Fiscal reform and federal relations: Brazilian deadlocks

Brazilian fiscal reform got to deadlocks because proposals tried to avoid considering federal relations. There are two main types of federal relations: the competitive and the cooperative. In both types is possible to observe coordination mechanisms. Brazilian federalism is a mixture of both types what leads to difficulties finding solutions. We argue that is more important to find mechanisms to facilitate cooperation than to discuss the qualities of any alternative fiscal structure. Fiscal reforms brings along a great deal of uncertainty. So it is important to discuss the reform timing and the compensation mechanisms before hand.

cause of an understanding that this could increase resistances to changes. However in a federation it is almost impossible to discuss fiscal reform without dealing with the conflicting issues of federal relations. In our view fiscal reform debate in Brazil has taken a route that does not conduct to a solution. Since the 1988 Constitution Brazilian fiscal structure have been changing in a way as to bring many problems to Brazilian economic structure. In this article we argue that a better way to approach fiscal problems would be to understand the elements that lead to deadlocks and hinders cooperation among actors involved. Institutional economics mainly the approach of North (1990North ( , 2005 and the extensions offered by Dixit (1996) and Acemolgu and alli (2004) provides an adequate framework to tackle this issue.
The article is divided as follows. The first section points the Brazilian fiscal problems and the failure of the proposals that where advanced to solve then. Next it is discussed how intergovernmental problems are better understood as a problem of allocation of property rights among government levels in the presence of distributional conflict. In the third section we identify the main conflicts that can be observed in inter-governmental relations and finally, on the fourth section, we point out which are the main determinants of the observed paralysis and alternative routes of solution.

BRAzILIAN FISCAL STRUCTURE PRoBLEMS AND PRoPoSALS
There is a substantial literature pointing the flaws of Brazilian federal fiscal structure 2 . This list includes: the inadequacy of the structure after structural changes like stabilization and commercial liberalization, excessive tax burden over firms, fiscal evasion, excessive payroll tax burden, fiscal war, regressive taxation, vertical imbalances, cascading taxation, unclearly defined competencies between levels of government in many areas of public policy, among the most important. Authors are also unanimous about the consequences: damaging effects over the productive sector specially the export sector, absence of coordination between levels of government on collection and expenditure, maintenance or even worsening income distribution concentration.
The origins of these problems could be traced back to the extensive fiscal reform undertaken by the Military Government in 1967 and the 1988 Constitution. Moreover a large number of small changes have significantly altered fiscal structure.
The main picture can be seen in Graphic 1, depicting the evolution of Bra-zilian tax burden and expenditures composition in three levels of government. First of all it is interesting to call attention to the fact that almost all along the period we observe fiscal unbalances. Federal government with rare exceptions showed fiscal surplus compared to primary expenditure while state and municipality levels showed deficit. These differences can be explained by two reasons: transferences and indebtedness. States and municipalities until 1998 where allowed to borrow including from their state owned banks. Transferences by the other side are a usual mechanism used by federal government since the advent of the republic in 1889 to cope with problems of regional imbalances. In 1967 they became constitutionally compulsory and federal government began to automatically transfer a significant amount of its revenues to states and municipalities. However what strikes most in this graphic is the close relation between main political Brazilian events and the changes in fiscal structures. Before 1929 when states where totally autonomous to tax and borrow we see an increasing trend of state revenues and expenditures while federal government faced decreases. Between 1930 and 1945, a period in turmoil because of internal and external events, it is possible to see a reversal in the observed tendency due mainly to measures adopted by the Vargas dictatorship in order to increase central government power in relation to states. During the re-democratization period between 1945 and 1964, we observe a relative stability. After the fiscal reform implemented by the Military Regime in 1966 state governments suffered a huge decline in their share of taxes while municipalities increased their share in expenditures due to larger transferences from federal tax collection. The re-democratization process, starting in 1982 with the first election for state government in 18 years, changed again the pattern. Now federal government began to loose participation while state and municipalities gain. After the 1988 Constitution, the trend is again altered with federal government trying to recuperate losses between 1982 and 1988. In short what this graphic shows is a century long distributional conflict between levels of government.
Brazilian public indebtedness has been in the centre of the fiscal problem since 1970. All three levels of government were allowed to borrow without restrictions even from their own financial institutions until 1998.
Traditional literature on Brazilian fiscal structure problems concentrates on proposals to solve them without giving much attention to the fact that economic policy is a matter of political outcomes. The usual diagnosis is about significant allocative inefficiencies with a technical design to solve it, if this does not happen it is because of a lack in political will. In our view, it is a much more complex problem of distributional conflict between levels of government, the inexistence of cooperation mechanisms in inter-government relations have been dampening the possibilities of reform, a better understanding of the conflict is necessary in order to tackle them with better expectations of success.
A good evidence of the point made above is the sequence of fiscal reform proposals made in the last 10 years. During this period Brazilian society witnessed at least five reform attempts by the federal government, none of them were adopted. They began in 1997, during the Cardoso government, Werneck (2002 and analyses these proposals: "They seem to share the same basic diagnosis on what is wrong with indirect taxation in the country. They are strikingly similar in what concerns taxes to be eliminated. And are not so dissimilar in what concerns taxes to be created" (Werneck, 2003: 9). The proposals mainly suggested a reform on the VAT tax which is today the main revenue source of state governments and the higher single tax collection and the elimination of a bundle of taxes considered damaging to the countries economy. Table 1 presents the main features of the proposals. However, we do not agree with Werneck that stated that: "… Considering that two of those proposals stemmed from the Executive and two from Congress that seems to point out to a surprising degree of agreement on the main line of the required reform" (Werneck, 2003: 9) 3 In our view there are significant differences between the proposals that emerge from the federal executive and those that came from the Congress. If we rearrange the information on Table 1 highlighting the sponsors the picture changes considerably. Table 2 presents the same information however it focuses on the origin of the suggestions. Beginning with the taxes to be eliminate what we see is that executive and legislative powers agreed that three taxes should be eliminated: Federal tax on manufactured products (IPI), Service tax charged by local governments (ISS) and State VAT. But this is the only agreement that can be seen in this table. They disagreed on the elimination of tax on financial transactions (CPMF) which is one of the cascading federal taxes. Legislative power proposes to eliminate it while the executive never accepted its elimination. Federal Government was not ready to renounce to one of its major source of revenues even if it is damaging to the productive structure of the country. When we turn our attention to taxes to be created it is not possible to find a single agreement between executive and legislative. The two legislative proposals maintained the idea of a Dual VAT that should be collected by both federal and state levels. The executive proposals transforms state VAT into federal VAT. only in the last one, that was not supposed to be a serious offer because it came too late in the legislature, federal government proposed a state VAT which should be nationally uniform. The idea of a nationally uniform VAT is something that in a way violates the very concept of federalism which should allow states some degree of power to collect and to legislate on taxes.

FISCAL FEDERAL STRUCTURE AS ALLoCATIoN oF PRoPERTy RIGHTS AMoNG GoVERNMENT LEVELS
The discussion above highlights the fact that fiscal reform in Brazil is a matter of intergovernmental relations. In this section we will argue that federations are institutional structures that need coordination mechanisms because of the important distributional conflict between federal members.
The inclusion of conflict into the analysis unveils a completely different world, improving the understanding of fiscal federal relations. In order to do so it is necessary to recall a short description of the fiscal federalism approach usually utilized in traditional economics. The literature on fiscal federalism is mainly worried about the efficiency of federal arrangements. Adopting a normative attitude it seeks to define what should be the optimal level of fiscal decentralization.
The main problem faced by fiscal federalism literature is the existence of externalities. oates (1972) was aware of the problem when stated that: "... the optimal degree of fiscal decentralization will vary substantially among different societies" (oates, 1972: 31). The solution proposed by oates (1972) was the inclusion of transferences from central government in the analysis because that could handle the externalities problem through matching revenues and jurisdictions. This suggestion is based in the traditional Pigouvian solution to the externality problem. Transferences from central to lower level governments would reallocate income from jurisdictions that impose negative externalities to those that suffer it. However this process is not costless. As Coase (1960) has demonstrated if the use of the price system has costs then the Pigouvian solution may not be the best. Transferences will demand negotiations between the involved jurisdictions that may compensate the welfare gain from decentralization. Even thought there are a number of other authors that tried to deal with the problem of externality related to federalism, none of then were able to present a reasonable solution within the limits of conventional economics 4 . oates approach to the problem continues to be the main contribution to the question.
It is always appropriate to remember that according to Coase (1937) the costly nature of the price system is mainly due to uncertainty. If the actors knew which would be the best agreement they would not need to negotiate to get to the best alternative. It will be automatic as in the case of the so called Coase Theorem. It is also important to remind that when talking about federalism we are dealing with a problem that involves politics.
This brings us to Dixit (Dixit, 1996: 20) contribution to the problem of transaction costs in political transaction. Laws and constitutions last long, economic circumstances change a lot during constitution life, conflicts about contingencies are much more frequent in the political arena, and there are a larger number of interests involved. Besides constitutions should be build in sufficient general term as to be applicable to many circumstances. on the other side there is a wide array of interpretation of the meaning of orderings and so constitutions allow much more manipulation than other kinds of contracts. These characteristics impose two important consequences. on one hand there is the problem of "time consistency in policy" 5 , in the other multi-agency problems. According to Dixit (1996) problems of temporal consistency arise because contracts in politics are very loose. Property rights are very blurred. Principals have no means as to make politicians or bureaucrats accountable. Moreover there are many difficulties to establish a clear relationship between agents and principals. Principals are usually the voters. This group however does not have a single interest, even some groups that could be recognized as having one common interest it is always possible to identify other kind of interest that may be of a conflicting nature. Agents are a heterogeneous group as well. Conflicts are also vivid among them. They are important not only between politicians and bureaucrats but also within politicians and bureaucrats groups. When the federalism problem is incorporated into the analysis the complexity increases enormously.
Applying this line of argument to federalism it is possible to say that a federal structure is a contract between society and government, what brings together all the problems pointed above but moreover it is a contract within the state among autonomous unities of policy making. In countries organized in unitary basis this characteristic does not arise. This means that policy decision depends not only of central government decision but also of lower level government decision, what make it necessary for government levels to coordinate their action in order to diminish inefficiencies. However, considering the problem of incompleteness of contracts in the political arena, this is an issue that should not be handled as a maximization problem as the conventional theory would do. Probably a more adequate approach to deal with the issue in this case is to use North's idea of uncertainty reduction. Following this line of argument we should look at the problem of establishing rights and competencies for various government levels, in a federal structure, as a means to reduce conflicts and to improve coordination. Institutional design in this case is not an issue of finding an optimal design. Instead it is a question of finding what set of rules could attain the objectives pointed above in a specific political and economic environment.
The conflict presented here is about distribution. Two kinds of distributional conflicts can be observed in this case. The first one is about distribution of tax collection and competencies between levels of government and the second one is about income distribution between private and public sectors. Acemolgu et alli (2004) discussing the importance of institutions for economic development presents a framework of analysis that incorporates distributional conflict and highlights the importance of political institutions. The argument's emphasis is on political power. Political power has two components: de jure political power and de facto political power. De facto political power is determined by distribution of resources while de jure political power depends on political institutions. The combination the two political powers determine both economic institutions in the period and political institutions in the next period. In doing so the authors make all the variables endogenous to the model and they highlight the importance of political institutions and political power to the determination of economic policy and economic performance.
Very important to the argument is the idea that the choice of economic institutions is a problem of collective decision. Different groups have different preferences for institutions, because different sets of institutions will be responsible for different resource distribution. In a situation like this there is no optimal solution possible. This is why the authors sustain that political power is crucial to the outcome. The question that arises at this point is: which will be the set of equilibrium institutions? And the answer to this question is: the set of institutions that is preferred by the group with the largest political power. This solution is not necessarily optimal in the sense that no one loose, it is however, an equilibrium solution for no group will think useful to devote additional resource to change the situation unless some of the initial conditions change. What changes in initial conditions could alter the result? To answer this question we have to understand better what constitutes political power.
Political power according to the definition of the authors has two elements. The first one is de facto power has two components: economic resources and capacity to solve problems of collective action. Economic resources are usually taken into consideration economic theory when dealing with questions like this. Capacity to deal with problems of collective action is not. This has to do with what olson (1971) called our attention in the Logic of Collective Action, meaning group's action efficiency in the presence of differences in common and individual interests. In addition there is a second important element in the analysis of the authors. This is the de jure political power. This element is not usual in economic theory. It deals with political rules and allocation of political rights. They establish the reach of the agents' action in the political arena and in doing so they are also responsible for determining the possible outcomes.
Using Dixit (1996) and Acemoglu (2004) insights it is possible to advance a framework of analysis for federal relations and distribution of rights among levels of government. For the sake of simplicity it is possible to start by considering that there are only two government levels that are independent units of decision making. However their actions are inter-dependent which means that there are externalities problems. Federal relation is not a natural cooperative relation. There is a huge potential for conflict in these situation. Federal members compete mainly for tax basis, but they also compete for the provision of public goods. So we have a situation of distributional conflict in an area where contracts are very loose as Dixit (1996) has called our attention. Each unity would prefer to have the largest share of tax receipts and, at the same time to provide the minimum amount of public goods that would not risk loosing votes. The rational behaviour would be to try to make the other government sphere to provide the goods without loosing taxes or votes. When this is the behaviour of all members the outcome would be a minimum provision of public goods. To solve this problem some kind of coordination mechanism should be introduced.
In this case it is possible to define two main ideal kinds of relationship between levels of government. We should call then: cooperative 6 and competitive. The difference between these two forms of federalism can be found in the nature of the coordination mechanisms observed in each of the cases. In cooperative federalism policy to be executed in the territory of a member of the federation should somehow be agreed between central and lower governmental levels. on the other hand competitive federalism is the situation in which levels of government compete for tax bases and supply of services for each federal member.
The important point to be made about this classification of federalism types is that they have very different consequences to the transaction costs of inter-governmental relationship. In the one hand cooperative federalism should give rise to high ex-ante transaction costs during the period prior to the agreement that determine decision rights among federal members and should show low ex-post transaction costs along the period of policy implementation and execution once it is supposed that in this phase members are not to compete. This of course will depend on the specific rules that conforms federal relations. If competencies are clearly established in the Constitution it should be expected that ex-post transaction costs should be at their minimum standards 7 . on the other hand competitive federalism should exhibit minimum ex-ante transaction costs, for there is no negotiation between federal government levels, however there should be high ex-post transaction costs due to the lack of coordination on the provision of public goods between federal members. This should produce inefficiencies in the provision of public goods due mainly to externalities.
In consequence there is no a priori possible conclusion about the superiority of cooperative federalism over competitive federalism or vice-versa. The trade-off between ex-ante and ex-post transaction costs may make both forms of federalism similar in terms of total transaction costs. However competitive and cooperative 6 Inman and Rubinfeld (1997) also have a definition of cooperative federalism which involves the unanimity rule. These authors defined cooperative federalism as the principle that "prefer the most decentralized structure capable of internalizing all economic externalities, subject to the constitutional constraint that all central government policies are agreed to unanimously by the elected representatives from each of the lower-tier governments." (Inman e Rubinfeld, 1997: 48) This cooperative federalism definition is based on the American federalism structure and is not useful for our purposes. federalisms are pure Weberian ideal types. They are just abstract representations of federal arrangements that help us think about current federal structures in terms of solving conflicts in the process of intergovernmental relations. We can also think of then as extremes of a continuum which allows many other federal organization forms. Each one of the extremes would be a situation where ex-ante or expost transaction costs would be in such a low level as to produce the lowest total transaction costs possible. This would be because the cooperation mechanisms at work would be the best possible. Intermediate positions along the continuum should represent total transaction costs higher then the extremes. This means that various combinations of elements of competitive federalism with cooperative federalism would be less efficient alternatives in various degrees.
Why should the intermediate positions be less efficient? Federal organizations are contracts network. As was said above they provide members with the basic rules of their relationship. When these rules offer conflicting signals the functioning of the system may be damaged. Cooperative federalism is an arrangement that is close to the idea of an organization, like a firm, in Coase's definition where things are settled through agreements. Competitive federalism is close to the idea of a market where solutions are reached through agent's individual effort to improve his or hers performance, independently from the others. If the arrangement allows agents to seek their individual interests within an agreement the result would be at the expense of cooperation.
There are a few basic characteristics that differentiate cooperative and competitive federalism and they can be observed in fiscal structure. Concretely speaking in economic terms cooperative federalism operates in a system of tax-sharing. Taxes could be collect by central and/or state governments but they are usually distributed according to the responsibilities that each member has. And no it does not matter much which level is collecting. The principle that orientates who gets what is the principle of equal capacity to provide public or collective goods in the territory of each member state and this principle is usually stated in the Constitution or in law. In competitive federalism members of the federation make individual efforts to collect taxes and there is no agreement on sharing the results of these individual collection efforts.
In cooperative federalism tax base is also a matter of negotiation. Each federal level is attributed a specific base which cannot be taxed by other levels. In competitive federalism, federal and state levels compete for a common tax base like income for example. Definition of competencies in cooperative federalism are stated in Constitution or in the law while in competitive federalism it is up to each member to decide what to do. Finally with respect to transferences in cooperative federalism there are compulsory transferences that could be expended freely by lower government level and they usually follow the principle of compensatory transferences to attain a similar capacity of expending in all states. Transferences in competitive federalism are voluntary, earn-market and they are mainly to individuals and organizations rather than to other government levels. Table 1 summarizes the main features of each federalism ideal type. Each one of these ideal federalism types has its own cooperation mechanism and the features described above are the rules that characterized then. In real world there is no such a thing as a perfect cooperative or competitive federalism. What is possible to be observed is a variety of combination of both competitive and cooperative federalism. Each of one of these combinations will show some degree of transaction cost that should be higher than the ideal types. Some countries exhibit a higher degree of competitive federalism, like USA, while others are mainly characterized by cooperative federalism, like Canada or Germany. The question so is what are the minimum transaction costs possible for a given country? The minimum transaction cost possible for a given situation would emerge from the combination of features that reduces conflicts between jurisdictions making it possible to internalize the higher amount of externalities possible and provide the higher level of public goods. To answer to this question it is necessary to turn our attention to the case here in analysis. We argue that in the Brazilian case there are certain features of both competitive and cooperative federalism that are combined and that are responsible for increasing transaction costs due to the weakening of the cooperation mechanisms. It will be shown that this result is a consequence of a path dependence process in fiscal structure that have been evolving since the Military Brazilian Government and whose deadlocks are still waiting to be solved.

BRAzILIAN FISCAL STRUCTURE EVoLUTIoN AND PATH DEPENDENCE
Since the event of the 1988 Constitution a series of fiscal changes were made even though it was not possible to undertake a major fiscal reform. 1988 Constitution transferred a significant amount of resources from federal to state and municipal levels without transferring competencies. The new situation was one of state and municipal enrichment in relation to federal level impoverishment. This situation was aggravated with the inexistence of controls over state and municipal indebtedness. In Graphic 2 we observe the trends of fiscal revenues and disposable revenues by government level.
Graphic 2 Revenue and Disposable Revenue % PIB by Government Level 1960Level , 1965Level , 1970Level , 1975Level , 1980Level , 1988Level -2004 0 5 10 15 20 25 1960 1965 1970 1975 1980 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Central From 1980 to 1991 federal government revenues share shows a variable pattern while state and local levels increase their shares. This was the result of changes in government level's property rights. Federal government transfers to both state and local government increase as well as transfers from state to local government. Local government was the level that received the largest gain. (See Graphic 3) This period was characterized by hipper-inflation. Even thought local level net gain was really amazing.
This new fiscal structure was a result of a well succeeded mobilization for decentralization made mainly by local government level politicians that managed to influence legislators. It imposed a heavy burden on the Union for the majority of responsibilities were still poorly defined and federal government maintained all its responsibilities. Graphic 3 gives a good picture of the inter-governmental distributional conflict. The principal trend shows a continuous federal loss, fairly stable tendency for state governments and a local level continuous gain. However there are some moments in this period when we see some reaction from federal government and others when state or local government succeed to regain terrain. This is the process by with Brazilian fiscal federalism evolves with periods of centraliza-tion and others of decentralization. The important point to investigate is the mechanisms utilized by the three government levels in this conflict. In other words how intergovernmental rights evolve. Level -1960Level - , 1965Level - , 1970Level - , 1975Level - , 1980Level - , 1988Level - -2004
To effectively change a situation like this in Brazil as in many federations it is necessary to change the Constitution. There are two ways to do so: constitutional revision or constitutional amendment. Constitutional amendment requires the approval of 2/3 of the Congress. Constitutional revision requires only 50% approval but could not be done at anytime and is a widespread process that is not limited to the issue at stake. In 1993 the Constitution was revised and federal government suffered another unfavourable impact on tax shares. The urgency of macroeconomic problems could have determined this apathy concerning fiscal problems from the part of federal government. An alternative explanation is that federal government was, by that time, envisaging other formulas to solve its financial crisis and decided not to allocate resources to enter into a difficult negotiation with congress that could even worse its situation.
From 1988 to 1990 federal government succeeded to reverse the situation mainly through reducing non constitutional transferences and increasing tax rates of non-shared taxes. This was a reaction to the trend established after the beginning of the 80's with the re-democratization process. Unfortunately there are no disaggregating data about federal revenues for this period. Even though Graphic 2 shows how federal government revenues grew much faster the other levels revenues in this period. This result can only be attributable to growth in taxes. What is clear is that federal success was obtained by imposing non negotiated loss in revenues shares to the other two government spheres.
From 1990 to 1994 there is certain stability. The possibility of changing Constitution in 1993 could explain this outcome. But instead of improving federal finances Constitutional revision imposed another loss to federal government 8 . This loss was mainly due to the prohibition that any other tax except states VAT could be collect from public utilities services like electric energy, inter-state and municipal transportation, communication, fuel and minerals 9 . This was another victory of lower government levels because this tax is collected by state government and shared with local governments.
It took three years for a federal government reaction. Meanwhile two other important events were evolving. The first was the state debt renegotiation and the other was the beginning of the establishment of clear responsibilities among federal members.
State and municipalities debt in 1991 was 11,5 % of PIB, in 1996 it reached 5,8% of PIB. After stabilization process it became inevitable to control state debt as part of stabilization efforts. Federal government try to force states to do it by offering states a debt negotiation. According to Rezende e Affonso (2000) state bonds were exchanged by federal bonds, real interests were set on a maximum 7,5% and there was also a provision that interests payments could not exceed 11% of states tax collection. The result was that after 1997 states interests' payments were reduced significantly while federal interests' payments increased sharply and states debt grew at a much lower rate than federal debt. once again federal government suffered further loss.
In 1996 a set new set of norms that intended to provide a clear definition of responsibilities of government levels concerning health services. The NoBSUS-96, as it was called, established that the provision of public health services should be done by local level governments and costs should be shared between the three levels. Later in 1998 another set of norms established responsibilities for education policy with the same principles presented in public health services policy.

Graphic 4 Union, States and Municipalities Interests Payments, % GDP
The process describe so far has two main undesirable consequences. The first is the burden imposed on productive sector of the economy as the increasing nonshared taxes were all cascading taxes which impose heavy costs to Brazilian economy and as total tax burden increases without favoring the perception by the public of an improvement in public services provision. The second is a non coordinated pattern of intergovernmental relation, characterized as tug of war; each time one actor reacts causing damage to a third party; general public cannot recognize the causal relation between improvement in public service provision and a generalized increase in tax burden, the reaction chain goes on indefinitely. About the first much has been written by Brazilian economists. About the second there is a widespread unawareness. We argue that the second consequence is the principal determinant of the costs determined by a larger tax burden.

DETERMINANTS oF DEADLoCKS
Intergovernmental relations in Brazil evolved in a way as not to favour cooperation. As states and municipalities accumulate victories over federal government in Congress they developed the belief that it was not necessary to negotiate and that they could always solve they financial problems by imposing further losses to federal government. From its part federal government reacted since 1991 by increasing and creating taxes. Both sides of the fight use their instruments as power mechanisms to reach their goals. The problem is that this is a zero or negative sum game to the public sector and the economy. Either one or more government sphere looses or either private sector pays the bill through tax burden increases. It is clear that this process involves two distributional conflicts. The first conflict is between spheres of government and the other between public and private sector.
In this case deadlocks will be found on the structure of intergovernmental relations. First, there is no mandatory provision that revenues' changes should be accompanied by responsibility changes. Lower level governments as well as federal government, each with the mechanisms at their disposal; have been changing tax structure without having to account for the uses of the additional resources. Brazilian federalism presented the main characteristics of cooperative federalism, as defined earlier. It has tax-sharing, definition of tax base per level of government, definitions of competencies, compulsory transferences and compensatory transferences. However it also has some competitive federalism features as: exclusive taxation for certain taxes, competition for tax base between federal and state governments and also among state governments, voluntary transferences. This structure opens room to a predatory competition between levels of government that has the consequences pointed in the literature.
The solution to the problem should emerge from negotiations in Congress on how to diminish potential conflict in intergovernmental relations. A promising route could be reached through property rights reallocation between federation members in order to prevent any one of them to impose losses on the other(s). For instances, a way of doing so would be to prohibit Union to increase tax rates without Congress' acquiescence. At the same time it would be necessary to allow federal government to interfere on states decision on tax changes, as those that conduct to fiscal war. These changes would induce federal units to enter into negotiation whenever changes are desired by one of them.
Any intense negotiation period between federal members is a process full of uncertainty. To cope with this problem three elements are essential: first it is not reasonable to expect that this process could be finished in a short span of time, it is wiser to define a schedule long enough to guarantee full awareness of the consequences of each step; second there should always be the possibility of decision revision; finally it will be important to establish a compensatory framework, what have been recurrently repudiated by federal government, to avoid heavy losses on any part involved. In doing so we believe that it will be easier to ensure participation and reach concrete solutions.