Federalism

Federalism, also referred to as federal government, a national or international political system in which two levels of government control the same territory and citizens. The word federal comes from the Latin term fidere, meaning “to trust.” Countries with federal political systems have both a central government and governments based in smaller political units, usually called states, provinces, or territories. These smaller political units surrender some of their political power to the central government, relying on it to act for the common good. In a federal system, laws are made both by state, provincial, or territorial governments and by a central government. In the United States, for example, people who live in the state of Ohio must obey the laws made by the Ohio legislature and the Congress of the United States. In Canada, residents of the province of Québec follow the laws made by Québec’s legislature and those made by the Canadian parliament. In addition to the United States and Canada, countries that are considered federalist include Australia, Brazil, Germany, India, Malaysia, Mexico, Nigeria, and Switzerland. Federal political systems divide power and resources between central and regional governments. The balance of power between the two levels of government varies from country to country, but most federal systems grant substantial autonomy to state or provincial governments. Central governments decide issues that concern the whole country, such as organizing an army, building major roads, and making treaties with other countries. Federalism varies in practice, however, and in some countries with federal systems the central government plays a large role in community planning, schools, and other local issues. Federal political systems are relatively uncommon around the world. Instead, most countries are unitary systems, with laws giving virtually all authority to the central government. The central government may delegate duties to cities or other administrative units, but it retains final authority and can retract any tasks it has delegated. The central government in a unitary system is much more powerful than the central government in a federal system. Cameroon, France, Italy, Japan, Kenya, Morocco, South Korea, Sweden, and Uruguay are examples of unitary systems. More from Encarta Offer: Live online homework help Math, Science, History and English Try Tutor.com for free! Educating mom Colleges reach out to single parents. Find out how. Cats Quiz Two parts furry, one part ferocious. Test your feline smarts! A confederation is similar to a federal system but gives less power to the central government. The loose alliances of countries or other political entities that make up a confederation seek to cooperate with one another while retaining ultimate control of their own internal policies. Unlike federal systems, confederations usually give each member nation absolute control over its citizens and territory. The central government decides only issues that affect all members of the confederation. In the 18th century the United States was founded as such a system under the Articles of Confederation. More recently, the Soviet Union dissolved in 1991, and many of the former republics formed a confederation called the Commonwealth of Independent States (CIS) to coordinate domestic and foreign policy. Confederations tend to be weak and unstable because member nations often resist relinquishing final authority on any matters and insist on their right to withdraw from the confederation at any time. Confederations are uncommon; most are international bodies with limited and specific responsibilities, such as the European Community (EC) and the British Commonwealth. THE BEGINNINGS FEDERALISM

OF

MODERN

II The United States began as a confederation. A weak central government ruled the country from 1783 to 1789 under the Articles of Confederation. Each state had an equal voice in Congress, but Congress could not collect taxes to operate the government. The confederation of states had no chief executive and no central body with enough power to make the states abide by the Articles of Confederation. Some states refused to follow the terms of the 1783 Treaty of Paris that ended the American Revolution, even though the Articles of Confederation gave Congress the right to make treaties for all the states. Trade disputes with Great Britain and other countries paralyzed the economy, but the Articles of Confederation left Congress powerless to take charge of international trade. Some states imposed heavy taxes on goods from neighboring states, further stifling commerce. The ensuing economic crisis threatened to destroy the young country, but no political authority had power to assume leadership. “The wheels of government are clogged,” future president George Washington remarked in 1785. Washington and other statesmen realized that the country could only survive if the central government had more power, but they also wanted to avoid trampling the rights of the states. In 1787 political leaders held the

Constitutional Convention to confront the crisis, and this historic meeting produced the principles of modern federalism. FEDERALISM IN THE CONSTITUTION II I Delegates to the Constitutional Convention at first sought merely to improve the Articles of Confederation, but this proved impossible. They wrote the Constitution of the United States, an almost entirely new document. The Constitution’s advocates, called Federalists, envisioned an energetic national government (see Federalist Party). The U.S. Constitution gave Congress broad powers, some of which are exclusive—that is, not shared with the states. For example, only Congress can make war, deal with foreign nations, issue money, and regulate interstate and foreign commerce. The laws of the national government prevail if they conflict with state laws. The Supremacy Clause in Article VI of the U.S. Constitution holds that the federal constitution, and all laws and treaties based on it, are “the supreme law of the land.” The Constitution preserves some powers for the states, however, making the United States what Federalists such as James Madison called a “compound republic.” The states share some powers with the national government, but they also retain some independence. The concurrent powers—those shared by both the national and the state governments —include taxing, spending for the public welfare, borrowing money, and eminent domain (taking private property for public use at a fair price). The Tenth Amendment to the U.S. Constitution specifies that powers not granted to the national government are reserved, or held, only by the states. The states regulate businesses and professions, conduct elections, provide for public schools, and protect the health and safety of their people. The states also retain sole power to establish local governments, including counties, cities, towns, school districts, and many kinds of special districts. These local governments provide a wide range of services, such as schools, streets and roads, elections, and police and fire protection. Most of these government bodies can impose taxes. CONFLICTS OVER FEDERALISM IV From the time the Constitution was adopted in 1789, questions about the exact boundary between state and national power have sparked frequent disputes. Using “states’ rights” as support for their cause, Southern states defended the practice of slavery as an issue for states to decide, while Northern states pushed to abolish slavery. Combined with economic and political tensions between the states, slavery led 11 Southern states to secede (withdraw) from the Union during 1860 and 1861. The American Civil War, which began in 1861, was partly a conflict over the proper role of national and state governments. The power of Congress to make laws affecting state and local issues grew steadily after the Civil War ended in 1865. The Supreme Court of the United States gave Congress more and more control over the states, often basing its decisions on the Commerce Clause in Article I, Section 8 of the Constitution. This clause empowers the federal government to regulate interstate commerce. By the late 19th century the Supreme Court interpreted this authority broadly, allowing Congress to take action on food quality, child labor, and other problems not specifically related to interstate trade. The Supreme Court further expanded the scope of congressional power under the Commerce Clause during the New Deal in the 1930s. The New Deal, which President Franklin Roosevelt created to confront the country’s economic depression, included laws affecting nearly every home and workplace. The Supreme Court upheld most of Roosevelt’s New Deal initiatives, including laws setting minimum standards for pay and working conditions, protecting labor unions, and regulating farm production. After World War II (1939-1945), national authority under the Commerce Clause continued to grow. The action of Congress against racial segregation stands as the most important expansion of national power in the postwar period. Some Southern states argued that the Tenth Amendment gave them the right to maintain segregation and that Congress had no authority to interfere in purely local matters. In 1964 the Supreme Court ruled in Heart of Atlanta Hotel v. United States that despite the rights reserved to the states by the Tenth Amendment, Congress had the authority to bar segregation because it could harm interstate commerce. The authority of the federal government has also grown as a result of a more gradual increase in power at all levels of government in the United States. Since the Constitution was adopted in 1789, national, state, and local governments all have assumed more powers and duties. They have been forced to do so by the increase in population, the growth of cities and towns, the rise of huge industries, and the ever-growing need for better roads, railways, and communication systems. Once strictly local, problems such as crime and transportation have become national issues. Many experts defend the growth in power of the federal government in the United States. They insist that the public interest demands federal control in cases involving more than one state. Other experts fear that the continued expansion of federal authority over state and local matters will create an inefficient and possibly dangerous concentration of power in federal hands.

When conflicts arise, the courts must decide how to balance states’ rights with the needs of the nation. Although federal courts tend to take a broad view of national powers, in the early and mid-1990s the Supreme Court issued several rulings that curtailed congressional power over the states. In 1992, for example, the court ruled that Congress could not require states to make laws controlling radioactive waste. The Court issued another important decision in 1997, ruling that Congress could not compel local law enforcement officers to conduct background checks on gun buyers. In 1999 the Court issued a series of rulings that further shifted power from the federal government to the states. In three 5-to-4 decisions that reflected the justices’ deep divisions over how to balance state and federal powers, the Court strengthened the principle of sovereign immunity, which gives states immunity from lawsuits arising from violations of federal law. In one case, the Court ruled that state employees cannot sue states for overtime wages due to them under federal labor laws. In the two other cases, the Court ruled that businesses cannot sue states in federal court for patent infringement or false advertising claims that violate federal law. In a similar decision in 2000, the Court ruled that states cannot be sued for violating a federal age-discrimination law. In 2001 the Court ruled that state employees cannot sue states for money damages based on employment-discrimination violations of the federal Americans with Disabilities Act. Combined with earlier rulings, the decisions reduced the ability of individuals to sue states for violating federal law. Fiscal Policy and Growth in Africa: Fiscal Federalism, Decentralization and the Incidence of Taxation Economic Commission for Africa Theories of Federalism By Ademola Ariyo By Stephen Njuguna Karingi, KIPPRA Page 2 Draft, not for citation Table of Contents 1. Introduction 1 2. Rationales for Decentralisation: Political and Economic Arguments 3. Forms of Decentralisation 4. Intergovernmental Relations and Financial Autonomy 5. Fiscal Federalism 7 5.1 The Principles of Fiscal Federalism 7 5.2 The Kenyan Lesson 12 Lessons for Kenya from International Best Practice in Fiscal Decentralisation 7. Conclusion 8.

1 2 5

6. 17 20 Page 3

1 Introduction Decentralisation is supported by political as well as economic arguments,

combining objectives as separate as enhancing economic efficiency and strengthening ‘democracy’. Devolution is one of the types of decentralisation. For that reason, any discussion on fiscal devolution would best be placed in context if there was consensus understanding of what decentralisation is and all about. In this paper therefore, we begin by seeking an understanding of the concept of decentralisation. We touch also briefly on the theoretical underpinnings of decentralisation and provide some motivation why decentralisation is popular. Attractive claims are made in support of decentralisation ranging from its capacity to enhance economic efficiency to strengthening democracy and promoting good governance. There are many forms of decentralisation, and it is important that it be appreciated that in practice, decentralisation processes need not be of any one neat type, but should draw elements from different forms. In fact, one would want to caution against easy generalisations of decentralisation that are made regarding its potential benefits. After the theoretical presentations we then give the Kenyan story of attempts to decentralise and draw lessons which should inform the current attempt to devolution in the on-going Constitutional reform. Rationales for Decentralisation: Political and Economic Arguments In a lay-man’s language, decentralisation is understood to mean the distribution and re-distribution of power and authority within the various levels of the modern government. But we also note that, decentralisation has also been recently associated with the transfer of certain types of power and authority to non-government and market based bodies (see Rondinelli 1986). Smith (1985) presents a discussion the political arguments for decentralisation are hinged on the liberal democratic tradition. Democrats see decentralisation as a way of increasing responsiveness of the government to the demands of individual citizens to the extent that one of the most compelling motivations for decentralisation is the strong correlation thought to exist between decentralisation and democracy. Decentralisation is therefore seen to promote such values as equality, responsiveness and accountability and in the liberal democratic tradition, it is believed that more responsive and accountable local governments would contribute to a more stable form of national democracy. Economic arguments for decentralisation on the other hand stem from the public choice theory and the new political economy school. The public choice theory advanced by authors such as Buchanan (1980) on its part is based in the school of rational individualism which believes that individuals are rational and are the best ‘maximisers’ of their own welfare making the markets very important Page 4 2 as it is through them that individual preferences are translated into choices that would in turn allow the satisfaction of individual wants. However, as a result of the existence of public goods, the Government is still recognised by the public choice theorists. The government to some should not restrict itself to the provision of public goods but should be in a position to control and regulate private economic activity as well. Nonetheless, due to the rational individualism operating in the case of those manning Government, rent-seeking behaviour is seen to prevail which not only impinges on individual freedom, but also causes the waste of social resources. The negative outcomes from the rent seeking behaviour provide the theoretical basis to the New Political Economy, which regards the state as predatory, and concludes that due to the government individual preferences are not maximised. Since governments are rent seeking, decentralisation is seen as a means to controlling them because it facilitates opportunities for the articulation of individual preferences. In other words, according to Mackintosh and Roy (1999) the heart of the economic argument for the decentralisation is that lower levels of government are more sensitive to individual preferences than are higher levels. Forms of Decentralisation

In discussing types of decentralisation, there are two important issues that come to mind. Firstly, one must distinguish between the nature of power that is being transferred and therefore classify the functions that the ‘decentralised unit’ can effectively perform. Secondly, it is important to consider the institutional nature of the decentralised unit, in order to differentiate between the extent of control that the central Government, continues to exercise over the decentralised unit, post the transfer of powers. There are four popular typologies of decentralisation: deconcentration, delegation, devolution and privatization as identified in Rondinelli and Nellis (1986). Administrative Decentralisation: Deconcentration and Delegation Administrative decentralisation seeks to redistribute authority, responsibility and financial resources for providing public services among different levels of the government. It forms an important aspect of ‘territorial decentralisation’, which specifically refers to the transfer of powers and resources, from higher levels of the government to local governments situated in different territories or regions within the country. Depending on the nature of the ‘decentralised unit’, administrative decentralisation is of two types, deconcentration and delegation. Deconcentration is the weakest form of (administrative) decentralisation, as it Page 5 3 distributes decision-making authority and financial and management responsibilities among the different levels of the central government. A more ‘extensive’ type of administrative decentralisation is ‘delegation’. Through delegation, central governments transfer decision-making and the administration of public functions to semi-autonomous organisations not wholly controlled by the central government, but ultimately accountable to it. Examples are public corporations, regional development agencies, special function authorities, semiautonomous project implementation units and a variety of parastatal organisations. None of the units to which powers are either ‘deconcentrated’ or ‘delegated’ are elected institutions. They also do not have powers to reformulate policies that affect their status or position within the broader decision-making structure, and can only take decisions on subjects transferred to them. Decentralisation on Fiscal and Economic Grounds The second type of decentralisation is that on fiscal and economic grounds. Today, there are various types of economic decentralisation that are recognised. Mackintosh and Roy (1999) distinguishes the various categories of what can be considered as economic decentralisation. One type of economic decentralisation is fiscal decentralisation, which concerns the transfer of fiscal and expenditure powers from the central government to lower levels of government. It includes intergovernmental transfers that shift general revenues from taxes collected by the central government to local governments for general or specific uses, the authorisation of municipal borrowing and the mobilisation of either national or local government resources through loan guarantees. Fiscal decentralisation can take many forms, and lower levels of government may be empowered to raise revenue through a range of methods including self-financing or cost recovery through user charges, co-financing or co-production arrangements through which users participate in providing services and infrastructure through monetary or labour contributions, and the levying of property or sales tax or indirect charges. While fiscal decentralisation primarily concerns transfers of powers from higher to lower levels of the government, at the other end of the spectrum of economic ‘decentralisations’ is a more controversial issue, i.e., privatisation. Privatisation of public sector economic activity, together with the liberalisation of sectors monopolised by government, is understood to be a form of decentralisation which is seen as the creation of ‘new decentralised economic activity within

markets, replacing public sector institutions with private firms as decisionmakers’. Another form of economic decentralisation is the decentralisation of public management decision-making, which basically involves decentralisation of the Page 6 4 management of public funds for the delivery of publicly funded services. This forms the core logic of the New Public Management. The second form, contracting is naturally one of the predominant forms that NPM assumes. This involves the contracting out of publicly funded service provision to private commercial or non-profit firms. It may be added that these different forms of economic decentralisation are extremely complex, as they involve institutions with varying institutional natures. These may be elected or non-elected institutions of the state, or non-state institutions from the market and civil society that work in ‘partnership’ with the state, in one guise or another. Embodying functional changes of a variety of sorts, different forms of fiscal decentralisation can be complementary either to ‘administrative’ decentralisation as explained above, or ‘political or democratic’ decentralisation, which we will describe next. Democratic Decentralisation The last major type of decentralisation is what is variously called political or democratic decentralisation. Some authors also use the term ‘devolution’ to loosely understand the same phenomenon. A key feature of this form of decentralisation is that it entails the transfer of power from higher levels of the government, both from elected institutions such as the national parliament or central bureaucracies, to local elected representatives of the people. Relying on electoral forms of representation, these types of decentralisation are theoretically derived from the basic proposition of democracy, which is simply that individuals by selecting representatives to their government from among themselves would be able to convey their preferences better to the latter. The argument for decentralisation is subsequently adapted from this core logic. Local governments comprised of representatives from individuals resident within their jurisdictions would be better informed of the preferences of their ‘constituencies’ and would be able to respond more effectively in order to meet their interests. Decentralisation in the Context of Public Sector Reforms: New Public Management ‘Reforming’ the public sector has been on the agenda in governments across the world for over two decades now, and different kinds of experiments in management and public-private partnerships have followed. These are important to regard because they form a part of the ‘good governance’ agenda of governments across the world. These decentralisation reforms of the public sector bear important implications for the continuing role of the state, and the position of citizens in their interrelationship with the state, given the importance of private ownership and provision within public sector reform. Public sector reform primarily involves different kinds of public-private partnerships. The common objective in these partnerships is to increase the Page 7 5 efficiency and effectiveness of ‘governance’. The public sector has long been criticised for being overstaffed, inefficient and corrupt. Reform initiatives mainly involve the adoption of market-based principles of functioning, with the underlying presumption that market based principles would allow for greater efficiency in the public sector. Other positive aspects such as an increase in responsiveness and accountability in ‘governance’ are also attributed to these processes of public sector reform. However, each form of public-private partnership involves a different kind of role for the state, which cannot simply be described as ‘minimal’.

The NPM historical origin is in the UK in the 1980s (see Ferlie et al. 1996 for discussions on the New Public Management). There is a very conscious and deliberate attempt to separate economic decision-making and the control over resources from the actual provision of services. The predominant form that the NPM assumes is one of contracting. Driven by objectives of efficiency and economy, the underlying belief is that contracting out of services by the central government to executive agencies, local government departments or private providers would embody the replacement of traditional hierarchical controls of government with the market as the central means of coordinating economic activity. Thus, contracting is driven by motives of ‘controlling social sector spending and reducing the leverage over the public services of both the trade unions and the local government’. Contracting encourages the ‘economic decentralisation of the management of public funds’. The underlying logic is that autonomous delivery agencies whether private providers, or even PSCs or Public Service Contractors, which are ‘those NGOs that function as market oriented non profit businesses’, would necessarily be more cost effective and efficient. This would result from their ‘relative isolation’ from the strong authoritarian and hierarchical social forms, which characterise and impede traditional government. Independent of these constraints, it is believed that private agencies, PSCs and NGOs would be able to deliver goods and services in ways that are regarded typical of market exchanges, i.e., neutral of considerations other than economic efficiency. Having considered the political and economic rationales for decentralisation and also having discussed albeit briefly the different types of decentralisation, this paper now shifts attention to fiscal devolution. But before that its important to note that it is always premised that decentralisation can be designed. In other words, each country can design its own form of decentralisation. So, what are the key issues in designing decentralisation? These are: territorial demarcation in design; nature and functions of institutions; the issue of ownership; intergovernmental relations; and financial autonomy. Fiscal devolution is related more to the question of intergovernmental fiscal relations and financial autonomy. Page 8 6 Intergovernmental Relations and Financial Autonomy Smith (1985) provides an exhaustive discussion on intergovernmental relations emphasising how critical they are in the framing of decentralisation policies. Intergovernmental relations basically refer to relations between different tiers or levels of government. Intergovernmental relations concern the links between different levels of government in a decentralised system, i.e., the centre, province and district. Decentralisation necessarily redefines relations between levels of a government to a greater or lesser degree. How effectively it does may have profound implications for the success of decentralisation, because despite the initiation of formal decentralisation policies, ‘unsuitable’ intergovernmental relations can engender tense relationships between central and local governments. Intergovernmental relations can affect the nature and extent of decentralisation, in its various forms, critically. For instance, at one extreme, central authorities might simply unload functions onto local government and abdicate any responsibility. At the other end, central government agencies act as facilitators and supporters of the local government. Strong proponents of decentralisation, sometimes view the centre’s role as overbearing particularly in the processes of devolution that entail the transfer of powers to elected local representatives. Others more suspicious of local politics and capacities suggest that accountability and service standards are best maintained by strong supervision, or at least monitoring, of the local government by the centre. The key issue then is regarding the attitude of the central government regarding the extent of control

that it wishes to relinquish, as well as the relative position of bargaining that different local governments may have vis-à-vis the centre. This last aspect can also depend on a range of factors like the respective political parties that are in power at the centre and the state, differing ideological rhetoric adopted by these parties and the significance of regional politics at the national level. The important thing to bear in mind is that within the same country, different subnational governments can have very different relations with the central government, which can potentially affect the uniform adoption of decentralisation policies. On the issue of financial autonomy, the nature of financial arrangements for local government functioning is a key index of the degree of decentralisation. Territories can be delineated, their functions allocated and institutions created, but if authority over revenue and expenditure is retained by the central government then decentralisation is partial at best. Fiscal decentralisation that concerns the transfer of fiscal and expenditure powers from the central government to lower levels of the government is a critical component of decentralisation policies. Page 9 7 We discuss the fiscal arrangements in detail in the following section. The point of raising this issue here is to emphasise their significance while devising new institutional arrangements that are to be viable financially. Fiscal Federalism The Principles of Fiscal Federalism The problems of a decentralised fiscal system have received much attention in the public finance literature. The fiscal structure in some countries like the US was designed to maximise the autonomy of the States and to ensure that the federal government only managed core fiscal functions (like foreign affairs and defence). In recent times, the World Bank, and aid agencies have begun to assert that decentralisation of economic powers - especially spending powers - from central government to provinces and cities is an essential part of an economic reform package. The shift in the thinking of multilateral organisations is cemented on arguments that are presented in World Bank (1997) and also in OECD (1997). These assertions are significantly influenced by theoretical developments in the discussion on fiscal federalism. Hence, it becomes important to discuss the basic theory of economic decentralisation. These theoretical issues begin with considering the basic fiscal functions of government and inquire whether each function, theoretically, belongs to central government, to the State government or to both. The functions of government require a theory of the State. So, what are these functions? The question to answer in thinking about fiscal devolution and economic development is the one which studies on fiscal decentralisation always begins with: Should the allocation, distribution and stabilisation functions be centralised or could they be decentralised? We consider each of these functions of the State below and try to answer this question which should help us in our discussions on the way forward for Kenya. The allocation function refers to the provisioning of public goods by the government, which are defined as goods that are produced by the public sector for the benefit of the community as a whole. Examples of such goods include justice, police, foreign affairs and defence. However, the allocation function may be broader in its remit depending on what is demanded by the community. Thus health education, water supply etc. may also be public goods. When discussing the allocation function it is important to distinguish between the provision of public goods and the production of public goods. For example a government may provide education either by running schools in its jurisdiction or by paying the fees of students who study in privately run schools. The allocation function is most directly related to the question of economic decentralisation: Should social goods and services be provided on a centralised or

Page 10 8 a decentralised basis? Some public goods are such that the incidence of their benefits is nation-wide (Like defence, foreign affairs etc) while others (rubbish collection, street lighting etc) are geographically limited in scope. The general conclusion that emerges from theories of economic decentralisation is that public goods and services should be provided by local government when their benefits accrue to a local population and by central government when the benefits accrue to the nation as a whole. This sounds fairly simple. However, in practice several problems may arise: in some cases, in the provision of public goods, the cost of providing a good may fall, the larger the number of consumers who benefit from it. There is the problem of congestion. Then problems may arise because of differences in preferences and income. In making the general proposition that allocation should be decided by the geographical remit of a public good we often forget that peoples preferences and incomes may differ across communities. Moreover, when we speak of “local” and “Central” units we also have to deal with the problem of optimum group size. And finally, there are two other complicating issues. The first one is that of mixed national-local goods. Some goods are both national and local in character. And the second one relates to the protection of minority preferences. It is assumed that the preferences of members of a community are those of its majority. But, what about minorities? The distribution function is defined as those activities undertaken by a government to bring about greater equality in income and wealth. The extent to which this function is a part of a government’s public policy remit differs according to country and historical context. However, in principle, it is accepted that if there is a concern with equalising distribution then that concern is addressed by the government and not by private agents. It would appear, in the first instance that policies to adjust the distribution of income amongst individuals should be conducted on a nation-wide basis because attempts to adjust within a local area can be frustrated by the “voting with the feet effect”. Hence it is a principle of economic decentralisation theory that fiscal redistribution should largely be a centralised function. However, this does not mean that the distribution question is not relevant to the theory of economic decentralisation. This is because even if one accepts that decentralised units cannot address the question of the National distribution of income, there still remains the question of equity between rich and poor regions. This brings in the key issue of fiscal equalisation. The stabilisation function refers to the use of economic policy to maintain a stable and smoothly running economy. This refers to the responsibility government has for ensuring an appropriate money supply, keeping the balance of payments under control, and ensuring that inflation is not excessive. Stabilisation too is viewed largely as a centralised function. Page 11 9 The major reason why stabilisation policies are considered to be best administered at the central level is this: regions and provinces in a country have no means of “insulating” themselves from policies adopted in other regions or provinces. Thus any policies applied to control inflation; unemployment or growth within a province will “leak” outside the province. For example suppose a region decides to control unemployment by increasing spending on public works and raising taxes. This may result in a situation where taxpayers emigrate to other provinces and the unemployed from other provinces come and find work in this province. This would frustrate the original aim of the policy. At the national level, controls on emigration make this phenomenon unlikely. Similarly when it comes to monetary policy it is impossible to imagine a situation where local authorities all have the right to print and distribute money, as this

would frustrate national control over the money supply. In summary the degree of fiscal and monetary co-ordination required to run stabilisation politicos is unlikely to happen in a decentralised setting. However, if such important parts of government policy - stabilisation and distribution - are centralised then that still raises the question of how these policies are executed at local levels. In the case of distribution, it is clear that solving or alleviating the problem of regional inequality involves multi-party negotiations between the Central government and local units. In the case of stabilisation, too, keeping inflation low, and ensuring solvency in the balance of payments involves ensuring that central and local policies are harmonised. Thus the problem of fiscal decentralisation does not end once we have decided on the appropriate assignment of functions between central and local governments. There still remains the question of how central and local governments interact with each other - their economic relationships are an integral part of the study of economic decentralisation. This is the issue of intergovernmental fiscal relations. Firstly, intervention by a higher level of government may be needed to correct spillover of benefits. In the model discussed above it was assumed that expenditures incurred on providing public goods in one area would only benefit the residents of that area. In practice, of course, this is not true. Economic theory cannot come up with any solution for this problem, except an extreme one: to centralise provision of all goods where spillovers occur! In practice however, this issue is often resolved through negotiation. In such a situation it becomes essential to work out some form of contractual arrangement between local areas that allow benefit spillovers to be recognised. Hence, we need inter-governmental contractual arrangements to deal with such spillovers. Such contracts may involve pooling resources, payments by provinces that benefit from spillovers to provinces that lose from them, or sharing of facilities to make spillovers redundant. Page 12 10 Secondly, the central government may consider certain local services to be of vital strategic importance and hence may wish to subsidise them. There are two kinds of strategic issues that arise in this context, where intergovernmental negotiation becomes necessary. The first has to do with cases where the use of a resource by one local area inhibits the use of a resource by another local area. A classic example of this is water supply. Suppose two provinces obtain water from the same river. This can become a major source of political contention, and some inter-jurisdictional negotiation is vital here. Clearly in such a situation some co-ordination becomes imperative. Provinces would need to come to an agreement regarding the appropriate property rights that each has to the common resource- the river. This is the most familiar example of a strategic issue, which may require central government negotiation or tripartite discussion between provinces and the central government. Another case could be that of pollution where, for instance, a city's pollutants are dumped in the countryside. The second type of strategic issue that arises in this context is that which occurs when it becomes impossible for local jurisdictions to come to agreements with each other because the number of “stakeholders” if an issue is too large. Here strategic intervention becomes essential. Thirdly, the process of allocation of public goods may itself raise issues of distribution equity across local jurisdictions. We have already recognised the principle that distributional issues are best dealt with at a central level. But is allocation independent of distribution? Can issues of equity be separated from issues of allocation? In economic theory, it is argued that governments may desire to ensure the availability of some essential goods and services in a quantity that is affordable to all for policy reasons. Hence, the “merit goods" argument may often warrant central intervention even in the allocation function. The other argument has to do with the assumption that all provinces have the same capacity to provide goods and services. This is patently untenable. Poorer

provinces have smaller tax bases and hence a lower capacity to provide goods and services compared to richer provinces. Hence, there may be a case of the central government intervening to ensure that poorer provinces do not seriously suffer in performing the allocation function on account of their relative poverty. For example, leaving the allocation function entirely to provincial governments may mean that children in richer provinces have a better education than children in poorer provinces. This may not be ethically acceptable to the population as a whole and hence there is here a case for central intervention to ensure common minimum standards. Second, it may be the case that in some provinces the cost of providing a service may be greater than in other provinces due to geographical or other considerations. Thus a province in a remote or hilly area may not be able to supply transport services at the same level of cheaper cost as a province in the plains. Provinces adjoining international borders may have to incur higher Page 13 11 expenditure on policing. Provinces prone to natural disaster may need extra assistance. Again, in such cases there may be a case for the central government to intervene in the allocation function. Lastly, intervention by the centre might be necessary because it may be, more advantageous to tax at central government rather than local government. Provinces may provide services. But should they also raise taxes (the question of tax efficiency)? As we have already seen this may give richer provinces an unequal advantage over poorer provinces. Benefit spillovers too complicate issues. Moreover by drawing on the entire economy as a tax base it may be possible to make the taxation system more efficient and cost effective. For these reasons it is often the case that central governments collect taxes and then distribute the proceeds to State governments according to some formula like population, per capita income etc. Thus, even if provinces provide public services, it may not be the case that provinces will finance them. Strategic, distributional and tax efficiency issues often involve pooling and devolution of resources by the Centre to the local government units. What happens is that a national tax base is created? This tax base is then devolved to local units taking account of distributional, strategic and efficiency considerations. The resources devolved to local units are called Grants. The term “Grants” should not be construed, as it often is, to mean hand-outs from the centre to local units, as finally all the resources of the State are equally the property of the centre and the local authorities. The theory of grants is concerned with how grants may best be devolved to address the problems of inter-governmental fiscal relations discussed above. One of the key issues is that of distribution which is a major plank when it comes to consideration of equitable development. In addition, the objective of fiscal equalisation is seen to be an important issue in a fiscally devolved system. The type of grant structures that best suits a country like Kenya thus depends largely on concrete circumstances. Economic theory of decentralisation does, however, provide some discussion of the relative merits of the different types of grants: matching versus non-matching grants; and general versus selective grants. We will only present here the conclusions that economic theory of grants arrive at and distil the recommendations that would be applicable to a country like Kenya. The most important policy implication on the choice between matching and nonmatching grants is that if a government wants to increase the consumption of specific public goods then matching grants are better than non-matching grants. This is because the scope for “leakage” of the grant into consumption of nonpublic goods is limited. However, an equally important policy implication is that, irrespective of whether a grant is matching or non-matching, there will always be some “leakage” from a grant! The general implication of this is that the scope for

Page 14 12 central government to “fine tune” public expenditure in local areas through the use of grants is limited. To this extent, since local governments are more in tune with local preferences, it can be argued using this theory that the bulk of grants should be non-matching grants, and matching grants should be used selectively. The other important issue in the case of a non-matching grant that is provided by government is whether it should be given as a general grant or as a selective grant. The policy implications of the ensuing analysis are interesting. Whether or not selective grants achieve their intended purpose depends on two things: the intensity of demand for public goods other than the good for which the selective grant is offered, and the magnitude of the selective grant. For example suppose a government wishes to use grants to increase the availability of primary education in a province. The analysis tells us that the extent to which this aim can be achieved though selective grants depends on (a) the extent to which the community values primary education, as opposed to other public goods like health, water supply etc. (b) the amount of money the centre is able to allocate as a selective grant. Small selective grants will not succeed in their aim. In general both these examples show us that community preferences are vital in determining the success of strategic central interventions using grants. Governments cannot assume that just because they “throw some money” at a project or a policy goal, that goal will be achieved. The Kenyan Lesson Decentralisation has been of central concern in Kenya. The attempts to undertake decentralisation in the form of devolution can be traced as far back as pre-independence time when the first Constitution was being prepared. The independence constitution contained a decentralisation system based on units identified as “Jimbo 1 ”. However, this system collapsed as soon as it was put in place and the country reverted to a Unitary State. The unitary state system (see Figure 1) consists of the Central Government, Local Authorities, State owned enterprises and extra-budgetary institutions. The Government structure can be split into two namely, administrative and economic structures. The administrative structure is exemplified by a provincial administration. Here the Government links to the grassroots or village level through the Provincial Commissioners, District Commissioners and the District Officers. The District Officers complete the chain through their link to the Chiefs and Assistant Chiefs. The economic structure on the other hand is more complex than the administrative one. This comprises Ministry headquarters at the top with Provincial Departmental heads and District Departmental Heads at the Province and District levels respectively. These link to the lowest levels through field 1 “Jimbo” is a Swahili word for region. The country was divided into regions by the colonial government as it left and each region had a significant level of political autonomy. Page 15 13 officers like extension workers, Hospital workers, and teachers. The Provincial Departmental Heads mainly provide an operational link to the Districts. Not many projects and programmes are implemented at the provincial level. Implementation mainly takes place at the District level. The Ministry Headquarters have a second line of service provision to the people through the state owned enterprises (Parastatals). Funds for service provision and budget implementation can therefore be channeled to the targeted areas through the Ministry Headquarters, through the District offices, or through Parastatals as grants. Figure 1 shows that what exists in Kenya today is an indication that the country never actually managed to design a decentralisation framework that would pass

the criteria for a sound decentralisation. However, decentralisation has remained an important issue as exemplified by the attention this is receiving in the new Constitution currently being drafted. Page 16 14 Figure 1: The Structure of Government in Kenya The current system of Government notwithstanding, there has been policy initiatives that have been undertaken over the years which have the semblance of an attempt to decentralise. The first example as noted includes the establishment of local authorities (even though all the powers are centred on the Minister for Local Government). The country’s decentralisation history can therefore be said to have began with local authorities. Even though the current local government system was laid in 1963, the local authorities have been in existence for much longer. A number of reforms have been made to strengthen local authorities. The reforms include the Sessional Paper No. 12 of 1967; the Transfer of Functions Central Government (Ministries/Departments) Province (Provincial Commissioner, Heads of Departments, Provincial Monitoring and Evaluation Committee) District (District Commissioner, Departmental Heads, District Development Committee, District Evaluation Committee) Division (District Officer, Divisional Heads, Sub-DDC) ) Location (Chief, Location Development Committee, Extension staff) Sub-Location (Assistant Chief, Extension staff, Sub-location Development Committee) Local Authorities Page 17 15 Act of 1969 and the Local Government Act of 1977. The local authorities at independence had various functions assigned to them. But the Transfer of Functions Act undermined the revenue base of local authorities by transferring major functions such as primary education, health, road maintenance as well as the Graduated Personal Tax to the Central Government. The result has been that apart from a few local authorities such as Nairobi and Mombasa which still have some schools and health facilities, almost all of the other authorities do not perform any of the functions that the Transfer of Functions Act touched on. The country has also enacted various White Papers to promote participatory development planning and management such as the highly acclaimed Sessional Paper No. 10 of 1965 on African Socialism, Sessional Paper No. 1 of 1986 on Economic Management, creation of Regional Development Authorities, and the formulation of the District Focus for Rural Development Strategy in the 1980s. The Sessional Paper No. 10 emphasised the importance of decentralised planning in Kenya in the hope of extending planning functions to provinces, districts, and municipalities. The objectives of the District Focus for Rural Development which became operational in July 1983 were to: shift decision making closer to the beneficiaries and the point of implementation; encourage and strengthen local participation; speed up development by prompting equity in resource allocation; and arrest rural-urban imbalance and increase employment opportunities. However, in spite of all these initiatives little seems to have been achieved in

Kenya in terms of devolution in particular. There has not since independence been significant restructuring of the existing Central Government and local government structures to facilitate communities participation in their own local development affairs. In this respect, the following are some of the weaknesses that can be drawn from the past decentralisation efforts: (a) There has been existence of different institutional arrangements which apply different approaches in service delivery; (b) Existence of local authorities with weak revenue bases and ineffective service delivery systems. This weakness has for the last few years been addressed through the scrapping of the local authority service charge which was being paid by the workers living in a given local authority with the Local Authority Transfer Fund (LATF) which is 5 per cent of the income taxes collected by the central government. The formula used in distributing this money recognises the variations in the demand for the quantity of service in different local authorities. (c) Poor governance characterised by corruption and political interference in the District Focus for Rural Development which led to improper prioritisation in the resource allocation process at the District level. Page 18 16 (d) Lack of a formal legal mandate and weak institutional framework rendering the District Development Committees decisions under the DFRD unenforceable. On the basis of the foregoing discussion, it is evident that Kenya has not achieved much in decentralisation and with respect to fiscal incidence and growth; the centre through the central government has had all the control. In this respect, the main conclusion that one can make is that attempts to decentralisation so far in Kenya have failed to achieve the types of decentralisations that have been witnessed elsewhere in Africa. But having said that, it is evident that the country has been keen for a decentralisation system that works and the current Constitution Review process has in the draft constitution proposed the principles of devolution; the organisation of these devolution principles; the powers of devolved governments and the intergovernmental relations. While the review process is not complete, the last leg of the discussion will start in November 2003 and consensus is starting to emerge on the devolution process that the country should adopt. The principles and objectives of devolution in the draft constitution are as follows among others: (a) To give powers of self-governance to the people at all levels and enhance participation of people and communities in the exercise of the power of the state; (b) Ensure democratic and accountable exercise of power; (c) Increase checks and balances and the separation of powers; (d) Ensure equitable sharing of national and local resources throughout Kenya, with special provisions for marginalised areas; (e) Facilitate the decentralisation of central government powers. There are four tiers of Governments below the national that have been envisaged based on the current existing administrative boundaries. The four levels of Government will be: village; location; district and province. The district will be the principal level of devolution powers. All the devolved authorities will be entitled to equitable share of revenue raised nationally to enable them to provide basic services. The devolved governments are also entitled to benefit from local resources to ensure that the benefit incidence to the local community is maximised. On the financial autonomy (fiscal decentralisation question), while the principle functions of the different tiers are yet to be fully determined, the National Government will be responsible for the collection of the major sources of revenue. The Districts may also impose taxes or levies. The principal of fiscal

Page 19 17 equalisation is inbuilt in the constitution such that the national revenue shall be shared equitably. More specifically, through a commission for local government finance, there will be unconditional grants; conditional grants; and equalisation grants that will be shared from the centre to the lower levels of Government. The intergovernmental relations are also defined in the draft and do not deviate significantly from the theoretical propositions discussed in earlier parts of this paper. Lessons for Kenya from International Best Practice in Fiscal Decentralisation Our discussion so far has argued that while there are both political and economic arguments for decentralisation, Kenya has been a failure in implementing decentralisation. However, the country is today determined to get it right as it undergoes a transition and as it redefines its institutions of government through a new Constitution. There are lessons that the country can draw from, which have been summarised in the literature by Roy Bahl as implementation rules for fiscal decentralisation. In this section we highlight what each of the implementation rules in decentralisation mean for Kenya. Fiscal decentralisation is a comprehensive system: What this implies for the country is that to implement, there is need to begin with the design of the system. The current exercise being in the right direction. But for it to work, it will be important to put in place the appropriate laws and being a system, piece-meal applications are likely to fail as has happened in the past. However, given the nature of the shocks that fiscal decentralisation are likely to have in the country, a ‘phased-in’ strategy is the best approach. Therefore, the gains from decentralisation may not necessarily be immediate. Finance follows functions: Essentially, the country will have to decide what are the functions to be carried out by the lower levels of government even before determining what taxes or grant formula to use. The function assignment must come first, followed by the revenue assignment, then the grant formula and finally whether the lower levels of government should be given powers to borrow. So far, the draft constitution has dealt with the expenditure and revenue assignment and also with the types of grants. The details of these require that the legislation to put it into effect follow the same order in dealing with the arising issues in fiscal decentralisation. There should be a strong central ability to monitor and evaluate decentralisation: It is anticipated that for fiscal decentralisation to work that the central government show leadership in financial accounting; auditing rules; disclosure requirements for borrowing; adjustment of grant formula; and technical assistance for lower levels of government. Moreover, the centre should have a strong fiscal analysis unit and an extensive data system. The reality of Kenya indicate that this is likely to be one of the greatest challenges to the success of Page 20 18 fiscal decentralisation in the country. The financial management systems in the country today are weak and the central capacity as it exists today is not sufficient both to carry out the required fiscal analysis and also monitor. There is therefore a significant risk that given the existing financial systems in the public sector and capacity for technical analysis of the issues in decentralisation the country’s decentralisation programme might turn out to be sub-optimal. Intergovernmental systems for urban and rural sectors are different: Due to differences in sub-national governments capabilities, the intergovernmental systems are different for different types. Experience even here in Kenya has been that higher levels of Government if one may call the City of Nairobi that are given more powers to tax and borrow while the lower or smaller levels of

government rely more on grants. The country will thus need to appreciate that intergovernmental relations can be different not only because of the likelihood of political differences in future between the centre and lower levels of government but also because capacities differ. Significant taxing powers should be given to local governments: It does not make much sense to assign expenditure functions without significant taxing powers. Decentralisation may not be complete if the taxing powers are limited. However, the country will have to develop capacity for property taxation, being the leading candidate of a tax that cannot easily be exported in terms of financing the activities at the lower levels of government. As the situation is today, most local governments in Kenya have very limited capacity in property taxation and the centre might have to be involved here initially in determining the valuation of properties in the different local governments. Central governments must make decentralisation rules that they make: The economic and political history of the country is that even where constitution and legislative provisions exist, the Government has tended to flout some of the rules. The rule of law has not always been seen to be sacrosanct and what this means is that there is a high risk that if there is no rule of law, decentralisation might not achieve its intended objectives. This is going to be a challenging area for the country but a risk that will need to be appreciated from the outset. Keep it simple: Given the diverse interests in the country and the fact that it is a young democracy, there has been a temptation to accommodate more and more concerns as they arise in the process of the constitution making. A case in point has been the number of districts that have been accepted as the recognised number in the draft constitution. The many demands mean that the country is likely to come up with a very complex decentralisation system. However, simple intergovernmental systems are easier to administer and they also minimise the political tensions that tend to be engendered by decentralisation processes. As a starting point, there is a strong case to keep Page 21 19 fiscal decentralisation simple since the current experience indicates that there is likely to be limited capacity at the local levels and it is only by keeping it simple that the costs of monitoring and evaluation can be minimised. Intergovernmental fiscal design should match decentralisation reforms objectives: Different designs have different objectives. In Kenya, it is clear that the political argument for decentralisation seem to far outweigh the economic argument. Indeed, there have been arguments that the territorial demarcation and the tiers of governments being proposed might result in unviable local government entities. This being the case, it will be important that the fiscal design match the political objectives. Financial autonomy both in the expenditure and revenue side will need to be designed in such a way that the political tensions that have led to the clamour for devolution are not engendered in the new government system. Fiscal decentralisation should consider all levels of Government: This rule has got more to do with a decentralisation situation where the territorial demarcation may result in some regions/provinces that are so large such that the provincial autonomy may compromise central control. This is unlikely to be a key issue in Kenya unlike the outcomes that have been witnessed in countries like India and China where decentralisation objectives with respect to the central control have been interfered with by the autonomy claims of larger provinces/regions. Impose a hard budget constraint: For decentralisation to work, it is important to have discipline at the lower level strongly enforced. A hard budget constraint which can take various forms such as – no deficit grants measures, no bailout of debt, or no direct coverage by central government of year-end shortfalls – should be put in place. Once again, this is likely to be a risk area for fiscal

decentralisation in Kenya given that the central government has always failed to operate under a hard budget constraint. Failure to do so at the local government levels would impose hard macroeconomic adjustment requirements for the country and may in the end lead to intergenerational inequality even for regions that are disciplined when the debts of undisciplined regions are taken up by the central government. Intergovernmental systems are always in transition: A realisation of this in the design process of decentralisation is likely to result in institutions that would facilitate the capturing of the transitional issues. For instance, the fiscal architecture or revenue capacity for different regions may change at different rates and in different directions. This would call for the adjustment of the grant formula in operation. Unless there is an institutional framework to address the challenges posed by fiscal architecture changes in different regions, tensions may emerge between different regions which might cause instabilities in the whole decentralisation system. Fortunately, the draft constitution for Kenya recognises this challenge and proposes to entrench a Page 22 20 local government fiscal commission which would deal with challenges such as changing fiscal architecture. There must be a champion for fiscal decentralisation: Theory and practice of decentralisation has shown that it is as much a political process as it is an economic process. Consequently, if there is no champion for fiscal decentralisation, there is a risk that the objectives may not be achieved. The failure of decentralisation attempts in the country can be attributed to lack a champion. Even today, there is a perception that the momentum that was there in the last few years have been lost, at least in the case of one of the wings of the ruling coalition. Unless the whole ruling coalition or the leading lights of the coalition from the highest level believe in fiscal decentralisation, there is a high likelihood that the process might fail. Lessons from other African countries will be useful here whereby the country can evaluate where decentralisation has succeeded and where it has failed, the level of involvement of political leadership. Conclusion This paper has looked at the theoretical arguments for decentralisation and some of the important theoretical issues that have to be dealt with in designing decentralisation. The paper then briefly looked at the history of decentralisation in Kenya and drew some of the weaknesses in this history. The conclusion that was reached in this brief discussion is that Kenya never really had any significant decentralisation and the various attempts made failed in their objectives. However, it was pointed out that currently the on-going Constitutional reform process is convinced about the need for decentralisation to succeed in the country. The paper noted that the political argument for decentralisation appears to be the main motivation for decentralisation as opposed to economic arguments. However, through a discussion of the rules of implementation for fiscal decentralisation based on international experience, it emerged that there are several risks that exist in particular related to capacity both at the national level and subsequent local governments which might significantly affect the nonattainance of an optimal outcome for decentralisation in the country. But having said that, there is no one neat type of decentralisation and it is possible that as the country continues to design its decentralisation, experience from the successes and failures in the African continent will inform greatly the final system that the country adopts. Page 23 21

© Metaphysics Research Lab, CSLI, Stanford University

Federalism First published Sun Jan 5, 2003; substantive revision Thu Oct 12, 2006 Federalism is the theory or advocacy of federal political orders, where final authority is divided between sub-units and a center. Unlike a unitary state, sovereignty is constitutionally split between at least two territorial levels so that units at each level have final authority and can act independently of the others in some area. Citizens thus have political obligations to two authorities. The allocation of authority between the sub-unit and center may vary, typically the center has powers regarding defense and foreign policy, but sub-units may also have international roles. The sub-units may also participate in central decision-making bodies. Much recent philosophical attention is spurred by renewed political interest in federalism, coupled with empirical findings concerning the requisite and legitimate basis for stability and trust among citizens in federations. Philosophical contributions have addressed the dilemmas and opportunities facing Canada, Australia and Europe, to mention just a few areas where federal arrangements are seen as interesting solutions to accommodating differences among populations divided by ethnic or cultural cleavages yet seeking a common political order.

•

1. Taxonomy

•

2. History of Federalism in Western Thought until 1900

•

3. Reasons for Federalism

•

•

○

3.1 Reasons for a federal order rather than separate states or secession

○

3.2 Reasons for preferring federal orders over a unitary state

4. Further Philosophical Issues

○

4.1 Sources of Stability

○

4.2 Allocation of Authority

○

4.3 Distributive Justice

○

4.4 Democratic Theory

Bibliography

○

Historical

○

Contemporary

•

Other Internet Resources

•

Related Entries

1. Taxonomy Several species of federal orders may be distinguished, and much valuable work contributes to distinguishing and explicating the central terms (cf. Wheare 1964, King 1982, Elazar 1987a, Riker 1993, Watts 1998). Federations are here taken to involve territorial divisions of authority, typically entrenched in the constitution which neither a sub-unit nor the center can alter unilaterally. In comparison, decentralized authority in unitary states can be revoked by the central legislative authority at will. Such entrenchments notwithstanding, some centralization often

occurs owing to the constitutional interpretations by a federal level court in charge of settling conflicts regarding the scopes of final legislative and/or judicial authority. In contrast, confederations have weaker centers than federations do. Typically, in a confederation a) sub-units may legally exit, b) the center only exercises authority delegated by sub-units, c) the center is subject to sub-unit veto on many issues, d) center decisions only bind sub-units but not citizens directly, e) the center lacks an independent fiscal or electoral base, and/or f) the sub-units do not cede authority permanently to the center. Confederations are often based on agreements for specific tasks, and the common government may be completely exercised by delegates of the sub-unit governments. Thus many would count as confederations the North American states during 1776-1787, Switzerland 1291-1847, and the present European Union. In asymmetric (con)federations the sub-units have different bundles of authority; some may for instance have special rights regarding language or culture. If the decisions made centrally do not involve sub-units at all, we may speak of Separate (Split or Compact) federalism. The USA is often given as example, since the two Senators from each state are not representing or selected by sub-unit (State) authorities but by electors voted directly by citizens — though this is by sub-unit decision (U.S. Constitution Art. II Section 1; cf. Dahl 2001). Federations can involve sub-units in central decision-making in at least two different ways in various forms of interlocking (or cooperative) federalism. Sub-unit representatives can participate within central bodies — in cabinets or legislatures - (collective agency/compositional arrangement); in addition they often constitute one central body that interacts with other such bodies, for instance where sub-unit government representatives form an Upper House with power to veto or postpone decisions by majority or qualified majority vote (divided agency/relational arrangements). Two processes of federalism may be identified (Stepan 1999). Independent states may come together by ceding or pooling sovereign powers in certain domains for the sake of goods otherwise unattainable, such as security or economic prosperity. Such ’coming together’ federations are typically arranged to constrain the center and prevent majorities from overriding a sub-unit. Examples include the present USA, Switzerland, and Australia. ’Holding together’ federations develop from unitary states, as governments’ response to alleviate threats of secession by territorially clustered minorities. Such federations often grant some sub-units particular domains of sovereignty e.g. over language and cultural rights in an asymmetric federation, while maintaining broad scope of action for the central government and majorities. Examples include India, Belgium, Canada and Spain. In addition to federations, other interesting alternatives to unitary states occur when non-territorial sub-units are constituted by groups sharing ethnic, religious or other characteristics. Karl Renner and Otto Bauer explored such arrangements for geographically dispersed cultural minorities, allowing them some cultural and ”personal” autonomy without territorial self rule (Bauer 1903; Renner 1907; Bottomore and Goode 1978; cf. Tamir 1993). Consociations consist of somewhat insulated groups in sub-units who in addition are represented in central institutions often governing by unanimity rather than by majority (Lijphart 1977). 2. History of Federalism in Western Thought until 1900 Johannes Althusius (1557-1630) is often regarded as the father of modern federalist thought. He argued in Politica Methodice Digesta (Althusius 1603) for autonomy of his city Emden, both against its Lutheran provincial Lord and against the Catholic Emperor. Althusius was strongly influenced by French Huguenots and Calvinism. As a permanent minority in several states, Calvinists developed a doctrine of resistance as the right and duty of "natural leaders" to resist tyranny. Orthodox Calvinists insisted on sovereignty in the social circles subordinate only to God's laws. The French Protestant Huguenots developed a theory of legitimacy further, presented 1579 by an author with the telling pseudonym "Junius Brutus" in Vindiciae Contra Tyrannos. The people, regarded as a corporate body in territorial hierarchical communities, has a God-granted right to resist rulers without rightful claim. Rejecting theocracy, Althusius developed a non-sectarian, non-religious contractualist political theory of federations that prohibited state intervention even for purposes of promoting the right faith. Accommodation of dissent and diversity prevailed over any interest in subordinating political powers to religion or vice versa. Fundamentally dependent on others for the reliable provision of requirements of a comfortable and holy life, humans require communities and associations that are both instrumentally and intrinsically important for supporting [subsidia] their needs. Families, guilds, cities, provinces, states and other associations owe their legitimacy and claims to political power to their various roles in enabling a holy life, rather than to individuals' interest in autonomy. Each association claims autonomy within its own sphere against intervention by other associations. Borrowing a term originally used for the alliance between God and men, Althusius holds that associations enter into secular agreements — pactum foederis — to live together in mutual benevolence.

Ludolph Hugo ((ca.) 1630–1704) was the first to distinguish confederations based on alliances, decentralized unitary states such as the Roman Empire, and federations, characterized by ‘double governments’ with territorial division of powers, in De Statu Regionum Germanie (1661) (cf. Elazar 1998; Riley 1976). In The Spirit of Laws (1748) Charles de Secondat, Baron de Montesquieu (1689-1755) argued for confederal arrangements to ensure the ideal scale of government required for political liberty understood as non-domination — that is, security against abuse of power. A ‘confederate republic’ with separation of powers secures the requisite homogeneity, identification and self-sacrifice within sufficiently small sub-units where the common good subdues private interests thus preventing tyranny and ‘internal imperfection’. The sub-units pool powers sufficient to secure external security, reserving the right to secede (Book 9, 1). Sub-units also serve as checks on each other, since other sub-units may intervene to quell insurrection and power abuse in one sub-unit. David Hume (1711-1776) disagreed with Montesquieu that smaller size is better. Instead, "in a large democracy ... there is compass and room enough to refine the democracy." In "Idea of a Perfect Commonwealth" (Hume 1752) Hume recommended a federal arrangement for deliberation of laws involving both sub-unit and central legislatures. Sub-units enjoy several powers and partake in central decisions, but their laws and court judgments can always be overruled by the central bodies, hence it seems that Hume’s model is not federal as the term is used here. He held that such a numerous and geographically large system would do better than small cities in preventing decisions based on “intrigue, prejudice or passion” against the public interest. Several 18th century peace plans for Europe recommended confederal arrangements. The 1713 Peace Plan of Abbé Charles de Saint-Pierre (1658-1743) would allow intervention in sub-units to quell rebellion and wars on nonmembers to force them to join an established confederation, and required unanimity for changes to the agreement. Jean-Jacques Rousseau (1712-1778) presented and critiqued Saint-Pierre’s proposal, listing several conditions including that all major power must be member, that the joint legislation must be binding, that the joint forces must be stronger than any single state, and that secession must be illegal. Again, unanimity was required for changes to the agreement. Immanuel Kant (1724-1804) defended a confederation for peace in On Perpetual Peace (1796). His Second Definite Article of a Perpetual Peace holds that the right of Nations shall be based on a pacific federation among free states rather than a peace treaty or an international state: "This federation does not aim to acquire any power like that of a state, but merely to preserve and secure the freedom of each state in itself, along with that of the other confederated states, although this does not mean that they need to submit to public laws and to a coercive power which enforces them, as do men in a state of nature." The Articles of Confederation of 1781 among the 13 American states fighting British rule had established a center too weak for law enforcement, defense and for securing interstate commerce. What has become known as the U.S. Constitutional Convention met May 25 — September 17 1787. It was explicitly restricted to revise the Articles, but ended up recommending more fundamental changes. The proposed constitution prompted widespread debate arguments addressing the benefits and risks of federalism versus confederal arrangements, leading eventually to the Constitution taking effect in 1789. The "Anti-federalists" were fearful of undue centralization. They worried that the powers of central authorities were not sufficiently constrained e.g. by a bill of rights (John DeWitt 1787, Richard Henry Lee) — which was eventually ratified in 1791. They also feared that the center might gradually usurp the sub-units’ powers. Citing Montesquieu, another pseudonymous ‘Brutus' doubted whether a republic of such geographical size with so many inhabitants with conflicting interests could avoid tyranny and would allow common deliberation and decision based on local knowledge (Brutus (Robert Yates?) 1787). In what has become known as The Federalist Papers, James Madison (1751-1836), Alexander Hamilton (1755-1804) and John Jay (1745-1829) argued vigorously for the suggested model of interlocking federal arrangements (Federalist 10, 45, 51, 62). Madison and Hamilton agreed with Hume that the risk of tyranny by passionate majorities was reduced in larger republics where sub-units of shared interest could and would check each other: "A rage for paper money, for an abolition of debts, for an equal division of property, or for any improper or wicked project, will be less likely to pervade the whole body of the Union than a particular member of it." (Federalist 10). Splitting sovereignty between sub-unit and center would also protect individuals’ rights against abuse by authorities at either level, or so believed Hamilton, quoting Montesquieu at length to this effect (Federalist 9). Noting the problems of allocating powers correctly, Madison supported placing some authority with sub-units since they would be best fit to address “local circumstances and lesser interests” otherwise neglected by the center (Federalist 37). Madison and Hamilton urged centralized powers of defense and interstate commerce (Federalist 11, 23), and argued for the need to solve coordination and assurance problems of partial compliance, through two new means: Centralized

enforcement and direct applicability of central decisions to individuals(Federalist 16, also noted by Tocqueville 1945). They were wary of granting sub-units veto power typical of confederal arrangements, since that would render the center weak and cause “tedious delays; continual negotiation and intrigue; contemptible compromises of the public good.” (Madison and Hamilton, Federalist 22; and cf. 20). They were particularly concerned to address worries of undue centralization, arguing that such worries should be addressed not by constraining the extent of power in the relevant fields, such as defense, but instead by the composition of the central authority (Federalist 31). They also claimed that the people would maintain stronger "affection, esteem, and reverence" towards the sub-unit government owing to its public visibility in the day-to-day administration of criminal and civil justice (Federalist 17). John Stuart Mill (1806-1873), in chapter 17 of Considerations on Representative Government (1861), recommended federations among "portions of mankind" not disposed to live under a common government, to prevent wars among themselves and protect against aggression. He would also allow the center sufficient powers so as to ensure all benefits of union — including powers to prevent frontier duties to facilitate commerce. He listed three necessary conditions for a federation: sufficient mutual sympathy "of race, language, religion, and, above all, of political institutions, as conducing most to a feeling of identity of political interest"; no sub-unit so powerful as to not require union for defense nor tempt unduly to secession; and rough equality of strength among sub-units to prevent internal domination by one or two. Mill also claimed among the benefits of federations that they reduce the number of weak states hence reduce temptation to aggression, ending wars and restrictions on commence among sub-units; and that federations are less aggressive, only using their power defensively. Pierre-Joseph Proudhon (1809-1865), in Du Principe fédératif (1863) defended federalism as the best way to retain individual liberty within ‘natural’ communities such as families and guilds who enter pacts among themselves for necessary and specific purposes. The state is only one of several non-sovereign agents in charge of coordinating, without final authority. Philosophical reflections on federalism during the 20th century has addressed several issues, including reasons for federalism to concerns for stability, the legitimate allocation of authority between sub-unit and center, distributive justice and challenges to received democratic theory. 3. Reasons for Federalism Many arguments for federalism have traditionally been put in terms of promoting various forms of liberty in the form of non-domination, immunity or enhanced opportunity sets (Elazar 1987a). When considering reasons offered in the literature for federal political orders, many appear to be in favor of decentralization without requiring constitutional entrenchment of split authority. Two sets of arguments can be distinguished: Arguments favoring federal orders compared with secession and completely independent sovereign states; and arguments supporting federal arrangements rather than a (further) centralized unitary state. They occur in different forms and from different starting points, in defense of ’coming together’ federalism, and in favor of ’holding together’ federalism. 3.1 Reasons for a federal order rather than separate states or secession - Federations may foster peace, in the senses of preventing wars and preventing fears of war, in several ways. States can join a (con)federation to become jointly powerful enough to dissuade external aggressors, and/or to prevent aggressive and preemptive wars among themselves. The confederate American states moved to a federation largely for the first of these reasons, since the center powers of the Confederacy were too weak for protection from external threats. The European federalists Altieri Spinelli, Ernesto Rossi and Eugenio Colorni argued the latter in the 1941 Ventotene Manifesto: Only a European federation could prevent war between totalitarian, aggressive states. Such arguments assume, of course, that the (con)federation will not become more aggressive than each state separately, a point Mill argued. - Federations can promote economic prosperity by removing internal barriers to trade, through economies of scale, by establishing and maintaining inter-sub-unit trade agreements, or by becoming a sufficiently large global player to affect international trade regimes (for the latter regarding the EU, cf. Keohane and Nye 2001, 260). - Federal arrangements may protect individuals against political authorities by constraining state sovereignty, placing some powers with the center. By entrusting the center with authority to intervene in sub-units, the federal arrangements can protect minorities’ human rights against sub-unit authorities (Federalist, Watts 1999). Such arguments assume, of course, that abuse by the center is less likely. - Federations can facilitate other ends where credible commitments, coordination or control over ‘spill-over’ effects and externalities by sovereign states is desired but difficult to ensure without transferring some powers to a common body. Since cooperation in some areas can create a need for further coordination in other sectors, federations often exhibit creeping centralization.

- Federal arrangements may enhance the political influence of formerly sovereign governments, both by facilitating coordination, and - particularly for small states — by giving these sub-units influence or even veto over policy making, rather than remaining mere policy takers. - Federal political orders can be preferred as the appropriate form of nested organizations, for instance in ’organic’ conceptions of the political and social order. The federation may promote cooperation, justice or other values among and within sub-units as well as among and within their constituent units, for instance by monitoring, legislating, enforcing or funding agreements, human rights, immunity from interference, or development. Starting with the family, each larger unit responsible for facilitating the flourishing of sub-units and securing common goods beyond the reach of sub-units without a common authority. Such arguments have been offered by such otherwise divergent authors as Althusius, the Catholic traditions of subsidiarity as expressed by pope Leo XIII (1891) and Pius XI (1931), and Proudhon. 3.2 Reasons for preferring federal orders over a unitary state - Federal arrangements may protect against central authorities by securing immunity and non-domination. Constitutional allocation of powers to a sub-unit protects individuals from the center, while interlocking arrangements provide influence on central decisions via sub-unit bodies (Madison, Hume, Goodin 1996). Sub-units may thus check central authorities and prevent undue action contrary to the will of minorities: "A great democracy must either sacrifice self-government to unity or preserve it by federalism. The coexistence of several nations under the same State is a test, as well as the best security of its freedom . . . The combination of different nations in one State is as necessary a condition of civilized life as the combination of men in society" (Acton 1907, 277). - Federal orders may increase the opportunities for citizen participation in public decision-making; through deliberation and offices in both sub-unit and central bodies that ensures character formation through political participation among more citizens (Mill 1861, ch. 15). - Federations may facilitate efficient preference maximization more generally, as formalized in the literature on economic and fiscal federalism - though many such arguments support decentralization rather than federalism proper. Research on ’fiscal federalism’ addresses the optimal allocation of authority, typically recommending central redistribution while local provision of public goods. Federations may thus allow optimal matching of the authority to create public goods to specific affected subsets of the populations. If individuals' preferences vary systematically by territory according to external or internal parameters such as geography or shared tastes and values, federal — or decentralized — arrangements that allow local variation are favored, for several reasons. Local decisions prevent decision-making from becoming overloaded, and local decision-makers may also have a better grasp of affected preferences and alternatives, making for better service than would be provided by a central government that tends to ignore local preference variations (Smith 1776, 680). Granting powers to population subsets that share preferences regarding public services may also increase efficiency by allowing these subsets to create such ‘internalities’ at costs borne only by them (Musgrave 1959, 179-80, Olson 1969, Oates’ 1972 ’Decentralization Theorem’). Federal arrangements also shelter territorially based groups with preferences that diverge from the majority population, such as ethnic or cultural minorities, so that they are not subject to majority decisions severely or systematically contrary to their preferences. Non-unitary arrangements may thus minimize coercion and be responsive to as many citizens as possible (Mill 1861 ch. 15, Elazar 1968; Lijphart 1999). Such considerations of economic efficiency and majority decisions may favor federal solutions, with "only indivisibilities, economies of scale, externalities, and strategic requirements ... acceptable as efficiency arguments in favor of allocating powers to higher levels of government" (Padou-Schioppa 1995, 155). - Federal arrangements may promote mobility and hence territorial clustering of individuals with similar preferences, and allow sub-unit autonomy to experiment and compete for individuals who are free to move where their preferences are best met. Such mobility towards sub-units with like-minded individuals adds to the benefits of local autonomy over the provision of public services — absent economies of scale and externalities (Tiebout 1956, Buchanan 2001). 4. Further Philosophical Issues Much recent attention has focused on philosophical issues arising from empirical findings concerning federalism, and has been spurred by dilemmas facing Canada, Australia and Europe. 4.1 Sources of Stability As political orders go, (con)federal political arrangements pose peculiar problems concerning stability and trust. Federations tend toward disintegration in the form of secession, or toward centralization in the direction of a unitary state. Such instability should come as no surprise given the tensions typically giving rise to federations in the first place. Federations are often marked by a high level of ’constitutional politics’: Political parties often disagree on

constitutional issues regarding the appropriate areas of sub-unit autonomy, the forms of cooperation and how to prevent fragmentation. Such sampling bias among states that federalize to hold together makes it difficult to assess claims that federal responses perpetuate cleavages and fuel rather than quell secessionist movements. Some nevertheless argue that democratic, interlocking federations alleviate such tendencies (Simeon 1998, Simeon and Conway 2001, Linz 1997; cf. McKay 2001, Filippov, Ordeshook and Shvetsova 2004). Many authors note that the challenges of stability must be addressed not only by institutional design, but also by ensuring that citizens have an ‘overarching loyalty’ to the federation as whole in addition to loyalty toward their own sub-unit (Franck 1968, Linz 1997). The legitimate bases, content and division of such a public dual allegiance are central topics of political philosophies of federalism (Norman 1995a, Choudhry 2001). Some accept (limited) appeals to considerations such as shared history, practices, culture, or ethnicity for delineating sub-units and placing certain powers with them, even if such ‘communitarian’ features are regarded as more problematic bases for (unitary) political orders (Kymlicka 1995, Habermas 1996, 500). The appropriate consideration that voters and their sub-unit politicians should give to the interests of others in the federation in interlocking arrangements must be clarified if the notion of citizen of two commonwealths is to be coherent and durable. 4.2 Allocation of Authority Another central topic is the critical assessment of alleged grounds for federal arrangements in general, and the allocation of authority between sub-units and central bodies in particular, indicated in the preceding sections. A related important issue is who shall have the authority to make such allocations and changes (Recent contributions include Knop et al. 1995, Kymlicka 2001, Kymlicka and Norman 2000, Nicolaidis and Howse 2001). The “Principle of Subsidiarity” has recently received attention owing to its inclusion in European Union treaties. It holds that authority should rest with the sub-units unless allocating them to a central unit would ensure higher comparative efficiency or effectiveness in achieving certain goals. This principle can be specified in several ways, for instance concerning which units are included, which goals are to be achieved, and who has the authority to apply it. The principle has multiple pedigrees, and came to recent political prominence largely through its role in quelling fears of centralization in Europe - a contested role which the principle has not quite filled (Fleiner and Schmitt 1996, Burgess and Gagnon 1993, Follesdal 1998). 4.3 Distributive Justice Regarding distributive justice, federal political orders must manage tensions between ensuring sub-unit autonomy and securing the requisite redistribution within and among the sub-units. Indeed, the Federalists regarded federal arrangements as an important safeguard against “the equal division of property” (Federalist 10). The political scientists Linz and Stephan may be seen as finding support for the Federalists’ hypothesis: Compared to unitary states in the OECD, the ‘coming together’ federations tend to have higher child poverty rate in solo mother households and a higher percentage of population over-sixty living in poverty. Linz and Stepan explain this inequality as stemming from the ‘demos constraining’ arrangements of these federations, seeking to protect individuals and sub-units from central authorities, combined with a weak party system. By comparison, the Constitution of Germany (not a ’coming together’ federation) explicitly requires equalization of living conditions among the sub-units (Art. 72.2) Normative arguments may also support some distributive significance of federal arrangements, for instance owing to legitimate trade-offs between sub-unit autonomy and redistributive claims among sub-units (Follesdal 2001). A central normative issue is to what extent a shared culture and bonds among citizens within a historically sovereign state reduce the claims on inter-sub-unit redistribution. 4.4 Democratic Theory Federalism raises several challenges to democratic theory, especially as developed for unitary states. Federal arrangements are often more complex, thereby challenging standards of transparency and accountability. The restricted political agendas of each center of authority also require defense (Dahl 1983; Braybrooke 1983). The power that sub-units wield in federations often restricts or violates majority rule, in ways that merit careful scrutiny. Federal political orders typically influence individuals’ political influence by skewing their voting weight in favor of citizens of small sub-units, or by granting sub-unit representatives veto rights on central decisions. Minorities thus exercise control in apparent violation of principles of political equality and one-person-one-vote — more so when sub-units are of different size. These features raises fundamental normative questions concerning why sub-units should matter for the allocation of political power among individuals who live in different sub-units. Federations are often thought to be ’sui generis’ , one-of-a-kind deviations from the ideal-type unitary sovereign state familiar from the Westphalian world order. Indeed, every federation may well be federal in its very own way, and not easy to summarize and assess as an ideal-type political order. Yet the phenomenon of non-unitary sovereignty is not

new, and federal accommodation of differences may well be better than the alternatives. When and why this is so has long been the subject of philosophical, theoretical and normative analysis and reflection. Such arguments may also contribute to the overarching loyalty required among citizens of stable, legitimate federations, who must understand themselves as members of two commonwealths.