Resume of Patterns of Democracy
Cours de Politique comparée
Professeur H-P Kriesi
Sémestre d'éte 2000
Jorge E. Viñuales
Dimensions allowing to compare and catalogue systems as majoritarian or consensual
1. Concentration of executive power in single-party majority cabinets VS coalition cabinet or executive power sharing
1. Centralized and unitary government VS decentralized and federal government
2. Executive-Legislative relationships dominated by the executive VS Executive-Legislative balance of powers
2. Unicameralism VS Bicameralism (two equally strong chambers differently constituted)
3. Two-party system VS Multiparty system
3. Flexible constitution VS rigid constitution
4. Disproportional and majoritarian electoral system VS proportional representation
4. Constitutionality control by the Legislature VS Judicial Review
5. Pluralist and competitive interest group system VS organized and corporatist interest group system tending to concertation
5. Central Bank depending on the executive power VS Independent Central Bank
Chapter 2: The Majoritarian or Westminster model
Ex: British system, New Zealand's system, Barbados' system
1. Concentration of executive VS executive power sharing
In the UK cabinets are usually composed of the members of the parti that has won the majority of parliamentary seats. Exceptions (either coalition cabinets or minority cabinets) occurred from 1918 to 1945, for instance the coalition cabinet during the war. Only two minority cabinets since 1945.
1. Centralized-unitary VS decentralized-federal government
Local governments in the UK perform a series of important functions, however, they are the creatures of the central government and their powers are not constitutionally guaranteed. Two exceptions should be noted : Northern Ireland was consistently autonomous from 1921 to 1972, and for about a month more recently ; Scotland and Wales since 1997.
2. Executive dominanceVS Executive-Legislative balance of powers
Although the vote of confidence, the cabinet is clearly dominant. Indeed, cabinets are composed of the leaders of a cohesive majority party in the house of commons, and therefore it is normally backed by the majority in the house of commons. This is why, strong cabinet leadership depends on the cohesiveness of the majority party and on majority support in the house of commons. It is the disciplined two-party system that gives rise to executive dominance.
2. Unicameralism VS Bicameralism
The UK deviates from the pure majoritarian model in that it has to chambers. However, this Bicameralism is clearly assymetrical and the house of Lords hold almost no power.
3. Two-party system VS Multiparty system
There are two large parties : Conservative party and Labour party. After the scission of the Labour party resulting in the creation of the Social Democratic party, Liberals, merged with socio-democrats have become a considerable third force. Liberals have been one of the two big parties until the Labour party replaced them in the interwar years.
Generally speaking, the two parties only differ in socioeconomic issues (which is a typical trait of two-party systems)
3. Flexible constitution VS rigid constitution
Britain has an unwritten constitution in the sense 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 (Magna Carta of 1215, Bill of Rights 1689 and the Parliament Acts of 1911 and 1949), common law principles, customs, and conventions. This has two important implications : it makes the constitution completely flexible because it can be changed by parliament by regular majorities ; the absence of judicial review
4. Disproportional and majoritarian electoral system VS PR.
Members of the House of Commons are elected in single-member districts according to the plurality method "first past the post". This system tends to produce highly disproportional results that lead to "manufactured majorities". The disproportional electoral system has been particularly disadvantageous for the Liberals and the Liberal Democrats.
4. Constitutionality control by the Legislature VS Judicial Review
There is no written constitutional document with the status of a higher law against which the courts can test the constitutionality of ordinary legislation. Parliament is in charge of changing and interpreting the constitution, therefore Parliament (the parliament majority) can be said to be the ultimate or sovereign autority. However, this parliamentary sovereignty is bound by European Community's laws and institution in several areas of policy. Britain is thus under the European Court of Justice. It is also under the competence of the European Court of Human Rights since 1966.
5. Pluralist interest group system VS corporatist interest group system
The system is clearly pluralist, which means that there is a multiplicity of competitive and uncoordinated interest groups exerting pressure on the government
5. Dependent VS Independent Central Bank
The Bank of England has not been able to act independently and has instead been under the control of the cabinet. However, in 1997, the Bank was given the independence to set interest rates.
Chapter 3: The Consensus model
Ex: Switzerland, Belgium, European Union
1.Concentration of executive VS executive power sharing
There is a clear executive power sharing in Switzerland (la formule magique)
1. Centralized-unitary VS decentralized-federal government
Switzerland is a federal state in which power is constitutionally divided between central government and the governments of 20 Cantons and 6 so-called half-Cantons. The half Cantons have only one instead of two representatives in the Swiss Federal Chamber and they carry only half the weigh in constitutional amendment procedures. In most of the other respects there is complete equality among Cantons.
2. Executive dominance VS Executive-Legislative balance of powers
There is separation of powers although the Federal Council is not popularly elected. However, once the election is done, each Federal Councilor stays in power for a fixed term. There is a balanced relationship between executive and legislative.
2. Unicameralism VS Bicameralism
The principal justification for instituing a bicameral instead of a unicameral legislature is to give special representation to minorities, including the smaller states in federal systems. This is the case of Switzerland, where the upper house is elected on a different base than the lower house, and it has real power (the absolute equality of the two chambers is a sacrosanct rule in Switzerland)
3. Two-party system VS Multiparty system
Multiparty system without any party coming close to a majority. The emergence of a multiparty system in Switzerland can be explained in terms of the plural character of the society and the proportional electoral system
3. Flexible constitution VS rigid constitution
Swiss Constitution is rigid. It needs the double majority of the people and the Cantons to introduce a constitutional amendment.
4. Disproportional and majoritarian electoral system VS PR.
The PR system has not inhibit the translation of societal cleavages (linguistic, religious, socioeconomic, etc.) into party-system cleavages. The National Council is elected by PR.
4. Constitutionality control by the Legislature VS Judicial Review
It is a particularity of Switzerland that it has no judicial Review. The Federal Tribunal is not entitled to review the constitutionality of ordinary laws (because they are adopted directly by the people as an organ of the direct democracy)
5. Pluralist interest group system VS corporatist interest group system
In Switzerland predominates liberal corporatism.
5. Dependent VS Independent Central Bank
The Swiss Central Bank is one of the most independent Central Banks in the World, together with the German Bundesbank and the Federal Reserve of the US.
Chapter 4: Thirty-six democracies
I. Definition of Democracy (According to R. Dahl in Polyarchy)
(a) The right to vote
(b) The right to be elected
(c) The right of political leaders to compete for support and votes
(d) Elections that are free and fair
(e) Freedom of association
(f) Freedom of expression
(g) Alternative sources of information
(h) Institution for making public policies depend on votes and other expressions of preferences
II. Delimitation of the object
(a) 36 democracies of at least 250.000 habitants.
(b) Starting point: the first democratic election after 1945 (included) or after independence.
(c) Democracies sufficiently consolidated with a life-span of at least 19 years
(d) Some precision about Switzerland (where women were allowed to vote only in 1971), USA (where the universal suffrage was not firmly established until the Voting Rights Act in 1965), most colonial countries such as France, UK, Belgium, the Netherlands (under the rule of whom the universal suffrage was also violated), Israel (that after 1967 keep control of occupied territories imposing special rules).
III. The set of democracies chosen includes representatives of the three democratization waves identified by S. Huntington on the base of the establishment of the universal suffrage (defined as the right to vote of at least 50% of the male population)
(b) Reverse wave (after the WWI)
(c) 1943-1962 (resulting of decolonization)
(d) Reverse wave (especially in Latin America)
(e) 1974-on (starting point: the end of the Portuguese dictatorial regime. Main area: Latin America and Eastern Europe)
Although the analyzed period starts in 1945, democracies analyzed from the 1940's belonged all to Huntington's first wave (with the only exception of Israel). However, about 15 of them suffered from the first Reverse wave.
VI. Countries whose governments have been continuously democratic since the 1940's have several common traits:
(a) Although pluralist, they are rather homogenous
(b) They are all economically developed, industrialized and urbanized
(c) They all belong to the Western Judeo-Christian world (except for Japan)
(d) They are almost all geographically concentrated in the North Atlantic area
Second and third waves added a great deal of diversity to the object studied. This diversities can be grouped into 3 different variables (which importance derive from their expected influence on the type of democracy adopted and it performance. We must however note that the correlation of one another is rather weak):
(a) The degree of pluralism (which is a composite measure including: ethnic divisions, religious cleavages as well as inter religious cleavages, linguistic differences, and the intensity of these cleavages) allows to divide democracies into three groups: plural, semiplural, non-plural. This measure should receive three considerations: all but one plural societies are linguistically divided countries ; the classification might also be taken as reflecting the situation of a longer time span ; non-plural is not the same as homogenous because most of non-plural societies are religiously divided to at least some extent and most contain at least one or more small minorities
(b) The level of socioeconomic development based on the human development index (which is a composite measure including income, life expectancy and educational attainment)
(c) The greatest difference is in the population size.
A) The executive-parties dimension
Chapter 5: Party Systems (two-party VS multiparty systems)
The classification of cabinets (single-party VS multiparty coalition cabinets / bare majority VS oversized majority VS minority cabinets) heavily depends on the Party System.
I. The typical system matching majoritarian democracies is the Two-party system (in opposition to multiparty systems which are typical of consensual democracies). This system claims several advantages:
(a) It allows a clear choice because there are only two parties
(b) It has a moderating influence because the two main parties have to compete for center swing voters, which means that they have to advocate center policies (it has been criticized that these two claims are contradictory since advocating rather similar policies makes the choice less clear)
(c) It is necessary to form single-party majority cabinets that will be stable and effective policy makers (although the assertion of a two-party system leading to a stable single-party cabinet is indeed confirmed by practice, policy success is not assured at all)
II. How to count the parties relevant to define a system? We must take into account the number of parties as well as their relative strength (J. Blondel). Laakso and Taagerpera have suggested a very useful and precise method called the "effective number of parties" and consisting of:
N = 1 / å S2i
(N is the number of parties and S the proportion of seats of the i-th party in the parliament)
III. Closely allied parties. How they should be treated? Four criteria are relevant to decide whether they should be consider as one or as two parties:
(a) Do they compete for votes in elections?
(b) Do they form a single parliamentary group?
(c) Are they either in the cabinet together or in the opposition together?
(d) Do they have a long standing collaboration?
If these four criteria do not resolve the matter we should split the difference, which means that we take calculate two "effective number of parties", counting the party as only one and then as two, and using these two numbers we obtain a meaninful average.
VI. Factionnalized parties should undergo the same procedure.
VII. The Party systems of 36 democracies
VIII. The issue-dimension of Partisan conflict: there is a very strong correlation between the number of issue-dimensions and the degree of multiplicity of the party system. This is for two reasons: societal groups in conflict need representation ; internal divisions of parties forming the two-party system are not taken as canalizing this need. Following are the issue dimension order by their importance:
(a) Socioeconomic dimension
(b) Religious dimension
(c) Cultural and ethnic dimension
(d) Urban dimension
(e) Regime support dimension
(f) Foreign policy dimension
(g) Materialist VS post-materialist dimension
Chapter 6: Cabinets (concentration VS sharing of executive power)
The concentration or the share of executive power is the more characteristic variable allowing to differentiate majoritarian and consensual democracies.
I. Cabinets can be classified according to two criteria: whether the cabinet is formed only by one party or by a coalition ; the kind of parliamentary support
- typical single-party majority cabinet
- presidential cabinets might be assimilated to this form
- American tokenism (although technically a coalition, it is substantially at least a minimal majority)
- Majority governments in disguised, Strom (when a minority cabinet have received clear support from other parties in the parliament that do not want to form the cabinet)
WHEN THERE'S NO MAJORITY
- minimal winning coalition theory
- minimum size coalitions
- coalitions with the smaller number of parties
- minimal range coalitions
- minimal connected winning coalitions
- (policy viable coalitions?)
- real minorities (single party) cabinet that is forced to negociate constantly to stay in office
- broad coalitions of Austria
If one party has a majority of the parliamentary seats it is likely to form a single-party cabinet or, which is less likely but also possible, it will form a coalition with one or more minority parties or with the other large party (ex: the Churchill Cabinet during the WWII)
If no party has a parliamentary majority, it is likely that a coalition cabinet will be formed.
But which is the most likely coalition? There are different theories:
(a) Minimal winning coalition theory: Political parties want to maximize their power, which means holding as many of the cabinet positions as possible. A party will team up with one or more parties until it reach majority, but it will resist unnecessary parties that could reduce its share of ministers. This theory predicts always more than one outcome (unless there's a majority party). There are alternative theories that precise the most probable outcome of this theory.
(b) Minimum size coalitions: A party that wants to maximize its share of power will chose (to form the coalition) the party providing the narrowest parliamentary majority
(c) Coalitions with the smaller number of parties: because bargaining with many parties is difficult, a party will choose the smaller number possible of allies to form the minimal majority (even if this choice implicate a larger minimal majority)
(d) Minimal range coalitions(policy based theory): if parties are separated from left to right, each position being a space, parties will choose to ally with like-minded partners, which means that the coalitions formed will tend to be those with the least number of spaces between allies (even if this choice implicate a larger minimal majority)
(e) Minimal connected winning coalitions(policy based theory): parties will form coalitions following two criteria - they must be adjacent (party A cannot ally with C without the inter-phase of B) - unnecessary parties are rejected
(f) Policy viable coalitions(policy based theory): What really matters is the pivotal party in the parliament.
These theories, however useful, do not result in good predictions. Indeed, a big part of the cabinets formed are in contradiction with the majoritarian assumption of minimal winning majority (119 out of 196 European cabinets are not minimal winning coalitions but minority or oversized cabinets).
II. Factors that explain the formation of undersized cabinets
(a) Time: a period in the opposition offer certain parties the opportunity of electoral gains for future enhanced cabinet participation
(b) Institutional factors: - the "constructive vote of no confidence" may maintain a minority cabinet in office if parties in the parliament are not able to reach an accord on the cabinet formula to come ; - right of the cabinet to make its legislative proposals matters of confidence. Proposals are automatically considered as adopted unless an absolute majority of the French National Assembly votes to dismiss the cabinet (only the dismiss votes are counted)
(c) The strength of parliamentary committees that give parties the ability to influence policy without being part of the cabinet (decreasing the incentive to enter the cabinet)
III. Factors that explain the formation of oversized cabinets
(a) Policy based theories explain pressures tending to enlargement of coalitions. Parties are naturally more likely to ally with a like-minded partner which may lead to the integration of unnecessary parties (invited by the original members of the coalition in order to equilibrate the balance and the tendencies of the coalition)
(b) An overriding objective of all parties (internal or external threats)
(c) Uncertainty about the loyalty of a coalition member or the stability of the coalition itself might lead to enlarged coalitions as insurance against defection
(d) Institutional factors: - prescription of linguistic balance in Belgium ; - special majorities for the adoption of certain acts, ex: constitutional amendments or economic legislation, might lead parties to form a coalition in order to obtain the necessary majority in Parliament (especially when the reforms that are sought and that need the special majority have been included in the electoral campaign)
VI. Presidential cabinets have special characteristics: - they do not depend on parliamentary support to stay in office (so from this viewpoint they can be assimilated to a minimal winning majority); - they need support for proposal laws (so from this viewpoint they can be assimilated to a minimal winning, an oversized or a minority depending on the party affiliations of the president and the ministers, and on the size of their parties in the legislature)
VII. Study of 36 democracies
VIII. Correlation between type of cabinet and Party system: As the effective number of parliamentary parties increases, the incidence of single-party minimal winning cabinets decreases. Prime minister power is also correlated with majoritarian cabinets
Chapter 7: Executive-Legislative Relations (Patterns of Dominance and Balance of Power)
While majoritarian democracies mainly are characterized by a dominant executive power, consensual democracies are characterized by a more balanced relationship between executive and legislative powers.
I. Governments can be classified roughly into two types, parliamentary and presidential, according to three criteria:
HEAD OF GOVERN. (Checks and balances)
Whatever his name is (prime minister, etc.) he and his cabinet are responsible to the legislature, which mean that they are dependent on the legislature's confidence and can be dismissed from office by a vote of no confidence or censure.
His name is president and he is elected for a constitutionally fixed period of time. In normal circumstances he cannot be removed from office (in special circumstances he can be removed by the procedure of impeachment). He usually has a power of veto to reject parliamentary legal proposals.
FORM OF ELECTION (Legitimacy)
Selected by legislature in a variety of forms:
Negotiation between main parties in parliament, formal election by the lower house, appointed by the king or sometimes the president.
Popularly elected, either directly or via a popularly elected presidential electoral college
IND. OR COLLEGIAL
Collegial executive. Prime minister's position in cabinet might vary from preeminence to virtual equality, but since the most important decision usually have to be made by the cabinet as a hole there is always a high degree of collegiality in the decision-making process
Individual. The president elects the cabinet members (following certain procedures that usually includes some kind of parliament participation) that are only advisors and are not formally need for the decision-making process which is entirely in the president's hands
Among the 36 democracies, some governments are difficult to classify: Switzerland has two aspects of parliamentary systems and only one of presidential (the Federal Council is a collegial organ, elected by the parliament, but it doesn't need the confidence of the parliament to stay in office) ; some semi-presidential governments such as Austria, Iceland, Irish, Portugal (these four can be for practical reasons be considered as ordinary parliamentary systems), France (depending on whether the president has or has not a majoritarian support in the parliament, the government can be considered as Presidential or Parliamentary respectively, and in practice this alternates), Finland (this system resembles the French one in its parliamentary phase and can therefore be considered as parliamentary), Israel (formerly Parliamentary but quite innovative since 1996 because prime minister is now popularly elected, for a fixed period of time, and with predominance over the cabinet member thanks to his popular legitimacy, these elements classify Israel's system as Presidential. Parliament and prime minister have the power to remove each other but the exercise of this power needs the removal of both)
We can note other not-so-essential contrasts between Parliamentary and Presidential systems: separation of powers (in Presidential systems a person cannot simultaneously serve in both executive and legislative posts while in Parliamentary they usually do it); power to dissolve (Presidents do not usually have it while Prime ministers do, but there are some exceptions like France); dual or single executives (Parliamentary systems usually have a dual executive, a head of State, for example a King with little or none political power and the head of government, on the other side Presidential systems concentrate both in one person. However this characteristic trait might and do vary from country to country).
II. Separation of Power and Balance of Power: the distinction between parliamentary and presidential systems is very important in many respects but albeit useful it is not a decisive indication of executive predominance (we might as well find an executive-legislative balance in Parliamentary systems such as Belgium as well as in Presidential like the US). The distribution of power in executive-legislative is not determinant to conclude on whether there's a predominant executive or a balance of power.
Presidential power derives from:
(a) Powers constitutionally defined: reactive power such as veto and proactive power such as the ability to legislate by decree in certain areas (these are very stable powers)
(b) Strength and cohesion of president's parties in the legislature (these relative powers might and does change abruptly and it is generally less stable than in parliamentary systems)
(c) Legitimacy derived from popular election (the powers derived from popular election vary according to the magnitude of the president's electoral victory)
III. How to measure Dominance and Balance of Power? For Parliamentary systems the best indicator is cabinets durability because a cabinet stays in power as long as it is dominant vis-à-vis the legislature. But how can we measure cabinet durability? There are two criteria (and two adjustments):
(a) To focus exclusively on the partisan composition of cabinets and to count a cabinet as one cabinet if its party composition does not change (large definition of cabinet duration)
(b) Taking into account four particular events: - changes in party composition - prime ministership - coalitional status - new elections (narrow definition of cabinet duration)
(c) Truncating any extreme value that surpasses the British Cabinet's value (which is considered the exemplar of cabinet dominance)
(d) Criterion (a) might lead to aberrant values, for example in the Swiss case (this case is arbitrarily given a low value because it is the exemplar case of executive-legislative balance).
The effective measure of 36 democracies shows a tendency of majoritarian countries (UK and most ex-colonies) to have predominant executives while consensual countries, especially those with constitutional separation of powers (presidential democracies and Switzerland but not France) correspond to an executive-legislative balance of power.
IV. Cabinet types and cabinet durability: How are cabinet types (single-party VS multiparty coalition cabinets / bare majority VS oversized majority VS minority cabinets) related to the degree of executive dominance (represented by cabinet durability). A positive relation between minimal winning and one-party cabinets on one hand and executive dominance on the other is expected because: - both belong to the same cluster - minority cabinets are at the mercy of (therefore they cannot be expected to dominate) parliaments - bare majorities have generally received solid support from their partisans in parliament while we can expect greater legislative independence when cabinets are oversized rather than minimal winning.
The overall pattern shown by the analysis of 36 democracies is that minimal winning one-party cabinets have the longest average life span. Both types of minimal winning (one-party and coalition) last longer than minority and oversized cabinets. These latter cabinets have very similar duration. Minority coalitions have the shortest life. (multiparty systems are usually temporary caretakers after a cabinet has fallen and while waiting a new election).
These results are translated into a clear correlation between minimal winning single-party cabinets and greater executive dominance. Presidential countries (USA, Costa Rica, Venezuela but not France) with minimal winning single-party cabinets havea much lower executive dominance than expected, this is due to the separation of powers.
V. Monarchs and Presidents: Different heads of State
(a) Monarchs have remained in power where the reigning family has been willing to withdraw from a political active role. Monarchs present some advantages (an apolitical and impartial head of state symbol of unity) and disadvantages (ex: the power to appoint the prime minister, which although it is essentially virtual, it may be concrete sometimes)
(b) Presidents as heads of state of a parliamentary democracy are much more likely to intrude on the powers of the head of government. Several methods are used to minimize this risk: presidents are not popularly elected (legitimated) ; Swiss system ; Botswana system (presidential title and function are given to the prime minister that is called president)
Chapter 8: Electoral Systems (Majority and Plurality Methods VS Proportional Representation)
The typical electoral system of majoritarian democracy (inspired of a majoritarian philosophy) is the single-member district plurality or majority system, and it is opposed to the typical system used in consensual democracies(inspired of both minority and majority representation conceptions): proportional representation (PR).
The main ideas opposed are over-representing or under-representing any parties VS to translate votes into seats proportionaly.
I. Classification and consequences of the electoral systems used in 36 democracies in terms of seven basic aspects of these systems:
(a) According to the electoral formula electoral systems can be classified as follows:
1. Plurality and Majority
1.1. Plurality: the first past the post, which means that the candidate who receives the most votes, whether a majority or a plurality, is elected (Many countries)
1.2. Mixed Majority-plurality: on the first ballot an absolute majority is required for election, but if no candidate wins a majority a plurality suffices on the second ballot, candidates failing to win a minimum percentage on the first ballot are barred from the second ballot (legislative elections in France)
1.3. Alternative vote: voters are asked to indicate their first preference, second preference, and so on among the candidates. If a candidate receives an absolute majority of the first preferences he is elected. If there are no such majority, the candidate with the lowest number or first preferences is dropped, and the ballots with this candidate as the first preference are transferred to the second preference. This procedure is repeated until a majority winner emerges (used in Australia and for presidential elections in Ireland)
2. Proportional Representation
2.1. List PR system: parties nominate lists of candidates in multimember districts, voters cast their ballots for one party list or another (they are sometimes allowed to split their votes among several lists), and seats are allocated to the party lists in proportion to the numbers of votes they have collected. List PR systems may be subdivided further according to the mathematical formula used to translate votes into seats, the most frequent is the Hondt formula. (used in many countries)
2.2. Mixed member proportional formula (MMP): each voter has two votes, one for a district candidate and one for a party list. The list PR seats compensate for any disproportionality produced by the district seat results. The exact degree of the overall results depends on how many list PR seats are available for the purpose of compensation
2.3. Single transferable vote (STV): voters vote for individual candidates instead of for party lists. The ballot contains the names of the candidates, and voters are asked to rank-order these. The procedure for determining the winning candidates is: first, any surplus votes not needed by candidates who already have the minimum quota of votes required for election are transferred to the next most preferred candidates on the ballots in question ; second, the weakest candidate is eliminated and his or her ballots are transferred in the same way. If necessary these steps are repeated until the available seats are filled (used for Senate elections in Australia)
3.1. Limited vote: voters casts their votes for individual candidates and those with the most votes win. But voters do not have as many votes as there are seats in the district and districts need to have at least two seats. (the more limited the number of votes each voter has, and the larger the number of seats at stake, the more the limited vote tends to deviate from plurality and the more it resembles PR
3.2. Single non transferable vote (SNTV): special case of limited vote where the number of votes cast by each voter is reduced to one
3.3. Parallel plurality: each voter has both a district vote and a PR vote. These features make it resemble MMP, but the crucial difference is that the PR seats are not compensatory.
(b) The magnitude of an electoral district is the number of candidates to be elected in the district (it must not be confused with the geographical extent or the population of a district). Majority-plurality systems can be applied both in single-member and multimember districts, while PR needs at least a two member district. District magnitude has a strong effect on the degree of disproportionality and on the number of parties. In two respects district magnitude is very important:
v It has a strong influence in both plurality-majority systems and PR (and SNTV) systems, but in opposite directions. Indeed, increasing the district magnitude in plurality-majority systems entails greater disproportionality and greater advantages for large parties (plurality: if A is slightly stronger than B in a particular area of 3 members, A will be likely to win the 3 seats, while if this area were divided into 3 areas of 1 member each, B would be able to win 1 seat. If the district magnitude is increased further the disproportionality also increases, this is why multimember districts have become rare.
v District magnitude has a strong impact on the degree of proportionality that the different PR systems attain (a party representing a 10% minority is unlikely to win a seat in a five-member district but will be successful in a ten-member district. Two member districts can therefore hardly be regarded as compatible with the principle of proportionality, conversely, a nation-wide district is optimal in terms of proportionality attainment.
(c) High-magnitude PR districts tend to maximize proportionality and to facilitate the representation of even very small parties. However, in order not to make it too easy for small parties to win election, all countries that use nation-wide districts have instituted minimum thresholds for representation (minimum number of seats or a minimum percentage of the total national vote). In fact, there are two types of barriers against small parties: explicit (or threshold) and implied (by district magnitude, the higher the magnitude the lower the barrier), and both are related:
T = 75% / M + 1
(T is the threshold and M the average district magnitude)
(d) Because electoral systems are methods of translating votes into seats, the number of seats available for this translation is clearly an integral part of the system of translation. This number is important for two reasons:
v The chances of a proportional allocation of seats improve with the size of the body to be elected or number of seats (if three parties win 43, 31 and 26 % of the national vote in a PR election and there are only 5 seats, the allocation will not be quite disproportional, but if we increase the number of seats to cover to 10, and then to 100 we will arrive to a proportionally perfect allocation)
v This is particularly important with respect to some countries where parliaments have relatively to the population too few members (these parliaments do not follow the general rule that states that parliament should have as many members as approximately the cube root of their population) which lead to a disproportionality.
(e) Indirect effect of Presidential systems on the effective number of parliamentary parties
Because only the largest parties are likely to win the presidency, these main parties have an advantage over smaller ones that tend to carry over into legislative elections. This tendency is especially strong when the presidential election is decided by plurality instead of majority-runoff (where small parties may want to try their luck in the first round) and when the legislative elections are held at the same time or shortly after the presidential elections.
(f) Malapportionment, which means that the districts have magnitudes that are not commensurate with their voting populations, also contribute to diproportionality. This is particularly difficult to avoid in plurality-majority systems with single-member districts, because equal apportionment requires that relatively many small districts be drawn with exactly equal electorates or populations
(g) Some list PR systems allow parties to have separate lists on the ballot but to formally "link" these lists, which means that their combined vote total will be used in the initial allocation of seats after what seats will be proportionally distributed according to the percentages of votes obtained by each party. This is called "apparentement" and because it helps smaller (usually under-represented) parties, it tends to reduce disproportionality and to increase the effective number of parties. The formation of these mutually beneficial interparty electoral links is also a logical consequence of some electoral systems, namely, the alternative vote, STV, French two-ballot system
II. How to measure the degree of disproportionality? Michael Gallagher suggests an index:
G = 1 / 2 å (vi - si)2
(differences between the vote percentages Vi and seat percentages Si)
III. Electoral disproportionality in Presidential Democracies. Presidential elections are inherently disproportional because there is only one seat to cover and therefore we need a pluralist or a majority formula. This is another reason (in addition to their inherent tendency to have majoritarian cabinets and their reductive effects on the number of parties) why presidential systems tend to be majoriatarian.
IV. Degrees of disproportionality in 36 democracies: There is a strikingly clear line dividing the PR parliamentary systems from the plurality and majority systems (even Spain with its low district magnitude PR and Greece with its reinforced majority PR and Japan with its semi-proportional system are still situated on the PR side). The four highest disproportionalities are Bahamas, Barbados, Mauritius and Jamaica (all small countries with plurality systems and unusually small legislatures). UK and NZ are still disproportional. The only exceptional cases of PR systems that are highly disproportional are presidential systems (Colombia, Costa Rica, Venezuela). Legislative disproportionality is low in the USA because of the primary elections (possibility for dissidents to try their luck in one of the major party primaries instead of establishing separate small parties). France has a high percentage of disproportionality because of its disproportional legislative election system in combination with presidentialism.
V. Electoral Systems and Party Systems: Plurality method favors two-party systems and PR and two-ballot systems encourage multipartism. There are two reason for this: - the mechanical effect of the plurality rule is that all but the two strongest parties are severely under-represented because they tend to lose in each district ; - psychological factors reinforces this because electors soon realize that their votes are wasted it they continue to give them to the third party and they transfer them to the less evil of the two adversaries, politicians are also less likely to engage in as third parties candidates and they seed a position in one of the two main parties (Duverger)
But whatever the electoral system might be: - all electoral systems tend to yield disproportional results ; - all electoral systems tend to reduce the effective number of parliamentary parties compared with the effective number of electoral parties ; - all electoral systems can manufacture a parliamentary majority for parties that have not received majority support from the voters (manufactured majorities mean a majority artificially created by the electoral system and that might be contrasted with earned majorities that happen when a party wins majorities of both votes and seats, and with natural minorities that occur when no party wins a majority of either votes or seats). Although all electoral systems support these 3 tendencies, plurality and majority give a much stronger support (especially to manufactured majorities) than PR systems (Rae).
There is therefore a strong correlation between disproportionality and plurality and majority electoral systems. There is also a negative correlation between the disproportionality of the electoral system and the effective number of parties (the more disproportional is an electoral system the fewer the effective number of parties), but the number of parties is also strongly affected by the issue-dimensions of each society.
Chapter 9 : Interest Groups (Pluralism Versus Corporatism)
I. There are two main types of interest group system:
a) Corporatism (Schmitter), which is compromised oriented system characterized by
- a small number of relatively large interest groups
- coordinated into national peak (hierarchical and monopolistic) organizations, especially those representing labor and management, (seeking concertation)
- this concertation means regular consultation by the leaders of peaks organizations, especially those representing labor and management, both with each other and with government representatives
- to arrive to comprehensive agreements that are binding (obligatory) on all three partners in the negotiations (so called tripartite pacts)
- Katzenstein adds another distinctive element which is an ideology of social partnership and the absence of a winner-take-all mentality
In fact Schmitter distinguishes two conceptually different meanings of Corporatism, the first refers to an interest group system in which groups are organized into national, specialized, hierarchical, and monopolistic peak organisations ; the second refers to the incorporation of interest groups into the process of policy formation (he calls it concertation). However, in practice, both tend to occur together because the first is a precondition for the second. That is why we can give a general definition using the aforementioned characteristics.
This system is typical of consensual democracies
b) Pluralism, which is characterized by
- a multiplicity of uncoordinated and competitive small interest groups
- absence or weakness of peak organizations
- little or no tripartite consultation
- absence of tripartite pacts
This system usually corresponds to majoritarian democracies.
II. Is Corporatism in decline?
During the 1970s the Corporatist system was regarded as having a high macroeconomic performance. It appeared to produce "a higher economic system". More recently research has tend to dissent from this interpretation and claim that corporatism is in decline, even in the most corporatist countries like Austria or Sweeden. What this "decline" means :
- that the efficacy of corporatist structures and the frequency of their use have decreased, and not that they have disappeared
- the decline is only and above all a matter of degree
- the proportion of decline is only rarely more than ten percent (Siaroff)
- the decline is an ambiguous concept that can be interpreted differently. H. Wiarda argues that corporatism, instead of declining, is simply developing into new areas (industrial phase of corporatist tripartite relationships is fading, new postindustrial issues such as education, health care, welfare, the environment and many others are coming to the fore and being negotiated in the familiar corporatist manner among the relevant groups and the government, this means that the policy process is still corporatist)
- economic globalization, which "limits governmental capacity to act effectively in economic policy" and therefore is claimed to be reason of the decline of corporatism, can also be invoked as a factor explaining the growth of corporatism (Katzenstein says that smaller european countries with open economies have been vulnerable to shifts in the world economy during the XX century and they have adopted corporatism as a protective device)
- the eroding level of integration of individuals with interest organizations and political parties weakens the ability of labor unions to act on behalf of large numbers of workers and hence also weakens their influence in tripartite negotiations. (Katzensteins' difference between liberal corporatism, in which business is the stronger force, and social corporatism, in which labor is dominant helps explain this situation. Indeed, the decline in the strength of labor unions does not necessarily mean an overall decline in corporatism but merely a shift from social to liberal corporatism
III. Siaroff has elaborated a comparative study that ranges many industrialized democracies according to their degree of pluralism (on the basis of a composite index which takes into account 8 factors including presence and strength of peak organization, process of concertation, the degree of centralization of wage bargaining, the strength and historical orientation-reformist vs revolutionary- of labor unions, levels of strikes and lockouts, etc. ). This study does not cover developing democracies. The reasons is that data is often not available and that, broadly speaking, developing countries tend to be more pluralist than corporatist because the organizational weakness of the relevant players, including interest groups and parties, makes tripartite concertation very difficult.
Lijhpart finds that in general terms the 36 democracies as a group are more pluralist than corporatist (an important reason for this is the presence of 12 developing countries).
IV. There is no clear causal connexion (although there is a correlation) that links the interest group system variable to the other four variables (which among them are clearly causally related. Indeed, electoral systems shape party systems, which in turn have a strong causal effect on the formation of cabinets, and types of cabinets are further causally related to cabinet duration). Therefore, the hypothesis that interest group systems are related to these other variables (belonging to the executive-parties dimension) rests entirely on the conceptual correspondence between the corporatism-pluralism distinction and the broad consensus-majoritarian difference.
There is however, a clear correlation between the interest group systems in the 36 democracies and their types of cabinets and party systems. Indeed as hypothesized, democracies that have more minimal winning one-party cabinets are also the countries that have more pluralist interest group systems (and vice-versa). In a similar way, democracies with a highly pluralist interest group system have small effective numbers of parliamentary parties (according to J. LaPalombara the deviant case of Italy is explained by the political coalitions which is an alternative way of achieving consensus. Strong interparty cooperation can therefore compensate for weaknesses in interest group coordination. However, this alternative way is very rare). Finally, the type of interest group system is also correlated with the electoral system and, though less strongly, with executive dominance.
B) The federal-unitary dimension
Chapter 10 : Division of Power (the federal-unitary dimension and centralized-decentralized contrasts)
I. The majoritarian model of democracy is characterized by the concentration of power in the hands of the majority. The consensus model is instead characterized by the non-concentration or dispersion of power whether it is dispersed to political actors operating together within the same political institution or dispersed to separate political institutions. In all democracies power is necessarily divided to some extent between central and non-central governments, but it is a highly one-sided division in majoritarian democracy. To maintain majority rule in the pure majoritarian model, the central government must control not only the central government apparatus but also all non-central, potentially competing, governments. Majoritarian is theresfore both unitary and centralized. On the other hand, the methods in consensus model are federalism and decentralization.
II. There are two sets of characteristics allowing to define Federalism. Federalism is firstly defined by :
- a guaranteed division of power between central and regional governments (Riker's definition of Federalism is "a political organization in which the activities of government are divided between regional governments and central government in such a way that each kind of government has some activities on which it makes final decisions". The component units here are regional governments so this definition does not account for Dahl's sociological federalism)
- the non-centralization of power (Elazar's definition of Federalism is "the fundamental distribution of power among multiple centers…, not the devolution of powers from a single center or down a pyramid")
These two main characteristics assume that the fundamental purpose of guaranteeing a division of power is to ensure that a substantial portion of power will be exercised at the regional level or, that the purpose of non-centralization of power is decentralization of power.
W e can classify the 36 democracies following a double criterion federalism-unitary and decentralized-centralized. We can use the constitutional arrangements as an expression of the will of each country (thus we have 5 categories : federal-decentralized, federal-centralized, semi-federal, unitary-decentralized and unitary-centralized). Concerning the chart of these 36 countries, two striking features should be noted : first, federalism is relatively rare (the mean score is 2.3 and the median 1.6 both much closer to the 1.0 score of unitary-centralized countries than to the 5.0 of federal-decentralized countries ; second, most federal systems are decentralized and most unitary systems are centralized which means that more than a half of the democracies can be classified in one of the two extreme categories (federal-decentralized or unitary-centralized)
Some deviant cases (for example Austria and India that are in the middle of federal-decentralized and federal-centralized) can be explained by the "President's rule" (right to dismiss state governments and to replace them with direct rule from the center with the purpose of dealing with grave emergencies) which is in practice used to remove state governments controlled by rival parties.
The semi-federal category includes the sociologically federal countries (Dahl) where central governments largely recognize, subsidize and delegate power to private associations with important semipublic functions, especially in such fields as education, culture and health care (ex: Israel, the Netherlands or Belgium). To test the degree of confidence of this index we can compare its results with those of other indicators (like the central government's share of a country's total tax receipts or the composite institutional autonomy index). Such comparison provides strong validation for the index of federalism undertaken.
III. Federalism tends to be used in two kinds of countries : relatively large countries and plural societies. In these plural societies federalism performs the function of giving autonomy to ethnic minorities. To analyze this function it is useful to distinguish between congruent federalism (territorial units with a social and cultural character that is similar in each of the units and in the federation as a whole) and incongruent federalism (units with social and cultural compositions that differ from one another and from the country as a whole). In incongruent federations social and political boundaries tend to coincide while in congruent federations they cut across each other. Incongruent federalism thus can make a plural society less plural by creating (grouping) relatively homogenous smaller areas. This is a recurrent pattern in federal systems that are also plural societies.
IV. As units in federal systems have their own constitutions (within certain limits set by the federal constitution) they can experiment with different forms of government or other institutions, what may be, if successful, beneficial both for other members of the federation and for the central government. In practice however, this experimentation is very rare. In the USA there has been experimentation about the electoral system.
Chapter 11 : Parliaments and Congresses (Concentration versus division of legislative power)
I. The pure majoritarian model call for the concentration of legislative power in a single chamber (typical example is the New Zealand's Parliament); the pure consensus model is characterized by a bicameral legislature in which power is divided equally between two differently constituted chambers (typical example is the Swiss Parliament). In practice, however, we find a variety of intermediate arrangements.
II. Unicameralism and Bicameralism : The most simple classification of Parliaments is the distinction between unicameral and bicameral (although there are some special cases like Norway and Iceland). We can note two features :
- Two thirds out of the 36 democracies studied have bicameral legislatures. The 13 countries with unicameral parliaments tend to be the smaller countries (Greece is the larger one with a population of about 10.000.000)
- The nine formally federal systems among the 36 democracies all have bicameral legislatures, whereas, as of 1996, the 27 formally unitary systems are evenly divided between unicameralism and bicameralism (Norway, 13 unicameral, and 13 bicameral)
III. Varieties of Bicameralism : the two chambers of bicameral legislatures tend to differ in mainly seven ways:
a) The original most important function of upper houses was to serve as a conservative brake on the more democratically (and progressive) elected lower houses. (this historical function has become obsolete, although Britain and Botswana still keep it at least theoretically)
The three of differences that follow, which can be considered as rather minor, do affect how the two chambers of the several legislatures operate, but they do not affect the question of whether a country's bicameralism is a truly strong and meaningful institution
b) Upper houses tend to be smaller than lower houses (with a wide range of variation and even some exceptions like Britain's House of Lords which in fact is only an apparent exception because if we exclude the members who rarely attend, especially many of the hereditary peers, the number is reduced to about 300)
c) Legislative terms of office tend to be longer in upper than in lower houses (Switzerland is the only relatively minor exception)
d) Upper houses are usually staggered elected. The modalities vary but this is a common feature (ex: One-third of the American and Indian upper houses is elected every second year while one-third of the French upper houses is renewed only every three years. Yet, Swiss upper house is selected in a staggered manner but at irregular intervals)
The three differences that follow are extremely important and crucial to determine the strength or weakness of bicameralism
e) Concerning the formal constitutional powers that the two chambers have, the general pattern is that the upper houses tend to be subordinated to lower houses. For instance, their negative votes on proposed legislation can frequently be overridden by the lower houses (exceptions to this pattern are for example Switzerland and the US).
f) The political importance of upper houses is also determined by the method of selection. Generally speaking, upper houses, unlike lower houses, are not popularly elected but indirectly elected or just appointed which mean that they have no democratic legitimacy (one important exception to this pattern are Switzerland and the US)
On the basis of these two criteria, bicameral legislatures can be classified as symmetrical (chambers have equal or only moderately unequal constitutional powers and democratic legitimacy) or asymmetrical (highly unequal in these respects). The German second chamber is an exceptional case because, although one of the strongest in the world, it does not owe its strength to either popular election or an absolute legislative veto but to the fact that it is composed of representatives (usually ministers) of the executives of the member states of the federation.
g) Upper houses may be elected by different methods or designed so as to overrepresent certain minorities. If this is the case, the two chambers differ in their composition and may be called incongruent. (for example in Switzerland there is equality of cantonal representation regardless of the cantons' populations. The US are another example of the same pattern. The German Bundesrat and the Canadian Senate are examples of federal chambers in which the component units are not equally represented but in which the smaller units are overrepresented and the larger ones underrepresented. The Austrian Bundesrat is a proper exception as its membership is roughly proportional to the population of the Länder. Several nonfederal upper houses must also be classified as incongruent. The French Senate is elected by an electoral college in which the small communes, with less than third of the population, have more than half of the votes. The Spanish Senate is incongruent for three reasons : the mainland provinces are equally represented ; most senators are elected by means of the semiproportional limited-vote system ; almost one-fifth are elected by the regional autonomous legislatures. Many other bicameral legislatures are congruent because their two chambers are elected by similar methods : list PR in Italy, the Netherlands, and prefederal Belgium etc.)
IV. Combining three distinctions or criteria (bicamerlism vs. unicameralism ; symmetrical vs. asymmetrical ; congruent vs. incongruent bicameralism) we can construct a classification (the bicameral-unicameral classification) of the 36 democracies in four principal categories : strong bicameralism (symmetric and incongruent), medium bicameralism (either symmetry or incongruency is missing), weak bicameralism (asymmetrical and congruent), unicameralism. Several countries can be placed in intermediate positions because they have changed their cameral structure during the period under consideration, in the case of Britain and Botswana upper houses are just relics of a predemocratic era, Iceland and Norway have one-a-half cameralism. Anyways, the mean (2.2) is well below the theoretical midpoint (2.5) between a strong bicameralism (4) and unicameralism (1).
V. There is a strong empirical relationship between the bicameral-unicameral and the federal-unitary classification (all formally federal systems have bicameral legislatures while some nonfederal systems have bicameral and other unicameral parliaments). This strong link is clearly shown by the positive correlation between federalism and bicameralism (as the degree of federalism and decentralization increases, first a shift from unicameralism to bicameralism takes place and then the strength of bicameralism increases. Deviant cases can be explained by the population size : small countries such as Austria, Finland, Denmark, Norway, Israel and Papua New Guinea tend to have unicameral or weakly bicameral legislatures in spite of their federal, semifederal or decentralized status while large countries such as Colombia, France and Italy have a relatively strong bicameralism in spite of their unitary and centralized systems. Population size is also strongly related to federalism).
Chapter 12 : Constitutions (Amendment Procedures and Judicial Review)
I. In the pure consensus model, the constitution is rigid and protected by judicial review, the pure majoritarian model is characterized by a flexible constitution and the absence of judicial review. Rigid constitutions tend to have more judicial review protection than more flexible constitutions.
II. Written and Unwritten Constitutions : this distinction is relatively unimportant for two reasons : almost all of the constitutions in the world are written (the absence of a written constitution is explained by a strong consensus on the basic norms Britain or, conversely, by the incapacity to reach such consensus Israel) ; it is more relevant to determine whether the constitution, written or unwritten, imposes significant restraints on the majority than to ask whether it is written or not.
There are, however, two counterarguments : if the written constitution is a single document, explicitly designated as the country's highest law, the parliamentary majority is likely to feel morally bound to respect it ; unwritten constitutions, because they do not have a formal status superior to that of other laws, logically entail both complete flexibility and the absence of judicial review.
III. Flexible and Rigid Constitutions : there are different devices that give constitutions different degrees of rigidity. These provisions can be reduced to four basic categories based on the type of majority required : ordinary majority (indicating complete flexibility it is typically majoritarian) ; approval by more than an ordinary majority but less than two-thirds majorities ; approval by two-thirds majorities ; approval by more than two-thirds majorities (supermajorities in plurality systems are clearly much less constraining than the same supermajorities in PR systems. As a consequence, plurality systems are classified one category below). Most countries fit the two middle categories.
IV. Judicial Review : If parliament itself is the judge of the constitutionality of its own laws, it can easily be tempted to resolve any doubts in its own favor, that is why it is necessary an independent body to judge this constitutionality. There are two remedies : to give this power to the courts (Judicial review is implied by the higher status of the constitution. We would have a decentralized judicial review). But this method presents the problem that such vital decisions as the conformity of law to the constitution should be made by the elected representatives of the people rather than by an appointed and frequently quite unrepresentative judicial body. So, as a compromise between these two logics (Judicial review is implied by the higher status of the constitution VS democratic logic) Kelsen proposed a centralized system of judicial review in which judicial review is the exclusive competence of a special body. The most common system is still the decentralize judicial review.
V. Judicial Review and Judicial Activism : The impact of judicial review depends only partly on its formal existence and much more vitally on the vigor and frequency of its use by the courts, especially supreme and constitutional courts. We can classify the different countries in four categories on the base of the existence or not of a judicial review and the degree of activism in the assertion of this power. There are only few countries where judicial review is very strong (US, Germany, India and since 1982 Canada). Almost half of the democracies are in the category of weak judicial review. However there appears to be a trend toward more and stronger judicial review : several countries moved to higher degrees of strength of judicial review and the four countries with written constitutions and no judicial review are older European democracies that have accepted the supranational judicial review of the European Court of Justice and/or the European Court of Human Rights). Another interesting point is that countries with centralized judicial review tend to have stronger judicial review than countries with decentralized systems (which might be explained by the fact that if a special body is created for the express and exclusive function of reviewing the constitutionality of legislation, it is very likely to carry out this task with some vigor)
VI. Constitutional Rigidity and Judicial Review : there are two reasons to expect that the variables of constitutional rigidity versus flexibility and the strength of judicial review will be correlated : both rigidity and judicial review are antimajoritarian device restricting the majority rule ; judicial review can work effectively only if it is backed up by constitutional rigidity and vice versa (if there is a strong judicial review but the constitution is flexible, the majority in the legislatures can easily respond to a declaration of unconstitutionality by amending the constitution. Similarly, if the constitution is rigid but not protected by judicial review, the parliamentary majority can interpret any constitutionally questionable law it wants to pass as simply not being in violation of the constitution). Indeed, the correlation is significant (0.39) although not particularly strong. A prominent outlier is Switzerland, that has no judicial review. Finally, both judicial review and rigid constitutions are linked with federalism as well as with the other two variables of the federal-unitary cluster : bicameralism and independent central banks.
VII. Referendums and Consensus Democracy: Referendum, which have usually been regarded as a majoritarian device is under certain circumstances clearly antimajoritarian. Indeed, the use of referendum in the process of constitutional amendment, as a requirement in addition to legislative approval, is more antimajoritarian than majoritarian, in particular if offers dissatisfied minorities the opportunity to launch a campaign against the proposed amendment. But there is an additional way that exists when referendums are combined with popular initiative (in Switzerland this gives even very small minorities a chance to challenge any laws passed by the majority of the elected representatives. Even if this effort does not succeed, it forces the majority to pay the cost of a referendum campaign. Hence the potential calling of a referendum by a minority is a strong stimulus for the majority to be heedful of minorities views. The referendum plus initiative has thus reinforced two Swiss traditions : la formule magique and the search for legislative majorities on particular bills that are as close to unanimity as possible)
Chapter 13 : Central Banks (Independence Versus Dependence)
I. When central banks are strong and independent they play a critical role in the policy process. Conversely, when central banks are dependent branches of the executive and hence relatively weak, this weakness is also a highly relevant attribute of the democratic system. Giving Central Banks independent power is yet another way of dividing power and fits the cluster of divided-power characteristics of the consensus model of democracy ; central banks that are subservient to the executive fit the concentrated-power logic of majoritarian democracy.
II. The duties and powers of central banks : the most important task is the making of monetary policy (regulation of interest rates and the supply of money). Monetary policy has a direct effect on price stability and the control of inflation, and it indirectly, but also very strongly, affects levels of unemployment, economic growth, and fluctuations in the business cycle. Other duties are managing the government's financial transactions ; financing the government's budget deficits buying government securities, making loans from their reserves, or printing money ; financing development projects ; regulating and supervising commercial banks. These other tasks may conflict with the task of controlling inflation, and the power of central banks over monetary policy can therefore be enhanced by not giving them these additional duties. Central banks and their role in monetary policy have become especially important since 1971 when President Nixon devalued the US dollar (breaking the fixed link of the dollar to gold and of nondollar currencies to the dollar, fashioned in the Bretton Woods agreement of 1944). In the more uncertain situation of floating exchange rates, central bank independence became an even more important tool to limit price instability. Central bank autonomy in most countries, however, was not enhanced until after 1990. There are two main reasons for this change : the Maastricht treaty requires central bank autonomy as a condition for participating in the single European currency ; the globalization of finance, which makes it important for developing countries to "signal their creditworthiness" to international investors.
III. Measuring independence of Central Banks : on the basis of bank charters or statutes Cukierman, Webb, and Neyapti have developed and index measuring central bank independence. This index includes four clusters of variables : the appointment and tenure of the bank's governor (the highest or more independent ratings are given to a governor whose term of office is eight years or longer, who cannot be dismissed, and who may not simultaneously hold other offices in government) ; policy formulation (the highest ratings go to banks that have exclusive responsibility to formulate monetary policy and play an active role in the government budgetary process) ; central bank objectives (the highest rating is accorded when price stability is the major or only objective in the charter, and the central bank has the final word in case of conflict with other government objectives. Medium ratings are given when price stability is one goal together with other compatible objectives, such as a stable banking system, and slightly lower, when price stability is one goal, with potentially conflicting objectives such as full employment) ; limitations on lending (they are independent when they are allowed to lend only to the central government and when they fully control the terms of lending). We average this measure with the index of political and economic independence of central banks by Grilli, Masciandaro and Tabellini, and with the rate of turnover in the governorship of the central bank (the higher this index the less independent the central bank. This measure is particularly accurate for less developed countries).
Most countries are concentrated in the lower half of the empirical range.
IV. Federalism and Central Bank Independence : Although it has been argued that central bank independence is mainly a function of the time horizons of the politicians who are in power or of the existence and influence of corporatist institutions, these hypothesis are not confirmed by empirical data. On the contrary, there appears to be a strong link between central bank independence and federalism (0.57). The five central banks with the greatest independence all operate in federal systems (Germany, Switzerland, US, Austria and Canada).
C) General findings of the book
Chapter 14 : The two-dimensional conceptual map of democracy
I. If we take a look at the correlation matrix of the 10 variables (correlation of each variable with each one of the other nine) we see that there are only strong correlation within each dimension (that is to say, the five variables of the executive-parties dimension are strongly correlated among them but not with the variables of the other dimension. The same happens with federal-unitary variables). All the correlations within the two clusters (dimensions) are statistically significant. The first cluster has somewhat stronger interconnections than the second cluster because de percentage of minimal winning one-party cabinets is a particular strong element (this is theoretically very important because this variable can be seen as the essence of the distinction between concentration of power and the joint exercise of power). The effective number of parliamentary parties is a second key component in this cluster. In the second cluster, the federalism and decentralization variable emerges as the strongest element (this is also theoretically very important because this variable can be seen as the essence of the federal-unitary dimension).
A similar result (the clusters with strong interconnections within each one and weak interconnection between them) can be found by means of factor analysis.
II. If we average for each country the characteristics on each of the two sets of five variables we can place each country on a two-dimensional conceptual map of democracy (Figure 14.1). The horizontal axe represents the executive-parties dimension and the vertical axe the federal-unitary dimension. The model countries, UK and Switzerland are clearly where expected. Switzerland, the paradigm of consensual democracy is on the federal-joint-power corner whereas the UK is on the unitary-power-concentration corner. This map also reveals prototypes of combination types, for example, the US are on the federal-power-concentration corner (as well as Canada etc.) while the Scandinavian democracies are on the unitary-joint-power corner.
There are some important guidelines that are shown be this graphic :
a) There is a strong correlation between British heritage and majoritarian type of democracy (at least with respect to the position on the concentration-joint power axe). It is interesting to note that exceptions to this pattern (India, Israel, Mauritius and Papua New Guinea) are also plural societies suggesting that it is the degree of pluralism that explains why these countries are consensual rather than majoritarian.
b) The degree in which countries are plural societies also appears to explain the location of the 36 democracies on the federal-unitary axe. Of the 13 countries situated below the middle (tending to a strong federalism) 10 are plural or semiplural societies. An additional explanation is given by population size. Indeed, of the fifteen countries with population greater than 10.000.000, ten are in the bottom part (tending to strong federalism). Although this variable has a stronger correlation with the degree of federalism than the plural character of these societies, plurality serves also to explain the position on the first dimension axe and has therefore a stronger value for the overall explanation.
c) It should be also noted that federalist regimes are fewer than centralized ones as shown by the graphic.
If we take a look at the shifts in the position on the map we do not find a clear direction towards either majoritarianism or consensualism (see figure 14.2). Instead, the pattern shown is one of great stability (changes have indeed changed the position of some countries but these changes do not have a clear direction).
Chapter 15 : Macroeconomic Management and the Control of Violence (Does consensus democracy make a difference)
I. Does the difference between majoritarian and consensus democracy make a difference for the operation of democracy, especially for how well democracy works ? The conventional hypothesis is that there is a trade-off between the quality and the effectiveness of democratic government (the more a government is representative the less it is effective for policy-making because of the lack of consensus within the government). This traditional hypothesis rests on the strong argument that concentrating political power in the hands of a narrow majority can promote unified, decisive leadership and hence coherent policies and fast decision-making. There are however several counter arguments that should be noted :
a) Majoritarian governments may be able to make decisions faster than consensus governments, but fast decisions are not necessarily wise decisions (in fact, the opposite may be more valid)
b) The supposedly coherent policies produced by majoritarian governments may be negated by the alternation of these arguments ; this alternation from left to right and vice-versa may entail sharp changes in economic policy that are too frequent and too abrupt (as S. E. Finer says : successful macroeconomic management requires not so much a strong hand as a steady one PR and coalition governments are better able to provide steady, centrist policy-making. Policies supported by a broad consensus, furthermore, are more likely to be carried out successfully and to remain on course than policies imposed by a decisive government against the wishes of important sectors of society.
c) For maintaining civil peace in divided societies, conciliation and compromise (goals that require the greatest possible inclusion of contending groups in the decision-making process- are probably much more important than making snap decisions.
Empirical evidence concerning this argument and its counter arguments is mixed. P. Katzestein (1985) and R. Rogowski (1987) have shown that small countries adopted PR and corporatist practices to compensate for the disadvantages of their small size in international trade ; that is, these consensus elements served as sources of strength. R. Rose (1992) and F. G. Castles (1994) find no significant differences in economic growth, inflation, and unemployment between PR and no-PR systems among industrialized democracies. On the other side, N. Roubini and J. Sachs (1989) do find a clear connection between, on the one hand multiparty coalition government and governments with a short average tenure (both characteristics of consensus democracy) and large budget deficits on the other hand. However their methods and conclusions have been very challenged. M.L.Crepaz (1996) finds that, in the OECD countries, consensual institutions are not related to economic growth but do have significantly favorable effects on inflation, unemployment, and the number of working days lost as a result of industrial unrest. G.Bingham Powell (1982) finds that representational democracies (consensual democracies) have a better record than majoritarian democracies with regard to controlling violence.
All these test have to do with macroeconomic management and control of violence, which are two excellent performance indicators because they involve crucial functions of government and because precise quantitative data are available.
From now on, Lijhpart adopts as his working hypothesis that consensus democracy produces better results (without expecting that the differences will be very strong because there are several variables, out of government control, that can influence upon the socioeconomic situation of a country. The existence of these other variables does not, however, mean that governments have no control at all. Furthermore, the effects of such fortuitous events as well as external influences that cannot be clearly identified and controlled can be minimized when economic performance is examined over a long period and for many countries) in macroeconomic management and control of violence.
II. Consensus democracy and successful macroeconomic management : Empirical test of the working hypothesis shows that the evidence with regard to economic growth and economic freedom is mixed, but with regard to unemployment, strike activity and budget deficits the consensus democracies have a slightly better record than majoritarian ones and with regard to inflation it has a significantly better record.
III. Consensus democracy and the control of violence : At first sight, consensual democracy appears to a significant negative correlation with the number of riots and deaths from political violence per million people (that is to say, the more a democracy is consensual, the less the number of riots and political violence is, so the better the control of violence is). However, these results are relative, because if we focus on the effect that the degree of development, the degree of pluralism, and the population size have on these results we find that consensual democracy has a minor impact or correlation with the number of riots and political violence. However, the correlations remain negative, which means that consensus democracy has a slightly better performance on the control of violence than majoritarian democracies.
IV. If we now apply the logic of the traditional argument to the federal-unitary dimension we find that empirical evidence is also very mixed with one exception : consensual democracy has a strong negative correlation with inflation (the more a democracy tends to have a joint-power, the less the level of inflation). This finding was clearly expected because one of the variables of the federal-unitary dimension for consensual democracies is central bank autonomy (which allows the central bank to concentrate on keeping a low inflation level).
V. The general findings of this chapter are the following :
a) On balance, consensus democracies have a better performance record than majoritarian democracies, especially with regard to the control of inflation but also, albeit much more weakly, with regard to most of the other macroeconomic performance variables and the control of violence ; majoritarian democracies do not have an even slightly better record on any of the six groups of performance variables
b) However, the overall results are relatively weak and mixed ; when controls were introduced and outliers were removed, few statistically significant correlations were found. Hence, the empirical results do not permit the definitive conclusion that consensus democracies are better decision-makers and better policy-makers than majoritarian systems
c) Therefore, the most important conclusion of this chapter is negative : majoritarian democracies (contrary to what the traditional argument states, namely that there is a trade-off between the quality and the effectiveness of democratic government) are clearly not superior to consensus democracies in managing the economy and in maintaining civil peace. This means that the second part of the conventional argument (the effectiveness of democratic government) is not empirically proven. It remains the question of knowing if consensual democracies have a better performance for the first part of the conventional argument (the quality or representativity of democratic government).
Chapter 16 : The quality of democracy and a "kinder, gentler" democracy (consensus democracy makes a difference)
I. The conventional argument admits that consensus democracy is better at representing, in particular, representing minority groups and minority interests, representing everyone more accurately, and representing people and their interests more inclusively. The extent to which consensus democracies perform this representation function better than majoritarian democracies can be shown through several measures of the quality of democracy, and the extent to which consensus democracy can be conceived of as kinder and gentler than majoritarian democracy can be shown through different elements (how likely they are to be welfare states, to protect the environment, to put fewer people in prison and avoid the death penalty, and to give economic assistance to developing countries.
II. The quality of democracy can be measured in several ways (when we refer to consensus democracy we include exclusively the concentration-joint-power axe which has been called executive-parties dimension) :
a) The Dahl's rating of democratic quality is strongly correlated to consensus democracy (to a higher degree of joint power corresponds a higher degree of democratic quality). The same happens with the Vanhanen's rating of democratic quality (based on the degree of competition among parties and the degree of participation). The correlation remains significant when the level of development is controlled for and the aberrant cases are removed.
b) The measure of women's political representation and the protection of women's interests are a strong indicator of the quality of democracy. Women are a minority that exists everywhere and can thus be compared systematically across countries. This comparison gives an insight on how well minorities are represented generally. The political representation of women has a strong positive correlation with the degree of consensus democracy, even after the degree of development has been controlled for. If we take into account women representation in the executive power (president or ministers) the positive correlation becomes even stronger. Concerning protection of women's interests, H.L.Wilensky proposes a rating the industrialized democracies according to their innovativeness and expansiveness of their family policies (a matter of special concern to women). In this respect, the correlation with consensus democracy is significant and unaffected by the level of development. France is a majoritarian deviant case with one of the highest scores in innovativeness and expansiveness of women's interest. When this outlier is removed, the correlation becomes stronger.
c) The basic goal of democracy is the degree of political equality. Political equality is difficult to measure directly but economic equality can serve as a valid proxy, since political equality is more likely to prevail in the absence of great economic inequalities. The rich-poor ratio is the ratio of the income of the highest 20% to that of the lowest 20 percent of households (income of the highest 20% households / income of the lowest 20% households). Consensus democracy has a very strong negative correlation with inequality thus measured even when the level of development is controlled for. A similar method that takes into account the highest and the lowest deciles (instead of the 20s %) shows similar results. Finally, Vanhanen's index of power ressources (indicator of equality based on several measures such as the degree of literacy and the percentage of urban population) is positively correlated with consensual democracy especially when the level of development is controlled for.
d) Voter turnout (electoral participation) is an excellent indicator of democratic quality because it shows the extent to which citizens are actually interested in being represented, and because, as turnout is strongly correlated with socioeconomic status, it can therefore serve as an indirect indicator of political equality (high turnout or participation means more equal participation and hence greater political equality and vice-versa). Consensus democracy and voter turnout are positively correlated. However, several controls have to be included to obtain a more reliable picture : compulsory voting which tends to increase turnout percentage ; frequency and multitude of electoral choices tend to depress the turnout percentage ; turnout tends to be higher in more developed countries. If these three variables are controlled for, the effect of consensus democracy on voter turnout becomes much stronger and significant.
e) Does the type of democracy affect citizens' satisfaction with democracy ? H-D. Kingemann reports that citizens in consensus democracies are significantly more satisfied with democratic performance in their countries than citizens of majoritarian democracies. An earlier study had shown that those who have voted for a winning party tended to be more satisfied. However, a more realist measure of satisfaction should also take into account those who voted for a party that lost. If we consider the distance the number of satisfied voters who won and those who lost we obtain a much meaningful measure. Once again, in consensus democracies the differences between winners and losers were significantly smaller than in majoritarian democracies.
f) Government-voter proximity. The distance between voters positions and government policies' position can be an interesting measure of democratic quality. The J.D. Huber and G.Bingham Powell's measures show this distance ("government distance" is the distance between the government's position on a left-right scale and the position of the median voter). In addition, they measure the "voter distance" (the percentage of voters between the government and the median citizen). The smaller these two measures are, the more representative the government is of the citizens' policy preferences). Contrary to the majoritarian claim (that the government's policy position is likely to be close to the bulk of the voters because in two-party systems parties are likely to be moderate), both distances are actually smaller in consensus than in majoritarian democracies.
g) Accountability and corruption : another claim of majoritarian democracy is that its typically one-party majority governments offer clearer responsibility for policy-making and hence better accountability of the government to the citizens (who can use elections either to renew the term of the incumbent government or to throw the rascals out). The claim is undoubtedly valid for majoritarian systems with pure two-party competition. However, in two-party systems with significant third parties, "rascals" may be repeatedly returned to office in spite of clear majorities of the voters voting for other parties and hence against the incumbent government. Moreover, it is actually easier to change governments (although partially) in consensus democracies than in majoritarian democracies where cabinet duration is much longer. A related measure is the incidence of corruption. Greater clarity of responsibility typical of majoritarian democracies is claimed to inhibit corruption whereas consensus systems' tendency to compromise and "deal-making" fosters corrupt practices. The actual relationship between consensus democracy and corruption is statistically insignificant. Moreover, it is slightly negative. This negative correlation becomes slightly stronger (but not significant) when the level of development is controlled for.
h) John Stuart Mill hypothesis : majority rule is the most fundamental requirement of democracy and the combination of plurality or majority elections and parliamentary government may lead to minority rule (indeed, a large minority may obtain the control in a plural election based on the motto "first past the post"). On the other hand, consensus democracies, which frequently use PR and which in addition tend to have more inclusive coalition cabinets, are most likely to practice a true majority rule than majoritarian democracies. Two measure can be used to test Mill's hypothesis : popular cabinet support ; the percentage of time that the majority-rule requirement (which means that the cabinet or president are supported by popular majorities) is fulfilled (this measure is called the Mill criterion). Empirical evidence does not show strong statistical correlations between consensus democracy and either measure and this for three reasons :
- As smaller units have fewer political parties even when they use PR (Dahl), the smallest majoritarian democracies have high popular cabinet support as a result of their almost pure two-party systems in which a winning party usually wins a popular majority. When population size is controlled for, consensus democracy and popular cabinet support become positively and significantly correlated and the same happens between consensus democracy and the Mill criterion.
- Presidential systems, although majoritarian, tend to secure popular support for the executive
- Consensus democracies with frequent minority governments, especially the Scandinavian countries, have relatively low cabinet support. But when counting votes we do not take into account the tendency towards strategic vote. If popular cabinet support could be calculated on the basis of the voters' sincere preferences instead of their actual votes, the consensus democracies would do much better on this indicator of democratic quality.
The general conclusion is that consensus democracies have a better record than majoritarian democracies on all of the measures of democratic quality, and that all except two are statistically significant. (correlations regarding the federal-unitary cluster are not significant in any of the measure of quality of democracy).
III. Consensus democracy and its kinder, gentler qualities : in addition the aforementioned indicator of quality of democracy, consensus democracies (on the executives-parties dimension) is associated with some other attributes that are attractive from a democratic perspective : a strong community orientation and social consciousness (the kinder, gentler qualities). There are especially four areas where the kinder gentler qualities of consensus democracy are likely to manifest themselves :
a) Consensus democracies are more likely to be welfare states (even after the level of development has been controlled for). This has been confirmed by two measures : the Esping-Andersen's measure (that has been severely criticized for understating the degree to which Australia, New Zeland, and the UK are welfare states. However, if we re-run the experience after removing these three states the overall results are the same) ; social expenditure as a percentage of gross domestic product (PIB).
b) Consensus democracies have clear better records on environmental performance than majoritarian democracies. This have been measured by means of two indicators : Monte Palmer's composite index of concern for the environment (based mainly on carbon dioxide emissions, fertilizer consumption, and deforestation ) which has a significant positive correlation with consensus democracy (even after the level of development has been controlled for) ; the energy efficiency (gross domestic product divided by total energy consumption. The most environmentally responsible countries produce goods and services with the lowest relative consumption of energy) which has an extremely strong positive correlation with consensus democracy unaffected by the introduction of the level of development as a control variable.
c) Consensus democracy is negatively correlated with incarceration. However, this result is strongly affected by the aberrant case of the US (four times more of inmates than the following country on the scale). If we remove the US, the negatively correlation between consensus democracy and incarceration becomes stronger. When in addition the level of development is controlled for, the correlation becomes clearly significant and strong. The result of these measures is that consensus democracies put about 26 fewer people per hundred thousand population in prison than the majoritarian democracies. Concerning the death penalty, the negative correlation between consensus democracy and the death penalty is strong and highly significant and is not affected by controlling for level of development.
d) In the field of foreign policy, three indicators (average annual foreign aid as a percentage of gross national product during a certain period before the cold war end ; idem in the post cold war period ; foreign aid in the latter period as a percentage of defense expenditures) are strongly correlated with consensus democracy although at different levels. Indeed, two important controls need to be introduced : the level of development (because wealthier countries can better afford to give foreign aid ) ; population size (representing the size of a country, because larger countries tend to assume greater military responsibilities and therefore to have larger defense expenditures, which limits their ability and their willingness to provide foreign aid). After the introduction of these two controls, the correlation remains significant. These measures show that the typical consensus democracy gave about 0.2 percent more of its gross national product in foreign aid than the typical majoritarian democracy in both the Cold War and post Cold War periods.
The federal-unitary dimension does not present interesting correlations to be noted. As a conclusion, we can say that consensus democracy (on the executives-parties dimension) makes a big difference with regard to almost all of the indicators of democratic quality and with regard to all of the kinder and gentler qualities.
Chapter 17 : Conclusions and Recommendations
Lijhpart closes his work indicating that the variety of institutions and arrangements found in the different democracies can all be reduced to the two dimensions defining two types of democracy : majoritarian and consensus type. Consensus type has the higher performance because, not only it has an at least similar performance to the majoritarian type as long as macroeconomic management and control of violence are concerned, but furthermore it has a clearly superior performance in terms of democratic quality and kindness and gentleness.
The corollary of these findings is the recommendation of the consensus type of democracy for the constitutional engineering. According to the findings of the book, the consensus type of democracy emerges (with very few exceptions which can be explained) from the combination of a Parliamentary system with a substantially proportional electoral system. The typical problems of Parliamentary regimes, cabinet instability and executive-legislative deadlock, can be solved by the adoption of a certain combination of mechanisms (the German style constructive vote of no confidence, which requires that parliament can dismiss a cabinet only by simultaneously electing a new cabinet. It may happen however that the internal divisions within the parliament impede the election of a new cabinet, and this situation renders the cabinet impotent because parliament rejects all its legislative proposals. In such case, a possible solution can be found by adding the French rule that gives the cabinet the right to make its legislative proposals matters of confidence, which means that parliament can reject such proposals only by voting its lack of confidence in the cabinet by an absolute majority). On the other hand, the most delicate problem of PR, the extreme fragmentation of the party-system, can be solved by adopting a certain minimal percentage of votes for having access to seats (threshold clause). Lijhpart adds to these eventual structural problems, some cultural problems. Indeed, he thinks that both institutional and cultural traditions may present strong resistance to consensus democracy (especially the ex-British colonies). However, he finishes his book pointing out, first that there is a great deal of interaction between culture and structure (mutual influence and not only culture as the cause and structure as the effect), and second that some common grounds exist already in critical regions (South-East and South-West) that are fertile for the developing of consensus democracies.