Introduction

In the first edition of his Patterns of Democracy, published in 1999, Lijphart (1999, 9) used the term ‘Westminster model’ ‘interchangeably with majoritarian model to refer to a general model of democracy’. In other words, he regarded the main features of the UK’s political system (the Westminster system) as a ‘prototypical instance of majoritarianism’ (Whitehead 2013, 9), called the ‘Westminster’ or ‘majoritarian’ model.Footnote 1 However, shortly before the book was published, the Blair Labour government had ‘set in train the most radical programme of constitutional reform that Britain had seen since 1911 or 1832’ (Bogdanor 2001, 143). It can even be regarded as more radical than the reforms of 1911 and 1832 (Bogdanor 2001, 143). Just over a decade later, the 2010 general election led to a hung parliament and to the first coalition government since Churchill’s wartime one and ‘the first peacetime coalition government since the 1930s’ (Curtice 2010, 623). Despite these developments, Lijphart (2012a, 20) concluded in the second edition of his book published in 2012 that ‘recent changes in British politics do not change the overall character of Britain as a prime example of majoritarian democracy.’ Both Lijphart’s interchangeable use of the terms ‘Westminster’ and ‘majoritarian’ model and his conclusion on the effects of the aforementioned developments are challenged by this article, which also examines the effect of the 2015 general election on the nature of the Westminster system through the prism of Lijphart’s Patterns of Democracy. This article aims to answer the questions of (1) whether the Westminster system represented the majoritarian ideal in the past; (2) whether the Westminster model has noticeably changed as a result of the various constitutional reforms (reforms of the UK’s political system) introduced since the change of government in 1997 on the one hand and the outcome and consequences of the 2010 general election on the other; and (3) how the 2015 general election has influenced the nature of the Westminster system.

The Westminster system and its changing nature

Lijphart contrasts two ideal types of democracy, reflecting the ‘theoretical dichotomy between . . . power hoarding versus power-sharing’ (Whitehead 2013, 11): the majoritarian and the consensus model, respectively. The majoritarian–consensus contrast is characterised by 10 variables (Lijphart 1999, 10–12) proportionally clustered along 2 dimensions: the executives–parties dimension and the federal–unitary dimension. The executives–parties dimension covers the party system, the cabinet, the executive–legislative relationship, the electoral system and interest groups, while the federal–unitary dimension covers the vertical division of power, the distribution of power in the legislature, the constitutional amendment procedures, the interpretation of the constitution with regard to the constitutional compatibility of laws and the central bank. In his first edition, Lijphart examined 36 countries—including quite diverse democracies, industrialised as well as developing countries, parliamentary as well as presidential systems, and European as well as non-European states spread all over the world—in the period from 1945 to 1996, while his second edition covered a nearly identical set of 36 countries in the period from 1945 to 2010.

The executives–parties dimension

The executives–parties dimension is predominantly influenced by the outcome and consequences of elections—which have a direct influence on the electoral system and the party system, as well as an indirect influence on the cabinet and the executive–legislative relationship. In the UK, the 2010 general election was of particular importance in this regard, as it led to a hung parliament and the first coalition government since 1945. The replacement of a single-party by a two-party coalition government influenced the nature and functioning of the Westminster system to a great extent. A single-party majority government in the UK provides the maximum potential for policy change (Tsebelis 2002, 78–9) and thereby maximum transparency and political accountability, as the governing party is the only bearer of parliamentary sovereignty and thus has—in the words of Dicey (1897, 38)—‘the right to make or unmake any law whatever’. Thus it is able to keep all the promises made in its election manifesto without making any compromises and, as a consequence, can be held solely accountable for the government’s record. A two-party government, by contrast, relies on compromises and concessions, with the accompanying negative effects on governability and transparency, as well as political accountability. However, not only the 2010, but also the 2015 general election significantly influenced the executives–parties dimension in the UK—albeit in a slightly different way.

The first variable of Lijphart’s Patterns of Democracy refers to the party system and basically contrasts two-party systems and multiparty systems (Lijphart 1999, 63). Lijphart counts the number of parties using the ‘effective number of parliamentary parties index’ (ENPP)Footnote 2 created by Laakso and Taagepera (1979). The average ENPP for all British elections between 1945 and 1996 calculated by Lijphart (1999, 76–7) is 2.11. Hence, the UK nearly represented the majoritarian ideal (2.00) in this period. The average score between 1945 and 2010 presented by Lijphart (2012a, 74–5) in his second edition is 2.16 and shows a marginal increase over time. However, this downplays the erosion of the British two-party system in the 2010 general election, which resulted in the highest ENPP score (2.57) since 1945. The 2015 general election produced a score of 2.54 and thereby stabilised the ‘effective two-and-a-half party system’ (Table 1). Thus, with regard to the party system, the UK has clearly moved towards greater consensus in both the 2010 and 2015 general elections (compared to the average score between 1945 and 2010), thereby entering a grey area between the majoritarian and the consensus models.

Table 1 The (average) ENPP in the UK
Full size table

The second variable concerns the cabinet and essentially refers to the difference between single-party majority governments and oversized multi-party coalitions (Lijphart 1999, 62). Lijphart (1999, 109) measures ‘the overall degree of majoritarianism in the formation of cabinets’ by the average time in office of one-party cabinets and minimal winning cabinets. The average time spent in office by these two cabinet types in the UK between 1945 and 1996 calculated by Lijphart (1999, 110–11) is 96.7 %. As a consequence, the UK nearly represented the majoritarian ideal (100 %) in this period. The average score between 1945 and 2010 presented by Lijphart (2012a, 99–100) in his second edition is 97.3 %, which indicates a marginal increase over time. However, this downplays and does not fully take into account the coalition government formed after the 2010 general election, which scored 50 %—and moved the UK into a grey area between the majoritarian and the consensus models. While opinion polls predicted another hung parliament and, as a consequence, a coalition or minority government, the 2015 general election led to a single-party majority government, resulting in a score of 100 %, which corresponds with the majoritarian ideal (Table 2). Hence, given the make-up of the cabinet, the UK clearly moved towards greater consensus in the 2010 general election (compared to the average score between 1945 and 2010) and then back towards greater majoritarianism—to the majoritarian ideal—in the 2015 general election.

Table 2 The (average) time spent in office by one-party cabinets and minimal winning cabinets in the UK (per cent)
Full size table

The third variable refers to the relationship between the executive and legislative branches of government and basically contrasts executive dominance with ‘a more balanced executive–legislative relationship’ (Lijphart 1999, 116). Lijphart (1999, 129–31) (continuously) measures the degree of executive dominance using the ‘index of cabinet duration’ suggested by Dodd (1976), drawing on only one criterion for cabinet termination: the change in party composition. The average cabinet duration in the UK between 1945 and 1996 calculated by Lijphart (1999, 132–3) is 8.49 years and ranks the UK in 30th place in Lijphart’s sample of 36 countries listed in ascending order. Hence, the UK was among the most majoritarian in Lijphart’s sample in this period. The average score between 1945 and 2010 presented by Lijphart (2012a, 120–1) in his second edition is 8.12 years and shows a marginal decrease over time. However, this on the one hand downplays and on the other does not fully take into account the cabinet duration of the Conservative/Liberal Democrat coalition government which ousted the Labour single-party government (after 13 years of Labour in power) as a result of the 2010 general election and, in turn, was ousted by the Conservative single-party government as a result of the 2015 general election—which resulted in a score of 5 years (Table 3). Thus, in terms of the executive–legislative relationship, the UK has clearly moved towards greater consensus in the 2010 and 2015 general elections (compared to the average score between 1945 and 2010), yet it has remained majoritarian. The likelihood of long cabinet durations after the 2015 general election has been reduced by the Fixed-term Parliaments Act 2011 which removed the power of the prime minister to call an election at a time when he has the best chance of winning.

Table 3 The (average) cabinet duration in the UK (years)
Full size table

The fourth variable concerns the electoral system and essentially refers to the difference between plurality/majority systems (~disproportional systems) and systems of proportional representation (Lijphart 1999, 143). Lijphart basically measures the overall disproportionality of electoral systems using the ‘least squares index’Footnote 3 proposed by Gallagher (1991). The average electoral disproportionality over all British elections between 1945 and 1996 calculated by Lijphart (1999, 162) is 10.33 and ranks the UK in 22nd place in Lijphart’s sample of 36 countries listed in ascending order. As a consequence, the UK was far from representing the most majoritarian type in Lijphart’s sample, but still fit the majoritarian model in this period. The average score between 1945 and 2010 presented by Lijphart (2012a, 150–1) in his second edition is 11.70 and indicates a marginal increase over time. However, this downplays the electoral disproportionality in the 2010 general election, which scored 15.10 and was nearly reached again in the 2015 general election, which scored 15.02 (Table 4). Hence, in terms of the electoral system, the UK moved towards greater majoritarianism in both the 2010 and 2015 general elections (compared to the average score between 1945 and 2010). The proposed introduction of the Alternative Vote system, rejected in the 2011 Alternative Vote Referendum, would not have increased, but probably even decreased electoral proportionality—as already observed by the Independent Commission on the Voting System (1998). Another attempt at electoral reform in the near future is highly unlikely, as the governing Conservative Party (2015, 49) announced that it would ‘respect the will of the British people, as expressed in the 2011 referendum, and keep First Past the Post for elections to the House of Commons’.

Table 4 The (average) degree of disproportionality in the UK electoral system
Full size table

The fifth variable refers to interest groups and basically contrasts ‘a competitive and uncoordinated pluralism of independent groups’ with ‘the coordinated and compromise-oriented system of corporatism’ (Lijphart 1999, 171). Lijphart measures interest group pluralism using the index created by Siaroff (1999). The average interest group pluralism in the UK between 1945 and 1996 provided by Lijphart (1999, 177) is 3.38 and ranks the UK in 34th place in Lijphart’s sample of 36 countries listed in ascending order. Hence, the UK was almost the most majoritarian in Lijphart’s sample of 36 countries in this period. The average score between 1945 and 2010 presented by Lijphart (2012a, 165–6) in his second edition is 3.02 and shows a marginal decrease over time. Thus, in terms of interest groups, the UK has slightly moved towards greater consensus over time, while still remaining majoritarian. However, in contrast to other variables, interest group pluralism was affected by neither the 2010 and 2015 general elections nor by political reforms in the more recent past (Lijphart 2012b, 2).

The federal–unitary dimension

The federal–unitary dimension is predominantly influenced by constitutional reforms. In the UK, constitutional reforms are actually in the hands of the elected government, as the British constitution can be changed ‘by means of a simple majoritarian decision’ (Lijphart 1999, 18). Despite this fact, the British governments’ enthusiasm for constitutional reform has rarely been high over the course of the last century. Even for the Labour Party, for a long time constitutional reform was not really a major priority (Bogdanor 2001, 139). In contrast, ‘Labour . . . has, at least since the 1920s, been distinctly sceptical if not downright hostile towards constitutional reform’ (Bogdanor 2001, 142). However, Labour’s attitude towards constitutional reform changed considerably in the late twentieth century and, after winning the 1997 general election, the Blair Labour government started an (almost) unprecedented constitutional reform programme, which has significantly, as well as sustainably, influenced the federal–unitary dimension.

The sixth variable in Lijphart’s Patterns of Democracy concerns the vertical division of power and essentially refers to the difference between unitary and centralised governments on the one hand and federal and decentralised governments on the other (Lijphart 1999, 186). Lijphart (1999, 189) measures the degree of federalism and decentralisation using an index based on a five-point scale from 1 for ‘unitary and centralised’ to 5 for ‘federal and decentralised’. The degree of federalism and decentralisation in the UK between 1945 and 1996 assigned by Lijphart (1999, 189) is 1. As a consequence, the UK represented the majoritarian ideal in this period. The average score between 1945 and 2010 presented by Lijphart (2012a, 178) in his second edition is 1.2 and indicates a marginal increase over time. However, this downplays the effect of the devolution reforms introduced since 1998. Nevertheless, while Lijphart assigned the UK a score of 2 (unitary and decentralised) after 1998, it seems to be appropriate to assign the UK an intermediate position between centralised and decentralised (‘semi-decentralised’) after the commencement of the first devolution reforms introduced since the change of government in 1997—such as the Scotland Act 1998, the Government of Wales Act 1998, the Belfast (‘Good Friday’) Agreement 1998 and the Greater London Authority Act 1999 which came into force between 1998 and 2000 and were followed by further devolution reforms. These reforms partially and asymmetrically transferred the power of the central government to regional or local governments (Table 5). Hence, with regard to the degree of federalism and decentralisation, the UK has moved slightly towards greater consensus since 1998 (although not all parts of the aforementioned devolution reforms came into force in 1998), yet it has remained majoritarian. As the governing Conservative Party (2015, 69) has announced that it ‘will continue devolution settlements for Scotland and Wales, and implement the Stormont House Agreement in Northern Ireland’, the UK is likely to be gradually pushed further towards full decentralisation and, as a consequence, in the end marginally further towards greater consensus (based on the proposed contemporary classification of ‘semi-decentralised’).

Table 5 The (average) degree of federalism and decentralisation in the UK
Full size table

The seventh variable refers to the distribution of power in the legislature and basically contrasts the concentration and the division of legislative power in the legislature (Lijphart 1999, 200). Lijphart (1999, 205–7) measures the distribution of power in the legislature using an index principally based on a four-point scale ranging from 1 for ‘unicameralism’ to 4 for ‘strong bicameralism’ (symmetrical and incongruent chambers). The distribution of power in the legislature in the UK between 1945 and 1996 assigned by Lijphart (1999, 212) is 2.5. Lijphart classifies the British Parliament generally as asymmetric and (‘technically’) incongruent, but reduces its score of 3 by half a point, as he regards the House of Lords as a ‘relic of a pre-democratic era’ (Lijphart 1999, 213). Hence, in this variable the UK was far from representing the majoritarian ideal (a score of 1), but was situated in a grey area between the majoritarian and the consensus model in this period. The score between 1945 and 2010 presented by Lijphart (2012a, 199) in his second edition is also 2.5. The House of Lords Act 1999—which removed all but 92 hereditary peers from the UK’s second chamber (Bogdanor 2001, 139)—affected the composition of the House of Lords, but did not change the congruency or, as a consequence, the distribution of power in the British legislature. Thus, with regard to the distribution of power in the legislature, the UK has continuously been a hybrid between the majoritarian and the consensus model since 1945. A mainly elected House of Lords with elections held on the basis of proportional representation, as recommended by the Joint Committee on the Draft House of Lords Reform Bill (2012), but not implemented in the last electoral period, would reduce the degree of congruency (as the House of Lords could no longer be regarded as a ‘relic of a pre-democratic era’) and push the UK towards greater majoritarianism. However, the chances of a major House of Lords’ reform after the 2015 general election were not increased by the electoral victory of the Conservative Party, which ‘see[s] a strong case for introducing an elected element into . . . [the] second chamber’, but regards this as ‘not a priority in the next Parliament’ (Conservative Party 2015, 49).

The eighth variable concerns the procedures to amend the constitution and essentially refers to the difference between ‘flexible constitutions’ and ‘rigid constitutions’ (Lijphart 1999, 216). Lijphart (1999, 218–20) measures the degree of constitutional rigidity using an index principally based on a four-point scale ranging from 1 for ‘approval by an ordinary majority’ to 4 for ‘approval by more than a two-thirds majority’. The degree of rigidity of the British constitution between 1945 and 1996 assigned by Lijphart (1999, 220) is 1. As a consequence, the UK represented the majoritarian ideal in this period. The score between 1945 and 2010 presented by Lijphart (2012a, 208) in his second edition is also 1. It might be argued that the frequent use of referenda on constitutional issues since 1997 has created ‘a persuasive precedent, if not a convention, . . . such that a referendum is required on any proposal to transfer the powers of Parliament’ (Bogdanor 2001, 78). However, from a legal perspective, referenda are generally not required (in addition to legislative approval) in the UK. Hence, with regard to the procedures required to amend the constitution, the UK has continuously represented the majoritarian ideal since 1945.

The ninth variable refers to the interpretation of the constitution with regard to the constitutional compatibility of laws and basically contrasts parliamentary sovereignty and the existence of judicial review (Lijphart 1999, 216). Lijphart (1999, 225–6) measures the strength of judicial review using an index based on a four-point scale ranging from 1 for ‘no judicial review’ to 4 for ‘strong judicial review’. The strength of judicial review in the UK between 1945 and 1996 assigned by Lijphart (1999, 226) is 1. Hence the UK represented the majoritarian ideal in this period. The score between 1945 and 2010 presented by Lijphart (2012a, 215) in his second edition is also 1 (Table 6). However, parliamentary sovereignty has been affected by the Human Rights Act (United Kingdom 1998), which came into force in 2000 and incorporated the European Convention on Human Rights (ECHR) into British law and provided British courts with the power to review the provisions of primary and subordinate legislation for compatibility with the ECHR. If certain courts (United Kingdom 1998, Section 4 (5)) issue a declaration of incompatibility concerning primary legislation, government and parliament in Westminster are called upon—but not committed—to repeal or amend the offending statute—which can be done by a specific fast-track method (United Kingdom 1998, Section 10; Bogdanor 2001, 89). It may be argued that the Human Rights Act did not change the strength of judicial review in the UK, as (1) the ECHR is far from representing a ‘single unified and legally binding constitutional charter’ (Whitehead 2013, 17); (2) courts are required by the Human Rights Act (1998, Section 3 (1)) to read and to give effect to ‘primary legislation and subordinate legislation . . . in a way which is compatible with the Convention rights’ as far as possible; and (3) a declaration of incompatibility ‘does not affect the validity, continuing operation or enforcement of the provision in respect of which it is given’ (United Kingdom 1998, Section 4 (6)(a)) and ‘is not binding on the parties to the proceedings in which it is made’ (United Kingdom 1998, Section 4 (6)(b)). However, due to the existence of a specific fast-track method in the event of a declaration of incompatibility and the fact that declarations of incompatibility (if not overturned on appeal) have nearly always been remedied and, as a consequence, have noticeably influenced British politics in the past (Human Rights Futures Project 2013), it seems to be appropriate to assign the UK an intermediate position between no judicial review and weak judicial review since 2000 (a score of 1.5). Thus, in terms of the interpretation of the constitution with regard to the constitutional compatibility of laws, the UK has slightly moved towards greater consensus since 2000, while still remaining majoritarian. The governing Conservative Party (2015, 60) has announced that it ‘will scrap the Human Rights Act, and introduce a British Bill of Rights’, but it remains unclear to what extent this reform would alter the strength of judicial review in the UK.

Table 6 The (average) strength of judicial review in the UK
Full size table

The tenth variable concerns the central bank and essentially refers to the difference between weak and dependent central banks on the one hand and strong and independent central banks on the other (Lijphart 1999, 232). Lijphart measures the independence of central banks by two different indices developed by Cukierman et al. (1994) and an index proposed by Grilli et al. (1991). The average degree of central bank independence in the UK between 1945 and 1996 provided by Lijphart (1999, 237) is 0.31—the mean of the (first) ‘Cukierman index’ (0.30) and the ‘Grilli index’ (0.32)—and ranks the UK in 27th place in Lijphart’s sample of 36 countries listed in descending order. As a consequence, the UK was among the most majoritarian in Lijphart’s sample in this period. The average score between 1945 and 2010 presented by Lijphart (2012a, 234) in his second edition is 0.33 and indicates a marginal increase over time. However, this downplays and undervalues the increased independence of the Bank of England that was initiated by the Blair Labour government, which transferred the ‘operational responsibility for setting interest rates’ to the Bank of England only a few days after winning the 1997 general election (Bank of England 1997, 9). Subsequently, it implemented the Bank of England Act 1998, which formally provided the Bank of England’s Monetary Policy Committee with the responsibility for setting interest rates (Whitehead 2013, 25). While the average degree of central bank independence in the UK between 1998 and 2009 presented by Lijphart (2012b, 6) is only 0.40, Arnone et al. (2007, 45) provided a more plausible average score of 0.69 (based on the ‘Grilli index’) at the end of 2003 (Table 7). Hence, with regard to the central bank, the UK has clearly moved towards greater consensus since 1997 (compared to the average scores between 1945 and 2010) and now fits the consensus model more than the majoritarian model.

Table 7 The (average) degree of central bank independence in the UK
Full size table

Conclusion

  1. 1.

    The Westminster system has never represented the majoritarian ideal: it has never been a system which perfectly matches the characteristics of the majoritarian model or is consistently the most majoritarian in Lijphart’s sample of 36 countries. It was strongly majoritarian between 1945 and 1996, but Lijphart (1999, 9) was wrong to use the term Westminster model ‘interchangeably with majoritarian model’, as there has always been a major ‘mismatch between the case and the model’ (Whitehead 2013, 10); in other words: there exist ‘serious discrepancies between mythology and reality’ (Whitehead 2013, 11)—between the (idealised) Westminster or majoritarian model and the (real) Westminster system.

  2. 2.

    The Westminster system has noticeably moved towards greater consensus as a result of the various constitutional reforms introduced since the change of government in 1997 on the one hand and the outcome and consequences of the 2010 general election on the other. Whereas the constitutional reforms caused a shift towards greater consensus along the federal–unitary dimension, the outcome and consequences of the 2010 general election predominantly led to a shift towards greater consensus along the executives–parties dimension. In the end, the Westminster system remained majoritarian, but Lijphart (2012a, 20) was wrong to conclude that the aforementioned developments did ‘not change the overall character of Britain as a prime example of majoritarian democracy’.

  3. 3.

    The 2015 general election pushed the Westminster system towards greater consensus with regard to the party system and the executive–legislative relationship on the one hand and towards greater majoritarianism with regard to the cabinet and the electoral system on the other along the executives–parties dimension (compared to the average scores between 1945 and 2010). As a consequence, the outcome and consequences of the 2015 general election have somehow stopped the general trend towards greater consensus. The contemporary Westminster system is not—as Whitehead (2013, 14) argued—a ‘hybrid-political system’. Even the outcome and consequences of the 2010 general election did not transform the Westminster system into a ‘hybrid-political system’. Taking everything into account, the contemporary Westminster system is still a majoritarian one, but it definitely does not represent the (idealised) Westminster or majoritarian model. On the contrary, it is less majoritarian than it has been in the past.

There is no undisputable academic answer to the question of whether the Westminster system should be pushed further towards greater consensus or back towards greater majoritarianism in the future, as neither the consensus nor the majoritarian model is principally superior—not in general and not even with regard to a single country like the UK. Lijphart (1999, 258–309) obviously preferred the consensus model, but both the majoritarian and the consensus models have advantages and disadvantages which can rarely be compensated for. For example, the majoritarian model leads to a high potential for policy change, but a low potential for policy consensus and, as a possible consequence, to some kind of ‘elective dictatorship’ (Hailsham 1978, 9), whereas the consensus model leads to a high potential for policy consensus, but a low potential for policy change and, as a possible consequence, to a form of ‘obstructed republic’ (Der Spiegel 2002; Strohmeier 2006). In a nutshell, both the majoritarian and the consensus models can be regarded as forms of government which are—in the words of Sir Winston Churchill (United Kingdom 1947)—‘the worst form of Government except all those other forms that have been tried from time to time’.