Patterns of Democracy Arendt Lijphart 1999. New Haven: Yale University Press. Chapter 2 The Westminster Model of Democracy In this book I use the term Westminster model interchangeably with majoritarian model to refer to a general model of democracy. It may also be used more narrowly to denote the main characteristics of British parliamentary and governmental institutions (Wilson 1994; Mahler 1997)-the Parliament of the United Kingdom meets in the Palace of Westminster in London. The British version of the Westminster model is both the original and the best-known example of this model. It is also widely admired. Richard Rose (1974, 131) points out that, "with confidence born of continental isolation, Americans have come to assume that their institutions-the Presidency, Congress and the Supreme Court-are the prototype of what should be adopted elsewhere." But American political scientists. especially those in the field of comparative politics, have tended to hold the British system of government in at least equally high esteem (Kavanagh 1974). One famous political scientist who fervently admired the Westminster model was President Woodrow Wilson. In his early writings he went so far as to urge the abolition of presidential government and the adoption of British-style parliamentary government in the United States. Such views have also been held by many other non-British observers of British politics, and many features of the Westminster model have been exported to [10] other countries: Canada, Australia, New Zealand, and most of Britain's former colonies in Asia, Africa, and the Caribbean when they became independent. Wilson (1884,33) referred to parliamentary government in accordance with the Westminster model as "the world's fashion." The ten interrelated elements of the Westminster or majoritarian model are illustrated by features of three democracies that closely approximate this model and can be regarded as the majoritarian prototypes: the United Kingdom, New Zealand, and Barbados. Britain, where the Westminster model originated, is clearly the first and most obvious example to use. In many respects, however, New Zealand is an even better example-at least until its sharp turn away from majoritarianism in October 1996. The third example-Barbados-is also an almost perfect prototype of the Westminster model, although only as far as the first (executives-parties) dimension of the majoritarian-consensus contrast is concerned. In the following discussion of the ten majoritarian characteristics in the three countries, I emphasize not only their conformity with the general model but also occasional deviations from the model, as well as various other qualifications that need to be made. The Westminster Model in the United Kingdom 1. Concentration of executive power in one-party and baremajority cabinets. The most powerful organ of British government is the cabinet. It is normally composed of members of the party that has the majority of seats in the House of Commons, and the minority is not included. Coalition cabinets are rare. Because in the British two-party system the two principal parties are of approximately equal strength, the party that wins the elections usually represents no more than a narrow majority, and the minority is relatively large. Hence the British one-party and bare-majority cabinet is the perfect embodiment of the principle of majority rule: it wields vast amounts of political power [11] to rule as the representative of and in the interest of a majority that is not of overwhelming proportions. A large minority is excluded from power and condemned to the role of opposition. Especially since 1945, there have been few exceptions to the British norm of one-party majority cabinets. David Butler (1978, 112) writes that "clear-cut single-party government has been much less prevalent than many would suppose," but most of the deviations from the norm-coalitions of two or more parties or minority cabinets-occurred from 1918 to 1945. The most recent instance of a coalition cabinet was the 1940-45 wartime coalition formed by the Conservatives, who had a parliamentary majority, with the Labour and Liberal parties, under Conservative Prime Minister Winston Churchill. The only instances of minority cabinets in the postwar period were two minority Labour cabinets in the 1970s. In the parliamentary election of February 1974, the Labour party won a plurality but not a majority of the seats and formed a minority government dependent on all other parties not uniting to defeat it. New elections were held that October and Labour won an outright, albeit narrow, majority of the seats; but this majority was eroded by defections and by-election defeats, and the Labour cabinet again became a minority cabinet in 1976. It regained a temporary legislative majority in 1977 as a result of the pact it negotiated with the thirteen Liberals in the House of Commons: the Liberals agreed to support the cabinet in exchange for consultation on legislative proposals before their submission to Parliament. No Liberals entered the cabinet, however, and the cabinet therefore continued as a minority instead of a true coalition cabinet. The so-called Lab-Lib pact lasted until 1978, and in 1979 Labour Prime Minister James Callaghan's minority cabinet was brought down by a vote of no confidence in the House of Commons. 2. Cabinet dominance. The United Kingdom has a parliamentary system of government, which means that the cabinet is dependent on the confidence of Parliament. In theory, because the House of Commons can vote a cabinet out of office, it [12] "controls" the cabinet. In reality, the relationship is reversed. Because the cabinet is composed of the leaders of a cohesive majority party in the House of Commons, it is normally backed by the majority in the House of Commons, and it can confidently count on staying in office and getting its legislative proposals approved. The cabinet is clearly dominant vis-A-vis Parliament. Because strong cabinet leadership depends on majority sup port in the House of Commons and on the cohesiveness of the majority party, cabinets lose some of their predominant position when either or both of these conditions are absent. Especially during the periods of minority government in the 1970s, there was a significant increase in the frequency of parliamentary de feats of important cabinet proposals. This even caused a change in the traditional view that cabinets must resign or dissolve the House of Commons and call for new elections if they suffer a defeat on either a parliamentary vote of no confidence or a major bill of central importance to the cabinet. The new unwritten rule is that only an explicit vote of no confidence necessitates resignation or new elections. The normalcy of cabinet dominance was largely restored in the 1980s under the strong leadership of Conservative Prime Minister Margaret Thatcher. Both the normal and the deviant situations show that it is the disciplined two-party system rather than the parliamentary system that gives rise to executive dominance. In multiparty parliamentary systems, cabinets-which are often coalition cabinets tend to be much less dominant (Peters 1997). Because of the concentration of power in a dominant cabinet, former cabinet minister Lord Hailsham (1978, 127) has called the British system of government an "elective dictatorship."l [13] 3. Two-party system. British politics is dominated by two large parties: the Conservative party and the Labour party. Other parties also contest elections and win seats in the House of Commons-in particular the Liberals and, after their merger with the Social Democratic party in the late 1980s, the Liberal Democrats-but they are not large enough to be overall victors. The bulk of the seats are captured by the two major parties, and they form the cabinets: the Labour party from 1945 to 1951, 1964 to 1970, 1974 to 1979, and from 1997 on, and the Conservatives from 1951 to 1964, 1970 to 1974, and in the long stretch from 1979 to 1997. The hegemony of these two parties was especially pronounced between 1950 and 1970: jointly they never: won less than 87.5 percent of the votes and 98 percent of the seats in the House of Commons in the seven elections held in this period. The interwar years were a transitional period during which the Labour party replaced the Liberals as one of the two big parties, and in the 1945 election, the Labour and Conservative . parties together won about 85 percent of the votes and 92.5 percent of the seats. Their support declined considerably after 1970: their joint share of the popular vote ranged from only 70 percent (in 1983) to less than 81 percent (in 1979), but they continued to win a minimum of 93 percent of the seats, except in 1997, when their joint seat share fell to about 88.5 percent. The Liberals were the main beneficiaries. In alliance with the Social Democratic party, they even won more than 25 percent of the vote on one occasion (in the 1983 election) but, until 1997 , never more than fourteen seats by themselves and twenty-three seats in alliance with the Social Democrats. In the 1997 election, however, the Liberal Democrats captured a surprising forty-six seats with about 17 percent of the vote. [14] A corollary trait of two-party systems is that they tend to be one-dimensional party systems; that is, the programs and policies of the main parties usually differ from each other mainly with regard to just one dimension, that of socioeconomic issues. This is clearly the case for the British two-party system. The principal politically significant difference that divides the Conservative and Labour parties is disagreement about socioeconomic policies: on the left-right spectrum, Labour represents the left-of-center and the Conservative party the right-of-center preferences. This difference is also reflected in the pattern of voters' support for the parties in parliamentary elections: working-class voters tend to cast their ballots for Labour candidates and middle-class voters tend to support Conservative candidates. The Liberals and Liberal Democrats can also be placed easily on the socioeconomic dimension: they occupy a center position. There are other differences, of course, but they are much less salient and do not have a major effect on the composition of the House of Commons and the cabinet. For instance, the Protestant-Catholic difference in Northern Ireland is the overwhelmingly dominant difference separating the parties and their supporters, but Northern Ireland contains less than 3 percent of the population of the United Kingdom, and such religious differences are no longer politically relevant in the British part of the United Kingdom (England, Scotland, and Wales). Ethnic differences explain the persistence of the Scottish National party and the Welsh nationalists, but these parties never manage to win more than a handful of seats. The only slight exception to the one-dimensionality of the British party system is that a foreign-policy issue-British membership in the European Community-has frequently been a source of division both within and between the Conservative and Labour parties. 4. Majoritarian and disproportional system of elections. The House of Commons is a large legislative body with a membership that has ranged from 625 in 1950 to 659 in 1997. The mem- [15] bers are elected in single-member districts according to the plurality method, which in Britain is usually referred to as the "first past the post" system: the candidate with the majority vote or, if there is no majority, with the largest minority vote wins. This system tends to produce highly disproportional results. For instance, the Labour party won an absolute parliamentary majority of 319 out of 635 seats with only 39.3 percent of the vote in the October 1974 elections, whereas the Liberals won only 13 seats with 18.6 percent of the vote-almost half the Labour vote. In the five elections since then, from 1979 to 1997, the winning party has won clear majorities of seats with never more than 44 percent of the vote. All of these majorities have been what Douglas W. Rae (1967, 74) aptly calls "manufactured majorities"-majorities that are artificially created by the electoral system out of mere pluralities of the vote. In fact, all the winning parties since 1945 have won with the benefit of such manufactured majorities. It may therefore be more accurate to call the United Kingdom a pluralitarian democracy instead of a majoritarian democracy. The disproportionality of the plurality method can even produce an overall winner who has failed to win a plurality of the votes: the Conservatives won 8 clear seat majority in the 1951 election not just with less than a majority of the votes but also with fewer votes than the Labour party had received. The disproportional electoral system has been particularly disadvantageous to the Liberals and Liberal Democrats, who have therefore long been in favor of introducing some form of proportional representation (PR). But because plurality has greatly benefitted the Conservatives and Labour, these two major parties have remained committed to the old disproportional method. Nevertheless, there are some signs of movement in the direction of PR. For one thing, PR was adopted for all elections in Northern Ireland (with the exception of elections to the House of Commons) after the outbreak of Protestant-Catholic strife in the early 1970s. For another, soon after Labour's elec- [16] tion victory in 1997, Prime Minister Tony Blair's new cabinet decided that the 1999 election of British representatives to the European Parliament would be by PR-bringing the United Kingdom in line with all of the other members of the European Union. PR will also be used for the election of the new regional assemblies for Scotland and Wales. Moreover, an advisory Commission on Voting Systems, chaired by former cabinet member Lord Jenkins, was instituted to propose changes in the electoral system, including the possibility of PR, for the House of Commons. Clearly, the principle of proportionality is no longer anathema. Still, it is wise to heed the cautionary words of Graham Wilson (1997,72), who points out that the two major parties have a long history of favoring basic reforms, but only until they gain power; then "they back away from changes such as electoral reform which would work to their disadvantage." 5. Interest group pluralism. By concentrating power in the hands of the majority, the Westminster model of democracy sets up a government-versus-opposition pattern that is competitive and adversarial. Competition and conflict also characterize the majoritarian model's typical interest group system: a system of free-for-all pluralism. It contrasts with interest group corporatism in which regular meetings take place between the representatives of the government, labor unions, and employers' organizations to seek agreement on socioeconomic policies; this process of coordination is often referred to as concertation, and the agreements reached are often called tripartite pacts. Concertation is facilitated if there are relatively few, large, and strong interest groups in each of the main functional sectors-labor, employers, farmers-and/ or if there is a strong peak organization in each of the sectors that coordinates the preferences and desired strategies for each sector. Pluralism, in contrast, means a multiplicity of interest groups that exert pressure on the government in an uncoordinated and competitive manner. Britain's interest group system is clearly pluralist. The one exception is the 1975 Social Contract on wages and prices con- [WESTMINISTER MODEL 17] cluded between the Labour government, the main labor union federation (the Trades Union Congress), and the main employers' federation (the Confederation of British Industry). This contract fell apart two years later when the government failed to get union agreement to accept further wage restraints and imposed wage ceilings unilaterally. The 1980s were characterized even more by grim confrontations between Margaret Thatcher's Conservative government and the labor unions-the very opposite of concertation and corporatism. As Michael Gallagher, Michael Laver, and Peter Mair (1995, 370) point out, Britain is "decidedly not a corporatist system II for two important reasons: "The first is the general lack of integration of both unions and management into the policymaking process. The second is the apparent preference of both sides for confrontational methods of settling their differences." 6. Unitary and centralized government. The United Kingdom is a unitary and centralized state. Local governments perform a series of important functions, but they are the creatures of the central government and their powers are not constitutionally guaranteed (as in a federal system). Moreover, they are financially dependent on the central government. There are no clearly designated geographical and functional areas from which the parliamentary majority and the cabinet are barred. The Royal Commission on the Constitution under Lord Kilbrandon concluded in 1973: "The United Kingdom is the largest unitary state in Europe and among the most centralised of the major industrial countries in the world" (cited in Busch 1994, 60). More recently, Prime Minister Tony Blair called the British system "the most centralised government of any large stale in the western world" (cited in Beer 1998,25). Two exceptions should be noted. One is that Northern Ireland was ruled by its own parliament and cabinet with a high degree of autonomy-more than what most states in federal systems have-from 1921, when the Republic of Ireland became independent, until the imposition of direct rule from London in [18 WESTMINISTER MODEL] 1972. It is also significant, however, that Northern Ireland's autonomy could be, and was, eliminated in 1972 by Parliament by means of a simple majoritarian decision. The second exception is the gradual movement toward greater autonomy for Scotland and Wales-"devolution," in British parlance. But it was not until September 1997 that referendums in Scotland and Wales finally approved the creation of autonomous and directly elected Scottish and Welsh assemblies and that Prime Minister Blair could proclaim the end of the "era of big centralized government" (cited in Buxton, Kampfner, and Groom 19970 1). 7. Concentration of legislative power in a unicameral legislature. For the organization of the legislature, the majoritarian principle of concentrating power means that legislative power should be concentrated in a single house or chamber. In this respect, the United Kingdom deviates from the pure majoritarian model. Parliament consists of two chambers: the House of Commons, which is popularly elected, and the House of Lords, which consists mainly of members of the hereditary nobility but also contains a large number of so-called life peers, appointed by the government. Their relationship is asymmetrical: almost all legislative power belongs to the House of Commons. The only power that the House of Lords retains is the power to delay legislation: money bills can be delayed for one month and all other bills for one year. The one-year limit was established in 1949; between the first major reform of 1911 and 1949, the Lords' delaying power was about two years, but in the entire period since 1911 they have usually refrained from imposing long delays. Therefore, although the British bicameral legislature deviates from the majoritarian model, it does not deviate much: in everyday discussion in Britain, "Parliament" refers almost exclusively to the House of Commons, and the highly asymmetric bicameral system may also be called near-unicameralism. Moreover, the Lords' power may well be reduced further. Especially in the Labour party, there is strong sentiment in favor of reforms [WESTMINISTER MODEL 19] that range from eliminating the voting rights of the hereditary members to the abolition of the House of Lords. The change from near-unicameralism to pure unicameralism would not be a difficult step: it could be decided by a simple majority in the House of Commons and, if the Lords objected, merely a one-year delay. 8. Constitutional flexibility. Britain has a constitution that is "unwritten " in the sense that there is not one written document that specifies the composition and powers of the governmental institutions and the rights of citizens. These are defined instead in a number of basic laws-like the Magna Carta of 1215, the Bill of Rights of 1689, and the Parliament Acts of 1911 and 1949 common law principles, customs, and conventions. The fact that the constitution is unwritten has two important implications. One is that it makes the constitution completely flexible because it can be changed by Parliament in the same way as any other laws-by regular majorities instead of the supermajorities, like two-thirds majorities, required in many other democracies for amending their written constitutions. One slight exception to this flexibility is that opposition by the House of Lords may force a one-year delay in constitutional changes. 9. Absence of judicial review. The other important implication of an unwritten constitution is the absence of judicial review: there is no written constitutional document with the status of "higher law" against which the courts can test the constitutionality of regular legislation. Although Parliament normally accepts and feels bound by the rules of the unwritten constitution, it is not formally bound by them. With regard to both changing and interpreting the constitution, therefore, Parliament that is, the parliamentary majority-can be said to be the ultimate or sovereign authority. In A. V. Dicey's (1915, 37-38 famous formulation, parliamentary sovereignty "means neither more nor less than this, namely, that Parliament. . . has, under the English constitution, the right to make or unmake any law whatever; and, further, that no person or body is recognised by [20 WESTMINISTER MODEL] the law of England as having a right to override or set aside the legislation of Parliament. " One exception to parliamentary sovereignty is that when Britain entered the European Community-a supranational instead of merely an international organization-in 1973, it accepted the Community's laws and institutions as higher authorities than Parliament with regard to several areas of policy. Because sovereignty means supreme and ultimate authority, Parliament can therefore no longer be regarded as fully sovereign. Britain's membership in the European Community now called the European Union-has also introduced a measure of judicial review both for the European Court of Justice and for British courts: "Parliament's supremacy is challenged by the right of the Community institutions to legislate for the United Kingdom (without the prior consent of Parliament) and by the right of the courts to rule on the admissibility (in terms of Community law) of future acts of Parliament" (Coombs 1977, 88). Similarly, Britain has been a member of the European convention on Human Rights since 1951, and its acceptance of an optional clause of this convention in 1966 has given the European Court of Human Rights in Strasbourg the right to review and invalidate any state action, including legislation, that it judges to violate the human rights entrenched in the convention (Cappelletti 1989,202; Johnson 1998,155-58). 10. A central bank controlled by the executive. Central banks are responsible for monetary policy, and independent banks are widely considered to be better at controlling inflation and maintaining price stability than banks that are dependent on the executive. However, central bank independence is clearly in conflict with the Westminster model's principle of concentrating power in the hands of the one-party majority cabinet. As expected, the Bank of England has indeed not been able to act independently and has instead been under the control of the cabinet. During the 1980s, pressure to make the Bank of England more autonomous increased. Two Conservative chancellors of [21] the exchequer tried to convince their colleagues to take this big step away from the Westminster model, but their advice was rejected (Busch 1994,59). It was not until 1997--one of the first decisions of the newly elected Labour government-that the Bank of England was given the independent power to set interest rates. The Westminster Model in New Zealand Many of the Westminster model's features have been exported to other members of the British Commonwealth, but only one country has adopted virtually the entire model: New Zealand. A major change away from majoritarianism took place in 1996 when New Zealand held its first election by PR, but the New Zealand political system before 1996 can serve as a second instructive example of how the Westminster model works. 1. Concentration of executive power in one-party and bare-majority cabinets. For six decades, from 1935 to the mid-1990s, New Zealand had single-party majority cabinets without exceptions or interruptions. Two large parties-the Labour party and the National party-dominated New Zealand politics, and they alternated in office. The one-party majority cabinet formed after the last plurality election in 1993 suffered a series of defections and briefly became a quasi-coalition cabinet (a coalition with the recent defectors), then a one-party minority cabinet, and finally a minority coalition-but all of these unusual cabinets occurred in the final phase of the transition to the new non-Westminster system (Boston, Levine, McLeay, and Roberts 1996,93-96). The only other deviations from single-party majority government happened much earlier: New Zealand had a wartime coalition cabinet from 1915 to 1919, and another coalition was in power from 1931 to 1935. 2. Cabinet dominance. In this respect, too, New Zealand was a perfect example of the Westminster model. Just as during most of the postwar period in the United Kingdom, the combination [22] of the parliamentary system of government and a two-party system with cohesive parties made the cabinet predominate over the legislature. In the words of New Zealand political scientist Stephen Levine (1979, 25-26), the "rigidly disciplined two-party system has contributed to the concentration of power within the Cabinet, formed from among the Members of Parliament. . . belonging to the majority party." 3. Two-party system. Two large parties were in virtually complete control of the party system, and only these two formed cabinets during the six decades from 1935 to the mid-1990s: the Labour party (1935-49, 1957-60, 1972-75, and 1984-90) and the National party (1949-57, 1960-72, 1975-84, and after 1990). Party politics revolved almost exclusively around socioeconomic issues-Labour represented left-of-center and the National party right-of-center political preferences. Moreover, unlike in Britain, third parties were almost absent from the New Zealand House of Representatives. In eleven of the seventeen elections from 1946 to 1993, the two large parties divided all of the seats; in five elections, only one other party gained one or two seats; and in 1993, two small parties gained two seats each (out of ninety-nine). New Zealand's two-party system was therefore an almost pure two-party system. 4. Majoritarian and disproportional system of elections. The House of Representatives was elected according to the plurality method in single-member districts. The only unusual feature was that there were four special large districts, geographically overlapping the regular smaller districts, that were reserved for the Maori minority (comprising about 12 percent of the population). These four districts entailed a deviation from the majoritarianism of the Westminster model because their aim was to guarantee minority representation. From 1975 on, all Maori voters have had the right to register and vote either in the regular district or in the special Maori district in which they reside. As in the United Kingdom, the plurality system produced severely disproportional results, especially in 1978 and 1981. In [23] the 1978 election, the National party won a clear majority of fifty-one out of ninety-two seats even though it won neither a majority of the popular vote-its support was only 39.8 percent-nor a plurality, because Labour's popular vote was 40.4 percent; the Social Credit party's 17.1 percent of the vote yielded only one seat. In 1981, the National party won another parliamentary majority of forty-seven out of ninety-two seats and again with fewer votes than Labour, although the respective percentages were closer: 38.8 and 39.0 percent; Social Credit now won 20.7 percent of the popular vote-more than half of the votes gained by either of the two big parties-but merely two seats. Moreover, all of the parliamentary majorities from 1954 on were manufactured majorities, won with less than majorities of the popular vote. In this respect, New Zealand was, like the United Kingdom, more a pluralitarian than a majoritarian democracy. 5. Interest group pluralism. New Zealand's interest group system, like Britain's, is clearly pluralist. Also, again like Britain, New Zealand has had high strike levels-indicative of confrontation instead of concertation between labor and management. In comparative studies of corporatism and pluralism, many scholars have tried to gauge the precise degree to which the interest group systems of the industrialized democracies are corporatist or pluralist. Their judgments differ considerably with regard to a few of these countries, but on Great Britain and New Zealand there is little disagreement: both belong on the extreme pluralist end of the pluralist-corporatist spectrum. New Zealand, moreover, is generally judged to be slightly more pluralist than Britain (Lijphart and Crepaz 1991). Hence in this respect, too, New Zealand is the somewhat better example of the Westminster model. 6. Unitary and centralized government. The " Act to grant a Representative Constitution to the Colony of New Zealand," passed by the British parliament in 1852, created six provinces with considerable autonomous powers and functions vis-a-vis [24] the central government, but these provinces were abolished in 1875. Today's governmental system is unitary and centralized not as surprising, of course, for a country with a population of less than four million than for the United Kingdom with its much larger population of about sixty million people. 7. Concentration of legislative power in a unicameral legislature. For about a century, New Zealand had a bicameral legislature, consisting of an elected lower house and an appointed upper house, but the upper house gradually lost power. Its abolition in 1950 changed the asymmetrical bicameral system into pure unicameralism. 8. Constitutional flexibility. Like the United Kingdom, New Zealand lacks a single written constitutional document. Its "unwritten" constitution has consisted of a number of basic laws like the Constitution Acts of 1852 and 1986, the Electoral Acts of 1956 and 1993, and the Bill of Rights Act of 1990-conventions, and customs.2 Some key provisions in the basic laws are "entrenched" and can be changed only by three-fourths majorities of the membership of the House of Representatives or by a majority vote in a referendum; however, this entrenchment can always be removed by regular majorities, so that, in the end, majority rule prevails. Hence, like the British parliament, the parliament of New Zealand is sovereign. Any law, including laws that "amend" the unwritten constitution, can be adopted by regular majority rule. As one of New Zealand 's constitutional law experts puts it, "The central principle of the Constitution is that there are no effective legal limitations on what Parliament may enact by the ordinary legislative process" (Scott 1962,39). 9. Absence of judicial review. Parliamentary sovereignty also means, as in Britain, that the courts do not have the right of judicial review. The House of Representatives is the sole judge of the constitutionality of its own legislation. 10. A central bank controlled by the executive. Andreas [25] Busch (1994,65) writes that historically New Zealand "has been a country with . . . a very low degree of central bank independence," and for the period until 1989, he gives the Reserve Bank of New Zealand his lowest rating-indicating even less autonomy than that of the Bank of England. This situation was changed radically by the Reserve Bank Act of 1989. Price stability was now defined as the primary aim of monetary policy, and the central bank was given the sole responsibility not to exceed the target rate of inflation, the precise level of which has to be negotiated between the central bank and the minister of finance. Inflation levels have decreased dramatically in New Zealand: measured in terms of the consumer price index, inflation was at double-digit levels during six years in the 1980s, but it was only 2 percent on average from 1991 to 1997 (OECD 1998, 240). Greater central bank independence must be given at least some of the credit for this success. With only two exceptions-the parliamentary seats reserved for the Maori minority and the earlier shift to central bank autonomy-democracy in New Zealand was, until 1996, more clearly majoritarian and hence a better example of the Westminster model than British democracy. In fact, especially in view of the minority cabinets and frequent defeats of cabinet proposals in Britain in the 1970s, Richard Rose could legitimately claim that New Zealand was "the only example of the true British system left" (personal, communication, April 8, 1982). However, the adoption of PR and the first PR election of parliament in October 1996 entailed a radical shift away from the Westminster model. The two major parties were opposed to PR, but they both unintentionally contributed to its adoption. The first impetus was the Labour party's unhappiness with the results of the 1978 and 1981 elections, mentioned above, in which the National party won parliamentary majorities not only with less than 40 percent of the popular vote but with fewer votes than the Labour party had received. When Labour was returned to power in 1984, it appointed a Royal Commission on the Electoral System [28 WESTMINSTER. MODEL to recommend improvements. The commission's terms of reference were very broad, however, and it recommended not just small adjustments but a radical change to PR as well as a referendum on whether to adopt it. The government tried to deflect this proposal by turning it over to a parliamentary committee, which, as expected, rejected PR and instead merely recommended minor changes. The election campaign of 1987 put PR back on the political agenda: the Labour prime minister promised to let the voters decide the issue by referendum, but his party retreated from this pledge after being reelected. Seeking to embarrass Labour, the National party opportunistically made the same promise in the 1990 campaign, and when they won the election, they could not avoid honoring it. The voters then twice endorsed PR in referendums held in 1992 and 1993 (Jackson and McRobie 1998). The form of PR that was adopted and used in the 1996 election was a system, modeled after the German system, in which sixty-five members are elected by plurality in single-member districts-including five special Maori districts-and fifty-five members by PR from party lists; a crucial provision is that this second set of fifty-five seats is allocated to the parties in a way that makes the overall result as proportional as possible. Therefore, although the New Zealand term for this system is the "mixed member proportional" (MMP) system, implying that it is a mixture of PR and something else, it is in fact clearly and fully a PR system.3 The first PR election instantly transformed New Zealand politics in several respects (Vowles, Aimer, Banducci, and Karp 1998). First, the election result was much more proportional than those of the previous plurality elections. The largest party, the National party, was still over-represented, but by less than three percentage points; it won 33.8 percent of the vote and 36.7 3. Each voter has two votes, one for a district candidate and one for a party list. To avoid excessive fragmentation, parties must win either a minimum of 5 percent of the list votes or at least one district seat to qualify for list seats. [27] percent of the seats. Second, the election produced a multiparty system with an unprecedented six parties gaining representation in parliament. Third, unlike in any other postwar election, no party won a majority of the seats. Fourth, an ethnic dimension was added to the party system: the New Zealand First party, led by a Maori and winning seventeen seats, including all five of the special Maori seats, became the main representative of the Maori minority (although it was not a specifically Maori party nor supported exclusively by Maori voters). The Christian Coalition almost succeeded in making the party system even more multidimensional by adding a religious issue dimension, but its vote fell just short of the required 5 percent threshold. Fifth, in contrast with the long line of previous single-party majority cabinets, a two-party coalition cabinet was formed by the National and New Zealand First parties. Because of these significant deviations from the majoritarian model, post-1996 New Zealand is no longer a good, let alone the best, example of the "true British system." Hence, in Kurt von Mettenheim's (1997,11) words, "the United Kingdom [now] appears to be the only country to have retained the central features of the Westminster model..' It should be noted, however, that all of the post-1996 changes in New Zealand have to do with the executives-parties dimension of the majoritarian model, comprising the first five of the ten characteristics of the model, and that, especially with regard to this first dimension, several other former British colonies continue to have predominantly Westminster-style institutions. A particularly clear and instructive example is Barbados. The Westminster Model in Barbados Barbados is a small island state in the Caribbean with a population of about a quarter of a million. It has a "strongly homogeneous society" that is mainly of African descent (Duncan 1994, 77). It gained its independence from Britain in 1966, but there [38 WESTMINSTER. MODEL] continues to be "a strong and pervasive sense of British tradition and culture" (Banks, Day, and Muller 1997, 69)-including British political traditions. Barbados is often called the "Little England" of the Caribbean. 1. Concentration of executive power in one-party and bare-majority cabinets. Since independence in 1966, Barbados has had single-party majority cabinets. Its two large parties-the Barbados Labour party (BLP) and the Democratic Labour party (DLP)-have been the overwhelmingly dominant forces in Barbados politics, and they have alternated in office. Unlike in the British and New Zealand cases, there are no exceptions or qualifications to this pattern that need to be noted. In fact, the pattern extends back to colonial times. Ever since the establishment of universal suffrage and cabinet government in the early 1950s, the sequence of single-party majority cabinets has been unbroken. 2. Cabinet dominance. Barbadian cabinets have been at least as dominant as those of the two earlier examples of the Westminster model. The term elective dictatorship, coined by Lord Hailsham for Britain also fits the Barbados system well (Payne 1993, 69). One special reason for the predominance of the cabinet in Barbados is the small size of the legislature. The Barbadian House of Assembly had only twenty-four members from 1966 to 1981; this number was increased only slightly to twenty-seven in 1981 and twenty-eight in 1991. Many of the legislators are therefore also cabinet ministers, which in turn means that, as Trevor Munroe (1996, 108) points out, almost one-third of the members of the legislature "are in effect constitutionally debarred from an independent and critical stance in relation to the executive." 3. Two-party system. The same two large parties have controlled the party politics of Barbados since independence, and they have formed all of the cabinets: the DLP from 1966 to 1976 and from 1986 to 1994, and the BLP between 1976 to 1986 and from 1994 on. These two parties differ from each other mainly [29] on socioeconomic issues, with the BLP occupying the right-of-center and the DLP the left-of-center position on the left-right spectrum. In five of the seven elections since 1966, no third parties won any seats; only one small party won two seats in 1966, and another small party won one seat in 1994. The strength of the two-party system is also illustrated by the fate of the four members of parliament who defected from the ruling DLP in 1989 and formed a separate party. As Tony Thorndike (1993, 158) writes, this new party "did not long survive the logic of the 'first past the post' Westminster system and the two-party culture of Barbados. In elections in January 1991 it lost all its four seats." 4. Majoritarian and disproportional system of elections. In the elections before independence, including the 1966 election, which was held several months before formal independence took place, Barbados used the plurality method but not in the usual single-member districts. Instead, two-member districts were used (Emmanuel 1992, 6; Duncan 1994, 78); these tend to increase the disproportionality of the election results because, in plurality systems, disproportionality increases as the number of representatives elected per district increases. Since 1971, all elections have been by plurality in single-member districts, but electoral disproportionality has remained high. For instance, in 1971 the DLP won three-fourths of the seats with 57.4 percent of the votes, and in 1986 it won twenty-four of the twenty-seven seats (88.9 percent) with 59.4 percent of the votes. In three of the elections since 1966, the parliamentary majorities were "manufactured" from pluralities of the vote, but in the other four elections the seat majorities were genuinely "earned" with popular vote majorities. On balance, therefore, Barbados has been less of a pluralitarian democracy than Britain and New Zealand. Moreover, unlike the other two countries, Barbados has not experienced any instances of a parliamentary majority won on the basis of a second-place finish in the popular vote. 5. Interest group pluralism. Again like the United Kingdom [30] and New Zealand, Barbados has had an interest group system that is pluralist rather than corporatist. In recent years, however, there has been a trend toward corporatist practices. In 1993, the government, business leaders, and labor unions negotiated a pact on wages and prices, which included a wage freeze. This agreement was replaced two years later by a new and more flexible tripartite pact. 6-10. The characteristics of the second (federal-unitary) dimension of the majoritarian model. Barbados has a unitary and centralized form of government-hardly surprising for a small country with only a quarter of a million people-but as far as the other four characteristics of the federal-unitary dimension are concerned, it does not fit the pure majoritarian model. It has a bicameral legislature consisting of a popularly elected House of Assembly and an appointed Senate that can delay but not veto a case of asymmetrical bicameralism. It has a written constitution that can be amended only by two-thirds majorities in both houses of the legislature. The constitution explicitly gives the courts the right of judicial review. Finally, the central bank of Barbados has a charter that gives it a medium degree of autonomy in monetary policy (Cukierman, Webb, and Neyapti 1994, 45). Anthony Payne ( 1993) argues that the former British colonies in the Caribbean are characterized not by Westminster systems but by "Westminster adapted." As illustrated by Barbados-but by and large also true for the other Commonwealth democracies in the region-this adaptation has affected mainly the second dimension of the Westminster model. On the first (executives parties) dimension, the Westminster model has remained almost completely intact. The fact that Barbados deviates from majoritarianism with regard to most of the characteristics of the federal-unitary dimension does not mean, of course, that it deviates to such an extent that it is a good example of the contrasting model of consensus democracy. In order to illustrate the consensus model, I turn in the next chapter to the examples of Switzerland, Belgium, and the European Union. Footnotes to Chapter 2 1. In presidential systems of government. in which the presidential executive cannot normally be removed by the legislature (except by impeachment), the same variation in the degree of executive dominance can occur, depending on exactly how governmental powers are separated. In the United States, president and Congress can be said to be in a rough balance of power, but presidents in France and in some of the Latin American countries are considerably more powerful. Guillermo O'Donnell (1994, 59-60) has proposed the term "delegative democracy"-akin to Hailsham's "elective dictatorship"-for systems with directly elected and dominant presidents; in such "strongly majoritarian" systems, "whoever wins election to the presidency is thereby entitled to govern as he or she sees fit, constrained only by the hard facts of existing power relations and by a constitutionally limited term of office." 2. The Constitution Act of 1852 and Electoral Act of 1956 were superseded by the two later acts. CHAPTER 3 The Consensus Model of Democracy [31] The majoritarian interpretation of the basic definition of democracy is that it means "government by the majority of the people." It argues that majorities should govern and that minorities should oppose. This view is challenged by the consensus model of democracy. As the Nobel Prize-winning economist Sir Arthur Lewis (1965,64-65) has forcefully pointed out, majority rule and the government-versus-opposition pattern of politics that it implies may be interpreted as undemocratic because they are principles of exclusion. Lewis states that the primary meaning of democracy is that "all who are affected by a decision should have the chance to participate in making that decision either directly or through chosen representatives." Its secondary meaning is that "the will of the majority shall prevail." If this means that winning parties may make all the governmental decisions and that the losers may criticize but not govern, Lewis argues, the two meanings are incompatible: "to exclude the losing groups from participation in decision-making clearly violates the primary meaning of democracy" Majoritarians, can legitimately respond that, under two conditions, the incompatibility noted by Lewis can be resolved. First, the exclusion of the minority is mitigated if majorities and minorities alternate in government-that is, if today's minority can become the majority in the next election instead of being [32] condemned to permanent opposition. This is how the British, New Zealand, and Barbadian two-party systems have worked. In Barbados, alternation has operated perfectly since independence in 1966: neither of the two main parties has won more than two elections in a row. In Britain and New Zealand, however, there have been long periods in which one of the two main parties was kept out of power: the British Labour party during the thirteen years from 1951 to 1964 and the eighteen years from 1979 to 1997, the New Zealand National party for fourteen years from 1935 to 1949, and New Zealand Labour for twelve years from 1960 to 1972. Even during these extended periods of exclusion from power, one can plausibly argue that democracy and majority rule were not in conflict because of the presence of a second condition: the fact that all three countries are relatively homogeneous societies and that their major parties have usually not been very far apart in their policy outlooks because they have tended to stay close to the political center. One party's exclusion from power may be undemocratic in terms of the "government by the people" criterion, but if its voters' interests and preferences are reasonably well served by the other party's policies in government, the system approximates the "government for the people" definition of democracy. In less homogeneous societies neither condition applies. The policies advocated by the principal parties tend to diverge to a greater extent, and the voters' loyalties are frequently more rigid, reducing the chances that the main parties will alternate in exercising governmental power. Especially in plural societies, societies that are sharply divided along religious, ideological, linguistic, cultural, ethnic, or racial lines into virtually separate subsocieties with their own political parties, interest groups, and media of communication-the flexibility necessary for majoritarian democracy is likely to be absent. Under these conditions, majority rule is not only undemocratic but also dangerous, because minorities that are continually denied access to power [33] will feel excluded and discriminated against and may lose their allegiance to the regime. For instance, in the plural society of Northern Ireland, divided into a Protestant majority and a Catholic minority, majority rule meant that the Unionist party representing the Protestant majority won all the elections and formed all of the governments between 1921 and 1972. Massive Catholic protests in the late 1960s developed into a Protestant-Catholic civil war that could be kept under control only by British military intervention and the imposition of direct rule from London. In the most deeply divided societies, like Northern Ireland, majority rule spells majority dictatorship and civil strife rather than democracy. What such societies need is a democratic regime that emphasizes consensus instead of opposition, that includes rather than excludes, and that tries to maximize the size of the ruling majority instead of being satisfied with a bare majority: consensus democracy. Despite their own majoritarian inclinations, successive British cabinets have recognized this need: they have insisted on PR in all elections in Northern Ireland (except those to the House of Commons) and, as a precondition for returning political autonomy to Northern Ireland, on broad Protestant-Catholic power-sharing coalitions. PR and power-sharing are also key elements in the agreement on Northern Ireland reached in 1998. Similarly, Lewis (1965,51-55,6584) strongly recommends PR, inclusive coalitions, and federalism for the plural societies of West Africa. The consensus model is obviously' also appropriate for less divided but still heterogeneous countries, and it is a reasonable and workable alternative to the Westminster model even in fairly homogeneous countries. The examples I use to illustrate the consensus model are Switzerland, Belgium, and the European Union-all multiethnic entities. Switzerland is the best example: with one exception it approximates the pure model perfectly. Belgium also provides a good example, especially after it formally became a federal state in 1993; I therefore pay particular attention to the [34] pattern of Belgian politics in the most recent period. The European Union (EU) is a supranational organization-more than just an international organization-but it is not, or not yet, a sovereign state. Because of the EU's intermediate statust analysts of the European Union disagree on whether to study it as an international organization or an incipient federal state, but the latter approach is increasingly common (Hix 1994). This is also my approach: if the EU is regarded as a federal state, its institutions are remarkably close to the consensus model of democracy. I discuss the Swiss and Belgian prototypes first and in tandem with each other and then turn to the EU example. The Consensus Model in Switzerland and Belgium The consensus model of democracy may be described in terms of ten elements that stand in sharp contrast to each of the ten majoritarian characteristics of the Westminster model. Instead of concentrating power in the hands of the majority, the consensus model tries to share, disperse, and restrain power in a variety of ways. 1. Executive power-sharing in broad coalition cabinets. In contrast to the Westminster model's tendency to concentrate executive power in one-party and bare-majority cabinets, the consensus principle is to let all or most of the important parties share executive power in a broad coalition. The Swiss sevenmember national executive, the Federal Council, offers an excellent example of such a broad coalition: the three large partiesChristian Democrats, Social Democrats, and Radical Democrats -each of which has held about one-fourth of the seats in the lower house of the legislature during the post- World War n era, and the Swiss Peoplets party, with about one-eighth of the seats, share the seven executive positions proportionately according to the so-called magic formula of 2:2:2:1, established in 1959. An additional criterion is that the linguistic groups be represented [35] in rough proportion to their sizes: four or five German-speakers, one or two French-speakers, and frequently an Italian-speaker. Both criteria are informal rules but are strictly obeyed. The Belgian constitution offers an example of a formal requirement that the executive include representatives of the large linguistic groups. For many years, it had already been the custom to form cabinets with approximately equal numbers of ministers representing the Dutch-speaking majority and the French-speaking minority. This became a formal rule in 1970, and the new federal constitution again stipulates that "with the possible exception of the Prime Minister, the Council of Ministers [cabinet] includes as many French-speaking members as Dutch-speaking members" (Alen and Ergec 1994). Such a rule does not apply to the partisan composition of the cabinet, but there have only been about four years bf one-party rule in the postwar era, and since 1980 all cabinets have been coalitions of between four and six parties. 2. Executive-legislative balance of power. The Swiss political system is neither parliamentary nor presidential. The relationship between the executive Federal Council and the legislature is explained by Swiss political scientist ]iirg Steiner (1974,43) as follows: "The members of the council are elected individually for a fixed term of four years, and, according to the Constitution, the legislature cannot stage a vote of no confidence during that period. If a government proposal is defeated by Parliament, it is not necessary for either the member sponsoring this proposal or the Federal Council as a body to resign." This formal separation of powers has made both the executive and the legislature more independent; and their relationship is much more balanced than cabinet-parliament relationships in the British, New Zealand, and Barbadian cases in which the cabinet is clearly dominant. The Swiss Federal Council is powerful but not supreme. Belgium has a parliamentary form of government with a cabinet dependent on the confidence of the legislature, as in the [36] three prototypes of the Westminster model. However, Belgian cabinets, largely because they are often broad and uncohesive coalitions, are not at all as dominant as their Westminster counterparts, and they tend to have a genuine give-and-take relationship with parliament. The fact that Belgian cabinets are often short-lived attests to their relatively weak position: from 1980 to 1995, for instance, there were six cabinets consisting of different multiparty coalitions-with an average cabinet life of only about two and a half years. 3. Multiparty system. Both Switzerland and Belgium have multiparty systems without any party that comes close to majority status. In the 1995 elections to the Swiss National Council, fifteen parties won seats, but the bulk of these seats-162 out of 200-were captured by the four major parties represented on the Federal Council. Switzerland may therefore be said to have a four-party system. Until the late 1960s, Belgium was characterized by a three party system consisting of two large parties-Christian Democrats and Socialists-and the medium-sized Liberals. Since then, however, these major parties have split along linguistic lines and several new linguistic parties have attained prominence, creating an extreme multiparty system: about a dozen parties have usually been able to win seats in the Chamber of Representatives, and nine of these have been important enough to be included in one or more cabinets. The emergence of multiparty systems in Switzerland and Belgium can be explained in terms of two factors. The first is that the two countries are plural societies, divided along several lines of cleavage. This multiplicity of cleavages is reflected in the multidimensional character of their party systems. In Switzerland, the religious cleavage divides the Christian Democrats, mainly supported by practicing Catholics, from the Social Democrats and Radicals, who draw most of their support from Catholics who rarely or never attend church and from Protestants. The [37] socioeconomic cleavage further divides the Social Democrats, backed mainly by the working class, from the Radical Democrats, who have more middle-class support. The Swiss People's party is especially strong among Protestant farmers. The third source of cleavage, language, does not cause much further division in the Swiss party system, although the Swiss People's party's support is mainly in German-speaking Switzerland, and the three large parties are relatively loose alliances of cantonal parties within which the linguistic cleavage is a significant differentiator (McRae 1983,111-14). Similarly, the religious cleavage in Catholic Belgium divides the Christian Social parties, representing the more faithful Catholics, from the Socialists and Liberals, representing rarely practicing or non-practicing Catholics. The Socialists and Liberals ,are divided from each other by class differences. In contrast with Switzerland, the linguistic cleavage in Belgium has caused further splits both by dividing the above three groupings, which used to be Belgium's three dominant parties, into separate and smaller Dutch-speaking and French-speaking parties and by creating several additional small linguistic parties (McRae 1986, 130-48). 4. Proportional representation. The second explanation for the emergence of multiparty systems in Switzerland and Belgium is that their proportional electoral systems have not inhibited the translation of societal cleavages into party-system cleavages. In contrast with the plurality method, which tends to overrepresent large parties and to underrepresent small parties, the basic aim of proportional representation (PR) is to divide the parliamentary seats among the parties in proportion to the votes they receive. The lower houses of both legislatures are elected byPR. 5. Interest group corporatism. There is some disagreement among experts on corporatism about the degree of corporatism in Switzerland and Belgium. mainly because the labor unions [38] in these two countries tend to be less well organized and less influential than business. This disagreement can be resolved, however, by distinguishing between two variants of corporatism: social corporatism in which the labor unions predominate and liberal corporatism in which business associations are the stronger force. Peter J. Katzenstein (1985, 105t 130) uses Switzerland and Belgium as two exemplars of the latter, and he concludes that Switzerland "most clearly typifies the traits characteristic of liberal corporatism." Both countries clearly show the three general elements of corporatism: tripartite concertation, relatively few and relatively large interest groups, and the prominence of peak associations. Gerhard Lehmbruch (1993, 52) writes that "the strength of Swiss peak associations is remarkable, and it is generally acknowledged that the cohesion of Swiss interest associations is superior to that of Swiss political parties." Moreover, Klaus Armingeon (1997) argues that, although the extent and effectiveness of corporatism in many European countries has been declining in the 1990s, it continues to be strong in Switzerland. 6. Federal and decentralized government. Switzerland is a federal state in which power is divided between the central government and the governments of twenty cantons and six socalled half-cantons, produced by splits in three formerly united cantons. The half-cantons have only one instead of two representatives in the Swiss federal chamber, the Council of States, and they carry only half the weight of the regular cantons in the voting on constitutional amendments; in most other respects, however, their status is equal to that of the full cantons. Switzerland is also one of the world 's most decentralized states. Belgium was a unitary and centralized state for a long time, but from 1970 on it gradually moved in the direction of both decentralization and federalism; in 1993, it formally became a federal state. The form of federalism adopted by Belgium is a "unique federalism" (Fitzmaurice 1996) and one of "Byzantine [39] complexity" (McRae 1997,289), because it consists of three geographically defined regions-Flanders, Wallonia, and the bilingual capital of Brussels-and three nongeographically defined cultural communities-the large Flemish and French communities and the much smaller German-speaking community. The main reason for the construction of this two-Iayer system was that the bilingual area of Brussels has a large majority of Frenchspeakers, but that it is surrounded by Dutch-speaking Flanders. There is a considerable overlap between regions and communities, but they do not match exactly. Each has its own legislature and executive, except that in Flanders the government of the Flemish community also serves as the government of the Flemish region. 7. Strong bicameralism. The principal justification for instituting a bicameral instead of a unicameral legislature is to give special representation to. minorities, including the smaller states in federal systems, in a second chamber or upper house. Two conditions have to be fulfilled if this minority representation is to be meaningful.: the upper house has to be elected on a different basis than the lower house, and it must have real power- ideally as much power as the lower house. Both of these conditions are met in the Swiss system: the National Council is. the lower house and represents the Swiss people, and the Council of States is the upper or federal chamber representing the cantons, with each canton having two representatives and each half-canton one representative. Hence the small cantons are much more strongly represented in the Council of States than in the National Council. Moreover, as Wolf Linder (1994, 47) writes, the "absolute equality" of the two chambers is a "sacrosanct rule" in Switzerland. The two Belgian chambers of parliament-the Chamber of Representatives and the Senate-had virtually equal powers in prefederal Belgium, but they were both proportionally constituted and hence very similar in composition. The new Senate, [40] elected for the first time in 1995, especially represents the two cultural-linguistic groups, but it is still largely proportionally constituted and not designed to provide overrepresentation for the French-speaking and German-speaking minorities.1 More over, its powers were reduced in comparison with the old Senate; for instance, it no longer has budgetary authority (Senelle 1996, 283). Hence the new federal legislature of Belgium exemplifies a relatively weak rather than strong bicameralism. 8. Constitutional rigidity. Both Belgium and Switzerland have a written constitution-a single document containing the basic rules of govemance-that can be changed only by special majorities. Amendments to the Swiss constitution require the approval in a referendum of not only a nationwide majority of the voters but also majorities in a majority of the cantons. The half-cantons are given half weight in the canton-by-canton calculation; this means that, for instance, a constitutional amend ment can be adopted by 13.5 cantons in favor and 12.5 against. The requirement of majority cantonal approval means that the populations of the smaller cantons and half-cantons, with less than 20 percent of the total Swiss population, can veto constitutional changes. In Belgium, there are two types of supermajorities. All con stitutional amendments require the approval of two-thirds majorities in both houses of the legislature. Moreover, laws pertaining to the organization and powers of the communities and regions have a semiconstitutional status and are even harder to adopt and to amend: in addition to the two-thirds majorities in [41] both houses, they require the approval of majorities within the Dutch-speaking group as well as within the French-speaking group in each of the houses. This rule gives the French-speakers an effective minority veto. 9. Judicial review. Switzerland deviates in one respect from the pure consensus model: its supreme court, the Federal Thibunal, does not have the right of judicial review. A popular imtiative that tried to introduce it was decisively rejected in a 1939 referendum (Codding 1961,112).2 There was no judicial review in Belgium either until 1984, when the new Court of Arbitration was inaugurated. The court's original main responsibility was the interpretation of the constitutional provisions concemiIig the separation of powers between the central, community, and regional governments. Its authority was greatly expanded by the constitutional revision of 1988, and the Court of Arbitration can now be regarded as a genuine constitutional court (Alen and Ergec 1994,20-22; Verougstraete 1992,95). 10. Central bank independence. Switzerland's central bank has long been regarded as one of the strongest and most independent central banks, together with the German Bundesbank and the Federal Reserve System in the United States. In contrast, the National Bank of Belgium was long one of the weakest central banks. However, its autonomy was substantially reinforced in the early 1990st roughly at the same time as the transition to a federal system, but mainly as a result of the Maastricht 1reaty, signed in 1992 and ratified in 1993, which obligated the EU member states to enhance the independence of their central banks. Robert Senelle (1996, 279) concludes that the Belgian central bank now enjoys a "high degree of autonomy. . . in the conduct of its monetary policy."[42] The Consensus Model in the European Union The principal institutions of the European Union do not fit the classification into executivet legislative. judicial. and monetary organs as easily as those of the five sovereign states discussed so far. This is especially true for the European Council, which consists of the heads of government of the fifteen member states, meeting at least twice a year; it can exert great political influence, and most of the major steps in the development of the European Community and, since 1993, of the European Union have been initiated by the European Council. Of the other institutions, the European Commission serves as the executive of the EU and can be compared to a cabinet; the European Parliament is the lower house of the legislature; and the Council of the European Union can be regarded as the upper house. The responsibilities of the European Court of Justice and the European Central Bank are clear from their names. 1. Executive power-sharing in broad coalition cabinets. The European Commission consists of twenty members, each with a specific ministerial responsibility. appointed by the governments of the member states. The five largest states-Germany, the United Kingdom, France, Italy, and Spain-appoint two commissioners apiece, and each of the other ten members appoints one commissioner. Because all fifteen nations that belong to the EU are represented on it. the Commission is a broad and permanent internation coalition. In practice, the Commission is also a coalition that unites the left, center. and right of the political spectrum in Europe. A telling example is that, in the mid 1990s. the two British commissioners were Conservative Leon Brittan and former Labour party leader Neil Kinnock-politicians who are extremely unlikely ever to serve together in a British cabinet. 2. Executive-legislative balance of power. After each five yearly parliamentary election, the new European Commission must be approved by a vote in the European Parliament. Parlia- [43] ment also has the power to dismiss the Commission, but only by a two-thirds majority. Parliament has strong budgetary powers, but although its other legislative powers were enhanced by the Amsterdam Treaty of 1997, they remain relatively weak. In com- , parison with the Commission, the Parliament's role appears to be subordinate. This judgment of executive-Iegislative relationships changes. however, when we add the Council of the European Union-composed of ministers from the goveniments of the fifteen member states-to the picture. George Tsebelis and Jeannette Money (1997, 180) call the Council "the European equivalent of [an] upper house." The Council is also clearly the strongest of the three institutions. Overall, therefore, the Commission is much more like the equal partner in the consensus model than the dominant cabinet in the Westminster model. 3. Multiparty system. The 626-member European Parliament had eight officially recognized parties (comprising the minimum of 18 members required for recognition) in 1996. The largest of these was the Party of European Socialists with about 34 percent of the seats in Parliament-far short of a parliamentary majority. The next largest was the European People's party (mainly Christian Democrats) with about 29 percent of the seats. None of the other parties held more than 10 percent of the seats. The political fragmentation is even greater than appears from this multiparty pattern because the parties in the European Parliament are considerably less cohesive and disciplined than the parties in the national parliaments. The partisan composition of the "upper house," the Council of the European Union, changes as the cabinets of the member countries change, and it also depends on the subject matter being discussed, which determines which particular minister will attend a particular session; for instance, if farm policies are on the Council's agenda, the national ministers of agriculture are likely to attend. In practice, however, the Council is also a multiparty body. 4. Proportional representation. The European Parliament has been directly elected since 1979. It is supposed to be elected [44] in each country according to a uniform electoral system, but the member countries have not been able to agree on such a system. Nevertheless, the prevalent method is some variant of PR, and PR is used in all of the member countries and in Northern Ireland. The only exception has been the election by plurality of the British representatives from the United Kingdom, but in 1997 the new Labour cabinet decided that the 1999 European Parliament elections in the United Kingdom would be entirely by PR Even then, however, there will still be a significant degree of disproportionality as a result of the overrepresentation of the small states and the underrepresentation of the large states in the European Parliament. At the extremes, Germany has ninety-nine and Luxembourg six representatives in the European Parliament, even though Germany's population is about two hundred times larger than Luxembourg's. In this respect, the European Parliament combines in one legislative chamber the principles of proportional representation and of equal national representation that, for instance, in Switzerland are embodied in two separate houses of the legislature. 5. Interest group corporatism. The EU has not yet developed a full-fledged corporatism, largely because the most important socioeconomic decisions are still made at the national level or subject to national vetoes. As the EU becomes more integrated, the degree of corporatism is bound to increase. In the title of Michael Gorges's (1996) book Euro-Corporatism? the question mark is deliberate, and Gorges answers the question mainly in the negative for the present situation, but he also sees significant corporatist elements in certain sectors as well as a clear trend toward greater corporatism. One important factor is that the European Commission has long favored a corporatist mode of negotiating with interest groups. For instance, it sponsored a series of tripartite conferences during the 1970s, and although these did not lead to the institutionalization of tripartite bargaining, "the Commission never abandoned its goal of promoting a dialogue between the social partners and of improving their partici- [45] pation in the Community's decision-making process" (Gorges 1996, 139). Another indication of the EU's inclination toward corporatism is that one of its formal institutions is the advisory Economic and Social Committee, which consists of interest group representatives appointed by the member governments. 6. Federal and decentralized government. Compared with other international organizations, the supranational EU is highly unified and centralized, but compared with national states even as decentralized a state as Switzerland-the EU is obviously still more "confederal" than federal as well as extremely decentralized. 7. Strong bicameralism. The two criteria of strong bicameralism are that the two houses of a legislature be equal in strength and different in composition. The EU's legislature fits the second criterion without difficulty: the Council has equal representation of the member countries and consists of representatives of the national governments, whereas the Parliament is directly elected by the voters and the national delegations are weighted by population size. In national legislatures, deviations from equal power tend to be to the advantage of the lower house. In the ,EU it is the other way around: the upper house (Council) is considerably more powerful than the lower house (Parliament)-not fully in accordance with the consensus model, but even less with the majoritarian model.3 8. Constitutional rigidity. The EU's "constitution" consists of the founding Treaty of the European Economic Community, signed in Rome in 1957, and a series of both earlier and subsequent treaties. Because these are international treaties, they can be changed only with the consent of all of the signatories. Hence they are extremely rigid. In addition, most important decisions in the Council require unanimity; on less important matters, it has become more common since the 1980s to make decisions by [46] "qualified majority voting,,' that is, by roughly two-thirds majorities and by means of a weighted voting system (similar to the weighted allocation of seats in the European Parliament). 9. Judicial review. A key EU institution is the European Court of Justice. The Court has the right of judicial review and can declare both EU laws and national laws unconstitutional if they violate the various EU treaties. Moreover, the Court's approach to its judicial tasks has been creative and activist. Martin Shapiro and Alec Stone (1994,408) write that "clearly the two most politically influential constitutional courts in Europe are those of Germany and the Community [EU]. . . . There are few instances as observable and as important as the ECJ [European Court of Justice] case of a court building itself as a political institution, and building the whole set of institutions of which it is a part." 10. Central bank independence. The European Central Bank, which started operating in 1998, was designed to be a highly independent central bank; indeed, the Economist (November 8, 1997) wrote that "its constitution makes it the most independent central bank in the world. " However, its independence was compromised to some extent when the first bank president was appointed in 1998. In order to maximize the president's authority, the appointment is formally for an eight-year term, but the first president had to pledge to resign well before the end of his term, probably after about four years, as part of a political deal between France, which had insisted on its own candidate, and the other EU members. In the beginning of this chapter, I emphasized that the majoritarian model was incompatible with the needs of deeply divided, plural societies. The EU is clearly such a plural society: "Deep-seated and long-standing national differences, of which language is only one, have not and will not disappear in Europe" (Kirchner 1994, 263). Hence it is not surprising that the EU's institutions conform largely to the consensus instead of the majoritarian model. Many observers predict that the EU will even- [47] tually become a federal state, especially as a result of the adoption of a common currency. For instance, Martin Feldstein (1997, 60) asserts that the "fundamental long-term effect of adopting a single currency [will] be the creation of a political union, a European federal state with responsibility for a Europe-wide foreign and security policy as well as for what are now domestic economic and social policies." If and when the EU develops into a sovereign European state, its institutions are likely to change the European Parliament, for instance, will probably ,become a more powerful legislative chamber-but it is not likely to stray far from the consensus model, and it is almost certain to take the form of a federal United States of Europe. Footnotes: for Chapter 3 1. Most senators-forty out of seventy-one-are directly elected from two multimember districts that are partly defined in nongeographical terms-one comprising Flanders and Dutch-speakers in Brussels and the other Wallonia and French-speaking Bruxellois. The remaining thirty-one senators are indirectly elected or coopted in different ways. The overall linguistic composition is: forty-one Dutch-speakers, twenty-nine French-speakers. and one German-speaker. A further curious provision is that any adult children of the king are "senators by right" 2. National laws can, however, be challenged in a different manner: if, within ninety days of the passage of a law, a minimum of fifty thousand citizens demand a referendum on it, a majority of Swiss voters can reject it. 3. Another notable example of at least a slight asymmetry favoring the house is the U.S. Congress in which the Senate has special powers over treaties and appointments.