatedto local cultural or leadership factors. The Transfer system 23, In Karnataka the transfer system is complicated, non-transparent, and non-equalizing. There are 428 schemes (conditional grants) to rural local governments from the state and the center, and it i s difficult to coordinate these or monitor their effectiveness. Because the conditions on the expenditure of these grant funds are so restrictive, local governments have little opportunity to structure local expenditure to match local preferences. A small portion of the total grant funds i s given to Gram panchayats as untied aid, and this does provide some measure of local discretion. However, the distribution of untied grants among GPs i s done on a basis o f equal amounts per GP. There i s no evidence of equalization inthe distribution of these grants, in fact, even the poverty alleviation grants do not go more heavily to panchayats with a greater concentration of poor familiesg. It would not be far off the mark to say that the present system of intergovernmental transfers in Karnataka i s inconsistent with the goal of developing fiscal decentralization inthe rural local government sector. 24. The transfer system in Kerala is less complicatedand more equalizing. Inthe Kerala system, the Gram Panchayats do receive grants through sponsored schemes where the local bodies act as agents of the central and state govemments. Plan grants and the sponsored schemes together form about 35-40 percent of the state plan. For the sponsored schemes broad guidelines are issued by the sponsoring govemment or department. Even so, the panchayat decides the location o f the project and selects the beneficiaries, besides talung up the responsibility for the implementation o f the project. But the most notable feature of Kerala's fiscal decentralization is the large share of untied grants, over 80 percent in 1998-99, in the structure of transfers to gram panchayats. Both inpercentage and per capita terms this represents a level that i s probably unique in all of India. The Kerala govemment grants to local bodies consist of non-plan and plan categories. Non-plan grants are broadly divided into statutory and non- statutory grants. Statutory grants are the largest of these two components and consist o f assigned taxes and vehicle tax compensation. The assigned tax consists o f a basic tax on land and a surcharge (4 percent) on stamp duty collected under the provisions of the Kerala Stamp Act, 1959, on the value of the property transacted. The entire amount i s meant to be transferred topanchayats after deductingcollection charges. Besides these taxes, 20 percent o f the Motor Vehicles Tax's net collection, also called Vehicle Tax Compensation (VTC), i s to be distributed to GPs on the basis o f the length and types o f roads in a panchayat. Statutory grants accounted for about 13 percent of all transfers to GPs in 1999, and there i s local government discretion as to the expenditure of these funds. Plan grants are the most important and were the preferredinstrumentto devolve more resourcestopanchayats inthe "big-bang" decentralization. The per capita state plangrant has increased from Rs. 12in 1993-94 to Rs.207 in 1998-99, an increaseof about 15 times in a period of six years. This corresponds to an increase from 3 percent to about 12 percent of state revenue. This increase i s probably overestimated though since panchayats have only been receiving a share of their entitlements". Overall, the distribution o f transfers among Gram panchayats in Kerala would appear to be somewhat horizontally equalizing. This i s largely due to plan grants, which appear to be objectively distributed, and to favor those GPs with larger shares of SC-ST population, and to discriminate against those GPs with larger populations and those having a greater share of better constructed housing. Essentially the same pattern i s found for central government grants. For more details, refer to Volume 111-Karnataka Case Study. loSee Volume 11, Chapter 111, Section F. 13 25. However horizontal disparities are widening. Inboth states, there are wide variations inper capita expenditures. InKarnataka, this disparity i s not only between different districts but within each of the districts analyzed. In general, per capita expenditure of GPs is highest in the more affluent districts though on average, 37 percent of the GPs had per capita expenditures of less than Rs. 20. Gram Panchayats with a greater proportion of scheduled castes and tribes spend significantly more, cet. par., and intergovernmentaltransfers have a greater impact on expenditures than do revenues raised from own sources. The story i s similar in Kerala, where there are also significant and growing disparities across gram panchayats. Statistical analysis of this variation suggests that per capita expenditures tend to be higher ingram panchayats with smaller populations, larger land areas and greater concentrations o f SCST population. It i s the distribution of plan expenditures, which are grant funded, that brings some degree of equalizationto this overall expenditure pattern. FiscalPressureson Gram Panchayat 26. The fiscal pressures on GPs are growing. This i s especially true in Kerala where the gram panchayats play a more significant budgetary role. The volume o f expenditures the panchayats incur or manage keeps expanding. As functional devolution expands, with several institutions already transferred to GPs and with many panchayats forced to take several "agency" functions on behalf of the state and central governments, their expenditure responsibilities have increased. The Kerala analysis examined finances of a representative sample of 26 gram panchayats for which detailed revenue and expenditure data were gathered. The fiscal balance inthis exercise was measured as the difference between the sum o f own source revenues, shared taxes, and assignedtaxes and recurrent expenditures. The results of this calculation show that there i s no panchayat that generated a surplus from own revenue (including assignedtax grants) consistently for the four years from 1999 to 2002 inthe sample of 26 GPs. Local GovernmentAccountingSystems 27. In both states, outdated local government accounting systems have been overwhelmed by the requirements of decentralization. In both states, control mechanisms o f the state government over financial management in the panchayats appear to be relaxed. The local government accounting system i s in danger of collapse and annual audits have fallen in arrears. It i s a cash-based accounting system. The registers and books of accounts are outdated and unable to account for the increased and diversified flow o f resources o f the present decentralized system. Moreover, there are no uniform accounting codes or rules for panchayats. GPs still rely on the old Panchayat accounting rules, while TPs and ZPs do not have designated accounting rules. As a result of this state o f affairs, ongoing monitoring of financial management i s difficult. It is not surprisingunder these circumstances that panchayats do not capture fully the extent ofthe financial resourcesthey handle. 28. In Karnataka, duplication and overlapping responsibilities confuse authority and accountability, as well as complicate delay reporting and reconciliation. No one knows - or could know - cashbalances on a day-to-day basis. A general ledger i s not part of the system, making it difficult to reconcile receipts and payments from multiple sources. The multiplicity o f accounts, accounting centers, accountants and reporting streams result in delays of several months in reporting, and limit oversight by authorized bodies. This i s further complicated by the many budget items (40-50) through which transfers are made to PRIs. Thepanchayat accounts themselves are incomplete, inpart because o f the various ways inwhich payments o f electricity bills are reflected. 29. Some consequencesof these accountinginadequaciesare serious. Budgeting seems to have been reduced to a compliance formality. There i s little use of the budget for fiscal planning. There i s no linkage between the annual plan and the budget, making monitoring and evaluation difficult. There i s a lack o f reliable information about the assets required to be maintained by the panchayats. It i s difficult to determine unit costs o f projects and various services like water supply, street lighting, sanitation, and so on. Also, the economic viability o f proposed projects is unclear, State Finance Commissions have had to 14 assessthe financial requirements of the PRIsby informed guesswork more than from data. 30. Meanwhile, in Kerala, old systems are under stress and new ones have not yet emerged to take their place. Since 1996-97, panchayat funds have tripled and the state government's grants to Gram panchayats have increased ninefold. Despite this substantial increase in the flow of resources, financial management i s still led by accounting rules that were framed more than 40 years ago. In addition, the state government has complicated the situationby issuingnumerous regulations for the use of these funds. The result is a complicated accounting system, and one that is yet to be computerized. 31. Consequences in Kerala of these inadequacies are also serious. Althoughpanchayats are required to prepare an annual statement of all receipts and expenditures, consolidation of accounts does not take place in any of the three tiers of rural local government. GPs prepare an Annual Financial Statement, but it does not include Plan funds or funds received for Centrally Sponsored Schemes. TPs and ZPs do not prepare an Annual Financial Statement. Inthe absence o f consolidated accounts and financial statements, itis difficult even for thepanchayat committee (not to mentionthe general public) to know the extent of receipts and expenditures as well as the financial position of the local body. InformationSystemon Local Finances 32. Neither state has an adequate information system in place on local finances. There i s no adequate information system which can be used by local governments to plan their revenue and expenditure programs, or which can be used by higher-level governments to monitor the fiscal performance of local governments. The district and block governments in Karnataka do not all keep budgets and the data available for gram panchayats i s incomparable at best and flawed at worst. A similar situation holds in Kerala. These issues of reliable information on finances and financial performance have important implications for planning at federal, state and local levels. As things stand at present, recommendations for fiscal reform may be made based on data that misrepresent the situation. Most important, assessment of the fiscal health of the local governments may be substantially incorrect, leading to misguided attempts at reform. 33. A result of this problem is that the high closing balances posted bypunchuyats in "official data" are misleading in both states. Local government data supposedly collected as part o f the normal fiscal monitoring process andused by the Central Finance Commissions and the State Finance Commissions are erroneous. The data inflates the amount o f resources at the disposal of panchayats, suggests a budget surplus, and may create temptations for fiscally strapped state governments to further reduce the already negligible funds at the disposal ofgrampanchayats. What seemedimplausible to the research team is that cash-strappedstates would allow this condition to continue for an extendedperiod. The data does not tell a transparent story on local financial performance for three reasons. First, official records o fpanchayat revenues are basedon budgeted allocations rather than funds received; second, the significant irregularity inthe timingof transfers received by local governments distorts both actual expenditure patterns and the reporting of these expenditures; and, third, weaknesses in record keeping systems and practices are prevalent. As a result the local finance data reported to State and Central Finance Commissions shows large closing balances inpanchayat accounts. If this were accurate it would imply that the panchayats were receiving considerably more revenue than they hadthe capacity to spend. InKarnataka and Kerala, these reported balances were equivalent to about 50 percent and 70 percent o f annual revenues, respectively. Such a situation would question the efficacy o f these local governments as development agencies. In fact these balances do not describe the realfinancial position. 34. Some steps to improve accountability have been introduced. Kerala is the only state in the country where full recognition has been given to the principle that the primary accountability o f the local government institutions i s downwards to the local community. Various innovative tools and methods introducedduringthe people's plan campaign to make the panchayat accountable to the community have not yet been perfected. These include social audits, the creation o f beneficiary committees, right to 15 information initiatives, citizen charters and an ombudsman. All of these initiatives have run into problems and are yet to mature. V. RECOMMENDATIONS 35. The Constitutional Amendments of 1993 commit the Indian federalism to development of a third tier of government with some degree of fiscal autonomy. Though Kerala and Karnataka are two states that have been aggressive in promoting decentralization, both still face significant problems with their programs. The structure of the local government financing systemwould seem more instep with the objectives o f decentralization inKerala, but implementation has been a problem here as well. Kamataka, on the other hand, despite important strides on other aspects of decentralization, has not yet moved to a financing structure that is consistent with the goals o f fiscal decentralization. Neither state has prepared a blueprint that can serve as a guideline for resolving such problems and for developing its rural local government financing policy. A statepolicy paper on fiscal decentralization goals, and an actionplan for achieving these goals, should be prepared by each state. The following are some of the basic recommendations that would be considered for this paper. A. Clarify ExpenditureAssignments andLocalGovernmentHierarchy 36. A first step towards developing a workable fiscal decentralization program for rural local governments i s a clearly designed set of expenditure assignments. The expenditure assignments shouldbe set by the state government, and should largely replace the "concurrent approach" that is now in place. The expenditure assignments should be spelled out in some detail at the level o f sub-functions. Only then can there be a basis for deciding on the proper vertical sharing of state revenues among the districts, blocks, and villagepanchayats. 37. State governments might rethink the roles of the three tiers of rural local governments. International comparisons do not reveal a clean pattem as to the optimal structure of local governments. Some countries prefer a smaller number and larger size, such as Indonesia and South Africa, while others seem to favor a small population for the lowest unit, such as Ukraine, Switzerland and Sweden (Table 1). No clear patterns separate developed from developing countries either. However, three tiers of rural local governments below the state level with overlapping responsibilities as in India i s costly, invites duplication of effort, and confuses the accountability objectives of the govemment's decentralization program. For all practical purposes, districts and blocks are now spending agents of the central govemment with relatively little autonomy to shape the expenditure budget and no taxing powers to affect the overall size of the budget. Moreover, these units of rural local government are considerably larger in population size than the Gram Panchayats, hence they may be less suitable as the basic units of decentralization. Thought might be given to definingmore clearly the role of the districts and blocks as either primarily deconcentrated arms o f the state government, or as autonomous govemments representing the preferences of their constituents. Several initiatives need to take place. First and foremost is to clarify lines of accountability o f functionaries. Are these accountable to the elected councils or to the line agencies? And if they are accountable to line agencies, what i s the role of elected members? Untilthese roles are clearly specified, district and block management will remain confused. 16 Germany 3 286 31,010 Switzerland 3 2,903 2,473 Canada 3 936 32,853 Japan 3 3,229 39,291 17 B. Strengthen LocalGovernment Budgetary Autonomy 40. Consolidate schemes in Karnataka. Rural local governments inKamataka are financed primarily by conditional grants or "schemes." Responsibility for many central and state programs has been passedto local govemments, but discretion over the use of these funds has been heldby higher-level govemments. There are over 400 such schemes now operating inKamataka. Those schemes that match the expenditure responsibilities o f the gram panchayats should be consolidated into a small number of general-purpose grants, and passedto the gram panchayats. All other schemes shouldbe consolidated within each o f the state departments so that there can be greater flexibility to prioritise spending decisions within the departments. This consolidation will serve several purposes: it will make the intergovemmental transfer system more transparent, it will allow a better match between expenditures for rural development and local preferences, and it will significantly augment the resources of the gvampanchayats. Ineffect, this i s the strategy that has been followed by Kerala. 41. Local governments should have power to hire, fire and determine the compensation level for their employees. Without this discretion, it i s difficult to see how local governments can control their budgets and be held to a hard budget constraint. Moreover, this would eliminate the dual responsibility that conflicts local govemment employees between accountability to the government unit for whom they work, and accountability to the level o f govemment that determines their well-being in the workplace. Effective implementation o f service delivery by rural local govemments would seem to require accountability of their officers. 42. At the same time, it is important to ensure clarity in the role of line agencies. These agencies today implement schemes for the state, district, block and village panchayats. This undermines accountability. Moreover with fiscal decentralization the roles of state line agencies for functions devolved to panchayatsneeds to change accordingly. C. Augment Panchayat Resources 43. Augment gram panchayat resources. While the gram panchayats have the constitutional mandate to perform the role of local self govemments, they do not have enough discretionary resources to do so in a meaningful way. InKarnataka, total revenues available to gram panchayats from all sources are Rs. 68 per capita (less than 0.4 percent of gross state product). The result is that the impact of the local budgets i s too small to attract the attention of the local population, and the goals o f better accountability and more involvement o f citizens ingovernance are not reached. Kerala is insomewhat better shape with per capita spending o f Rs. 328 (1.4 percent of gross state product), but even this amount i s not likely to stimulate great local interest inbudgetpriorities and outcomes. Therefore, a keypriority inbothstates i s to augment grampanchayat resources. There are many ways to do this. 44. In Karnataka, consolidate some of the schemes into untied funds for gram panchayats. This consolidation should include state and central schemes as well as centrally sponsored schemes. To the extent this could include a redirectiono f some funds away from districts and blocks, it would augment the revenues o f grampanchayats. 45. Increase own revenues of grampanchayats. Reforming the tax system and user charges levied by the GPs i s one of the most important elements of the fiscal decentralization reform. Revenue performance i s very weak inboth states. The problem i s not primarily the absence of revenue sources, or the absence of discretion to increase revenues. In fact, the results for both Kamataka and Kerala show a wide variation inper capita revenue, with some GPs performing better than others even under equivalent fiscal capacities. Moreover, both Kamataka's and Kerala's empirical results reject the hypothesis that grants reduce the revenue collection of local bodies. The problem seems to be an unwillingness of the GP to enforce the taxes available to them or even to levy the maximum allowable tax rates. While this i s not the case for every GP, it is the case for many local govemments inboth states. The underlying problems 18 that lead to a low revenue productivity have to do with an inadequate tax structure, a weak tax administration and perhaps with the disconnect between higher taxes and what voters see as the benefits o f higher taxes. A number of reform options discussed below might be considered to address these problems. 46. Improve the structure of the property tax. At present, revenue collections are low and inadequate to cover even the cost of basic local public services. InKerala, flaws in the design o f the property tax system have preventedPanchayats from benefiting from the real estate boomtaking place inthat state, or from significant foreign remittances. In Karnataka, the growth in property values in more prosperous areas has not been captured. A major problem in both states i s the present system o f valuation. Houses are valued at their estimated rental value, and this amount i s not changed frequently. The values are set rather than being based on some scientific approach. The move to a more systematic system of valuation, even if very basic, and regular revaluation, would significantly increasethe revenue yield o f the property tax in both states. (Box 2). How to do this? One approach, already used in Bangalore, i s to fix value against location, area and construction materials of structure to develop a simple system o f determining area-based values. This i s not a perfect system for valuing, but it can be used to increase the size of the property tax base inan approximate fair manner". 47. Improve collection of the property tax. The GramPanchayats do not collect even the low amounts o f taxes that are levied. Analysts in both states see the property tax as more voluntary than mandatory, and place the blame on a weak enforcement system. The state governments may take the lead here by helping to improve the capacity o f GP tax administration personnel, providing technical assistance to improve property tax administration systems, increasing the penalty rate for non-payment, and providing incentives for increased collection efficiency. A first step might be to provide technical assistance to local governments to do a complete update o f their tax rolls, and to enhance the capacity o f local officers to maintainthis roll. 48. Improve collection of user fees and charges. The KarnatakaandKerala case studies show that non- tax revenues (user charges, rents, licenses and fees) are as large inamount as tax revenues and in some cases are growing faster. This may signal an enhancedability to collect revenues when a direct linkage to a service is present. However, inbothstates there i s a wide variation inthe use of non-tax revenues. This i s due to some combination o f different opportunities, different willingness to levy such charges, and differences in the capacity of the local tax administrations to put such a system inplace. The two states could provide a technical assistance program to organize the assessment and collection of user charges, and it could facilitate the transfer o f information among the thousands o f GPs involved. If an incentive program i s offered to encourage increased local government revenue mobilization, it i s important that it include non-tax as well as tax revenue. "Another approach,whilenot the best approach, is that states couldincrease property tax revenues by simply charging the maximum rates available. 19 Box 2: Local Property Taxation Most local governments in developing and developed countries are given the right to tax property, and most use this power. Intemational "expert" opinion generally holds that this is a "good" tax for local governments. We may also note that most local governments in developing countries are fiscally strapped. They are usually assigned few tax bases, and they often are unable to effectively administer those taxes that they are given. This seems especially true for the property tax. So, casual observation would suggest that the "good" localtax i s not getting the job done. There are significant advantages to the property tax. The property tax can be revenue productive, and is often the mainstay o f local government fiscal systems. It accounts for about 0.4 percent o f GDP in developing countries, a share well below that inindustrialized countries. While this may seem a small amount, it shouldbe pointedout that it is large enough to finance about 11percent o f sub-national government spending. Some have argued that there is faimess inthe property tax inthat property values, and therefore property taxes, rise with the provision o f better local services to the property. There would seem to be faimess in asking the benefiting property owners to pay for these services. The property tax is most probably not regressive, and may vary from proportional to progressive. Inan ability-to-pay sense, it couldbe viewed as equitable. A significant portion o f the residentialproperty tax, andsome o f the non-residential tax, is bome by landowners who usually sit atop the income distribution. Many o f the poor are screened out o f the tax, by explicit exemption of properties o f value below a certain amount. Moreover, the property tax captures in its base those who hold wealth but may be missed by the income tax. Perhaps the biggest disadvantage o f the property tax is its unpopularity with local residents and voters. Why is the property tax so unpopular, and why does it draw out such protest? First, it is a very visible tax. Unlike the Value Added Tax (VAT), for example, where no one reallyknows how muchthey pay, the property tax is paid annually or inperiodic submissions, and the taxpayer is quite clear about hisiher liability. Second, assessment isjudgmental, unlike sales or income taxes. Taxpayers are uncomfortable with this. Third, the property tax is a levy on accrued income, not on realized income, which seems unfair to many taxpayers. Another major disadvantage is that administration costs are high if the tax is to be well administered. Valuation, done properly, i s an expensive proportion, as is effective enforcement. The major problems with the property tax becoming a significant revenue source for local governments in developing countries are problems related to valuation and enforcement. With respect to the former, most countries just do not have the procedures or staff inplace to cany out the job. With respect to the latter, there i s some question about the willingness o f local officials to enforce such an unpopular tax, especially when the burdenoften falls most heavilyon localelites. This set of advantages and disadvantages has led countries around the world to use the property tax as a revenue source, to widely varying degrees. Generally, industrialized countries make greater use o f the property tax than do developing economies. As noted inthe table below, the reported intensity o f use o f the property tax inIndia is quite low. Sub-national Property Tax as a Share o f GDP in Selected Large Countries Country Property Tax/GDP Country Property Tax/GDP Canada 4.07 Indonesia 0.32 United States 2.87 Mexico 0.31 Australia 2.49 China 0.17 Russian Federation 1.24 India 0.10 Argentina 0.92 20 49. Broaden the tax base. The main problems with revenue performance at the panchayat level are administration, capacity of the tax offices, and willingness to tax. However, there are some possibilities for broadening the tax base that are worth considering as policy reforms. There would seem to be two good options for broadeningthe tax base. The first i s for the states to allow rural local governments to tax agricultural land as part o f the property tax regime. While enforcement may be difficult, the revenue potential i s significant, and this may be a more appropriate tax base in communities that have a more developed agricultural sector. A secondpossibility for base broadening i s a rudimentary form o f business tax. This is used to good effect in Kerala (though it could be strengthened), and could add value to the revenue systemof Karnataka's rural local govemments that are more inthe nature of "commerce towns." 50. The states should appoint a technicalassistanceteam to begin improvementof at least property tax administration. At a minimum,the focus of this technical assistancecould includethe development and implementation of statewide standard procedures for valuation and revaluation, maintenance o f the tax roll, and training. Some incentives for more intensive use o f the property tax also need to be developed. D. RedesignTransfers 51. Transfers should be determined in a transparent manner. The state should establish a clear and transparent way inwhich the aggregate transfers, the relative shares of urban and rural local governments, and the allocations to each of the tiers of urban and rural local governments are determined. This does not happen now inKamataka. The present allocation evolvedhistorically, based on the expenditures incurred before the schemes were transferred to the local bodies. There has been little attempt to estimate the expenditures neededto carry out the functions assigned to the local governments, hence the determination of the vertical share ofpanchayats may be significantly flawed. 52. Unbundle the transfer system from scheme basedtransfers into broad programs. The critical element of reform i s to unbundle the transfer system from scheme-based transfers into broad programs and impart flexibility and autonomy for the local governments to make allocations according to their priorities. This i s particularly so for the grampanchayats, which, as discussed above, i s perhaps the only meaningful tier of local government. If states choose to maintain districts and blocks as essentially deconcentrated arms of higher-level governments, even here, consolidating schemes and increasing flexibility would go a long way towards reducing administrative burden and improve expenditure efficiency. InKerala, significant strides have already been made inthis area by converting many of the schemes into untied grants and giving significant budgetary discretion to especially the grampanchayats. 53. Relate transfers to need and capacity. The distribution o f transfers among the gram panchayat governments inKerala i s done on a basis o f an objective formula that contains some elements that suggest equalization. InKamataka, by contrast, the allocation among gram panchayats for untied grants i s done on a basis of "equal" total amounts, that is, the same rupee amount to each GP irrespective o f population size or composition, economic structure, level of development, location, and so on. As such, it does not take into account either the capacity to raise revenues or expenditure needs (and poverty) o f the GPs. While it i s necessary that the transfer system should be simple, it should not ignore the equalization principle altogether. We can note from the quantitative analysis carried out inKamataka, that even when tied grants are considered, there is no relationship between the distribution of these grants and the concentration o f scheduled castes and tribes in the GP population. A high priority reform i s for the Kamataka system to be changed to reflect whatever degree of equalization the govemment desires with respect to fiscal capacity, or expenditure needs. Inview o f the great advance Kerala made in measuring poverty at the GP level, the question o f refining and usingthe poverty index as a criterion may also have to be examined. (Box 3). 21 Box 3: Allocating Transfers By Formula Many countries that are decentralized and many o f those that are deepening their fiscal decentralization, distribute intergovernmental transfers by formula. A formula is transparent and objective, and once established, does not change regularly. However, an effective formula-based distribution o f transfer requires up-to-date and reliable information on the formula elements. The smaller the recipient unit o f government, the more difficult it is to identify a formula. While the use o f formulae have been fairly widespread for distribution from central to state governments, the practice is much newer for state-local revenue sharing and is much more constrained by data limitations. Some governments intransition and developing countries do use formulae to allocate grants to the bottom tier local governments. Insome cases the formulae are more complex, and insome cases a significant amount o fjudgment is required in developing the indicators to be used. An interesting example is the "equitable shares" grant in South Africa, which is an unconditional central government grant to urban and rural local governments. I t is distributed according to (a) the number o f poor households in the municipality (that is, those spending less than a "threshold" amount), and (b) the cost o f delivering a standard level o f basic services to a household (electricity, water, sanitation and rehse removal). The central government sets both the poverty threshold and the cost for basic services. Both are determined as uniformnational amounts. Indonesia introduced a big-bang fiscal decentralization in 2001, withthe centerpiece being a formula grant system. The central government distributions of this grant were made directly to the lowest level local governments. The formula elements are population size, poverty rate, land area, and the construction price index. The weights on each o f these formula components have changed, depending on political and economic considerations. Population and land area were given larger weights inthe 2002 distribution. Many countries with three levels o f government distribute grants to the lowest tier using a variety o f formula methods. Generally, there are many different grant programs in a single country, and these are distributed among local governments inmany different ways. Among the objective (and subjective) indicators that countries have used for distribution to local governments are population size and poverty (Mexico), the value o f real estate (Brazil), the balance between expenditure needs and fiscal capacity (Russia and China), and equal shares (India). 54. The States should obey their own rules. A major problem inbothKamataka and Kerala is that the state governments do not pass full transfer entitlements to their local governments. In effect, state deficits are being passed down. This practice eliminates transparency, discourages the objective of decentralization and moves the system of intergovernmental fiscal relations toward a yearly ad hoc arrangement. The state should budget for that level of transfers that it can make, and distribute the full amount ina timelyway. 55. Make transfers predictable. Due to the system o f funds flow, multiplicity o f accounts, many accounting centers, and antiquated recording of information, no one is sure about how much money actually flows to the panchayats. The timing o f this flow, with "bunching" at the end o f the fiscal year, and the failure of the treasury to release full entitlements, subjects local fiscal planning to great uncertainties. This situation undermines any attempt at planning or budgeting at thepunchayat level, and confounds attempts to monitor the fiscal performance o f local governments. Reducing the uncertainty in the transfer system would go a long way towards improving fiscal management at the Punchayat level. And this can be achieved easily through states transferring panchayat entitlements on agreed upon dates and amounts. 56. Determine the vertical share of rural local governments objectively. An objective determination of the vertical share of rural local governments (the entitlement of the local government sector inthe state budget) should be established and heldto. There are several choices for the determination o f this vertical share, but an ad hoc, annual determination i s least consistent with the goals of fiscal decentralization and improved responsibility and accountability o f local governments (Box 4). 22 Box 4: Systems for Vertical Sharing of Intergovernmental Transfers A crucial feature of the designof intergovernmentaltransfers is the determination of the vertical shares, that is, the entitlement of the rural local government sector. It i s important that this be transparent, so as to encourage efficient fiscal planning by local governments and to make possible the imposition o f hard budget constraint on them. In theory, in an equalizing grant system, the vertical share i s the amount necessary to fund that part o f minimum expenditure needs that local fiscal capacity cannot fund. Inpractice, such a calculation is rarely made. Generally, countries use three methods to establish the vertical share entitlement. The first is an ad-hoc arrangement. Each year, or for a period o f years, the higher-level government decides on the amount o f the entitlement. This gives the higher level government maximum discretion in adjusting the vertical share to fit the exigencies o f the current budget situation. A variant of this i s to appoint a commission that makes or recommends a decision on allocation for a period o f years, such as i s done inIndia andAustralia. The second approach is to fix the vertical share as a percent o f revenues o f the higher-level government. This is perhaps the most transparent o f the approaches, and (if the higher-level government follows through on its commitment under this approach) gives the local governments a guarantee o f resources to be received. Indonesia allocated 25 percent of internal taxes to the revenue sharing pool for local level governments, and the Philippines allocates more than one-third o f total internal central revenue collections, lagged by two years. China, Russia, Mexico, Argentina and Brazil are countries that dedicate percents o f their income tax and VAT revenues to the revenue sharing pool. Many countries mandate that a percent o f the funds distributed to provinces or states be passedthrough to municipalities (Nigeria, Mexico and Brazil, for example). The third approach is more in the vein o f cost reimbursement. The higher level government determines those functions of local governments that it will support financially, andprovides conditional grants that are determined in amount by the higher level government. Grants by states to local government inmuch o f India ("schemes") are o f this type. The most common functions that are singled out for conditional grants are basic inflastructure projects, teacher's salaries, and welfare-health related outlays. Virtually every country in the world has some form o f conditional grants inits system o f intergovernmental transfers. Which is the better way o f vertical sharing? I t depends on the objectives that the government most wants to achieve with its system o f intergovernmental transfers. A guaranteed share of revenues will probably give the local governments the most predictable revenue stream, and the most insulation from the political impacts on budget- making by the higher-level government. While this would seem to be the approach most consistent with fiscal decentralization, it has the drawback o f limiting the fiscal flexibility o f the higher-level government. 57. Relate transfers to state revenues, not to state outlays. The present approach to vertical sharing calls for some re-thinking. If the entitlement of local government i s largely linkedto plan size, then it will vary from year to year depending, inpart, on expenditure choices madebyhigher-level governments. Ifitis linkedto total staterevenues, itwillbe subject to the fluctuations instatereceipts. Thebetter route, and perhaps the more transparent and less politically-driven approach might be to base the vertical share of local governments on total state revenues rather than on planned expenditures. This share could be changedperiodically by the SFC. 58. Institute an incentive for increased revenue mobilization. The transfer system could be amended to provide an incentive for increased revenue mobilization. While this will be difficult to design and monitor, it might be a useful way to stimulate GPs to increase their tax effort. A special pool of funds might be set aside for distribution among those local govemments that reach a certain tax collection plateau set out by the state government, or even to those that implement certain desirable features intheir tax administration system. This will almost certainly not be an equalizing feature of the intergovernmental fiscal system, but it could encourage increasedrevenue raisingefforts. 23 Box 5: Problems in Redesigningthe Transfer System: the Kerala Example The design o f a transfer system i s difficult, especially if the subjects are rural local governments. First, the quality of the data to be used in the distribution formula may be suspect, or the information may be dated. For Kerala, several indicators o f service needs and fiscal capacity are available, but for many o f these indicators, the data is missing for a significant number o f GPs. These indxators would have to be eliminated as possible candidates for inclusioninthe formula, or the data set will need to be improved. The situation i s no better inthe case o f measuring taxable capacity. There is no data on personal income or economic output at the GP level. Some data on housing might be gathered regularly (for example, number o f housing units o f various types) and could be suitable for measuringproperty tax capacity. However, gathering such data could be a costly proposition. Measuring poverty i s problematic. Officialpoverty rate indicators are not available at the GP level. Ideally, the state will allocate its resources more heavily to those GPs where the concentration o f poor families is greatest. The measure used inthis study, the percent o f population in scheduled castes and tribes, is imperfect at best. Yet it i s poverty as much as any indicator that the state would like to factor into its distribution formulae. A second constraint is that an equalization formulae may allocate significantly more funds to smaller, poorer gram panchayats. Will these local governments have the capacity to effectively use the additional funds? A thirdconstraint to grant reform is that any new formula will result inboth "winners" and ``losers.'' That is, some GPs will receive more than under the previous system, and some will receive less. Those who lose will object, probably strenuously, and will point out that at a minimum,there i s needfor an adjustment period. The result is that any new formula will need to be phased inover a period o f time inorder to protect the losers from a "shock effect." There are many ways to design a hold-harmless system, but an essential ingredient must be that the phase-in period has a clearly defined end point. In order to identify the extent o f this constraint, we need to know how much the present system o f distribution would be "shocked" by the introduction o f a new system. The greater the shock, the more complicated the design o f the hold-harmless system. A fourth constraint is that related to building an intergovernmental transfer system. The systemfeature implies that all the components (grants, statutory grants, and so on.) fit together to achieve the same objectives. The easiest way to do this is to merge all transfers into one unifiedgrant system. However, some will argue for more than one component and more than one objective, and this will raise the possibility o f offsetting effects (for example, component A will be equalizing while component B will be counter equalizing, and so on). What this discussion points to is the needfor the State to begin with some realistic decisions about the objectives it wants to acheve with its intergovernmental transfer system. In fact, there is no one best system, and everything depends on the goals that are laid down. The study examined some options for the design o f an optimal transfer system for Kerala (details in Volume IV). Below i s a recommended transfer formula. From the simulations carried out there i s much to recommend a per capita consolidated grant. It i s simple and transparent and a straight population allocation recognizes needs and can be calculated with a minimum o f arbitrariness. The data is more available than for other allocation schemes. The "shock" o f transition to the consolidated per capita grant from the present system would not be so great. There would be about as many GPs that gain as would lose, and the average increase for a "winner" would be 14 percent while the average reduction for a "loser" would be 13 percent. Only one GP would gain more than 50 percent on a per capita basis, and only six would lose more than 50 percent. A kind o f `hold-harmless' system to protect such outliers could be easily devised. There are also disadvantages to the per capita consolidated grant, The one shortcoming that most will call out is the failure o f such a systemto recognize the special needs o f places that are poor, or that must serve large land areas. Moreover, there is no provision for rewarding those GPs that make a greater tax effort. 59. Strengthen State Finance Commissions. There is no question that the State Finance Commissions can make a major contribution towards strengthening intergovernmental fiscal relations. InKarnataka, however, the SFCs have not been as successful as one would have hoped. There are a number o f reasons why: an unwillingness of the state to accept the recommendations of the Commissions, the absence of staff to support the work of the SFCs, and the absence of complete and accurate fiscal data on which to base the analysis. Moreover, it i s not clear that the state government has a cell that regularly tracks the 24 performance of the local governments. While Kerala appears to have developed a system of intergovernmental fiscal relations that is more suited to decentralization, it i s not clear that they are any better off in terms of having a supporting analytical cell or a usable fiscal information system for rural local governments. The following are steps that the states should take inredressingthese problems. First, establish an analytic cell within each state that is charged with tracking the performance o f the local governments. This cell could serve as staff to the SFCs at the time of the work o f the SFCs. Second, develop and maintain a census o f local government finances, based on a uniform set of local government accounts. Third, produce an annual report on intergovernmental fiscal relations that describes and evaluateslocal government fiscal performance. E. Improve Accounts andAccountability 60. Local governments should prepare budgets and financial accounts. This implies a number o f prerequisites: staff with a capacity to do the job, certainty of revenue flow from the state level, and an accounting system that will enable a tracking of budget implementation. At the close of the fiscal year, allpanchayats should prepare financial statements ina common format that reflects standardinternational practices for local governments. 61. Ensure timely consolidation of accounts. Preparation of monthly accounts and annual accounts should emanate from consolidated receipts and payments or t i a l balances of ZPs and TPs based on the statements of all drawing officials and not from records o f the Treasury. The Treasury should function only as a banker to ZPs and TPs. The closing balance arrived at from the books of accounts of ZPs and TPs should be reconciled on a monthly basis with the figures of the Treasury. The responsibility of preparation of monthly and annual accounts of TPs should bethat of TPs and not that of ZPs. 62. Bring in mainstream accounting. The Accountant General's Office has recommended revised accounting systems and a chart of accounts for PRIs. There is needto bringinmainstream accounting by the introduction of general ledgers, sub-ledgers, and simplification of charts o f accounts, including reduction in number of heads of account. Annual accounts should have schedules o f various assets and liabilities, and income and expenditure including a list of properties and fixed assets owned by PRIs, and so on. Also there is a need for schedules for other major accounts like loans, deposits and so forth. It is preferable to have sub-ledgers or subsidiary ledgers for the suppliers and contractors accounts. This will enable the PRIs to effectively monitor and administer the individual contracts. 63. Overhaulthe accounting and the budgetingsystem of the PRIs. Serious thought may be givento introduce a modified double entry system of accounting. The accounting system and the manner o f presentation of the budget should be redesigned in such a manner that appropriate financial reporting i s made possible for various purposes, such as on-going monitoring, performance evaluation or cost o f providing individual services. For keeping up-to-date accounts and for ensuring transparency, it i s necessary to computerize the book keepingfunction ofpanchayats. 64. Redesignthe internalmanagement structure of Grampanchayats. As the president i s executing many executive functions, it i s necessarythat he shares the responsibility associated with such functions. Considering the volume o f work ingram panchayats, it seems necessary to introduce a kmd of cabinet system of functioning with the president and the chairpersons o f the standing committees forming a team. The line of control of the panchayat bureaucracy, especially that of those from the transferred departments, should be made unambiguous. 65. Improve auditing. All efforts shouldbe made to bringthe statutory audits up-to-date. The accounts of the majority of PRIs have not been finalized for years. The final accounts should be inplace within two months after the close o f the financial year. Ideally, the audit should be done within six months thereafter. The Performance Audit Authority needs to carry out its two roles: one, to act as the internal auditor of panchayats; two, to evaluate the performance of the panchayats and to provide them with 25 suitable advice for improvement. The time has come to introduce value-for-money audit for the local government institutions, so that they may be made accountable not only for observing rules and norms, butalso for results. 66. Provide positive incentives for mechanisms for downward and upward accountability. International experience shows that, to have viable local governments, and a positive impact on service delivery, people's empowerment and community development i s important. Both states have a legislation supporting this approach. For example, right to information, overseeing hnctions o f the Village Assembly (gram sabha), social audit, transparent method of beneficiary selection, social mobilization, citizen's charter etc. and implementation of these functions should be taken into consideration in evaluating the performance of the panchayats. This seems to be the right direction forward for mutually reinforcing approaches of service delivery mechanisms through local governments and community participation. F.CreateanInformationSystemon LocalFinances 67. It i s crucial to compile a census of government finances for the panchayats. Building and maintaining a reliable information system i s key to designing and implementing an efficient intergovernmental fiscal system. As this study has shown, it i s possible to develop a reasonable set of fiscal statistics that more or less portray the fiscal condition o f local governments. These efforts will have to be made regular and systematic to develop a robust data bank on rural fiscal decentralization in the state. While the effort to maintain a data base might beginwith the present system, and follow the steps taken in this study, it will be necessary in the long run to move towards a standardized set of local government financial accounts. This will require developingthe capacity, particularly at the level o f GPs, to collect and report the proper information. This is a high priority need of the states. Such an information systemwill take time to develop, but without it fiscal decentralization cannot be managed. VI. CONCLUSION 68. This study focused on Karnataka and Kerala because these states are perceived to be ahead on rural decentralization. However, even in these states, and even 10 years after the Constitutional Amendment, the agenda and implementation plan for rural local government finance i s not clear. Karnataka has burdened the district and the block levels with state personnel and earmarked the use of resources such that elected governments are treated as service agents. The tier closest to the people, the gram panchayat, has been left with hardly any resources and very little expenditure discretion. Kerala, on the other hand has arguably a better designed and more mature decentralization system, but is captive to the fiscal situation o f the state government, and the result is again under-fundingof the decentralization initiative relative to its entitlements. Other states would appear to share several of the problems outlined above, consequently it i s plausible that many o f this study's conclusions and recommendations should apply across India. 69. The fiscal decentralization system is flawed. The study concludes that despite India's commitment to rural decentralization, panchayats cannot live up to their potential because the structure of the fiscal decentralization system i s flawed. While Karnataka continues to make efforts on moving ahead on this policy agenda, what i s inplace today i s a patchwork o f legislation and measures that reflect what has been politically and bureaucratically possible. While many initiatives have tried to move the agenda forward, they were pursued in a fragmented and narrow way. InKerala the system i s better structured to achieve the goals of fiscal decentralization, particularly because it gives more expenditure autonomy to gram panchayats, but it is still flawed inimportant ways. 26 70. The State Finance Commissions, and newly formed cells within the Finance Department, could make an important contribution. Inboth states, they could lead the development o f a policy plan that spells out the goals for decentralization, the policy and legislative agenda, an implementation plan, a schedule of technical assistance for rural local governments, and a timetable. This could serve as the muchneededblueprint for fiscal decentralization. 71. Official data on finances is not reliable. I t i s difficult for the data contained in local financial records to tell a transparent story on local financial performance. First, official records o f panchayat revenues are based on budgeted allocations rather than cash received; second, the significant irregularity inthe timing of transfers and their amount distorts both actual expenditure patterns and the reporting of these expenditures; and, third, weaknesses inrecord keeping systems and practices are prevalent. Data on revenues and expenditures are publishedregularly and are used to make policy prescriptions, namely by the Central Finance Commissions and the State Finance Commissions. Butthis data i s badly flawed, and inflates the funds actually managed bypanchayats considerably. In fact, neitherpanchayats, nor SFCs, nor GoI, nor state Rural Development and Panchayat Departments know the actual amount of funds available to panchayats, or the expenditure patterns o f panchayats. Gathering the data necessary to obtain an accurate picture o f the finances of rural local governments was, o f necessity, an important objective o f this study. 72. Rural decentralization is far from complete. The rural local governments for the most part do not have adequate capacity to deliver services, collect taxes, or keep track o f their financial affairs. The state governments are not yet up to implementing the intergovernmental system, or to a full tracking and monitoring of the financial performance o f their local governments. The central government still plays a role in sub-state fiscal relations, and this both helps and confounds the problems. On top of this, the overall government deficit limits how much can be done by way o f devolving resources. All of these factors suggests that it i s a time for planning the development of the finances of the rural local sector, and that there i s a premium on capacity building at the state and the local levels. The progress in fiscal decentralization to the ruralgovernment sector can produce quite remarkable results. 73. In fact, fiscal decentralizationis a process that needs to be followed in a programmatic manner. There are some guidelines that might be used to structure such a program, as indicated by the following sequence ofreform options. 0 Develop a fiscal decentralization planthat includes all of the necessarycomponents, and i s basedon the objectives which the government has set. 0 Begin investinginupgradingthe capacity of the state government to lead andmonitor the reformprogram and inthe capacity of the local governments to implement the refom. 0 Clarify expenditure assignments 0 Improve local government revenue mobilization through structural reform and improved implementation. 0 Restructure the transfer system to realize equalizationobjectives, and to allow local governments more autonomy and discretion inthe use of funds according to their own priority setting and planning mechanism, jointly with improving local government capacity. 0 Fix the financial reporting systemo f local governments. 27 REFERENCES Anand,.S and Kothari, S. 1999. Minor Forest Produce and Community Control over Natural Resources " under Adivasi Self-Rule", UnpublishedPaper, New Delhi:Lokayan. Bahl, Roy W. 1983. "The Assignment of Local Government Revenues inDeveloping Countries," in Tax Assignment in Federal Countries, edited by Charles E. McLure, Jr. Canberra: Center for Research on Federal Financial Relations, Australian National University. Bahl, Roy W. 1994. "Revenues and Revenue Assignment: Intergovernmental Fiscal Relations in the Russian Federation," in Russia and the Challenges of Fiscal Federalism, edited by Christine Wallich. World BankRegional and Sectoral Studies. Bahl, Roy W. 2002. " Implementation Rules for Fiscal Decentralization" in M. Govinda Rao (ed), Development, Poverty and Fiscal Policy: Decentralisation of Institutions, New Delhi: Oxford University Press. Bahl, Roy W., and Linn, Johannes. 1992. Urban Public Finance in Developing Countries. New York: OxfordUniversity Press. Bardhan, Pranab and Mookherjee, Dilip. 1990. " Capture and Governance at Local and National Levels ", American EconomicReview Vo1.90, No.2. Bird, Richard and Gendron, Pierre-Pascal. 1998. "Dual VATSand Cross-Border Trade: Two Problems, One Solution?" University of Toronto, InternationalCenter for Tax Studies. Bird, RichardM.,and Vaillancourt, Francois, eds. 1998,Fiscal Decentralization in Developing Countries. Cambridge: Cambridge UniversityPress. Bird, Richard M., Ebel, Robert D. and Wallich, Christine Ieds. 1995. Decentralization of the Socialist State. A Regional and Sectoral Study. Washington, D.C.: World Bank. Boadway, Robin and Flatters, Frank. 1982. Equalization in Federal State: An Economic Analysis, Economic Council of Canada, Ottawa: Canadian Government PublishingHouse. Breton, Albert. 1995. Competitive Governments, Cambridge University Press, Cambridge, Delivery. Urban Management Program Series 16. Washington, D.C.: World Bank. 28 Dillinger, William. 1991. "Property Tax Reform: Guidelines and Recommendations." Working Paper, Urban ManagementProgram. Washington, D.C.: World Bank. Government o f India. 2001. "Report of the Task Force on Devolution of Powers and Functions upon Panchayat Raj Institutions", New Delhi: Ministryof Rural Development. Inter-American Development Bank. 1997.Latin America After A Decade of Reforms: Economic and Social Progress. Distributedbythe Johns Hopkins University Press for the Inter-American Development Bank. Litvack, Jennie, Ahmad, Junaid and Bird, Richard.1998. Rethinking Decentralization in Developing Countries.The World Bank Sector Studies. Washington D.C.: World Bank. Mahipal. 1997. Panchayat Raj inIndia: Issues and Challenges", Kurukshetra, August. " Mander, Hand Naik, V. 1999, Gram Sabhas and the Delivery of Justice: Some Issues ", " Paper presented at the UNDP conference, July 1999, New Delhi. Martinez-Vazquez, Jorge, and McNab, Robert. 1997. "Fiscal Decentralization, Economic Growth, and Democratic Governance." Paper prepared for USAID Conference on Economic Growth and Democratic Governance, Washington, D.C., October 9-10, Martinez-Vazquez, Jorge, Guess, G. and Loehr, B. 1997. "Fiscal Decentralization: A Methodology for Case Studies." CAER ConsultingAssistance on Economic Reform 11:Discussion Papers, no. 3, May. Mathur, O.P. 1999. Decentralization in India: A Report Card, National Institute of Public Finance and Policy, New Delhi,March. Musgrave, Richard A. 1983. " Who Should Tax, Where, and What?" in Tax Assignment in Federal Countries. Editedby Charles McLure, Jr. Canberra: Center for Research on Federal Financial Relations, Australian National University. Oates, Wallace, E. 1968. " The Theory of Public Finance in a Federal System", Canadian Journal of Economics, Vol. 1, (2). Oates, Wallace, E. 1972. Fiscal Federalism, Harcourt, Brace and Jovanovich, New York. Oates, Wallace, E.1999. ``An Essay on Fiscal Federalism", Journal of Economic Literature, Vol.XXXVII, September. 29 Olson, Mancur. 1967. " The Principle o f Fiscal Equivalence: The Division of Responsibilities Among Different Levels of Government", American EconomicReview,Vol. 59 (5). Parker, A.N. 1995. " Decentralization: The Way Forward for Rural Development", Agricultural and Natural ResourcesDepartment, The World Bank. Peterson, George E. 1997. "Building Local Credit Systems." UrbanManagement ProgramPolicy Paper. Washington, D.C.:World Bank. Pillai G. Karunakaran. 1986. Local Finance in a Developing Economy, Delhi: B R Publishing Corporation. Pillai, Velu, T.K. 1940. The Travancore State Manual, Trivandrum: The Government of Travancore. Prud'homme, Remy. 1995. "On the Dangers o f Decentralization", in The WorldBankResearch Observer 10 (2). Pushparajan, K. 2002. "Panchayat Administration: Machinery and Strategy", Trivandrum: Thunchan Smaraka Samithy. Rai, Manoj, et a1 (eds). 2001. The state of Panchayats: A Participatory Perspective, PRIA, New Delhi: Samskriti. Rajaraman, Indira. 2003. The Fiscal Domain for Panchayats, New Delhi: OxfordUniversityPress. Rani, P.G. 1999. " State Finance Commissions and Rural Local Bodies: Devolution of Resources", Economic andPolitical Weekly, June 19. Reports o f the First State Finance Commissions of Assam, Andhra Pradesh, Madhya Pradesh, Himachal Pradesh, Karnataka, Kerala, Maharashtra, Orissa, Punjab, Rajasthan, Tripura, Uttar Pradesh and West Bengal. Santha, E.K. 1993. Local Self-Government in Malabar (1800-1960), New Delhi: ISS Occasional Paper, No. 12. Spahn, P. Bernd. 1997. "Decentralized Government and Macroeconomic Control." paper prepared for the International Institute of Public Finance 53 rdCongress, Kyoto, Japan. 30 Srivastava, K.B. 1999.Panchayats inScheduled Areas: An Analysis of Provisions of the Panchayats (Extension to the Scheduled Areas) Act 1996(Central Act 40) and Extent o f its Adaptation by the States having FifthScheduled Area, Hyderabad: National Institute of Rural Development Tanzi, Vito. 2001. " Pitfalls on the Road to Fiscal Decentralization", Working Paper No. 19, Camegie Endowment for International Peace. Tanzi, Vito. 1996. "Fiscal Federalism and Decentralization: A Review of Some Efficiency and Macroeconomic Aspects", inAnnual World Bank Conference on Development Economics 1995. Edited byMichael Bruno andBoris Pleskovic. Washington, D.C.: WorldBank. Ter-Minassian, Teresa, ed. 1997. Fiscal Federalism in Theory and Practice. Washington, D.C.: InternationalMonetary Fund. Tiebout. 1956. " A Pure Theory of LocalExpenditures", Journal of Political Economy, Vol. 64, October. Vyasulu, Vinod. 1998. DecentralizationFromAbove: Panchayat Raj in 1990s. Wong, Christine. 1995. "Overview o f Issues inLocal Public Finance inthe PRC" inFinancing Local Government in the Peoples Republic of China,Edited.by Christine Wong, Chapter One. Unpublished manuscript.