Legislative, Executive, and Judicial Institutions: A Synthesis Legislative, Executive, and Judicial Institutions: A Synthesis CHERYL SAUNDERS1 COMPARING FEDERATIONS: SIMILARITY AND DIFFERENCE The significance of the structure and operation of the institutions of government within both spheres of government in all federations was noted in the introductory chapter to this volume. The subsequent chapters described and analyzed the legislative, executive, and judicial institutions of 11 of the world’s 25 federations:2 Argentina, Australia, Austria, Canada, Germany, India, Nigeria, Russia, South Africa, Switzerland, and the United States. The purpose of this chapter is to draw the results together and to assist in understanding them by placing the experiences of the countries in context and by making comparisons between them. Effective comparison requires awareness of both similarity and difference. In comparing federal institutions, the similarities are clear enough. Each country has a federal political system, and each has key institutions that, at one level, are broadly similar in kind. The differences are more complex and demand more careful attention. This introductory section groups some of the principal contextual differences by reference to the size and wealth of the respective federations, characteristic features of their federal arrangements, and the general constitutional and political framework of which their institutions of government are a part. These should be borne in mind in evaluating the institutional arrangements of the 11 federations. Size and wealth First, the 11 federal polities vary in geographic area, population size, and economic prosperity. Table 1 illustrates the range. It shows one group of countries that are territorially vast, ranging from Russia with 17 million square kilometres to Australia with 7.7 million; a medium-size group, the largest of which is Argentina with 2.8 million square kilometres and the smallest of which is Germany with 357,023; and two states, Austria and Switzerland, with 83,870 and 41,290 square kilometres, respectively. Very large territorial areas present particular challenges for the coverage of both national institutions and the institutions of larger constituent units. It is more difficult in these circumstances, for example, to balance effective local representation with regular and consistent meetings of the national, or even a constituent unit, legislature.3 Differences in population size are equally marked. They do not necessarily correlate with geographic area, although, unsurprisingly, the populations of Austria and Switzerland are at the smaller end of the scale while Russia and the United States have substantial populations of 144 million and 293 million, respectively. But Canada and Australia, both of which have large or relatively large territories, also have small population sizes, with at least some communities dispersed over lightly populated areas, while India, with a medium-size territory, has one of the largest populations in the world. Both the overall size of the population and the degree of population concentration or dispersal affect the tasks that must be performed by the institutions of government. Population size is not necessarily reflected in the number of constituent units in a federal political system, although all else being equal, it must be a relevant consideration. There are, however, significant variations in the number of units in the various federations in this volume, as table 1 shows.4 The total number of units in a federation may have some implication for the capacity of unit institutions, and it will necessarily affect the design of national institutions in which the units are represented as well as the conduct of intergovernmental affairs. The economic prosperity of a federation is a factor of a different kind, with the capacity to affect the design and operation of institutions and the services they provide. In this regard also, in this sample of federations, the differences are substantial. Per capita GDP admittedly provides only a single guide. It shows, however, a disparity of 40 to 1 between Nigeria (with a per capita GDP of $US1,000 in 2004) and the United States (with a per capita GDP of $US40,000). Five other federations are clustered with the United States at the more prosperous end of the scale: Switzerland, Canada, Austria, Australia, and Germany. The latter has the lowest GDP of this group, at $US28,700. All the remainder have a GDP that is less, and in some cases substantially less, than half the German figure.5 Federal systems A second group of differences goes to the nature of each federal political system: the manner of its formation; the viability of its constituent units; the form of its federal division of power; and the depth of its federal culture. The significance of some of these factors is obvious. In a federation formed by uniting existing polities, both spheres of government are more likely to have their own constitutions, with the accompanying potential for institutional innovation on the part of the constituent units, than would be the case in a federation formed from a polity that is already united, in form or effect, or by a process that combines aggregation and devolution.6 This distinction may also have consequences for the extent of the autonomy of the constituent units and for the duplication of certain institutions between the two spheres. The causal link is not invariable, but it helps explain the relatively greater institutional autonomy of the constituent units in Argentina, Australia, Switzerland, and the United States as opposed to those in India, Nigeria, and South Africa. A contrast of another kind, which also has implications for autonomy, can be drawn between federations in which some constituent units lack the capacity to deliver government services and federations in which the basic capacity of all units is deemed acceptable. The case for a national power to intervene is stronger in the case of the former, at some cost to federal principles. Unless problems of capacity are of a transitional nature, this factor suggests that, all else being equal, constituent units should be delineated with capacity in mind. This consideration has particular relevance in Nigeria, where the original three regions, one of which (the northern region) was disproportionately large, were progressively subdivided into what are now 36 states, weakening the state sphere in the process.7 The form of the federal division of power has a significant impact on institutions, in a manner that will be elaborated later in this chapter. The principal contrast is between federations in which power is divided vertically, by reference to subject matter, and those in which a horizontal division, by reference to function, is also used. In the former, of which Argentina, Australia, Canada, and the United States are examples, each jurisdiction enacts and administers its own legislation and has institutions to match. In their institutional structure, although not necessarily in their formulation and implementation of policy, such federations may be described as dualist. In the latter, of which Germany, Austria, Switzerland, and South Africa are examples, power is divided both vertically and horizontally, leaving to the constituent units both the responsibility and the right to administer much federal legislation. A division of this kind increases the need for each sphere of government to rely on the other and is typically marked by a greater integration of the institutions of the several spheres. The participation of the governments of the German Länder in the national law-making process through the Bundesrat, or Federal Council, is a good example. The division of financial resources between the spheres of government also has institutional implications. The requirement for a legislature to approve taxation and expenditure is one of the principal checks and balances in a system of representative government. Where constituent units depend heavily on transfers from the national government, their institutions are necessarily affected to a degree, requiring the adaptation of traditional principles and practices to these altered circumstances.8 Formalized tax-sharing is one common response, which also constrains the ability of the national government to exert undue influence over the constituent units through its control of financial resources (as occurs, for example, in Argentina). The structure and operation of the institutions of both spheres may be affected by the country’s commitment to federalism. Federal culture is the product of many factors,9 one of which is the diversity of the people served by the federation. Population diversity is principally relevant to an understanding of federal institutions where it is linked with the delineation of the constituent units, creating what has been described as a pluri-national,10 or multinational, federation.11 Switzerland, Canada, India, Nigeria, and Russia are examples. By contrast, Australia, South Africa, and the United States can be distinguished as “territorial” federations, despite the multicultural composition of their peoples.12 The nature of the impact of population diversity on federal institutions varies. In Switzerland diversity has been central to the growth of a political culture that is consensual rather than majoritarian; in which the twin poles of self-rule and shared-rule are valued; and in which national minorities can share in the ownership of the state.13 Interest in consensus is more spasmodic in the other four pluri-national federations, to the extent that it exists at all. In each of them, however, population diversity manifests itself in ways that affect the institutions of government. Thus, the legal system and institutions in francophone Quebec are significantly different from those in the rest of Canada; in Russia and Canada a degree of asymmetry in the structure and operation of the federation is a response to substantial cultural difference between constituent units;14 in Nigeria population diversity underpins efforts to mandate equity in appointments to the institutions of government through the federal-character principle. In India the linguistic diversity of the people has driven successive changes to state boundaries and has been a prime cause of the transition of a highly centralized federation into one that Rajeev Dhavan and Rekha Saxena describe as “negotiatory” in character. Constitutional and political framework Finally, there is a host of relevant differences between other aspects of the constitutional and political systems of these 11 federations. Those that are central to the design and operation of the institutions of government are the subject of this book and are examined in detail in the remainder of this chapter. Others, however, contribute to the context within which the institutions can be understood. These include the legal system, the configuration of political parties and the patterns of their support, the stability of the political system and the depth of the constitutional culture, and the availability and use of emergency power.15 Of the federations in this study, three – Australia, India, and the United States16 – have common-law legal systems; another three – Canada, South Africa, and Nigeria – have mixed legal systems that, in the field of public law, are predominantly common law in character. The remainder have civil-law systems. The background legal system affects the institutions of government in a variety of ways. In civil-law systems legislative codes are the predominant source of law. In common-law systems the law derives from either legislation or judicial decisions, and the latter have precedential value. This difference informs the federal distribution of powers and has institutional consequences for the organization and operation of the courts. Other differences between legal systems affect intergovernmental relations. In particular, they explain why civil-law countries are more likely to require prior legislative consent to intergovernmental agreements than are common-law countries.17 Political parties are another dimension of the political and constitutional context that affects the operation of government institutions, both within and between the spheres of government. Parties have significance for the federal balance itself. The dominance of the African National Congress (ANC) in all spheres of government in South Africa reinforces centralized decision making, discouraging independent action in the provincial sphere; the dynamics of Indian federalism changed once the rise of regional parties eclipsed the authority of the Congress Party; the regionalization of parties in Canada adds weight to the influence of the provinces; and political alignments across jurisdictional boundaries encourage collaboration between governments that, taken to an extreme as in Argentina, undermine constitutional safeguards. Levels of party support also affect the operation of particular institutions. Different political majorities in the two chambers of a national legislature will lead to disagreement and may lead to deadlock, as the chapter on Germany shows. Over a long period of time, identical majorities may result in the need for a second chamber to be questioned altogether, as Anna Gamper notes in the Austrian context. The consequences of party configuration are easy to see; their causes are less easy to trace. Electoral systems are one, but only one, variable. Of the common-law federations in this study, all but India have majoritarian electoral systems, but the outcomes vary dramatically, from the tightly disciplined, effective two-party system in Australia, to the more loosely disciplined two-party system in the United States, to the significantly more regionalized parties in Canada. Similarly, forms of proportional representation produce both a relatively small and stable range of parties in Germany, Austria, and Switzerland and the hegemony of the ANC in South Africa, although the latter can be explained by reference to distinctive local factors. Just as the depth of federal culture was described earlier as a significant, if sometimes nebulous, influence on the operation of federal institutions, so political stability and constitutional culture are influences as well. The federations in this study vary significantly in these respects. Australia, Canada, Switzerland, and the United States have enjoyed established constitutional systems for well over a century, and Germany and Austria have been constitutionally stable for more than 50 years. By contrast, in both South Africa and Russia the constitutional system is relatively new and follows decades of undemocratic or authoritarian rule, while both Argentina and Nigeria have had long periods of dictatorship, interspersed with constitutional government (which, in both cases, has been re-established relatively recently). It is not surprising, in these circumstances, to find more serious problems of capacity and performance in the institutions of this latter group of countries. One further factor is interwoven with these differences. The constitutional arrangements of all of the four less established federations provide for the exercise of extraordinary power in emergencies in ways that affect both federalism and democracy. All except South Africa, moreover, have a history of using such powers, sometimes for long periods and always with significant effect. A fifth country, India, occupies the middle ground between the two groups of federations in this respect. India has also been a federal constitutional democracy for more than 50 years, but constitutional stability has been punctuated by exercises of emergency power, sometimes affecting only particular states and sometimes affecting the country as a whole. Reaction against the excesses of emergency rule in the 1970s has now diminished the likelihood of its widespread use, but the this power continues to be used in relation to individual states, with destabilizing effect. The case studies thus illustrate a familiar conundrum. Provision of a specific emergency power brings with it the temptation to use and abuse it. In the absence of specific provision, however, a genuine emergency may cause the constitution to be flouted or abrogated altogether. The preferable course depends on the circumstances of the country concerned, bearing the lessons of both history and comparative experience in mind. A comparative model that merits attention is the Constitution of South Africa, which strictly confines the circumstances in which an emergency may be declared and the period for which it may last, requires the declaration to be made by an act of the Parliament, subjects the decision to judicial review, protects core fundamental rights from the exercise of emergency powers, and precludes indemnity for unlawful action taken during this time.18 Table 1 Contextual differences between the 11 federations* Arg Aust Austria Can Germ Ind Nig Russ SA Switz US Geog.size (m. sq kms) 2.8 7.7 0.08 10 0.35 1.2 0.9 17 1.2 0.4 9.6 Pop.size (m)** 37.5 20.3 8.2 32.8 82.4 1027 130 144 44.8 7.3 293 GDP 2004 (’000)# 12.4 30.7 31.3 31.5 28.7 3.1 1.0 9.8 11.1 33.8 40.0 Nos of units## 23 6 9 10 16 28 36 88 9 26 50 Legal system civ Com civ com+ com com com+ civ com+ civ com Stable since~ 1983 1901 1945 1867 1949 1950 1999 1993 1994 1848 1789 Emergency power yes No no no no yes yes yes yes no no Political system pres Parl parl parl parl parl pres pres parl mix pres * As at 2005, unless otherwise indicated. ** Approximate: based on most recent census or other recent estimates. # Per capita GDP in USD. ## Constituent units only; self-governing and other territories are not included. ~ The date from which (broadly unbroken) constitutional government runs. NATIONAL INSTITUTIONS General The federal character of a system of government is generally reflected in the form and operation of the institutions of the national government in a variety of ways. This part surveys that range before I go on to examine particular institutions in detail. It shows that the federal form of the state typically has a profound impact on national institutions and that there is likely to be considerable interaction between the institutions of the spheres of government. In some cases, the impact of federalism on national institutions is a natural consequence of the territorial division of the country into federal units and the development of political activity by reference to them. In most federations, for example, the electoral divisions for the chamber of the national legislature that is elected broadly by reference to population numbers (the “popular” house or chamber) are influenced in some way by the boundaries of the constituent units. The organization and operation of political parties are also likely to reflect the federal structure of the country, causing greater attention to be paid to regional concerns in some federations (e.g., Canada, India, and the United States), although having little or no effect in others. The boundaries of constituent units are used for organizational purposes in relation to other national institutions as well. In most federations, they effectively dictate the placement of the regional offices of the national administration. In federations in which there is a distinct federal court hierarchy (e.g., the United States, Australia, and Argentina), the design of judicial districts is likely to be guided by the boundaries of the constituent units. In a second category of cases, the influence of federalism on national institutions is attributable to a concern for unit representation. Famously, this is the principal rationale for the composition and powers of federal chambers in national legislatures (see below). As the chapters in this volume show, however, concern for unit representation can affect the composition of popular chambers as well. In some federations, of which Argentina and Australia are examples, units are guaranteed a minimum number of seats, notwithstanding their population size. In an extreme manifestation of this concern in India, constituency divisions have been frozen until 2026, denying effect to the altered patterns of population distribution revealed by census data, so as to limit the domination of the populous northern Indian states in the Lok Sabha. Constituent units are frequently represented in other national institutions too, although the details vary considerably between federations. In Canada appointments to the Supreme Court are apportioned between provinces or regions in historical shares; in most federations there is some kind of consultation with units regarding appointments to the principal constitutional court; in many federations cabinets are formed with an eye to unit representation as a political imperative. Unusually, Nigeria has formalized this procedure through a constitutional requirement for the appointment of at least one cabinet minister from each of the 36 states. In an example of another kind, the constituent units are usually somehow involved in decisions to amend the national constitution, or at least those parts of it that involve the federal structure. In some cases they are involved directly through a double-majority referendum requirement, as in Switzerland or Australia; approval of a proposed change by a proportion of unit legislatures, as in Canada; or an option either to initiate or to approve proposed changes, as in the United States. In most other cases their involvement is indirect, through the consent of the federal legislative chamber. In cases of this kind, the units have an effective voice only to the extent that the federal chamber performs a substantive federal function. In multinational federations, where unit boundaries coincide with linguistic, religious, or other cultural divisions, requirements for national institutions to reflect cultural diversity may also, de facto, involve unit representation. The representation of the three major language groups in the Swiss Federal Council is a case in point. The impact of the federal-character principle on the composition of institutions in Nigeria is an example of a different kind. In a smaller number of cases the constituent units may play an instrumental role in the composition of national institutions. This is most obviously the case where a federal legislative chamber comprises representatives of unit governments (as in Germany) or legislatures (as in Austria) or a mixture of representatives of both the provincial legislature and executive (as in South Africa). Even where the federal chamber is directly elected, however, unit institutions may have control over some aspects of the process of selection. In Switzerland, for example, the cantons prescribe the electoral system for the Council of States, and in Australia and the United States the states fill casual Senate vacancies. The United States offers other examples of this wider point. The states still have responsibility for drawing electoral boundaries for the US House of Representatives, for example, although within an increasingly prescriptive framework of judicial and congressional regulations. Similarly, the states retain the power to decide how their electoral college votes will be allocated for the purposes of presidential elections.19 Finally, the challenge of accommodating the institutional requirements of two spheres of government within a single state may cause some institutions to be shared. Two instances are particularly common. First, many federations have a single court system in which the lower courts are under the authority of the constituent units but also deal with federal cases. Germany, Austria, Switzerland, India, and Canada are examples. Second, in some federations the administration of federal legislation by the constituent units is a deliberate component of the federal design, linked to the federal division of power. To a greater or lesser degree, this is the case in Germany, Austria, Switzerland, South Africa, and India. Political executives One of the principal points of contrast between the 11 federations represented here concerns the relationship between the executive and legislative branches of government, involving, in most cases, the choice between a presidential and a parliamentary system of government. Argentina, Nigeria, Russia, and the United States have presidential systems in which an executive president is elected separately from the legislature and does not rely on the support of the legislature to continue in office. The Russian system differs from the others to the extent that, theoretically, executive power is shared between the president and a prime minister or “chairperson of the government,” and ministers are appointed by the president on the proposal of the prime minister along lines broadly reminiscent of the structure of government in France. The difference carries through to other aspects of the system. Unlike in the other three states, in Russia here is no provision for a vice-president, and the Russian Duma may pass a vote of no-confidence in the government, which can be removed by the president on other grounds as well. By contrast, Australia, Austria, Canada, Germany, India, and South Africa have parliamentary systems in which the executive is drawn from the elected legislature and depends on the continued confidence of that legislature. Switzerland is an unusual case, designed to encourage consensus-style democracy through a collegial and inclusive structure. The members of the Swiss Federal Council are elected individually by a joint sitting of both houses of the national legislature. Once elected, however, they cannot be removed by a no-confidence vote. The choice of a presidential system by the framers of the Constitution of the United States was compatible with federalism, in the sense that both involved limited and divided powers. As the presidential system operates in the United States, it has enabled Congress to develop as a legislative institution with a will of its own, in a major point of contrast to many parliamentary systems. Presidentialism is not necessarily either synonymous with limited government, however, or a neat complement to federalism. In Argentina, Nigeria, and Russia concentration of power in the presidency has enabled the office to dominate the system of government at the expense of both the independence of the legislature and the autonomy of the federated units. In Argentina and Russia attempts have been made to restrict presidential power by creating a chief of cabinet answerable to the Congress in the former and by limiting the circumstances in which the president can dismiss the chairperson of the government in the latter. Neither has made a significant difference to what the Argentinean chapter describes as “hyperpresidentialism” and the Russian chapter describes as “superpresidentialism.” In Nigeria Osieke Ebere reports that there has been occasional consideration of reverting to a parliamentary form of government (abandoned in 1979 in favour of a presidential system in a brief return from military to civilian government) as a means of avoiding tension between the positions of president and prime minister. The dramatically different consequences of the establishment of presidential systems in these four cases are attributable to a variety of factors. The first and most obvious is the extent of the power vested in the president. Despite the prominence of the office in the United States, the powers exercisable independently by the president are relatively confined. By contrast, in the other three presidential federations, the president tends to have broader formal powers, including a limited law-making function and access to emergency powers of various kinds. A second factor concerns the extent to which the president can make use of extraconstitutional political or economic levers to augment the effective power of the office. Thus in Argentina the dominant role of the president is reinforced by extensive powers of political patronage within party machines, facilitated by a closed-list voting system and by the fiscal dependence of the states on the national government. A third possible factor is the shallow commitment to constitutionalism, understood in this context as compliance with constitutional limitations on power. Just as there are differences between presidential systems so, too, do the parliamentary federations differ in various ways. Of the six represented in this volume, South Africa is the most distinct. The leader of the South African government is drawn from the Parliament and depends on the confidence of Parliament in order to hold office. This justifies describing the system as parliamentary; however, it has some presidential features as well. The leader is described as a president rather than as a prime minister and plays the role of both leader of the government and head of state. He or she may be impeached as well as removed through a no-confidence procedure. Once elected, the president withdraws from the Parliament, although most ministers must be members of Parliament. This unusual structure thus offers an interesting combination of the key advantages of both systems. The presidency can provide the focus for national unity, but it remains under parliamentary control, and the efficiency of decision making that is a hallmark of parliamentary government is preserved. The remaining parliamentary federations differ from each other in less obvious respects. Where the electoral system is based on proportional representation, as in Austria and Germany, more parties are likely to have significant representation in the legislature, making coalition government common or even the norm. Where a majoritarian system of one kind or another is used, as in Australia and Canada, elections are somewhat more likely to produce a clear majority for one party, although, as the Canadian experience shows, regional parties may be a complicating factor. Other differences concern the procedures for forming and removing governments. Australia, Canada, and India, all of which are common-law federations, tend to follow British constitutional practice, conferring broad formal power on the symbolic head of state, who is kept in check by largely unwritten conventions.20 In both of the continental federations, the procedures for forming and removing governments are somewhat more specifically prescribed and include, in the case of Germany, a requirement that lack of confidence be expressed by a positive vote of majority support for an alternative government leader. Heads of state One further point of distinction between parliamentary and presidential systems concerns the position of head of state. In presidential systems the president is both the leader of the national government and the head of state. In this case he or she exercises not only the full range of national executive power but also performs the symbolic, ceremonial, and representative functions of the head of state, including those connected with the position of commander-in-chief. Parliamentary systems, with the exception of South Africa (for reasons explained earlier), typically have a separate office of head of state, with limited substantive power, performing most functions on government advice. Of the federations in this category three – Austria, Germany and India – are republics. Another two – Australia and Canada – are monarchies in which the functions of the Queen, based in the United Kingdom, are performed locally by representatives of the monarch who have become the de facto heads of state. Switzerland offers a model of an entirely different kind. The logic of the deliberately consensual framework for the Federal Council precludes a focus on any one of its members as head of state; instead, the council as a whole fills the position of head of state, while the less exalted position of chairperson of the council rotates annually between council members. Switzerland aside, the position of head of state is held by a single person. In these circumstances there is a question about how to ensure that the head of state has a sufficiently broad support base – which is not associated exclusively with a particular party, social group, or sphere of government – to be able to effectively symbolize the unity of the state. The problem is particularly challenging in presidential systems, where the head of state is also leader of the government. Several techniques are employed for this purpose in the presidential federations analyzed in this volume. In both Argentina and Russia the president is directly elected on the basis of a second-round voting system so as to ensure that the successful candidate receives the support of at least a majority of voters. Argentina moved away from an electoral college system in 1994; Hernandez reports that the tendency of the change to lessen the breadth of a candidate’s regional support was one of the controversial aspects of the electoral college. In Nigeria the president is directly elected as well, but the voting system requires the successful candidate to receive at least one-quarter of the votes in at least two-thirds of the states. Consideration has also been given to mandating the rotation of the presidency around the country’s principal zones. In the United States the president is indirectly elected through an electoral college in which 48 of the 50 states each casts its electoral votes as a single block. This system has its drawbacks in majoritarian terms, but it encourages broad presidential campaigns during which explicit attention is paid to state interests. The symbolism of the position of head of state is a factor that affects the design of the office in parliamentary federations. Austria has a directly elected president and, like Argentina and Russia, uses a second-round voting system. In both Germany and India the president is chosen by an assembly representing both the national legislature and the legislatures of the constituent units. In Canada the governor general, representing the Queen, is chosen alternately from among francophone and anglophone Canadians. In Australia the Queen herself tends to be regarded as a neutral symbol, and each jurisdiction has its own representative of her. The question of how the head of state should be chosen, if and when Australia becomes a republic and the common symbol of the monarchy is removed, has not yet been resolved. Legislatures National legislatures in federations are frequently bicameral. Bicameralism has become accepted as the most obvious mechanism by which the constituent units can play a role in national institutions and thus be a key component of the arrangements for shared rule. Bicameralism takes a wide variety of forms, however. There are also some federations (e.g., Venezuela since 1999) that have no national bicameral legislature.21 Nevertheless, the national legislatures of all the federations in this volume have two houses or chambers, subject to the caveat that, for reasons of history and composition, the German Bundesrat is “not a parliamentary organ in the strict sense” (Oeter, this volume). One house in each national legislature is directly elected by reference to population numbers, giving broad effect to the democratic principle of one vote, one value. These houses perform the usual functions of any popularly elected legislature. They make laws, they approve budgetary measures, and they scrutinize the executive branch. In parliamentary systems this is the house on which the government principally depends and from which the leader of the government is drawn. In presidential systems this is the house in which impeachment proceedings typically start.22 As described earlier, even this house is generally affected by the federal form of the polity in some respects: in procedures for the formation of constituencies; in requirements for the minimum representation of constituent units; and, occasionally, in the power of unit legislatures to prescribe aspects of the electoral process. From the perspective of federalism, however, the principal focus of attention is on the second house or chamber. In all 11 federations, with the possible exception of Canada (see below), this chamber has a federal purpose that is representative, functional, or a combination of the two. This purpose, in turn, provides the rationale for conferring upon it extensive powers. In some federations, including Argentina, Australia, Switzerland, and the United States, the powers of the second chamber are almost co-extensive with those of the first. The latter usually has somewhat greater authority in financial matters, in recognition of its greater democratic legitimacy, although here, also, Switzerland is an exception. Further, in addition to any specifically federal functions, some second chambers have general, national powers that the other house lacks: to consent to treaties, to approve key executive and judicial appointments, or to try impeachments. In Germany, Austria, and South Africa, by contrast, the powers of the federal chamber are tailored more precisely to its intended functions and are markedly less than are those of the other house. Nevertheless, the powers of the second chamber are significant in all these federations, creating the potential for conflict between houses constituted in different ways, to give effect to different principles. The representative purpose of a federal chamber is achieved through its composition. Two variables are particularly important: (1) the manner in which its members are chosen and (2) the proportions in which the constituent units are represented. The manner of choice ranges between chambers that are directly elected and those that are indirectly elected (or appointed in some way). As a generalization, there has been a trend towards direct elections unless indirect elections serve some systemic purpose, although it is not entirely consistent. Thus, while the members of the federal chamber in Argentina, Australia, Nigeria, Switzerland, and the United States are directly elected by the people of the respective constituent units in three of these federations – Argentina, Switzerland, and the United States – members of the federal chamber were originally elected by unit legislatures. On the other hand, in Russia the Federation Council was directly elected when it was first formed in 1993, but the two members from each constituent unit are now, respectively, appointed by the unit governor and elected by the unit legislature. Of the remaining federations, Austria, India, and South Africa elect members of the federal chamber indirectly, in a way that involves the legislatures of the states concerned, while the German Bundesrat comprises representatives of the Länder governments. In Canada, exceptionally, senators are appointed by the national government until retirement at age 75, and the Senate is barely relevant to federalism, making Senate reform a subject of perennial debate. Even in Canada, however, Senate appointments are apportioned by reference to a regional formula, giving it a representational role of a kind.23 In many federations there are additional links between the units and the federal chamber of the national legislature that tend to reinforce its representational character. These are most obvious in Germany and South Africa, where the function of the federal chamber complements the federal division of power, and members of the chamber act, in effect, as delegations from the unit government or government and legislature, bound collectively by their instructions. Other examples can be found, however. In Switzerland elections to the Council of the States are controlled by the cantons. In Russia, in an example of an entirely different kind (reflecting the sheer geographical size of the Russian federation), each month members of the Federation Council are supposed to spend a stipulated period of time in the unit that they represent, although they need not be residents of it. Elsewhere, however, while links continue to exist, there are signs that they are diminishing as national politics gradually becomes ascendant. In Australia, for example, as in Argentina and the United States, casual vacancies in the Senate are still filled by state governments or legislatures, but a change to the Constitution in 1977 restricts the choice to members of the same party as that of the retiring senator. In India, since 2003, members of the Rajya Sabha no longer need to live in the state they represent, easing the way to filling the seats on the basis of party considerations alone. In Argentina two states retain the theoretical capacity to instruct and recall their senators, but the procedures are never used, and doubts have been raised about their constitutional validity. The manner in which members of the federal chamber is chosen also affects the duration of their terms. In all the federations in this volume, with the exception of Nigeria,24 the terms of members of the federal chamber vary from the terms of members of the more popular house. Where the federal chamber is directly elected, the term served by its members typically is fixed for a period of six years, with a proportion of the members facing election every two or three years, so that the chamber has a perpetual existence, at least in normal circumstances.25 The link between direct election and fixed, rotating terms is not invariable, however. The Council of States is directly elected in Switzerland, but the terms of its members are regulated by legislation of the respective cantons. Conversely, members of India’s Rajya Sabha have fixed six-year terms, even though they are elected indirectly. In all other cases of indirect election, the terms of members of the federal chamber depend on the term of the legislature of the constituent unit they represent. In these circumstances, the composition of the federal chamber may be in continual flux, complicating the development of a working relationship with the other house. A second, important point of difference in the composition of federal chambers lies in the proportions in which the constituent units are represented. Equal representation gives effect to the principle of the equality of constituent units, possibly reflecting the lingering influence of conceptions of sovereignty attributable to the derivation of the federal idea.26 It necessarily ignores differences in the population size of the constituent units, however, and to this extent detracts from the democratic principle (if the people are considered as a whole). It is not possible to reconcile the two sets of principles, and most chapters record some tension between them. Nevertheless, a clear majority of the 11 federations in this volume give effect to the equality principle: Argentina, Australia, Nigeria, Russia, South Africa, Switzerland,27 and the United States. In all except Switzerland, arrangements for chairing the federal chamber are also designed to preserve the equal voting strength of each unit, either by providing for the national vice-president to act as chair or through the manner in which the votes are counted.28 Austria, Canada, Germany, and India adjust the representation of the constituent units by reference to the distribution of population in ways that depart in varying degrees from a calculation based on population alone. The representative function of a federal chamber is significant. It symbolizes the participation of the units, or the people organized in units, in the decisions of the federation as a whole. At least where the equality principle applies, it also gives less populous units, individually and collectively, a proportionately greater voice in the national legislature than they would otherwise have, and it facilitates the formation of parliamentary committees and cabinets representative of the country as a whole. As many of the chapters show, the composition of the chamber by reference to the constituent units does not necessarily mean that it functions in a distinctively federal way by, for example, representing unit interests or defending any particular federal principle. Where specific responsibilities, relevant to federalism, are conferred on the federal chamber, it may be possible to make a compelling case that the chamber fills a substantive federal role. Germany, South Africa, Austria, and, to a lesser extent, Switzerland illustrate the point. In these federations the role of the federal chamber complements the horizontal division of powers, whereby the constituent units administer federal legislation. In these circumstances, whatever other functions it may perform, the federal chamber becomes a forum though which such arrangements are mediated. Legislation to be administered by the units is approved; attempts to avoid or control unit administration are considered; and the adequacy of implementation resources are taken into account. In South Africa Christina Murray reports that the National Council of the Provinces (NCOP) was designed as “a concrete expression of the commitment to cooperative government” in the Constitution. Germany, South Africa, and Austria are also cases in which the members of the federal chamber are elected indirectly. In each case the composition of the federal chamber, the powers allocated to it, and the federal division of competences can be seen as an integrated package. Germany’s Bundesrat comprises members of the governments of each of the Länder and thus institutionalizes a form of executive federalism. The Bundesrat can veto specific categories of laws of particular significance to the Länder, including, at the time of writing, laws dealing with the allocation of financial resources and with prescribing administrative procedures to be followed by the Länder, although some adjustments may be under way.29 In relation to other types of laws, however, the veto of the Bundesrat can be overridden by the Bundestag. In South Africa and Austria the federal chamber comprises representatives of the legislatures of the constituent units, chosen to reflect the party balance in each legislature.30 They thus avoid entrenching executive federalism, although at a possible cost to the clarity of the role of the federal house. In both cases the federal house has particular authority in relation to legislation affecting the powers of the constituent units and more limited authority in relation to legislation of other kinds. In Austria, however, proposed laws affecting the administrative functions of the Länder also need the direct consent of the Länder. In South Africa the NCOP is already being criticized for not adequately reflecting provincial interests. An intriguing question is raised by South Africa in this regard. The requirement for the NCOP members who represent the same province to cast a single bloc vote on provincial questions, in accordance with instructions from provincial legislatures, is consistent with the functional role of this house. Accepting that NCOP members may cast an individual vote on matters that do not directly affect the provinces, however, may have encouraged the members to attach greater significance to such issues, despite the relative ease with which an NCOP decision can be overridden.31 India provides an interesting contrast. Members of the Rajya Sabha, also, are indirectly elected by state legislatures,32 and authority to administer federal legislation may be conferred on the states. The states have no entitlement in this regard, however, and there appears to be no conception that the federal chamber might provide a link between the two spheres of government that would facilitate state administration of federal law.33 The powers of the Rajya Sabha are not confined to questions of particular relevance to the state but are equal to those of the Lok Sabha (except in relation to financial legislation). The Constitution itself prescribes the manner in which administrative authority can be devolved and mandates the allocation of corresponding resources. In these and other federations specifically federal functions of other kinds are sometimes conferred on the federal chamber. These may include the nomination, or consent to the appointment, of members of the constitutional or supreme court, as in, for example, Germany and the United States; consent to treaties, as in the United States and Argentina; and approval of emergencies or other exceptional extensions of national power over the affairs of the constituent units, as in India and South Africa. In Argentina, novel requirements for the federal chamber to initiate intergovernmental agreements for tax-sharing and other forms of fiscal cooperation came into force as a matter of law in 1994, but no action has yet been taken in relation to them. Whatever their composition and functions, the operation of all federal houses is affected by the allegiance of their members to political parties, with the possible exception of Russia, where the Constitution forbids factionalism in the Federation Council. Where members of the federal chamber are elected directly, decisions are made almost exclusively on party lines. The United States is an exception, where relatively weak party discipline makes majorities more fluid in both houses. Even where members are indirectly elected and have specifically federal functions to perform, party interests tend to predominate. The role of parties does not necessarily preclude a federal chamber from fulfilling its representative role. It impairs any substantive functional role, however, where federal considerations are subordinated to party interests. The party political character of the house that is rationalized by reference to federalism also further complicates its relationship with the house that is popularly elected and that has party preferences of its own. If decisions in the federal house are made on party political grounds, it is in danger of becoming redundant where party majorities in the two houses coincide and inappropriately obstructionist where they do not. The problem is illustrated by most of the chapters. Thus, in relation to Austria, Gamper draws attention to the ineffectiveness of the federal chamber from the standpoint of federalism and to moves to abolish or reform it. In Switzerland Wolf Linder and Isabelle Steffan argue that the popular chamber performs at least as significant a federal function as does the federal house. Dhavan and Saxena note the failure of India’s federal chamber to protect state power from national incursions, despite its specific authority to do so. In Germany, Stefan Oeter points to the frequent political differences between the two houses, prompting the current search for ways to limit opportunities for disagreement. Where two houses disagree, even for reasons of party differences, informal negotiations often succeed in resolving the question in some way. In some federations, of which Argentina and the United States are examples, no other form of resolution is possible; so if the houses do not agree, the proposal does not proceed. Arguably, this approach reflects a view that each house plays a different role in the legislative process so that the consent of each is required. In most federations, however, formal mechanisms are provided to resolve disagreements between the houses. To the extent that such mechanisms favour one house over the other, they also enhance its bargaining position in any negotiations. The mechanisms differ between federations. Germany uses a joint mediating committee in which the houses are equally represented, forcing compromise between the parties and giving rise to a degree of “hidden consociationalism” in the political system. In India a joint sitting of the houses of the legislature provides a rarely used fall-back mechanism, the outcome of which depends on the strength of the various parties in the two houses. In South Africa a super-majority of the popular house can override a decision of the NCOP, even on matters of specifically provincial interest. In Australia the centrepiece of the deadlock mechanism is an election for both houses, which has become a strategic weapon in the struggle for political power between the parties and is relatively useless for the purpose for which the procedure was originally designed. Administration It follows from what has already been said that there are significant differences between federations in the organization of the national administration. The principal contrast lies between dualist systems, typified by the United States, and integrated systems, typified by Germany. In the former the national government has a complete administration of its own and normally administers its own legislative programs; in the latter a substantial portion of national legislation is administered, as of right, by the constituent units. Many federations exhibit elements of both approaches and lie somewhere between these two paradigmatic cases, which, in certain respects, also show some small signs of drifting closer to each other. Of the 11 federations in this volume, five are essentially dualist: Argentina, Australia, Canada, Nigeria, and the United States. India also has some dualist characteristics, but its arrangements are so distinctive that it deserves specific consideration in its own right. In the dualist federations, by definition, the national administration covers all areas of national constitutional responsibility. In addition, however, at least in common-law federations, national activities generally extend well beyond areas of national legislative power to areas of state responsibility that are funded by the national sphere or are the subject of intergovernmental cooperation. Typically, national administrative bodies are established for these areas as well, so as to develop policies, monitor expenditure, and evaluate outcomes. In such federations there may be extensive duplication of administrative agencies between the spheres of government and an elaborate network of relations between them. In Australia and the United States questions have arisen about whether the administrative functions of one sphere of government can constitutionally be conferred upon the other and, if so, in what circumstances. The Supreme Court of the United States has held that Congress cannot impose on states the obligation to administer its laws without the consent of the states concerned.34 Australian courts have cast doubt on the extent to which state administrative functions can be conferred on national officers or agencies, even with the consent of the Commonwealth legislature.35 These decisions have encountered some criticism in both countries for being excessively rigid interpretations of typically meagre constitutional provisions that detract from administrative convenience without good cause. There may be more to be said in their favour, however. In both countries the mechanisms for legal and political accountability are structured on the assumption that each government is responsible to its own legislature for the administration of its own laws. These mechanisms are disturbed by the conferral of administrative authority across jurisdictional boundaries, unless alternative provision is made. It should be noted, however, that in neither case does the constitution preclude the use of financial incentives, in the form of conditional grants, to induce the constituent units to adopt and to implement national policies and programs.36 India is distinctive in at least two respects. The first concerns the all-India Services: a traditionally independent, highly professional, administrative structure whose officers serve both spheres of government and are held to account within the sphere in which they are working at any one time. The all-India Services are a product of Indian history, inherited by independent India from the British Raj and turned to the use of building and preserving national unity. Recruitment, training, and overall management of the services lie with the Union government, however, and Dhavan and Saxena note the influence thus exercisable by the Union over administration in the states. Additions to the all-India Services require the consent of a super-majority in the Rajya Sabha, which generally has not been forthcoming because of state opposition. The second point of distinction in relation to India concerns the allocation of responsibility for the administration of national legislation between the Union and the states. State administration of Union legislation is the norm in areas of concurrent power,37 and the states may be required to administer Union legislation in other areas as well.38 The Constitution provides a framework of necessary principle that prescribes how the administrative authority of the Union is conferred on the states, requires the allocation of corresponding resources, and seeks to ensure that the authority is exercised in accordance with Union directions. State authority, conversely, may be conferred on Union administrative bodies with Union consent. Other federations in which administration of national legislation by the constituent units is a structural feature of the system, linked to the federal division of powers, include Austria, Germany, South Africa, and Switzerland. In Austria, Germany, and Switzerland there is a national administration as well, but it is limited to a few areas in which the national sphere administers its own legislation, and the administration of the constituent units is correspondingly more extensive. In South Africa there is a single administration serving both spheres of government, for which the national legislature provides a fairly prescriptive legislative framework that extends to salaries and terms and conditions of employment. Murray draws attention to the impact of these decisions on budgetary management by the provinces. In these federations the categories of national legislation to be administered by the constituent units in their own right are indicated in the constitution, in the same manner as is any other aspect of the federal division of competences. It is convenient to describe the power exercisable by the units in this case as an administrative competence. In addition, the constitutional scheme may authorize the conferral of additional authority on the constituent units to administer national legislation, if the national legislature, including the federal chamber, so decides, which is described here as administrative authority. Typically, national institutions exercise greater control over the administrative process where the units exercise only administrative authority, and, in Germany, additional funding is also required. Inevitably, the detail of these arrangements differs between federations. In Germany, for example, direct administration by the Länder, in the exercise of their constitutional competences, is the norm unless otherwise indicated in the Constitution. The position is broadly similar in Switzerland. In Austria there is greater use of the technique of conferring administrative authority, although the Länder have competence to execute some national laws, in addition to their own; and they have both law making and administrative competence in areas of framework legislation. In South Africa the provinces administer as of right national laws enacted in areas of concurrent legislative power, subject to the Constitution and any national law to the contrary. Power to administer other national legislation may also be conferred on the provinces specifically. In each case the federal chamber plays a role in approving or disapproving departures from the basic scheme, to the extent that departure is possible.39 In Austria, however, direct consent of the Länder may also be required, and in Switzerland direct democracy provides an alternative sanction. In a system of this kind the national government must rely on the good faith and competence of the constituent units and must also accept some diversity in the implementation of its legislation. Remedial action is generally available in the event of administrative failure, through political intervention or judicial proceedings. South Africa’s Constitution, in particular, makes careful provision for national supervision and intervention in the face of concern about provincial capacity.40 Even so, such procedures are rarely used. These federations necessarily operate on the assumption that the constituent units are able and willing to perform their constitutional responsibilities, reinforced by the expectation that both spheres of government will act in good faith towards each other or, in South African terms, in the spirit of cooperative government. There are signs in at least two of the chapters, however, that the model is under pressure. In Switzerland, Linder and Steffan report that the national administration is exerting greater influence in the formulation and implementation of national policy and that national legislation is becoming increasingly prescriptive, limiting the autonomy of the cantons, in the interests of uniformity. In Germany, similarly, an increase in the number of national statutes imposing more detailed administrative obligations on the Länder has increased the proportion of bills that are subject to the absolute veto of the Bundesrat and, thus, occasions for disagreement between the two houses. The present proposal is to diminish the potential for conflict by restricting the powers of the Bundesrat to a degree in relation to laws of this kind, compensating the Länder with new substantive powers.41 Some remaining aspects of the national administration can be dealt with more briefly. In dualist federations in particular, where the national administration is more extensive, there is a question about its geographic distribution. Head offices are often located in the national capital but may be placed elsewhere. The distribution of regional offices is likely to follow the configuration of the constituent units, but smaller units may miss out and regions comprising groups of units may be used instead. In either case there is potential for argument about whether national authorities have been sufficiently even-handed. In culturally diverse federations questions may also arise about whether the national administration and other national agencies are sufficiently representative. The Swiss chapter notes a requirement for proportionate representation of ethno-linguistic groups in most national departments. In Nigeria the composition of national agencies is required to “reflect the federal character of Nigeria” so as to preclude the domination of particular national groups.42 In Canada official bilingualism gradually helped to secure a more proportionate representation of francophone Canadians. In all federations the national government has other agencies as well, established for particular purposes. These may have no federal significance beyond the questions about location and composition. Two categories have some relevance from the perspective of federalism, however. The first comprises agencies created to play a specific role in relation to the federal system. Examples include the Nigerian Federal Character Commission, the Australian Grants Commission, and the South African Financial and Fiscal Commission. Second, the federal system is specifically relevant to the way in which some agencies are constituted. Central banks are, or were, an example, and although the influence of federalism on central banks has diminished, it has continuing significance in some federations, including Switzerland and Germany. Some other such agencies are the product of intergovernmental arrangements and are considered in that context below. Courts Every federation has a national court system of some kind. The interesting points of comparison lie in the scope of this system, the influence of federalism upon it, and its relationship with the courts of the constituent units, where separate and distinct courts exist. As a generalization, a point made by Osieke in relation to Nigeria applies to most other federal countries: there tends to be greater interdependence of courts in federations than there is of any other institutions of government. Nevertheless, court systems in federations can also be characterized broadly as dual or integrated. Dual systems involve largely separate and parallel court hierarchies for each sphere of government, exercising the jurisdiction assigned to the respective spheres. Integrated systems involve a single court hierarchy, authority over which is likely to be divided between the national government and the constituent units. It may, however, be assigned to the national government alone. Austria, Russia, and South Africa are examples of federations in which the courts are entirely a national responsibility, although in Austria there has been some discussion of creating Länd administrative courts in order to spread the burden of adjudication and to further strengthen the Länder. Federations with a dualist judicial structure include Argentina, Australia, Nigeria, Switzerland, and the United States. Those with an essentially integrated system include Austria, Canada, India, Germany, South Africa, and Russia. As a generalization, federations with a dualist administrative structure tend to have a dualist judiciary too, but the correlation is not complete, as in the example of Canada. This is an area in which distinctions between common-law and civil-law systems are also relevant, further complicating the accommodation of the structure of the judiciary to the realities of federally divided power. Thus, in Argentina, the need to allocate jurisdiction between different court hierarchies within a dualist court system potentially threatens the integrity of the interpretation of the legislative codes, as the principal source of law, necessarily enacted by the national legislature. Conversely, in Australia the integrating feature of a single, final court of appeal has had the effect of consolidating the potentially pluralist common law of the several states into a single common law. These categories of dualist and integrated systems are not watertight. Most dualist systems involve some degree of integration or at least departures from the paradigm of parallel court systems. In Argentina an “extraordinary appeal” to the Supreme Court of Justice, alleging arbitrariness by a provincial court, is a device for improving consistency in the interpretation of federal codes, even when a dispute falls within provincial jurisdiction. In Australia federal jurisdiction can be conferred on state courts, and the apex court, the High Court, hears appeals in matters of both federal and state jurisdiction. In Nigeria, also, appeals lie from state to federal courts. In Switzerland, as a logical corollary of the horizontal division of power, most legal questions are handled in cantonal courts, with an appeal to the Federal Court only in relation to federal law, including, significantly, the consistency of cantonal law with federal law. Conversely, most integrated systems enable the constituent units to have some ownership of the courts most relevant to them. In Germany judges of the first two levels of courts are appointed and paid by the Länder. In India judges of the highest state courts are appointed by the president, but consultation occurs with the governor of the state concerned, and the lower judiciary is subject to state control. These formal procedures are now significantly affected by the control over appointments exercised by the courts themselves, in a manner described more fully below. In Canada the superior court judges in each province are appointed and paid by the national government, but the decision about the number of judges to be appointed lies with the province. Even in South Africa, where the courts are nationally controlled, the premier of a province is included on the Judicial Service Commission when appointments to the high court of the province are under consideration. Arrangements to secure the independence of the courts vary with the structure of the court system. Federations with a dualist judicial structure, including Argentina, Australia, and the United States, are more likely to leave the protection of judicial independence to the respective spheres. Federations such as Germany, with a more integrated judiciary, typically have a common framework for the organization of the judiciary, sometimes allowing variation in matters of relative detail. Cutting across these considerations of federal design, however, is the effect of the experimentation with new mechanisms for protecting judicial independence that has taken place in many countries, federal and non-federal. Thus in Argentina, Nigeria, and South Africa specialist bodies have been created to deal with aspects of the organization of the judiciary, including the appointment and removal of judges, discipline, and, in some cases, funding: the Council of the Magistrature in Argentina, the National Judicial Council in Nigeria, and the Judicial Service Commission in South Africa. Here again, however, federalism has an effect. In Argentina the national council deals only with national courts, leaving provinces to establish their own councils if they so wish, as half have chosen to do. By contrast, in Nigeria the council has responsibility for the courts of both spheres of government. Appropriately in these circumstances, its membership includes some state representation, and advice on appointments goes to state, rather than national, institutions. Another small trend of some potential significance is the role of the courts themselves in protecting judicial independence. The development is most marked in India, where Dhavan and Saxena describe the effective takeover of the process of appointment to the Supreme Court and to the state high courts by the Supreme Court itself and the corresponding control acquired by state high courts over appointments to the lower judiciary. But in Canada and Australia also, interpretations of the constitution in recent decisions have had the effect of reinforcing the independence of the courts of the constituent units.43 In all federations, with one exception, the compliance of both spheres of government with the federal constitution is underpinned by judicial review. The exception is Switzerland, where the constitutionality of national law is left to the processes of direct and representative democracy. Comparison of the remaining ten federations included in this volume reveals two broad approaches to the organizational arrangements for constitutional review. In one group of federations review can be carried out by all courts or, at least, all courts above a certain level, subject to final appeal to a court with both constitutional and ordinary jurisdiction at the apex of the national judicial hierarchy. Six federations are in this group: Argentina, Australia, Canada, India, Nigeria, and the United States. This is a typically common-law approach to constitutional adjudication, and all these federations, with one exception, have common-law systems. The exception is Argentina, where the US constitutional model was followed closely, even in this respect. The alternative approach is to establish a specialist constitutional court to carry out the function of judicial review, precluding, or at least restricting, consideration of constitutional questions in other courts. Federations with arrangements of this kind include the remaining civil-law federations: Austria, Germany, and Russia. South Africa also has a specialist constitutional court. Consistent with its common-law heritage in matters of public law, however, other South African courts can also deal with constitutional questions, subject to a final appeal to the constitutional court, and there has been intermittent discussion for some time about establishing the constitutional court as a final, general, court of appeal. Whichever approach is adopted, the court with final authority to interpret the constitution is an important actor in the federation. Responsibility for it usually lies with the national government, but the constituent units are often involved in some way, reflecting the role of the court in umpiring the boundaries of constitutional power between the two spheres, however contested its effectiveness may be. In the presidential systems of Argentina, Russia, and the United States, the federal chamber of the national legislature must consent to appointments to the court. In the fourth presidential system, Nigeria, appointments are made by the president on the advice of the National Judicial Council, whose members include five state judges. In Germany, also, the federal chamber plays a role in the composition of the Constitutional Court, but in this case the role involves the appointment by the Bundesrat of one-half of the members of the Court in recognition, as Oeter explains it, of the “overtly political relevance of such broad judicial control.” In Canada and Australia the traditional power of the national executive to appoint judges is qualified by federal principle in relation to this particular court in yet other ways. In Canada, by convention, certain proportionate shares of the regions are observed in making appointments to the Supreme Court; three of the judges always come from Quebec, three from Ontario, two from the western provinces, and one from the Atlantic provinces. In Australia there is a legislative requirement for the national government to consult state attorneys-general before making appointments to the High Court. INSTITUTIONS OF THE CONSTITUENT UNITS Relations between spheres By definition, the constituent units of a federation must have some degree of autonomy to organize and operate their own institutions. By definition also, however, membership in the same polity implies some common standards and some interdependence, so that autonomy cannot be absolute. In each federation, therefore, the arrangements for the institutions of the constituent units strike a balance between self-rule and shared rule in much the same way as has become familiar in other aspects of federal design. A wide range of variations exist, therefore, between the extremes of autonomy and interdependence. As a generalization, constituent units with their own constitutions are likely to have greater institutional autonomy than are those without, and a greater ability to experiment with institutional design. This feature of a federation tends to reflect the way in which the federation was initially formed and the extent of the territorial integrity of the constituent units. Constituent units that were independent of each other before the federation was formed usually join the federation with existing constitutional traditions of their own, which are maintained. Argentina, Australia, Switzerland, and the United States are examples. The German and Austrian Länder and the ethnic republics in Russia have also had their own constitutions but for reasons that are explicable by reference to the complex history of all three countries and that are consistent with the way in which territorial boundaries are changed. By contrast, in India and Nigeria the constitutions of the states are prescribed in the national constitution. Many of the Indian states had no autonomy of their own before independence. More significantly, however, the states have been refashioned and their boundaries repeatedly redrawn in a manner that would have made a regime of separate state constitutions difficult, if not impossible. Nigeria is an example of a somewhat different kind. In the early years of Nigeria’s federation, the three, and later four, constituent units had constitutions of their own, as had been the case in colonial times. After the first period of military rule, however, from 1966, the proliferation of states began, precluding the use of separate state constitutions for much the same reason as was the case in India. Canada and South Africa offer variations on this approach. In both countries the national constitution sets out a framework of government for the constituent units, which may be altered by the units themselves, within national constitutional constraints. The mechanisms are different, however. The Canadian provinces may either alter sections of the national Constitution that apply to them or enact legislation of their own of a constitutional kind.44 In South Africa a province may substitute a provincial constitution for certain provisions in the national Constitution. The scope of the discretion of South Africa’s provinces is significantly limited by the terms of the national Constitution, which also provides that the constitutionality of a provincial constitution must be certified by the Constitutional Court. Only two provinces have sought certification for a provincial constitution: Kwazulu-Natal and the Western Cape. Both initially were refused,45 although a further application by the Western Cape ultimately succeeded. In every federation there are national constitutional controls of some kind over the institutions of government of the constituent units, representing shared constitutional standards for the polity as a whole. Sometimes they are expressed in very general terms. Examples include the guarantees of “a republican representative system” in Argentina,46 a “republican form of government” in the United States,47 and a “republican, democratic and social state governed by the rule of law” in Germany.48 Somewhat more stringent national constitutional controls often come in the form of constitutional protections of rights, which typically now apply to both spheres of government. Canada is a particularly interesting example. The Canadian Charter of Rights and Freedoms binds the governments and legislatures of the provinces as well as the federal government. It also, however, provides a mechanism by which a legislature can make a law “notwithstanding” a provision in the bill of rights that, consistent with federal principle, is available to both spheres of government as well.49 As the chapters in this volume demonstrate, however, there is a range of more specific ways in which the institutions of constituent units are constrained by the national sphere, which are less readily reconciled with federal principle. Thus, in Canada and India the titular heads of the constituent units are appointed by the national government, although in Canada practices of consultation with the constituent units have developed over time.50 In Russia the national president can nominate and effectively appoint the governors of the constituent units. In South Africa the premier of a province that is controlled by the governing party is effectively chosen by the central party organization. As Murray argues in relation to South African practice, controls of these kinds not only detract from the checks and balances that federal governance may provide but also may impede democracy by inhibiting the development of political accountability within the constituent units themselves. One link of a different kind between national institutions and those of the constituent units concerns the timing of elections. In most federations there is no correlation between the timing of the elections in the two spheres. In a few cases, however, elections for the two spheres of government tend to be held on the same day. The United States is an example, although some states depart from the typical pattern. Such a practice is not possible unless legislatures have fixed terms – a common phenomenon in presidential systems but achieved with greater difficulty in parliamentary systems.51 Thus in South Africa, as Murray reports, while national and provincial elections are currently synchronized, early dissolution of a provincial legislature is constitutionally possible,52 triggering differences in the electoral cycle of the province in question. Synchronized elections have some advantages in terms of cost, time, and diversion from the ordinary business of government. Where the composition of the federal chamber depends on the governments or legislatures of the constituent units, they may have another advantage as well. As the example of Germany shows, staggered elections may mean shifting majorities, further complicating relations with the popular house. In South Africa the NCOP’s composition also depends on the constituent units. For the moment, however, the practice of synchronized elections ensures that NCOP members are chosen at the same time. In most federations there is some capacity for the national government to depart from the usual constitutional order in ways that sometimes involve intervention in the affairs of constituent units, in the face of external aggression or, less frequently, internal insurrection. In addition, as noted earlier, where the constituent units administer national legislation, national institutions may have residual powers of intervention for that purpose as well if a unit is unwilling or unable to carry out a task effectively. There is also a third category of cases, however, in which the national government has more general emergency powers in relation to the federation as a whole or to the affairs of the constituent units, individually or collectively. The dilemma for federal institutional design that is thus presented was described earlier. On one hand, intervention is sometimes necessary in the interests of the security of the entire community, the integrity of the state, or the effective performance of government functions; on the other hand, intervention is inconsistent with the normal federalist assumption that constituent units have final responsibility for the conduct of their own affairs. At least four of the federations in this volume allow national intervention of this somewhat more extraordinary kind: Argentina, India, Nigeria, and Russia.53 With the possible exception of India, these are also among the least stable federal democracies. One question raised by the chapters in this volume, however, is whether the provision of emergency power is not only a response to instability but also a contributing factor. The details of emergency arrangements differ between federations, in relation to the circumstances in which intervention is authorized, the action that may be taken, the checks and balances that are provided, and the extent to which the procedures are actually used. Nevertheless, some generalizations may be made. First, the circumstances that trigger intervention are typically stated broadly. In Nigeria, for example, a state of emergency may be proclaimed when there is a “clear and present danger of an actual breakdown of public order and public safety.”54 The consequences of the imposition of a state of emergency on grounds of this kind may include the suspension of the federal division of powers, the abrogation of some or all constitutional rights, and the expansion of executive power to make laws. Second, in most of these federations, extraordinary action can also be taken in the face of the failure, or perceived failure, of government in the constituent units. Thus, in India, “President’s rule” can be given effect where “a situation has arisen in which the government of a State cannot be carried on in accordance with the Constitution.”55 In cases of this kind, the institutions of a unit government may be suspended and their functions assumed by representatives of the national government. A state of emergency generally augments the power of the national executive. The potential for abuse is obvious, and attempts are made to safeguard against it in various ways. Techniques in common use include a requirement for ministers to countersign presidential emergency decrees; requirements for approval by the legislature or by the federal chamber; or the consent of institutions of the constituent units. But it is the evidence of these chapters that none is completely reliable. In presidential systems, such as Argentina and Nigeria, ministers are dependent on the president for office. Legislatures are often in recess, and even when they are not, they may be controlled by the executive. Reliance on consent among unit institutions to provide a check and balance will be unlikely to be effective where the institutions themselves are the appointees of the national sphere, as is the case with the governors in India. The failure of preventative mechanisms is suggested by the frequency of recourse to emergency procedures in India and Argentina. These cases suggest that, if specific emergency powers are provided, the circumstances in which they can be used and the consequences of their use should be more carefully circumscribed, weighing the values of federal democracy in the balance. Institutional arrangements In all federations the constituent units have legislatures, executives, and administrations. These are primarily responsible to the people of the respective units, who may be conceived as citizens of their units, whether they are so described or not. In some federations constituent units also have a judicial branch. They may have an array of other institutions as well, such as ombudspeople, human rights commissions, and electoral commissions. By contrast, in other federations, the courts and other specialist institutions of government of this kind are common to the two spheres with the consequence, in most cases, of greater national control. For the most part, the institutions of the constituent units are broadly similar to those of the national sphere and, thus, also to each other. Sometimes, as in Nigeria and India, and to a lesser extent in South Africa,56 this occurs because the institutions of the constituent units are prescribed in the national constitution. Even in other cases, however, the principal institutions of the two spheres of government are likely to resemble each other. There are some notable exceptions, however. The legal system and institutions of Quebec are influenced by the French civil-law model, while the rest of Canada adheres to the common law and institutions of government that fall broadly within the British parliamentary tradition. In Nigeria there are sharia courts in some states and not in others. In Australia the referendum to establish a republic, which was rejected in 1999, would have left the states to break their own ties with the Crown, at their own speed, creating the possibility that some states would have become republics while others would have retained the monarchical system, at least for a period of time. There are many lesser variations in the institutional arrangements of constituent units, some of which are common to a range of federations. First, it is relatively rare for the largely ceremonial position of head of state of the kind found in most parliamentary systems to be replicated in the constituent units. Australia is an exception, where governors are appointed by the Queen on the advice of state premiers in a manner almost identical to that involved in the appointment of the governor general, and they play a very similar role. In Canada and India the head of state of the constituent units is appointed by the national government. Austria and Germany have no equivalent position, necessitating other minor adjustments to the operation of the parliamentary systems of the Länder. Second, bicameralism is somewhat less common in the legislatures of constituent units. In Austria, Germany, Nigeria, South Africa, and Switzerland bicameralism is effectively precluded altogether. In Canada provincial second chambers have gradually disappeared. In Argentina, India, and Russia there is mixed use of bicameralism, but only a minority of unit legislatures now have two chambers. Only in Australia and the United States is bicameralism still a prevailing feature of the state governments. The rationale for bicameralism in this sphere is difficult to establish, however, beyond a general belief in the value of a second opportunity for deliberation, whatever the party composition of the second chamber. In this context the suggestion by Dhavan and Saxena that second chambers might develop as a forum within which local government is represented is of interest, particularly in view of the trend, noted in the next section, for local government to be accorded constitutional status. Third, direct democracy is more likely to be used in the constituent units than in the national sphere, not only for constitutional change but also for other purposes as well. Switzerland, Austria, Germany, the United States, and Argentina are examples. In a related phenomenon, some constituent units of some federations, of which Switzerland and the United States are examples, also provide for the election of judges, although the practice can be controversial because of its implications for judicial independence. It is apparent from the chapters in this volume that a degree of experimentation with institutional design takes place within the constituent units of some federations. The potential necessarily varies with the degree of national control. Thus in South Africa, for example, an attempt by the province of the Western Cape to change its voting system was rejected by the Constitutional Court.57 Similarly, with regard to Austria, Gamper describes how the Constitutional Court has limited the extent to which direct democracy can be used in order to retain the character of the polity as a representative democracy. Other chapters, however, show a degree of experimentation in the sphere of the constituent units in other federations in relation to, for example, electoral systems, legislative terms, and new mechanisms for accountability and transparency. A successful experiment in one unit may be adopted in others and even in the national sphere. LOCAL GOVERNMENT Just as the institutions of the constituent units are broadly similar to those of the national government in each federation, so the institutions of local government have some of the same general features as do the other two. There is, however, considerable variation both between and within federations in relation to the manner in which the head of each local government unit is chosen. Some are elected indirectly from the council or other elected body of local government representatives in each area. Increasingly, however, in many federations the mayor or other equivalent position is chosen through direct election, particularly in larger local government areas. There is also a trend towards according local government constitutional status and increasing levels of autonomy in areas for which it is responsible. Local government is now recognised in a form that ostensibly gives it a degree of autonomy in the national constitutions of eight of the 11 federations in this volume: Argentina, Austria, Germany, India, Nigeria, Russia, South Africa, and Switzerland. In some cases it is recognized in the constitutions of the constituent units as well, and in Australia and the United States it is recognized only in state constitutions. In Canada it is mentioned in passing in the national Constitution, as a component of exclusive provincial power, in terms that have no implications for the autonomy of local government.58 The strength of the guarantees of local autonomy varies between federations. In some cases the national constitution provides only a framework for local government, leaving the constituent units considerable residual authority in relation to the structure of local governing institutions; in other cases the constitution is more prescriptive. In any event, constitutional promise does not necessarily equate to performance, as the chapters on Argentina and India show. There is continuing tension between the traditional autonomy of constituent units in relation to local government and the relatively new-found autonomy of local government. It may be that these difficulties are transitional and that a systemic change is under way, to which the institutions of the other spheres will adapt in time. INTERGOVERNMENTAL RELATIONS It has become a truism that the spheres of government in all federations work together in various ways, no matter how powers are divided between them and whatever the logic of their institutional design. Institutional design is significant for the particular forms that intergovernmental interaction takes, however. The converse is also true. The operation of institutions originally developed in the context of unitary systems of government may be adversely affected by intergovernmental activity unless deliberate countervailing provision is made. In many federations intergovernmental activity is envisaged from the outset, and provision is made for it in the national constitution. In this case intergovernmental forms and procedures are likely to be compatible with the institutional design and may even be integral to it. Arguably, federations of this kind are increasing. It is almost impossible to imagine the establishment of a new federation in the twenty-first century in which some provision is not made for intergovernmental activity. Formally planned arrangements for intergovernmental interaction take a particular form in federations with a horizontal division of power, separating legislative from administrative competence. Germany is a striking example in which relations between governments are systematized through the Bundesrat. There are variations on this model in both Austria and South Africa, where members of the federal chamber do not represent governments but are chosen by legislatures, enhancing its legitimacy as a legislative chamber but detracting from its usefulness as a forum in which governments, representing their respective constituencies, can reach agreed solutions. Thus, in Austria, laws that deny the Länder their administrative authority require the direct consent of the länder as well as the federal chamber. In South Africa the functions of the NCOP are complemented by constitutional principles of cooperative government and intergovernmental relations. Assessment of this model is complicated by both the dominance of a single party and the challenges of transition. Murray suggests, however, that the NCOP is still a long way from fulfilling an intergovernmental role. In federations with a more dualist structure, planned intergovernmental arrangements take a variety of other forms. In the United States and Argentina treaties or compacts between states provide a basis for joint action on matters of mutual interest, subject to the approval of the national legislature. Such arrangements may, inter alia, provide for the establishment of joint regulatory or administrative agencies. In the absence of specific constitutional authority, it has proved more legally complex in Australia to establish bodies of this kind; many such bodies exist, but they sit less easily in the federal constitutional framework. However, the Australian Constitution specifically authorizes intergovernmental arrangements of another kind, in the form of a power for state legislatures to refer matters within their authority to the Parliament of the Commonwealth, thus providing a degree of flexibility in the federal division of powers while retaining the essentially dualist institutional design. Nigeria and India make specific provision for forums, outside the normal institutional structure of government, in which intergovernmental consultation can occur. In Nigeria the constitutional body is the Council of State, which acts as an advisory body to the president on matters that include questions of particular interest to the states,59 and it includes the state governors in its membership, together with a range of functionaries of other kinds. In India the comparable body is the Inter-State Council, envisaged in the Constitution as a forum through which the various governments might meet. The council was not established until 1990, however, and has not yet fulfilled its apparent potential. Planned intergovernmental relations thus contemplate a range of new institutions: meetings of ministers or other executive officers, joint administrative agencies, or agreements between governments of various kinds. These same institutions also form the core of the myriad of intergovernmental arrangements that exist in all federations for which no specific constitutional provision is made. In these circumstances intergovernmental institutions are a pragmatic response to the experience of federalism in action, further fuelled by the pressures of globalization and internationalization and, often, by an imbalance in financial resources between the spheres of government. Intergovernmental institutions rarely have coercive authority or effect in their own right; rather, they are mechanisms to facilitate coordination between the spheres, dependent for any regulatory effect on implementing action by governments, legislatures, and courts. Systematic arrangements for meetings of ministers are a common feature of parliamentary systems in which, in normal circumstances, legislatures can be expected to faithfully implement commitments made by their respective governments. Successive chapters in this volume detail the proliferation of such meetings, not only in Canada, Australia, and Switzerland, where the federal chamber does not provide a forum for cooperation within the institutions of the federal sphere, but also in Austria and Germany, where it does. In systems with a separation of powers, where the legislature is not necessarily controlled by the executive branch, meetings of this kind are likely to be confined to the exchange of information and coordination of executive action, not least in lobbying the national government. The National Governors Association in the United States is an example. Even so, the chapter on Argentina suggests that such an institution could be useful in that country as a forum in which the governors of the provinces could draw strength from each other through collective action, presenting a united front to the national government in the interests of the provinces as a whole. Other intergovernmental arrangements may be the product of deliberations between governments in forums of this kind. Decisions may be enshrined in formal agreements, which may also be called treaties, concordats, or compacts. Agreements, in turn, may have a variety of purposes, ranging from joint funding of projects to uniform legislation to shared administration of a program of mutual interest. In this case the agreement may call for the creation of a joint agency or the conferral of administrative authority on other spheres of government, across jurisdictional lines. Experience in both Australia and the United States suggests that there may be constitutional obstacles to mixing authority in this way, unless it is specifically authorized by the constitution, as is the case in India. Some intergovernmental arrangements require the involvement of national institutions – typically, the national legislature and in some cases the federal chamber. Argentina is an example, where legislation to give effect to an intergovernmental agreement on fiscal co-participation must be initiated by the Senate, which has, so far, failed to act. Even where there is no constitutional requirement for national involvement, the institutions of the national sphere often participate in intergovernmental arrangements, either because the arrangement in question requires national resources or authority or because the national sphere has historically assumed a coordinating role. In Canada and Australia ministerial councils typically involve ministers from the national government, even when dealing primarily with questions that fall within the competence of the constituent units. National involvement in such matters is not universal, however, and there are some signs, briefly described below, of a trend towards the creation of coordinating bodies involving the constituent units alone. Intergovernmental relations in federal systems are sometimes associated with the practices of “executive federalism.” The name reflects the reality that, in many federations, intergovernmental outcomes are precipitated by agreements between the governments of the participating spheres, relying on the compliance of their legislatures. However, there is no single form of executive federalism, and the term has different meanings in different contexts. In Germany it refers to the activities of representatives of the Länd governments in the Bundesrat. In common-law parliamentary federations it refers to the activities of ministerial councils. In some presidential federations, of which Argentina is an example, it refers to the dominance of the executive branch vis-à-vis the legislature, reinforced by the relationships between the president and governors. In each case the practice creates some difficulties for the regular operation of democratic institutions and sometimes, also, for the rule of law. Most obviously, it further subordinates legislatures to the executive arm of government. In doing so it exacerbates a familiar difficulty for the chain of democratic accountability in many systems, which continues to be based on the premise that elected legislatures hold the executive branch to account and themselves take the decisions of greatest importance: to make law, to tax, and to spend. The typical complexity and lack of transparency in intergovernmental relations further compounds the difficulty. Nevertheless, intergovernmental relations have a relatively good press as an efficient and effective response to the complexity of governance in a system in which power is divided and institutions duplicated. There is a case, therefore, for them to be managed and refined so as to improve their compatibility with the institutions of federal democracy. The evidence of several of the chapters in this volume is that the process is under way. To the extent that intergovernmental arrangements have a centralizing tendency, it may be countered to a degree by the emergence of such bodies as the Council of the Federation in Canada and the Conference of Cantonal Executives in Switzerland, representing the constituent units alone. Gamper notes that, in any event, intergovernmental arrangements are not invariably centralizing; in the context of the centralized design of the Austrian federation, cooperation between the Länder provides a counterweight to the authority of the national government. Problems of transparency and accountability may be more difficult to overcome, given the inherently executive character of these processes. Even here, however, there are signs of change. Steps to involve unit legislatures in intergovernmental arrangements have been taken in some federations, presaging the emergence of a new form of “legislative federalism.” In relation to Switzerland, for example, Linder and Steffan show how the collaboration of cantonal legislatures can work to the benefit of the constituent units, in that case by instigating a popular vote, which rejected proposed taxation changes that would have disadvantaged the cantons. In a development of a different kind, in Canada in 2004 the first ministers meeting on health care was televised in a gesture towards transparency that could, however, as Thomas Hueglin observes, compromise the goal of securing a negotiated solution. TRENDS AND CHALLENGE One clear conclusion of all the chapters is that there is a remarkable degree of interaction between federalism and the institutional structure of government. The interaction is most evident in relation to the federal chamber of the national legislature. Federalism affects most other institutions as well, however: the popular legislative chamber, the executive branch, the administration, and the courts. In addition, it has been the catalyst for the emergence of new forms of institutions serving specifically federal purposes. A second conclusion is that the institutions of government and the federal design are integral parts of the same constitutional and political system and cannot fully be understood in isolation from each other. To a degree, it is possible to explain and understand the resulting composite types by reference to some of the standard categories that have been used in this chapter. Thus it is possible to characterize federations variously as dualist or integrated; dividing powers only vertically or also horizontally; with a parliamentary or presidential form of executive government; or operating within the framework of the common law or the civil law. But the categories are not discrete. In many cases a federation that ostensibly belongs to one category for a particular purpose in fact has characteristics drawn from another, or country-specific features of its own, defying the standard categories altogether. This is a time of considerable experimentation in the design of federal institutions, and the degree of intermixture is likely to increase. The categories, therefore, are useful for analytical purposes and for understanding the implications of certain institutional combinations, but their use should not be pushed too far. It is difficult to generalize about challenges and trends across such a diverse range of federations. At one level, it is inevitable that the challenges differ with the circumstances of the federations concerned. In the less stable or emerging federations of Argentina, Russia, Nigeria, South Africa, and. in some respects, although to a lesser extent, India, the foremost challenges are to develop capacity, find a basis for stability, increase and share the benefits of economic prosperity, tame emergency power, and safeguard the rule of law. In the other long-standing and relatively prosperous federations, the principal preoccupations are more likely to be efficiency and economic competitiveness; the fine-tuning of institutions to improve accountability; and the impact of the growing influence of international and supranational arrangements on the balance of authority between the national sphere and the constituent units. At another level, however, one challenge at least is common: that of realizing the potential of a system of government that is complex and sophisticated but that nevertheless holds considerable promise for the governance of diverse and demanding societies in an increasingly interdependent world. In response to this challenge, certain trends may, tentatively, be observed. One is a revival of interest in the possibilities as well as the limitations of bicameralism, in both spheres of government. Another, which may conceptually be connected to it, is the gradual emergence of local government as an autonomous sphere of government with constitutional status of its own. A third is the design of intergovernmental structures and processes to facilitate federal governance in a manner that safeguards the democratic accountability upon which the legitimacy and effectiveness of institutions ultimately depend. Necessarily, there is some tension between federalism and majoritarian democracy, particularly if attention is confined to the national government alone. Each federation resolves this tension for itself, in its own version of federal constitutional democracy. An assessment of the strengths and weaknesses of the arrangements that result cannot be judged by the institutions of the national sphere alone.60 Federal democracy necessarily involves the institutions of both spheres, their relations with their respective peoples, and the interaction between them. Notes 1 I am grateful to Ronald L. Watts, Christina Murray, and John Kincaid for their helpful comments on earlier versions of this chapter. 2 For the complete list, see Forum of Federations, List of Federal Countries, < http://www.forumfed.org/federalism/cntrylist.asp?lang=en >, viewed 22 February 2006. 3 See, for example, Domrin’s description in this volume of the meeting pattern of the Russian Federation Council, further balanced by a requirement for members to spend a mandatory period each month in the region they represent. 4 Unless otherwise indicated, discussion of constituent units is confined to units that are full members of the federated polity, with a degree of constitutional autonomy in their own right. It should be noted that, in many federations, including Argentina, Australia, Canada, India, Nigeria, and the United States, there are self-governing territories with significant de facto autonomy that perform the role of a constituent unit in relation to their respective populations. The institutions of such territories are likely to be broadly similar in design to those in the constituent units, although this is not necessarily so. 5 2005 CIA World Factbook, < http://www.photius.com/rankings/economy/gdp_2005_0.html >, viewed 24 May 2006. 6 Of the federations in this volume, Canada and India are examples of a combination of aggregation and devolution. For elaboration of this distinction, and its application in other cases, see Ronald L. Watts, “Comparing Forms of Federal Partnerships,” Theories of Federalism: A Reader, ed. Dimitrios Karmis and Wayne Norman (New York: Palgrave Macmillan, 2005), 233, 249. 7 A table detailing the subdivision is provided in Ignatius Akaayar Ayua and Dakas C.J. Dakas, “Federal Republic of Nigeria,” Constitutional Origins, Structure and Change in Federal Countries, ed. John Kincaid and G. Alan Tarr (Montreal and Kingston: McGill-Queen’s University Press, 2005), 253. 8 Cheryl Saunders, “Budgetary Federalism: Balancing Federalism and Representative Government,” Mensch und Staat: Festschrift for Thomas Fleiner ed. Peter Hanni (Frieburg: Universitatsverlag Freiburg Schweiz, 2003), 175-199. 9 See Michael Burgess, Comparative Federalism, Theory and Practice (London: Routledge, 2006), 140-144. 10 Brendan O’Leary, “Power-Sharing, Pluralist Federation, and Federacy,” The Future of Kurdistan in Iraq, ed. Brendan O’Leary, John McGarry, and Khaled Salih (Philadelphia: University of Pennsylvania Press, 2005), 49. O’Leary identifies “pluri-national” federations as states “in which there are multiple recognised nations, whose respective nationals may be both concentrated and dispersed,” in contrast with multinational federations, “often interpreted as indicating spatially discrete and homogeneously adjacent nations.” 11 Will Kymlicka, “Federalism, Nationalism and Multiculturalism,” Theories of Federalism: A Reader, ed. Dimitrios Karmis and Wayne Norman (New York: Palgrave-Macmillan, 2005), 276. 12 Ibid. 13 Thomas Fleiner, Walter Kalin, Wolf Linder and Cheryl Saunders, “Federalism, Decentralisation and Conflict Management in Multicultural Societies,” Federalism in A Changing World: Learning from Each Other, ed. Raoul Blindenbacher and Arnold Kohler (Montreal and Kingston: McGill-Queen’s University Press, 2002), 227-258; Thomas Fleiner, “The Challenge of Ethnic Diversity to Federalism,” Le Fédéralisme dans tous ses Etats, ed. Jean-Francois Gaudreault-DesBiens and Fabien Gélinas (Québec: Editions Yvon Blais, 2005), 184. 14 For a critical analysis of the difficulties of asymmetry in these circumstances, see Kymlicka, “Federalism, Nationalism and Multiculturalism,” 269-292. 15 Another, which might have been included, is the use of direct democracy. Apart from Switzerland, however, where direct democracy has a profound effect on the operations of institutions in both spheres, its influence so far has been muted (where it has been used at all). For this reason I don not treat it here, but I draw attention to its effect in considering particular institutions below. 16 Louisiana, with its civil-law system, makes the United States technically a mixed legal system. In this context, however, the exception is too minor to cause the categorization to be altered. Even so, the use of a French run-off system of election in Louisiana, in contrast to the rest of the country, is a development of some interest for present purposes. See Louisiana, Revised Statutes, title 18, secs. 481, 511. 17 Johanne Poirier, “Les Ententes Intergouvernementales et la Gouvernance Fédérale: Aux Confins du Droit et du Non-Droit,” in Gaudreault-DesBiens and Gélinas, Le Fédéralisme dans tous ses Etats, 463-472. 18 South Africa Constitution, sec. 37. 19 In the wake of the contested results in the 2002 presidential election Congress enacted the Help America Vote Act, 2002. See also, Century Foundation Working Group on State Implementation of Election Reform, Balancing Access and Integrity (Washington: Century Foundation, 2005) < http://www.reformelections.org/publications.asp?pubid=542 >, viewed 5 March 2006. 20 India Constitution, Article 74, is more explicit than are the other two in this regard, requiring the president to act in accordance with the advice of the council of ministers. 21 Allan R. Brewer-Carias, “Some Problems of the Centralised Federation and Subnational Constitutionalism in Venezuela,” < http://www.camlaw.rutgers.edu/statecon/subpapers/brewer.pdf >, viewed 12 May 2006. Other federations without bicameral national legislatures are St Kitts-Nevis, Micronesia, Comoros, and the United Arab Emirates. 22 Nigeria is an exception. Both houses initiate impeachment proceedings; the investigation is conducted by a separate panel, whose report must be accepted by both houses. 23 The Canadian Senate also has extensive powers, which are rarely used, reflecting its lack of electoral legitimacy. 24 In Nigeria both houses have fixed four-year terms. 25 Australia’s Senate can be dissolved as a whole for the purpose of resolving deadlocks between the houses. The procedure is cumbersome, however, and has been used only six times in more than 100 years of federation. 26 See, however, Burgess, Comparative Federalism, drawing on Federalist 62 to argue that, whatever the origins of the notion, “Equality of representation was ultimately a circumstantial compromise,” 196. 27 While Switzerland has six half-cantons, with one rather than two representatives in the Council of States, this is the consequence of historical circumstances in which three full cantons were divided and, in that sense, is not a departure from the equality principle. 28 In Australia the president of the Senate has a deliberative vote, and, in the case of a tied vote, the question passes in the negative. In South Africa each province has a single vote. 29 In February 2006 Chancellor Angela Merkel and the premiers of the Länder reached agreement on a package of federalism reforms that would reduce the authority of the Bundesrat in return for greater policy authority to the states in relation to, for example, education. German Info, “Long-Sought Deal Paves Way for Federalism Reform,” 24 February 2006, < http://www.germany.info/relaunch/info/publications/week/2006/060224/politics1.html >, viewed 24 May 2006. 30 In Austria the members are elected by the Land legislature but need not be members of it. In South Africa the premier of the province leads its delegation in the National Council of Provinces. 31 Note, however, the suggestion in the chapter on the United States that the absence of a bloc vote requirement for the US Senate inhibited its evolution as a distinctively federal house, even during the period when senators were elected by the state legislatures. 32 Twelve of the 250 members are also appointed by the national executive for their contribution to certain aspects of public life. 33 Note, however, the requirement for a super-majority in the Rajya Sabha to consent to legislation extending the all-India Services and, thus, effectively detracting from the administrative independence of the states, even in relation to their own legislation. 34 Printz v. United States, 521 US 898 (1997); New York v. United States, 505 US 144 (1992). 35 R v. Hughes 202 CLR 535 (2000). 36 South Dakota v. Dole, 483 US 203 (1987); State of Victoria v. Commonwealth (1957) 99 CLR 575. 37 India Constitution, Article 73(1). 38 India Constitution, Article 258. 39 The authority of the federal chamber for this purpose is not necessarily comprehensive. In South Africa, for example, additional administrative responsibility can be conferred on the provinces without attracting the requirement to the consent of the NCOP, acting in its functional federal capacity, sec. 125. 40 South Africa Constitution, sec. 100. See also sec. 125(3), restricting the executive authority of each province by reference to the extent to which it “has the administrative capacity to assume effective responsibility.” 41 German Info, “Long-Sought Deal Paves Way for Federalism Reform,” 24 February 2006, < http://www.germany.info/relaunch/info/publications/week/2006/060224/politics1.html >, viewed 24 May 2006. The proposed changes were still under consideration in May 2006, with an expectation that some of them at least would be made by the end of the calendar year. 42 Nigeria Constitution, Article 14(3). 43 Reference re Remuneration of Judges of the Prov. Court of P.E.I.; Ref re Independence and Impartiality of Judges of the Prov. Court of P.E.I.  3 SCR; Kable 44 Peter W. Hogg, Constitutional Law of Canada, 4th ed. (Toronto, ON: Carswell, 1997), 90-92. 45 Certification of the Kwazulu-Natal Constitution CCT15/96; Certification of the Constitution of the Western Cape CCT6/97. 46 Argentina Constitution, Article 5. 47 US Constitution, Article 4(4). 48 Germany Constitution, Article 28. 49 Canadian Charter of Rights and Freedoms, sec. 33. Sec. 31 of the Charter also has federal significance: “Nothing in this Charter extends the legislative powers of any body or authority.” 50 Similarly, in Canada, constitutional powers of the national sphere to require the reservation of provincial legislation and to disallow provincial legislation have fallen into disuse, under the effect of constitutional convention. 51 Fixed terms in a parliamentary system require provision to be made for the rare but important case in which a government loses the confidence of the legislature and no other government can be formed. 52 South Africa Constitution, sec. 109, allows dissolution within the last two years of a five-year term if so resolved by a majority of the provincial legislature or if there is a vacancy in the office of premier, which has not been filled by the legislature within a stipulated period. 53 South Africa falls (just) outside this group because of the more controlled and targeted nature of its emergency powers. 54 Nigeria Constitution, sec. 305(3)(d). 55 India Constitution, Article 356. 56 South Africa Constitution, sec. 143, provides a default constitution for the provinces but enables them to adopt their own constitutions with legislative and executive (or traditional monarchical structures) that differ from those in the constitutional template. 57 Certification of the Western Cape Constitution 1997 9 BCLR 1167 (CC), 1997 4 SA 795 (CC). 58 For a discussion of the significance of these developments, see Nico Steytler, ed., The Place and Role of Local Government in Federal Systems (Johannesburg: Konrad-Adenauer-Stiftung, 2005). 59 These include appointments to the National Electoral Commission and the National Judicial Council, both of which perform functions for the state as well as the national sphere. 60 See the important distinction drawn by Stepan between federal institutions that are “demos-enabling” and “demos-constraining” in Alfred Stepan, “Toward a New Comparative Politics of Federalism, Multinationalism and Democracy,” E.L. Gibson, ed., Federalism and Democracy in Latin America (Baltimore: Johns Hopkins University Press, 2004), 29.