68367 Strengthening local governance Defining the roles of Upazila and Union Parishads SUMMARY The current political transition in Bangladesh offers a unique opportunity to strengthen the country’s system of local governance – i.e. creating local governments that are accountable to their citizens. This note focuses on the rural sector. It takes as its starting point the premise that the Upazila and the Union Parishads can jointly form the backbone of a system of rural local governments in Bangladesh. The key interventions required are to strengthen the Upazila as an independent tier of government – especially relative to national parliament – and to clarify the relationship between the Upazilas and Union Parishads so that the two can jointly work as one system of local government. In this context, holding elections at Upazila level will be an important step to take, but this action alone will not result in a sustainable or effective system of local governance. Upazilas require clarity on their functions and their relations to Union Parishads, adequate untied resources for which they can be held accountable, and appropriate checks and balances on their powers. Why focus on Upazilas and Union Parishads as the core of a system of rural local governance? International experience suggests that a proliferation of tiers of government is both costly from an administrative point of view and often leads to confusion over roles and accountability in terms of who is responsible for what functions. Upazilas and Union Parishads provide both the proximity to citizens – government close to the people—and offer the appropriate size for scale efficiencies to be captured in the delivery of rural services. Strengthening a third tier above the Upazila, such as the districts, may therefore not be needed. History also offers an important guide. Upazila and Union Parishads have acquired legitimacy in the eyes of citizens and are even accepted by the diverse political parties in the country. Finally, Bangladesh is urbanizing very fast. In addition to Upazila and Union Parishads on the rural side, it is urban municipalities and corporations rather than district governments that will need to be strengthened. What is needed to strengthen the Upazila and Union Parishad system? A minimum program to strengthen local governance requires: (a) basic reforms to the legal framework for local governance to clarify powers and functions and to institute an effective system of oversight over Upazila performance; (b) holding of elections at Upazila level; and (c) allocating adequate discretionary resources to Upazilas and Union Parishads. – both as fiscal transfers from national government and from own taxes. In other words, the holding of local elections should be anchored in a legal and fiscal system that strengthens local governments. One without the other will not be effective in creating a strong rural local government system. Such a system, however, is on its way to being created. The Local Governance Support Program (LGSP) introduced through the PRSP and currently being implemented by the Local Government Division (LGD) is strengthening Union Parishads. It offers an approach that has important lessons for strengthening local governments. Firstly, it already provides increased discretionary resources to UPs and encourages the raising of own-taxes, while strengthening systems by which UPs are held accountable to communities. Secondly, it introduces important changes to the system of intergovernmental relations in Bangladesh, through its emphasis on financial reporting, universal auditing and direct disbursement of funds. Importantly, this program was supported by civil society leaders and endorsed by both leading parties. It has received financial support from the World Bank, UNDP, EU, SDC and others. What is missing is a parallel initiative to strengthen Upazilas as a tier of government. Currently, the Upazila tier does not have an elected Chair; its functions are not clear and its Strengthening local governance in Bangladesh Page 1 source of revenue highly centralized. It is in effect a delegated tier of national government and often influenced by the politics of Parliament and individual parliamentarians. In the context of direct elections for Upazila Chairs, a program similar to the LGSP can be crafted and funded for the Upazila tier to establish an effective Upazila and Union system of rural local government. This note provides greater detail of what an Upazila program could look like. Taking these steps will begin the process of strengthening sub-national governance in Bangladesh. Further reforms will be necessary over time. The interim government will also need to outline the process through which the reform program should be developed and sustained over time as an elected national government comes back to the helm of the country’s political affairs. A final word on urban local governments. The strengthening of urban Pourashavas and City Corporations would follow the same principles outlined in this note for Upazilas. The capacity of the urban local governments suggest that the reform process may be easier to formulate and implement. But the politics of urban local government reform maybe more complex. For this reason, the interim government may wish to discuss a sequencing process whereby the elections of Upazila Chairs and the strengthening of Upazilas may well precede the reform of urban local governments. But, the time may well be opportune to launch an overall national debate and discussion about local governments in general covering both the rural and urban areas. . Strengthening local governance in Bangladesh Page 2 A. Introduction 1. There is a broad social consensus for strengthening local governance in Bangladesh. This is largely expressed as support for strengthening the role of Upazila and Union Parishads, through elections and greater political decentralisation. However, there is a disjuncture between the desire for political decentralisation and the extent to which the current design of the local government system provides for the decentralisation of real power, functions and resources to local governments, and allows meaningful citizen oversight of local governments and services that are being delivered. Thus, while representatives may be elected at local government level, they do not have the de facto authority or resources available to successfully undertake the activities that their constituents have elected them to perform. This mismatch undermines the extent to which citizens are really able to hold local government representatives to account for their performance, and also limits the capacity and ability of elected representatives to oversee activities that are being performed by other (non-elected) agencies at the local level. 2. A long standing manifestation of this debate in Bangladesh has been on the role of the Upazila (see below). Most recently, discussion has focused on holding elections for the position of the Chairman of Upazila Parishads. This post is currently vacant, and while indirectly elected members are drawn from Union Parishads (UP) and Pourashava’s, the de facto current position is that Upazila’s are operating as purely administrative unit of central government rather than as an autonomous and accountable tier of government within the local governance system. This is because only the elected Chairman has executive authority at Upazila level, with the other Parishad members playing an advisory role. In the absence of an elected chairman the executive function is performed by UNOs – the administrative representative of central government. 3. Holding elections for the position of Upazila Chairmen would certainly advance the extent to which political authority is democratically decentralised in Bangladesh. This brings with it many advantages, including (a) ensuring a greater degree of responsiveness of Upazila administrations to the needs of the people of an area (b) counterbalancing the power of national MPs, and (c) mitigating the effect of a “winner takes all� system at the national level through allowing other political parties access to the levers of governance (there is no bar in the ordinance to Chairman being elected on party platforms, though this may exist elsewhere in legislation),. 4. However, expanding the extent of political decentralisation (as provided for in the Upazila Ordinance) brings with it some challenges. Firstly, merely holding elections does not automatically expand the actual executive authority of those who are elected, as this is prescribed in legislation outlining the powers, functions and resources available to them for the exercise of their authority. Without sufficient legal authority they will be unable to respond to citizen needs, and citizens will not be able to use their votes to influence the activities of service providers that operate at the local level (but who report to the central government). Secondly, increasing the number of elected representatives at the local level (which includes UPs and Pourashava’s as well) can lead to confusion over their respective roles and functions, unless this is clearly spelt out. For example, elected UPs have a long history in Bangladesh, which has brought with it broad acceptance, and increasingly improved leadership. Recent government programs which have transferred fiscal resources and physical assets to them have also conferred on them the image of a tier of governance that is close to the people. Thirdly, providing Upazila Chairmen with executive authority requires that a strong system of checks and balances is introduced both within and outside of the Upazila to mitigate the risk of any Strengthening local governance in Bangladesh Page 3 abuses of this authority. The central government, Parishad members and citizens all have a role to play in providing check and balances on the Chairman, and it is important that the actions of these stakeholders complement one another. 5. Both local and international experience emphasizes that just the act of holding Upazila elections, while an important step, is not enough to strengthen local governance in Bangladesh. Indeed, it may have the opposite result if careful thought is not given to the political system adopted, and the relative assignment of political, administrative and fiscal powers between Upazila’s and Union Parishads. Thus, carefully reviewing the overall structure and technical design of the local government system is a pre-requisite for ensuring that steps towards greater political decentralisation become entrenched over time. 6. To do so requires that policy makers give careful attention to: a) The allocation of powers and functions between central and local government b) The respective roles of the different tiers of local governments especially in relations to the delivery of local services (such as Upazilas and Union Parishads) c) The size and nature of the resources allocated to local governments through own revenues and the national budget d) The role of elected local representatives relative to member of parliaments (MPs), local officials, such as the UNO’s and heads of line agencies at Upazila level e) The system of checks and balances by which elected representatives are held accountable, both by communities and higher tiers of government 7. These are some of the key policy issues that need to be addressed to ensure the political decentralisation has any meaningful or sustainable impact. B. The evolution of Upazila’s in Bangladesh 8. Upazilas were created through the promulgation of the Local Government (Upazila Parishad and Upazila Administration) Ordinance on November 7, 1982, during the period Lt General Ershad was in power. In essence the ordinance reconstituted Thana’s and upgraded Thanas as Upazilas. Each Upazila was presided over by a Parishad, consisting of indirectly elected and appointed members, and a directly elected Chairman. The Chairman would hold the executive authority of the Upazila, while the Parishad members held regulatory and oversight roles. 9. Upazila Parishads were formed in 1982, but elections for the post of Chairman were not held. The chief official in each Upazila by default therefore assumed executive authority, alongside administrative authority. During this period there was some de-concentration of service delivery functions to Upazilas that resulted in the growth of staff at this level. The first elections for Upazila Chairman were held in 1985, and were perceived by many as part of a strategy by the military government to build a political base at local government level. Considerable resources were channeled to Upazilas in this period, but without the parallel development of robust systems of financial, administrative and political accountability for their use. This created a culture of impunity in some Upazilas, where the Chairman and senior staff literally “ran the show� without fear of oversight. Most disturbingly, Upazila’s started to encroach on the powers and resources of Union Parishads and on each other (and to some extent with the Zilla tier), as functional and geographic boundaries were not clearly demarcated. This began to significantly weaken the system of local governance as a whole. Thus, in 1991, the first elected national government after military rule abolished the Upazila Parishads, citing their misuse of funds and assets. This move appeared to have bi-partisan support from MPs at the time, who felt it encroached on their role and functions within their constituencies Strengthening local governance in Bangladesh Page 4 10. However, surveys done in the early 1990’s indicate that around 75% of citizens supported the elected Upazila system, as it brought services and opportunities for redress of grievances closer to communities. The national government thus gradually began the process of returning functions, resources and staff to the Upazila level, which strengthened a popular demand for Upazila elections. By some counts around 80% of those surveyed supported Upazila elections. Most importantly, local people started believing in their own abilities to make things happen. 11. In 1991, with the abolition of the Upazilas, the government established a Local Government Commission. Its 1997 report stated that the existing Upazila system (as introduced in 1982) had an adequate description of their functions and adequate provision human and financial resources. In 1998, a revised Upazila Act was published, which developed slightly more robust (though still very rudimentary) rules and procedures and demarcated boundaries for election. It retained the provision for the direct election of the Upazila Chairman, with all UP chairpersons as members and 3 women from UP as members as well. 12. However, in April 2001, a new Government Order (GO) was issued. The GO states that if, for any unforeseen reason, Upazila elections do not take place in some areas, the Government (instead of Election Commission) will declare the election date within the stipulated time and hold elections. This has fundamentally altered the operation of the Act, and introduced perceptions that this provision may be used to avoid elections in areas where an opposition party may win or in marginal constituencies where MPs may feel their electoral base will be undermined. This GO is now subject to public interest litigation that is ongoing. C. Powers and functions of Upazila’s 13. The current Upazila Ordinance creates a “strong chairman� model of political governance at the Upazila level, through conferring executive authority on the elected Chairman, and giving him / her control over the administration. This system has led to conflict with national MPs, who may feel threatened by the greater resources and authority available to Upazila Chairmen in their constituencies. The bulk of the Ordinance, however, is devoted to regulating election procedures and activities of the Chairman, rather than the Upazila Parishad or administration. Although the Chairman is accountable to the Parishad, this does not include the direct ability to dismiss him / her, as this is assigned to the central government. 14. Expenditure authority is vested in the elected Chairman, subject to the approval (and amendment) of an annual budget by the members of the Parishad. However, considerable powers exist for central government to both impose or amend a budget, or require certain expenditures to be charged to the fund. In addition, central government retains extensive discretion to formally intervene in the affairs of the Upazila, and quash decisions of the Parishad. 15. The Act provides for only a very weak model of accountability for Upazilas. There are no specific provisions requiring consultation with citizens, and only limited procedures requiring annual financial statements to be made available on request at Upazila offices. Although financial statement require auditing, the regularity of this, or consequences of audit findings, are not laid out in the Act (and are thus subject to GOs) 16. The functions assigned to Upazila’s are described in Schedule 2 of the Ordinance, and include: a) Undertaking the planning and execution of “all� development activities at Upazila level, specifically including the planning and execution of all rural public works programmes, and their monitoring and evaluation. The ordinance does not define which “development activities� are being referred to, so this could include all or some of typical local service Strengthening local governance in Bangladesh Page 5 delivery functions such as water, sanitation, roads, etc. It does also not specify whether development activities are related to asset creation only, or includes the delivery of associated services and the maintenance of the assets. b) Promoting health, welfare, cultural, educational, agricultural, livestock, fishing, forestry and employment generating activities and the co-operative movement. These provisions appear to limit the ability of Upazilas to actually undertake the delivery of any services associated with these functions c) Implementing government policies and programmes, with provisions for the further transfer of functions at the initiative of the central government d) Provision for management of the environment, although it is not specified what this might mean. It is not clear, for example, if this refers to environmental health regulation. e) Supervising Upazila officers, except Magistrates, Munsifs and regulatory staff. Upazilas are permitted to hire staff necessary for the performance of their functions without recourse to higher tiers f) Training, assisting and coordinating the activities of Union Parishads, and preparing their development plans based in UP plans g) Assisting Zilla Parishads in their development activities Unfortunately, the schedule provides only a broad definition of the actual functional assignments to Upazila’s, which can easily be read to include all those functions assigned to Union Parishads. It does not therefore remove uncertainties in the relative division of powers and functions between Upazila and UPs. 17. The local source revenues of Upazila’s are defined in Schedule 3 of the Ordinance and include leases for hats, bazaars and ferries (at a rate to be determined by government), a street lighting tax, the ability to levy service charges, and a set of licence fees that are likely to be largely of nuisance value. Importantly, Upazila’s may not raise a levy that is similar to that used by a lower tier of government (the UP or Pourashava). The bulk of revenues of Upazila’s are provides through transfers from the central government. These are made on a largely conditional basis, although a small Upazila Block Grant has been more recently introduced. 18. In sum, the history of Upazila’s have generated several important lessons for the further evolution of the local governance system: a) The strengthening of Upazilas can occur at the expenses of lower tiers of local governments, such as UPs, despite these lower tiers being well established and understood, and close to the people. This is due to the ambiguity of the definition of functions of each tier b) The extent of authority vested in the Upazila Chairman has not been counter-balanced by adequate systems of formal oversight and regulation. This has led to instances of improper use of authority, that has created a backlash against the Upazila system c) Very limited forms of citizen oversight have prevented the development of downward accountability in the Upazila system to the extent that is found at lower tiers d) Very limited revenue sources have been provided to Upazilas, some of which probably would have negative economic effects if their rates were raised to meaningful levels. The poor selection of revenue sources has undermined the credibility of the local tax system, which is largely regarded as under-performing e) Limited local revenue has left Upazilas reliant on central government transfers which are generally earmarked for specific purposes, and some of which are transferred as assets rather than cash. This considerably limits the discretion of Upazilas, and negatively impacts on the deepening of their accountability relationship with communities as they must respond to the demands of central government rather than community needs. These problems all highlight the importance of the need for an appropriate definition and alignment of political, functional and fiscal assignments in the local government system. Strengthening local governance in Bangladesh Page 6 D. Some comparative experience 19. Many other countries have faced similar problems, and embarked on lengthy programmes to reform their local government systems. Given the similarity of the problems (despite the difference in contexts) it is useful to reflect on these experiences for Bangladesh. The examples are not meant to provide a comprehensive assessment but rather narrowly focus on key challenges and mismatches that are being faced in the intergovernmental systems of different countries that have relevance for the design of upazila and unions: a) South Africa is a unitary state that has a “wall-to-wall� system of LG. Outside of the big cities theire is a two-tier system of LG, with about 8 local municipalities making up a district municipality. While all local municipalities have directly elected representatives, only about 40% of district representatives are directly elected with the remainder being delegated from local municipalities. Provincial governments are assigned health, education and welfare functions, leaving LG with basic local functions such as water, sanitation, electricity distribution, local transportation and land use planning. There has been considerable confusion over functional assignments between district and local municipalities, which has caused significant disruption to the service delivery system. In some provinces, a repeated and confusing set of functional transfers has resulted in the largely indirectly elected district tier being assigned the bulk of basic local service delivery functions, but without sufficient own revenue sources or direct de facto political accountability for its performance. The process of transferring functions has also disrupted service delivery. The misalignment of political, administrative and fiscal powers between these tiers continues to pose a major challenge to the credibility of the local government system. b) Pakistan is a federal state with strong provincial governments. The big bang process of decentralisation has been driven from the top down, with the central government seeking to weaken provincial control. A two tier system of district and local (tehsil) governments has been established, drawing on the South African model. While good progress has been made on some aspects of functional and fiscal assignments, there are concerns that the political system has not yet rooted itself in strong forms of community-based accountability – as reflected in low levels of citizen support for the local government system. c) Vietnam’s system of sub-national government includes 3 tiers – provinces (64), districts (some 600) and communes (about 10,000). At all three levels, directly elected People’s Councils operate as “legislatures�, and People’s Committees (some of the members of which are from the Councils) act as “executives’. Democratic governance is obviously constrained by the one party nature of the state – but there is nonetheless considerable electoral competition amongst Party members at the local levels. Vietnam is one of the most highly decentralized countries in Asia. Approximately 65% of public expenditure is managed at sub-national levels – with real discretionary powers being exercised by provinces. However, the recent State Budget Law (which came into effect in 2004) only specifies provincial expenditure responsibilities – and leaves it to individual provinces to then assign expenditure responsibilities and revenues to districts and communes. This was intended to allow for variation between different types of province. In practice, what has happened is that provincial authorities (in both the Provincial People’s Councils and Committees) have retained most expenditure responsibilities at the provincial level – and have devolved only very limited responsibilities (and revenues) to districts and communes. This is clearly a problem in that it is at these lower levels that downward accountability (however Strengthening local governance in Bangladesh Page 7 constrained by Vietnam’s one party system) is greatest. As a result, local needs are inadequately captured in sub-national budgets – which tend to be dominated by provincial “needs� and which are often skewed towards large infrastructure investments, rather than towards local infrastructure and service delivery needs. The failure to adequately define LG powers, functions and resources has allowed province s to dominate even local service delivery functions. d) In India, a similar lack of clarity on the role of LG has been compounded by the formation of a large number of tiers in the local government (panchayat) system, including District, Block, Gram and Gram Sabha. State governments have resisted any significant devolution of power or functions, which has left elected LG without any meaningful degree of discretion over its activities. The result has been a disempowered LG system where the very lowest tier, the Gram Sabha, which is intended to be the forum for community consultation and oversight, is unable to perform its role. e) Indonesia is a unitary state, like Bangladesh, with both provinces and local governments. In 1999, two laws were passed that devolved certain sectors (rather than specific functions) to LG, and assigned the resources to LGs to fund these sectors. This broad brush approach to functional assignments created ambiguity and confusion over the role of provinces, central agencies and LGs. Although the responsibility of provinces is to coordinate, evaluate, accredit and standardize, they continue to prepare and implement programs - often large scale programs that distort LG efforts and sent out wrong incentive signals. As a result, central governments and LGs view provincial authorities as competitors rather than partners. Although there were attempts to subsequently clarify these issues, through laws on cooperation, GOs and regulations, the problem has remained. For example, even today, there is no clarity over which level of government has the authority to establish, accredit, or close a public health center. Given this situation, the budget process, monitoring, and reporting arrangements are also not aligned. Line ministries continue to play a key role in implementing programs, alongside provinces and districts. There is substantial confusion, lack of coordination, and mistrust among the three tiers of government. For example, senior local government officials in the district offices are often unable to understand what is happening in their sector, who is doing what, and when actions are being taken. The Indonesian experience thus underlines, for Bangladesh, the need to resolve the role of local governments within the broader intergovernmental system, where the system itself is greater than the sum of its parts. 20. These international experiences emphasize the importance of a carefully structured and sequenced program of political, administrative and fiscal decentralisation – regardless of the pace at which reforms are intended to proceed. Simply establishing an elected LG system does not necessarily result in improved governance or service delivery. The functions of LG, and of each tier of LG, need to be carefully defined, and adequate resources explicitly assigned to them. A failure to do so will undermine the emergence of an accountable tier of governance, and will continue to create problems with LG performance until they are resolved. Indeed, these experiences demonstrate that a failure to adequately assign powers and functions can actually undermine the sustainability and ongoing development of the LG system as this issue will continue tot dominate the LG agenda going forward. In other words, simply holding Upazila elections will not in itself strengthen LG, and may actually serve to weaken on undermine important aspects of the emerging LG system. Strengthening local governance in Bangladesh Page 8 E. Principles guiding political, functional and fiscal assignments 21. International experience suggest that there are some key principles that need to guide the basic design of the emerging system of local governance in Bangladesh: i) Subsidiarity, or ensuring that functions are originally assigned to the tier of government that is closest to citizens. It is important to note that there are few functions that will ever be solely performed by LG, as central government will always retain a monitoring role to ensure that functions are being exercised efficiently and accountably, and other tiers of LG can play support roles in the performance of many functions. This principle thus needs to inform a detailed process of mapping functional assignments, both through the identification of functions that the LG tier can perform, and the allocation of the various activities within a function between Upazila, UP and central government (including ministries and line agencies). ii) Accountability, as expressed through an adequate system of formal checks and balances, mechanisms of citizen participation and oversight, and sufficient own revenues to provide the ability for LGs to respond to citizen preferences. Strengthening local governance requires the central government to reduce its control of the activities of other tiers (“upwards accountability�, while taking measures to ensure that the is replaced by adequate local control (“downward accountability�). It is important to note that central government will never withdraw from its responsibility to ensure that LGs remain accountable, but rather that it needs to reorient its controls to support and complement greater downward accountability. iii) Efficiency incentives, as expressed through both capturing economies and scale and scope, and designing institutions that encourage universalising citizen access to service delivery and governance systems. Although greater subsidarity and accountability can improve allocative efficiency at LG level (through allowing citizen preferences to be reflected in local decisions), they do not guarantee that LGs will be able to perform their functions in an efficient manner. In particular, LGs need incentives to exploit economies of scale and scope in the performance of functions, while good governance practices and basic national development standards need to be pursued by all local bodies. Greater efficiencies can be sought both in the assignment of functions to different tiers of LGs (which must be balanced against the principle of subsidiarity) and through assigning own revenue sources and transfers in a manner that improves citizen oversight of performance and rewards efforts to improve performance. iv) Managing the transition, through providing specific fiscal and administrative support measures – a role in which the centre is of paramount importance. The strengthening of sub- national governance is a long term agenda, which will face many obstacles. Continually increasing levels of support for local governance are important for the long term success of reforms. However, all reform programs of this nature are difficult, sometimes leading to situations where internal restructuring leads to temporary declines in the quality of services or governance. Central government can play an important role in managing these “shocks�, though financing the transition in a manner that reduces these impacts on citizens. 22. More operationally, in refining the legislative environment for local government policy makers need to ensure that the system is: f) Comprehensive, in dealing with the relative allocation of powers and functions across tiers of government and within the local government system. Strengthening local governance in Bangladesh Page 9 g) Sequenced, to ensure that proposed reforms complement one another without overloading the capacity of local governments, and are adequately financed through own revenue and central government transfers h) Transparent, to reduce uncertainty or conflict over respective roles, to prevent excessive interference by other tiers in the activities of local governments, and to stimulate community engagement with local governments. In particular, this requires shifting the focus of oversight from regulating the inputs (through complex rules and earmarked transfers) to managing outputs and outcomes of local governments. i) Consistent, in allocating political and legal authority over public functions to democratically elected institutions. F. Key questions for strengthening local governance in Bangladesh  What functions should be assigned to local government? 23. The current legal definition of functions assigned to local governments is vague and non- specific. When this is read in the context of the legal principle of “ultra vires� (which has not been specifically overturned in respect of LG in Bangladesh and which requires specific authorisation for government to perform any function) the legal scope for the discretionary performance of most functions at LG level appears limited. 24. In theory, functional assignments can be mapped out between tiers of government based on the principle of subsidiarity (devolution of each function to the lowest possible level). In reality, this mapping is constrained by: a) The concurrent role that tiers of government play in most functions. For example, higher tiers will always be required to play a role in the establishment and monitoring of national standards in the specific functions, such as in water services. b) The extent to which certain functions contain strong redistributive elements that are best funded off the national tax base. For example funding safety nets should remain a national function, but its administration can be well suited for performance by the lowest tier of government. c) Perceived capacity constraints at LG level. It is important to emphasize that no tier of government will develop the capacity to perform functions that are not already assigned to it, so that this argument will always apply in centralised environments such as Bangladesh and does not constitute a sufficient reason not to assign a function. However, it does imply that the decentralisation of functions does need to be sequenced appropriately, to avoid overloading LGs through requiring the development of multiple capacities simultaneously. d) Multiple tiers of local government that may exist, which requires a broader evaluation of the necessity and role of each tier in each function 25. Two approaches can be suggested that should guide the discussion of functional allocations in Bangladesh: a) The identification of the broad functional areas that may be most appropriately performed by LGs, with a focus on the provision of basic local public goods; b) The development of an activity map that identifies:  Whether each function is appropriate for an original allocation to LG (decentralisation) or for administrative delegation; and Strengthening local governance in Bangladesh Page 10  The extent of concurrency involved through identifying each activity associated with the performance of a function (for e.g. planning, financing, asset creation, management / project execution, monitoring, evaluation)  To which tier of local government should functions be assigned? 26. The structure of the local government system should be guided by the functions that it will perform. There is little sense in having a broadly drawn LG jurisdictions or a large number of tiers of LG if they are only assigned responsibility for very local services. At present Bangladesh has 4 sub-national tiers: Zilla, Upazila, Union and Gram Sarkar, which may be more than necessary for the current set of functional assignments (and may be in part why they largely act as agents of the central government). The cost and administrative overheads of this model need to be continuously weighed against the benefits it brings in accountability and efficiency. 27. While in theory the assignment of functions between tiers within the LG system (Zilla, Upazila, Union) will need to balance the principles of subsidiarity and efficiency, in practice this will need to complement other important choices in the design of the system. In particular, the model of political representation is important in determining the outcome here, as assigning original functions to indirectly elected tiers of government will undermine accountability.  How should local governments be financed? 28. The financing of the local government system should, in theory, reflect the choice on the functions that they will perform (“finance follows function�). In reality, it becomes increasingly complex to assign revenue sources (whether local taxes, user charges or national transfers) as the number of tiers in the system increases. This is because: i) There are few truly productive and non-distorting local tax sources that are available for LGs, which is a severe problem across the world. The primary “good� local tax is that on immovable property. These few tax bases thus need to be shared across tiers of LG, which is administratively complex and can often undermine incentives for effective revenue collection. ii) There is a limit to the extent to which national transfers can be divided up between tiers without introducing considerable uncertainty for them, and increasing the extent to which these transfers are earmarked for specific purposes (thus undermining their discretion). Given high levels of poverty across Bangladesh, a key role for transfers will be to ensure that all citizens have access to a basic package of local services. iii) Most LGs also rely on user charges for services that they deliver. This has a very positive effect of ensuring that the services are truly demanded by citizens, who may otherwise refuse to pay. However, under-funding of LGs can result in surcharges being added to user charges, which can be both anti-poor and economically damaging. Another common set of charges are licence fees, which in the most part have little more than a nuisance value and whose administrative costs can outweigh their revenue gains.  How can we strengthen the systems for citizens hold their local governments to account? 29. Simply devolving functions and finance to local governments does not guarantee an improvement in the quality or coverage of services. The intention of strengthening LGs is to Strengthening local governance in Bangladesh Page 11 bring decision-making closer to people, so that choices made by LGs more closely reflect the preference of a specific community. For this to work requires measures that strengthen the ability of citizens to participate in decision-making, and oversee the performance of local governments. There are four issues that need to be addressed here: i) Information: the quality, timeliness and scope of information on the performance of LGs need to be improved. Citizens themselves require this information, so that they can make judgements on the performance of their elected LGs. At present there are few requirements for regular public disclosure of important information, as the emphasis of GoB monitoring is on LGs reporting to the central government on progress with expenditures on specific programs. Restructuring the reporting system to support citizen oversight of LGs will actually strengthen GoB oversight of LGs. This requires the introduction of disclosure requirements (some of which LGs already use informally), and reorienting the reporting system to focus on LGs outputs and outcomes, rather than the inputs provided by central government. ii) Participation: few structured opportunities exist for citizens to participate in planning, decision-making and performance review at LG level. This results in local elites often dominating the decision-making process, through informal means of participation. Government policy and legislation can play an important role in requiring consultation at ward and LG levels, through requiring consultative development planning processes, open budget hearings and meetings, and consultative meetings to review information disclosed by LGs (such as annual audit reports). iii) Oversight: a system of checks and balance is a key ingredient of all democratic systems. At the local level this means that a balance must be struck between the power of the executive (e.g. the Upazila Chairman) and that of the local “legislature� (the Parishad members). At present, considerable executive authority is given the Chairmen, without a well developed framework for Parishad oversight. There are various ways to strengthen the oversight powers of the Parishad, such as only indirectly electing Chairman, or increasing the scope of authority of standing committees in the Parishad, and its legislative authority as a whole. iv) Monitoring and Intervention by central government: the role of the central government also needs to be reoriented so that it complements, rather than replaces, the oversight of communities. At present, the extensive system of input controls, reporting requirements and intervention procedures imposed by central government serve to replace community oversight. Central government can reorient its role through providing greater discretion on the use of resources, while strengthening its capacity to monitor and evaluate LG performance. This requires a shift from a system of input controls to one of output monitoring and outcome evaluation, in line with international best practice. Central monitoring tools and basic public financial management standards (such as accounting practices and procurement rules) can be used to improve the availability of information for communities to assess LG performance, while independent formal auditing procedures can provide necessary oversight of LG compliance and performance for both central government and communities. Intervention procedures, while important in severe cases of institutional failure, need to be used solely to enforce the accountability of local government, and protected from any political interference. G. Options for Bangladesh 30. Options exist for how to proceed with political, administrative (functional) and fiscal decentralisation respectively, in ways that strengthen accountability in the LG system. What is Strengthening local governance in Bangladesh Page 12 most important is that the way forward considers these factors in an integrated manner. For example, if citizens are to be truly able to hold their elected local representatives to account then the political structure of Upazila Parishads should be matched by the functions assigned to them and their fiscal powers.  Option 1: Federating UPs into Upazilas: Upazilas would be indirectly elected, constituted by the elected representatives nominated by respective UPs (eg all UP chairman). The majority of functions and fiscal powers would be assigned to UP’s, who could delegate these upward to UPZ, and pay for services rendered. Only a small portion of functions (eg land use regulation) would be originally assigned to Upazillas. Citizens would hold their UPs accountable for the extent to which they received appropriate services from Upazilas, as UPs would be able to exercise considerable control over Upazilas. o Pros: subsidiarity which builds on a long established tier at the UP level; coordination issues minimised as Upazila’s have a limited independent power base o Cons: administrative capacity is located at the Upazila level, which to some extent remains outside of direct democratic oversight.  Option 2: Original assignments to Upazilas: Upazilas would be directly elected and assigned the bulk of functions, and associated financing. They could choose to delegate this to UP’s, along with the necessary resources. o Pros: economies of scale may be achieved in some services; functional allocation would align with administrative capacity o Cons: centuries old UPs would be considerably weakened, subsidiarity would be undermined.  Option 3: Hybrid option: Part direct and indirect elections, split in functional assignments and finance. Upazillas and UPs would share local government functions and revenue sources, supported by a clear regulatory framework that assigns powers, functions and finance between them. o Pros: greater ability to align functional assignments with administrative capacities without overburdening the system o Cons: careful process of legislative and regulatory reform would be required to guide assignments and monitor success. H. Towards a minimum programme for strengthening local governance 31. The interim government faces a conundrum with regard to local governance. On the one hand there is a clear mandate and support for holding Upazila elections. On the other hand it is clear that simply holding elections will not significantly develop or strengthen the system of sub- national governance in Bangladesh (and may indeed just create conditions for failure). Additional reforms to the structure and functions of the LG system are required to ensure its sustainability, particularly when an elected government (and Parliament) returns to office. However, the interim government has neither an extensive timeframe nor mandate to undertake the full set of reforms required in the LG system, and all of its actions will be subject to ratification by an elected Parliament. 32. However, even given these constraints, there are important steps that can be taken over the next few months, and which can be explicitly linked to a longer term process of reforms: Strengthening local governance in Bangladesh Page 13 a) A first step is to focus its effort on Upazila elections, rather than other forms of LG. The Upazila appears to have the greater chance of emerging as a separate tier of governance at this point; b) It is imperative to undertake a rapid review and assignment of core local service delivery functions to Upazila’s and UPs respectively. This will create an environment in which the new system can stabilise, through ensuring that they perform complementary and non- overlapping roles, and do not immediately begin to enter into conflict with each other or with higher tiers of government; c) Sufficient discretionary resources need to be allocated to Upazila’s so that they can perform their assigned functions in line with the preferences of their communities. Specifically, untied discretionary resources need to overshadow all other transfers from central government, and be allocated and disbursed in a fair, transparent and efficient manner; d) The system of checks and balances needs to be reviewed to ensure that Upazila’s are responsive and accountable to communities, while being provided with adequate discretion. 33. Precedent exists for taking this approach in Bangladesh. The Local Governance Support Project (LGSP) of the Local Government Division (financed by the World Bank). The LGSP uses the same principles of expanding discretionary resources to Ups while strengthening citizen oversight and output monitoring. Government, at the highest level, has committed to its implementation through the financing agreement with the World Bank, and is actively implementing the program through allocating block grants to participating UPs in the next financial year. Replicating this approach at Upazila level is possible, and would strengthen the LGSP through adding momentum to the process of strengthening local governance and clarifying the assignment of functions between Upazila and UP. 34. Each of these steps will require legislative amendments to achieve, but constitute only the basic minimum program necessary to begin the process of strengthening sub-national governance. The elected Parliament may need to ratify these amendments, which provides an opportunity to give ongoing momentum to the broader set of reforms that are necessary. The amendments might, therefore, also establish a process to review policy, legislation and budget practices to support ongoing reforms that might be undertaken by a future government. For example, a permanent commission on local governance could be established with a specific mandate to report to the future government on additional reforms. I. Managing risks in the transition 35. Key risks that are likely to be evident in any of the options are: a) The process of reform: The evolution of the local government system in Bangladesh will require an ongoing process of policy development that addresses a far wider set of issues, including for example, the procedures to demarcate local bodies, and to deal with rapid urbanisation. The interim government may neither have the time nor the resources to address all these issues and its decisions may be subject to ratification in the first session of the elected Parliament. Government may therefore find it more to their advantage to carefully select those issues that need to be addressed in the current context, while establishing a process to address outstanding issues later. This may require both outlining the process for legislative ratification of reforms, as well as using legislation to set in motion the process to undertake further policy development. Strengthening local governance in Bangladesh Page 14 b) Interference in local affairs by higher tiers: Both the bureaucracy and national MPs currently have considerable power to influence decision-making (through “mandatory� advice) at local level. Specific, systemic reforms will be required to ensure that the scope of local autonomy is expanded to prevent this. For example, sector ministries will have to withdraw from issuing detailed instructions on the implementation of programs (thereby confining themselves to determining required outputs), while the role of MP Constituency Funds and their de facto executive role at LG level would require review. c) Failure of individual local governments: Any initiative that provides more discretion to LGs must also anticipate that at least some LGs will make bad choices. In some instances these choices may lead to financial failure or problems in the delivery of basic services. Although this is no different from the current situation, there is a need to develop mechanisms for government to manage the effects of these failures. At present legislation provides only a rudimentary framework for formal interventions in LG affairs, without providing an enabling fiscal environment (a hard budget constraint) or oversight procedures to ensure that LGs have an incentive to avoid these interventions. The current framework also provides considerable room for political interference. Refinements to the legal framework will thus be required to ensure GoB is able to rapidly and effectively deal with LG failures in an integrated manner. Strengthening local governance in Bangladesh Page 15