\V iJ \ 7--H POLICY RESEARCH WORKING PAPER 1714 Fiscal Federalism A description of Bosnias current arrangements irn fiscal in Bosnia-Herzegovina federalism and the challenges proposed by the Dayton The Dayton Challenge system - as well as general lessors for the design oF fiscal federal systems in ethnically William Fox diverse economies. Christine Wallich The World Bank Europe and Central Asia Central Europe Department January 1997 POLICY RESEARCH WORKING PAPER 1714 Summary findings The challenge of fiscal federalism in Bosnia is perhaps rights be? How would the entities, in turn, be structured unique in the world. The Dayton talks held in October internally, and what would be their fiscal governance? 1995, immediately after a cease-fire, assumed a totally These questions were open in October 1995, when the blank slate: How would the new nation that emerged as international community worked with experts and a result of the peace talks be structured, fiscally? How political leaders to forge for Bosnia a new constitution would be the roles of the central state and of the two and the new fiscal system that would be laid out in it. subnational units (the "entities") that constituted it? How Fox and Wallich describe Bosnia's current would the three previously warring communities of arrangements in fiscal federalism, outline the unique Croats, Bosniacs, and Serbs work together to form a challenges that the Dayton system proposed, and draw central government, what would the entity governments lessons for the design of fiscal federal systems in look like, and what would their fiscal functions and ethnically diverse economies. This paper - a product of Europe and Central Asia, Central Europe Department -is part of a larger effort in the department to analyze fiscal federalism issues in transition economies. Copies of the paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contactYasminJiwa, room Hi 1-075, telephone 202-473-4848, fax 202-477-1034, Internet address yjiwa@worldbank.org. January 1997. (26 pages) The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paperare entirely those of theauthor. They do not necessarily represent the view of the World Bank, its Executive Directors, orthecountries they represent. Produced by the Policy Research Dissemination Center Fiscal Federalism in Bosnia-Herzegovina The Dayton Challenge William Fox Christine Wallich Fiscal Federalism in Bosnia-Herzegovina: The Dayton Challenge William Fox and Christine Wallich' Introduction The challenge of fiscal federalism in Bosnia is perhaps unique in the world: The Dayton talks held in October 1995 immediately after the cease-fire, began with a fully blank slate: "How would the new nation that emerged as a result of the peace talks be structured from a fiscal perspective?" "What would be the role of the central State and what would be that of the two subnational units (the "Entities") that constituted it?" "How would the three previously warring communities of Croats, Bosniacs and Serbs work together to form a central government, what would the Entity governments look like, and what would be the fiscal functions and rights of these Entities?" And then, "How would the Entities, in turn, be structured internally, and what would be their fiscal governance?" All these questions were open in October 1995, when the international community worked together with experts and political leaders to forge, for Bosnia, a new constitution, and the new intergovernmental fiscal system that would be laid out in it. The new system of fiscal federalism should, it was agreed, be able to withstand the stresses that would be a natural consequence of the centrifugal forces still present in the country, be economically sensible, and yet also obtain the consensus of the three parties who would live with and implement it. The purpose of this paper is to describe Bosnia's current arrangements in fiscal federalism, outline the unique challenges that the Dayton system proposed, and draw some lessons for the design of fiscal federal systems in ethnically diverse economies. I. The Dayton Fiscal and Governance Challenge Under the Dayton-Paris accords, Bosnia and Herzegovina (BiH) is set up as a single sovereign state composed of two Entities, a Federation (comprised of two major ethnic groups, and split between a Croat-majority area and a Bosniac-majority area) and the Republika Srpska. The Federation itself is structured as a three-tier fiscal system -- the Federation, cantons, and municipalities (see Diagram 1). While most cantons in the Federation are mono-ethnic, there are two multi-ethnic canLons, and many multi-ethnic municipalities. Republika Srpska is a largely mono-ethnic entity as a result of the war (although this could be partially reversed after the elections). It has a two-tier structure, comprised of the Republika Srpska and municipalities (see Diagram I). Freedom of movement is intended to prevail and right of return is to be respected under the Dayton Agreements. 'William Fox is a Professor of Economics at the University of Tennessee. Christine Wallich is the Country Director for Bosnia and Herzegovina at The World Bank. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors, and do not necessarily represent the views of the World Bank, its Executive Directors, or the countries they represent. Very useful comments and assistance to this paper were provided by Wei Ding, Sweder van Wijnbergen, Sebnem Akkaya, Luis Alvaro Sanchez, Jennifer Keller, Richard Bird, Francois Vaillancourt, Stanley Weiner. Prepared for the Conference on Fiscal Decentralization in Developing Countries, Montreal, Canada, September 19-20, 1996. Map: Bosnia and Herzegovina Ethnic Majority by Canton Republia Srpska ^ Federation Cantons IUna-Sana (Bihac Region) 2. Posava }'i3l1i 3. Tuzla-Prodrinja Bosniac 4. Zenica-Doboj 5. Gornjedrinski (Gorazde) cmat 6. Central Bosnia (Travnik-Vitez) 7. Neretva (Mostar-Konjic) Serb 8. West Herzegovina (Posusje-Grude) J 9. Sarajevo 10. West Bosnia (Glarnoc-l'omislavgrad As a result of the war, the unified fiscal system of the old Republic of BiH was broken up, and until recently there were three totally different fiscal systems in the country -- one in the Croat-majority area of the Federation, one in the Bosniac-majority area of the Federation and one in the Republika Srpska. While all three systems bore the imprint of their Yugoslav origins, the system in the Croat-majority area, came, over the course of the 4 years following Bosnia and Herzegovina's declaration of independence, to be harmonized with that of Croatia, while that of the Republika Srpska became harmonized to that of the Federal Republic of Yugoslavia -- Serbia/Montenegro. The Croat and Bosniac systems have now begun to be merged, as new Federation structures are being put in place. Customs structures, the tax administration, customs levies, excises and several other taxes are now largely harmonized. However, the divergence between the Federation and the Republika Srpska systems remains significant (Section III). 2 Post Dayton (1996) Structure of Governments in . osnui and Herzegovina State c f osn ia and Herzegovina [ Fed e r ati on j R e pu blika S rp sk a j C anton. (1 0s ) -Municipalities (64 M u n ic ip alitie s (7 3) Income and Spending Differences. There are significant income differences between ethnic groups in Bosnia and Herzegovina, and these differences are reflected in income differentials between the different parts of the Federation and betveen the Federation and the Republika Srpska. Recent (1995) estimates suggest a per capita income of about $500 in the Bosniac majority area of the Federation (accounting for some 2.5 million of Bosnia's total population currently in the country of 3.5 million 2), a per capita income of some $1800 in the Croat areas of the Federation (some 400,000 population) and a per capita income of $1000 in the Serb area - Republika Srpska (700,000 population). This disparity of per capita incomes suggests major differences across these areas in terms of tax bases and in taxable capacities, and, absent equalization policies, in terms of service provision levels. Even prior to the war in Bosnia, there were differentials in public service levels across the regions in BiH. The war, which broke the fiscal structure into three separate systems, has significantly exacerbated spending differences. In the Bosniac-majority area, for example, the pension level was 11.4 DM per month in 1995, while the scheme operating in the Croat-majority 2Bosnia's total pre-war population was 4.5 million, of which some three million have had to leave their towns and villages, with one million estimated to be refugees abroad, mostly in Western Europe. Another 250,000 people are dead or missing. 3 area paid pensions of about 65 DM per month. World Bank estimates put the damage from the war at over $20 billion (World Bank, 1996). However, the destruction of infrastructure and public utilities, as well as of the capital stock in the productive sectors, hit the Bosniac areas the hardest, since industry and most core infrastructure facilities had been strategically located in central Bosnia ever since the early days of Yugoslavia. The tax bases of the Bosniac-majority area and its ability to generate income have thus been significantly impaired.3 By contrast, the Croat-majority area's strategic location near the coast enabled this region to control trade into the interior of the country during much of the war, and despite its limited resource endowment, to generate substantial revenues from customs and other trade related levies. The Republika Srpska's economy has been badly damaged by the embargo, and is expected to recover only slowly. A Rolefor Equalizing Policies? These differences in taxable capacities and public service provision levels would suggest, other things equal, that from both an equity and efficiency perspective, inter-Entity (and intra-Entity, in the case of the Federation) fiscal equalization policies could have a constructive role to play in the new fiscal system. Such policies would prevent the large differences in access to fiscal resources currently available to the subnational Entities and the groups within them from being translated into large differentials in the supply of public services and infrastructure.4 From both perspectives, one might be especially concerned about those public services with major externalities to the nation as a whole, such as health and education, which are the domain, under the Dayton constitution (see section III) of the subnational Entities (or the cantons, in the Federation). Implicit in this view of equity and efficiency is that the relevant domain of concern (McLure 1994) for efficiency and equity-focused redistribution policies is the nation as a whole - - i.e., all of Bosnia. Under such a view, the population of Bosnia sees itself as "Bosnian," instead of as "Croat," "Bosniac" or "Serb," and supports policies that address inter-regional disparities. On the other hand, if people's primary allegiance is to the national groups within the Entities, there will be a more limited view of the domain of concern, and differences in the average incomes of groups within an Entity or between Entities, or in the average level of public service provision, may not be the predominant concern of these groups.5 A Strong Role of the Central State? The vast challenge of recovering from the destruction of war would also suggest that a strong central government, capable of maintaining macroeconomic stability (see Tanzi; Prud'homme) that creates a "common economic area" throughout Bosnia and Herzegovina and mobilizes resources for reconstruction through effective budgetary and foreign borrowing policies, could also bring substantial benefit. With the near- 3It could be argued that the concentration of industrial conglomerates, made non-viable due to transition shocks in Bosniac- majority areas, would have led to some decline in the tax bases of this area even without the war. 41n the short term, this is expected to be donor-financed, but as the economy recovers, a domestic contribution to infrastructure finance is anticipated. 5McLure (1994) has made these observations about the service provision differentials in the Russian Federation, and the issue of whether natural resource revenues "belong" to all Russians, or to the territories (oblasts) in which they are located. The same observations about preferences and tolerances for redistribution also apply to BiH. 4 complete destruction of the productive sectors' capital stock, and a major portion of the country's infrastructure destroyed, effective use of scarce domestic budgetary resources and foreign flows could not be more critical. However, as this paper will suggest, the legacy of war, and the legacy of the intergovernmental fiscal system in Yugoslavia, make a strong central government difficult. The State that has been created out of Dayton, far from being empowered and fiscally robust, is one with no taxing powers of its own, and therefore is dependent totally on the Entities for resources via "contributions" to the State budget (section III). Its spending authorities are virtually nil. From a macro perspective, its ability to manage a stabilization policy using fiscal instruments is highly constrained. The State's discretionary macroeconomic management tools are limited further by the fact that the central bank is to be run as a currency board for the next six years.6 With a weak fiscal authority at the center, the potential for regional equalization is also limited. And, after the national elections, the State may not borrow abroad or indebt itself, without the consent of the Entities. In a nutshell, Dayton set up a "bottom-heavy" system with most responsibilities and rights in the hands of the Entities. Clearly, setting up an intergovernmental system that makes sense, and in this fraught political context, is a major challenge. II. The Legacy of Yugoslavia The legacy of Yugoslavia has significant relevance for the design of Bosnia's intergovernmental fiscal system; some would say that it is the specter that still hangs over the design of such policies. Yugoslavia was always the most decentralized of the socialist economies, both in terms of its economic and fiscal management and its political structures. The republics had significant autonomy from the very beginning, unlike in the rest of the socialist world, and the power of the republics was further increased in the 1963 constitution, creating the basis for the emergence of centrifugal forces early on. Tax reforms in 1971 gave greater revenue raising powers to the republics and local authorities, and the 1974 Constitution transferred many functions to the republics, with "areas of common interest" remaining the responsibility of the state. These included military expenditures, administration, and economic interventions, including support to enterprises. Yugoslavia's six republics were highly differentiated in income level and cultures. Slovenia was by far the most prosperous, with a per capita income in 1990 of $6,500, with Croatia the next best-off. Bosnia and Herzegovina, Macedonia, and Montenegro were considered the least developed republics. Bosnia's per capita income in 1990 was only 29% of that in Slovenia (and Macedonia's only 22%). Further, the level of per capita budget expenditures in these republics reflected these large income differences. In the social sectors, for example, Slovenia's per capita social spending was twice the Yugoslav average, while social spending in 6There will be a one-time opportunity to issue currency when the currency board is established, from which the State may accumulate seignorage revenue. However, the revenue is expected to be limited, and cannot be relied upon as a repeating revenue source. While the possibility of seignorage revenue is intercepted by the establishment of a currency board, the State may earn some revenue through interest earnings on foreign exchange reserves. In the interim and until a new currency is in place, the seignorage from the circulation of the deutsche mark in the country accrues to the Bundesbank in Germany (and of the kuna to the Croatian Central Bank). 5 Bosnia and in Macedonia was 70% of the Yugoslavia average (OECD, 1992). These spending differentials were the result, inter alia, of different initial endowments and incomes, but equally important, the strong reliance, peculiar to the socialist world, on "derivation-based" systems of revenue sharing. Most fiscal revenues were notionally "owned" and retained by the areas in which they were collected, so that the socialist revenue systems typically benefited those areas with the more robust tax bases. Regional Redistribution Policies and Whto Gainedfrom Cross-Subsidies? Addressing the huge regional imbalances in income and expenditures was an objective of Federal policy very early on. Federal support to the republics came through two channels: a fiscal channel and a financial channel. The fiscal channel had several dimensions. The "Fund for the Development of the Least Developed Regions," established in 1965, introduced explicit support to the less- developed regions in Yugoslavia -- Bosnia and Herzegovina, Montenegro, Macedonia and Kosovo.7 At its peak, this fund mobilized some 2 percent of GSP -- from the better off regions who contributed a fixed proportion of their GSP to fund its resources.8 Federal budgetary resources were also directed to raising the level of social outlays in the poorer republics. The financial channel operated through the Republic-level commercial banking systems and the central bank,9 and contributed both to Yugoslavia's hyperinflation and external indebtedness. Many attempts have been made to estimate the overall magnitude and net effects of these fiscal and financial transfers between republics. Vodopivec's (1991) empirical estimate of the net redistribution among republics of all the subsidies and taxes suggests that the three better off republics (Slovenia, Croatia, and Serbia) were net "taxpayers" and the three less developed republics (Bosnia and Herzegovina, Macedonia, and Montenegro) were significant net beneficiaries. Such estimates, however, can only be suggestive at best, given that -- as in much of the socialist world -- the inter- republic redistributions through the fiscal and financial systems were, in the end, a series of ad hoc, bargained and very nontransparent agreements, whose effects and incentives were not at all well understood, and whose very murkiness gave rise to much dissatisfaction amongst Republics. Structural Vulnerabilities. The better off republics' dissatisfaction with Federal redistributive policies was compounded by another feature of Yugoslavia's intergovernmental fiscal system -- the "bottom up" system of intergovernmental revenue flows, which Yugoslavia shared in common with most other socialist countries, and which still holds in Bosnia and Herzegovina today (Bird, Ebel and Wallich, 1995). This system conferred a very special 7Kosovo along with Voivodina, which is largely Hungarian-populated, is now a dependent province within Serbia- Montenegro. 878 percent of Federal Fund resources were mobilized from the three better off republics in the 1970's, falling to 75 percent in the 1980s (Federal Statistical Office of Yugoslavia). 9The financial channel was effected in two ways. First, commercial banks in each republic could supplement their deposit resources by borrowing from their republic branch of the National Bank of Yugoslavia. Empirical studies suggest that the central bank did play a role in the regional allocation of credit through this mechanism, although it is difficult to get a consistent picture of who were the gainers and who were the losers. (World Bank, 1989). Second, the commercial banks in each republic financed the financial losses of enterprises which amounted to 5-6 percent of GSP in the latter part of the 1980's. The inflation (tax) this process gave rise to was felt by all, but the "benefits" of the system (financing of losses) were concentrated in Bosnia and Herzegovina, Croatia and Serbia (World Bank, 1989). 6 vulnerability on the Federal budget, since all tax revenues were collected at the republic level (or below) and transferred upwards to the Federal budget in agreed shares (the Federal government received all of the federal sales tax and all import duties). In addition, fiscal "contributions" from the republics, over and above those they made to the Federal Fund for Underdeveloped Regions, were negotiated annually, to balance the federal government budget. The Role of Fiscal Factors in the Break-up. There were many reasons for the breakup of Yugoslavia, but it would be incorrect to suggest that fiscal, quasi-fiscal and redistributional factors played no role. Redistribution policies of this scope were bound to nourish political separatism, as prosperous republics became more and more reluctant to release and transfer resources to poorer republics. The lack of transparency in the system made it possible for each republic to argue that it bore the costs of the intergovernmental system, while the benefits accrued to others, so that each thought the others gained, at its expense. Woodward (1995) argues, compellingly, that the centralizing budgetary and financial policies, designed to give the central government more control over stabilization policy, supported by the international community in the late 1980's, further exacerbated such tensions, since they reduced the budgetary autonomy of the better off republics and empowered the central government, fiscally and financially, with greater command over resources. The vulnerability conferred by the design of the system itself finally surfaced, as the three most prosperous republics -- Slovenia, Croatia and Serbia -- withheld federal tax revenues from the federal budget in 1991. It was with this history that the partners at Dayton began their work. Some lessons? What lessons can be drawn from Yugoslavia's breakup for the creation of fiscal federalism in Bosnia-Herzegovina? And what lessons would one hope to incorporate into the new structure that emerged from Dayton? There would seem to be several that represented major constraints: First, because of the Yugoslav legacy, the designers at Dayton had to take into account the legacy of suspicion of a strong central state and its perceived potential for abuse. This limited the role that could be given to national-level taxing powers or centralized spending. Second, the designers found that the notion of "ownership rights" over revenue collected on one's territory was hard to overcome -- this was perhaps even more so in the ethnic regions of Bosnia, recently emerging from war, than in Yugoslavia, where territorial claims to revenues were very strong indeed. Third, and related, the designers had to contend with a prima facie need, but apparent virtual intolerance for, fiscal transfers along the lines of the Federal Fund or other cross-subsidies between areas, and sometimes even within multi-ethnic municipalities. But, if no other fiscal lesson was learned from the demise of Yugoslavia, the vulnerability of the Federal fisc to transfers from below was clear to the designers of the new structure. And yet, the State that emerged had no revenue powers of its own. The vulnerability of the federal fisc to a decentralized tax administration was similarly evident. And yet, the bottom up system of revenue administration remained. This reflects the fact that the design on Bosnia's new system was driven very strongly by the interests of the Entities, whose agreement had to be obtained to the new structure. 7 III. The Dayton Rules Under the Constitution, major government budgetary responsibilities are divided among the four levels of government -- the State, the Entities, the Cantons and the municipalities (World Bank, 1996). Under the State Constitution, the State government has exclusive responsibility for foreign policy, foreign trade policy, customs policy, monetary policy, immigration, refugee and asylum policies, air traffic control, payment of international financial obligations of Bosnia and Herzegovina incurred with the consent of both entities, regulation of inter-Entity transport, communications, and international and inter-Entity law enforcement (but not the raising of an army, which is an Entity-level responsibility). All government functions not explicitly assigned by the State Constitution to the State are assigned to the Entities (the Entities may agree to delegate some of their responsibilities upward). The State government has no independent tax sources under the constitution (however, the Constitution provides that with the approval of the State Parliament, the State may levy its own taxes in the future), and is expected to finance its activities almost entirely from transfers from the two Entities. Two-thirds of the state's revenues are to come from the Federation and one-third from the Republika Srpska, to cover the budget approved by the State Parliament. Under the Federation Constitution (June 1996) the Entity governments have a wide range of expenditure responsibilities, including exclusive responsibility in their territories over defense (there are currently three armies: one in the Republika Srpska, and two in the Federation, one Bosniac and one Croat), internal affairs, police, environmental policies, economic and social sector policies, agriculture, industry, refugees and displaced persons, reconstruction programs and justice, tax, and customs administration. The constitution provides that some of these functions may be exercised jointly with the cantons, or separately, or by the Cantons as coordinated by the Federation Government. Thus, the Federation and cantons exercise joint responsibility for health, environmental policy, infrastructure for communication and transport, social welfare, and interior affairs, among others. The cantons (in the case of the Federation) are responsible for all other matters not granted expressly to the Federation. These include education, culture, housing, public services, local land use, and social transfer expenditures. Each canton is authorized by the federal constitution to delegate its responsibilities to the municipalities in its territory. The municipal governments in the Federation are granted "self-rule on local matters," including all responsibilities delegated to them by the cantons. In the case of municipalities with a majority population that is different from that of the canton as a whole, the canton must delegate all responsibilities described in the preceding paragraph to the municipal government. Municipal governments in the Republika Srpska are responsible for all matters not explicitly the responsibility of the Entity -- and largely parallel the combined responsibility of cantons and municipalities in the Federation. 8 IV. Can Dayton Work? Issues and Problems Can the Dayton Agreements work? More appropriately, is the Dayton structure the best for Bosnia, and if not, how should the structure be altered? These are complex questions to address, particularly since the equity implications associated with government structures, and the particular importance linked to cross subsidies in BiH, make it very difficult to determine what is the best system. The complicated, multi-level government structure arising from the peace arrangements reflects an ardent aversion to a strong State government that derives from experiences as part of Yugoslavia. Better off minority groups are particularly concerned that their rights will not be upheld, and as a result they envision being economically disadvantaged. The fear is that a strong central government would continue the Yugoslavian tradition of cross subsidies providing greater benefits to the less well off majority group. The war has further diminished the willingness of some nationalist groups to view BiH as a single country and to work for the better of the whole. Creating a government with most expenditure responsibilities and revenue raising power at the sub-national level was expected to limit the opportunity for transfers to the majority. The paranoia is sufficiently strong that favoritism of the majority is seen behind every program. For example, apprehension exists within the Federation that government-run economic development programs will be designed to support industries that predominate in the majority areas. The difficulty of creating governments in this environment can be seen in the relationship between the Bosniacs and Croats in structuring the Federation. Croat majority areas favor relatively strong canton governments and the Bosniac areas favor a relatively stronger Federation. The Croats, as the better-off, but minority population, want to continue decentralizing responsibilities to lower levels of government, where ethnicity and incomes are relatively more homogeneous, and therefore opportunities for cross subsidies are lessened. The Bosniacs argue for service delivery and revenue collection concentrated in the Federation where they might better be able to effect cross subsidies. Economies of scale would seem to indicate that many functions would be better performed at the Federation level given the relatively small geographic area and population size of the cantons. However, the discussions are heavily driven by equity and not efficiency. Less than one year has elapsed, so it is too early to judge fully the Dayton-Paris Agreements based on actual events. Still, some general problems can be seen. First, the tax and expenditure assignments are imprecise, leaving the potential for disagreements and requiring difficult political negotiations to reach accords. Even where there is greater precision in the rules, some of the assignments are inefficient and need to be reconsidered. Second, little room is allowed for cross subsidies to exist anywhere in the structure. This means large inequities can be expected to develop since economic conditions differ widely across the country. Third, most of the government structures appear potentially sustainable, even if not efficient. However, the State's roles are very limited, and its funding sources are precarious. As a result, the State's sustainability is in question. Finally, government's size in the economy after the many transitions is still to be determined. The remainder of this section provides a detailed 9 examination of specific issues in the Dayton rules. Imprecision of Expenditure Rules. Despite the appearance of clear guidelines regarding which government is responsible for delivering which services, the Dayton Rules are confused and create overlapping responsibilities. Not surprisingly, very different perspectives exist on what the final outcome should be. Overlap occurs when the State is charged with one aspect of service delivery and the Entities with another. For example, the State is to do "customs policy" and the Entities "customs administration." These assignments leave in question which level is to set tariff rates, which is to determine exemptions, which gets control over the revenues, and a number of other practical issues. Similar overlap occurs at the Federation and canton levels, which are assigned joint responsibility for human rights, health, environmental policy, transportation and communication infrastructure, social welfare policy, immigration and asylum, tourism and natural resources. Little is said about municipal responsibilities. The relatively centralized structure already in place will remain in the Serb Republic. In the Federation, the issue is what are cantonal versus municipal responsibilities. Municipalities are granted self rule and also are permitted to perform functions delegated by the cantons. The probable outcome is a different relative set of duties in each canton, given their geographic and demographic diversity. Differing patterns of responsibility will reflect varying sizes, ethnic compositions and political strengths in the cantons. No problems should arise from assigning provision to different government levels as long as governments are creative in achieving efficiency in production (see page 20). A series of other specific problems with the expenditure rules can be identified. In some cases there is no clear assignment of functions. Pensions are such an example, though pensions might be categorized under one of the general functions listed in the Dayton Agreements. In other cases, assignments are politically difficult, such as having defense delivered by the Federation. In a third set of cases, the assignments may be economically inefficient; higher education being set as a cantonal rather than a Federation responsibility is such an example. Expenditure responsibilities being imprecise or being assigned to the wrong level of government does not require the assignments be inefficient, since governments could negotiate efficient arrangements, such as cooperative service delivery. However, the recent history of events in Bosnia make such negotiations particularly difficult, meaning the likelihood is for economic inefficiencies over the short to medium term. Expenditure Policy. The war arrested the transition to a market economy and the corresponding relative reduction of the public sector. The war years led to ad hoc and very distorted budgeting and expenditure practices in the Bosniac, Croat, and Serb areas. Expenditures were heavily focused on military purposes, with about 45 percent of each Entity's budgeted expenditures going for military uses. Data are unavailable on the total extent of military expenditures because large off-budget military purchases and spending financed by foreign governments occurred as well. Budgets for non-military purposes only were provided as necessary and as resources were available. Actual expenditures were made essentially on a cash basis. Governments made payments when resources were available, and suspended payments 10 when they were not. Pensions and wages -- for military and non-military workers -- often went unpaid. Actual factor payments bore little relationship to market value. Expenditure rationalization is an important area for reform. The size of the public sector needs to be limited to make room for private sector development, and the structure of public expenditure programs needs to be improved to facilitate growth, while maintaining an acceptable level of social protection. Public expenditures currently represent more than 40 percent of GDP, and this, where there are substantial public sector arrears in wages, pensions, military, and debt service. Reduction of overall public sector expenditures to sustainable levels as a percentage of GDP will require major cuts in many areas of public expense; key areas for reduction are defense, subsidization of the economy, and entitlement programs. Pension payments cannot be based on pre-war earnings and expectations, and must be scaled back to a level that is consistent with reasonable wage tax rates. (Tax rates are currently in excess of 90 percent on net wages.) A rationalization of public expenditure programs will necessarily be required to accompany this process of downsizing the government, as well as to providing better targeting of public expenditure. Highest priority must be given to public expenditures providing a framework that allows the private sector to flourish -- appropriately selected infrastructure is essential. Subsidization of outdated, unproductive public enterprises must be given the lowest priority. Great care must be taken to avoid building large administrative structures in the public sector -- a potential problem that can be exacerbated by the large number of overlapping governments. Early experience evidences a propensity for cantons to overstaff their central offices with political appointees. For example, Tuzla Canton already has over 300 employees performing administrative functions. These practices must be stopped at the canton level and avoided in all governments. Efforts to rationalize salaries for government employees have resulted in large percentage salary hikes during 1996, expanding the budgetary size of administrative staffs. The raises, mandated by central governments, have placed many cantons and municipalities in the difficult position of having insufficient resources to finance their expected salary payments. As a result, employees in some areas receive raises and in others do not. Finance and Tax Administration. The Dayton Agreements offer fewer specifics about financing arrangements after 1996 than about expenditures. Three general guidelines are given: the Entities and their constituent governments are given responsibility for finance and tax administration; the State is to receive two-thirds of its funding from the Federation and one-third from the Serb Republic; and there is to be a Federation Tax Administration. All three create significant issues in practice. Financing of the State. The Constitution indicates that the State is to adopt a budget for the expenditures necessary to finance its responsibilities. No explicit own source revenues are provided for the State, but the State can generate revenues through charges such as passport fees. Remaining financing is to be provided two-thirds by the Federation and one-third by the Serb Republic, a bottom up arrangement, reminiscent of the former Yugoslavia. The State's sustainability is a concern in this environment, where a strong degree of distrust exists. No limitation is placed on State spending, other than that expenditures are to be the ones necessary to finance its responsibilities. The financing structure differs from a conventional grant system, since the recipient is telling the grantor how much to provide. Disagreement between the State and the Entities regarding what are necessary expenditures should be expected, and the Entities are likely to balk at times about paying their share. Difficulties with funding of the United Nations may be an appropriate parallel. Tax rates will need to be higher in the Serb Republic than in the Federation to generate the respective revenue contributions because of the greater aggregate population and incomes in the Federation. As a result, pleas of unfairness can be expected. The State may be less accountable for spending revenues raised by the Entities. Also, variability in revenue flows will be borne solely by the Entities, which in principle are unable to reduce contributions to the State to help lessen effects of unexpectedly low revenues. There are two sides to this coin, since Entities do benefit if better than anticipated revenues result. Federation Finance and Tax Administration. Tax structures, emanating from the Yugoslav tradition, are similar in each of the ethnic areas. The Bosniac structure as of November 1995 is given in Table 1. Ownership of tax revenues, control over tax rates and bases, and tax administration are separable. Division of these, with say the Federation administering taxes that are owned by both the Federation and cantons, normally is based on conditions that may not be fully in place at this point. A level of trust must exist, so that the government owning revenues is certain that the administering government will transfer the appropriate amounts in a timely fashion.10 In some cases, local offices of the Russian State Tax Service are passing resources to the regions and municipalities and giving the Russian Federation government what remains, a pattern that is the Bosnian concern. Cantons fear Federation collection will be to their detriment and the Federation feels the same in reverse. The owning government also must feel that the administering government will make a good faith effort to collect revenues. Pressures and incentives in the administering government may be to focus more of its limited resources on collecting revenues that it can keep, than on taxes going to other governments. Allowing the administering government to retain part of revenues as a collection fee can help, but it still gets more revenue benefits from collecting taxes that it totally retains. In such a small country, administrative and compliance efficiencies should arise from Federation (or State) control over most tax rates and bases, since uniform rates and bases will be more difficult to achieve with devolved control. However, political incentives can cause Federation officials to make concessions on tax bases where the revenues are owned by the cantons. Granting concessions against another level of government's tax bases offers all of the same political gains but fewer costs because the concession granting government does not have to face the budgetary consequences of its actions. Tuzla has already complained that concessions given to reduce the electric company's wage taxes have significantly reduced Tuzla's receipts. 10 I'he peculiar Yugoslav institution, the "SDK," the Social Accounting Office, which exists in a somewhat truncated form in Bosnia and Herzegovina today, makes it possible to pinpoint closely where consumption and employment and production take place and where taxes should be collected. So far this system has been insufficient to generate confidence in the governments getting their appropriate shares. 12 Again, willingness to entrust control to another level of government requires a strong degree of trust, despite the incentives and history creating a perception that the trust would be misplaced. Separate tax administrations, with offices located in each municipality, currently are operating in the Bosniac and Croat areas. A significant next step in development of the Federation must be creation of a Federation Tax Administration, as called for in the Dayton Agreement. Legislation was recently passed establishing most tax collection at the Federation level, but the Tax Administration has not become operative as yet. A director and deputy director of the tax administration have been appointed, office space has been acquired and staffing plans have been developed. TABLE I Tax Structure for the Republic of Bosnia and Herzegovina, November 1995 Tax Type of Tax (Description) Rates 1. Taxes on Individuals and Corporate Profit 1.1 Taxes on Individuals Schedular taxes on certain components of income are supplemented by a further tax on gross personal income exceeding a specified threshold. a) payroll tax Withheld by employer 10 percent b) farming tax 5-10 percent c) self-employment tax 36 percent d) tax on income from property Progressive rates depending on the type of income e) tax on total personal income Progressive rates, scale from 5 percent to 50 percent 1.2 Republic reconstruction tax Base is net wages; withheld by employer 10 percent 1.3 Tax on Corporate Profit The tax is paid by enterprises and other 36 percent legal entities realizing profit, wilether or not domiciled in the republic 1.4 Tax on Games of Chance Tax paid by individuals 15 percent 2. Social Security Contributions 2.1. Employees 2.1.1 old age insurance The gross wage is the base for payment 17 percent of the contribution 2.1.2 health insurance 14 percent 2.1.3 unemployment insurance 2 percent 2.2 Employer 2.2.1 old age insurance The gross wage is the base for payment 1 7 percenlt of the contribution 2.2.2 health insurance 14 percent 13 2.2.3 unemployment insurance 2 percent 3. Tax on Property 3.1 Tax on Real Estate Tax paid by legal entities and individuals 15 percent Transactions at the time of a real estate transaction 3.2 Gift and Inheritance Tax Paid on real estate and personal property 0 percent to 12 percent paid over a certain value by second degree, third 10 percent more, others 20 percent more 3.3 Motor Vehicle Tax Tax paid by individuals on vehicles less 1500 dinars per year than five years old and engine displacement of 1.6 liters 3.4 Boat Tax Paid by individuals on boats exceeding 5 900 dinars per year meters in length 3.5 Sign Tax Paid by legal entities and individuals on a DM 50 to DM 250 annually sign displayed 3.6 Tax on Weekend Cottages Paid by individuals on the construction from .10 to I percent on the value of the structure value of construction 4. Domestic Taxes on Goods and Services 4.1 Sales Tax on Products Tax paid on final consumption, but on 5 percent to 20 percent alcoholic and nonalcoholic beverages, beer, tobacco products, coffee, petroleum and petroleum products, paid on first sale 4.2 Sales Tax on Services Paid by legal entities and individuals on 10 percent services subject to the tax 4.3 Sales Tax on Surcharge A general surcharge on goods, levied on 10 percent the sales-tax inclusive price 4.4 Excise Taxes 4.4.1 Tax on Petroleum Products Paid by the producer or importer per liter DM .08 to DM .40 depending on the product 4.4.2 Tobacco Tax Paid by the producer or importer DM 0.22 to DM 2.30 per pack 4.4.3 Beer Tax Paid by the producer or importer DM 20 to DM 30 per hectol iter 4.4.4 Alcoholic Beverage Tax Paid by the producer or importer DM 8 to DM 10 per liter of pure alcohol content 4.4.5 Nonalcoholic beverage Tax Paid by the producer or importer from DM I to DM 20 per hectoliter 4.4.6 Tax on Imported Paid by the legal entity or individual on a DM 1900 to DM 8000 Automobiles new imported automobile exceeding power of 75 KW 4.4.7 Tax on Imported Coffee Paid per kilogram by the legal entity or from DM I to DM 4 per kg individual who brings coffee into the republic 5. Taxes on International Trade 5.1 Import Duties from 0 to 21 percent Source: International Monetary Fund. 14 Structural conditions in BiH suggest that grants should be an important financing source for cantons and municipalities. One reason is that any assignment of revenue ownership is likely to lead to both vertical and horizontal imbalances because data for estimating expected tax collections are poor and the weak economies in some areas will leave them unable to provide minimum service levels. Some mechanism, presumably operating through a grant system, needs to be in place to smooth financing problems during the interim as assignments are finalized and regional economies rebound. Also, a system of grants could be used to exploit economies of scale in tax collection, enhance equity in revenue distribution, and account for geographic externalities in consumption. However, the very limited tolerance for cross subsidies between geographic areas suggests that grants must play a small role in overall finance during the near term. Unlike transfers made during the Yugoslavian years, any grants that are introduced must be transparent and rely on objective indicators. A system of grants based on negotiated formulas or ad hoc procedures would be particularly inappropriate for Bosnia and Herzegovina, given the lingering distrust still prevailing among different groups in society. Debt Policy. The IMF and World Bank require that their lending be to the internationally recognized government, which is the State, but no agreement has been reached on how external debt service is to be financed. The Constitution says that "Each Entity shall provide all necessary assistance to the government of Bosnia and Herzegovina in order to enable it to honor the international obligations' l Like many Dayton rules, implementation of this statement is unclear. One possible interpretation is that each Entity is to make contributions according to the two-thirds Federation, one-third Serb Republic rule. However, blind adherence to this rule is likely to generate controversy that could lead to one Entity refusing to pay its share. At this point, even getting agreement from the Entities to borrow through the State has been very difficult. Determining a means for financing debt service is more complicated. The Dayton-Paris Agreements allow little flexibility for macroeconomic policy, but the governments may seek to operate at a deficit, both to ease financing problems and to stimulate the economy. Some examples are already apparent. The original 1996 Republika Srpska budget assumed a deficit equal to one-third of spending. The 1996 State budget included authorization for borrowing and for letters of guarantee. In practice, the ability to borrow is extremely limited. The Republika Srpska budget did not pass when it was recognized that there was no means of financing the deficit. Thus, with the Central Bank operating as a currency board and a weak banking sector, concerns about deficits may be unimportant in the near term. Governance During 1996. The Dayton Agreements were drafted with expectations that the complete, anticipated government structure would not be operative throughout 1996. In any event, a transitional period is necessary because time lines in the Agreements were impractical, a difficulty exacerbated by political problems in getting legislative approval for many facets of the Agreements. Revenue assignments in the Federation for 1996 were negotiated between the IMF, the World Bank and the government. The State was to receive a share of customs tax revenues. The ''Constitution of Bosnia And Herzegovina, Annex 4. Article 1II, Section 2(b). 15 Federation was given access to excise taxes and the share of customs taxes not passed to the State. Other taxes in the Federation were to go to the cantons. To date, the State has performed minimal operations, in as much as broad based acceptance of the State's role was hampered in the pre-election period. The presidency was elected on September 14, 1996, but the government has yet to be formed. The State budget is a relatively small DM 61.4 million for 1996, and is mostly used for wages to operate Ministries and set up costs for providing diplomatic functions and issuing documents. The State has been receiving about 30 percent of customs tax revenues from the Federation, but there has been considerable controversy over how much it should get. The Republika Srpska is not contributing to State finance in 1996. The structures of the new State government was formed 12 weeks after the elections of September 14, 1996, but its members have not yet been identified, as of early December 1996. It is expected that the State budget and other macro institutions will be established following the appointment of this new Council of Ministers. The Republika Srpska government structure was not altered by the agreements, so no transition has been necessary. However, the initial budget proposed by the Ministry of Finance was 50 percent above expected revenues. The Parliament sent the budget back for adjustment, and the government has been operating under an interim budget. Spending was limited to available funding during the first three quarters, and revenues are running about 15 percent below estimates. The Serb Republic had an explicit deficit equal to about one percent of GDP in 1995, and a payment to cover some of the 1995 shortfall is included in the 1996 budget. The majority of revenues are collected at the Entity level, and shares of the revenue are transferred to finance municipal functions. Banja Luka has argued strongly that for political reasons it is receiving less than its legislated share of taxes. The Federation has not begun to operate as a full fledged government. Ministries have been formed and Ministers and some staff appointed, but only the Federation Customs Bureau is fully operative. A combined payments bureau has just begun to function. Most responsibilities are still being performed by the former governments in the Bosniac and Croat majority areas, or by cantons and municipalities. Tax revenues continue to be collected by the former Bosniac and Croat governments, with some of the revenues being transferred to the Cantons and municipalities. The Federation has been receiving customs and excise tax revenues since May. By the end of September, the Federation had received income of DM 183.9 million and had expended DM 173.4 million. Only DM 13.0 million has gone for salaries and materials. All 10 cantons had been formed by the time a law establishing their geographic boundaries was passed in June. However, most of the cantons have elected a legislature and president, and are establishing a working framework that includes constitutions. Only the cantons including Tuzla, Zenica, and Bihac can be described as functioning. Organization of these three was well underway before peace, in part because these regions were cut off from Sarajevo during the war and had, de facto, to set up their own institutions. 16 V. Options for the Future It is difficult to be fully prescriptive about the best government structure for a country where so much transition is under way. The complexity of making such prescriptions is exacerbated by the recent history, that includes being part of Yugoslavia and participating in the war, and by the various agreements that already have been made. Still, a series of recommendations that would enhance the public sector's operational efficiency can be made. Shifting Functions to the State. A number of functions could be managed much more effectively at the State than at the Entity level. The Entities would be well served to identify where efficiency gains are likely, and to seek agreement to transfer the functions to the State. Examples are administration of value added (should one be legislated) and enterprise profits taxes and of customs administration. The Entities could delegate the functions to the State, but both Entities would need to agree. Three conditions need to be in place for delegation to occur: (1) significant efficiency gains must result, (2) transfer must be Pareto Optimal and (3) there must be reasonable security that the agreements will not be violated. Nonetheless, political discord, fears that agreements will not be adhered to, and envy could lead the Entities to reject efficient delegation. Coordination Between the Entities. Open borders that allow the free movement of goods and services and permit the economy to expand are a requirement of the Dayton agreements, but oblige the Entities to coordinate in setting tax administration and policy. For example, a value added tax cannot be easily imposed at a sub-national level without internal borders, but the borders would violate the Dayton agreements. Even a tax or boundary check between the Serb Republic and the Federation is inconsistent with rapid economic growth, and creates ample opportunity for abuse. One form of abuse would be imposing a recording fee at the checkpoint that could quickly be escalated to effectively become a customs duty. The potential for abuse is equally large for customs administration. Close coordination must substitute for borders to ease administration of these taxes, unless administration is delegated to the State. Given the open economy (even at the borders with Croatia and Yugoslavia), the broad tax policy of each Entity also must be carefully coordinated. Differentials between areas in customs, consumption, and capital tax rates13 provide significant opportunities for tax avoidance and evasion and will result in concentration of imports, production and sales in the lowest tax areas.'4 The best way to limit tax competition and the resulting resource migration is for the Entities to set tax rates on mobile factors, consumption, and customs within a small range of each other. The compensation could be informal, based simply on observing the other Entity's tax policies, or explicitly, through negotiations. More efficient tax policy is likely to result from explicit cooperation, but today there is no mechanism in place where the dialogue can occur. Coordination between the Serb Republic and Yugoslavia and between the Federation and Croatia is already well underway. 13The Dayton Agreements can be interpreted as requiring common customs tax rates, but the Entities are free to set the other tax rates, 14Labor migration in response to tax differentials is not a significant issue in the current environment, given the absence of freedom of movement across entities. 17 Size of Government. Government must be reduced substantially to a size that is appropriate for facilitating the private sector's operation, providing a sustainable, reasonable safety net, and ensuring delivery of services that are essential to an acceptable quality of life. The multi-layer government structure raises the challenge of attaining a smaller size while achieving public sector goals, but emphasizes the importance of being vigilant in seeking efficiency in government. In recent years, budget control has been exercised by limiting spending to available revenues, with no expectation that planned expenditures were to be financed. Budgetary practices need to be rationalized to where expenditures are set to meet appropriate goals, with the expectation that all expenditures, including wages and pensions, normally will be met. Rationalization of spending will require that expenditures for defense, the safety net, and subsidization of the economy will be reduced. In addition, the budget cannot be expanded to accommodate arrears. A number of outstanding claims on BiH arise from frozen foreign exchange deposits and unpaid pensions and wages. The total value of these claims is more than DM 13 billion in the Federation, or about 10 years of current own source revenues. Recently, foreign exchange bank books worth nearly DM 5 billion were issued allowing former Republic of BiH soldiers to make explicit claims against the Republic. The recognition of explicit claims on a case by case basis must be curtailed, and previous decrees and laws repudiated. It is essential that any settlement of these liabilities occur using revenues from the sale of assets (such as could arise from flats and enterprise privatization) and that no mechanism for settling liabilities have implications for the budgets of the State, Entities, cantons or municipalities. Political and equity based concerns will be the key factors in deciding the relative compensation paid for these claims.'5 Budgetary savings can be obtained through more efficient operations. First, better delivery mechanisms must be identified. Diseconomies of small scale must be overcome through creative means of consolidating certain functions. The State or inter-entity cooperation should be used wherever significant efficiency gains are possible. Similarly, cantonal or municipal cooperation could allow the scale necessary to efficiently deliver certain services. Health care, higher and secondary education, transportation, and environmental services are obvious areas for cooperation. Privatization of service delivery is an option that permits both the benefits of private sector incentives and the capacity to expand the geographic area for service provision. Second, public sector employment must be scaled down to acceptable levels. In certain circumstances public employment can be a stopgap for job shortages, but it cannot serve as a long term strategy for job creation or as the safety net. Public sector employment must be determined by the number necessary to efficiently deliver services. It is essential that administrative staffs be streamlined, and every service be examined for potential areas to reduce. Assigning Expenditure Functions in the Federation. Four criteria are used for '5These claims do not include the international debt of BiH. 18 recommending the optimal assignment of public functions between the Federation, cantons, and municipalities: (1) the potential for economies of scale, (2) the existence of interregional spillovers, (3) variation in preferences, and (4) the desire for equalization and tolerance for cross- subsidies. A recommended assignment of functions within the Federation (and in some cases including the State) based on the tradeoffs among these criteria is shown in Table 2. The Federation should be responsible for areas mainly with Federation-wide impacts. The Constitution has already assigned several functions to the Federation, including defense, federal police, justice, and customs administration. It would be reasonable to assign additional functions to the Federation, including university education, some medical care, intercanton transport, environmental control, and part of social welfare. Cantons and municipalities will have responsibility for many functions including much of health care, education, housing, fire protection, and utility services. Despite the substantial responsibilities assigned to cantons and municipalities, Federation expenditures are a significant share of total (see Table 5 below). Agreements should be reached between cantons and their constituent municipalities regarding the specific sharing of responsibilities. The Constitution requires that municipalities that are predominantly composed of a minority population in the canton be permitted to deliver services locally. A key problem is determining the revenues, that otherwise would go to the canton, that are to flow with the expenditures. Minority municipalities must receive a revenue allocation that legitimately allows them to meet their service responsibilities, without creating unreasonable fiscal strain on local citizens. Careful design of the revenue arrangements is important because Ethnic problems often arise locally and frequently can be solved if much of the power for decision making is local (Tishkov, 1993). Cantons could devise transfer schemes that make it very difficult for minority municipalities to finance service delivery, which could lead to strong disagreements. Tax Assignment in the Federation. As a rule, canton and municipal revenues should be based on a combination of assigned revenue instruments and grants, with the goal of ensuring vertical balance -- meaning that each level of government would have sufficient revenue capacity to fund the services for which it is responsible. Furthermore, tax and grant assignments must be made with an eye on both the funding of current service levels as well as future service needs. That is, revenue buoyancy must be carefully considered when assignments are made. Each level of government -- the Federation, cantons, and municipalities -- must have sufficiently buoyant revenue sources to provide the revenues needed to pay for the delivery of services over time. 19 TABLE 2 Recommended Assignment Of Public Functions Service Category Type of Service Level of Government Heath Care: Primary Municipality Secondary (e.g., hospitals, curative) Canton Tertiary (infectious disease, research) Entity Education: Primary Municipality Secondary Cantonal/Municipality University Entity Transportation: Roads/Highways (intracity) Municipality Roads/Highways (intercity) Canton/Entity/State Airports Entity Public Transportation: Intracity Municipality Public Transportation: Intercity Entity Private Transportation, Taxis Canton/Municipality International State Environmental Air/Water Pollution Canton Water/Forestry Entity/State Housing ALL Canton Solid Waste, Water, Sewer, Fire ALL Municipality/Canton Land Use/Zoning ALL Municipality Licensing/Regulation ALL Canton Cultural Policy ALL Canton Tourism ALL Canton Social Welfare ALL Canton/Entity Telecommunications ALL Entity/State As noted above, it is important to differentiate among the separate steps of the taxation process when making tax assignments. These steps include tax administration; definition of tax bases; setting of tax rates; and tax ownership. The level of government in charge of collecting a tax need not be the same level that defined the tax base, set the tax rate, or owns the tax revenues. Locally owned taxes (cantonal or municipal taxes) need not be locally administered, and federally imposed and collected taxes can be-- and, in other countries often are-- shared with other levels. Alternatively, taxes could be collected at the canton level and shared with the Federation. In fact, it is recommended that for the most important taxes, definition of the tax base and possibly setting of the tax rate be done at the Federation (and ultimately State) level, while allowing cantons and municipalities some flexibility over revenue instruments. Economic integration, tax compliance, and tax administration would be greatly enhanced by nationally uniform bases and rates for customs, excise, corporate and personal income, and sales or value- added taxes. However, cantons must be protected from the Federation exempting taxpayers or narrowing the base for taxes where the Federation has little or no revenues at stake. A proposed set of assignments for the Federation is in Table 3. 20 Three criteria must be weighed in determining the appropriate tax ownership assignments. First, tax assignment must bear a strong correspondence to the expenditure responsibilities of each government level to ensure that adequate financing is available to deliver required services. A perfect relationship between expenditure responsibilities and tax revenue ownership is not required, and probably is not desirable, because grants can be made from a government that owns revenues to assist another government in financing its expenditure responsibilities. Second, tax assignment must not violate the Constitution, Dayton Pacts and other international agreements. Third, tax ownership must be consistent with the administrative ability to measure where the taxable activity occurs, if revenues are to be distributed on a sites of collection basis. For example, ownership of customs taxes based on where the imports are consumed is administratively impractical at the municipal or canton level. The Federation needs to take a number of important steps before tax ownership can be determined precisely and implemented effectively. Expenditure assignments must be structured before final tax assignments can be made. Also, the full set of Federation tax laws need to be legislated and enforced. A mix of laws from the Federation and old Croat and Bosniac areas currently are being enforced on the Federation side. Tax assignments that would result in vertical balance given expected 1997 expenditure assignments were estimated (see Tables 4 and 5). 6 The assignments assume that no intergovernmental transfers occur, other than those implicit in the assignments. The State needs to receive 7.4 percent of Federation and 17.3 percent of Serb Republic customs tax revenues to finance its expenditures. The difference in shares illustrates the effect of having one-third of the State financed by the Serb Republic and the other two-thirds financed by the Federation. No other detail was estimated for the Serb Republic. In the Federation, the remainder of customs taxes and excise, special import and enterprise profits taxes are assigned to the Federation. Wage income and profits taxes are assumed to be split evenly between the Federation and the cantons.17 The cantons and municipalities receive the range of smaller taxes including the real estate, judicial and advertising taxes. The sales tax is assigned as a shared tax, with 64.5 percent going to the Federation and the rest to the cantons and municipalities. 16Revenues and expenditures were estimated separately. Expenditure requirements were based on budgeted and actual 1996 expenditures and assumed growth rates for the State and Federation. Additional transfers to the State are necessary if the State is to remit debt amortization payments. Canton and municipal expenditures were estimated using actual 1995 expenditures for the various services, and assumed growth rates. Revenue estimates for 1997 are based on actual collections in 1995, the first nine months of 1996, and assumed growth rates. No attempt is made to estimate revenues separately for municipalities versus cantons in the Federation. Defense, social programs and debt service are used to balance expenditures and revenues and do not include off-budget expenditures. '7The split wage tax could be designed as a piggybacked wage tax, as has been proposed by the IMF. 21 TABLE 3 Summary of Suggested Tax Assignments Tax Type Tax Tax Base Tax Rate Revenue Ownership Administration 1. Customs Federal or State Federal or State Federal or State Federation and State 2. Excises Federal level Federal legislation Federal legislation; surcharges set Federation. Cantons or by cantonal or municipal municipalities may assemblies. impose retail level taxes. 3. Wage and Federal level or Federal legislation Basic rate set by the Federal Shared by the Persona canton level legislation; surcharges set at the Federation, cantons, Income. cantonal level. and municipalities. 4. Corporate Federal level Federal legislation Federal legislation Federation. Income 5. Sales Tax-- Federal level; Federal legislation Rates to be set by the Federation Federation, cantons and Retail Level feasible but legislation; surcharges can be set municipalities complicated at by cantons. the cantonal level 6. Property Tax Municipal level Federal legislative Municipal level Municipal level framework 7. Real Estate Municipal level Federal legislative Municipal assemblies Municipal level Transfers background 8. Motor Vehicles Municipal level Federal legislative Cantonal level Cantons and framework municipalities 9. Social Security Federal level Federal legislation Federal legislation Funds Contributions 10. User Fees. Service Cantonal legislation Cantonal or municipal providers Service providing Social Services providing agencies agencies 11. User Fees. Service supply Regulatory Service providing companies Service providing Utilities. companies Frameworks subject to regulatory oversight companies 12. VAT Federal level Federal legislation Rates to be set by the Federation Shared between the legislation Federation and the cantons. * Tax assignments are suggested between the State and Federation Government. The tax assignments between the State and Republika Srpska Entity could be similarly aligned. 22 TABLE 4 1997 Tax Revenue Assignments Federation Federation Revenue Distribution (Percent) Revenues State Federation Cantons Total (%) Customs Duties 7.4 92.6 0.0 100.0 Excise & Special Import Taxes 0.0 100.0 0.0 100.0 Wage & Income Taxes 0.0 50.0 50.0 100.0 Profits Tax 0.0 50.0 50.0 100.0 Sales Tax 0.0 100.0 100.0 100.0 Other Taxes 0.0 0.0 100.0 100.0 Total Revenues 1.3 40.0 58.7 100.0 The revenue assignments were proposed to create vertical balance in the system, but several important cautions should be noted. First, the assignments will allow horizontal imbalances at the canton and municipal levels. Estimates suggest several cantons would have significant deficits, arising both from the expenditure and revenue sides. Revenues are constrained in some areas because of very weak economies and expenditure needs are larger in some places than others. Second, the expenditure assignments used here are similar to those in place for 1996, except most spending for health, education, and police services are passed to the cantons. Changes in the expenditure assignments will necessitate changes in revenue assignments. Third, very minimal grants from the Federation to the cantons are allowed for in the financing scheme. Own source revenues should be the major financing source because of accountability, concerns about cross subsidies, differences in service demands, and other reasons. Still, a greater role for grants exists, and development of a transfer system would require additional tax assignment to the Federation. Finally, the assignments fail to recognize differences in revenue buoyancy, and will have to be altered over time if expenditures grow at different rates than revenues. Sub-national governments often are assigned taxes with low buoyancy. However, about 60 percent of the revenues assigned here are from the relatively buoyant sales and wage taxes. Most taxes, including the corporate and personal income, excise, customs, and value added taxes would be best administered at the Entity level, and in several cases at the State level. A limited number of revenue sources, such as the property tax and user fees, could be administered at the sub-Entity level. Taxes are administered at the Entity level in the Republika Srpska, and legislation has been passed for administration at the Federation level as well. Advantages in terms of lower operational costs, coordination of administration, and uniform practices all argue strongly in favor of the Federation. The key is to design an administrative and revenue distribution system that safeguards the cantons and municipalities so that they will be willing to accept Federation administration. 23 TABLE 5 1997 Tax Revenue Distribution Bosnia and Herzegovina Total Revenue Distribution (Mil. DEM) T otal (Mil. DEM) Revenues State RS Federation Cantonis Customs Duties 25.00 39.70 208.30 0.00 273.00 Excise & Special Import Taxes 0.00 0.00 173.00 0.00 225.00 Wage & Income Taxes 0.00 60.00 20.00 20.00 100.00 Profits Tax 0.00 0.00 70.00 70.00 140.00 Sales Tax 0.00 78.00 0.00 500.00 578.00 Other Taxes and Fees 15.00 21.00 0.00 179.00 215.00 Total Revenues 40.00 198.70 523.30 769.00 1,531.00 Expenditures Expenditures (Except Army, 40.00 116.70 163.30 769.00 1,089.00 Social Prog. & Debt Serv.) Expenditures: Army, Social 0.00 82.00 360.00 0.00 442.00 Programs & Debt Service Total Expenditures 40.00 198.7 523.3 769.00 1,531.00 The Federation and RS data do not include special funds for health care, unemployment, and pensions. The RS data do not include municipal revenues. Conclusion Progress in developing the governments in Bosnia has been slower than was envisioned when the Dayton Agreements were penned. This is not surprising since there is little or no world experience with developing governments on the short time line presumed in the Agreements. There are positive signals in Bosnia, such as a number of government institutions becoming operative and revenues being passed to both the State and the Federation. Still, the challenge of achieving fully operational governments remains in place. Traditional economic models of federalism, like the one on which this paper's analysis is based, suggest a government structure assuming there is an intent to achieve Pareto-efficiency for the entire country. Current attitudes in Bosnia appear consistent with the views that many people in each ethnic group see themselves as members of their group, rather than as Bosnians, and are not broadly concerned about welfare of the entire country or access to public services outside the group. The motivation for the fiscal federalism structure proposed in the Dayton Agreements is 24 better interpreted as an effort to manage conflict between the ethnic groups. Federalism in a conflict management sense does not require that each group be given its own state, but does lead to the conclusion that institutions of power should be brought closer to people so that decision making can be more sensitive to the different ethnic groups (Tishkov, 1993). Decentralization in this context is a means to lessen the points where disagreement exists, rather than a structure to obtain economic efficiency. Common institutions at the State, Entity or canton levels are maintained, but only for functions that must be broader in scope. The fiscal (and other) interdependencies from common institutions at the State. Entity. and canton levels provide opportunities to build relationships and trust over time. The government structure included in Dayton is workable. However, unless governments negotiate other arrangements, the likely outcome in the short to medium term is diseconomies of scale in provision of certain services, as some that can be provided at lower unit costs at the State level are provided at the Entity level, and some with lower unit costs at the Entity level are delivered at the canton or municipal level. Services with geographic spillovers will be underprovided because governments will fail to adequately account for the benefits received by members of other ethnic groups. Further, little consideration will be afforded to equity between groups. resulting in widely different access to services across the country. Better service delivery mechanisms from a national, Pareto-efficient perspective will not be selected given the very strong distaste for cross subsidies and the fear by minority groups that they will be dominated by a larger group. More efficient arrangements can be expected to evolve over time as confidence in the government structures grows. Horizontal imbalances will be large in the Federation, and grants will be used to a very limited extent to offset these imbalances. The resulting problems will be lessened in the next several years to the extent the former Bosniac or Croat governments are able to provide some resources to the more financially strapped areas or the donors help fund infrastructure and other needs. But donors will not finance operating expenditures, and maintenance of new infrastructure investments will place greater demands on limited budgets, so financial problems in the places least able to pay could expand. Economic growth that increases revenue raising capacities across Bosnia is one solution to offsetting imbalances. Also, differences in access to services could be lessened with a system that allocates more grants to places with the greatest expenditure needs and with the least revenue capacity. 25 Bibliography Bird, Richard M., Robert D. Ebel, and Christine 1. Wallich, eds. 1995. Decentralization of the Socialist State: Intergovernmental Finance in Transition Economies. Washington, D.C.: World Bank. Ding, Wei. 1991. Costs and Benefits of Union and Interdependence of Regional Economies. Mimeo. World Bank. Farsad, Mansour. 1991. Yugoslavia -Economic Integration and Viability of a Breakup. Draft Working Paper, Eastern Europe Dept. World Bank. Federal Statistics Office of Yugoslavia. 1945-1985. Yugoslavia. Annual Publication. McLure, Charles E. Jr. 1994. The sharing of taxes on natural resources and the future of the Russian Federation, in Russia and the Challenge of Fiscal Federalism, Christine I. Wallich, ed. Washington, D.C.: World Bank. OECD. 1992. Regional Developments and Developments in the Republics of Yugoslavia. March 1992 Mimeo. p 36. Prud'homme, Remy. 1994. On the Dangers of Decentralization. Policy Research Working Paper No. 1252. Washington, D.C.: World Bank. Tanzi, Vito. 1995. Fiscal Federalism and Decentralization: A Review of Some Efficiency and Macroeconomic Aspects. Annual World Bank Conference on Development Economics 1995. Michael Bruno and Boris Pleskovic, editors. The World Bank. Tishkov, Valery A. 1993. Nationalities and Conflicting Ethnicity in Post-Communist Russia. Working Paper Series. Conflict Management Group. Harvard Law School. Vodopivec, Milan. 1991. Redistribution through the Soft Budget Constraint. Mimeo. World Bank. October. Woodward, Susan. 1995. Balkan Tragedy. Washington D.C.: Brookings Institution. World Bank. 1996. Bosnia and Herzegovina: Toward Economic Recovery. World Bank Country Study. World Bank. World Bank. 1989. Yugoslavia - Financial Sector Restructuring: Policies and Priorities. World Bank Report #7869-YU. November. 26 Policy Research Working Paper Series Contact Title Author Date for paper WPS1692 Regulating Market Risk in Banks: Constantinos Stephanou December 1996 P. Infante A Comparison of Alternative 37642 Regulatory Regimes WPS1693 Famines and Economics Martin Ravallion December 1996 A. Ramirez 85734 WPS1694 What Can New Survey Data Tell Us Martin Ravallion December 1996 A. Ramirez about Recent Changes in Distribution Shaohua Chen 85734 and Poverty? WPS1695 Are There Dynamic Gains from a Jyotsna Jalan December 1996 A. Ramirez Poor-Area Development Program? Martin Ravallion 85734 WPS1696 South-North Migration and Trade: Maurice Schiff December 1996 M. Patena A Survey 39515 WPS1697 Strategies to Develop Mortgage Dwight M. Jaffee December 1996 R. Garner Markets in Transition Economies Bertrand Renaud 37670 WPS1698 Regulatory Structure and Risk and Ian Alexander December 1996 R. Schneiderman Infrastrcuture Firms: An International Colin Mayer 30191 Comparison WPS1699 The Maturity Structure of Debt: Fabio Schiantarelli January 1997 B. Moore Determinants and Effects on Firms' Alessandro Sembenelli 38526 Performance-Evidence from the United Kingdom and italy WPS1700 Child Nutrition, Child Health, and Harold Alderman January 1997 A. Ramirez School Enrollment: A Longitudinal Jere R. Behrman 85734 Analysis Victor Lavy Rekha Menon WPS1701 Monitoring Environmental Standards: Catherine Dion January 1997 B. Laplante Do Local Conditions Matter? Paul Lanoie 85878 Beno7t Laplante WPS1702 Nontradables Inflation and Hana Polackova January 1997 H. Polackova Macroeconomic Policy Mix: A Model 30182 with Policy Application to Transition Economies WPS1703 Financing of Government Liabilities Hana Polackova January 1997 H. Polackova as the Population Ages in New 30182 Zealand WPS1704 Citizen Complaints as Environmental Susmita Dasgupta January 1997 E. de Castro Indicators: Evidence from China David Wheeler 89121 Policy Research Working Paper Series Contact Title Author Date for paper WPS1705 The Polish Experience with Bank Fernando Montes-Negret January 1997 T. Ishibe and Enterprise Restructuring Luca Papi 38968 WPS1706 Monetary Policy during Transition: Martha de Melo January 1997 C. Bernardo An Overview Cevdet Denizer 37699 WPS1707 Trade Reorientation and Productivity Simeon Djankov January 1997 J. Ngaine Growth in Bulgarian Enterprises Bernard Hoekman 37947 WPS1708 Has Latin America's Post-Reform William Easterly January 1997 R. Martin Growth Been Disappointing? Norman Loayza 31320 Peter Montiel WPS1709 Poverty Comparisons with Jean Olson Lanjouw January 1997 A. Ramirez Noncompatible Data: Theory and Peter Lanjouw 85734 Illustrations WPS1710 Why Paper Mills Clean Up: Raymond S. Hartman January 1997 D. Wheeler Determinants of Pollution Abatement Mainul Huq 33401 in Four Asian Countries WPS1711 Issues in Comparing Poverty Trends Christine Jones January 1997 R. Martin Over Time in Cote d'lvoire Xiao Ye 31320 WPS1712 Demand Elasticities in International Arvind Panagariya December 1996 J. Badami Trade: Are They Really Low? Shekhar Shah 80425 Deepak Mishra WPS1713 Why Did Colombian Private Savings Alejandro L6pez January 1997 E. Khine Decline in the Early 1 990s? 37471 WPS1714 Fiscal Federalism in Bosnia- William Fox January 1997 Y. Jiwa Herzegovina: The Dayton Challenge Christine Wallich 34848