Theoretical approaches to electoral system change

Much of the scholarly literature on institutional change (or the lack thereof) assumes that rational political actors try to maximize their utility – often votes or seats. According to the rational choice perspective, parties with a lot of electoral support prefer majoritarian electoral systems and small constituencies, while parties with less support prefer PR, ideally in large constituencies – as Josep Colomer (Reference Colomer2005: 2) puts it, ‘the large prefer the small, and the small prefer the large’. Small constituencies (particularly with plurality or majority electoral rules) create high barriers for candidates, favouring larger parties (and smaller parties with highly concentrated local support), while large constituencies electing many candidates under PR rules lower the threshold and help smaller parties.

As new parties enter the electoral arena and threaten the established parties (perhaps as the suffrage is extended or new issue dimensions develop), the establishment might fear that electoral support will flow more heavily to these new parties. With majoritarian electoral rules, SMP (single-member plurality, colloquially known as ‘first past the post’ in Britain) or the two-round system, the privileged classes feared future electoral wipe-out under such winner-take-all systems once mass suffrage arrived, leading to the argument that the establishment chose PR to defend itself against the socialist threat (Boix Reference Boix1999; Grumm Reference Grumm1958; Rokkan Reference Rokkan1970). Colomer (Reference Colomer2005: 18) argues that there has been a worldwide trend towards PR as party systems have increased in size (despite the use of majoritarian electoral rules) in many countries, though if the number of electoral parties (measured by vote share, not seats won) remains relatively low, then the pressure for changing to PR will be low (Colomer Reference Colomer2005: 14).

While institutional approaches emphasizing rational assumptions are very common in the literature, some have argued that there are limits to their application to real-world cases. The rational choice approach often assumes that parties are unitary actors (Downs Reference Downs1957: 24–5), which may not be the case when factionalism is present, and that there is a lack of uncertainty when political actors make their calculations (Downs Reference Downs1957: 77). Unfortunately, in the real world, politics is fluid, so actors might not have ‘perfect information’ about their electoral context (Green and Shapiro Reference Green and Shapiro1994: 19), meaning that they can make serious miscalculations when choosing an electoral system (Andrews and Jackman Reference Andrews and Jackman2005; Katz Reference Katz2005). Gideon Rahat concludes that rational choice models do not work well with complex political situations, in which there is much uncertainty and instability, and actors are not unitary (Rahat Reference Rahat2004: 476). Indeed, it is much easier to assess actions as rational (or not) after the fact (Rahat 2006).

Choices that may appear quite irrational reflect shortcomings in the rational choice approach. Alan Renwick points out that recent (post-1962) electoral reforms in established democracies actually contradict Colomer’s (Reference Colomer2005) argument, with shifts to more proportional changes occurring in cases where party systems were smaller than where shifts towards less proportional electoral systems took place (Renwick Reference Renwick2010: 248–9). Jean-Benoit Pilet (Reference Pilet2008: 48–9) argues that, contrary to rational expectations, the larger parties of Belgium support the PR status quo because they are satisfied with the system and are cautious about the uncertainty surrounding anticipating what a major upheaval in the electoral system might entail, despite the election simulations of political scientists. Other research on the attitudes of politicians also indicates that those who win tend to be satisfied with the system (Bowler et al. Reference Bowler, Donovan and Karp2006).

Another constraint on rational action is whether the system change that is proposed by elites is actually perceived as legitimate by the wider public. Matthew Shugart (Reference Shugart2008: 10) argues that ‘it is precisely at the intersection of normative critiques of the existing rules and rational interest of political actors that reform is most likely to occur’, with political actors seeing electoral reform as a selling point. While much of the current research into electoral system choice focuses on recent case studies, this question of the democratic character of various electoral systems is quite an old one. In the early twentieth-century European context, André Blais and his colleagues argue that the transition to PR was not always a defence mechanism against socialism: the widespread ‘view that PR was the only truly “democratic” system that ensured the fair representation of various viewpoints’ also mattered (Blais et al. Reference Blais, Dobrzynska and Indridason2005: 189).

In other words, there are values beyond those linked to seeking power that motivate political actors; as Renwick (Reference Renwick2010: 37–46) argues, principles associated with democracy, stability, governance, policy outcomes, constituency service, identity and practicability may drive electoral reform. Furthermore, while electoral reform is often seen as a strictly elite affair, Renwick (Reference Renwick2011: 457) points out that the wider public might get involved in the process, noting that much reform activity takes place in between the two extremes of mass imposition (involving the public in some way) or elite imposition (where politicians can introduce the electoral system of their choice without having to take significant account of others’ views). While Renwick concludes that power-seeking elite preferences matter when studying the reasons for electoral reform, he argues that other values held by elites – and the wider public – can influence the process. In addition, Renwick points out that while Colomer’s (Reference Colomer2005) argument that places with larger party systems are more likely to adopt PR does not stand up well when analysing events in established democracies, it does apply much better to new democracies (Renwick Reference Renwick2010: 250–1).

The creation of the Scottish Parliament and ‘new politics’: two conflicting perspectives

Between 1999 and 2007, Scottish Labour and the Scottish Liberal Democrats governed together in coalitions. They had previously participated in the Scottish Constitutional Convention, during which the Scottish Parliament’s electoral system was deliberated over and ultimately chosen. The system was to embody six criteria that would reflect a ‘new politics’ distinct from Westminster’s supposedly less enlightened model:

The above criteria are difficult to maximize simultaneously, but a working group took advice from electoral systems experts, civil society groups and political parties. The Liberal Democrats wanted PR in the form of the preferential single transferable vote (STV) and Labour was reluctant to accept any form of PR, while the Scottish Greens wanted PR in the form of the additional member system (AMS), based upon the West German Bundestag’s electoral system (Jones Reference Jones1992: 76–7).

AMS is known as the mixed-member proportional (MMP) system by most political scientists, who have adopted the name used by New Zealand, which introduced this system of PR for parliamentary elections in 1996. In MMP systems, some representatives are elected in single-member constituencies while others are elected from regional or national party lists, with the latter compensating, on a partisan basis, for disproportionality arising from the constituency races so that the overall result (adding constituency and list representatives together) is proportional on a partisan basis (Lijphart Reference Lijphart2012; Reynolds et al. Reference Reynolds, Reilly and Ellis2005; Shugart and Wattenberg Reference Shugart and Wattenberg2001).

We analyse the process of electoral system choice using two broad perspectives: rational choice and ideological commitment to ‘new politics’.

Methods and data

We test the relative persuasiveness of the above perspectives using a range of sources. The main source of data is the archive of the Scottish Constitutional Convention, held at the National Library of Scotland (NLS GD489: Records of the Scottish Constitutional Convention 1989–1996). We use this to map the sequence of events and to provide an account of decision-making that could be compared with other sources. We also examine the reports released by political parties and the secondary literature about the creation of the Scottish Parliament and triangulate (Rothbauer Reference Rothbauer2008: 893) our findings from archival and secondary sources with new interviews (conducted 2014–16) with politicians and members of the Scottish Constitutional Convention (and later Commission). We interviewed Jim (now Lord) Wallace (then leader of the Scottish Liberal Democrats), Jack (now Lord) McConnell (then general secretary of the Scottish Labour Party) and Murray (now Lord) Elder (then a senior adviser to the Labour shadow secretary of state for Scotland, Donald Dewar). In addition, we interviewed two academic advisers to the Scottish Constitutional Convention (who remain anonymous) and two members of the Scottish Constitutional Commission (its chair, Joyce McMillan, and John Lawrie).

Following Rahat’s (Reference Rahat2011) advice, we use both rational choice and historical-comparative approaches in our analysis in an example of process tracing, defined as ‘the analysis of evidence on processes, sequences, and conjunctures of events within a case for the purposes of either developing or testing hypotheses about causal mechanisms that might causally explain the case’ (Bennett and Checkel Reference Bennett and Checkel2015: 7). Although much of the data analysed is qualitative in nature, some quantitative evidence, in the form of election results from the era, will be used.

The basic architecture of the Scottish Parliament was designed before the Labour government came to power in 1997, through the Scottish Constitutional Convention which was founded in 1989. It involved civil society groups, church leaders and politicians. The SNP and the then-governing Conservatives did not participate, but the Convention included Scotland’s two other main political parties, Labour and the Liberal Democrats. After taking power, Labour accepted the Convention’s recommendation of MMP for the Scottish Parliament in the Scotland Act 1998.

Labour’s calculations

As noted above, Labour could have expected, at least initially, to do very well in Scottish Parliament elections under SMP. Yet the conventional explanation of Labour’s acceptance of PR – that the party feared a future electoral shift towards the SNP that would be magnified by SMP into a seat majority – involves longer-term considerations that included an expectation of declining support and a larger effective number parties (ENP) based on vote share (Colomer Reference Colomer2005: 18). There is some evidence of both a decline in Labour’s support and a more crowded party system in the local election results from the time of the introduction of unitary authorities in 1995 (Table 1). Labour’s vote share at local elections Scotland-wide declined from 43.8% in that year to 32.6% in 2003, the last year SMP was used. Furthermore, the ENP rose from 3.5 to 4.8 in this period, as Table 1 shows, suggesting that Colomer’s argument (2005) about increased multipartyism and the fear of future defeat could have played a role in Labour’s acceptance of STV for local elections in the second Labour/Liberal Democrat coalition government after devolution.

Table 1

Effective Number of Parties (ENP), Scottish Unitary Local Authority Elections, 1995–2003 (% vote shares below party names)

Sources: Election data: McConnell (Reference McConnell2004: 139). Calculations of ENP by the authors.

The evidence supporting a rational choice explanation for PR in Scottish Parliament elections, however, is less convincing. The ENP figures from election results for Scotland’s Westminster seats, from 1970 to 1992 (the last election held before Labour agreed to PR for the Scottish Parliament), do reveal the rise of multiparty politics in Scotland. Table 2 shows the ENP in 1970 at 2.8, rising to 3.4 in both 1974 elections, dropping slightly to 3.2 in 1979, but rising again to 3.6 in 1983 and staying at this level in 1992 (despite falling back slightly to 3.4 in 1987).

Table 2

Effective Number of Parties (ENP), Scottish Voting in UK House of Commons Elections, 1970–92

Sources: Election data: Bennie et al. (Reference Bennie, Brand and Mitchell1997: 50). Calculations of ENP by the authors.

This level of multipartyism, however, is not quite up to the four effective parties that Colomer (Reference Colomer2005: 8) argues constitute the threshold for a change to a proportional or mixed system. Furthermore, while Labour’s vote share in Scotland declined somewhat during this period, the Conservatives’ share dropped considerably, and the support for the SNP and Liberal Democrats (and their predecessors, the Liberals and the Social Democratic Party) was quite volatile, as Table 3 illustrates. This parliamentary electoral context did not show Labour in any danger of imminent decline and the level of multipartyism was within the tolerance limit for the continuation of a majoritarian electoral system suggested by Colomer (Reference Colomer2005: 8).

Table 3

Vote Shares (%) for Parties Gaining More than 1% of the Vote, Scottish Voting in UK House of Commons elections, 1970–92

Source: Bennie et al. (Reference Bennie, Brand and Mitchell1997: 50).

From a rational seat maximization point of view, therefore, Labour should have insisted on SMP for the Scottish Parliament, but this was not the case. There were, firstly, practical concerns about the Scottish Constitutional Convention itself. Labour wanted to keep a broad consensus behind the proposals; as Gerry Hassan and Eric Shaw (Reference Hassan and Shaw2012: 58) point out, ‘without it there was the risk the Lib Dems might withdraw, leaving the Convention looking even more like a Labour talking shop’. Lord Elder subsequently confirmed that this was a key consideration of Donald Dewar and the Scottish Labour leadership at that time (interview, 2016). For Lord Wallace, ‘the Labour Party always knew they would have to concede the principle to get this off the ground’ (interview, 2016).

Secondly, although it was not a central concern, there is some evidence that Labour perceived that PR might prevent the SNP governing in the future Scottish Parliament on a minority of the vote (Taylor Reference Taylor1999: 56–7). Thus, for Liberal Democrat MP Malcolm Bruce, ‘There was a combination of . . . not exactly blackmail, let’s call it hard bargaining . . . plus evidence that PR would be helpful’ (quoted in Taylor Reference Taylor1999: 57). However, Lord Elder said that this was not an overriding concern for himself or Donald Dewar, then the shadow secretary of state for Scotland (interview, 2016).

However, thirdly, beyond these practical concerns there was also recognition that the party wanted the Scottish Parliament to be different from Westminster. Lord McConnell said:

Previously he had commented that ‘Gradually the Labour Party in Scotland moved in the direction of AMS because there were people who believed it was the right system and others who believed it was a compromise which could be accepted by everybody who was in favour of devolution in Scotland’ (quoted in Thomson 2009: 73). Labour conceded the principle of PR due to a variety of concerns and seat maximization considerations did not entirely trump the others. The exact form of PR that was adopted had to be subsequently hammered out through elite bargaining among the Scottish Constitutional Convention’s participants.

The Scottish Constitutional Convention and Commission

Although the principle of PR was accepted in the original report of the Scottish Constitutional Convention, there was no agreement on its precise form, so the Convention established a Scottish Constitutional Commission in 1993 to look into this matter. The Commission eventually selected MMP as a compromise between two competing forces: the commitment to a new form of politics (defined, in particular, in contrast to the majoritarian Westminster system) and the need to secure the approval of the Labour Party, for whom the adoption of PR would involve a measure of electoral self-sacrifice. Thus, MMP was the only form of PR that would be acceptable to Labour. Such a compromise may be characterized as ‘satisficing’ – it involved finding a mid-point between optimization and pragmatism (Simon Reference Simon1956). For both Joyce McMillan and Lord Elder (interviews, 2016), MMP was attractive because it retained a strong single-member constituency element. STV, consisting entirely of multi-member constituencies, appeared much more radical.

The Scottish Constitutional Convention was forged in the context of a Conservative UK government whose claims to legitimacy were under pressure in Scotland. Voting patterns in Scotland and England diverged increasingly throughout the 1980s. The Conservative vote share generally went down in Scotland while the share for other parties rose. However, the ‘union state’ structure of the UK (Watts Reference Watts2007), in which territorial diversity was accommodated through the organs of central government, meant that Scotland continued to be governed by the UK government through the secretary of state for Scotland. The SMP electoral system gave the Conservatives a UK-wide seat majority in 1979, 1983, 1987 and 1992. Through territorial management that critics found to be inept (Mitchell Reference Mitchell1990), the Conservatives were increasingly perceived as a threat to Scottish distinctiveness at a time when Scottish identity was becoming more important.

Thus, for the members of the Convention, the whole ‘Westminster’ system (including its majoritarian, confrontational elements and its electoral system) was increasingly discredited and associated with an old-fashioned style of politics that was insensitive to Scottish distinctiveness and alien to a more ‘consensual’ Scottish tradition. From the outset, it sought to define its proposals against the metonym ‘Westminster’. It also included a variety of participants determined to use the convention to further their aim of introducing greater pluralism to Scottish democracy, with women’s groups especially prominent. This was consistent with the wider aims of a ‘new politics’ that was more consensual, diverse and inclusive, a move away from the male-dominated, majoritarian House of Commons (Brown Reference Brown2000; Mitchell Reference Mitchell2000).

The Convention’s electoral system criteria, as noted above, called for equal representation of men and women in the Scottish Parliament (Scottish Constitutional Convention 1992: 91–2). Scottish Labour also demanded this, with the Scottish Labour Conference in 1990 calling for equality (Labour Party, Scottish Council 1991). There was an attempt to make equal representation for women a statutory requirement, with a member of the Convention’s Working Group on the Electoral System, Isobel Lindsay, pushing for this in her letter to the Convention’s secretary; she proposed that parties failing to nominate equal numbers of male and female candidates be denied ‘additional member’ (party list) seats (Lindsay Reference Lindsay1992).

However, in the words of one academic participant (interviewed in 2014 and who wishes to remain anonymous), there was a clash of two cultures in some of these discussions. On the one hand, all of those in the Convention were convinced that Scotland’s place in the UK on these terms was damaging and unsustainable. However, while they were rhetorically committed to ‘new politics’, they were not all convinced that PR was required to achieve it. In particular, sections of the Labour Party were strongly against the introduction of a new electoral system that would involve them diluting their electorally strong position in Scotland. Malcolm Bruce, then leader of the Scottish Liberal Democrats, noted that ‘Some of the Labour MPs and party activists seemed to be asking themselves, “Why are we dealing with the bloody Liberals?”’ (quoted in Taylor Reference Taylor1999: 54). Within Scottish Labour, a group set up by George Foulkes opposed the party’s drift towards PR, though Foulkes ended his two-year PR opposition in December 1991 (Macwhirter Reference Macwhirter1991).

The Convention, therefore, had to find a path between these perspectives in order to secure consensus. The Liberal Democrats and the Greens argued strongly in favour of PR while Labour initially opposed it. In a mailing to all Convention members, the Scottish Green Party’s pamphlet entitled Fair Votes for Scotland invoked fairness, not only in terms of proportionality, but also of women’s representation, noting that ‘the geographical link between voter and MP is retained, and smaller parties can gain representation (Scottish Green Party 1990). Scottish Labour’s executive committee saw that in the party there was ‘no enthusiasm for reform based entirely on a list system, with no constituency base and all the elected Members dependent for their success on their Party list position. Equally there is virtually no backing for the Single Transferable Vote, which is seen as destroying the link between the elected Member and his or her constituency’ (Labour Party, Scottish Council 1991). Scottish Labour’s opposition to STV put the party into conflict with its Convention partner, the Liberal Democrats, who strongly favoured this preferential system (Scottish Liberal Democrats 1993).

Due to the above differences among the parties, only the principle of PR, and not the specific form, was conceded by the Convention in 1992. Its final report contained an entire section entitled ‘Making the Scottish Parliament Truly Representative’, in which SMP was heavily criticized by references to how it was possible for a member of parliament ‘to be returned with little more than one third of the votes cast’ and that ‘Mrs Thatcher has dominated for a decade although in the most respectable of her three election victories her Party polled only 42% of the vote and a very much smaller share of the qualified electorate’ (Scottish Constitutional Convention 1992: 91). The Convention report did not decide upon a specific form of PR, however, allowing itself some leeway by calling for a system that ‘produces results in which the number of seats for various parties is broadly related to the number of votes cast for them’ (Scottish Constitutional Convention 1992: 91) and the number of MSPs was not specified.

The remaining differences over the electoral system had to be resolved by the Scottish Constitutional Commission, appointed by the Convention in October 1993. Its terms of reference included taking the six electoral system principles agreed by the Convention and then choosing an electoral system,

These three points moved on from the Convention’s principles. One major point of contention was over a statutory requirement (point (c) above) for gender equality, which nearly led to the Liberal Democrats, who opposed the proposal, walking out of the Convention in November 1993 (Clark Reference Clark1993).

The Commission reported in 1994 and noted the emerging consensus about MMP, but left the system’s mechanics open for further discussion. In particular, there were still disagreements about the size of the parliament and the balance between constituency and list members. An agreement was eventually hammered out between Jim Wallace and George Robertson, then a Labour MP and shadow secretary of state for Scotland, at the home of Menzies Campbell, then a Liberal Democrat MP, in Edinburgh. Wallace and Robertson agreed to a compromise between their different positions on size (the Liberal Democrats wanted 145; Labour wanted 112) and arrived at a parliament elected by MMP with 129 members. Wallace recalls: ‘George and I looked at each other and said “yeah, let’s just split the difference” . . . So that’s how we got 129’ (interview, 2016). This compromise was included in the final report of the Scottish Constitutional Convention (1995) and implemented by the Labour government elected in 1997. The developments in this long process are summarized in Table 4.

Table 4

Timeline of Developments in the Choice of MMP

Implications for electoral system choice theory

The analysis presented here chimes with previous studies that have emphasized the need to move beyond strictly rational choice explanations for electoral system change. MMP was chosen for the Scottish Parliament as a result of both rational choice (Labour trying to prevent a future SNP majority) and ideological considerations (the Scottish Parliament defining itself against Westminster’s majoritarian politics). We therefore categorize the choice of MMP as an example of elite imposition, where politicians can introduce the electoral system of their choice without having to take significant account of others’ views (Renwick Reference Renwick2011: 457).

Although the Scottish Constitutional Convention was an instance of a civil society-led process driving reforms beyond the influence of politicians, it implicitly accepted the requirement to secure the backing of (particularly Labour) party elites. Moreover, its membership was not randomly representative in the manner of a citizens’ assembly. Its recommendation of MMP did not emerge out of a strong desire from the public for PR. While a proportional electoral system at the UK level would have reduced some of the territorial divergence exaggerated by SMP (by, for instance, giving the Conservatives more seats in Scotland), the real desire that the Convention channelled was for some form of devolution of power. The anomaly of continuing to be governed by a party that did not have majority support in Scotland and whose vote share was declining was not to be corrected by a change in the electoral system, but by the introduction of a Scottish Parliament. Although, as the Conservatives predicted, this led to other anomalies in the British constitution, it solved the problem of Scottish domestic policies being decided by a party perceived to lack a democratic mandate.

Thus, it is difficult to argue that MMP emerged as a result of elite engagement with an electorate or civil society demanding PR. Instead, it emerged from elite bargaining about how far, on the one hand, proponents (particularly women’s groups) of a ‘new politics’ could reasonably be asked to dilute their ideas and, on the other hand, how far it was reasonable to ask Scottish Labour to engage in what many in the party saw as electoral self-harm. Labour elites wanted to go along with the spirit of the Convention; other members of the Convention, conscious of Labour’s electoral dominance and that it would likely be the party of government that came to implement devolution, wanted to help them carry their party with them. Devolution on these terms, with Labour bound into the process, was a better prospect than devolution designed solely by Labour after the party entered government. Labour was also conscious of the need to keep a broad base of support for devolution (interview with Lord Elder, 2016).

MMP for the Scottish Parliament was, therefore, an elite-imposed reform. In Renwick’s (Reference Renwick2011) terms, it is an example of a mixture of elite settlement and elite bargain. There was a two-stage process of principle and mechanics. First, Labour politicians in the Convention viewed some form of PR for the Scottish Parliament as a ‘generally good’ principle that would differentiate it from Westminster. This also coincided with some longer-term Labour thinking about the dangers for the Union of a future SNP majority, but, as we have demonstrated, this was not the primary motivation. Second, when deciding the exact form of PR, politicians thought about their own interests. By choosing MMP with its closed-list PR component, rather than the preferential STV form of PR, parties would retain more control over the candidates to be elected. They also preserved a strong constituency link that reflected their British view of how democracy should work.

Overall, the same dynamics found by scholars at the national level applied here to this sub-state case. Seat maximization is one of several priorities for political elites and may not necessarily trump other political and, crucially, normative concerns. In this case, moreover, the subsequent majority won by the SNP in 2011 demonstrates that politicians’ best-laid plans can be disrupted by the electorate and by not fully grasping the implications of PR. In choosing a sub-state electoral system, perceptions of the quality of governance at the national level also feed into normative considerations about democracy. Is the national system a model to be copied or can a diverging system be used as a badge of distinctiveness and a break with the past?