Federalism reform imperatives, restructuring principles and lessons for Pakistan.

Shah, Anwar

In spite of several decades of sustained growth, Pakistan remains in the company of "failed" states on account of its quality of governance. This paper presents an institutional framework for improving the quality of governance in Pakistan. This framework suggests that a major challenge facing the public sector in Pakistan is to harmonise its authorising environment and operational capacity with its mission and values. It provides a broad outline of how such a difficult task can be accomplished. The paper provides an analysis of the fiscal federalism dimensions of the authorising environment and presents directions for reform. The paper recognises various political, bureaucratic and institutional impediments that stand in the way of such reform. Finally, the paper distils lessons on institutional reform of relevance to Pakistan and other non-industrialised countries.

**********

Over the past several decades Pakistan's economic performance, as measured by its rate of growth of real per capita GDP, has been better than most countries in the region. It doubled its per capita income over the past two decades [see Malik et al. (1994)]. Yet this growth performance did not translate into improved public sector performance in delivery of public services. Instead, Pakistan today represents a classic case of decay in the public sector's enabling environment and operational capacity. This is borne out by the index of governance quality developed by Huther and Shah (1996). These authors rank governance quality in Pakistan as "poor" and comparable to those of the so-called "failed" or "predatory" states in Sub-Sahara Africa. On several of the individual subcomponents of the governance quality index such as citizen participation (comprising political freedom and political stability), government orientation (a composite measure of judicial efficiency, bureaucratic efficiency, absence of corruption), social development (consisting of human development and distributive justice), and economic management (measured by central bank independence, debt management and outward orientation), Pakistan consistently ranks amongst the worst performers among a sample of 80 nations studied by the same authors. Pakistan's record in public services delivery in general is quite dismal and appears to show little improvement over time (see Figure 1). The source of this malaise has not received any serious attention by academics and policy-makers in Pakistan and elsewhere. Pakistan and the development assistance community instead has continued to focus their attention on policies to raise revenues and reduce the deficit and Pakistan continues to take half a step forward, two steps backwards and several steps sideways to skirt the issue and retain eligibility for external financing to try yet one more time to support unsustainable policies of its successive governments. This paper argues that as Pakistan aspires to earn its place among nations with a strong record of serving their citizens, it must focus its attention to reforming its institutions (including political and bureaucratic cultures) of governance. In the following a framework for improving public sector performance in a federal system is presented and its implications for reform imperatives in Pakistan are drawn.

A SIMPLE FRAMEWORK FOR IMPROVING PUBLIC SECTOR PERFORMANCE IN A FEDERAL SYSTEM

The starting point for this framework is to develop a deeper understanding of three contextual dimensions of public sector organisation in any country. These are:

(a) Public Sector Mission and Values: Societal values and norms as embodied in the constitution and more specifically reflected in annual budget policy statements serve as the formal points of reference for public sector mandates and its values in delivering these mandates. Unwritten societal norms that are widely shared or acknowledged should also be taken into consideration. In industrialised countries, public sector mission and values are realistically spelled out in terms of a medium term policy framework. For example, there is a formal requirement in New Zealand that such a policy statement must be tabled in the parliament by March 31 (about 2-3 months in advance of the budget statement). Pakistan's constitution specified reasonably precise but ambitious goals in certain areas such as attainment of adult literacy by 1966 but the accomplishment of such goals has not been seriously attempted. Similarly central planning made lofty promises that were never fulfilled. Public sector values in Pakistan have never been addressed. This is because the public sector orientation remains on "command and control" rather than serving its citizenry. For an official trained in 'command and control', need to develop a code of conduct of client orientation, may appear quite futile.

(b) Authorising Environment: Formal (budgetary processes and institutions) and informal institutions of participation and accountability are of interest here. Do these institutions and processes work as intended in providing an enabling environment for the public sector to meet its goals? Do various levels of government act in the spirit of the constitution in exercising their responsibilities? What are the checks and balances against deviant behaviour? Is independence of central bank, judiciary, and auditor general guaranteed? Is the central bank focused solely on price stability or is it expected to pursue multiple objectives? Are there formal rules to ensure fiscal discipline? Is the design of transfers consistent with their objectives? Are their private agencies that rate various levels of government for their credit worthiness? Is public sector borrowing subject to financial market discipline? How is the government performance measured? Are output and outcome indicators for public services monitored by any one?

In industrialised countries, institutional norms are strictly adhered to and there are severe moral, legal voter and market sanctions against noncompliance. Pakistan's constitution with a few exceptions such as central bank independence, provides an excellent blueprint for a sound authorising environment for the public sector but regrettably the practice in Pakistan almost completely negates the constitutional intentions.

(c) Operational capacity and constraints: What is authorised is not necessarily what may get done as the operational capacity may not be consistent with the task at hand. Some key questions to a better understanding of the operational capacity include: Do the agencies with responsibility lot various tasks have the capacity to undertake them? Do they have the right skills mix and also have the incentive to doing the right things and doing them right'? Is the bureaucratic culture consistent with the attainment of societal objectives? Are their binding contracts on public managers for output performance? Does participation by civil society helps alleviate some of these constraints? To what extent these constraints can be overcome by government reorganisation and reform? Whereas, in industrial countries, answers to most of the above questions are expected to be in affirmative it is not true in the case of Pakistan.

[FIGURE 1 OMITTED]

The challenge of public sector reform in any country is to harmonise public sector mission and values, its authorising environment and its operational capacity so that there is close if not perfect correspondence among these three aspects of governance (see Figure 2). Such a task is daunting for any developing country. It is probably a less daunting but nevertheless quite a complex task in Pakistan. This is because while most of the fundamental institutions are already in place yet a political and bureaucratic will to restore these institutions their due role may not be forthcoming.

Scope of the Paper: The framework outlined above suggests that in Pakistan there is an urgent need to start a process of institutional renewal and reform. Addressing such a large agenda is beyond the scope of this paper. Instead, the remainder of the paper simply addresses a subset of those issues dealing with federalism dimensions of the authorising environment only. The paper is organised as follows. First, an introductory overview of the federal system is presented. This is following by an analysis of taxing and spending responsibilities of various levels of government. Subsequently an evaluation of fiscal transfers is carried out. The two concluding sections discuss the impediments to reform efforts in Pakistan and draws lessons from Pakistan's experience for other developing countries.

[FIGURE 2 OMITTED]

AN INTRODUCTORY OVERVIEW OF THE FEDERAL STRUCTURE OF THE GOVERNMENT IN PAKISTAN

Pakistan is a federation of four provinces, Punjab (56 percent of total population), Sindh (23 percent of total population), Northwest Frontier Province (13 percent of total population) and Balochistan (5 percent of total population). In addition to the four provinces, parts of Kashmir which acceded to Pakistan are organised as the State of Azad Kashmir and enjoy self-government. Federally administered northern areas include the administrative district of Daiamir, Ghanche, Ghizer, Gilgit and Skardu. There are also seven self-governing federally administered tribal areas where the laws of Pakistan do not apply consistent with the tradition fostered under the British Empire. For administrative purposes, the provinces are subdivided into divisions and each division is divided into districts and each district is subdivided into tehsils. At the district and tehsil levels, elected councils and provincial civil administration have overlapping administrative responsibilities.

Local governments in Pakistan do not have any distinct constitutional status; they are established by provincial governments and powers are determined by provincial statutes (see Table 1 for an impressionistic review of the fiscal system in Pakistan). At the present time, there are 5195 local government units in Pakistan of which only 512 units are in urban areas. For local government purposes, urban areas are organised by population size class into metropolitan corporations, municipal corporations, municipal committees, town committees and cantonment boards. Metropolitan corporations have some deconcentration of metro functions by zonal offices. Karachi metropolitan corporation has in the past experimented with a zonal lower tier of local government called municipal committees. Land development and water and sewerage functions in large urban areas are assumed by provincial agencies called Development Authorities (DAs) and Water and Sanitation Agencies (WASAs) respectively, and for smaller municipalities, development expenditures are undertaken directly by provincial line departments. Karachi Metropolitan Corporation represents the exception in water and sewerage with its own Karachi Water and Sewerage Board (KWSB). Rural areas are organised into a two tier local government structure with the lower tier union council representing several villages and a larger rural area called a district having a district council comprising several union councils. Provincial statutes regarding local government organisation are fairly uniform and provide for elected local councils to have broad powers over a wide range of local public services. In practice however, for reasons to be discussed later, the existing system of local government has become dysfunctional.

In terms of programme spending, the federal government dominates in defense, debt servicing, general administration, fuel and power, transportation and communications, industrial development, population planning, manpower management, water supply, and sanitation and subsidies. Provincial government spending dominates in law and order, food and agriculture, rural development, education, and health. Local government spending is pre-eminent in social welfare, parks and recreation, and animal care. Overall, federal government spending accounted for 67.2 percent of consolidated programme spending, with the provincial government spending share at 28.8 percent, and local government spending representing, only 4.1 percent of total in 1994-95. It is also interesting to note that federal and provincial governments accounted for 73.3 percent of total spending on local public services in the same year. At the same time, the federal government collects 90.7 percent of the total tax revenue, with the remaining 9.3 pelvent collected in roughly equal proportions by the provinces and local governments.

The 1973 Constitution delineates the responsibilities of federal government (Federal Legislative List) and the areas of shared responsibilities (Concurrent Legislative List). All residual functions are the domain of provincial governments. Local governments functions are defined under provincial laws through the use of Local Government Ordinances. The federal government is responsible for such key areas as defense, foreign affairs, international trade, macroeconomic management and industrial development, among others. The provinces share with the federal government responsibility for population planning, curriculum and syllabus planning, and social welfare, and are solely responsible for roads, highways, police protection and justice, agriculture and post-primary education. Local governments are given responsibility for local public services as basic education and health, fire protection, parks and recreation, and water and sewerage.

Despite the extent of decentralisation of legislative responsibilities as set out in the constitution and provincial statutes, the actual assumption of responsibilities is more centralised. For example, though education is a sub-national responsibility in Pakistan, the federal government nevertheless takes responsibility in national policy formulation, planning, curriculum, and managing centres of excellence. Similarly, although health care in Pakistan is a sub-national responsibility and much of it assigned to the local level, in practice, it is more centralised. The federal government is responsible for drug regulation and quality control, regulating the standard of professional education, undertaking measures for containing communicable disease, and interacting with external organisations.

Implications of the Existing Structure

A quick review of the federal structure in Pakistan suggests that the legislative division of powers in Pakistan as set out by the constitution and provincial statutes compares favourably to the assignment based on fiscal federalism principles especially the subsidiary principle i.e. assignment should be at the lowest level of government unless a convincing case can be made against such an assignment based upon efficiency, equity or economic union arguments see Shah (1994), for an overview of the fiscal federalism principles. See Appendix Tables AI and A2 for a stylised assignment of responsibilities). However, the actual assumption of responsibilities by various levels of government during the past two decades have consistently departed from the de jure assignment. Higher level governments have increasingly encroached not only upon lower level government responsibilities, but also upon traditional private sector functions. There appears to be a trend towards centralisation of responsibilities with the federal government assuming either an exclusive role (e.g., population planning, electricity, curriculum development, syllabus planning, centres of excellence) or a dominant role (e.g., social welfare, vocational/technical training, employment exchanges, historical sites and museums) for some of the functions on the concurrent list, and even for some specified to be purely provincial. It should be noted that while majority of social welfare expenditures (87.7 percent in 1994-95) are carried out by local governments/religious institutions/community groups, these entities simply administer federally defined and financed programmes. Similarly, provincial governments have gradually assumed a large number of functions previously carried out, and intended to be carried out, by local governments (e.g., curative health, land development, primary education, farm-to-market roads, rural development, preventive health, water supply, drainage and sewerage) and/or the private sector. Superior institutional capacity at higher levels of government and the attendant efficiency in service delivery was cited by the federal and provincial governments as main reasons for this centralisation of responsibility for public services. And, market failures have been cited as the reason for higher level governments in Pakistan to increasingly have assumed responsibility for direct provision of private goods and services. Such an involvement is quite widespread, and encompasses such key economic areas as manufacturing, wholesaling, retailing, banking and financial intermediation functions.

In general, wider interpretations of their constitutional mandates by federal and provincial governments have had the following implications for public sector" management in Pakistan.

(i) Aggregate fiscal discipline is lacking inspite of centralised controls. This arises from a lack of appropriate institutional environment to force governments at all levels to face financial consequences of their decisions. Such an institutional environment requires (a) an independent central bank entrusted with the sole task of price stability alone; (b) arms length relationship between the financial sector and governments; (c) fiscal rules and gate-keeper committees for self-discipline; and (d) no bailouts of any levels of government by the central bank or by higher" level governments see Shah (1997) for a detailed overview of experiences with macroeconomic management under centralised and decentralised fiscal systems). Regrettably, Pakistan continues to follow failed policies of fiscal discipline through centralised controls.

(ii) The private sector is crowded out. Federal and provincial governments during the 1970s acted aggressively to curtail private sector participation in education and health services and in industry, banking and finance resulting in limiting citizens' choices and lowering the quality and quantity of private and quasi-private goods and services. It is useful to note here that recent survey results indicate that the public sector in Pakistan is less efficient than the private sector in the provision of education and health services [see Hassan (1995)].

(iii) Decision-making is highly centralised. In areas of concurrent responsibility, the higher level of government has often adopted a dominant position. Decisions on most large capital projects require federal approval. And, as most local functions are being assumed by the provincial government, local participation in major" public policy decisions that affect them is by and large non-existent.

(iv) Efficiency and equity in public service provision are undermined. Since the decision-making is significantly divorced from the people on whose behalf decisions are being made, there is a greater likelihood of imperfect matching of local public services with local preferences, and of inadequate levels of response to local needs. These problems are particularly acute in primary education and basic health. In primary education, enrolment rates are low (42 percent), dropout rates are high (50 percent between grades 1 and 2) and female literacy rates are low (22 percent). In rural areas, fewer than 15 percent of girls complete 5 years of education, and only 1 percent of girls remain in school by age 14. Absenteeism by teachers goes unreported. Teachers are not accountable to local communities or parents or even to school heads, but are transferable at will by the Provincial Department of Education. Centralisation of primary education has also worked to create "ghost" schools which either exist on paper but could not be located or have a presence in the form of school buildings but cannot find students. In basic health, performance under provincial administration is no better. Most health centres lack medicines and supplies and in many instances proper sanitary conditions. In rural areas, health centres often lack both medical personnel and supplies. Medical personnel are frequently transferred against their will. There are no channels for local participation and input in the provision of health care. While no welfare cost estimates associated with centralisation are available for Pakistan, for industrialised countries such losses on account of simply misjudging preferences are expected to range 9-20 percent of expenditures. Once one considers the lack of local accountability, participation and delinking of benefits and costs, welfare losses are expected to be of much greater magnitude.

(v) Public sector accountability is impaired. Separation of taxing and spending decisions at the provincial and local level compromises accountability as citizens do not know with any precision who is responsible for what services. The level of government raising most of the revenues, i.e. the federal government, may be blamed for provincial-local service delivery problems. Accountability is also impaired because of a lack of clarity regarding the roles of various levels of government in areas of shared responsibility.

(vi) The local public sector is weakened. Pakistan started out with a well thought out and well developed local government organisation structure. During the 1950s and 1960s, local governments were a vibrant part of the public sector and delivered most of the local public services. During the 1970s and 1980s, some of the more important local functions, such as primary education and basic health, were shifted upwards to provincial governments. Provincial and the federal governments currently dominate the provision of local public services. As an example, consider water and sewer where the federal and provincial governments together account for 95 percent of total public sector expenditures. This centralisation of local service responsibilities has led to deterioration in service quality, as accountability is significantly impaired.

The implication of this is that the current degree of centralisation of expenditure responsibilities have not served the country well and may not be sustainable in the long run.

Reorienting Government Functions to Strengthen Accountability and to Improve Service Delivery

To overcome above-mentioned problems, a number of options may be worthy of consideration by Pakistan. These include:

(a) Fostering enabling environment for aggregate fiscal discipline. Critical elements for such an enabling environment would be: (a) granting independence to the State Bank of Pakistan but having the Bank governor under contract to ensure price stability with an independent oversight board to review performance under this contract; (b) prohibiting ownership and preferential access to financial markets by any level of government; (c) establishing golden rules for borrowing (i.e. borrowing permitted for long term capital projects only) and guidelines on the permissible levels of deficit and debt as a percent of GDP and having gatekeeper federal-provincial committees to enforce compliance; and (d) strengthening private sector oversight of all governments through requirements for credit ratings and private financial audits.

(b) Disengaging federal and provincial governments from provision of private goods and services and from production of public goods and services that can be delivered more efficiently by the private sector. This process has been underway at a snail's pace. The successful experience of the Czech Republic and not-so-successful experience of Romania demonstrates that the success critically depends upon the rapidity of pace at which reforms are executed. Deregulation of economic activities started seriously in Pakistan only in the early 1990s. The redistributive role of the public sector requires that the provision of basic social services due to their positive redistributive impacts should be a fundamental public sector responsibility. Nevertheless, the scope for greater private sector involvement to assist the public sector in the delivery of social services is substantial. This would conserve scarce public resources and improve the quantity and quality of services.

(c) Restoring federal budgetary flexibility to ensure fulfilment of its mandates in macroeconomic stabilisation, regional and personal equity, and preservation of internal common market. It entails focusing federal government role primarily to its constitutional mandate. This requires retrenchment of direct federal roles in private and local public goods. It also requires rethinking current system of revenue sharing and discretionary transfers and replacing these by fiscal equalisation and conditional (per capita block transfers) conditional on standards of services to be achieved as opposed to current conditions on spending levels. This will allow the federal governments to use its spending power (through fiscal transfers) to influence subnational governments priorities in meeting national standards in education and health and basic infrastructure as well as ensuring comparable levels of basic services at comparable levels of taxation across the country.

(d) Clarifying and coordinating roles of federal and provincial governments. One option is to eliminate or at least curtail the federal role in a number of areas that are either provincial oz" local responsibility such as education or where central functions can impede progress in seeking innovative solutions e.g. central planning. Such candidates for curtailment include: central planning and development, irrigation, tourism, justice, transport and communications, population control, curriculum development, syllabus planning, and establishing centres of excellence. Such a retrenchment would have welcome effects on federal budgetary flexibility and is expected to cut federal deficit in halt'. In social services, the federal government's main interest would be to ensure that norms of national equity are satisfied. As such, it would mainly be concerned with setting national minimum standards and providing financing to the provinces conditional on their complying with these standards.

(e) Clarifying provincial-local responsibilities and strengthening local governments. Provincial governments have increasingly adopted an intrusive and unhelpful role in local governance. Local governments are being denied an opportunity to succeed and instead their roles are being subsumed by provincial governments. Arresting this trend could restore public confidence in the public sector. Provincial governments simply do not have the capacity to oversee, monitor, and manage the performance of a vast institutional network of basic services like primary education and basic health. It would therefore be desirable to curtail provincial role in land development, urban and rural water supply and drainage, primary and secondary education, basic health, preventive care and local roads. For all these services, however, the provincial oversight role and assistance for technical and financial matters needs to be strengthened.

The provincial role in financing primary and secondary schools and basic health could be large if not complete. Management of schools and health centres including staff recruitment, however, should be a local responsibility. In general, all local public services could be devolved to local governments in an asymmetric fashion depending upon the population and area served and local fiscal capacity. Large metropolitan areas with population in excess of one million could be given broad local powers. These may be subdivided into municipal committees (responsible for neighbourhood services) of say 500,000 residents each with regional/metropolitan functions (water and sewer trunk lines, intermunicipal roads, regional planning, land use planning etc.) assigned to a metropolitan corporation comprising of nominated members (possibly chairpersons) from member municipalities.

Finally, consideration may be given to amending local government ordinances to introduce safeguards against unwarranted provincial interference in local affairs and to strengthen provincial financial oversight by requiring municipalities to submit commercially audited financial statements to the provincial government in a timely fashion. Private commercial auditing of federal and provincial governments would further strengthen government accountability. Further, mechanisms may be developed to streamline provincial technical assistance to municipalities. Provincial officials at the local level may be required to submit an annual review of municipal governance issues and a strategy for provincial financial and technical assistance in their jurisdiction to be shared with the province as well as with local government officials. Alternate means to strengthen the supportive role of provincial officials to local government institutions may also be explored. Consideration may be given to eliminating the dual administrative structure at the local level by withdrawing the federal/provincial role in direct administration at local levels [see Government of Pakistan (1994)]. This can be accomplished as part of the civil service reform. It is interesting to note that local governments in India faced similar circumstances prior to 1993 as state governments often either encroached upon their responsibilities or simply disbanded elected councils as is the current common practice in Pakistan. To rectify this situation, Indian Parliament passed two constitutional amendments--the Seventy-third (The Panchayats) and the Seventy-fourth (The Municipalities) Amendments to the Constitution Act, 1992. By these amendments, local governments in India have acquired a constitutional status and their powers and relationships with states have been clarified [see Bakshi (1994)]. In addition, these amendments require states to establish a State Finance Commission for distribution of state fiscal transfers to local governments and establishment of Metropolitan Planning Committee for planning and coordination of development activities comprising of at least two-thirds of the elected members from member urban and rural municipalities in proportion to their population and the appointment of remaining members may be specified by the state legislature. The responsibilities of rural (Panchayats) and urban municipalities are specified in the Eleventh Schedule (Article 243G) and the Twelfth Schedule (Article 243W) of the Indian Constitution Act 1992 (see Table A2.10).

Special Issues for Administration of Rural Municipalities

Local administration of rural areas is heavily concentrated at the district level with union council having negligible or minimal responsibilities. The district level, however, is typically unsuited as a local government unit because of the size of the area and population to be served. For Pakistan as a whole, a district averages 883,000 persons. Comparable figures for Sindh and Punjab are about one and 1.3 million people. These are very large populations being under one government by any standards especially when the union council have minimal or negligible responsibilities. In western Canada, rural districts typically serve only 20,000 people. Not only that, the district areas in Pakistan are very large. They average 100 km by 100 km for the nation as a whole and almost that size in the densely populated province of Punjab. Given their large size and clientele they are not suitable for delivery of basic services. Instead, their role could be of a regional government that would take care of planning, coordinating, and of supplying services that encompass areas covering many union councils. They could be responsible for major link roads, sewerage connectors, waste treatment plants and disposal sites, water supply (i.e. mains and not intra-community distribution systems), a second tier of health facilities and perhaps secondary schooling. Rural property assessments (using a provincial base) could also be suitably assigned to them. A potentially useful role for the districts would be providing technical and other support tot union and tehsil councils such as (often occasional demands for road, water, sanitation) engineering advice and the training to develop accounting and management skills of union and tehsil council employees.

The size of the union councils (averaging 16,000 persons and 191 sq. km) compares favourably with that of the local governments that are responsible for the bulk of local services and functions in most countries. In almost all of the developed countries surveyed in McMillan (1995), the size of local governments fell into the 10,000 to 30,000 population range. Although there are wide variations within countries around these numbers, yet this range says a lot about a size of basic local unit thin can be effective in providing the broad range of local services and effective in a political representation sense also. These authorities could certainly play a much larger role. In communities this large, there is ample decision-making capacity and these are the people best aware of local needs, conditions and preferences. Decisions demanding technical skills often do not demand continuous input but rather good advice when a decision is to be made; e.g., road, water or sewer line construction and the supervision of such work when done. This is regularly handled in smaller municipalities throughout the world by hiring that expertise when needed. Local decision-makers are best able, given well defined alternatives, to select the facilities and allocate the resources to best meet local needs and utilise the available resources. Advisory services might come from the district level, but with the development of demands by union council governments, private and competitive expertise would soon emerge.

The magnitude of the district governments and the very reasonable size but minimal responsibilities of the union councils is an anomaly suggesting that the roles of the district and union councils require a careful review. Given the size of the rural district councils, they could assume responsibility for functions like those of the larger urban municipalities and even some of the upper tier metropolitan government. Indeed, it would not be surprising if the tehsil council (combinations of union governments possibly at the tehsil level) might not be more effective at some of those functions than the district councils which may be too large (geographically) for some of these activities.

THE FEDERAL-PROVINCIAL ALLOCATION OF TAXING POWERS

The allocation of taxing powers to various levels of government has an important bearing on the character of federalism and on accountability. For example, if the ability of sub-national jurisdictions to raise their own tax revenues is restricted, this can impose significant constraints on the ability of these jurisdictions to fulfil their proper expenditure responsibilities, and can therefore compromise the potential benefits to be had from decentralisation. The allocation of taxing powers is determined jointly by the constitutional assignment of tax sources by level of government as well as by the extent to which each level, especially the federal has chosen to exploit the major tax bases. The higher the tax rates chosen by the federal government, the less room there will be for raising revenues by sub-national levels, and the more dependent will the latter be on the federal government for their revenues.

The fiscal federalism literature advances the following principles for tax assignment [see Shah (1994) for further discussion].

Efficiency of the Internal Common Market. For free mobility of factors and goods and services, taxes on mobile factors of production or tradable goods and services should either be assigned to the federal government or harmonised and coordinated among subnational governments.

National Equity. Progressive redistributive taxes should be assigned to the level having primary responsibility for redistribution.

Administration Cost. To minimise collection, administration and compliance costs, a tax should be assigned to the level likely to be best informed about its base.

Fiscal Need. To ensure accountability revenue means should be matched as closely as possible with expenditure needs.

The 1973 Constitution of Pakistan specifies the following as areas of taxing responsibility for the federal government:

* customs duties including export duties;

* excise duties including duties on salt, but excluding duties on alcoholic liquors, opium and other narcotics:

* duties in respect to succession of property;

* estate duty in respect of property;

* income taxes other than on agricultural income;

* corporation tax;

* tax on sales and purchase of goods imported, exported, produced, manufactured or consumed;

* taxes on the capital value of assets, not including taxes on capital gains on immovable property;

* taxes on mineral oil, natural gas and minerals for use in generation of nuclear energy;

* taxes and duties on production capacity of any plant, machinery, undertaking, establishment or installation in lieu of any or more of them; and

* terminal taxes on goods and passengers carried by railway, sea or air and taxes on their fares and freights.

All other forms of taxation not specified in the above list fall under the purview of provincial governments. The provincial governments then allocate some of these to local governments on the basis of local government ordinances. The major sources of revenue for provincial governments are in fact transfers from the federal government in the form of revenue sharing. The major own sources of provincial revenues are stamp duties, motor vehicle registration taxes and entertainment taxes. Local governments rely heavily on provincially administered real property and motor vehicle registration taxes and on self-administered import/export (octori/export) taxes. The use of these assigned taxing powers by the various levels of government in Pakistan results in actual revenue raising to be heavily concentrated at the federal level, much more so than is required to finance its expenditure responsibilities. The federal government collects 90.7 percent of consolidated current revenues and retains only 59.3 percent lot own use. It is also interesting to note that while the federal government collects resource taxes and royalties primarily for transfers to provincial governments, the latter governments also reciprocate by collecting taxes on capital values and for the religious taxes, zakat and ushr, at rates determined by the federal government, and turning the proceeds over to the federal government. Such an arrangement whereby revenues are passed upwards is unusual in federations, the notable exceptions being China (a unitary country) and Russia where for historical and institutional reasons the major revenues are collected by sub-national governments and shares are transferred to the central government. In Pakistan, these are relatively minor revenue sources, and the reason for provincial administration, at least for the zakat and ushr (1) owes to the religious nature of them.

One important and distinguishing feature of the constitutional assignment of taxing powers to the federal government and the provinces in Pakistan is the tendency to assign major tax bases exclusively to one level of government or the other. Moreover, those bases that are often the largest revenues sources, such as sales and excise taxes and income taxes, are assigned to the federal government. In many federations, such taxes are co-occupied by the two levels of government so both are able to obtain revenues from them independently, though preferably in a harmonised fashion. This tendency to exclusive assignment has been a factor leading to the highly centralised revenue raising system observed in Pakistan, as well possibly to the fact that revenue sharing is resorted to as a means of transferring funds to the provinces. As we note later, revenue sharing has the disadvantage of not fostering accountability at the provincial level. This feature of exclusive assignment is not present at the provincial/local level. Entertainment taxes and taxes on cinemas, hotels and professions are co-occupied by provincial and local governments usually with independent collection.

Tax assignment in Pakistan emphasises efficiency in tax administration and uniformity of the tax regime across the nation to the neglect of provincial and local fiscal autonomy and fiscal needs and accountability considerations. Thus provincial governments are almost completely dependent upon revenue transfers from higher levels of government to finance their own expenditures. This degree of separation of taxing and spending responsibilities may not be in the interest of either the federal or the provincial level of government. Provincial governments have no control over their major sources of revenue, and may not have any incentives for cost efficiency or for raising revenues from own sources as additional efforts may not be worth the political costs.

A possible option for reform would have the federal government retain exclusive responsibility for taxation of international trade, corporate income and for value-added taxation of sales of goods and services. The federal government would maintain primary responsibility for taxing personal income, though with the ability of the provinces to piggyback at flat rates. The current practice of the federal government levying excises on production and royalties on natural resource exploitation and turning over the proceeds to the provinces could be reviewed. This is anomalous for three reasons. First if these are provincial revenue sources, it is not clear that the federal government should be involved in their administration. Accountability would be enhanced by allowing the provinces to act as their own tax authorities. Second, given that crude oil and natural gas are distributed unevenly across provinces, the practice of returning them to the province of origin leads to differences in fiscal capacity across provinces. And third, fiscal federalism principles suggest that major resource revenues should be retained as federal revenues. Provinces could have control over most excise taxes, taxes on agricultural land, and over wage (payroll) taxes, and could levy a broad-based single sales tax in parallel with the federal VAT with its base harmonised with the Federal VAT at retail level. Local governments may be given a role in setting property tax rates, and access to electricity duties, school lees, water and sewer rates and taxes on cinemas, hotels and entertainment services. Taxes on interprovincial and intermunicipal trade should be considered for elimination. This tax assignment would be consistent with the fixed constitutional arrangements for sharing fiscal powers.

It is interesting to note that in Pakistan in 1994-95, over 90 percent of all taxes were realised through withholding and voluntary payments, and 100 major corporations account for 75 percent of income, 80 percent of excises and 60 percent of sales taxes. Further, the record of federal tax administration bureaucracy in tax enforcement is not enviable. Surveys of businesses have indicated that the tax collection machinery appears to be more zealous in collecting bribes than taxes [see Stone (1995)]. Therefore, institutional capacity for tax administration should not be considered to be a serious impediment to any changes in tax assignment

These suggestions in aggregate would lead to a significant change in own revenues generated at various levels of government in Pakistan and would result in greater conformity of revenue means with expenditure needs for these governments.

FEDERAL-PROVINCIAL FISCAL TRANSFERS

Matching revenue means as closely as possible to expenditure needs of various levels of government serves to strengthen accountability in a federal system. Stronger tax performance and better cost recovery policies at the sub-national level will help in this task. It is however, desirable in federal systems for higher level governments to have access to more revenues than those dictated by their direct programme responsibilities alone. These additional revenues can be used to further national or provincial economic objectives such as setting national standards, securing economic union and ensuring interregional and intermunicipal fiscal equity. The design of these transfers, however is critical to achieving the objectives sought (see Table 2). The issues pertaining to the design of these transfers are the subject of focus in the following paragraphs.

Federal fiscal transfers to the provinces are the dominant source of financing operating expenditures for provincial governments in Pakistan, accounting for 87.5 percent of their operating expenditures in 1994-95. They financed nearly 99 percent of operating expenditure for the two smaller provinces. The design of these transfers have important implications for the fiscal behaviour of provincial governments and for the efficiency and equity of public service provision in Pakistan. Provincial transfers to local governments on the other hand are of minor significance in local finances as the provinces have provincialised major local public services such as basic health, education and water and sanitation. In the following, the existing structure of federal-provincial transfers is first described and then evaluated for consistency of its design with the objectives sought.

FEDERAL-PROVINCIAL FISCAL TRANSFERS: CURRENT STRUCTURE

Federal government transfers to the provinces consist of both unconditional and conditional components. These are briefly discussed below.

A. Unconditional Transfers: These transfers are advised by an intergovernmental body, the National Finance Commission (NFC), that is appointed every five years to conduct a review of federal transfers. Federal and provincial finance ministers, federal finance secretary and one additional member from each of the provinces are represented on the Commission. This Commission has a chequered history, with many instances of either not meeting, or meeting and not achieving a consensus view. During the past two decades, only the 1991 and 1997 NFC recommendations were made public and implemented. As a result of 1997 Award, the following federal unconditional transfers are currently available to the provinces.

(i) Constitutionally Mandated Revenue Sharing: The constitutionally mandated revenue sharing programme consists of three elements

(a) Revenue Sharing by Origin--1997-98 (22 billion rupees)

Revenues shared on the basis of collection 14.2 percent
Excise duty and royalty on gas 3.6 percent
Surcharge on gas 4.3 percent
Royalty on crude oil 1.0 percent
Profits on hydroelectricity 5.5 percent

(b) Revenue Sharing by Population--1997-98 (171.4 billion rupees)

For this element, the NFC has defined a divisible pool that is shared on 62.5:37.5 basis among federal and provincial governments. The divisible pool consists of the following taxes: Taxes on income including corporation tax, but not including taxes on income consisting of remuneration paid out of Federal Consolidated fund: wealth tax: capital value tax: taxes on the sales and purchases of goods imported, exported, manufactured or consumed: export duties on cotton; customs duties; federal excise duties, excluding the excise duty on gas charged at well head: any other tax by the Federal Government. With the 1997 Award, the following is the contribution of individual taxes to this divisible pool.

Revenue Sources Shared on the Basis of Population 81.5 percent
Income Taxes 21.1 percent
Other Direct taxes 00.8 percent
Import duties 26.8 percent
Excise duties 13.2 percent
Sales tax 19.5 percent

(c) Special grants: The 1997 NFC Award recommended the following transfers:

* To NWFP Province: Rs 3.3 billion in 1997-98 escalated by changes in the Consumer Price Index (projected annual rate of 11 percent) for a period of five years.

* To Balochistan Province: Rs 4.1 billion in 1997-98 escalated by changes in the Consumer Price Index (projected annual rate of 11 percent) for a period of five years.

B. Conditional Transfers: These transfers are of lesser significance in relative magnitudes but nevertheless about a dozen large grant programmes have been in vogue in recent years.

Closed-end Matching Grants for Provincial Resource Mobilisation. As part of the 1997 NFC Award, this programme provides federal matching assistance at a 50 percent rate, upto a limit, for provincial revenue effort in excess of the historical average growth rate of 14.2 percent. The limits on these grants are Rs 500 million each in case of Punjab and Sindh and Rs 100 million each for NWFP and Balochistan. The programme recognises fiscal effort only on account of increases in tax rates, withdrawal of exemptions, imposition of new taxes, and revision in rates of user charges.

Development Grants. These grants are based on approval of provincial annual development plans (ADPs) by the Federal Government.

Federal Contribution for Social Action Programme. The federal Government provides matching transfers on a 75:25 basis to finance provincial development expenditures in education, health, water supply and sanitation associated with the Social Action Programme.

Flood and Disaster Relief. These grants are by their very nature ad hoc in character, and are usually given to the Provinces for emergency relief and repair and renovations to basic infrastructure as a result of damages arising from natural disasters.

Physical Planning and Housing Project Assistance. These transfers finance federally approved provincial projects to upgrade urban infrastructure and housing projects. Federal assistance is given in the form of conditional non-matching grants.

Tameer-e-Watan Programme. This programme provides block funds allocations to federal legislators (in 1994-95, Rupees 6 million each for members of the National Assembly and Rupees 5 million each for senators) to be used for development project of their choice through Provincial Ministries of Local Government and Rural Development.

Tameer-e-Sindh Programme. This programme provides federal financing on an ad hoc non-matching basis tot rural development initiatives in the Province of Sindh.

Other Federal Initiatives. Federal line ministries also fund various provincial projects usually in the areas of social welfare, population planning, health, irrigation and drainage.

Prime Minister's Discretionary Funding. The Prime Minister occasionally uses discretionary allocation to provide provincial/local governments with funds to finance special programmes.

Federal Transfers to Universities. While higher education in Pakistan is a provincial responsibility, the Federal Government through the University Grants Commission has traditionally provided financing for university education. The funding mechanism used by the UGC is ad hoc and is primarily guided by budgetary needs for salary expenditures.

Almost all of these conditional grants programmes are discretionary in nature rather being than formula-driven. This has the disadvantage of requiring that the grants be allocated by administrative decision, which can be both costly and detract from of provincial decision-making autonomy. As well, funds available may be unpredictable thereby hampering long-term planning.

AN ECONOMIC EVALUATION OF FEDERAL TRANSFERS IN PAKISTAN

Revenue Sharing and Unconditional Transfers

The revenue sharing programme suggested by the 1997 NFC Award allocates selected revenue sources on the basis of collection whereas others are distributed to the provinces on the basis of population. In addition, special lumpsum transfers are provided to the two smaller provinces. These three components of the current programme are reviewed in the following.

(a) Revenue Sharing by Origin: This component, which distributes roughly 14 percent of total revenues, returns resource revenues (oil, gas excises and royalties and hydro profits) by point of collection. Hydroelectricity profits of federal public enterprises are also returned to the province where such plant may be located. The formula used for a determination of these profits makes no provision for capital consumption allowances. The existing programme of resource revenue sharing by origin seems difficult to justify as the federal government collects revenues at centrally determined rates and then returns them to provinces on the basis of point of collection. The reason for doing so may be motivated in part to give resource rich provinces feelings of entitlement and in part political accommodation needed to maintain national unity. To be sure, this treatment remains contentious as the largest province is now seeking the same treatment for federal royalties/excises on cotton (see The Dawn. Pakistan, Internet Edition, January 12, 1997). There seems neither to be an efficiency nor an equity argument for doing so unlike population or needs-based revenue sharing. Thus asymmetric treatment of resource based revenues is difficult to justify and therefore, they could also be considered for inclusion in a general revenue sharing pool to be distributed on a per capita and/or need basis or better still such revenues go to the federal general revenues pool for financing fiscal equalisation transfers to the provinces. Furthermore, the assignment of hydroelectric profits to the province of location of the facilities could be reassessed as it undermines the financial ability and autonomy of the national power enterprise (WAPDA).

(b) Revenue Sharing by Population: About 85 percent of revenue sharing funds are distributed by population. This programme is to be commended for its simplicity, objectivity and transferring a large pool of resources in a predictable fashion to bridge vertical fiscal imbalances. Moreover, the use of population implies that the funds are distributed in an equalising manner. Although population is used as the sole criterion to achieve regional equity, it is a rough indicator of the need for funds since many provincial expenditures escalate in rough proportion to population. Overall, the transfers are strongly equalising with respect to own tax collections (rank correlation -1.0), and mildly redistributive with respect to provincial GDP (rank correlation. = -0.4). The programme, nevertheless, is subject to a number of limitations which are discussed below.

Limitations of the Existing Revenue Sharing by Population

(i) Accountability Concerns: Large transfers to reduce vertical fiscal gaps have the potential of introducing unintended adverse incentives (Table 3). It is desirable to examine first the possible sources for this imbalance, and alternative means of dealing with these issues before agreeing on the best system of inter-governmental transfers. Vertical fiscal imbalance in Pakistan can be traced to a number of reasons.

* Mismatch of Revenue Means and Expenditure Needs. As we [lave seen, this indeed is an important source of the problem, especially the over-centralisation of taxing responsibility. We earlier noted that centralisation of expenditures has been used as a response to correct fiscal imbalances arising at the subnational levels. Tax decentralisation on the other hand has remained an unexplored option. Such an option needs to be seriously examined to deal with this issue.

* Provincial Tax Bases' are Limited. While significant untapped potential exists that could raise provincial revenues from own sources, these efforts may not alone be sufficient for meeting provincial expenditure needs given the set of tax bases currently assigned to the provinces. Therefore, both tax decentralisation (say, excises) and joint occupancy of some tax fields (e.g., personal income taxation) could be looked at to reduce provincial fiscal imbalances.

* Even the Existing Tax Assignments are Inadequately Used. This is partly due to excessive dependence on Federal transfers, which reduces collection incentives, and encourages weak tax administration. An explicit manifestation of these incentives is the non-exploitation of tax on agricultural lands by the provinces.

Thus, in dealing with vertical fiscal imbalances, reassignment of some responsibilities, tax decentralisation, tax base sharing and tax abatement are options which could take precedence over unconditional grants or general revenue sharing. The latter options, by separating taxing and spending responsibilities, impair accountability and should be relied on only to the extent that decentralisation of fiscal responsibility is not appropriate. This is clearly not the case in Pakistan. Here, there has been a trend to centralise both taxing and expenditure responsibilities and then rely on revenue sharing to deal with fiscal gap. These revenue sharing transfers finance up to 95 percent of provincial expenditures. Such an overwhelming dependence of provincial governments on federal transfers has tended to undermine federal budgetary flexibility as well as impairing sub-national public sector accountability. Furthermore, these transfers may discourage the provinces in realising the full potential of own tax bases. Citizens do not see a link between taxes paid to a particular level of government and public services offered by the same level of government.

(ii) Regional Fiscal Equity: The programme ignores fiscal capacity in addressing regional equity issues. In doing so, it lacks an explicit equalisation standard against which programme achievements can be measured. Adoption of a formal fiscal capacity equalisation programme, on the other hand, has the potential of determination of total amounts of transfers, and their allocation among provinces by a formula that sets a specific standard of equalisation to be achieved.

(c) Special Fiscal Need Grants: These transfers are intended to compensate the two fiscally disadvantaged provinces for their weak fiscal capacities but higher expenditure needs. In the absence of significant tax decentralisation and having a formal equalisation programme, it represents a pragmatic approach to deal with expenditure need differentials.

CONDITIONAL TRANSFERS: AN EVALUATION

Conditional transfers to sub-national governments are advocated to ensure certain minimum standards of services across jurisdictions (conditional block grants): to pursue higher level government objectives (could be conditional matching or non-matching); or to address inter-jurisdictional spillovers of benefits (conditional matching transfers). With the exception of a selected few programmes discussed below, in general federal-provincial specific purpose transfers in Pakistan are ad hoc and primarily used for agency functions or to advance political objectives. Moreover, they are awarded on a discretionary basis rather than being determined by objective formulas.

The programme of matching transfers for resource mobilisation is not well conceived. It rewards provinces for higher tax effort due to any changes in structure and rates of taxation but provides no incentives for revenue increases due to improvement in efficiency in tax collection and administration. By doing so, it potentially opens up the possibility of a province to shift a significant burden of its taxation to non-residents by going lax on collection of existing taxes but introducing newer (possibly nuisance) taxes. The programme also shows the federal government's lack of concern with additional burden of taxation and deteriorating quality and quantity of provincial public services. This is particularly worrisome as the effective burden of the public sector (taxation inclusive of bribes) in Pakistan is considered very high and the net benefit (benefits of public spending minus burden of taxation) of the public sector as a whole may be negative. The closed ended commitment of this programme is to be welcome as it will put a limit to any potential abuses and also restore the credibility of federal commitment tarnished by the unfilled open-ended commitment of a similar programme under the 1991 NFC Award.

Box 2. A Summary of Options on Federal-Provincial-Local Transfers

Discontinue all discretionary federal transfers including those to
MNA and senators and consider the following design of federal
transfers.

Dealing with vertical and horizontal fiscal imbalances

Short run (in the absence of tax decentralisation): Per capita
transfers

Long run (with tax decentralisation). Fiscal capacity equalisation
programme. This programme could be funded out of federal consolidated
fund revenues and could upgrade provincial per capita fiscal
capacities to a national norm.

Setting National Standards through Conditional Block Transfers:

* School grant. Per capita transfers using school-age population
(age 5-17) for primary and secondary education intended for local
governments (other than metropolitan areas) to pass-through the
provinces. Provinces provide grants to educational institutions based
upon the number of graduating students. Federal government makes
direct transfers to metropolitan areas based upon school age
population. Metropolitan governments provide grants to public and
not-for-profit private schools based upon graduations.

* University education grant. Per capita transfer to the provinces.
Provincial distribution to post-secondary institutions (public and
not-for-profit private) based upon graduation by programme type.
Higher weights for engineering and medicine, lower for sciences and
social sciences and still lower for language and literature.

* Health grant. Per capita transfers for basic health based upon
weighted population with higher weights for population aged 0-5 and
65 and over. Provincial distribution to local governments based upon
the same criteria. Direct federal transfers to metropolitan areas
using the same criteria.

* Link Roads grant. Per lane mile grant based upon the requirements
for provincial-connector roads for access to federal highways.

The federal programmes of matching ratio for transfers for the Social Action Programme could be reassessed. Consideration may be given to instituting conditional block (per capita) transfers for primary education and basic health, whose magnitudes are determined by formula (e.g., equal per capital rather than by discretion. These transfers may be made directly from the federal government to metropolitan corporations. For other local governments, these grants would pass through provincial governments. The primary school grant could be made conditional on provinces meeting targets on decentralisation of primary education and basic health, ensuring private sector financing on par with the public sector (based on number of graduating students) and overall targets on access to such education.

The basic health grant could be linked to standards of access to such services. Similarly, federal transfers to universities through the University Grants Commission could be eliminated and instead substituted by per capita transfers for post-secondary education to provinces. Provinces in turn may be encouraged to substitute own transfer programmes to both public and private universities and colleges utilising graduation by programme type as criterion. Various programmes could be weighted differently but financing of public and private education would be on an equal footing.

The use of development grants could be restricted and must ensure maintenance funds. Ill this regard, grants given to MNAs, Senators and MPAs could be eliminated as such grants could undermine local and provincial government priorities.

SUB-NATIONAL GOVERNMENT ACCESS TO CAPITAL MARKETS

The 1973 Constitution of Pakistan (Article 167) allows unfettered capital market access to provincial governments guaranteed by Provincial Consolidated Fund Revenues (which includes federal transfers) and subject only to any limits imposed by the provincial assembly provided that provinces do not owe any debt to the federal government. All the provinces are heavily indebted to the federal government, so access to capital markets for financing provincial-local capital project is not available to sub-national governments in Pakistan. In 1994-95, debt charges accounted for 23 percent of provincial current expenditures. This amounts to 165.8 percent of own source revenues. Most of this debt arises from capital investment in social sectors and, therefore, accumulation of assets results in little increase in debt carrying capacity. Currently, the federal government does not encourage any additional borrowing power by sub-national governments until they have stabilised their budgets. It is important to ensure, however, that the federal government does not: (1) supply debt; (2) subsidise debt; (3) guarantee debt; and (4) bails out defaulting government. Beyond this, credit market discipline is the key to responsible subnational borrowing.

The 1997 NFC provision which allows provincial borrowing from financial markets or Ways and Means Advances from the State Bank of Pakistan to finance operating deficits could inadvertently encourage fiscal mismanagement. Sub-national borrowing to finance current expenditures should be prohibited. On the other hand, if the provinces and quite possibly large metropolitan area governments meet quantitative guidelines on fiscal discipline and are also rated as credit worthy by the private sector. could be permitted in the long run to have access to private capital markets at their own risk with clear understanding that the federal government will not backstop such debt. To ensure that the provinces do face market discipline and that their borrowing strategies do not run counter to the federal government macro-economic stabilisation objectives, the federal government may use the National Economic Council or another intergovernmental forum as an intergovernmental consultative panel on macroeconomic policy coordination and more importantly to disseminate information on provincial and metropolitan finances and capital spending strategies to the private sector. This body could also serve as a catalyst in encouraging the private sector in developing credit ratings of federal, provincial and metropolitan governments to facilitate bond and loan finance. Needless to say, unless credit worthiness of provincial governments are substantially enhanced with a strong own resource mobilisation effort, provincial access to capital markets would remain largely closed. Therefore, strengthening provincial-local revenue sources should receive priority attention. Access to private capital finance by smaller municipalities does not appear feasible in the near future. For such governments, provincially directed access either through municipal finance corporations run on commercial principles or special arrangements like the FINDETER experiment in Colombia might be examined.

Why the Road to Reform Remains a Field of Dreams?

We noted that the fiscal systems in Pakistan requires significant restructuring. In the past, such reform efforts have not made much headway. A number of factors impede the progress of reform.

Political Factors

In Pakistan, political instability and feudal interests have contributed to setting aside constitutional dictums and introducing a system of centralised governance. Pakistan has been under military rule for a major part of its existence and past military regimes did not accommodate decentralised decision-making. During the periods, political activities have been permitted, feudal influences have dominated the political system and these influences favour either a centralisation or provincialisation of authority. This is because, while Pakistan has experienced heightened urbanisation in recent years with over forty percent of the total population as urbanised, electoral system still recognises 1981 basis of population distribution (17 percent urban). In rural areas of Sindh and Balochistan, and to a more limited extent in rural areas of Punjab and NWFP provinces, feudal lords do not allow effective political participation. The use of antiquated basis of population allows feudal lords to dominate politics at the federal and provincial levels. A centralised system allows these lords to have greater effective control than would be possible under a decentralised system where the urban sector would have a more significant voice. To further entrench feudal powers, local governments were disbanded in all metropolitan areas even though the courts found this practice to be in contravention of the law. Grants to members of national and provincial legislatures for development projects also work against the development of local governments as these members enjoy a greater degree of autonomy in project execution in the absence of a well functioning system of local government.

Bureaucratic Factors

The British during their colonial rule instituted a system of bureaucratic control to achieve with maximum efficiency colonial objectives of a predatory state. The system created a core of civil service elite which was highly educated and highly dedicated to serving the colonial rulers. Their loyalty to rulers and detachment from the common man was duly rewarded by allowing them preferential access to all public services through elite institutions and by ensuring them financial security through a system of cash rewards and land grants. Thus Pakistan inherited a civil service regime that was highly centralised, efficient, accountable, professional yet completely detached from local population.

After independence, the civil service retained its "professional" orientation for a while benefiting from an increasing array of perks such as almost costless acquisition of prime real estate, free membership in sports and entertainment complexes financed by public funds, privileged access to elite educational institutions for children etc. Political purges of the civil service in late sixties and early seventies set the civil service on a road to administrative decline [see Haque (1996)]. With insecurity regarding the lack of tenure, areas of public intervention expanded beyond limits and perks and bribes mushroomed. Corruption enabled officials to insure their careers against political risks. Thus the system became a highly centralised yet dysfunctional system of administration. A key feature of this system has a special bearing on local governance. Key positions on provincial (provincial secretary) and district governance (divisional and district commissioners) are held on assignment by officers of elite "Central Superior Services--Civil Service of Pakistan". While performing these duties at subnational levels, they remain primarily accountable to the federal government only. This system negates federalism and re-enforces federal control over local decision-making. For local governments, an especially worrisome aspect of this system of governance is that if the local governments function well, district commissioner's powers are considerably curtailed. On the other hand, if the local government is not operative, the district commissioner becomes the sole discharger of judicial and executive functions at the district level. No wonder, one sees that local governments are not allowed to succeed in Pakistan.

Institutional Factors

Institutional factors also impede effective decentralisation. Traditional institutions and mechanism of governance and accountability over time have withered away but these have not been replaced by newer institutions. Instead, all pervasive role of the state have retarded critical look at public policies and institutions. There is almost complete monopoly of the government on institutions of critical thought and media in both countries. Any critical review of government policies and programmes invites a government backlash. In Pakistan, rural sell-government worked well in earlier days of its independence. This system was abandoned in favour of a more centralised system which has resulted in denial of access to basic services by rural population. While lack of institutional capacity was cited as a reason for disbandment of the participatory system, the newer system left a majority of citizen with no voice and participation and no access to basic public services.

External Participants

External participants may also unwittingly impede development of a decentralised public sector in developing countries. A multitude of factors contribute to this development. First, a centralised system lowers transaction costs for external assistance and enlarges the comfort zone for external participants in terms of monitoring the utilisation of their funds for intended purposes. Second, some external participants have concerned themselves primarily with the revenue performance of developing countries. Such concerns may lead to larger centralised bureaucracies that pay little attention to efficient delivery of public services. For example in Pakistan, improved revenue performance of governments have been accompanied by ever deteriorating quality and quantity of public services. Third, centralised systems are more prone to a lack of internal policy agenda due to a lack of citizen participation and more dependent on external advice on policy reform. Typically this leads to quick policy fixes with little sustained reform. For example, in Pakistan, while the 1956 Constitution stated achievement of universal literacy as a goal for the next decade. Forty years later, there has been little change in literacy rates. In population planning, with US assistance, Pakistan established a goal to reduce population growth rate to 2 percent by 1975. Twenty years later, growth rate has increased to 3.5 percent. Similarly public deficit reduction has been an elusive goal for the past several decades. External assistance, contrary to its intentions, may have helped the country avoid facing difficult choices in reducing public sector interventions in marketplace. Availability of generous external assistance might have played a part in motivating the federal government in assuming some provincial responsibilities and the provincial governments in overtaking local government mandates.

SOME LESSONS FOR DEVELOPING COUNTRIES

The following important lessons for reform of fiscal systems in developing countries can be distilled from a review of Pakistan's past experiences.

* To ensure fiscal discipline, governments at all levels must be made to face financial consequences of' their decisions. This is possible if the central government does not backstop provincial and local debt and the central bank does not act as a lender of last resort to the central government. A central bank that is independent of all levels of government and solely concerned with price stability is more likely to support such a discipline. The integrity and independence of the financial sector also contributes to fiscal prudence in the public sector. To ensure such an integrity and independence, ownership and preferential access to the financial sector, as continues to be the case in Pakistan should not be available to any level of government. In such an environment capital markets and bond rating agencies would provide an effective fiscal policy discipline. Finally, fiscal rules accompanied by "gate keeper" intergovernmental councils/committees provide a useful framework for fiscal discipline and fiscal policy coordination. In this context, one can draw upon industrial countries' experiences with 'golden rules', Maastricht type guidelines and "common budget directives' to develop country specific guidelines. To ensure voluntary compliance with the guidelines, appropriate institutional framework must be developed. Transparency of the budgetary processes and institutions, accountability to electorate and general availability of comparative data on fiscal positions of all levels of government further strengthens fiscal discipline. Pakistan's experience demonstrates that unless institutional arrangements for imposing such discipline on governments as discussed above are established, sustained fiscal discipline is not possible as centralised controls simply do not work.

* Societal norms and consensus on roles of various levels of governments and limits to their authority are vital for the success of decentralised decisionmaking. In the absence of such norms and consensus, direct central controls do not work and intergovernmental gaming leads to dysfunctional constitutions.

* Tax decentralisation is a pre-requisite for sub-national credit market access. In countries with highly centralised tax bases such as is the case in Pakistan, unrestrained credit market access by subnational governments poses a risk for macro stabilisation policies of the national government as the private sector anticipates a higher level government bailout in the event of default and does not discount the risks of such lending properly. However, without significant tax decentralisation, higher" level institutional assistance may be needed for financing local capital projects. This assistance can take the form of establishing municipal finance corporations run on commercial principles to lower the cost of borrowing by using the superior" credit rating of the higher level government and municipal rating agencies to determine credit worthiness.

* An internal common market is best preserved by constitutional guarantees. The Federal government in Pakistan has failed in this role.

* Periodic review of jurisdictional assignments is essential to realign responsibilities with changing economic and political realities. With globalisation and localisation, national government's direct role in stabilisation and macroeconomic control is likely to diminish over time but its role in coordination and oversight is expected to increase as regimes and subnational governments assume enhanced roles in these areas. Constitutional and legal systems and institutions must be amenable to timely adjustments to adapt to changing circumstances.

* Enabling Environment for Decentralisation i.e., Institutions of citizen participation and accountability must be addressed in arty serious reform of fiscal systems. Even in primitive societies such as pre-British India, systems of local governance worked effectively to deliver local services and collect local charges due to well understood mechanisms of citizen participation and accountability. More modern systems of local governance such as those run by elite Pakistani bureaucrats with training in management including financial management have failed due to a lack of citizen voice and accountability checks. The reform effort must embody appropriate provisions of recall of elected officials for negligence or misconduct. Independence of judiciary and a free media can play an important part in political and bureaucratic accountability. These elements have not been addressed in the Pakistani reform efforts.

* Civil Service reform is critical to the success of a decentralisation programme. Bureaucratic ownership of a reform programme is critical but such ownership would not be forthcoming in most developing countries where decentralisation is seen as an attempt to weaken the powers of central bureaucracy. To overcome this, the reform of fiscal systems must embody reform of central bureaucracies. Such a reform must ensure that the Centre has no direct say in the recruitment and promotions of civil servants, other than overseeing that standards of transparency and fairness are met, at the subnational levels and that remunerations of subnational services must be competitive with the central government. Further, civil service incentive structure should reward service orientation and performance and discourage command and control and rent seeking. This can be accomplished through performance contracts, stay with it culture, recognition of specialised skills and evaluation systems that link performance, rewards and budgeting.

* Evaluation capacity development is of fundamental importance in public sector reform in developing countries. Formal ex-post evaluation nurtures a climate of listening, learning and accountability in the public sector. This is of greater importance in the context of developing countries where "government failures" have been spectacular but resulted in little afterthought on appropriateness of development objectives and strategies as institutions for accountability are weak or non-existent. In such an environment, evaluation can nurture a "bottom-line" or "development effectiveness culture" as Picciotto (1995) has argued that "evaluation is to the public sector what accounting is to the private sector". It is interesting to note that the 1996 Constitution of the Republic of South Africa has imposed stringent monitoring and evaluation requirements on higher level governments to ensure proper monitoring and oversight of local governments.

* Traditional administrative capacity, matters but should not be considered as an impediment to decentralisation. Administrative capacity to develop and maintain modern organisational practices such as budgeting, auditing and accounting systems is no doubt important but should not be considered as a barrier to decentralisation provided citizen participation and transparency in decision-making is ensured. This is because technical capacity can be borrowed from supportive higher level governments and elsewhere

* Asymmetric Decentralisation as provided under the Provincial Local Government Ordinances in Pakistan offers a thoughtful approach to Decentralisation. Regardless of the availability of help from higher level governments, lack of institutional capacity should never be considered as an excuse not to decentralise. Instead, an objective programme of decentralisation which recognises the nature and type of local government, its Clientele and its fiscal capacity can be developed and various local governments can be assigned differential powers by taking into account the above mentioned factors as was done in Pakistan in the past and more systematically being done in Indonesia by rating each local government.

Improperly designed intergovernmental transfers can undermine fiscal discipline and accountability while building transfer dependencies that cause a slow economic strangulation of fiscally disadvantaged regions. Properly designed intergovernmental transfers on the other hand can enhance competition for the supply of public goods, fiscal harmonisation, subnational government accountability and regional equity. In Pakistan, federal revenue sharing transfers finance upto 99 percent of expenditures in some provinces. This delinking of taxing and spending responsibilities have led to accountability problems at the provincial levels. In the event of such de-linking, role of conditional (conditional on standards of services and access to such services and not on expenditures) block transfers is worth examining to enhance accountability and private sector participation.

* The role of fiscal transfers in enhancing competition for the supply of public goods should not be overlooked. For example, transfers for basic health and primary education could be made available to both public and not-for-profit private sector on equal basis using as criteria, the demographics of the population served, school age population and student enrolments etc. This would promote competition and innovation as both public and private institutions would compete for public funding. Chile permits Catholic schools access to public education financing. Canadian provinces allows individual residents to choose among public and private schools for the receipt of their property tax dollars. Such an option has introduced strong incentives for public and private schools to improve their performances and be competitive. Such financing options are especially attractive for providing greater access to public services in rural areas of Pakistan.

* Finally, contrary to a common misconception, a developing country institutional environment calls for a greater degree of decentralisation than needed for an industrialised country. For an efficient working of a centralised bureaucracy, advanced information gathering and transmittal networks, an efficient and dedicated civil service, and well developed institutions of citizen participation and accountability are needed. This is possible in the setting of an industrialised country environment. A more primitive public sector environment is more suited to a decentralised form of governance. This is because information requirements and transaction costs are minimised by moving the decision-making closer to people who are affected by those decisions. Closeness also serves to enhance better participation, preference matching for public services, transparency and greater accountability. The experience of Pakistan demonstrates that public sector performance is significantly improved by decentralised decision-making even when enabling environment is quite weak.

Appendix Table 1
A Representative Assignment of Expenditure Responsibilities

 Policy,
 Standards Provision/ Production/
Function and Oversight Administration Distribution

Interregional and
 International U U N,P
Conflicts Resolution
External trade U U,N,S P
Telecommunications U, N P P
Financial Transactions U,N P P
Environment U,N,S,L U,N,S,L N,S,L,P
Foreign Direct
 Investment N,L L P
Defense N N N,P
Foreign Affairs N N N
Monetary Policy,
 Currency, U, ICB ICB ICB, P
Banking
Interstate Commerce Constitution, N P
 N
Immigration U,N N N
Transfer payments N N N
Criminal and Civil Law N N N
Industrial Policy N N P
Regulation N N,S,L N,S,L,P
Fiscal Policy N N,S,L N,S,L,P
Natural Resources N N,S,L N,S,L,P
Education, Health and
 Social N,S,L S,L S,L,P
Welfare
Highways N,S,L N,S,L S,L,P
Parks and Recreation N,S,L N,S,L N,S,L,P
Police S, L S,L S,L
Water, Sewer, Refuse,
 Fire L L L,P
Protection

Function Comments

Interregional and
 International Benefits & costs international in scope.
Conflicts Resolution
 External trade ...
Telecommunications National regulation not feasible.
Financial Transactions ...
Environment Externalities of global, national,
 state and local scope.
Foreign Direct
 Investment Local infrastructure is critical.
Defense Benefits and costs national in scope.
Foreign Affairs ...
Monetary Policy, Independence from all levels essential.
 Currency, Banking some international role for common
 discipline.
Interstate Commerce Constitutional safeguards important
 for factors and goods mobility.
Immigration U due to forced exit.
Transfer payments Redistribution.
Criminal and Civil Law Rule of law, a national concern.
Industrial Policy To avoid beggar-thy neighbour policies.
Regulation Internal common market.
Fiscal Policy Coordination is possible.
Natural Resources Promotes regional equity and internal
 common market.
Education, Health and
 Social Transfers in kind.
Welfare
Highways Benefits and costs of various roads
 vary in scope.
Parks and Recreation ...
Police Primarily local benefits.
Water, Sewer, Refuse, ...
 Fire Protection

Note: U is supranational responsibility. ICB is independent
central bank, N is national government, S is state/provincial
government, L is local Government and P is non-government
sectors/civil society.

Appendix Table 2
Conceptual Basis of Tax Assignment

 Determination Collection
Types of Tax of Base and Rate Administration

Customs F F F
Corporate Income F, U F,U F,U
Resource Taxes
 Resource Rent
 (Profits/Income) Tax F F F
 Royalties, Fees,
 Charges: Severance
 Taxes: Production,
 Output, and Property
 Taxes S,L S,L S,L
 Conservation Charges S,L S,L S,L
Personal Income F F,S,L F
Wealth Taxes (Taxes
 on Capital, Wealth,
 Wealth Transfers,
 Inheritances, and
 Bequests) F F,S F
Payroll F,S F,S F,S
Multi-stage Sales Taxes
 (Value-added Tax,
 [VAT]) F F F
Single Stage Sales Taxes
 (Manufacturer/
 Wholesale/Retail)
 Option A S S,L S,L
 Option B F S F
"Sin" Taxes
 Excises on Alcohol
 and Tobacco F,S F,S F,S
 Betting, Gambling S,L S,L S,L
 Lotteries S,L S,L S,L
 Race Tracks S,L S,L S,L
Taxation of "Bads"
 Carbon F F F
 BTU Taxes F,S,L F,S,L F,S,L
 Motor Fuels F,S,L F,S,L F,S,L
 Effluent Charges F,S,L F,S,L F,S,L
 Congestion Tolls F,S,L F,S,L F,S,L
 Parking Fees L L L
Motor Vehicles
 Registration, transfer
 Taxes, and Annual
 fees S S S
 Driver's Licenses
 and Fees S S S
Business Taxes S S S
Excises S,L S,L S,L
Property S L L
Land S L L
Frontage, Betterment S,L L L
Poll F,S,L F,S,L F,S,L
User Charges F,S,L F,S,L F,S,L

Types of Tax Comments

Customs International trade taxes.
Corporate Income Mobile factor, stabilisation tool.
Resource Taxes
 Resource Rent Highly unequally distributed tax
 (Profits/Income) Tax bases.
 Royalties, Fees,
 Charges: Severance
 Taxes: Production,
 Output, and Property Benefit taxes/charges for state-local
 Taxes services.
 Conservation Charges To preserve local environment.
Personal Income Redistributive, mobile factor.
 stabilisation tool.
Wealth Taxes (Taxes
 on Capital, Wealth,
 Wealth Transfers,
 Inheritances, and
 Bequests) Redistributive.
Payroll Benefit charge, e.g., social security
 coverage.
Multi-stage Sales Taxes Border tax adjustments possible
 (Value-added Tax, under federal Assignment: potential
 [VAT]) stabilisation tool.
Single Stage Sales Taxes
 (Manufacturer/
 Wholesale/Retail)
 Option A Higher compliance cost.
 Option B Harmonised, lower compliance
 cost.
"Sin" Taxes
 Excises on Alcohol
 and Tobacco Health care a shared Responsibility.
 Betting, Gambling State and local responsibility.

 Lotteries State and local responsibility.
 Race Tracks State and local responsibility.
Taxation of "Bads"
 Carbon To combat -local/national
 pollution.
 BTU Taxes Pollution impact may be
 national, regional or local.
 Motor Fuels Tolls on federal/provincial/local
 roads.
 Effluent Charges To deal with interstate.
 intermunicipal or local pollution
 issues.
 Congestion Tolls Tolls on federal/provincial/local
 roads.
 Parking Fees To control local congestion.
Motor Vehicles
 Registration, transfer
 Taxes, and Annual
 fees State responsibility.
 Driver's Licenses
 and Fees State responsibility.
Business Taxes Benefit tax.
Excises Residence-based taxes.
Property Completely immobile factor,
 benefit tax.
Land Completely immobile factor,
 benefit tax.
Frontage, Betterment Cost recovery.
Poll Payment for local services.
User Charges Payment for services received.

Source: Shah (1994).

Note: U is supranational agency, F is federal. S is state or
province, L is municipal or local.

Appendix Table 3
Principles and Better Practices in Grant Design

Grant Objective Grant Design Better Practices

To bridge fiscal * Reassign Tax abatement in
gap responsibilities Canada and tax base
 * Tax abatement sharing in Canada, and
 * Tax base sharing Brazil

To reduce regional General Non-matching Fiscal equalisation
fiscal disparities Fiscal capacity programmes of
 equalisation Australia, Canada and
 transfers Germany

To compensate for Open-ended matching RSA grant for teaching
benefit spillovers transfers with hospitals
 matching rate
 consistent with
 spillout of benefits

Setting national Conditional non- Indonesia roads and
minimum standards matching block primary education
 transfers with grants Colombia, Chile
 conditions on and South Africa
 standards of service education transfers
 and access

Influencing local Open-ended matching Matching transfers for
priorities in areas transfers (with social assistance as in
of high national preferably matching Canada
but low local rate to vary
priority inversely with fiscal
 capacity)

Stabilisation Capital grants Limit use of capital
 provided maintenance grants and encourage
 possible private sector
 participation by
 providing political and
 policy risk guarantee

Grant Objective Practices to Avoid

To bridge fiscal Deficit grants
gap Tax by tax sharing as in
 India and Pakistan (till
 1996 only)

To reduce regional General revenue sharing
fiscal disparities with multiple factors

To compensate for
benefit spillovers

Setting national Conditional transfers with
minimum standards conditions on spending
 alone ad hoc grants

Influencing local ad hoc grants
priorities in areas
of high national
but low local
priority

Stabilisation Stabilisation grants with no
 future upkeep requirements

Source: Shah (199a).

Author's Note: The views expressed in this paper are those of the author alone and should not be attributed to the World Bank Group. The author is grateful to Drs Sadiq Ahmed, Shahid Amjad Chaudhry, and Sarfraz Qureshi, and Professors Robin Boadway, Melville McMillan and Richard Bird for helpful discussions.

REFERENCES

Bakshi, P. M. (1994) The Constitution of India. Delhi: Universal Book Traders.

Huther, Jeff, and Anwar Shah (1996) A Simple Measure of Good Governance and Its Application to the Debate on the Appropriate Level of Fiscal Decentralisation. World Bank, Washington, D. C. (Report).

Haque, Nadeem-ul (1996) Appropriate Reform Strategies in Pakistan. IMF, Washington, D. C. (Report).

Hassan, Aynul (1995) Optimal Mix of Health Sector Expenditures. Paper presented at the Conference on Resource Mobilisation and Expenditure Planning, Lahore.

Malik, Suhail, Safiya Aftab, and Nargis Sultana (1994) Pakistan's Economic Performance, 1947 to 1993: A Descriptive Analysis. Lahore: Sure Publishers.

Pakistan, Government of (1994) Good Governance. Report of the Task Force on Social Contract.

Pakistan, Government of (1997) Report of the National Finance Commission. National Finance Commission Secretariat, Islamabad.

Picciotto, Robert (1995) Putting Institutional Economics to Work. Washington, D.C. (World Bank Discussion Paper No. 304.)

Shah, Anwar (1994) The Reform of Intergovernmental Fiscal Relations in Developing and Emerging Market Economies. Washington, D.C.: World Bank.

Shah, Anwar (1997) Fiscal Federalism and Macroeconomic Governance: For Better or For Worse? In Kiichiro Fukasaku (ed) Decentralisation, Intergovernmental Fiscal Relations, and Macroeconomic Governance. OECD, Paris. (Forthcoming.)

Shah, Anwar, et al. (1995) Supporting Fiscal Decentralisation in Pakistan. World Bank, Washington, D. C. (Report).

Stone, Andrew (1995) The Climate for Private Sector Development in Pakistan: Results of Enterprise Survey. World Bank, Private Sector Development Department. Washington, D. C. (Report).

(1) Islamic Welfare Tax which is not deposited into the Consolidated Fund and does not form part of the budget.

Anwar Shah works for the World Bank, Washington, D.C.

Table 1
An Impressionistic Review of Selected Aspects of the Fiscal
System in Pakistan

Selected Indicators Pakistan

Population (end-1997) 140 million
Area 0.8 million square kilometers
Per capita GDP (1997) US$ 540
Fiscal constitution (de-jure) Decentralised federation
Fiscal constitution (de facto) Multi-tiered centralised
 federation
Sub-national governments 4 provinces, 15 municipal
 corporations, 156 municipal
 committees, 301 town
 committees, 40 cantonment
 boards, 118 district
 councils, 4565 union councils
Provincial government constitutional
 status strong
Local government constitutional status weak
Institutions of federal-provincial
 coordination strong
Institutions of critical thought numerous but ineffective
Financial market discipline on
 public sector non-existent
Fiscal rules for aggregate fiscal
 discipline non-existent
Public sector interference in
 financial markets extensive and unconstrained
Actual provincial control of local
 governments strong
Range of local government
 responsibilities limited
Citizen participation low (rural) to moderate
 (urban)
Bureaucratic orientation command and control
Political and bureaucratic corruption high and growing
Red tape high and growing
Central bank independence weak
Output performance contracts for
 civil service none
Private provision of public services extensive
Provincial fiscal autonomy strong
Local fiscal autonomy weak
Local administrative autonomy weak
Federal-provincial role in local
 public services provision extensive and unconstrained
Local government role in local
 public services provision constrained
Transparency and predictability of
 the system of central-provincial
 transfers excellent
federal-provincial conditional block
 transfers to set national minimum
 standards non-existent
Fiscal capacity equalisation to
 a standard non-existent
Provincial-local transfers non-existent
Quality and quantity of public
 services poor/unreliable
Overall quality of governance poor

Table 2
Federal Transfers to Provinces 1994-95--Summary Indicators

 Punjab Sindh NWFP Balochistan All

Operating Transfers
 (as percent of Gross
 Operating Expenditure) 82 84 99 99 87
Capital Transfers
 (as percent of Gross
 Capital Expenditure) 15 12 9 8 12
Total Transfers
 (as percent of Total
 Expenditure) 71 72 75 72 72
Development Loans
 (as percent of Gross
 Capital Expenditure) 31 26 18 17 25

Source: Shah et al. (1995).

Table 3
Vertical Imbalances in Pakistan, 1994-95 (Percent)

 Revenue Expenditure
 Share Share Surplus/Deficit

Tax Collection
 National 90.2 67.1 23.2
 Sub-national 9.7 32.9 -23.2
 Provincial 4.9 28.8 -23.6
 Local 4.8 4.1 0.7
 All Levels 100.0 100.0 0.0