Britain’s Experience of Electoral Systems

Foreword

When Labour won the 1997 general election, its manifesto included a commitment to a referendum on how MPs are elected. Following a pre-election agreement with the Liberal Democrats, a commission was set up to choose the voting system that would be presented as the alternative to the existing First Past the Post system, but by the time the commission reported it was clear that, with a very large majority, Labour no longer had an interest in changing the voting system. In Labour’s 2001 manifesto the commitment to a referendum was replaced with a promise of a review of Britain’s experience of electoral systems. That promise was repeated in the party’s 2005 manifesto, although by that time officials in the new Department of Constitutional Affairs had already begun work on the review. The report on that review, however, has yet to be published.

But although it reneged on its referendum commitment, Labour introduced more constitutional reforms than any previous government had done and for that it is to be congratulated. These reforms included proportional voting systems for European Parliament elections, for elections to the new devolved bodies in Scotland, Wales and Northern Ireland, for the re-created Greater London Assembly and for the direct elections of mayors. This has provided us with rich experience for evaluating the performance of different voting systems and for selecting which would be best for future elections of the Commons.

It is our view that the choice of electoral systems should involve the electorate and not just the politicians. This is particularly important for the election of MPs because, while MPs have chosen the electoral systems for other tiers of government, we believe it would be wrong for MPs alone to choose the system by which they are elected.

We therefore believe that the publication of the Government’s review of electoral systems should be followed by wide, public consultation.

In preparation for the debate on the Government’s review, we have written this report with our own observations and analysis of Britain’s experience of different voting systems

This report has been prepared by Lewis Baston, Director of Research at the Society, with contributions from the Society’s Chief Executive, Ken Ritchie and with assistance from other members of the Society’s staff, particularly Stuart Stoner and Christine McCartney. We would also like to thank Paul Davies, Kieron Boyle for his assistance late on, and Tom Carpenter of Texture for undertaking another speedy and large design project. We have endeavoured to represent the policies of the Society as accurately as we can, but it should not be assumed that all opinions expressed are necessarily the views of the Society.

Introduction

The Government has promised to review Britain’s experience of electoral systems. This commitment followed Labour’s 1997 pledge to hold a referendum on how MPs are elected. No action was taken, however, on that pledge, although the Government has introduced proportional systems for European Parliament elections, elections in Northern Ireland, Scotland and Wales and for the Greater London Assembly, as well as a system other than first-past-the-post for the election of mayors. In its 2001 and 2005 manifestos, the Government undertook to review experience of new systems and stated that, if there was a case for changing the electoral system for the Commons, the correct way of introducing a change would be through a referendum.

The Electoral Reform Society believes that there should be wide debate on the Government’s review of electoral systems, when published. The Society offers this report – its own review of the experience of different electoral systems in Britain – to inform and promote that debate.

Evaluating electoral systems

Where now?

The government should redeem its 1997 promise to consult the electorate on a replacement for the FPTP electoral system to be used in the House of Commons. The case for changing the electoral system for the Commons in compelling. FPTP badly fails to meet the criteria the Government gave to the Jenkins Commission, as well as failing to meet the other criteria we have used in this report.

A constitutional change as major as the way we elect MPs should be decided in a referendum. This would be in keeping with past British and international practice. However, a process needs to be derived for choosing the system that will be offered as an alternative to FPTP in a referendum. We have considered three possible approaches:

A Speaker’s Conference is a device that has been used in the past. However, while it would involve politicians of different parties, there is a strong case for a process that involves people other than MPs so that MPs themselves are not in the position of determining the way in which they will be elected. We do not therefore regard this as a satisfactory option.

A new Commission, similar to the Jenkins Commission, is another alternative. Although much of the Jenkins Commission’s analysis stands the test of time, new evidence has emerged, particularly on the operation of MMP in Scotland and Wales. However, the reports of commissions have not always led to action by governments and there is an argument for a process that would bind government more tightly to the recommendations that are made.

There is a strong case for a process that involves representatives of the electorate, such as the Citizens’ Assembly that was set up in British Columbia. A Citizens’ Assembly is deliberative as well participatory, and enables an argument to be developed beyond the Yes/ No simplicities of a referendum. Control of the agenda of politics is just as crucial as the process of making a decision, and the Assembly mechanism turns the agenda behind the referendum over to the people. It is capable of catching the imagination, and democratising politics in a way that an official review cannot. It is capable of exploring and making accessible complex issues, in a way a referendum alone cannot.

In this report we have examined the pros and cons of the different electoral systems used, or considered for use, in Britain. These systems, which cover the realistic alternatives to FPTP, are:

• List systems, that can provide good proportionality but weaken the links between politicians and their electorates, as well as seriously reducing voter choice if closed lists are used.
• The Alternative Vote, although extending voter choice, does not guarantee a proportional outcome. It would be an inadequate reform, perhaps one not even meriting a referendum. p Mixed Member Proportional systems – including AV+ as recommended by the Jenkins Commission – have some merits. The preserve single-member constituencies, but at the expense of having ‘additional members’ with little accountablility to the electorate, and the constituency contests suffer from the same defects as FPTP. Morevoer, the experience in Scotland and Wales has shown the problems that arise from having two classes of elected members.
• The Single Transferable Vote would overcome most of the problems that have arisen with other systems:
  • It would allow voters to express their views in a simple, but more sophisticated, way.
  • It would remove the need for most forms of tactical voting.
  • It would make all elections competitive.
  • It would ensure that most votes count towards the result.
  • It would strengthen the link between members and their constituents.
  • It is more likely to lead to a more diverse range of candidates in terms of gender, age and ethnicity, and can produce broad proportionality by these characteristics should the voters consider them important.
  • It has the potential to produce a more positive and more attractive approach to campaigning.

The case for STV is strong, having been recommended by the Kerley, Sunderland and Richard Commissions and, to a degree, by the Arbuthnott Commission.

Although this report has been primarily concerned with the electoral system for the Commons, we have noted changes proposed and/or needed at other levels of government:

• Scottish Parliament: the Arbuthnott Commission has proposed changes to the MMP system in Scotland, and has recommended that, if these changes do not produce the desired results, STV should be considered. It has also recommended STV for Scottish European elections. b.
• Welsh Assembly: the Richard Commission proposed an increase in the size of the Assembly and that, if the size in increased, STV rather than MMP should be used.
• Scottish local government will use STV from May 2007.
• Welsh local government: the Sunderland Commission’s recommendation of a change to STV has remained on the table since 2002, but this may need to be revisited following the Assembly election in May 2007.
• On English local government, although the Government has made a number of proposals for improving local democracy, it has shown a disappointing unwillingness to consider the advantages reform of the electoral system would bring.
• Mayoral elections have not been well served by the Supplementary Vote – hence our proposal that it should be replaced with the Alternative Vote.
• GLA elections could be improved by a change to STV. This would also be easier for voters to use, particularly if the London Mayor were elected by SV.
• European Parliament elections: the Arbuthnott Commission has recommended that Scotland, like Northern Ireland, uses STV. We believe there is a strong case for reviewing the use of the closed list system.

We believe there is a need for a new start for politics. While we do not claim that electoral reform will resolve all the problems of political disillusionment and disengagement, we believe it is a change that is necessary. However, the process by which change is achieved is in itself important – reform should not come just arise from a deal between political parties imposed from above but should involve the electorate. We therefore call for a wide debate on the issues.

Introduction

We have produced this analysis of the performance of our electoral systems, with our recommendations for change, in anticipation of a wide, public discussion on how we elect our MPs. That consultation will, we trust, lead to a referendum offering a choice of electoral systems.

In its 2005 election manifesto, the Government promised a review of electoral systems and it subsequently emerged that work on the review was already underway even before the general election was held. That review has so far remained an internal Government one, and it has proceeded at what seems to have been a leisurely pace. While we accept that it is reasonable that the Government begin by considering internally the evidence at its disposal, we believe it essential that this should lead to a full and open debate on the issues. We have therefore produced this report as an input to that debate.

How we elect our MPs is a central feature of our democracy. It determines the composition of our Parliament and the nature of our Government; it affects the extent to which we as citizens can influence through our votes the outcome of elections; our relationship with our MPs and the extent to which we can hold them to account depends on it; and an intelligent choice of electoral system can even influence our political culture for the better. Electoral reform is therefore something that affects us all and the debate on electoral reform should not therefore be confined to corridors of Westminster.

The case for a wide consultation on our electoral system is a strong one. Although we have a representative democracy in which MPs are elected to take decisions on our behalf, it cannot be right to leave to MPs the decision on how they themselves are elected: that would be akin to asking a candidate for a job to devise the selection process. MPs have vested interests in the choice of system – different interests from candidates who might have been defeated as a consequence of the workings of the present system – and while we respect the many politicians who approach this issue from positions of principle, there is an understandable tendency for political parties to assess alternative systems on the basis of how it might affect their fortunes rather than on how they might best serve the interests of electors.

The Government has promised that any change it might wish to make to the electoral system will be put to the electorate through a referendum. We welcome this commitment, although it falls far short of the unqualified commitment to a referendum of the Government’s 1997 manifesto, and it appears to leave the critical questions of whether there should be a referendum and what the question should be entirely in the hands of Government. It is therefore particularly important that the consultation process is an open one and that any decisions arising from it are informed by, and seen to be informed by, the views of electors.

Debates on how MPs are elected are not new, and looking back at the history of the electoral reform movement demonstrates that issue is not the property of any single party – all parties, and as a consequence those who vote for them, have at some time suffered from the vagaries of the electoral system, and we all have an interest in finding a system that makes our democracy work better.

The history of electoral reform is one of opportunities not grasped. In March 2006 we celebrated the 200th anniversary of the birth of Thomas Hare who, concerned that the voting system did not offer voters effective choices and did not guarantee a voice for significant minorities, proposed reforms a century and a half ago. In the early years of the twentieth century the newly formed Labour Party sought proportional representation, as did the trade union movement, and in 1918 an opportunity was lost when the Commons overturned a Lords proposal for the use of the STV by a single vote. Again in 1930 electoral reform came close when the then Labour Government set up a Speaker’s Conference on the issue, but disagreements between the Lords and Commons prevented even a modest change to AV before the government fell in 1931.

Interest in electoral reform returned in 1974 when the Conservatives won more votes than Labour but fewer seats: Conservative Action for Electoral Reform, founded in the aftermath of the 1974 elections, at one time could count 70 Conservative MPs amongst its members. That year was also a turning point for reform in that it marked the re-emergence of the Liberals as a significant third party. The voting system which had not produced fair representation when there were only two major parties was wholly unsatisfactory in handling three. However, it was not just the growing band of Liberal – and later Liberal Democrat – MPs who were demanding reform. In the 1980s concern at what the Conservative government could do with just 42 per cent of the popular vote re-awakened the debate within Labour to the extent that in 1990 it established the Plant Commission to examine options for reform. Although the Commission was cautious in its recommendations, from the 1980s onwards the movement for electoral reform within Labour grew in strength

Electoral reform was firmly put back on the political agenda with the election of the Labour government in 1997. Prior to that election Labour had already agreed with the Liberal Democrats that a commission would be established to select an alternative system for electing MPs and that, through a referendum, the electorate would be asked whether they wanted to make the change. The commission was set up and did its job, but with a majority of 177 Labour had no need for Liberal Democrat support and the Government abandoned its commitment to hold a referendum. Labour’s 2001 manifesto promised only a review of the experiences of new electoral systems in Britain, noting that if there was a case for changing the voting system, the correct way to do it would be through a referendum. No review appeared, however, during the following four years, and Labour’s 2005 manifesto merely repeated, in a weaker form, the previous commitment to a review. Once again the opportunity for reform was missed.

However, although we cannot be other than critical of the Government for reneging on its pledge, we must also give it credit for introducing a more far-reaching package of constitutional reforms than any of its predecessors. Not only were powers devolved to the Scottish Parliament and Welsh Assembly (and indeed the Northern Ireland Assembly although that body has had a more difficult birth) but proportional voting systems were used to elect them. Proportional systems were also introduced for electing MEPs and the newly created Greater London Assembly, and a limited form of preference voting has been used in areas that have opted for an elected mayor. All of which goes to show that change can be made.

Questions of representation and electoral systems go to the heart of what it means to be a democracy. Few can fail to have been alarmed at the steep decline that has taken place in turnout in recent general elections, or of the more general decline in engagement, trust and belief in politics as a whole. It matters, profoundly. If the machine is not in working order, it is unable to produce the output that the people want and deserve.

Although Labour promised in 1996 that it would build a ‘stakeholder democracy’ in which people would be more involved in the decisions that affect them, public involvement in formal politics has, paradoxically, declined. Never in modern times had general election turnouts been below 70 per cent, but in 2001 it plummeted to 59 per cent and in 2005, in spite of a dramatic increase in postal voting and a divisive issue in the form of the war in Iraq, it only rose to 61 per cent. Opinion polls show that distrust of, and even contempt of, politicians is rising and membership of political parties has slumped. It is not that people are losing interest in political issues, but that they no longer see party politics as providing answers to the key problems facing society. This increasing disengagement of the electorate from formal politics (or perhaps the increasing disengagement of politicians from the electorate) makes a debate on our electoral system all the more urgent. We do not believe that a change in our electoral system will resolve all the problems of our democracy but, without a system that makes people feel that their votes matter, there is little prospect of improvement. Moreover, as we will argue in this review, an intelligent choice of voting system can encourage a change to a more positive and engaging political culture. Electoral reform we therefore see as a necessary, even if not a sufficient, step if we are to revitalise our democracy.

Without electoral reform there can be no guarantee that elections will produce representative government and there will be little chance of moving to a more vibrant and participative political culture. The question we address in this report, however, is what type of change in our electoral system is needed if we are to achieve not just fairness to candidates and parties but also if we are to restore the health of our democracy. We have analysed the performances of different electoral systems as objectively as we can and, as well as offering our own recommendations, we present the evidence on which others can draw their own conclusions.

We also acknowledge the considerable amount of work done by others. The report of the Independent Commission on the Voting System (the Jenkins Commission) in 1998 was a very significant contribution to the debate, and the Independent Commission on Proportional Representation (ICPR), hosted by the Constitution Unit and co-chaired by David Butler and Peter Riddell, assembled important evidence. However, since the publication of the ICPR report in 2004 there have been further developments to consider:

• Publication of the report of the Commission on the Powers and Electoral Arrangements of the National Assembly for Wales (the Richard Commission) in 2004 which noted the problems arising from the Additional Member System (AMS) and recommended that, if the Welsh Assembly were to be given more powers and an expanded membership, then it should be elected by the Single Transferable Vote (STV);
• The election of the London Mayor and Greater London Assembly in 2004 using the Supplementary Vote and AMS/ MMP respectively;
• The decision of the Scottish Parliament in 2004 to use STV for local government elections from 2007; p the 2005 general election, described by our Society as “the worst election ever” on the grounds that a government was elected with a clear majority but with only 35 per cent of the votes on a 61 per cent turnout;
• Publication of the report of the Commission on Boundary Differences and Electoral Systems (the Arbuthnott Commission) in 2006, which noted problems with the way the electoral system (AMS/ MMP) was working in Scotland, and recommended a change in the voting system for Westminster elections and a move to STV for European elections;
• The publication in 2006 of the report of the Power Commission which called for a “more responsive” electoral system as part of a much wider package of reforms.

We believe that a comprehensive review of the experiences of using the new voting systems the Government has introduced must now proceed. We call on the Government to publish its own promised review without further delay and to initiate a wide, public consultation on how our MPs should be elected in future. We offer this report – our own review of Britain’s voting systems – as a contribution to the debate that must now take place.

How to evaluate electoral systems

Elections are the method we use to select our parliamentary representatives and through them our government. Electoral systems convert votes into seats and there are many ways in which this can be done. In choosing an electoral system we therefore need to consider what we want to achieve – how we want to be represented and what sort of parliament we want.

Some may see the primary purpose of elections as being to choose the people who will represent electors in geographic (or other) constituencies; some see the aim as being to produce a legislative body which broadly represents the diversity of our society; other will focus their attention on the choice of an executive based on the policy positions of parties and their leaders as set out in party manifestos. These objectives are all important, but not every electoral system will meet them to the same extent.

What system is used?

Results

The 2005 election results are summarised opposite. Results of this and previous elections are available in more detail in a number of publications.¹For 2005, see House of Commons Library Research Paper 05/33. See also the Electoral Reform Society’s own The UK General Election of 5 May 2005: Report and Analysis. For 2001, see Electoral Commission, 2001, Election 2001. For previous elections see the long series of reference volumes compiled by F.W.S. Craig and latterly Colin Rallings and Michael Thrasher.

There were boundary changes affecting Scotland prior to the 2005 election, which reduced Scotland’s representation from 72 to 59, and the size of the House of Commons from 659 to 646. This reflected the creation of a devolved Scottish Parliament in 1999 with primary legislative powers; Scotland had previously been over-represented relative to the size of its electorate. Wales, with 40 seats, is still rather over-represented while England has 529 seats and Northern Ireland has 18.

Assessment

Participation

When the nearly universal franchise was introduced in 1918 there was a considerable fall in turnout from the last elections under the old restrictive franchise. The chaotic conditions of the end of the war in 1918 aggravated the problem and at 58.9 per cent the 1918 election holds the record for the lowest turnout (January 1910 is highest with 86.6 per cent). Turnout then rose to around 71 per cent in some inter-war elections (1922, 1923, and 1935). It was 76 per cent in other higher-temperature contests (1924, 1929, and 1931). Turnout in 1945, on fairly sketchy registration, was a creditable 73 per cent.

Turnout in Britain reached a high point in the early 1950s, with over 80 per cent voting in the elections of 1950 and 1951. It fell quite sharply in 1955 and then fluctuated around 75 per cent until the late 1990s. Elections that were close-fought or interesting, such as 1959, 1964, February 1974 or 1992 attracted more participation. Dull elections like 1955, 1970 and October 1974 attracted less. Turnout then plunged in 2001 and hardly recovered in 2005.

The chart below shows the recorded level of non-voting at each general election since 1950.

Percentage of electors not voting in general elections

In 1974 the political scientist Richard Rose produced a formula for adjusting turnout to take account of variations in the age of the register at the time of general elections. Electoral registers were based on an annual canvass of electors in October, coming into force in February. The register will therefore, under the old system, be between 4 and 16 months old when used. During this period there will be a deterioration in the accuracy of the register. A turnout figure for an October election, therefore, divides the number of votes cast by too large a registered electorate, as some of the registered names cannot vote because they are dead. Changes to the electoral registration system in 2000, notably the introduction of ‘rolling registration’ should have improved the accuracy of the register and brought turnout figures closer to the ‘real’ level.

Using Rose’s calculations, one can identify three broad periods since 1950. From 1950 to 1964 turnout was consistently high, over 80 per cent, except in 1955. There was a step change in turnout to the high 70s in 1966, with variations about 78-79 per cent until 1997. Since 2001 there has been a third period characterised by low turnout. Rose’s figures smooth out a number of variations but tell essentially the same story as the unadjusted figures.

Academic analyses of turnout at general elections tend to centre around two sorts of explanation – perceived competitiveness of the election and perceived differences between the parties. If the result of the election is perceived as uncertain or close, this attracts voters to the polls, while a foregone conclusion turns people off. Supporters of the likely winner may assume that their vote is not necessary, and their opponents may become demoralised. This explains many post-war variations in turnout – at the level of the fluctuations rather than the trend. Another factor has been how important people think it is that their party wins – what difference there is to their lives if the other party wins. The elections of February 1974 and 1992 had many voters who saw an important difference between the parties and a competitive national contest. In some elections, such as 1983, these two factors worked in different directions – there was a big difference between the parties but no doubt about who was going to win the election. 1970 was an unusual election, in that the expected result (a Labour win) did not happen and turnout sank – in part, the expectation of a Labour win deterred half-hearted Labour supporters from going to the polls.

Part of the architecture of post-war turnout patterns was that Labour and the Conservatives were both mass parties with some organisational presence nearly everywhere, and clear classbased loyalties among the electorate. There was also an ethos in most sections of society that voting was something one ought to do, out of civic duty or loyalty to one’s more immediate sense of community. In very working class Welsh valley seats there was extremely high turnout and monolithic Labour voting. This survived, with some modification, the rise of third parties in politics in 1974 and the increasing importance of national media campaigning during elections.

However, there was an unexpectedly large slippage in turnout in the 1997 election. Instead of being broadly in the middle of the range of 1955-92 turnout, it was at the bottom. Then turnout collapsed in 2001. The reasons are partly the familiar ones of perceived competitiveness and perceived difference. Few voters have expected anything other than a Labour victory in the last three elections, and comparatively few voters have regarded the result as being of great importance to their lives. But the fact that it was so much lower than before suggests that there was a ‘secular’ downward shift in participation in 2001.

The Electoral Commission sponsored some research on the reasons for not voting in 2001.

Study of the attitudes of non-voters showed that the decline from 1997 to 2001 was concentrated among people who had a weak sense of knowledge of, and connection to, politics. In previous elections many people in this position voted, out of a feeling that it was part of the duty of citizenship or from conforming to what everyone else was doing. In 2001, a lot of them suddenly stopped voting. Part of the reason may be deep and sociological, in that deteriorating social networks leave more people adrift of the sort of community identity that encourages people to vote. Part may also be the increased acceptability of apathy, cynicism and indifference in popular culture and the media.

Turnout in 2005, at 61.3 per cent, recovered a little from the depths it had reached in 2001. But there is small consolation in this being the second worst since 1918. Turnout should have increased even more under the circumstances. The polls and the dynamic of the campaign suggested more of a contest than 2001, when opinion surveys found that a foregone conclusion turned many electors away from voting. There was a large increase in postal voting, which in theory should have helped turnout. The political temperature, in the new circumstances after the 2003 Iraq war, seemed higher than in sleepy June 2001, several months before the attack on the World Trade Center. And yet there was only an improvement of 2 percentage points.

In historical terms, 2005 is clearly in the relegation zone of the turnout league table of elections in the last century. The table below gives the turnout in all the elections in the last century in which turnout has fallen below 75 per cent.

Even in relatively recent elections turnout has been much higher – 77.7 per cent in 1992 for instance. 2005’s turnout, like 2001, was about 60 per cent, in contrast to previous elections such as 1970 when a turnout of just over 70 per cent was regarded as poor. The problem of public disengagement from this most important arena of politics has clearly not been solved.

Differential turnout – who voted and who didn’t in 2005?

Before the 1980s election studies showed that social class made relatively little difference to who voted. A 1977 analysis on 1966-74 British Election Survey data by Ivor Crewe and colleagues found that social class, sex and education had no measurable effect on an individual’s propensity to vote. What mattered was age and mobility, with young people who rent in the private sector and move frequently being the least inclined to vote. A study using 1987 data from the same source by Anthony Heath and colleagues found that there was a weak relationship between class and voting, with manual workers a little less inclined to vote than non-manual workers. The differential seems to have widened at each successive election since 1987, particularly with the step-change fall in turnout in 2001.

It is notable that the 1987 election, when the class-turnout gap started to open up, was also the start of a period in which headline opinion poll findings started to become less and less accurate – inclined in particular to overstate the Labour vote. Part of the reason was sampling C2DE voters in proportion to their strength in the electorate, not among those actually voting (there having been no difference prior to 1987). Since the opinion poll debacle of 1992, the pollsters have tried all sorts of adjustments for likely voting and differential response rates, and seemed by 2005 to have got it more or less right.

(Source: MORI post-election survey)

Although the statistics are less reliable than for class, it is clear that ethnicity makes a difference to turnout. Those least inclined to vote are African, followed by Afro-Caribbean, voters (this is also correlated with social class, youth and other determinants of non-voting). The highest turnout is often found among Jews and well-established Asian communities (particularly Hindus), correlated with education and professional middle class status. Less established Asian communities (Bangladeshis) and Christian and secular whites are somewhere in the middle.

As the table above indicates, Labour voting and non-voting are positively correlated, i.e. the less a social group is inclined to vote, the more it tends to support Labour. This suggests that the current pattern of turnout disadvantages Labour – voters are skewed towards being middle class home-owners compared to the electorate. The relationship of cause and effect between this and New Labour’s political focus on the groups that do vote is an interesting question.

The age differential in turnout is also important. In 2005 the over-60s accounted for 28 per cent of the voting age population (VAP), but 34 per cent of those voting because of their higher propensity to turn out.

If turnout remains the same in 20 years’ time among each age group, demographic change will result in a radical shift in the composition of the voting population. The 2026 census projection suggests that the over-60s proportion of the VAP will rise to 36 per cent, and of those actually voting to 42 per cent. By contrast, the under-40s cast 38 per cent of votes in 2005, but on unchanged turnout patterns will account for only 33 per cent by 2026. A 4-point lead for the under-40s now will become a 9-point lead for the over-60s by 2026.

However, it is likely that turnout patterns will change because the initial assumption is that the difference is all life-cycle and no cohort. Today’s 25-year olds are 2026’s 45 year olds, so it seems likely that the present relatively healthy turnout among the middle-aged will deteriorate as the cohort made up by the current generation of abstainers moves through the life cycle. Unless something is done, or a most unlikely spontaneous revival in voting among young people takes place, it is possible that nearly half the votes in the 2025 election will be cast by those aged over 60.

There are consequences for intergenerational equity in terms of what the political process can deliver, with an ageing population and a strong turnout skew towards older voters. This might be seen in the allocation of public sector resources (with older voters having a stronger political claim on material benefits), pensions of course, in housing and planning policy (the voice of older voters in areas of planned growth will be stronger than that of the under-housed younger non-voters) and culturally (older voters are more likely to be white than younger nonvoters). The difference between population and VAP, and those actually voting, can hardly but distort policy and increase inequality to a greater and greater degree through 2025.

Differential turnout – by geography

One means of illustrating the differential turnout effect is to imagine what would have happened if turnout in every constituency had been equal but the proportions of the vote had remained exactly the same as in reality. The outcome in this case would have been a Labour lead of nearly 5 per cent in the popular vote, rather than just under 3 per cent, without any change in the number of seats for each party. This illustrates the extent to which the drop in turnout concentrated in safe Labour seats in 2001 was reflected in increasing electoral system bias rather than Labour losing many seats.

However, this applies differential turnout only at the level of the constituency. Few constituencies are composed entirely of one sector of society. Looking at election results at a smaller scale it is clear that within constituencies, the areas least inclined to vote are the poorer areas. Wealthy middle class areas turn out, deprived council estates do not. It is hard in most areas to be precise about this, because election results are not published in areas smaller than the constituency. Local elections took place on the same day in some areas, but in many of them the boundaries were not exact or there were differences between local and national voting. There are relatively few truly comparable cases.

In the Labour ultra-marginal seat of Crawley (2005 majority: 37 votes) turnout was 58.4 per cent in the general election, but was less than 50 per cent in the Labour heartland of the Broadfield ward. If turnout had been equalised across all county wards within Crawley, instead of a majority of 0.1 per cent, Labour’s lead would have been something like 1.3 per cent in the constituency. If Crawley is taken as typical, within-constituency differential turnout cost Labour something like 6 seats in 2005. This is at the extreme low end of possibilities.

Examining another seat, Burton in Staffordshire, produces a 7.2 per cent Labour lead with equalised turnout in each ward compared to 3.0 percent in reality in 2005. Burton is divided between a low turnout, safe Labour town and a Conservative, high turnout rural area. If Burton is typical, within-constituency differential turnout cost Labour something like 23 constituencies. This is more realistic, although Burton is a particularly divided seat, because a smaller unit of analysis than the county ward would probably increase the observed differential turnout effect because county wards combine several different sorts of area.

ChileDec 0587.7List PR – two member districts

Source for data: www.electionworld.org and International IDEA (CV) Compulsory voting.

This, assuming that voters are representative of the electorate nationally, means that there was no differential turnout and a full turnout would have given Labour the same 2.9 per cent lead in 2005. Assuming that voters are representative of the electorate in the constituency in which they live, Labour’s notional lead expands to 4.8 per cent if everyone voted. Assuming that voters are representative of the large county ward in which they live gives Labour a notional lead of 6.1 per cent (using Crawley as the base) to 9.0 per cent (using Burton as the base) on a universal (or even) turnout.

Turnout – international comparisons

Britain’s turnout also compares unfavourably with most other countries (see page 35). It is the lowest for a national parliament among the original 15 EU states, although in most post-communist states in eastern and central Europe there seems to be a low rate of this form of democratic participation. Britain, Canada and the USA are all notable in being English-speaking (or in Canada’s case bilingual) democracies with FPTP and particularly low rates of turnout.

Fairness / proportionality

There are four broad dimensions to the argument about fairness and proportionality in Westminster elections:

1. Government power – are governments elected on a reasonable share of the vote?
2. Is there an even playing field between aspirants to government?
3. Does the system get the right result in close elections?
4. Does the system represent significant minority parties and points of view adequately?

Does it produce representative government?

Labour won an overall majority of 66 seats, or 55.1 per cent of seats, with 35.2 per cent of the vote, in 2005.

No majority government in British history has ever rested on a flimsier base of public support – or, more accurately, none has since the extension of the franchise in 1918. In terms of active public consent for government, Britain is almost back in the pre-reform era of rotten boroughs. The table below gives the votes and seats for each of the majority governments in the last century which had less than 45 per cent of the vote at the outset.

The only remotely comparable election is 1922, although Labour in 2005 still polled 3 percentage points worse than the Conservatives did in 1922.²

It is notable that no election since 1970 has produced a government with over 45 per cent of the vote, and that the trend in the most recent elections has been to produce significant majorities with ever lower shares of the popular vote. A Commons majority has enormous power, and this power has now been awarded on the basis of only 35.2 per cent of the vote. The case for electoral reform has become stronger with each successive election.

Labour’s share of the vote in 2005 can also be compared unfavourably to the support enjoyed in past elections by losing parties. Attlee’s share of the vote in 1955 when Eden’s Conservatives won a majority of 58, comparable to Blair’s majority in 2005, was an amazing 46.4 per cent. Blair’s winning 35.2 per cent is scarcely higher than Neil Kinnock’s share of the vote in 1992 (34.4 per cent) and less than Jim Callaghan scored in 1979 in his unsuccessful bid for a third Labour term (36.9 per cent).

The government’s level of support among voters is therefore small. But taking the electorate as a whole, the proportion of eligible people who cast a vote to return the government is extremely small – only 21.6 per cent, or 9.6 million out of an electorate of 44.4 million. In terms of votes actually cast for Labour, this is the lowest total of any post1945 election with the single exception of 1983.

The table below shows the cases in the last century when a majority government has been returned with the votes of less than a third of the electorate. Again, the 1922 election (although note that this might be a misleading comparison) and the string of recent elections since October 1974 have seen record low shares of the electorate giving support to a government. Even among this company, 2005 stands out as producing a government with exceptionally few votes.

The present electoral system may allow the government to carry on regardless for a 4-5 year term, despite its low poll in 2001 and the withdrawal of enthusiasm signified by the drop in the Labour Party share of the vote of 5.5 percentage points from 2001 to 2005.

Labour’s control of the 2005 Parliament is somewhat less lopsided than its dominance in the Parliaments elected in 1997 and 2001. But Parliament is still a grossly distorted version of what Britain’s voters chose in 2005. This is all the more important given the concentration of executive power that accrues to majority governments in British constitutional arrangements.

Is it fair between the major parties?

There are two sorts of bias. One is the systematic ‘winner’s bonus’ which FPTP tends to produce. In the past this has taken the form of the ‘Cube Rule’, i.e. between the two major parties, if the votes are divided A:B, the seats will be divided A3:B3. Thus, a party with a 55:45 lead over its main rival will win seats in the ratio of 113:93, i.e. 1331:729, or about 9:5. This equates to 386 seats for the winner and 214 for the loser if there are 600 major party seats. Thus, the winner’s bonus produces working majorities from narrow popular vote leads, and landslides from more comfortable mandates.

Over the decades, the scale of this bonus has tended to fall because of the growth of third parties and a declining number of marginal seats.

However, while a defence of a winner’s bonus can be made, the other sort of bias is even more troubling. The electoral system can be systematically skewed rather than evenhanded. If it produces a lopsided majority for Labour when that party has a 40-30 lead, and for the Conservatives if the popular vote were reversed, that is one thing. But a system that produces a massive majority for Labour on a 40-30 lead (as it did in 2001) and a tiny majority for the Conservatives if they achieved the same margin in votes (as would happen on a uniform swing projection from 2005), is something else.

If the Conservatives had drawn level with Labour, with each party polling the same share of the vote (33.8 per cent), they would have gained an additional 19 seats from Labour and 3 from the Liberal Democrats, and one would flip from Labour to the Lib Dems.³ This would have meant 336 Labour seats to 220 Conservatives – an advantage of 116 seats despite equal numbers of votes. The Labour majority would have been 26 – not very comfortable, but more than the majority of 21 John Major achieved with a lead in share of the vote of 7.5 per cent in 1992.

For Labour to lose their overall majority it would require a uniform 2.2 per cent swing to the Conservatives (taking account of ‘collateral damage’ in terms of Labour losses to the Lib Dems), i.e. a Conservative lead of 1.4 per cent.

For the Conservatives to draw level with Labour in terms of seats, overcoming the deficit of 158 seats, they would need a swing of 5.2 per cent, i.e. a lead of 7.5 per cent in share of the vote, if it took place entirely from Conservative gains from Labour. However, they would be helped by collateral effects which would reduce the required swing to only 4.6 per cent, i.e. a vote share lead of 6.3 per cent.

For the Conservatives to win outright with a majority of 2, with 324 seats, they would need 126 gains. If these were all to come from Labour this would require a swing of 8.3 per cent, i.e. a national lead of 13.7 per cent. Taking account of collateral Conservative gains from other parties, the target is a scarcely less daunting 7.3 per cent swing, implying a national lead of 11.7 per cent.

The boundary changes coming into force in England and Wales make the Conservatives’ task a little easier, reducing the required swing to perhaps 6 or 6.5 per cent (and less if there is an unwinding of anti-Conservative tactical voting), but differential turnout and the spatial distribution of the parties’ votes make the electoral map very biased.

Does it get the right result in close elections?

A discussion of future possibilities may meet the objection that these are hypothetical. However, the record of FPTP in getting the ‘right’ result in close elections, i.e. giving the party with more public support a larger number of seats, is poor. There have been four elections since 1918 that have had one party or the other with a lead of two percentage points or less.

Only in one British election in the last century has such a small vote lead produced a majority sufficient to last for a full parliament, and that was in 1951. In 1929 and February 1974 the elections produced hung parliaments and governments that were unable to handle conditions of economic crisis. The simple fact is that when there is nearly a tie in the popular vote, FPTP does not reliably produce an artificial majority and stable government.

Even more anomalous is that in these four close British elections, the party with the most votes has received fewer seats than its main competitor in three of them! The ‘wrong winner’ elections included 1951, when the secondplaced Conservatives were awarded a clear majority and parlayed that into thirteen years of government. The same thing happens occasionally in other jurisdictions using FPTP, such as – spectacularly – the US Presidential election of 2000 when Gore narrowly but clearly defeated Bush in the popular vote.

The reason FPTP does this is that the actual national popular vote totals are almost irrelevant. Representation is determined by factors which have a much more debatable claim to relevance, such as the geographical distribution of each party’s vote.

Does it give fair representation to other parties?

Small parties with relatively evenly spread support are penalised under FPTP. UKIP was clearly the fourth party in terms of votes in the 2005 election, but it failed to win or even come close in a single seat. Seven parties with fewer votes than UKIP, and two Independents, elected members of parliament. The Green Party stood many fewer candidates, conscious that the system was unfair to them, and many voters were deprived of the chance of choosing them. Small parties with concentrated support, like the SNP and Plaid Cymru, did better. The failure to represent small but significant points of view contrasts the Westminster Parliament with the more proportional bodies in Northern Ireland and Scotland where minorities have a voice.

The lack of representation for minority parties affects both the largest parties as well – each of the big two is a minority party somewhere. Labour is under-represented in rural areas and the south, and the Conservatives are shut out in the big cities. This lack of national representation means that each has, socially and geographically, a skewed set of perceptions and priorities.

Constituency links

One of the advantages often claimed for single member district systems is that there is a direct relationship between the elected representative and his or her ‘patch’, making for easier access for citizens to government and better accountability. There is something to this argument, although it must be noted that the advantage of single-channel representation are more celebrated by MPs than the electorate, and it is clear that multi-member systems such as STV or MNTV (as in local authorities in England – see page 117) can produce strong constituency ties.

A factor which brings the legitimacy of the constituency links argument into question is the declining proportion of MPs who can command majority support in their own constituency. Only 34 per cent of MPs were elected with over half the vote in their own constituencies in 2005, the lowest proportion in British history. As well as the low share of the vote for the leading party at a national level, individual MPs were also elected on minority votes to a record extent. While nearly half of MPs had 50 per cent or more of the local vote in 2001, only just over a third had this level of support in 2005. This exceeded the proportion of minority mandates in February 1974 (408 MPs out of 635, i.e. 64.3 per cent), the previous high point.

53 MPs were elected with less than 40 per cent of the vote in 2005, with the lowest share being the Labour MP for Ochil and South Perthshire, who won only 31.4 per cent of the vote. Only 4 MPs exceeded 70 per cent of the vote, with the Labour member for Bootle having the highest share (75.5 per cent).

Because of the generally low turnout, as in 2001 no MPs polled a majority of the electorate in their own constituency in 2005 or even came particularly close. Only three polled more than 4 voters out of every 10 registered.

At the other end of the scale, three MPs had less than 20 per cent of the electorate.

FPTP can make constituency contests a guessing game. During the 2005 election campaign various journalists and websites attempted to advise on how voters might accomplish what they considered a desirable overall result. The results were mixed. For some voters, the choice threw up a dilemma of whether to vote for the party they really preferred, or to cast a vote for the party with the best chance of keeping out the candidate they most disliked. The pattern of 1997 and 2001 was that of tactical votes coalescing about the candidate most able to defeat the Conservatives. The partial unwinding of this effect has led to an increase in minority winners and smaller majorities. Seats such as Shrewsbury and Shipley ended up changing hands less because of changes in opinion than because of the shifting pattern of tactical voting.

Voter choice

Large parts of the country have not seen their parliamentary representation change hands since the 1970 general election (and often since long before). The urban central belt of Scotland, south Wales, the urban area between Liverpool and Manchester, the North East and the Yorkshire coalfield, stand out as areas of continuous Labour dominance. The Conservatives have exercised comparable domination in rural southern England (with the exception of the south western counties) and East Anglia.

A monopoly in representation ‘Safe Seat’ constituencies held by the same party since 1970

Large parts of the country are in effect under a long term monopoly supplier of representative services. Any form of PR, but particularly STV, would increase the choice available to the elector all over the country. It is generally considered unacceptable for the poor to be denied choice in public services and deprived of opportunities to make their voices heard. Why is it permitted for the quality of democracy and political engagement to be superior in St Albans than it is in St Helens?

The competition between the two largest parties to form a majority government has depended on a shrinking set of voters in the marginal seats who can be relied upon to vote, but whose choice can be affected by parties’ campaigns and policies. This rather small group therefore has considerable agenda-setting power and can prevent the development of policies in areas in which it is not interested, or more powerfully affect policy outcomes in areas where its interests might be affected. The wishes of this target group have forced a convergence between the main parties. Supporters of positions that might conceivably alarm this group tend to be silenced and marginalised, even if those policies are legitimate propositions for debate such as:

• Moderate (but still greater than currently takes place) redistribution of wealth and income through the tax system;
• Reducing the size of the state by cutting the scope of benefits and transferring public services to the private commercial sector;
• Increasing the level of public provision;
• Disengagement from the European Union;
• Legalisation of drugs;
• Curbing car use.

This forced convergence, at least in campaigning language, gives the impression to voters that there are no really distinctive policies and that the voting choice is basically not a very important one (which, as we have seen, drives turnout downwards). Concentration in campaigns on the issues which resonate with the atypical swing voters (interest rates, taxes, crime, immigration) bores whole new swathes of people into abstention. Voters have increasingly opted out of the choice of a government, as shown by the growth in support for third and minor parties, even when they have continued voting.

The concentration on ‘Middle England’ marginals dictated by FPTP is a distorting factor in the terms of political debate, in that the outcomes are affected and that large swathes of the country (mainly areas and people already suffering from social exclusion) are marginalised in national politics. There is a postcode lottery in democratic power, but it is worse than that – the lottery is rigged.

Other verdicts

The Jenkins Report, and our own negative findings on the operation of FPTP, have been echoed in every recent serious inquiry into electoral systems. There is a collective weight of considered opinion that FPTP is an unsuitable system for electing a legislature.

Making Every Vote Count: the case for Electoral Reform in British Columbia Citizens’ Assembly on Electoral Reform (December 2004):

However, we are convinced that the simple nature of majority governments should not override the basic values of fair election results, effective local representation, and greater voter choice. Most other western democracies do not depend on majorities, yet have stable and effective governments, governments that often are both inclusive of different interests and consensual in making decisions… We believe that our electoral system should not override fairness and choice in favour of producing artificial single-party majority governments.

We believe local representation must be a fundamental objective of any British Columbian electoral system. However, although local representation based on the FPTP system has worked in the past, it is now seen as too easily compromised in at least two ways.

Citizens wishing to support a particular party must vote for the single candidate the party offers and not necessarily for the local candidate they may prefer. This often means that the real competition is for a party’s nomination and not for the voter’s support on election day.

Party discipline quickly turns members of the Legislative Assembly into party advocates rather than local advocates. Many British Columbians now see MLAs as providing ‘Victoria’s’ voice to the people, rather than the people’s voice to Victoria.

FPTP is a simple system – voters need only place an ‘X’ beside the name of an individual. However, FPTP does not promise or provide fair election results. There is no logical or systematic relationship between a party’s total share of the votes cast and its seats in the legislature. Local candidates do not have to win a majority in their district to win a seat. In exceptional cases – for example, in British Columbia in 1996 – this meant that the party with the most votes lost the election. Governments elected with fewer votes than their opponents are not legitimate in a modern democracy.

The FPTP system can produce other undesirable outcomes. In the 2001 election, the opposition was reduced to two of 79 seats in the legislature, despite winning 42 per cent of the popular vote. Not only is this obviously unfair, it weakens the opposition so greatly that the legislature cannot hold Government to account. The very principle of responsible government, the heart of our constitution, is thrown into question. Many citizens understand that the current system is responsible for these results and believe that they are neither fair nor acceptable.

Power to the People: The report of Power: An Independent Inquiry into Britain’s Democracy (March 2006), page 189:

From the point of view of the Power Commissioners, the need to change the electoral system is not based on arguments about what might make for fairer representation but on the fact that we have now reached a point in our political history where democracy is at risk because our electoral and party system has become such a major block to popular engagement with political decision-making. The argument for change is now as much about what is expedient for the future of democracy in Britain as it is a matter of principle.

Putting Citizens First: Boundaries, Voting and Representation in Scotland (the Arbuthnott Report), Commission on Boundary Differences and Voting Systems (January 2006) para 2.17:

During the course of our work, the Westminster Election in May 2005 illustrated continuing voter disengagement. The relatively low turnout and the consequent low percentage of the overall population who voted for the winning party are causes for concern. We have not directly addressed the question of voting systems for the Westminster Parliament but do believe that the case for introducing a more proportional system for those elections is now very strong, since after 2007 they will be the only ones held in Scotland which do not involve a significant degree of proportionality. We note the research which suggests that the Scottish public shares this view.

Changed Voting, Changed Politics: Lessons of Britain’s Experience with PR since 1997, Independent Commission on PR (April 2004) para 14.46:

The question of relevance to Westminster comes down to the distribution of power. Do we want it concentrated under the winner-take-all system of First Past the Post, or do we want it spread between parties as is probable under PR?

Report of the Commission on Legislative Democracy, New Brunswick, Canada (December 2004):

If fairness or equality of votes in an electoral system is most important, for example, then it would suggest a change from our present single member plurality system which places more emphasis on effectiveness and accountability. If voter choice rates high for citizens, then a mixed member PR system allowing voters to cast two votes, one for the local candidate of their choice and one for the party of their choice, would likely come out on top…

It is clear to the Commission that the current single member plurality electoral system is not meeting the democratic values and needs of New Brunswickers. Fairness and equality of the vote, which are central to democratic satisfaction, must be given more weight when votes are translated into seats. Fortunately, it is not necessary to discard the values of effectiveness and accountability – key benefits of our current system – when making a change. The Commission’s made-in-New Brunswick, regional mixed member proportional representation system would continue to produce effective single party majority governments while maintaining the direct link between voters and their riding MLA – a link that helps keep them accountable to voters.

Voting Counts: Electoral Reform for Canada, Law Commission of Canada (2004) page 72:

The First Past the Post system performs poorly on many criteria for evaluating electoral systems, including criteria that are considered traditional strengths of the First Past the Post system. This negative scorecard conforms in large part to the majority sentiments of individuals whom we heard from during our consultation process – that it is time to seriously consider reforming Canada’s electoral system.

Prince Edward Island Commission on Electoral Reform Report (2003) page 82-83:

It has been stated that the key to the enduring appeal of FPTP is to be found in its simplicity. There is no doubt that the simplicity of the present system is a strong feature which resonates with many people. However, everyone needs to realize that this simplicity is due, in part, to the fact that the FPTP System was designed in an era when many people could not read or write and it, therefore, had to be a simple system to accommodate such conditions. Similar conditions do not exist in Prince Edward Island today…. Everyone needs to concentrate on the fact that people are now entitled to an electoral system which ensures that all citizens have an effective representation which promotes the diversity of interests and opinions found in today’s society.

Concluding note

The evidence assembled here presents an unattractive picture of the way FPTP operates in electing the House of Commons. We believe that the evidence supporting change is compelling, and FPTP’s failure on the Jenkins (and other) criteria is clear.

An alternative needs to be put in the place of FPTP for the House of Commons. But the question remains as to which alternative should go forward, and how the alternative should be decided upon. We return to these issues when we ask ‘Where Now?’in our final chapter.

Introduction

If there is any system of proportional representation favoured by the Government, it would appear to be the proportional brand of Additional Member System (AMS) known as Mixed Member Proportional, MMP. MMP was chosen for the devolved tiers of government in Scotland and Wales. The Greater London Assembly is elected by MMP and the proposal for elected English regional government also involved MMP.

We can understand the superficial attractions of MMP: for those resistant to, or at least apprehensive about, changes in the electoral system, MMP at least preserves something of FPTP. For some it might have seemed a safer and less radical option than other systems considered in this study – it can be portrayed as FPTP, with which people are familiar but with a correction factor to overcome the distorted representation that FPTP can produce. In this chapter we will look at the experience of MMP in Scotland and Wales, the benefits it has undoubtedly provided but also the difficulties that have been encountered.

This chapter is inevitably a long one as MMP in Scotland and Wales provides the most relevant British experience of systems other than FPTP when it comes to considering options for the Commons. The nature of representation and inter-party debate in the Scottish Parliament and Welsh Assembly (particularly in the Scottish Parliament because of its greater devolved powers) is much closer to the Commons than is the case with the European Parliament. While Northern Ireland’s experience of different electoral systems is longer, having different political parties and its own political culture, more care may be needed in drawing general conclusions from that experience. And although MMP has been used in London, the experience there has been in some respects different because of the different nature of the Greater London Assembly (which will be considered in chapter 9).

The MMP (AMS) system in Scotland and Wales

MMP is a hybrid of a FPTP election with singleseat constituencies and a list system in a wider region. In allocating list seats, however, account is taken of the number of FPTP seats won by each party in the region in such a way that the total seats for a party in the region is as close as possible to its seat entitlement based on the list vote.

In both Scotland and Wales, the MMP constituencies used for the elections were the Westminster constituencies, except that Orkney and Shetland (grouped in a single Westminster constituency) were given separate seats in the Scottish Parliament. In Scotland this produced 73 MMP constituencies and these have been retained in spite of the reduction in the number of Scottish Westminster seats in 2005: as a result MMP and Westminster constituency boundaries are not coterminous. The regions used for the list seats are the former (pre-1999) European Parliament constituencies – eight in Scotland and five in Wales. In Scotland most regions contain nine constituencies, but two have ten and one has only eight, while in Wales three regions have eight constituencies but one has seven and another has nine. However, the most significant difference between Scotland and Wales is in the number of list seats per region – seven in Scotland compared with only four in Wales.

Why this system?

The use of a proportional system for the Scottish Parliament was part of the devolution settlement, and indeed without the promise of proportional representation it is questionable whether there would have been sufficient support for the proposed new Parliament to get the project off the ground. Scotland’s electoral history made it clear that Labour would sweep the board in any FPTP election. The Liberal Democrats and the SNP were not going to mobilise support for a parliament that could have a permanent Labour majority and only if there was going to be a voting system that ensured a broadly representative parliament would they be willing to back it.

Labour was committed to devolution, no doubt conscious of its experience in March 1979 when a referendum on devolution was ‘won’ but with the ‘yes’ vote failing to reach the target set by the then Labour Government: the political consequences resulted in the fall of the Government. If a new referendum were to be won, it would need more than the votes of just Labour supporters, and the price of the support of other parties was the promise of a Parliament which was not Labour-dominated but in which they would have strong voices. The devolution project was therefore dependent on the use of a proportional voting system.

The choice of system was the result of long negotiations between the parties that participated in the Scottish Constitutional Convention and was necessarily a compromise between the Liberal Democrats and the SNP who would have preferred the Single Transferable Vote and Labour which, as the dominant party, was understandably less committed to proportionality. MMP was chosen as a system that, at least in part, preserved the single-member constituencies favoured by many Labour supporters.

Wales followed Scotland in adopting MMP. In contrast to Scotland, however, Wales was much more hesitant in moving towards devolution. While Scotland had always had its own institutions – and in particular its educational and legal systems – that set it apart from the rest of Great Britain, there was little other than a strong sense of national identity (important though that might be) to separate Wales from England. And while there had been significant support for Scottish devolution in all the main parties from the 1979 referendum to the 1999 election, the same was not true in Wales. A majority of Scottish voters had supported devolution in 1979, but in the equivalent referendum in Wales devolution was rejected by a margin of four to one.

In 1997, plans for a Welsh Assembly thus appeared to come in the wake of demands from Scotland – there was no Welsh body with the status of the Scottish Constitutional Convention through which a broad, supportive, cross-party consensus could be formed, and the entire project almost died at birth. While Scots supported devolution in 1997 by a margin of three to one, in Wales the referendum outcome was practically a dead heat: with a turnout of only 50 per cent, this meant that only one Welsh elector in four voted for the creation of an Assembly. However, in spite of this unpromising start, in the six years that followed the referendum support in Wales for devolution of one form or another grew significantly.³1. Table from John Osmond ‘Nation Building and the Assembly’, in Trench (ed), Has devolution made a difference?, Constitution Unit/Imprint Academic, 2004

Results

Scotland

In each election Labour has enjoyed stronger support than any other party, but has been far from having an absolute majority of the votes. As a result, the seats it has won have been well short of a majority and coalitions have been necessary to provide an Executive with a working majority. Unsurprisingly, in each case the coalition has been Labour-Liberal Democrat: parties have been reluctant to join forces with the SNP because of their policy on independence, and seemingly intractable policy differences have made the Conservatives an unlikely coalition partner. In 2003 however, even the coalition barely had a majority of seats (67 out of 129) and the partners’ combined list vote was only 38.8 per cent.

That Labour and the Liberal Democrats were able to form a two-party coalition with a majority was largely due to the advantage the system gave Labour in terms of seats. The reasons for this can be seen if we look at constituency and regional list seats separately

Wales

Labour was the dominant party in each election – as in Scotland, but more so.

As noted for Scotland, the party vote gives a better, although imperfect, measure of voter preferences than the share of constituency votes.

As in Scotland, that Labour won a much more generous share of seats than its share of the party vote merited was a result of Labour’s success in the constituencies, as can be seen if we look separately at the constituency and list results.

Assessment

Fairness / proportionality

Scotland

No party got near to half the votes, and no party got near to half the seats. Although Labour, the largest party in terms of support, received a slight winner’s bonus, it was not nearly enough to relieve them of the need to seek a coalition partner. If FPTP had been used they could have come close to two-thirds of the seats on barely a third of the votes. The Conservatives won 18 seats in each election – more or less their proportional share – while they won no seats in the 1997 general election and only one in each of the 2001 and 2005 elections.

The size of the electoral regions allowed smaller parties to win seats. In a typical electoral region of 16 seats (nine constituency and seven list) a party (or independent list candidate) with at least 6 per cent of the list vote is generally guaranteed a seat, although seats can be won with a smaller percentage as a result of ‘wasted’ votes. This made it possible for the Green Party and the SSP to win seven and six seats respectively in 2003, and a further two independent list candidates were also successful (Margo MacDonald, who resigned from the SNP shortly before the election and John Swinburne of the newly-formed Scottish Senior Citizens’ Unity Party) Additionally, two independents were also successful in constituency contests – former Labour MP Dennis Canavan’s victory in Falkirk West was not a surprise, but in Strathkelvin and Bearsden, Jean Turner, a retired GP campaigning on local health service issues, took the seat from Labour; however, these wins were a result of local factors rather than of the proportionality of the system. As a result, MMP has produced a Scottish Parliament with a very broad range of representatives.

Labour’s winner’s bonus arose primarily from regions in which it won more constituency seats than its proper regional share and consequently there were not enough list seats to compensate the other parties.

Consider, for example, the results for Glasgow Region in 2003, shown in the table opposite. The final two columns show the parties’ shares of the list votes and the number of seats they would have won on a purely proportional allocation. However, although Labour’s share should have been eight seats, it had already won ten constituency seats and there were not sufficient list seats to restore near proportionality. Thus while the SNP and SSP might both have expected three seats, they ended up with only two seats each. Labour, with 37.9 per cent of the list vote, won 58.8 per cent of the seats in the region.

Wales

As in Scotland, the use of MMP gave Wales a much more representative Assembly than would have been the case if FPTP had been used. While in each election Labour has won around half of the seats, in the constituency contests Labour won two-thirds of the seats in 1999 and three-quarters in 2003: an Assembly elected under FPTP would therefore have been dominated by Labour in spite of it polling far short of half of the votes.

We have already noted that the MMP system in Wales was not designed to be as proportional as that in Scotland. Other than the constituency seat won in 2003 by Labour defector John Marek, smaller parties and independents have failed to win any seats. In 2003 both the Green Party and UKIP won 3.5 per cent of the list vote across Wales, and 4.8 per cent and 4.4 per cent respectively in South Wales West but were

denied representation. The effective threshold in Wales is much higher than in Scotland, being around 9 per cent for a seat in South Wales West (which has 11 members) and 7.7 per cent in North Wales (13 members). In Scotland the effective threshold is around 6 per cent in the average region.

Measuring proportionality

Measuring the degree of proportionality in the Scottish Parliament and Welsh Assembly is a more difficult matter. In the constituency contests it can be assumed that votes cast do not represent the ‘true’ preferences of voters, as many would have confined their choices to potential winners rather than ‘wasting’ their votes on candidates of minor parties – this can be seen in the differences between the percentages of constituency and list votes for each party (and, as a further complication, the Green Party did not stand any constituency candidates in Scotland or Wales in 2003).

List votes are likely to provide a better picture of voters’ preferences, but here again care is needed for two reasons. First, in regions such as Glasgow and Central in Scotland and in South Wales West, some Labour voters might have recognised that a list vote for Labour would be ineffective as it could not win additional Labour seats. Secondly, there is evidence that some voters mistakenly believed that their ‘second vote’ (i.e. their list vote) was a second preference. Thus true support for Labour might be slightly higher than the list votes indicate.

However, the indices of disproportionality in the two countries calculated using shares of party votes have not been very different:

Constituency links

A principal defect of MMP is that it creates two different types of elected members – those elected in constituencies and those elected from the regional lists. This problem is examined in more detail below. In considering constituency links, we need to look at the two categories separately.

Constituency members

Constituency MSPs and AMs are elected in FPTP contests which suffer from the problems described above. Although there is the apparent simplicity of there being a single member for each constituency, and as a consequence each elector having a single constituency member, that gives no guarantee that links between constituency MSPs and AMs and their electorates will be strong.

In the 2003 election only 10 out of the 73 constituency MSPs had received the support of half the local voters. Eight of these were Labour and one each for the SNP and Dennis Canavan standing as an Independent in Falkirk West. This amounted to only 13.7 per cent of constituency MSPs. The number of majority mandates had fallen from 16 (21.9 per cent) in 1999 (12 Labour, 2 Lib Dem and one each for SNP and Canavan). As well as a shortage of strong constituency support, some MSPs were elected on very small shares of the vote. In 2003, 7 constituency MSPs had less than a third of the vote, and one had less than 27 per cent (which means that those supporting the MSP were outnumbered by supporters of other candidates by almost 3 to 1).

In Wales there were rather more elected with a majority in 2003, 15 out of 40 AMs (11 Labour, 2 Plaid Cymru and one each for the Conservatives and Liberal Democrats), or 37.5 per cent. But the trend was in the other direction from the decrease in majority mandates in Scotland – in Wales in 1999 there were only 9 out of 40 (5 Labour, 4 Plaid Cymru) or 22.5 per cent of constituency AMs with an absolute majority.

Thus across the two countries, only a small minority of constituency members in 2003 or 1999 were majority winners. It does not of course follow that all who did not vote for them opposed them – only that they would have preferred to have had someone else elected. As noted in chapter 3, people are unlikely to feel a strong link to politicians they did not want, and while those elected might enjoy the uniqueness of their positions in their constituencies, that does not mean they are regarded with respect or affection by their constituents.

Regional members

While the strength of the constituency link of constituency members might be a matter for debate, this is not the case for regional members. The latter are linked to regions comprising 7 to 9 constituencies and it is simply not possible for them to know, or be known, across their regions in the same way as constituency members. They may develop close relations with particular organisations or in areas in which they set up their offices and focus their campaigns, but for most purposes they are “additional”. Many regional members have not faced the electorate as individuals in election contests (under new legislation banning ‘dual candidacy’ – see the section on the Government of Wales Act 2006 below – none in Wales will have faced the electorate in this way) but only as names on lists submitted by parties. Consequently most electors do not know who they are, and many have only a hazy idea of their very existence.

Voter choice

As well as introducing broad proportionality, MMP has increased voter choice. Rather than voting with a single ‘X’, voters have two ‘X’s, allowing them more options and a more sophisticated vote than FPTP can offer. MMP, it is claimed, allows people to choose who they want as their constituency member as well as indicating which party they would like to see in government.

In practice it is not quite that straightforward. Constituency seats are decided through a FPTP contest, with all the disadvantages of FPTP. Many voters will have faced the choice between casting a worthless vote and voting tactically. For example, in most of the Glasgow and Central regions, constituency contests have been between Labour and the SNP (with Falkirk West as a special case). While voters can choose to vote for constituency candidates of other parties, they do so knowing that they have little chance of influencing the result. Indeed, in many constituencies in these regions it is a foregone conclusion that Labour will win, just as there are constituencies in rural Aberdeenshire (North East region), in which Labour candidates have no conceivable chances of election. In the constituency contests voters therefore have had choice, but not always effective choice, and the differences between constituency votes and list votes for the parties give at least an indication that tactical voting was widespread.

A further aspect of voter choice is the extent to which after elections people have a choice of members to represent their concerns. MMP, like other proportional systems, extends choice in this way. In Scotland, at least in theory (the practice is different for reasons discussed below), an elector can approach one of eight MSPs – their constituency MSP or one of seven regional list MSPs and in Wales each elector has five AMs (constituency plus four list). For many, approaching the constituency member may be the default option, but people preferring to deal with a member of their own political persuasion, sex or generation may choose one of their list members.

However, one of the paradoxes of MMP is that although all Conservative and SNP supporters have at least one MSP of their preferred parties, the same is not true of Labour supporters. In about a fifth of the constituencies, Labour supporters have no Labour representation at Holyrood. In most regions Labour has no list seats, and in constituencies which Labour does not win in these regions there is consequently no Labour member. In Wales, Labour did not win any list seats in 2003 and therefore have representation only in the 30 constituency seats the party won. Conservative and SNP support is spread widely in Scotland, not being focused enough to win many constituency seats but giving them representation on all of the regional lists. In Wales, the same is true for the Conservatives and Plaid Cymru.

In Scotland, Green and SSP supporters have representation in about three-quarters and two-thirds of the constituencies respectively, but the Liberal Democrats, having done much better in the constituency contests, have list seats in only four of the eight regions. Consequently, in nearly half of the constituencies Liberal Democrat supporters do not have MSPs of their party. The Welsh Liberal Democrats have a milder case of the same phenomenon, although because the system is less proportional only one party outside the big four (John Marek Independents – now ‘Forward Wales’) has any representation at all.

Stable Government

The history of devolved government in Scotland and Wales has run rather differently in each country. In Scotland there has been an Executive formed by a coalition between Labour and the Liberal Democrats from the beginning of devolution in 1999 at least until the 2007 elections. At the start of each parliament there was a period of negotiations in which the parties forged a partnership agreement containing a policy agenda for that parliament, and of course in which ministerial offices were allocated. Although there have been occasional strains in the coalition, what instability there has been in governing arrangements in Scotland has been for other reasons, notably the death of the first and resignation of the second First Minister.

In Wales the minority Labour administration that took office in 1999 was always unstable because it lacked a majority. From 2000 onwards government was put on a more stable footing with the agreement of a coalition between Labour and the Liberal Democrats, which lasted until the Assembly election in 2003. Labour achieved a bare overall majority in 2003 and formed a singleparty majority government that lasted until the resignation of Peter Law from the party in 2005. Labour continued in minority government at least until the 2007 election. The principal policy changes and innovations came during the two periods of stable government from 2000 to 2005. Wales differs from Scotland in that political culture has been more confrontational and dominated by a single party – in some ways the most unstable position is where a party is just short of an overall majority, finds it difficult to reach agreement with smaller parties, and continues to govern with majoritarian aspirations and behaviour patterns.

The problem of two categories of elected members

While it was predicted (at least by those who advocated alternative forms of proportional representation) that having two categories of elected members would create problems, the extent of the problems was unforeseen. A feature that differentiates MMP in Scotland and Wales from MMP elsewhere is that one party – Labour – has dominated the constituency results in both countries.

Thus, with very few exceptions, list members have been opposition members and tensions between the two categories have therefore been compounded by party rivalries.

Regional list members as second class members

While the principle of MMP is that all members, however elected, have equal status, that is not how list members have always been seen by their Labour opponents. For some Labour politicians opposed to proportional representation, the regional lists provide an “assisted places scheme” for the opposition. A survey of MSPs⁴Bradbury,J., Russell,M., Mitchell,J. The Constituency Roles of Members of the Scottish Parliament and Welsh Assembly: Member Types and the Politics of Constituent Casework. ESRC Seminar on the Constituency Role of Parliamentarians in Scotland and Wales: Comparative Perspectives on Devolution and Additional Member Systems, London. 2005. Table 38. found 43 per cent of constituency MSPs disagreeing that the two categories of MSPs should have “equal formal and legal status”. This view was mirrored in an opinion poll commissioned⁵‘Poll on Electoral Issues’ conducted by TNS System Three for the Electoral Reform Society, November 2005. in Scotland in 2005 by the Electoral Reform Society. It was found that:

• Only 2% of electors knew how many regional members represented them;
• 40% believed that constituency members are more important than regional members.

Thus MMP gives us two groups of members – constituency members whose constituency links are no better than under FPTP, and regional members who suffer from being only very weakly linked to the electorate.

Whatever the formal position, perceptions are clearly important to the status of list members and their political influence: that they are seen as of inferior standing by many of their electors as well as their political opponents clearly puts them at a disadvantage.

The status of list members is not of course helped by the manner of their election. With MMP the constituency contests are counted first and only when these results are known are list seats allocated. As the list seats compensate parties for their failure to win in constituencies, it is not surprising that list seats are seen as consolation prizes for losers, and although the Arbuthnott Commission has recommended a change in terminology to ameliorate the negative perceptions of list members, it is hard to see how this problem can be overcome under MMP in Britain given the predominant tradition of the primacy of constituency representation in British political culture.

List members have a further problem in that, covering a much larger area, it is harder for them to develop public profiles locally. For many list members, particularly those of the larger parties, the nature of election campaigns may add to their disadvantage: the focus of the election campaign tends to be on the constituencies with list candidates only playing a supporting role – the campaigns do not therefore give them the same public prominence as constituency candidates (although the situation of candidates of small parties standing only for list seats may be different).

Roles of constituency and list members Even if the perceptions of list candidates differ from those of their constituency colleagues, we need to ask whether they are able to perform similar roles as representatives. As legislators they are clearly equal – both categories of members are as able as each other to speak in debates and cast their votes – but as representatives of electors there appear to be differences.

Research by Bradbury and Russell found that although constituency MSPs do not necessarily receive more communications from constituents than list MSPs⁶They actually found that list members receive more communications. What is not clear from their research, however, is whether there is more overlap in communications received by list members. It is possible that a constituent deciding to approach a list member will approach all list members in the region, or at least all of a particular party., not only do they attach a higher importance to dealing with casework, working with interest groups within their constituencies and promoting business and government funding, but they also spend more time on these tasks than the average list member. They also conduct more surgeries.

The differences in the work done by constituency and list members are statistically significant, but are not huge. It might therefore be argued that this does not matter if constituents are receiving the service they want and list members, with smaller casework burdens, are able to devote more time to parliamentary business. However, it does not seem satisfactory that constituency members (who are predominantly Labour) spend more time on constituency matters (possibly giving them more opportunity to increase their profiles in their constituencies) while others can spend more time as legislators.

Interestingly, Bradbury and Russell found differences between ‘pure’ list members who had not also fought constituency contests and those who had done. The latter were found to be much more involved in work in the constituencies they had contested. This is not particularly surprising. Those who developed connections and became known in a constituency could be expected to receive more casework in that constituency, and it is natural that such list members should seek to develop their constituency bases. That, however, has led to controversies which we will now examine.

Cherry-picking and preying on constituency opponents

A major complaint of constituency members is that list members ‘cherry pick’ the casework issues that are likely to win them electoral support and that, rather than offering their services evenly across their regions, they will set up offices in the constituencies they hope to fight in the next election and effectively establish a campaigning presence in opposition to the constituency member.

For many constituency members this is naturally irritating. Lundberg⁷Lundberg, Thomas Carl, ‘Second-class representatives? Mixed Member Proportional Representation in Britain’, in Parliamentary Affairs, Vol 59, No. 1, January 2006. quotes the views of an MSP:

Unnecessary and against the intention of proportionality. Surrogate constituency MSPs were not the intention of MMP. Furthermore, the target seats are the ones shadowed. It also suggests that a list MSP is not as worthwhile as a constituency one.

However, Lundberg points out that the problem is not particular to Scotland and Wales but has also been encountered in New Zealand and Germany where ‘Losing’ constituency candidates… who end up being elected by the party list often ‘shadow’ the candidate who defeated them, setting up shop in the same constituency.

While some complaints may only be a manifestation of constituency members’ irritation at not having a monopolistic position, it is undoubtedly the case that some list members have deliberately targeted vulnerable constituency members whom they hope to depose in a future election. For example, Lundberg quotes an AM who in a 2003 survey maintained that:

“I do encounter resentment because my work enables me to threaten the constituency AM’s success at the next election” (our emphasis).

In Wales a leaked memo from a Plaid regional AM revealed her strategy in more explicit terms:

Deciding against casework as the main priority for regional AMs could mean the freeing up of staff resources … Could the AM employ someone for 2–3 days a week with the remaining time used by the party (locally or centrally) or another elected representative.

Each regional AM has an office budget and a staff budget of some considerable size. Consideration should be given to the location of their office – where would it be best for the region? Are there any target seats … within the region?

As a constituency AM for the Rhondda, Geraint Davies dealt with 2,500 cases over his four-year term. A very small proportion of those people indicated that they would be voting Plaid Cymru in telephone canvassing prior to the election. This begs the question.

We need to be thinking much more creatively as to how we better use staff budgets for furthering the aims of the party.

[Regional AMs] need not be constrained by constituency casework and events and can be more choosy about their engagements, only attending events which further the party’s cause… On receipt of every invitation, ask ‘How can my attendance at this event further the aims of Plaid Cymru?’ If the answer is ‘very little’ or ‘not at all’, then a pro forma letter of decline should be in order.⁸Leanne Wood memorandum, 2003, cited in evidence of Chris Ruane MP to the Welsh Affairs Select Committee in 2005. http://www.publicat ions.parliament.uk/ pa/cm200506/cms elect/cmwelaf/551/ 551.pdf

In such situations we must extend some sympathy to the constituency members, not because they are facing competition, but because the competition is not on a level playing field. List members have much more flexibility than constituency members in deciding where they are going to put their efforts, giving them a tactical advantage. However, to expect list members not to take account of where the returns to their work in terms of electoral support are likely to be greatest is asking politicians not to behave like politicians. Here we are not dealing so much with the misdemeanours of list members as with a serious defect of MMP.

In Scotland, but not in Wales, there has been an attempt to place constraints on list members. They have been required to demonstrate that they have conducted casework in more than two constituencies, to ask clients whether they would prefer to refer their cases to the relevant constituency member, and to inform the constituency member of cases handled in his or her constituency. To us, these are quite unacceptable requirements that undermine the equality of the two categories of members. However, as Bradbury and Russell note, they have been largely ineffective.

It also must be said, however, that not all targeting of constituencies in this way is necessarily bad. Within an electoral region it is more than likely that some constituencies will contain areas that generate more casework and political issues than others and it is quite understandable that such areas will receive greater attention from list members. Thus targeting need not always be motivated by thoughts of future election campaigns and not all complaints of constituency members are justifiable.

Participation

Scotland

Although turnout in 1999 might at the time have been regarded as a little disappointing, it was only 1 per cent less than the turnout in the devolution referendum of 1997 and was 1 per cent higher than the turnout in the 2001 general election in Scotland. However, turnout fell to only 49.4 per cent (48.4 per cent for regional list votes) in 2003.

A turnout of less than 50 per cent cannot be regarded as satisfactory, but it needs to be viewed against the dismal turnouts in other elections over the same period. As a ‘secondorder’ election, it might have been expected that Scottish Parliament election turnouts would be less than those of general elections, but the difference (49 per cent against 58 per cent) was not a huge one. Turnout was significantly higher than that of English local government elections, again as expected as the Parliament has a higher profile and greater powers (comparison cannot be made with Scottish local government as elections were held on the same day – a factor that might have improved turnout in the Parliamentary elections). Thus there is no evidence that MMP resulted in a reduction in turnout, but it would be difficult to identify any positive effects it might have had.

Turnout in most regions was around 50 per cent, the only exception being Glasgow region where the turnout for list votes was only 38.3 per cent.

As the above table demonstrates, turnout followed a pattern of being fairly consistent across all of Scotland except in Glasgow where it has been significantly lower. One might speculate that the dominance of Labour in Glasgow, both in local and Westminster elections (and now in Scottish Parliament constituency elections), has resulted in a situation where those in this Labour heartland feel little incentive to use their votes.

Wales

Turnout in Welsh Assembly elections has been more disappointing, perhaps reflecting lower levels of support and enthusiasm for the Assembly and the Assembly’s lesser powers. In the 2003 elections it was only a little more than in many local authority elections; thus far the electorate seems to see it in this light rather than as an important national election. Perhaps over the longer term a stronger role for the Assembly and members of the executive will lead to an increasing turnout in elections, but it would be foolhardy to bet on this.

Gender representation

One of the successes of the 1999 elections was the number of women elected – 46 out of 129 (35.7 per cent) and 24 out of 60 (40 per cent) in Wales. This compared with only 11 women (15 per cent) being elected in Scotland’s 72 Westminster seats and 3 out of 40 (7.5 per cent) in Wales in the 2001 general election.

In 2003 women did even better. In Scotland the number of successful women candidates increased by two to 48 (37.2 per cent) while in Wales 30 (50 per cent) were elected, making the Welsh Assembly the first legislative body in the world to achieve parity between the numbers of men and women elected

While this might be a cause for some satisfaction, further analysis shows a more complex picture. The number of women elected from constituencies and from lists in 1999 and 2003 is as follows:

In both elections a higher proportion of women were elected in constituency seats than through lists. One of the (few) arguments in favour of closed lists is that they allow parties to promote a better balance of the sexes by placing women in winnable positions on the lists. However, these figures show that women actually did better in the FPTP constituency contests than they did in lists.

Although women did well in constituencies, there were big differences in the gender compositions of parties’ successful candidates. In 2003 the percentages of party winners who were women were as follows:

It is thus clear that the high number of women elected was almost entirely due to the selection policies of the Labour Party. With the exception of Labour and the SSP, parties did not make use of the list seats to promote women. The performance of the Conservatives and the Liberal Democrats in seeking better gender representation was particularly bad.

With respect to ethnic minority communities, however, the Parliament is not representative. Not only is there a total lack of MSPs from ethnic communities, but also there were very few ethnic minority candidates and none in positions in which they were likely to be in contention, perhaps indicating a failure of the parties in Scotland to engage fully with ethnic communities.

Other issues arising with MMP

‘Wasted’ votes

A feature of most FPTP elections is the number of ‘wasted’ votes – wasted in the sense that they are either cast for losing candidates or merely add to the majorities of winners. In the constituency contests in Scotland in 2003 the level of ‘wasted’ votes was an exceptionally high 73.2 per cent.

One reason for the high number of ‘wasted’ constituency votes was the number of candidates – 64 out of 73 – who won with less than half of the votes in their constituencies. The voting system provides no assurance that these winners were the most popular candidates – indeed, with FPTP it is possible for the least popular candidate to win if the votes of those opposed to the winning candidate are spread across a number of other candidates.

While the regional lists offer a much wider choice of potential winners, even here there are minor snags. What does the intelligent Labour voter do in the Glasgow and Central regions? In 2003 it could have been safely predicted that Labour would win its share of seats in the constituencies and would not therefore be entitled to list seats: Labour voters might well have decided to use their list votes for other parties. However, even if that were the logic of the situation, over 183,000 people in these two regions in effect ‘wasted’ their list votes on Labour. Across Scotland, about two out of every five list votes were cast for regional lists that did not receive any seats, and about half of that total was the Labour vote in the Glasgow and Central regions.

Here it is worth noting a vulnerability in MMP. Some Labour politicians have suggested Labour should not contest list seats where it is strong, but should allow the Co-operative Party – or some other close partner – to do so. If most Labour list votes were to transfer as planned, Labour and its surrogate could have won, for example, 13 out of the 17 seats in the Glasgow region, converting 38 per cent of the vote into 76 per cent of the seats. This would in effect convert MMP into MMM (Mixed Member Majoritarian), in which the two modes of election are parallel rather than compensatory). This was used in Italy in the elections that took place under the mixed member system between 1994 and 2005 – constituency candidates would declare allegiance to dummy lists in order to prevent their list colleagues being penalised through the linking calculation known as the scorporo. While we prefer to believe that such manipulations lie outside accepted practice in Britain, it is a concern that at present there is no rule to prevent such fixes.

Do voters understand the system?

There is much evidence that many in Scotland lack an understanding of how MMP works. That voters might not know the details of how seats are allocated with the D’Hondt rules is perhaps not important, but that voters might not use their votes effectively is surely a cause for concern. The 1999 and 2003 Scottish Social Attitudes Surveys, conducted after the Scottish Parliament elections, found that 40% of respondents thought it “very” or “fairly” difficult to understand how seats were allocated in the Scottish Parliament.

Voter understanding was also tested in the Scottish Social Attitudes Surveys through a knowledge quiz, the results of which were as follows:⁹Scottish Parliament Election Study, 1999 and Scottish Social Attitudes, 2003, as cited in the Report of the Arbuthnott Commission p32.

Tactical voting with MMP

First Past the Post elections provide overwhelming encouragement for tactical voting, and therefore the FPTP element of an MMP election is subject to the same incentives to the voter, and the same disadvantage for smaller parties with evenly spread support. The Scottish Green Party has come to the perfectly rational conclusion that it is not worth standing candidates in the constituency contests in Scottish Parliament elections because they will not be elected. This, however, means that supporters of the Green Party are able to vote in the constituency elections for whichever candi-

date they feel will advance the Greens’ agenda most. If deployed (as it has not been in 1999 or 2003) in a calculated way, this enables the party’s supporters to have two bites at the cherry – one through the regional list vote, and another by encouraging constituency candidates of other parties to support Green ideas.

Party hopping

List members owe their election to their place on the party list and the votes cast by supporters of that party in the election. The question of how much personal autonomy a list member can and should have is a complicated one, brought to its most acute when such members change parties. Debate in New Zealand has raged over ‘party hopping’ and attempts to prevent this by legislation. There have been no issues that have developed within Scotland and Wales to a comparable extent because governing majorities have not been affected. However, the anomalies of the position were given some attention with the withdrawal of the whip from Brian Monteith MSP, who had been elected as a Conservative list member. As well as the position of the individual involved, party hopping runs against the principle of proportionality which underlies the list component of MMP by moving the composition of the legislature further away from representing the votes cast in the original election.

There is a genuine dilemma between the freedom of conscience of the representative and responsibility to the original wishes of the electorate. One way a few states have handled the issue has been to pass legislation that forces MPs out of parliament if they resign from their party mid-term. However, anti-defection legislation is relatively untested and has been passed in only a handful of states including India, Bolivia and South Africa. New Zealand provides an interesting example as it has recently banned party hopping shortly after moving to MMP.

Party hopping is nothing new in New Zealand politics. The composition of the parliament prior to the first MMP election in 1996 looked markedly different from the one voted in just three years earlier. Thirteen MPs had resigned to form new parties or to become independents. However, during the first parliamentary term under MMP, 11 MPs left their parties – not an insignificant number for a Parliament consisting of only 120 MPs – which served to undermine the principle of proportionality that the new voting system was supposed to introduce. In an attempt to maintain the proportionality of the parties between general elections, the governing Labour-Alliance coalition passed the Electoral (Integrity) Amendment Act in 2001. The Act stipulates that the seat of any non-independent MP would become vacant if the member resigned from the political party for which the member was elected. An MP can also be expelled if the parliamentary party leader and a two-thirds majority of the party membership believe the actions of the MP will distort the proportionality of political-party representation in Parliament. A constituency MP would face an immediate by-election while a list MP would be replaced by the next candidate on the party list.

The anti-defection legislation continues to be controversial in New Zealand. At issue is the balance between strict adherence to the principle of proportionality and freedom of conscience. Freedom of conscience is seen as a key constitutional right for any elected member to leave his or her party on matters of policy or principle, which is seen as an important safeguard against overly powerful party leaderships. It is for this reason that Germany, after which the MMP system was modelled, has no anti-defection legislation and freedom of conscience is written into Germany’s basic law. Against the principle of anti-defection legislation it can also be argued that it puts too much power in the hands of the party leadership, who can then use it to enforce unduly harsh discipline on elected members who after all have a duty to the nation as a whole.

It can also be argued that the law is unnecessary when one considers that every single party-hopper lost their seat in the 1999 election, which suggests that voters do not tolerate such behaviour. Although the law makes no distinction between constituency and list MPs, it inevitably has a greater bearing on list MPs. A resigning constituency MP can at least defend their behaviour at the ensuing by-election, whereas a list MP cannot.

The party hopping issue is made particularly difficult in list-based systems (including MMP), but it of course exists in all electoral systems. It is moderated under STV because at least then voters can be made aware of individual candidates’ positions and likely behaviour, so that if there are intimations that an MP has divergent views from the party mainstream constituents can choose to vote for or against accordingly.

Possible changes to MMP in Scotland: the Arbuthnott recommendations

MMP has come under pressure in Scotland as a result of changes in Scotland’s representation at Westminster. Scottish Parliament constituencies were the same as Westminster constituencies (other than Orkney and Shetland which each have a seat at Holyrood) until 2005 when the number of Scottish MPs was reduced from 72 to 59, bringing the ratio of electors to MP in Scotland more into line with England. As a result, Scotland now has two sets of constituency boundaries.

This situation is unsatisfactory from several points of view:

• Every elector lives in a Westminster constituency and a different Holyrood constituency: in some cases they may be centred on the same town, but in others they will be quite different. There is evidence that even with single constituencies many electors do not know their constituencies – living in two can only exacerbate this problem resulting in confusion among voters over who their representatives are, with implications for accountability.
• Parties organise by constituency, but now a party in a Westminster constituency must liaise with other constituency parties in arranging the selection of candidates, running election campaigns, etc. At a time when many constituency parties are very weak, this further imposition on overworked party officers cannot encourage the development of the strong party organisations which our democracy needs.
• MPs previously had to deal with a single constituency MSP (as well as with regional MSPs), referring relevant casework to them and consulting them on constituency matters. Now MPs will need to deal with at least two constituency MSPs.
• Running elections on two different sets of constituency boundaries is more complicated for electoral administrators (although this may be a less serious concern).
• Every elector lives in a Westminster constituency and a different Holyrood constituency: in some cases they may be centred on the same town, but in others they will be quite different. There is evidence that even with single constituencies many electors do not know their constituencies – living in two can only exacerbate this problem resulting in confusion among voters over who their representatives are, with implications for accountability.
• Parties organise by constituency, but now a party in a Westminster constituency must liaise with other constituency parties in arranging the selection of candidates, running election campaigns, etc. At a time when many constituency parties are very weak, this further imposition on overworked party officers cannot encourage the development of the strong party organisations which our democracy needs.
• MPs previously had to deal with a single constituency MSP (as well as with regional MSPs), referring relevant casework to them and consulting them on constituency matters. Now MPs will need to deal with at least two constituency MSPs.
• Running elections on two different sets of constituency boundaries is more complicated for electoral administrators (although this may be a less serious concern).

The government set up the Commission on Boundary Differences and Voting Systems under the chairmanship of Sir John Arbuthnott in July 2004. Its remit was to look at the consequences of having four different voting systems in Scotland (as will be the case when STV is introduced for local government in 2007) and of having different boundaries for Westminster and Holyrood elections. In particular, it was asked to make recommendations on how the Scottish Parliament is elected.

The Commission reported on 19th January 2006, recommending the retention of MMP. This was a little surprising as there had been much speculation that it would recommend a change to STV, thus allowing Westminster constituencies to be combined to form multi-member STV constituencies, thereby overcoming the problem of non-coterminous boundaries. However, the Commission took the view that, while having different boundaries for the two elections was undesirable, the problems created were ones electors and politicians could live with. Nevertheless, it recommended that if, after two elections, the more modest changes to MMP it proposed did not produce the benefits they hoped for, then STV should again be considered.

The Commission’s scope for revising MMP was constrained by decision of the government – supported by the Scottish Parliament – that the number of MSPs should remain 129. It did, however, consider two widely canvassed changes that would have retained the size of the Parliament and rejected them both. These were:

1. A proposal that each Westminster constituency should elect two MSPs (presumably by FPTP or the very peculiar Transferable Vote – SNTV). Some suggested that this could be done in such a way that one woman and one man would be elected in each. This would have given 118 MSPs and the remaining 11 could be elected from a national list. However, the Commission’s research showed that this would seriously reduce the proportionality of the Parliament which, it maintained, had been a feature of the devolution settlement. 2. A second proposal was that Westminster constituencies could be used to elect 59 MSPs and that the reduction in constituency MSPs could be compensated for by an increase from 56 to 70 list members. While this would have improved the proportionality of the system, as the major problem with MMP was its creation of list MSPs whose roles and accountability were unclear, increasing their number could perhaps have exacerbated the difficulties. It would also reduce the direct personal accountability of MSPs as a whole as long as closed lists remained.

The Commission did, however, propose two changes to MMP in Scotland: it recommended that lists should be ‘semi-open’ and that regional boundaries should be changed to take more account of local authority areas.

Semi-open lists

The switch to the method of electing list MSPs could be the much more significant change. The Commission’s arguments were (1) that voters should be allowed to choose their political representatives – they should be able to vote for individual regional candidates and not just for parties, and (2) that giving list MSPs a direct, popular mandate would improve their legitimacy and visibility. The lists, however, would be semi-open in that voters could either vote for a particular candidate on a list, or simply cast their vote for the list as at present.

Semi-open lists have not been used in the UK but are common in Europe. Votes for particular candidates count as votes for the list (as do votes simply for the list) and seats are allocated to lists as at present. A party’s ordering of its list, however, can be changed by the number of individual votes cast for the candidates. For example, a party may win three seats, but if the fourth-placed candidate on the list received sufficient votes then that person might win one of the seats rather than the third-placed candidate.

However, there are different ways in which semi-open lists can work. In the form that gives most power to the voters, if a party wins, say, three seats then the three candidates with most individual votes win the seats. But in a more common form of the system, votes cast just for the list are considered votes for the list as ordered by the party and this makes it much more difficult for list candidates to improve their position on the list through personal votes. The European experience is that large numbers of voters are content to vote for the list as presented by their party, making it relatively rare

that personal votes will affect the outcome. Semi-open lists of this type can therefore be criticised for giving only an illusion of voter choice, and can lead to apparent anomalies: the fourth candidate, for example, of a party winning only three seats might receive more personal votes than any of the first three but nevertheless fail to be elected.

The Commission considered the mechanisms of the semi-open list a detail for the Electoral Commission to devise, but it is a very important detail. Were the Electoral Commission (if it were the body charged with taking the decision) to opt for a form of semi-open lists in which votes for a party (rather than individual candidates) were taken as support for a party’s ordering of the list, then many of the benefits of the change anticipated by Arbuthnott might be lost.

While the use of semi-open lists would certainly be an improvement to MMP, what the Commission did not say is that their introduction would be a further complication to an already complex system.

Regional Boundaries

Present regions are the former (pre-1999) European Parliament constituencies, but now that European elections are held using a list system with the whole of Scotland as a single electoral area, these ‘regions’ have no relevance outside the MMP system. Adjusting boundaries to make them coincide with local authority boundaries therefore makes sense – it would link each local authority to one, and only one, group of regional MSPs.

It would, however, result in regions with different numbers of MSPs. In the scheme devised by the Arbuthnott Commission, Fife and Central Scotland would have a total of 16 MSPs while several other regions would have only 11: as a result the proportionality of the system would differ from region to region, with smaller parties facing higher thresholds to gain representation in the smaller regions. Nevertheless, thresholds would vary only between 6 and 8 per cent and this may therefore be regarded as a very small price to pay for more meaningful regional boundaries.

Possible Changes to MMP in Wales: the Richard Commission proposals

In July 2002, First Minister Rhodri Morgan established the Commission on the Powers and Electoral Arrangements of the National Assembly for Wales, under the chairmanship of Labour peer, and former Leader of the House of Lords, Lord Ivor Richard.

In its report, published in March 2004, the Richard Commission made a number of recommendations concerning the extension of the powers of the Welsh Assembly. It also reflected upon the experience of MMP in Wales since 1999. The Commission concluded that, although MMP had been successful in terms of securing broader representation than would have been possible under FPTP, it did have serious drawbacks. Principally, as has been reflected upon above, it found that MMP produced serious tensions between AMs in terms of the two classes of elected member. The Commission concluded that, with the new set of powers that it proposed for the Assembly, it would be necessary to increase membership of the Assembly from 60 to 80 AMs. Such an increase would place an intolerable strain on the electoral system, since, if the degree of proportionality was to be maintained, it would be necessary to double the number of list AMs. Tensions between constituency and list AMS could only thereby increase. As a result, Richard recommended that MMP be replaced by STV. The Commission identified the following positive features of STV:

• All elected Members are on an equal footing – being elected the same way – and have the same constituency responsibilities;
• It encourages a genuine contest in every constituency;
• Multi-Member constituencies could be created relatively easily by grouping Westminster seats, or by using local authority boundaries;
• It is straightforward for voters to operate: the system works smoothly in Australia, the Republic of Ireland and Northern Ireland; the Scottish Executive plans to introduce STV for local authority elections in Scotland and the system was recommended by the Sunderland Commission for use in local authority elections in Wales;
• It maximises voter choice (between candidates of different parties or of the same party, or candidates with no party label) and the incentive to vote and campaign;
• Constituents have a choice of elected representatives to approach with problems;
• Few votes are wasted: voters know that their second preference will help elect someone – should their first choice not be elected;
• It creates opportunities for independent candidates – because the electorate vote for the individual, not necessarily a party;
• It creates opportunity for diversity;
• More or less every vote counts equally and the result is broadly proportional – but the degree of proportionality is affected greatly by constituency size;
• Because individual Members do not have ‘safe’ seats, STV increases their accountability to their constituency.
• All elected Members are on an equal footing – being elected the same way – and have the same constituency responsibilities;
• It encourages a genuine contest in every constituency;
• Multi-Member constituencies could be created relatively easily by grouping Westminster seats, or by using local authority boundaries;
• It is straightforward for voters to operate: the system works smoothly in Australia, the Republic of Ireland and Northern Ireland; the Scottish Executive plans to introduce STV for local authority elections in Scotland and the system was recommended by the Sunderland Commission for use in local authority elections in Wales;
• It maximises voter choice (between candidates of different parties or of the same party, or candidates with no party label) and the incentive to vote and campaign;
• Constituents have a choice of elected representatives to approach with problems;
• Few votes are wasted: voters know that their second preference will help elect someone – should their first choice not be elected;
• It creates opportunities for independent candidates – because the electorate vote for the individual, not necessarily a party;
• It creates opportunity for diversity;
• More or less every vote counts equally and the result is broadly proportional – but the degree of proportionality is affected greatly by constituency size;
• Because individual Members do not have ‘safe’ seats, STV increases their accountability to their constituency.

The Government of Wales Act 2006

However, the Labour Party was not minded to accept the Commission’s recommendations – either in terms of its model for increased powers for the Assembly, or in its proposed changes to the electoral system. The Government argued that an increase in AMs was neither necessary nor desirable, thereby making a change in the electoral system unnecessary. Instead, in 2005’s Welsh Labour manifesto, and in the Better Governance for Wales White Paper that followed, it set out the case for a more limited increase in powers, through Orders in Council, and a solitary change to the electoral system – a ban on ‘dual candidacy’, which would prevent candidates standing for election both in a constituency and on a party list.

Over the following months, a heated debate ensued as to the merits or otherwise of the proposed reform. Although such a proposal was not, in and of itself, particularly controversial, the political circumstances in Wales meant it proved incredibly divisive. The electoral arithmetic, whereby Labour is dominant in the constituency contests under MMP, meant that the change would have minimal effect on Labour (who stand little chance of winning more than one or two list seats at most), whilst putting all the opposition parties at a disadvantage. Unsurprisingly therefore, the proposal was met with a chorus of opposition from Conservatives, Liberal Democrat and Plaid Cymru politicians alike. However they were joined in this opposition by a host of independent academics and experts, including the Electoral Commission, the Electoral Reform Society, and the aforementioned Arbuthnott Commission. Meanwhile, the head of the respected Constitution Unit thinktank, Professor Robert Hazell, described the change as “nasty, spiteful, and seemingly driven by partisan motives.”

The Arbuthnott Commission’s conclusions were particularly telling. The dual candidacy debate, although less intense than in Wales, had emerged as an issue in Scotland also, where, as we have seen, Labour also enjoyed a dominant position in constituency contests (and a corresponding lack of representation amongst list MSPs). Yet the Arbuthnott Commission reached the following conclusions:

• There is no survey evidence to suggest that dual candidacy is an issue for voters, or a disincentive to their participation in the political process. Few of our consultation responses raised dual candidacy as an issue, nor was it raised spontaneously in our focus groups.
• It may encourage parties, particularly small ones, not to field strong candidates in constituency seats, where they have less chance of success, keeping them instead for the region where they would be more likely to be elected. This could have a negative effect on the quality of constituency contests and unduly favour incumbent candidates. Barring dual candidacy could also create tensions between constituency and regional candidates within the same party, since the chances of a regional candidate being elected would be enhanced when candidates from his or her party perform poorly in the constituency election. It might therefore be in the interests of a regional candidate of any party to see colleagues lose constituency elections.
• Dual candidacy is a common and accepted feature of mixed member proportional systems across the world – indeed, in some cases candidates are expressly required to stand in both contests.
• The Commission believes that preventing dual candidacy would be undemocratic and agrees that it would place ‘an unnecessary restriction on the democratic rights of potential candidates, parties and local electors to have as unrestricted a choice as possible in an election.’

As these quotes indicate, particular offence was taken at the manner in which the Government set out the case for change, claiming that there was widespread dissatisfaction with the system as it was, and that it wasn’t ‘right’ that a losing constituency candidate could be elected by the ‘back door’. Yet, as the Arbuthnott Commission perceptively observed:

We suggest that dual candidacy only seems problematic to some people here because of the legacy of constituency representation within British political culture and the hegemony which this has secured for some parties. Candidates coming in second or third place who are then elected through the regional list are only ‘losers’ in the context of a first past the post, ‘winner takes all’, electoral system. This logic does not sit well within a proportional system and introducing it devalues and undermines the concept of proportionality. The criticism, and the pejorative terms in which it is sometimes put, does little to enhance the legitimacy of regional MSPs.

This indeed touched on the nub of the issue – the change was only brought forward because Labour AMs were uncomfortable with the level of competition (admittedly, as discussed, on an uneven playing field) that MMP creates. Yet the case for change was further undermined when Labour backbench peer and former MP Lord Foulkes put forward a private peer’s Bill proposing the self-same change for the Scottish Parliament. The Government did not support the Bill, saying there was no prospect of such a change at the present time. This only served to highlight the fact that the proposal, in the Welsh context, was partisan in the extreme, in its effect if not its intention. At the present time, Labour has four list MSPs, whereas it has no list AMs at all. The Government were happy to push a measure through that would have little if any impact on the party in Wales, but where it would have some negative impact (however limited) as in Scotland, no such proposal was made.

The subsequent legislative debates, during the passage of the Government of Wales Act, therefore left a sour taste in the mouth. A Lords amendment removing the ban was passed, only to be rejected by the House of Commons. When Liberal Democrat peers agreed not to continue their opposition in return for concessions on the formula for deciding membership of Assembly Committees, the dual candidacy ban was finally agreed to and the Bill passed into law. 2007’s Assembly elections will be the first held under the ban.

The irony is that, in spite of all the controversy over the proposal, its effect is likely to be minimal. List members will still be able to compete with constituency members, parties will still be able to encourage their list members to target particular constituencies, and a list member will even be allowed to stand as a constituency candidate next time around, so long as he or she does not stand on the list. The only real impact of the change is to force the opposition parties to choose whether to risk their best candidates in constituency contests, or whether to ‘play safe’ and put them on the list. The ban is unlikely to achieve its stated aim, namely to reduce the tension between constituency and list AMs. Indeed, by depriving list members of any direct involvement in constituency contests it could arguably create even more problems. The only way to solve such tensions is by electing all AMs (or MSPs for that matter) on the same basis. If proportionality and a constituency link are to be retained, then STV is the only solution, as the Richard Commission concluded.

Be that as it may, further reform looks unlikely in the present political climate. The Government of Wales Act was consciously designed as the ‘final stage’ of the development of the Welsh Assembly, at least in the short term. Liberal Democrat and Plaid Cymru amendments calling for the introduction of STV were defeated during the progress of the Act, and neither Labour nor the Conservatives have shown much interest in reviving the Richard Commission’s proposals. Should the Liberal Democrats or Plaid Cymru be in a strengthened position after the 2007 Assembly election, or, further along the line, after the next General Election, then the case for STV may be reexamined. In the meantime, the reformed MMP system, for all its faults, remains in situ.

Conclusion

Underlying a lot of Arbuthnott’s caution about further changes to the electoral system was a consciousness that the Scottish Parliament was a young institution and that there was merit in not making too many changes too soon to the founding settlement under which it had been established. While there might be some logic in this position, after a third MMP election in 2007 one can start making conclusions about whether further change is desirable. Arbuthnott indicated that the case for replacing MMP with STV might need to be reviewed in the future – there is therefore a constitutional ‘alarm clock’ to restart the debate. What Arbuthnott could not resolve without a move from MMP is the complication of two sets of constituency boundaries and the organisational strains arising from them will not be easily overcome. The experience of local government under STV will no doubt inform future discussion of the arrangements for the Scottish Parliament.

In Wales, however, the strains within MMP seem to be even more acute. Political culture remains heavily adversarial, there are high barriers to entry for parties other than the big four (and Independents who flake off from those parties) and a lack of consensus on the basic principles of the system. Conflict between mainly Labour constituency AMs and mainly opposition regional AMs has been more intense and partisan than that found in Scotland. Although Labour’s working majority in 2003 was won with 40 per cent of the constituency vote and 36.6 per cent of the regional vote, it appears that many Labour members still see themselves as the natural party of government in Wales and have therefore been less willing to accept a more pluralist approach. There is a pressing need, as there is not in the same way in Scotland, to heal the conflict between different sorts of members and make the system work in a more proportional and consensual way. Lord Richard’s report in 2004 remains the most coherent statement of a way forward.

What system is used?

The Single Transferable Vote (also known in Ireland as ‘proportional representation by the single transferable vote’) is the most commonly used system in Northern Ireland. There is broad consensus around its use for Northern Ireland elections among most political parties, the British and Irish governments and voters. STV is used for electing local authorities, the Northern Ireland Assembly and Northern Ireland’s three MEPs. FPTP is used for electing members of the House of Commons.

How the system works

For the Assembly, six Members of the Legislative Assembly (MLAs) are returned for each of 18 constituencies, corresponding to the 18 Westminster constituencies. Councillors on Northern Ireland’s 26 local authorities are elected in local electoral divisions (composed of several ‘wards’) which return between five and seven councillors each. For the European elections, Northern Ireland forms one constituency which returns three MEPs.

Each Assembly Member declares a substitute who will take the seat should it become vacant between elections. This does away with the need for by-elections. On local councils, a byelection is held using STV for a single seat (i.e. the Alternative Vote).

All representatives are elected on the same basis and, outside of election time, little attention is paid to which candidate topped the poll or came last.

Why STV? Origins of the system

In many ways, STV is Northern Ireland’s native electoral system, used to elect the first Northern Ireland Parliament at the formation of the state in 1921 and again in 1925 before the system was scrapped to help the Unionist Party retain its dominance. In the early 1970s, Northern Ireland’s institutions again returned to STV and it has been used for elections to local government, European Parliament and the various Northern Ireland-wide assemblies since. The exceptions are Northern Ireland’s Members of the House of Commons who are elected by FPTP and the 1996 Northern Ireland Forum which used a list-based system.

Since the 1970s it has been broadly accepted that any solution to the Northern Ireland situation cannot rest on majoritarian rule, but must provide fair representation to minorities and foster power-sharing. This is not just the view of politicians and peace-makers: the 2003 Northern Ireland Election Survey found that 84 per cent of recipients supported power-sharing between the communities. STV is understood in Northern Ireland as a system which guarantees the representation of significant minorities, whereas FPTP is perceived to under-represent minorities with unhelpful consequences for inter-communal relationships. As Donald Horowitz noted “To both nationalists and unionists, [STV] seems a natural choice for an accommodative regime.”¹⁰Donald Horowitz, 2002 ‘Explaining the Northern Ireland Agreement: The Sources of an Unlikely Constitutional Consensus’ in British Journal of Political Science, Vol. 32, Part 2.

The 1996 elections to the Forum used a novel electoral system: a constituency-based list PR system with a Northern-Ireland-wide top-up of two seats for the ten largest parties. This institution was elected for a very specific purpose: to choose delegates to the peace talks. The priority was to ensure that all significant opinions were represented around the talks table and particularly to ensure that the small loyalist parties could access negotiations. Every party represented in the Forum was entitled to a place at the table, while the Forum itself met in public with no powers and debated issues related to Northern Ireland’s future. The system was not designed to achieve a stable executive or to provide constituency representation, but it was intended to ensure that all significant parties were able to send delegations that could speak with authority on their behalf.

The negotiations provided the opportunity to evaluate and consider the most appropriate system for the new democratic institutions that would be set up under a future agreement. The Forum formed a Standing Committee on Electoral Reform, which considered the question of a suitable electoral system for a new Northern Ireland Assembly and also made a submission to the UK-wide Jenkins Commission. Party submissions to the committee doubtless reflect their negotiating positions in the behind-closed-doors peace talks.

Every party which made a submission recommended a proportional representation system; none felt that First-Past-the-Post would best serve the interests of Northern Ireland. Five out of seven parties recommended STV or a modified version of STV (small parties calling for the retention of the small top-up used in the Forum elections). The DUP did not rule out STV, but suggested an Additional Member System as proposed for Wales and Scotland. The Ulster Democratic Party was not specific about what kind of PR system they preferred.

The parties therefore were in broad agreement that STV should form the basis of Northern Ireland’s future electoral system. However, disagreement over the details threatened to hold up agreement in the early hours of Good Friday morning. The smaller parties which had helped to form the Agreement were concerned that they would be frozen out of the new institutions because they needed 16.7 per cent of the vote in one constituency to guarantee election. This was a particularly high hurdle for parties with dispersed support. As a compromise, the larger parties to the talks agreed to increase the size of the STV constituencies from five seats to six, lowering the threshold from 16.7 per cent to 14.3 per cent. The Assembly therefore has 108 members (many more relative to population size than the Scottish Parliament or the Welsh Assembly). This proved to be a sufficient guarantee that all those with significant support would be elected in 1998.

In addition, the Forum’s Standing Committee on Electoral Reform made a submission to the UK-wide Jenkins Commission, proposing that STV be adopted for elections to Westminster throughout the UK.

Assessment

Is there fair representation?

STV is not a pure PR system, in that the number of seats is broadly rather than exactly proportionate to voters’ party preferences. Part of the reason is that while some seats are determined by first preference votes cast alone, the last few seats tend to be the result of transfers made to lower preference candidates. This allows the voters views to be taken into account more fully: if their first preference candidate is eliminated, they can still influence the outcome of the election with their secondand lower-preference votes.

Politically speaking, some parties will be more ‘transfer-friendly’ than others; we can see this working against Sinn Fein, for example, and in favour of smaller cross-community candidates who pick up transfers from throughout the political spectrum. Northern Ireland’s voters remain reluctant to transfer across the political divide between nationalist and unionist candidates. Unionist politicians often benefit from intra-unionist transfers – so once a voter has given their top preferences to all the candidates from their preferred unionist party, they will transfer to other unionist candidates, then in some cases to small-party or independent candidates. On the nationalist side, some SDLP voters can be more reluctant to transfer to Sinn Fein.

Despite some variation in the number of electors, each constituency elects six Assembly Members. The results would be slightly more proportional if the number of MLAs varied between five and seven, but there is little demand for such a change.

With all these factors, the relationship between seats and votes in the Northern Ireland Assembly is fairly close.

In 2003, the DUP, the UUP and the Alliance Party all fared slightly better in terms of their proportions of seats than in their proportions of the vote (2.2, 2.3 and 1.8 percentage points difference respectively). Sinn Fein and the SDLP did slightly worse (1.3 and 0.3 percentage points difference respectively). The relationship between vote share and seats was similar in 1998. On the unionist side, the DUP had a more proportionate result while the UUP were the main beneficiaries of the system, winning almost 5 per cent more seats than was strictly proportionate. The Alliance Party and Sinn Fein were very slightly disadvantaged.

On the whole, small parties were not more adversely affected by STV than the larger ones.

In fact, the Alliance Party profited from a well-managed vote to secure six seats (5.5 per cent) on only 3.7 per cent of the vote in 2003, although in 1998 they could have expected one more seat under a PR list system. The Progressive Unionist Party was slightly disadvantaged in both years. In 2003, the Northern Ireland Women’s Coalition lost their two Assembly seats when their vote dipped to 0.8 per cent, although most strictly proportional systems would not have guaranteed them representation either.

In 2003, no party with more than 1 per cent of the Northern-Ireland-wide vote was denied representation and in 1998 the largest vote share which did not result in any seats was 1.07 per cent, which is a low threshold. The fears of small parties about being excluded by STV were not realised. Although voter behaviour changes under different systems, if first preference votes are an indication of party support, small parties would only have done better under List PR in one or two cases. At a constituency level candidates with around 10 per cent of the vote could be reasonably hopeful of securing enough transfers to gain a seat. In 2003, the candidate not elected with the highest share of first preference was a Sinn Fein candidate in South Antrim who received 11.5 per cent.

In 2005, Northern Ireland’s electorate cast votes for both Westminster MPs (under FPTP) and local councillors (under STV) on the same day, providing an opportunity to compare the results under the two electoral systems in terms of proportionality. The graphs below compare the votes and seats each party won in the two elections.

It is clear that the differences between votes and seats is much more pronounced in the general election (yellow) than council (black). The DUP won half of all Northern Ireland’s Commons seats on just one third of the vote, while the UUP won only one seat, despite receiving nearly 18 per cent of the vote. If the UUP had won seats in proportion to its vote, it would have been entitled to three seats. Sinn Fein were a little advantaged by the FPTP system, winning five seats with a little less than a quarter of the vote. The SDLP held their three seats, while candidates from other parties and independents secured 40,000 votes and no seats.

While there are some small differences in the local council election, to the advantage of the DUP and UUP and to the disadvantage of Sinn Fein and smaller parties, the ratio between votes and seats is much closer in every case than in the general election.

It is also clear, looking at the results, that some voters cast their vote differently in the two elections. In the local council elections, 45,000 voters deserted the DUP and Sinn Fein, at least two-thirds of them opting for a different party. The SDLP also saw a fall in their share of the vote at the local council level.

Some voters may have chosen to ‘lend’ their votes on a tactical basis to leading candidates from their community in the Westminster constituency contests. Others may be evaluating the significance of the two institutions and may feel more inclined to elect a higher profile candidate for their Commons constituency.

However, it is clear that some voters were aware that voting for their preferred party in the FPTP election would be of no value. The combined vote for small party and independent candidates in the local council elections was 11.9 per cent, while in Westminster, it was just 6.6 per cent (1.7 per cent of the province-wide vote going to an independent runner up in West Tyrone standing on a healthcare platform). The Electoral Commission’s report into the 2005 Election in Northern Ireland also noted that: “voters differentiated between the UK

Parliamentary election and the local elections in determining how to cast their vote”, choosing individual candidates or issues for the local poll and choosing a party for the general election, despite the higher profile of the Westminster candidates. When this effect is taken into account, it is possible that the general election results may be an even less accurate reflection of the wishes of Northern Ireland’s people.

Votes and seats in the 2004 European Elections

Unlike the rest of the UK’s representatives, Northern Ireland’s MEPs are elected by Single Transferable Vote, with the whole province being one constituency electing three representatives.

As there were only three seats, the relationship between seats and votes is somewhat artificial. It is clear that STV with smaller constituencies is not as proportional. The margin between the thirdand fourth-placed candidates was a mere 0.7 per cent, though the Ulster Unionist candidate received a vast majority of the Democratic Unionist transfers and also a majority of transfers from the three eliminated candidates.

Participation

Traditionally, turnout by Northern Irish voters in general elections has been below the UK average, and only looks healthier in the past couple of elections when compared with a decreasing UK-wide turnout. This is perhaps not surprising when we consider that Northern Ireland’s voters cannot influence the formation of the UK government as they elect only 18 members of the Commons and cannot cast votes for the leading UK parties.¹¹Neither the Labour Party nor the Liberal Democrats organise in Northern Ireland while the Conservatives are not a serious political force and have no representatives. (An unresponsive electoral system may also encourage them to stay at home, especially in safe seats and where voters from one community are in a clear minority.

The biggest contrast with the rest of the UK is the enthusiasm of voters in Northern Ireland for other elections, namely: local council, European and Northern-Ireland-wide institutions:

• European Election Turnout between 1979 and 2004 averaged 33 per cent in the UK as a whole and 55 per cent in Northern Ireland. Despite the distance of the institutions and the limited impact that Northern Ireland’s three MEPs can make on these international institutions, a majority of voters still turn out to vote. It is true that Northern Ireland voters may be more aware of the impact of Europe on their daily lives, with very visible examples like the European peace funding programmes and infrastructure investment.
• The Devolved Institutions in Scotland, Wales and Northern Ireland were established around the same time, yet voters in Scotland and Wales were less inclined to turn out to elect their Parliament and Assembly. All three institutions saw a decline in turnout between the inaugural election and the second poll which took place in 2003. In Scotland there was a 9 per cent drop, in Wales an 8 per cent drop and despite the fact that the Assembly had not met for a year and faced a continuingly uncertain future, the Northern Ireland turnout dropped only 6 per cent.

and Wales were less inclined to turn out to elect their Parliament and Assembly. All three institutions saw a decline in turnout between the inaugural election and the second poll which took place in 2003. In Scotland there was a 9 per cent drop, in Wales an 8 per cent drop and despite the fact that the Assembly had not met for a year and faced a continuingly uncertain future, the Northern Ireland turnout dropped only 6 per cent.

• Even in Local Government, a majority of Northern Ireland’s voters continue to turn out and cast their votes, despite the fact that local councils have far fewer powers and smaller budgets than English local authorities. When local elections coincide with general elections, a tiny proportion of Northern Ireland’s voters record a national vote without filling out the local ballot (as in the rest of the UK). When local elections were held on their own in 1989 and 1993 the turnout was 56 per cent in Northern Ireland, compared to 39 per cent and 37 per cent respectively in the English county council elections.

There are a number of factors which may account for turnout in Northern Ireland and differences with the rest of the UK:

• The nature of the institution being elected and its perceived impact on people’s lives.
• The political context, particularly the sense of political crisis, the relative likelihood that stable politics will result from the election and the perception of what is at stake.
• Grassroots party activity in Northern Ireland tends to be much higher: 55 per cent of people surveyed after the 2005 election had a canvasser from a political party call at their house, for example. In addition, local media coverage of election campaigns is greater.
• Safe seats are largely unheard of in most Northern Ireland elections.
• The perceived efficacy of voting: the public’s confidence that casting a vote will influence the outcome of the election and that the electoral system will correctly deliver what the voters want. Northern Ireland voters in general feel relatively positive about the efficacy of their vote. Almost two-thirds agreed with the statement ‘Voting is the only way people like me can have any say about how the government runs things’. Twice as many felt a sense of satisfaction at having voted as felt dissatisfied (52 per cent to 27 per cent).
• The political context, particularly the sense of political crisis, the relative likelihood that stable politics will result from the election and the perception of what is at stake.
• Grassroots party activity in Northern Ireland tends to be much higher: 55 per cent of people surveyed after the 2005 election had a canvasser from a political party call at their house, for example. In addition, local media coverage of election campaigns is greater.
• Safe seats are largely unheard of in most Northern Ireland elections.
• The perceived efficacy of voting: the public’s confidence that casting a vote will influence the outcome of the election and that the electoral system will correctly deliver what the voters want. Northern Ireland voters in general feel relatively positive about the efficacy of their vote. Almost two-thirds agreed with the statement ‘Voting is the only way people like me can have any say about how the government runs things’. Twice as many felt a sense of satisfaction at having voted as felt dissatisfied (52 per cent to 27 per cent).

Lastly there is the extent to which every Northern Irish election, whatever the institution, is seen in terms of the constitutional question. This has a mobilising effect on both parties and the public, and gives significance or even urgency to turning out and voting, no matter what the poll is. In referendums on the future of Northern Ireland and those elections which determine a mandate to negotiate or to govern Northern Ireland, this dynamic is more apparent. Even with this motivation, voters and parties must have some faith in the voting system to communicate their views on this issue.

It is worth noting that recent trends in Northern Ireland elections have not been affected by technical changes such as the greater use of postal voting, nor has the fall-off in the recorded electorate caused by the introduction of individual registration changed turnout noticeably.

4. 8.7 hours to 5.1 hours, according to the Northern Ireland Local Government Association, 2005, National Census of Local Government Councillors in Northern Ireland 2005.

The MP-constituency link

STV leads to a different kind of link between constituents and their elected representatives because of its multi-member constituencies. Each voter has between five and seven local councillors and six Assembly Members representing their area.

In every constituency and most District Electoral Areas, this will mean a number of different political parties or independent candidates will be elected to represent the same area. In Northern Ireland, this is particularly beneficial as various sections of the electorate may be particularly reluctant to approach a representative from the other community. For some this may be seen as a negative thing – permission for representatives to ignore those sections of the community that they are less familiar with and concentrate on their own patch. For others, it is certainly a benefit, as members of the public have the opportunity to approach a representative that they helped to elect and who they best identify with. Most unionists may prefer to go without the support of an MP for their case if that MP represented Sinn Fein for example.

In submissions to the Forum Standing Committee on Electoral Reform, a number of parties mentioned the benefits of the constituency link under STV as a plus for that system. In oral evidence, the Ulster Unionist Party called for ‘a PR system that maintains a strong connection between the individual and the constituency and enables the individual to cope with the constituency work’, indicating that STV would be their preference for such a system. The Alliance Party argued that:

STV not only preserves, but actually enhances constituency representation. Since a range of representatives are elected from multi-member constituencies, voters have the choice of approaching a member with whom they feel the most comfortable.

Only the DUP expressed a concern that STV might somewhat dilute the nature of the relationship between voters and representatives, compared with the Westminster model.

Critics of STV raise two conflicting claims about constituency representation: that multi-member constituencies weaken representation, and that they lead to the duplication of casework. In terms of local government, Northern Ireland’s councillors do spend more time on casework than their English and Welsh counterparts¹²8.7 hours to 5.1 hours, according to the Northern Ireland Local Government Association, 2005, National Census of Local Government Councillors in Northern Ireland 2005. although it is difficult to read much into the comparison given the different sets of functions of local authorities in each area.

Diversity of representation?

In Northern Ireland, top of the equality agenda among political parties, public bodies and the general public is fair representation for both the Protestant and Catholic communities. This means that equality for other, under-represented, groups has sometimes been sidelined or seen as less crucial. For example, policing reforms included the provision to recruit 50 per cent Protestants and 50 per cent Catholics to the policing service. Catholics were significantly under-represented in the service (8 per cent in 2001), as were women (13 per cent in 2001). Thanks to the special measures for recruiting Catholics, the proportion had more than doubled by 2004. Without a similar gender quota system, the proportion of women also increased, but at a much slower rate.

In terms of identity, STV results in accurate representation of all communities and centre-ground parties which have sufficient levels of support (over 1 per cent nationally or around 10 per cent in one constituency). In their 1997 submission to the Forum committee, the Alliance Party highlighted that STV was a good system for guaranteeing everyone was represented in a divided society:

In divided societies, the requirement for proportionality is essential to ensure that no significant group is neglected by the electoral system. Proportionality is an essential part of any power-sharing approach to decision-making.

Northern Ireland has a poor record in terms of women’s representation, and this is true across all electoral systems that have been used in recent years. Currently, three of Northern Ireland’s 18 MPs are women. Considering that the election of these three women ended a thirty-year period when not a single woman was elected as an MP, it is clear that the recent record of Westminster representation is poor. Similarly, no women were elected to the European Parliament between 1979 and 2004, with one woman being elected in that year.

On Local Councils, just 21 per cent of councillors are women: the same rate as in Scotland and Wales (where councillors are elected by FPTP). There are 18 women in the 108 member Assembly (16.7 per cent), slightly higher than the first Assembly which had 14 women. This is an equivalent rate to the Northern Ireland Forum elected by List PR, which had just 15 women out of 110 members. Only one out of every nine delegates elected on the constituency lists was a woman, while women made up only four of the 45 candidates at the top of regional lists (in addition to the Women’s Coalition, which only promoted female candidates).

Whatever the opportunities afforded by electoral systems, Northern Ireland has few women in politics. Possible explanations include the particularly combative nature of politics in a divided society; perceptions of negative treatment of women in politics and the lack of importance attached to this issue by political parties. No party in Northern Ireland has ever used positive action to select candidates for election (though Sinn Fein has used gender quotas for internal party positions).

Northern Ireland has no Assembly Members or local councillors from ethnic minority communities. Ethnic minorities make up 0.85 per cent of the population. The age profile of local councillors is much broader in Northern Ireland, with 12 per cent under 35 years of age, compared to 4.3 per cent in Wales.¹³Welsh Local Government Association, 2004, National Census of Local Authority Councillors 2004

Government stability

In the Northern Ireland Assembly, the Executive is not formed on the basis of coalition discussion. Instead ministerial positions are allocated using the D’Hondt mechanism to allocate at least one ministry to every party with over seven seats. This was designed to create an inclusive government and to encourage power-sharing between parties, though there are a number of difficulties in its operation. This somewhat artificial government formation process means there is no direct relationship between the electoral system and the kind of government that is formed.

While the Northern Ireland Assembly and wider peace process has been beset by difficulties, it is not apparent that STV has had a part to play in this. In local government, despite the lack of trust in Northern Ireland’s political culture and the sharp differences between the parties, stable arrangements have usually been formed and there has been co-operation on local issues.

There was a hope that STV would encourage the growth of the centre ground in Northern Ireland, but instead it has coincided with its demise. Political crises and logjam have eroded confidence in the Agreement and its institutions and people have expressed their scepticism at the ballot box, although this decline has been grossly exaggerated in Westminster FPTP elections. However, the parties on the extremes have themselves moved closer to the centre. The key feature of STV is that it is a responsive election system: if Northern Ireland’s voters were to vote for centre ground parties or to vote cross-community, the election results would reflect that. In terms of voting patterns, Northern Ireland’s voters have proved reluctant to cross the community divide in transferring their votes (each stage of the STV count provides a wealth of information about voting preferences).

Alternatives

Fewer STV members

The main debate in Northern Ireland centres around the number of Assembly Members there should be in the future, with some of the larger parties seeking to reduce it again (a view which has gained popular support due to the considerable costs of running a suspended Assembly). Alongside this, the current Review of Public Administration will reduce the number of local councils in Northern Ireland and the total number of councillors.

Alternative Vote

Two commentators, Robin Wilson (Democratic Dialogue) and Rick Wilford (Queen’s University) have questioned whether STV in six-member constituencies has in fact encouraged parties to concentrate only on their community, because they can win sufficient support from a section of the electorate with less effort. They recommend a switch to the Alternative Vote, because every candidate would need a majority of votes or transfers to be elected. AV could result in parties appealing to the lowest common denominator in their constituency and would totally remove minority representation. The Alliance Party has responded that AV ‘would be disastrous in Northern Ireland as it would further polarize election results … [and] should be avoided in a divided society as the elections essentially become sectarian head counts.’

MMP

The Democratic Unionist Party has seen some advantages in a mixed member system. Its submission to the Forum suggested this as a way of freeing some prominent party members who would not want to be ‘burdened down with constituency work’. The idea has attracted little support from others, and in the current context might advantage the larger political parties, including the DUP. Drawing boundaries and the detail of electoral system design may prove very contentious in the Northern Ireland.

MMP elections can use systems other than FPTP as the base for electing the constituency members. If AV were to be used it would probably have to be accompanied with a proportional top-up to ensure that all significant parties are represented, which seems an arduous way of replicating the proportional and preferential features of the existing STV system.

Another possibility would be a proportional topup for an STV system with a smaller number of members per constituency, to ensure that the smaller parties with widely spread support are not squeezed out. While this has some merits, it is essentially trying to replicate the effects of six-member STV at the cost of introducing additional complications and two tiers of representatives. In the Northern Ireland political context, the temptations to manipulate a mixed-member system by running stooge independent candidates or artificially distinct list and constituency slates may be too great to resist.

Reforming Westminster elections

In February 1974 the anti-Sunningdale Unionists won 11 seats out of 12 with 51 per cent of the vote, which was a body blow to the emergent power-sharing system.

In 2005 the FPTP elections for Westminster polarised representation around the two more hard line parties, the DUP and Sinn Fein. The electoral system magnified the victory of the DUP and the fall of the UUP. The DUP was the luckiest party in UK politics, in that it has the fewest votes per MP of any of them, while the UUP was treated worst by the system of any party that won seats. FPTP also, to a much lesser extent, helped Sinn Fein in its contest against the Social Democratic and Labour Party (SDLP).

The DUP’s victory in half the Northern Ireland seats rested on only a little over a third of the votes. While the result ended up roughly proportional on the nationalist side, it distorted the voice with which the unionist electorate of Northern Ireland spoke. For every two DUP voters, there was one UUP voter, but this 2:1 ratio of votes became a 9:1 ratio of seats.

The result in individual seats can also be peculiar. The SDLP’s victory in Belfast South was caused by a split in the Unionist vote between the UUP and DUP (who stood for the first time here in a recent Westminster election); a nationalist now represents a majority unionist constituency. This is the reverse of what happened in West Tyrone in the 1997 parliament, when a UUP MP represented a constituency with a heavy nationalist majority split between SDLP and Sinn Fein.

Using STV for Westminster elections would harmonise voting in Northern Ireland, so that every election would be conducted using the same system. It would also enable Northern Ireland’s MPs, collectively, to represent the views of the voters of the province more accurately. A model STV result in 2005 would have given a fairly exact proportional result, with the DUP winning 33.3 per cent of the seats for 33.7 per cent of the vote, and the others each winning 22.2 per cent of the seats for 17.5, 17.8 and 24.3 per cent of the votes.

Even in the absence of wider electoral reform for House of Commons elections, there is a case for changing the system of election for Westminster seats in Northern Ireland to STV. It may be objected that this means having two systems used to elect members

What system is used?

An appointed European Parliament existed prior to 1979, but since June 1979 the institution has been directly elected with a five year term. Agreement between the nine member states to hold direct elections was reached in September 1976, and in July 1977, the Callaghan Labour government introduced the European Assembly Elections Bill. The Bill proposed two alternatives of a regional open list PR system and a First Past the Post system of election, with STV to be used in Northern Ireland. The Bill fell because of lack of time but was reintroduced in November 1977. The Government recommended the adoption of a regional list PR system. However, on a free vote, the House of Commons voted for First-Past-the-Post to be used. The Bill became law as the European Parliamentary Election Act 1978, and the first elections to the European Parliament were held under First-Past-the-Post in Britain and STV in Northern Ireland in June 1979. Subsequent rounds of European elections were held under this arrangement in 1984, 1989 and 1994.

Since June 1999 European Parliament elections in Britain have been conducted using the closed list system of proportional representation. There have been two elections under this system.

There have been 75 MEPs elected from Britain, plus three from Northern Ireland, since the 2004 elections. In 1999 there were 84 British and three Northern Irish MEPs but the number was reduced after EU enlargement in 2004. Elections for British MEPs take place in eleven large constituencies – England is divided into nine regions and Scotland and Wales elect one group of MEPs per nation.

The number of MEPs per electoral division is related to population, so that the smallest region (North East) has three, and the largest (South East) has ten.

The voter must choose a single party and signify approval of the whole list with a cross. The votes are counted at regional level and once the regional result is finalised the distribution of seats begins. The parties are allocated seats in turn according to the D’Hondt formula, and candidates for each party are declared elected in the order in which they appear on the party list.

If vacancies arise during the term of the European Parliament, they are filled by the next-placed member of the list of the party of the MEP who has died or resigned – there are no by-elections.

Why was the system changed in 1999?

Article 138(3) of the Treaty of Rome contained a commitment to introduce a uniform electoral system for the European Parliament, and the UK was a party to this undertaking from her accession to the European Community in 1973.

Reform of the European Parliament electoral system was a Labour Party policy commitment prior to 1997. Following a vote at its Annual Conference in 1990, the Labour Party established a Working Party on Electoral Systems, under the chairmanship of Raymond Plant. The Plant Report recommended that a proportional regional list system should be adopted for European elections, in the light of the “growing influence of the regional idea” and to establish a more uniform electoral procedure.1

The Report concluded that either open or closed lists would be the best system. The Plant proposals were accepted at the 1993 Labour Party conference and the March 1997 Cook-Maclennan Joint Consultative Committee on Constitutional Reform reached agreement between Labour and the Liberal Democrats on a regional list system.

In October 1997, the European Parliamentary Elections Bill received its Second Reading in the House of Commons. The Bill provided for the MEPs for each region to be elected by a closed regional list system, along the lines of that used in France, Germany, Greece, Portugal and Spain.

The then Home Secretary Jack Straw set out the rationale behind the Bill:

There is a world of difference between a constituency that one can comprehend in terms of its size – such as Crewe, Blackburn and similar places or parts of cities, which broadly accord with the majority of communities – and the vast constituencies that form the basis of representation in the European Parliament, where the direct connection between the Member and his constituents is very much more tenuous…

A list system is used to elect Members of the European Parliament in every other member state of the EU, apart from the Republic of Ireland [sic – Northern Ireland does not either]. The particular simple list electoral system that the Bill would introduce is the same type of system as is used in France, Germany, Greece, Portugal and Spain, so nearly two-thirds of EU citizens already elect their MEPs in this way, and the figure would rise to over 70 per cent if Britain also adopted that approach…

The system that is set out in the Bill has the great virtue of simplicity. As now under the First-Past-the-Post system, the elector has to mark a single cross on the ballot paper. As I mentioned, simple list systems of that sort are already used by nearly two-thirds of voters in the European Union to elect their MEPs, so we are in reasonable company…

We are not a country that can lecture other European nations about turnout. Generally speaking, turnouts are higher in Europe. One argument in favour of change is the fact that turnouts are so low. It is difficult to argue that the current electoral system has gripped the electorate. It has not. My guess – and it can be only a guess – is that turnout will improve under the proposed system.

We said in our manifesto that we would introduce a proportional voting system for elections to the European Parliament. The system that we are proposing has many advantages. It is appropriate for the nature and functions of the body that is being elected. It will be simple for the electorate to use. It allows independent candidates to stand. It produces fair results so that in each region, a party or independent candidate will win a share of the seats proportionate to the share of the vote in that region. The Bill is another major step forward in the modernisation of our constitution. I commend it to the House.

The proposed use of closed lists proved the most controversial element of the Bill, arousing criticism from opposition parties in the Commons and the Lords. The government considered a Belgian-style ‘flexible’ or ‘semiopen’ list system early on during the Bill’s passage, but decided against on the grounds that such systems would often mean that candidates supported by large numbers of voters would still fail to be elected ahead of their colleagues further up the list, and this would discredit the system among voters. The personal votes for candidates in such large seats, it was argued, would probably not be significant anyway.

In October 1998 the House of Lords voted for a fully-open list system as used in Finland, an amendment that was not accepted by the government. The Bill was then subject to ‘ping-pong’ between the Houses, until on 19th November the Lords refused to back down for the fourth time, and the Bill was lost. However, the Leader of the House of Lords, Baroness Jay, told peers that the Bill would be introduced in the next (1998-99) session under the Parliament Acts. It had to achieve Royal Assent by mid-January 1999 in order for the new system to be implemented in time for the 1999 European elections. The new Bill, identical to its amended predecessor, passed all its Commons stages on 2nd December. The Bill was then rejected outright by the House of Lords on 15th December, and so in accordance with the Parliament Act became law in January 1999. The new closed regional list system was used for the first time at the European parliamentary elections of 10 June 1999.

Results

Results of elections in 1999 and 2004 (GB)

How has reform affected participation?

The introduction of list PR for the European Parliament elections cannot be said to have transformed participation. In the four elections between 1979 and 1994 under FPTP, turnout averaged 34.2 per cent in Britain, and in the two since reform it has averaged 30.7 per cent. However, the 1999 election attracted an exceptionally low turnout, taking place during a period in which turnout in general fell to very low levels, while the 2004 election saw the highest turnout of any European Parliament election in Britain to date. Taken over a longer period, European Parliament elections in Britain are unusual in having seen an increase in turnout between 1979 and 2004.

The low turnout in 1999 in Britain certainly lent no support to the idea that PR in and of itself automatically leads to higher turnout in all circumstances. However, as the Department for Constitutional Affairs quantitative survey of the election found, ‘95% surveyed said that it [the new electoral system] had no effect on their decision to vote.’ The new system led to neither an increase nor a decrease in participation, which was strongly correlated with age and a sense of duty to vote. Among the reasons for the low turnout was that ‘voters seemed generally uncertain of the function of the European Parliament and also were largely unaware of why the European Parliamentary elections were being held.’¹⁴http://www.dca.go v.uk/elections/euro/ euro99/13.htm.

The context of the 1999 election, in which political debate was becalmed in comparison with the climate in the two previous elections (in 1994 the election was perceived as a test of John Major’s survival as Prime Minister), also reduced interest. The sharp drop in turnout between 1994 and 1999 was mirrored in the fall in turnout in the general election of 2001, so there were also general trends working to reduce turnout.

The 2004 election, with the highest turnout of all Great Britain’s European elections, had two unusual features that encouraged participation. The first was the postponement of the May 2004 local elections until June so that polling day coincided with the European Parliament election. The second was the widely publicised use of all-postal ballots in the four northern regions of England. Postal balloting led to a significant increase in turnout in and of itself, but the publicity generated by its problems drew nationwide attention to the fact that the election was taking place. The 1999 elections had attracted next to no publicity.

Bodies with a scrutiny function sometimes find it difficult to connect with the electorate because public experience has been shaped by legislative and policy-making bodies such as Parliament, traditional local authorities and now the devolved assemblies. The idea of scrutiny in the European Parliament and local government is not a familiar one to the public, however much of a role it plays in good government.

Very few MEPs have been household names – only Barbara Castle, Glenys Kinnock and Robert Kilroy-Silk have made much of an impact on the public, and in each case they were well known before they entered the Parliament.

In summary, the effect of list PR on turnout in European Parliament elections cannot be said to be very powerful, although it seems to be positive rather than negative. The European Parliament is rare in being an institution whose elections attracted higher turnout in the mid2000s than it did in the late 1970s. The reasons for low turnout in European Parliament elections are for deeper reasons to do with the lack of knowledge or understanding of the institution involved, its perceived lack of importance relative to other elected bodies, and the general factors that caused particularly low turnout in the 1998-2003 period.

Assessment

How fair is the European Parliament electoral system?

The 2004 Euro election result faithfully reflected how people voted. The proportional system allowed different party representation in each region, gave representation for the surge in voting for UKIP, allowed the Liberal Democrats’ distinctive position on Europe to be represented, and did not award a ‘landslide’ when none existed. A comparison between the real result and how it might have looked under FPTP, or how results looked before 1999, reveals just how seriously FPTP can distort an election.

Figures in the table in brackets show the actual distribution under List PR. The other figures are the estimated seats won under FPTP on a notional distribution of single-member seats.

Some properties of the notional FPTP result include:

• The votes cast in the European Parliament election were notable for producing the first instance where the combined forces of Conservative and Labour fell below 50 per cent. This would not have been reflected in seats, given that FPTP would continue to award them 71 out of 75 seats (95 per cent).
• The Conservatives would gain 45 seats, that is 60 per cent, on 26.7 per cent of the vote.
• The 14.9 per cent who voted Liberal Democrat would receive no representation at all.
• Neither would the 6.2 per cent who voted Green.
• The 16.2 per cent who voted for the UKIP would be represented by only two MEPs (3 per cent), one of whom abandoned the party before 2004 was over.
• Representation of the two main parties would be polarised by region. The Conservatives could claim to speak for the South, and Labour for the North, Scotland and Wales, despite the dominant parties’ very low shares of the vote.

Before 1999 FPTP in European elections often produced very disproportional results. When either of the two main parties was well ahead, as the Conservatives were in 1979 and Labour were in 1994, the result was often a large majority of seats and the main competitor being reduced to a regionally specific rump. 50 per cent of the vote won the Conservatives 77 per cent of seats in 1979, and in 1994 Labour won 74 per cent of seats with 44 per cent of the votes. In 1979 Labour won two seats in London, none in the south and east outside London, and only one seat in the Midlands (Nottingham). In 1994 half of the remaining 18 Conservative MEPs represented seats in the South East, with only one in the Midlands, one in the North and none in Scotland or Wales.

Under FPTP, parties other than the main parties had extreme difficulty in winning European Parliament seats, except for the SNP whose support was concentrated in the Highlands and Islands. In 1989 the Green Party won 15 per cent of the vote but no seats and in 1979 the Liberals won 13 per cent of the vote but no seats. In each case the SNP share was 2-3 per cent but the party was represented in Strasbourg.

Voter choice

The lack of voter choice of individual candidate is one of the most obvious features of closed list PR, and it has proved one of the most controversial and unpopular. However, as Jack Straw noted in introducing the Bill that brought in the system:

To reinforce the point, lest that is somehow presented as a radical departure, let us not forget that that is essentially the way in which hon. Members are chosen. Candidates are selected by the parties, and voters have no influence over who the representative of their chosen party will be in their constituency. If they do not like the representative of the party, they have to vote for a different party. That is the simple way in which the system operates and the way in which this system will operate.

Single-member FPTP is essentially a system in which parties present a list of one in each constituency, and voters have no choice over who that might be. Introducing closed list PR did not harm voter choice of candidate. It has enhanced voter choice by enabling minority parties to make a realistic appeal for votes in all areas. In 2004 every voter had between eight and 13 options to choose from, with a wider range of parties having a realistic chance of winning representation than before. The range of choice at the lower end of the scale in 2004 would have been generous for a FPTP constituency in 1994.

Are there implications for government stability?

Government stability is not a relevant criterion for the European Parliament election. The European Parliament does not have the function of sustaining a government, and the British delegation forms a relatively small part of that body. There is no argument at all to distort the pattern of representation in order to provide an artificial majority. As Jack Straw noted:

We are dealing not with the election of a Parliament which then sustains a government, but with the election of a representative body in Europe. The role of a Member of the European Parliament is different from that of a Member of Parliament, yet they are both directly elected by the same method.

The United Kingdom returns only 87 Members to the European Parliament and, quite apart from the disconnection between the constituencies and the Members, the huge constituencies that we currently have mean that disproportionate results are greatly magnified. As a result, even a small swing in the vote can lead to a substantial difference in the number of seats, as the fluctuating composition of the UK delegation to the European Parliament over the past 20 years has demonstrated…

Under FPTP, the distortions caused by the British electoral system could affect the entire balance of power in the European Parliament. The Party of European Socialists (PES) was disproportionately under-represented in 1979 because of Labour’s poor showing, while the big swings in representation in Britain in 1989 and 1994 inflated the strength of the PES relative to the centre-right grouping.

How has the electoral system affected constituency links?

MEPs work predominantly in specialist subject committees and their main functions are scrutiny of European institutions and legislation. Their role generates less casework than other levels of government, and their focus of attention is usually on Europe-wide issues. The implications of European policies at a local level are often handled by MPs rather than MEPs, given the interactions of European and national policies, although there is still a constituency dimension to the work of MEPs. However, this often involves liaison with organised interest groups rather than individual constituents.

The change in 1999 involved moving to multimember regions rather than single member constituencies as the basis of representation and this has affected the way representation works.

Some politicians and commenters in 1999 regretted the loss of single-member constituencies in the European Parliament, but there is very little evidence that there was a great deal of individual accountability, or voter knowledge about the identity and record of their MEP. One must be very cautious about any claims made for the virtues of the single-member constituency link that existed prior to 1999. Each of these constituencies had over 500,000 electors and contact between the MEP and individual electors was rare. Votes in European Parliament elections were predominantly cast on the basis of party rather than the personal record of an MEP.

The FPTP constituencies, as well as being very large and remote units, were also frequently changed and rather arbitrary. They crossed county and regional boundaries willy-nilly, and occasionally even divided cities into different parts each of which was swamped in its constituency (Plymouth in particular suffered from this phenomenon). They had to change every time the Westminster boundaries were redrawn and every time there was a change in the national allocation of MEPs. There was only one case (1989) in which the election took place on the same boundaries as the previous election. If FPTP had continued to be used there would have been boundary changes in both 1999 (because of the redistribution of Westminster constituencies) and 2004 (because of a reduction in the size of the British delegation to the European Parliament). List PR and STV are relatively easily adjusted for changes in legislature size without disrupting local links, unlike FPTP.

On constituency links, the switch in the European Parliament electoral system was an interesting test case for academics. The political science literature cautions against assumptions that electoral system changes “generate a mechanistic response from representatives.”¹⁵Farrell, David and Scully, Richard Electoral Reform and The British MEP p7 and their citations for Katz, Richard Electoral Reform and its Discontents BEPR 9, 1999 and Representational Roles EJPR (1997) 32:211-26. A comparative study in the 1980s noted that “electoral systems are not fundamental in determining parliamentarian/ constituency relationships… electoral systems are, perhaps, rather more passive elements… than either supporters or opponents of electoral reform tend to believe.”¹⁶Bogdanor, Vernon (ed) Representatives of the People? Parliamentarians and Constituents in Western Democracies (Aldershot: Gower, 1985, p299)

Since moving from FPTP to closed list, British MEPs report spending less time on constituency centred matters and more on scrutiny and oversight. But British MEPs are still more inclined even than open list MEPs from other countries to see representing individuals and constituency concerns as an important aspect of their job. Culture and ethos have dominated mechanical effects – the casework ethos of British representation does still seem to carry over into the European Parliament delegation, while MEPs from countries with more technocratic cultures have carried that approach over as well. If there is a culture of casework and an emphasis on the importance of locality, even under closed list PR elected representatives will tend to maintain constituency links and parties will tend to select local candidates.

How does the electoral system affect diversity of representation?

The representation of women in the British delegation to the European Parliament has tended to be higher than at the House of Commons since direct elections began in 1979. However, progress was slow under FPTP. The introduction of List PR in 1999 led to a sharp increase in the proportion of women MEPs which fell only slightly in 2004.

At 24 per cent, Britain’s delegation in 2004 was only a little more balanced than the House of Commons in 2001 (18 per cent) or 2005 (20 per cent).

As in the outgoing 1999-2004 delegation to the European Parliament, there were four MEPs from minority ethnic communities elected in 2004 – two Labour (Neena Gill, West Midlands and Claude Moraes, London) and one Conservative (Nirj Deva, South East) were reelected. One Conservative stepped down (Bashir Khanbhai, Eastern) and one new Liberal Democrat was elected (Saj Karim, North West). Ethnic minorities accounted for 5.3 per cent of the British (and 5.1 per cent of the UK) delegation in 2004, compared to less than 2 per cent of MPs and 7 per cent of the population. If it were representative in proportion, there would be five or six ethnic minority MEPs rather than four, but if Khanbhai had remained third on the Conservatives’ Eastern list he would have been re-elected and would have brought the total up to five. Britain’s MEPs are therefore relatively representative of ethnic minorities. Before the introduction of PR in 1999, there was only one ethnic minority MEP (Mark Hendrick, Labour).

Closed list systems are potentially the most efficient way of ensuring that women and ethnic minorities are adequately represented, in that it is possible for the party to ensure that equality policies directly affect who is selected and elected. This is because closed lists:

1. Give parties (either through a central ruling or the wishes of party members, or a combination of the two) control over who is chosen as a candidate and in which order candidates are elected. 2. Involve selecting a slate of candidates to stand at the same time. Decisions about selecting a group of candidates raise obvious issues about how representative that group is, in a way in which a series of individual decisions in single member constituencies may not. An imbalanced ticket is a source of embarrassment for a party that expects to elect more than one candidate.

However, for closed lists to serve the aim of representativeness parties need to give a priority to the objective. If all parties do this, there is no problem, but if parties that do not regard it as a priority do well in the election, the overall balance of the sexes suffers. This was the case in the 2004 European Parliament elections, in which the Conservative Party and UKIP, both of which had overwhelmingly male slates of candidates for the top positions, did relatively well. Those elected for these parties comprised 37 men and two women (thus 95 per cent male). Those elected for the other parties combined comprised 21 men and 15 women (58 per cent male).

A stronger performance from the parties whose slates had more women would have resulted in an improved balance of the sexes; The success of UKIP and its overwhelmingly male candidate lists contributed to a poor overall result from the point of view of gender equality.

This result makes it clear that closed lists are not a panacea for a lack of gender and ethnic representation. Voters who favoured the policies of a party such as UKIP or the Conservatives in 2004 were unable to vote accordingly and ask the party concerned for women or ethnic-minority candidates. If there were open lists or STV, voters would be able to do more to affect which candidates were elected under each party banner. The argument for closed lists sometimes seems to assume that the voters would use their choice in a discriminatory manner, a proposition for which the evidence is lacking.

Possible changes to the list PR system

The total number of seats is not a matter that the UK government can change unilaterally, and the pattern of regions is now well established and is unlikely to be altered. This does however mean that elections in the smaller regions (North East with three seats, Wales with four seats) produce quite disproportional outcomes, although the results nationally are not severely affected.

The most important change that could be made to the list PR system is to give voters a choice of which candidates they wish to see elected – to move from closed lists to open or flexible lists. Voters dislike not having a choice over who is to represent them. In the public opinion study for the Independent Commission on PR (March 2004, fieldwork May 2003), 57 per cent expressed a preference for a choice of candidates.¹⁷Independent Commission on PR, Changed Voting Changed Politics Appendix C p137.

Closed lists give power to the party organisation, and even if there is internal party democracy, the party selectorate. They therefore encourage MEPs to look first to satisfying the party leadership, then to members who vote in ballots, and only then to distinguishing themselves to their constituents.

Most EU countries use one sort of list system or another to elect their European Parliament delegation – the principal exception is Ireland, both north and south, which uses STV. Larger countries often stick with closed list systems, but other countries such as Austria and Belgium do give voters the opportunity to vote for an individual candidate (although in practice these votes may have little effect on the result).

Open lists would be an improvement because they would meet some of the popular objections to not having a choice of candidate, and therefore bring European Parliament elections and candidates closer to the people. However, there is a price in terms of greater complexity and in many ‘flexible’ (or ‘semi-open’) lists voter choice can end up being illusory because it rarely affects the result.

During the Bill’s passage, both the Conservative spokesperson, Sir Brian Mawhinney, and Alan Beith, his Liberal Democrat counterpart, criticised the fact that voters would not be able to vote for individual candidates. Mawhinney also criticised the system for severing the link between constituent and individual MEP. In defending the proposal, Straw told MPs:

The electoral regions will be very large, and individual candidates are unlikely to be known by more than a small – not to say tiny – fraction of the electorate. Voters cannot, therefore, be expected to make an informed choice between individual candidates from the same party. If the electorate were required to rank candidates of the same party, such choices could be arbitrary. I also fear that it could disadvantage women and ethnic-minority candidates. I understand that Lord Plant shares that view…

I realise that this is an important issue that goes to the heart of the Bill, and I have listened carefully to all the representations that have been made to me on the point… As with any other system, there are arguments both ways. The modification of what we propose in the Bill would provide some more direct voter preference… On the other hand, that is marginally more complex than the present simple system that we propose, and it is possible for a candidate low down on a party list to receive many personal votes – perhaps more than those of his or her party colleagues – yet still not be elected, because the weight of party votes helps those higher up the list.

Nevertheless, I am prepared to listen to the arguments for adopting a Belgian-type system and to give them careful consideration.

However, at the Bill’s Report Stage in March 1998, Straw rejected the proposal:

I have concluded that there is no advantage in adopting in Great Britain a system of the kind used in Belgium. The type of system which is in use in Belgium has some superficial attractions. An elector may express a preference for a particular candidate, rather than simply endorsing all the candidates on a party’s list. However, the system suffers from a fundamental and incurable weakness, in that voters’ preferences for individual candidates are not necessarily translated into electoral success… even where votes for individual candidates amount to as much as 40 per cent of a party’s total vote, those candidates receiving the fewest individual votes can be elected while those receiving the most are not. I believe that such an outcome could lead to substantial disillusionment among the electorate following an election.

Alternatives

The option of returning to FPTP is not available. The UK is bound by European treaties and agreements to having a proportional system in European election. It should be ruled out in any case for its highly unrepresentative properties.

Northern Ireland elects its MEPs (as well as local authorities and Assembly members) by STV, as it has done since 1979 while two systems (FPTP 1979-99 and closed list PR since 1999) have been used in mainland Britain. The province’s MEPs have a higher profile than in the rest of the United Kingdom and more personal accountability, and turnout has been higher than in Britain in every European election.

Further evidence for the ability of STV to forge greater connections between electors and MEPs comes from the Republic of Ireland. In most EU countries, other than those with compulsory voting or elections on the same day as national elections, turnout in European Parliament elections is much lower than for the domestic parliament. Ireland has one of the smaller gaps between national and European Parliament turnout, and the 2004 ICPR ventured that “it is therefore possible that the range of choice offered by STV encourages more people to vote.”¹⁸ICPR, page 81.

In January 2006 the report of the Arbuthnott Commission inquiry into electoral systems in Scotland recommended that STV be introduced to elect Scotland’s MEPs. The Commission observed:

By requiring voters to select a party rather than an individual, the existing system makes it difficult for candidates to develop a profile among the electorate. The Commission therefore concluded that the status and legitimacy of Scotland’s MEPs would be improved if the closed list was replaced with the Single Transferable Vote for Scottish elections to the European Parliament.

Scotland’s seven MEPs are ambassadors for Scotland – they are ‘Scotland’s team’ in the European Parliament, a role the MEPs themselves value. We believe that introducing the Single Transferable Vote to elect them would allow Scottish voters to select the best team of parliamentarians to represent the country.

Under both closed list and FPTP elections for the European Parliament there have been frequent complaints about the lack of personal profile for MEPs and a lack of public knowledge of the functions and activities of MEPs. The Commission recognises that STV would help to remedy this problem:

If candidates for the European Parliament have to ensure that they are given sufficient first and second preferences from the Scottish electorate, they might be more likely to engage with the electorate and have a higher profile on the domestic stage. As a consequence the public could be expected to have a better idea of who they are, what they do, and the relevance of the institution in which they serve. ¹⁹Commission on Boundary Differences and Voting Systems (Arbuthnott Commission), 2006, Putting Citizens First paras 4.96, 4.98, 4.99.

STV is the system most likely to encourage MEPs to prioritise constituency work within their overall remit.

The recommended use of a single national constituency electing seven MEPs for the whole of Scotland avoids an artificial division of the country, would enable MEPs to take the national strategic approach which is appropriate for the European Parliament, confirm the existence of a Scottish ‘team’ as at present and also ensure an adequate degree of proportionality.

The Scottish model outlined by Arbuthnott can be easily transferred to Wales, whose four MEPs could be easily elected by nation-wide STV. The logic of Arbuthnott’s arguments apply just as strongly to Wales as to Scotland, although it should be noted that there is less of a sense of team identity in Wales than Scotland.

England presents more of a problem, in that its regions are of varying size and less well established entities with a sense of team representation than Scotland or Wales. However, all but three are the same size (seven seats) as Scotland or smaller and therefore amenable to forming single STV constituencies. The exceptions are the North West (nine seats), London (nine seats) and the South East (ten seats). It would make sense to subdivide these larger regions into constituencies electing 4-6 members.

What system is used?

Executive mayors in some local authorities have been directly elected since May 2000. The issues involved in electing a single individual are different to those involved in electing a group of people. For instance, there is no way of achieving ‘proportionality’ when only one elected post is available. Candidates win or lose rather than achieving a share of representation.

Mayoral elections use the Supplementary Vote (SV) system. This allows the voter to cast a first and a second preference, but not more than two preferences. The first preferences are tallied and if a candidate has over 50 per cent of the vote, that candidate is declared elected. If no candidate has a majority, then all but the top two candidates are eliminated. The second preferences marked on the ballots for eliminated candidates are then tallied, and those second preferences recorded for either of the top two candidates are added to the first-round totals for those candidates. Whichever of the two remaining candidates has the most votes after this reallocation of second preferences wins the election.

The design of the ballot for mayoral elections consists of two adjacent columns, one for the first preference and one for the second preference. A voter marks an ‘X’ in each column (although a ballot is not invalidated if the second preference is not recorded).

Why this system?

In July 1997, the newly elected Government released a Green Paper, entitled New Leadership for London: the Government’s proposals for a Greater London Authority. It set out a number of possibilities for the electoral systems to be used for both mayoral and GLA elections. First-Past-the-Post, the ‘second ballot’ or the Alternative Vote system were options for mayoral elections.

The Supplementary Vote had been the preferred option of the Labour Party’s Plant Report for elections to the House of Commons and following the Green Paper Patrick Dunleavy and Helen Margetts proposed using it for mayoral elections. Dunleavy and Margetts’ rationale for favouring it over AV was that it guaranteed that only the top two candidates after first preferences had been counted would be able to win. This would prevent a candidate without much positive support from coming through because they are a lowest-common-denominator second or third-choice for everyone.

The government’s March 1998 White Paper recommended SV for Mayoral elections and AMS for the Assembly, with 14 individual constituencies and 11 seats taken from a London-wide list. The government claimed the following advantages for SV:

• It is simple and easy to use and can produce a clear winner who would enjoy the support of a large number of Londoners.
• It is a simpler form of AV, but it is quicker to operate and count.
• It retained the familiar use of an ‘X’ in the ballot box, rather than unfamiliar preferential voting as with AV.

Plant (and other backers of SV) thought that SV had advantages over AV for Westminster elections in that it was simpler and that it also avoided the use of weak preferences. The simplicity was seen as being from the voters’ point of view, in that it preserved continuity from the familiar ‘X’ voting, and in counting. An SV count involves the familiar process of counting ‘X’ votes, albeit in two columns, as opposed to preference voting with which voters, other than in Northern Ireland, are unfamiliar. The simplici-ty argument for voters can be assessed with reference to the rate of ballot spoilage.

The argument on the point about weak preferences is that voters may have clear ideas about their top two candidates but below that preferences might be given relatively lightly and end up having an importance out of all proportion when the votes come to be counted.

During debate on the Local Government Bill, the Conservatives put forward an amendment to replace SV with First-Past-the-Post, whilst the Liberal Democrats proposed to replace SV with AV. Both amendments were opposed by the Government and defeated. Explaining the Government’s stance, then Local Government Minister Hilary Armstong told MPs:

The Conservative amendments would simply return us to a first-past-the-post system. That would not assist the election of directly elected mayors because we want to ensure that we do not allow the election of a mayor who was opposed by the majority of those voting. That could easily happen under the first-past-the-post system. If five people stood and the highest percentage of the vote that any of them got was 29 per cent that would be a small mandate. Indeed, it would not be a mandate on which to be elected to the mayoral role that we discussed earlier. The mayor will represent a large number of people and will take important decisions on their behalf. We have all talked about mandates and a simple majority system does not deliver the sort of mandate that the public would expect for a mayor.²⁰Proceedings of the Standing Committee on the Local Government Bill –, Tuesday 6 June 2000 http://www.publicat ions.parliament.uk/ pa/cm199900/cms tand/a/st000606/p m/00606s02.htm

Results

There have now been a significant number of mayoral elections in English local authorities, and the London Mayor who is in a rather different category. The complete list of mayoral elections is as follows:

May 2000 London Ken Livingstone (Independent) May 2002 Doncaster Martin Winter (Labour) Hartlepool Stuart Drummond (Independent) Lewisham Steve Bullock (Labour) Middlesbrough Ray Mallon (Independent) Newham Robin Wales (Labour) North Tyneside Christopher

Morgan (Conservative) Watford Dorothy Thornhill (Liberal Democrat) October 2002 Bedford Frank Branston (Independent) Hackney Jules Pipe (Labour) Mansfield Tony Egginton (Independent) Stoke-on-Trent Mike Wolfe (Independent ) June 2003 North Tyneside Linda Arkley (Conservative) May 2004 London Ken Livingstone (Labour) May 2005 Doncaster Martin Winter (Labour) Hartlepool Stuart Drummond (Independent) North Tyneside John Harrison (Labour) Stoke-on-Trent Mark Meredith (Labour) October 2005 Torbay Nick Bye (Conservative) May 2006 Hackney Jules Pipe (Labour) Lewisham Steve Bullock (Labour) Newham Robin Wales (Labour) Watford Dorothy Thornhill (Liberal Democrat)

Assessment

Participation

Given that directly elected mayors are an entirely new institution, and the electoral system is only one aspect of the creation of this sort of elected office, it is not possible to measure the impact of SV as such on turnout. However, part of the rationale for elected mayors was that it would encourage public interest and participation in local elections.

While elected mayors are an interesting innovation, they do not seem to attract extra voters to the polls. It did not raise turnout much from that attained in normal local elections, and when an entirely separate election turnout was abysmal.

Fair representation: does SV create majority winners?

By definition, SV winners have a majority among the votes counted in the second round. But perhaps surprisingly, most of the winners of SV mayoral elections have not, even after the redistribution, had a majority among those voting in the election. This is because so many of the votes cast at the first stage end up not transferring and being counted in the second round.

Winner’s final Majority round vote (% of all voters) London 2000 45.3 No Doncaster 2002 44.0 No Hartlepool 2002 37.8 No Lewisham 2002 55.1 Yes Middlesbrough 2002 62.8 Yes Newham 2002 50.8 Yes North Tyneside 2002 42.9 No Watford 2002 48.6 No Bedford 2002 43.6 No Hackney 2002 49.3 No Mansfield 2002 42.4 No Stoke-on-Trent 2002 28.9 No North Tyneside 2003 49.6 No London 2004 44.4 No Doncaster 2005 42.0 No North Tyneside 2005 47.7 No Hartlepool 2005 50.0 Yes Stoke-on-Trent 2005 44.6 No Torbay 2005 29.4 No Hackney 2006 54.4 Yes Lewisham 2006 42.8 No Newham 2006 56.9 Yes Watford 2006 51.2 Yes

Bold indicates first round winners.

In three elections – Mansfield and Stoke-onTrent in 2002 and North Tyneside in 2005 – the operation of the SV system has resulted in a candidate’s first-round lead being overturned by the transfer of preferences in the second round.

Of the 23 mayoral elections, only seven have produced a winner endorsed by a majority of those voting, three of which were first-round winners anyway. This is directly related to the huge wastage of votes caused by the rules of SV. The failure of SV to produce majority winners has been an unexpected feature of the system. The Jenkins Commission in 1998 thought that the winner would have over half the vote ‘in the majority of cases’ but the evidence of SV in practice shows that it has not fulfilled the initial hopes of those who introduced it.

Fair representation: spoiled papers

Completely invalid ballots

Mayoral elections have a relatively high incidence of spoiled ballot papers. Single FPTP elections have spoilage at between 0.1 per cent and 0.2 per cent; combined FPTP elections (as in May 2005 with county councils and Westminster) get 0.3 per cent to 0.4 per cent, 1.8 per cent at the combined European and local elections in 2004 and 2.9 per cent in the London mayoral election. Even in Torbay in 2005 the incidence of rejected ballot papers was high, at 1.7 per cent. In both Torbay and London voting for more candidates than entitled was a much more widespread phenomenon than in most elections.

The evidence suggests that the Supplementary Vote has not proved as simple for voters to use as its designers intended. There are high rates of spoiled ballots and ballots that are not transferable to the second count. Ease of counting may have been a relevant consideration when the legislation was going though in 1999 but it is much less so now. The London mayor and assembly elections have been counted electronically in both 2000 and 2004 but since the original decisions were taken there has been considerable progress in electronic counting of preferential votes in elections. Scottish local elections in 2007 are to use electronic counting for Single Transferable Vote ballots. With that experience a year before the London elections of 2008, there seems no reason why an Alternative Vote count should present any difficulties.

Second round invalid ballots

Some voters will choose not to cast a second preference because they are indifferent between all the candidates other than their favoured one, and that is of course their right. However, the scale and pattern of wasted votes indicates that this is a relatively small element of the non-transfer of votes in the second round of mayoral elections. In London 14.2 per cent of valid first-preference votes were not valid for second preference, which seems a very high level of indifference if taken at face value. On top of these voters (who mostly left the second column blank) there was another widespread phenomenon that invalidated second votes.

‘Double voting’ (i.e. casting firstand secondpreferences for the same candidate) was measurable in the 2004 London mayor election. Overall in London, 7.7 per cent of valid first-preference votes had the same candidate as their second preference, something that could not possibly help that candidate. The incidence of double voting was related to social exclusion and not having English as a second language, as is apparent from the distribution between the borough-based GLA constituencies.

Double voting was particularly prevalent among Respect and Labour voters. Particularly in the context of an election with as many ballot papers as in London in 2004 (mayoral, GLA constituency, GLA list and European), voters who are not particularly aware of the different systems may decide to vote for the party they support at every opportunity. The second column of the mayoral ballot may seem an extension of that principle.

Voter choice

Mayoral elections have seen some unusual patterns of electoral competition, but this seems to owe more to the fact that mayoral elections are a new, localised and candidate-centred sort of election that seems to encourage voters to examine independent and other options. The nature of the office, rather than the system, seems to have produced new sorts of voter choice.

However, the amount of choice available when voting is constrained by the system (although SV still marks an important extension of voter choice compared to FPTP). SV wastes votes on a large scale. Voters are allowed to use only their first and second preferences and all but the top two candidates are eliminated for the second count. This means that if the voter’s first preference candidate is eliminated, the only way their vote can affect the final result is if it is cast for one of the two candidates who have gone through to the second round.

Leaving aside double voting and void second preferences, there is a large proportion of ballots where people have filled the paper out completely correctly but just chosen two candidates who did not make it to the second round. In order to cast an effective second preference, voters need to guess which of the candidates will be in the final round, and which out of those two they would prefer.

In several mayoral elections the battle for second place has been extremely close. In Stoke-on-Trent in 2005, only 432 votes (0.5 per cent) separated the second candidate and the fourth candidate. In Hackney in 2002 the gap between second and fourth was 317 votes (1.0 per cent), and in Bedford in 2002 it was only 597 votes (2.2 per cent). In seven of the 23 mayoral elections to date the gap between secondand third-placed candidates has been less than 1.5 per cent. It is virtually impossible for voters to make accurate guesses about which candidates are going to be in the second round in these circumstances.

Elimination of all but the top two candidates has created an arbitrary element in the choice of elected mayors and reduced the influence that voters’ choice has been having on the final results.

In only one authority, North Tyneside, has a majority of the eliminated candidates’ vote been transferred in the second round. This is facilitated by the way that mayoral contests in the borough have been dominated by candidates of the Labour and Conservative Parties, therefore requiring less guesswork from supporters of other candidates about who will be in the run-off. In the third mayoral election in North Tyneside, very few votes were wasted.

However, North Tyneside is unusual among mayoral elections. In most of the authorities having mayoral elections, strong independent and local political candidates have contested the position. In this situation it is less clear who the top two candidates will be, and in general the more candidates who might get into the top two, the fewer transferred votes will count in the second round. In all four mayoral elections in May 2006, the gap between the secondand fourth-placed candidate was less than 10 percentage points, indicating that while one candidate had a clear lead it was far from clear who would be second placed.

In some circumstances, if it is unclear who will be in the run-off, voters may end up defeating the candidate they really prefer. Imagine, for instance, a voter for one of the minor independent candidates in the Stoke-on-Trent mayoral election of 2005, who vaguely likes the Conservatives but strongly dislikes the BNP. Given that it was unclear whether the Conservatives or the BNP would be in the runoff, it would be logical for such a voter to cast a second preference for Labour. This is because Labour were virtually certain of a place in the runoff, and a vote for Labour would be the best way of making absolutely sure the BNP lost. However, as it happened the Conservatives just made it to second place and that person’s vote would actually have helped Labour defeat the Conservative candidate whom the voter would have preferred.

Is SV simple to use?

The Supplementary Vote is unduly complicated and restrictive compared to AV, although it is still better than First-Past-the-Post because it greatly reduces tactical voting and produces winners with more support. It had attractions when it was first introduced and the concept of continuity with ‘X’ voting under FPTP appeared important. It was a step forward from FPTP, in that it does enable people to cast a sincere first preference and then a second preference between two main parties. It can work well in conditions where political competition is relatively simple, as in North Tyneside.

Concern over the very high rate of double voting and spoilage in the City & East GLA constituency in 2004 was one of the reasons for a pilot scheme tried in the mayoral election in the borough of Newham in 2006. Instead of a two column SV ballot, voters were given a single-column ballot in which they were asked to give their first and second-preferences with ‘1’ and ‘2’ respectively. This made the system closer to AV, although the restriction to two choices and the rule for eliminating lower-placed candidates were of course just as under SV.

The Newham mayoral pilot was not entirely successful, in that there was a high incidence (6.1 per cent) of spoilage caused by voters marking two X-votes in the column. Such votes were naturally void, because it was impossible to discern which was a first and which a second preference. However, as the table shows, Newham is a borough that suffers a relatively high spoilage rate in other sorts of election, connected to its high proportion of non-English-speaking residents.

*The figure for the London mayor second choice includes double votes and blank second columns but not ‘ineffective’ second votes, i.e. those not being counted in the second round.

The Electoral Commission’s report on the Newham pilot warned that ‘it may be unwise to judge the success of this change on the basis of a single pilot scheme, since electors may benefit from repeated exposure to new processes to improve understanding and awareness.’ While many electors needed to have polling station staff explain the new ballot design, this was because of unfamiliarity rather than inherent complexity: “there is no evidence to suggest that the new design was more complicated to use than the papers used in previous elections, just different.”²¹Electoral Commission, 2006 Electoral pilot scheme evaluation, London Borough of Newham August 2006

It might also be noted that the Newham pilot ran alongside an MNTV (multi-member FPTP, or Multiple Non-Transferable) election for the council, with three members to be elected for each ward. Therefore the voters were faced with a council ballot on which they were specifically told to use three ‘X’ votes, and a mayoral ballot on which a completely different principle applied. A future pilot of an AV-style ballot without this confusing context might reduce the rate of spoilage.

Alternatives and conclusion: SV and AV

There is something to the logic of SV not using lower preferences. AV is sometimes criticised on the grounds that it can produce a ‘lowest common denominator’ winner without much positive support of their own, and winning under SV does necessitate having a certain amount of positive support.

However, the idea that all preferences below the top two are weak needs further examination. Voters may often not care very much about the middle of their order of preference, but there are cases where the later preferences are ones that they consider very important. An example would be a voter who strongly opposes the BNP and would wish to place that party last, without being too bothered about their middle preferences. It seems a valid use of an electoral system to ensure that such strong aversions are given a chance, if the election came down to it, to affect the result. It would also seem, in a multi-party political environment, unreasonable to assume without evidence that preferences below the second are weak.

Perhaps the basic flaw of SV is that it is implicitly based on a two-party model of politics. The assumption is that people may vote for anyone in the first round to make a point but the important choice of who governs is between two of the established parties (usually Conservative and Labour). This is decreasingly the case in national politics and has been comprehensively invalidated in the mayoral elections that have taken place to date. It has become the norm for there to be strong non-party challenges at mayoral elections, from the first surprising monkey-suited success of Stuart Drummond in Hartlepool through Ray Mallon in Middlesbrough to other less famous independent victories in Bedford and Stoke. Torbay saw a further development of this trend with 59.1 per cent of the vote going to independent candidates, although it was so scattered that the two candidates in the run-off were Conservative and Liberal Democrat.

The Alternative Vote, instead of requiring voters to guess who will be in the run-off election, allows voters to give as many preferences as they like. It would have enabled the majority of Torbay voters who favoured independents, if they so wished, to ensure that an independent rather than a party candidate was elected mayor. AV is a much more appropriate system to accommodate the plural nature of mayoral politics. It would be a modest and sensible improvement for the government to change the mayoral electoral system to the Alternative Vote.

What system is used?

The Greater London Assembly (GLA) has a Mixed Member Proportional (MMP) – also known as Additional Member (AMS) – electoral system. It has 25 members – 14 elected from constituency seats based on combining London boroughs, and 11 from a London-wide list. As in Scotland (at least before 2007) and Wales, voters choose their constituency and list members separately, on separate ballot papers, but the list members are allocated to compensate for the disproportional constituency results. Unlike in Scotland and Wales, there is a threshold of 5 per cent of the vote before a party is entitled to a list seat.

No other directly elected English regional assemblies exist – the proposal to create one in the North East was rejected in a referendum in November 2004. English regional government, as envisaged, would have involved the creation of bodies with 25-35 members elected using the MMP system.

Unelected regional assemblies were established in the eight English regions other than London in 1999, following the Regional Development Agencies Act 1998. In 2004 they acquired the status of regional planning bodies. The assemblies are composed of delegates from local authorities in the region (around 70 per cent of members) and appointed representatives of other organisations. Their composition is fluid but there are rules to ensure that their membership reflects the political balance of the region.

Why was this system chosen?

Prior to 1997, Labour had expressed no firm views on the electoral system to be used to elect London’s mayor and the London Assembly. In July 1997, the newly elected government released a Green Paper, entitled New Leadership for London: the Government’s proposals for a Greater London Authority. It set out a number of possibilities for the electoral systems to be used. For the Assembly the possibilities were single-seat constituencies (elected by First-Past-the-Post or the Alternative Vote) or multi-member constituencies (elected by First-Past-the-Post, a list system, the Additional Member System or the Single Transferable Vote).

In response, Professor Patrick Dunleavy and Dr Helen Margetts wrote a report recommending the Supplementary Vote for elections for the Mayor, and either AMS (with 14 or 16 local seats), STV (using five multi-member constituencies) or a list PR system for the Assembly. It concluded that it was important to choose voting systems that complemented one another, and therefore favoured either SV for mayor with AMS or list PR for the Assembly, or AV for Mayor and STV for the Assembly. The government accepted the logic of this position.

The government’s rationale in favour of MMP/ AMS was as follows:

• It was the electoral system being used for elections to the Scottish Parliament and the Welsh Assembly.
• It would produce an Assembly which closely reflects the views of Londoners.
• It provided a balance between constituency representation and London-wide representation: “This will ensure that there are assembly members with whom the electorate can identify on a geographical basis and that there is a more proportional outcome.”

In May 1998, Londoners voted in a referendum in favour of an elected mayor, and legislation was introduced to bring the proposals into effect. At the Second Reading of the Greater

London Authority Bill on 14 December 1998, Deputy Prime Minister John Prescott outlined the case for the voting systems chosen:

The mayor will be elected by the Supplementary Vote system, which will provide a clear mandate from the people of London. The 25-member assembly will be elected by the Additional Member System, so the views of Londoners will be closely reflected in its political make-up. We have made a judgment, and chosen this system. That means that the mayor will get an overwhelming vote. It is important to have a mayor with a majority of support in London. That is why we have used the Supplementary Vote in this way. The method also means that only two counts will take place – the main count and the additional one for distributing the supplementary votes between the top two candidates. I think that that is the fairest way. No doubt points will be made about that, but that is the formula we put in the referendum, and the Bill implements what we agreed in that.²²House of Commons Debates 14 December 1998.

In spite of Conservative amendments proposing First-Past-the-Post, and Liberal Democrat amendments proposing STV or open lists for Assembly elections, the government won the day and SV and AMS were adopted as planned. The first elections took place in May 2000 and the second in June 2004.

Assessment

Participation

Greater London Assembly elections are closely connected with, and always concurrent with, elections for the much more high profile post of London Mayor. The mayoral election is the principal determinant of participation in Assembly elections.

Is the system fair?

The London Assembly electoral system has produced results that reflect the votes cast by Londoners with a considerable degree of accuracy, somewhat surprisingly so given the small size of the Assembly.

However, the threshold of 5 per cent of the vote to qualify for a list seat does distort representation a little. The distortion is small because with a 25-member assembly a vote share of 4 per cent would otherwise qualify a party for a seat. In 2000 no parties were affected by the need to win 5 per cent to qualify for a list seat, but in 2004 two parties (the BNP and Respect) would have won a seat each if the threshold rule had not been in place.

The compensatory principle of distributing list seats can create some strange results, as in 2004 when Conservative group leader Eric Ollerenshaw, a list member, lost his place on the GLA because his constituency colleagues had done too well.

Constituency representation

The constituency role of GLA representatives is smaller than that of councillors, MPs and members of devolved assemblies. The principal functions of the GLA are to exercise scrutiny over the Mayor and to provide checks and balances for what would otherwise be an extremely powerful and unaccountable executive. The size of constituencies for the GLA is also an issue in reducing the local leadership

role of constituency GLA members. Each constituency is composed of one or more boroughs and has an electorate of towards half a million people. These features of the GLA have prevented the potential problems of rivalry between list and constituency members from becoming too significant as yet.

Alternatives

The principle enunciated by Dunleavy and Margetts about the mayoral and Assembly electoral systems being complementary is worth bearing in mind. This noted, the current arrangements are not quite as complementary as they seem. It is true that the election for Mayor using the Supplementary Vote takes the form of two ‘X’ votes on one ballot paper, while the GLA ballot is superficially similar because the voter uses two ‘X’ votes (albeit on separate ballots). The difference arises in what these votes mean.

It is perfectly logical for someone to choose a constituency representative and a London-wide list from the same party – this pattern of voting is the one most likely to advance the voter’s political views in London government. However, it is pointless to use the second mayoral ‘X’ vote for the same candidate as the first vote. The imperfect relationship between the two forms of voting may help to explain the very high rates of ballot spoilage and under-voting at GLA elections.

In ‘Electing Mayors’ we discuss alternative electoral systems for mayoral elections, and conclude that the Alternative Vote would be a more effective way of achieving the aims outlined when the Supplementary Vote was devised.

If this suggestion on mayoral elections is adopted, it would allow a better harmonisation of voting methods. Voters filling in a single pref-

Greater London Assembly and the English regions Chapter 8 Britain’s experience of electoral systems

erential ballot for mayor could also be offered a single preferential ballot for Assembly representatives. This would entail the adoption of the Single Transferable Vote for the GLA.

Given the established nature of the GLA as a multi-party assembly in which smaller parties are represented, continuity would suggest adopting a relatively large district magnitude of perhaps five seats with five members each, or even more members per seat.

English regional government: the proposal

Before the ‘No’ vote on the proposal for a directly elected North East regional assembly in November 2004, some preliminary work had been done on the electoral system to be used for regional government. This envisaged using MMP to elect future regional authorities. The ODPM consultation on the North East assembly suggested a 25-seat body with 17 members elected from constituencies and eight elected from top-up lists. As the smallest region, the North East would have had the smallest regional authority and other regions would have had more elected members.

The 2002 Government White Paper, Your Region, Your Choice, made the case for a more proportional system, and MMP in particular, as follows:

The voting system for regional assembly elections must:

• promote inclusiveness, so that assemblies reflect the interests of the range of communities across all parts of their region;
• strike a balance between ensuring that voters have an identifiable constituency representative and encouraging assembly members to take a region-wide view of their responsibilities.

The Government accepted that a form of proportional representation would be the best way of achieving these objectives.

The government favoured AMS/ MMP using the arguments that:

• it ensures that all voters have an identifiable constituency representative, and that counties and sub-regions which have a distinct identity within their region can elect at least one constituency member to represent their interests;
• but at the same time the relatively large constituencies and top-up members should avoid tensions between assembly members and MPs representing the same constituencies and encourage assemblies to take a region-wide view of their responsibilities;
• it gives voters a wider choice than most forms of PR, so that, for example, they can vote for a popular independent candidate in their constituency whilst still supporting their preferred party with their top-up vote;
• it is relatively simple to understand.

The Government proposed that the proportion of top-up seats in any region would be around 33 to 35 per cent of all seats in the assembly, and that constituencies would be based on existing local authority areas. Also, ‘on grounds of simplicity and to produce the most proportional outcomes’, a single top-up constituency in each region was proposed, with a minimum threshold for party representation in an assembly of five per cent of the vote, following the GLA precedent.

As the government acknowledged, broad proportionality was considered particularly important in the north east because without it, one party could have overwhelming dominance of the assembly with the support of only a minority of voters. If the assembly were perceived to be the property of one party, it would not be seen as a legitimate voice of the north east.

English regional government: the alternatives

The selection of MMP/ AMS as the system of choice for regional assemblies should not be an unexamined choice. It has some features that may cause problems should the regional government programme be revived at a later date.

MMP/ AMS creates two types of members – constituency and list members. Tensions can, and do, arise between those who see themselves directly elected by voters in a particular area and those whose links with the electorate are more distant. As has been seen in Wales in particular, where one party wins most of the constituency seats and representation of the opposition is almost entirely through list seats, these tensions can lead to frictions. This would be likely to be replicated in the North East, South East and North West, and quite possibly in other regions depending on election results.

The mission of regional assemblies would be less about constituency representation than a broader view across the entire region. A system in which the ruling group will tend to represent constituencies, while the opposition parties will tend to be drawn from all-regional lists seems to offer the worst of both worlds. The executive may be unable to take the broad regional view required in the system of regional government because of their single-constituency focus, while the opposition are encouraged to take a broad view but lack a direct channel of representation on how policies affect individual communities.

The ratio of list to constituency members as proposed for the regions would be closer to the Welsh Assembly’s system than the more proportional variants in Scotland (43 per cent list) and London (44 per cent list), so that single party majority government on a minority of the vote would certainly be within the bounds of possibility in many regions. The less proportional variants of MMP would tend to exclude the smaller parties and reduce the potential for assemblies to represent all the communities and interests in the region. At the very least, there is a case for increasing the proportion of regional list members to 40 per cent, although this would exacerbate some of the problems of differences between list and constituency members. The proportionally-based system used to allocate places on non-elected regional bodies is significantly more representative than that proposed for the directly elected North East assembly in 2004.

FPTP or AV

It would be technically possible for the regional assemblies to be elected on more majoritarian systems than proposed by the government, such as AV or FPTP, but the government’s case in Your Region, Your Choice demolished the arguments for such a position. In short, it argued:

• Because voting behaviour is differentiated by region, FPTP would mean that several regions would be under permanent party control – the Conservatives in the South East and Eastern, and Labour in the North East, North West and Yorkshire & Humber. The East Midlands and West Midlands would probably alternate between Conservative and Labour control, with the South West being the only region that would be unlikely to generate a single party majority.
• In some regions – such as the North East – the largest party would generally be so dominant under FPTP that it would be impossible for the other parties to elect sufficient members to perform the functions of opposition without making the overall assembly too large.
• PR would encourage assembly members to take a region-wide view by avoiding representation in the assembly being linked to relatively small constituencies. It would also mean that regional assembly constituencies will not mirror parliamentary or local government constituencies and will thus avoid potential tensions.
• PR can facilitate diversity in the selection of candidates by political parties and therefore better represent all sections of society in the regional assemblies than a system based on single districts.

Single Transferable Vote

STV avoids several of the problems that arise over MMP/ AMS. All members are elected equally and problems of different status do not arise. All members would be elected from subregional units, almost certainly based around whole local authorities, which would be intermediate in focus between small single-member constituencies and large regional lists. Members would be responsible directly to the electorate rather than to the party bodies responsible for ordering lists. It accomplishes the blend of local and region-wide without creating different categories of member.

Conclusion

The GLA is different from the previously proposed regional assemblies in that it is part of a system of government which includes the Mayor as well as the executive. It is logical that the GLA be seen in the context of electing the Mayor.

While English regional devolution is not currently a live issue and the whole programme has been shelved, it is one that may well revive in future should demand increase and a better case made for it than was advanced during the North East referendum in 2004. In that eventuality, as well as looking again at powers and composition to answer some of the objections that arose in the North East case, there should be a fresh look at the electoral system and a further exploration of the arguments relating to MMP/ AMS, open and closed lists, and the case for STV.

What system is used?

Local government elections in Britain take place under several different electoral systems already.

• In Scotland, single-member plurality (FPTP) has been used in all local government elections since 1975, although it is being replaced in May 2007 with the Single Transferable Vote.
• In most English county, metropolitan and unitary authorities the vast majority of seats are elected one at a time using singlemember FPTP.
• In some county, Welsh unitary and English unitary, and most English shire district authorities there are wards electing more than one member at the same time (multi-member FPTP, or Multiple Non Transferable Vote – MNTV). In the London boroughs threemember MNTV is used in nearly all wards.

Northern Ireland, where the Single Transferable Vote (STV) has been used some 1973 in all local government elections, was discussed separately in chapter 5.

The electoral cycle for local government elections is also complex. Different categories of authority have elections in different years. Another distinction is between authorities having partial and ‘all-out’ elections. In the former, terms of office are staggered so that only a proportion of councillors – usually a third – is up for election in a given year. In metropolitan councils, for example, each ward has three members elected one at a time in 2006, 2007 and 2008 while the fourth year of the cycle (2009) sees no local elections. When boundary changes take place authorities that normally have partial elections have all-out elections, as the metropolitan boroughs did in 2004.

2006 London Boroughs (all members), metropolitan districts (one-third), some unitary and district councils. 2007 Scottish local authorities (all members), most district councils (all-out and one-third), some unitary councils, metropolitan districts (one third) 2008 Greater London Mayor and Assembly, metropolitan districts (one third), some unitary and district councils, Welsh local authorities (all members) 2009 English county councils, Northern Ireland local authorities, a few unitaries.

Results

Because of the diversity of local election results it is not possible to encapsulate them in short tables, as it is for other sorts of election. Full statistics on local government elections back to 1973 are available courtesy of Colin Rallings, Michael Thrasher and the Local Government Chronicle Elections Centre at the University of Plymouth. Their annual Local Elections Handbook and Local Elections in Britain: A Statistical Digest (1973-2002) are essential source material. Many of the tables below derive from their hard-accumulated data and anyone writing on local elections owes Rallings and Thrasher a debt of thanks.

In general terms, there are several factors that affect local election results:

• Party contestation. The normal pattern in local elections is for some form of party politics to prevail in the way councils are organised and seats are contested, although in some rural areas elections still tend to be non-party.
• Basic political composition. Voting behaviour in local and national elections is related, so that areas that are strongly for one party on the basis of national factors, class, ethnic and regional identity, will tend to see that reflected in local elections. Thus, County Durham is strongly Labour in local elections and Buckinghamshire strongly Conservative. Shifts among different groups are reflected too – for instance the local elections in 2003 when Labour did badly in Muslim areas indicated the same shift as took place in the 2005 general election.
• Timing of the election. Local elections follow, to some degree, the national ups and downs of popularity for the political parties. Local elections taking place in years when a party is particularly strong nationally will see good results for that party. For instance in 1995 Labour performed exceptionally strongly because of the unpopularity of the Major government at that time. Labour losses in 2003, 2004 and 2006 were aggravated by dissatisfaction with national government policies. Local elections are widely viewed as a test of the popularity of the government and national political leaders.
• Local political forces. Increasingly, locally based political movements have contested local elections and the extent to which these are organised will affect local elections. An early case was the rise of Kidderminster Hospital & Health Concern (KHHC) on Wyre Forest council in 1999, but there are local parties in other areas such as Stoke-onTrent, Wigan and Canvey Island. p Local issues. Although national factors have a very strong influence on local results, these are always blended with voters’ views about the record of the local authority or feelings about the individual candidates standing for election. The national picture sets the trend, but there are always some striking variations around the trend caused by local factors.

Assessment

Participation: turnout

Until 1985, Metro and London figures include metropolitan county and GLC elections; since 2000, London figures include GLA. Local elections coinciding with general elections (in 1979, 1997, 2001 and 2005) have been excluded. In 1999 Scotland and Wales, and in 2003 Scotland only, coincided with Scottish Parliament and Welsh Assembly elections.

The broad trends in participation in local elections are clear. From the late 1970s onwards turnout settled into a pattern of fluctuations in a fairly tight band just over 40 per cent. In the late 1980s it trended upwards to a peak in 1990 (the year the poll tax was introduced in England and Wales) before falling back below 40 per cent in the mid 1990s. A further downward shift took place in the late 1990s, reaching bottom between 1998 and 2002 at around 30 per cent. There has since been a recovery to somewhere around 37 per cent as political debate has become more heated and postal voting more available. Turnout has tended to be higher for local elections in Wales and Scotland (although in Scotland this is now because of the concurrent Parliament elections), and lowest in the metropolitan regions of England.

A subsidiary pattern can be seen in the shire district elections, at least until 1995. Turnout tended to be higher in years when the shire districts having all-out elections were polling, such as 1976 and 1983. Academic research has tended to support the contention that all-out elections tend to attract higher turnout than elections for only part of the council, perhaps because control of the council depends entirely on those election results. However, in the three years with the most notable increase in turnout at this point of the cycle (1983, 1987 and 1991) it coincided with the end of the government’s fourth year in power and therefore an increase in speculation about a general election. The results in each case were widely seen as a test of whether an election would be called. Subsequent elections in 1995, 1999 and 2003 have taken place squarely in mid term and have not had such obvious upward blips in turnout. 1992, a low-turnout local election, followed a month after the general election and suffered from ‘voter fatigue’ and the demoralisation of opposition party supporters.

Overall, local election turnout has fallen much less steeply than general election turnout, suggesting that declining general election turnout has been concentrated among people already weakly attached to electoral politics who were probably not voting in most local elections anyway. The gradual decline probably reflects the erosion of political engagement in general, although the radical reduction in the fiscal autonomy, powers and political salience of local government in the early 1990s and subsequently cannot have helped.

Participation: uncontested seats

A seat not being contested is a sign of the failure of electoral politics. It may be because the area covered by that seat is so monolithic in its political allegiance, or a candidate is so popular as an individual, that there is never any doubt about who will win. But no area is 100 per cent of one opinion. There will be voters who support positions locally in the minority who are literally left disenfranchised and unable to record their view. There will also be people who support the majority view but would like a choice of candidates. Uncontested seats deprive voters of choice, and arguably the representative of the legitimacy that the endorsement of the electorate is able to confer.

The last time a parliamentary vacancy was left uncontested was in 1951. One of the merits of party politics in local elections is that party organisations often try to nominate candidates everywhere and thereby give voters an opportunity to express their views. In highly political areas such as London and most cities, or those covered by parliamentary marginals, uncontested seats are nearly unknown. In recent London and metropolitan areas they have run at less than 1 per cent. Every ward in London in 2006, and all but one in 2002, had an election. The main exception to the pattern of nearly every metropolitan ward being fought is in the very safe Labour borough of Knowsley. In the more urban shire districts (particularly those with elections by thirds) and unitary authorities uncontested wards are rare.

The picture is somewhat different in the smaller shire district authorities in areas of England dominated by one or other party (or independents) and electing in the 1999/2003/2007 year of the cycle. Between 0 and 2 per cent of seats in the districts electing in other years go uncontested, but in the 2003 local elections 460 wards (9.4 per cent of the total) had no elections. This was not unusual. The proportions for 1999, 1995, 1991, 1987 and 1983 were 6.6 per cent, 8.3 per cent, 12.5 per cent, 11.3 per cent and 17.4 per cent respectively. Contestation seemed on an upward path until 2003.

In 2002 the Sunderland Commission expressed concern over the particular problem of noncontestation in Wales. Since the creation of the unitary authorities, contestation seems to have paralleled the previous trends for the county councils. Between 1983 and 1995, a handful of mainly urban districts had elections by thirds, explaining the relatively low figures in 1984, 1986, 1988, 1990, 1992 and 1994. In these councils there was more political organisation and it was also easier for a limited number of activists to ensure that seats did not go uncontested. Going unitary may have helped reduce the proportion of uncontested seats from the very high levels seen in the all-out district council elections.

In Scotland, having one ward in five go uncontested was not unusual, but a notable downward trend set in during the mid-1980s that (with the assistance of the move to unitary status) produced a record low rate of uncontested seats of only 2.8 per cent. In 2003 this rose again. The rise in the proportion of contested seats in Scotland reflects the increase in the political parties’ willingness to fight seats. In the 1970s a contest between all four main Scottish parties was very unusual (fewer than 5 per cent of wards) and around 40 per cent of wards had no party competition. By 1999 nearly a third (32.5 per cent) had four-party contests and fewer than one in ten (8.4 per cent) had no party competition.

In several authorities the pattern of uncontested seats made it difficult or impossible for the voters in the remaining wards to affect control of the council. In 2003 the Conservatives won Fenland in Cambridgeshire at close of nominations – 14 of their candidates were unopposed and they could rely on seven more because there were fewer opposing candidates than seats available. This made 21 councillors out of 40. For deciding who ran the council, the election was superfluous – although as recently as 1995 Labour had won a majority on the authority. The same year in Easington, County Durham, Labour had 28 candidates unopposed and another 12 returns guaranteed. There was only one ward out of 20 in which Labour was not guaranteed at least one seat.

As the Sunderland Commission noted in 2002, there is a pattern of uncontested elections in particular parts of Wales, especially the rural authorities of Gwynedd and Powys. In June 2004 16.5 per cent of wards were uncontested in the Welsh local elections. Gwynedd council filled 35 out of 75 seats without a contest in 2004; 45 out of 73 seats in Powys went similarly unchallenged. The Sunderland Commission noted (para 2.28):²³Report of the Commission on Local Government Electoral Areas in Wales (2002), para 2.28.

Local authority representatives and the Welsh Local Government Association itself told us that they viewed with concern the numbers both of members unopposed and of turnout. Some councillors said that they considered their positions to have greater legitimacy, and that they could exercise greater influence, if they had been elected (as distinct from returned unopposed) and on a high turnout.

How fair is the local electoral system?

The Electoral Reform Society’s report on the 2006 local elections²⁴Electoral Reform Society, 2006, The Great Local Vote Swindle. makes it clear that the local electoral system is not working fairly. A more detailed analysis of local elections is available in that publication.

Wrong winners

Under FPTP there is no guarantee that a party with the most votes wins the most seats, let alone obtains a majority. Factors such as the distribution of the vote, and differential turnout, can cause severe distortions in the relationship of seats to votes. At every set of local elections there are clear examples of parties winning control of councils despite polling fewer votes than another party. In the 2006 London borough elections there were six such cases out of 32 London boroughs:

• In Kingston-upon-Thames the Liberal Democrats won a majority of seats despite the Conservatives having clearly won the largest share of the vote.
• In Islington the Liberal Democrats maintained effective control with 24 seats out of 48 despite Labour winning more votes.
• In Haringey Labour won a narrow but clear majority despite the fact that whether they won the popular vote depends on what method is used to estimate shares of vote in MNTV elections.
• In Brent and Camden Labour won the most votes but the Liberal Democrats have most seats, but not overall control. p In Hounslow the Conservatives won the most votes but Labour has most seats, but not overall control.

There were five such boroughs in 2002 and three in 1998. In 2006 there were also six cases out of the 36 metropolitan boroughs, and several examples in unitary and shire district elections. In 2006 the Conservatives even got more than 50 per cent of the vote in Nuneaton & Bedworth but still won fewer seats than Labour. In Gravesham Labour retained control in all-out elections in 2003 despite having 43 per cent of the vote compared to the Conservatives’ 55.3 per cent.

No system based on constituencies can absolutely guarantee that wrong winners will not occur from time to time, but most electoral reforms would reduce it to a rare phenomenon rather than a relatively common pattern at each set of local elections.

Minority winners

As the chart shows, the popular mandate for parties with a majority of seats in London borough elections has tended to fall over time. Until 1982, the average share to win control of a borough was over 50 per cent. A distinct downward trend has set in since 1994 and by 2006 parties in control of councils averaged only 43.5 per cent support. This fall has paralleled the tendency of local election votes to fragment as voters have exercised more choice.

Another way of looking at the same phenomenon is to examine how many borough councils had a ruling party with more than 50 per cent of the local vote. Before 1982 there were 15 or more in each election, but since then it has tended to fall and in 2002 and 2006 only five boroughs had winners with a majority of the vote. Conversely, winning control with less than 40 per cent of the vote was very uncommon prior to the last two elections, but in 2006 there were nine boroughs in which the majority group failed to reach 40 per cent.

Democracy deserts

Every election throws up a number of areas in which a party has significant support but does not win any seats on the council. The council therefore does not properly represent the range of interests and points of view among the local community. Examples in the 2006 local elections included:

• Cambridge, where the Conservatives polled 22.1 per cent but lost their one remaining seat.
• Peterborough, where Labour won no seats in 2006 despite polling 21.1 per cent of the vote.
• Rotherham, where the Liberal Democrats had no seats for 23.5 per cent of the vote.
• Wandsworth, where the Green Party had 13.1 per cent of the vote and no seats.

These parties were punished by the electoral system for having a point of view that attracted broad support across each local authority, rather than one which appealed most particularly to localised ward interests.

Almost as bad as complete non-representation is the situation where a large share of the vote gives a party only a token voice on the council. A single councillor cannot hope to follow every aspect of council business in full, or offer an effective degree of political scrutiny and competition for the ruling administration. In many local authorities, the efforts of parties reduced to such artificial minorities are hampered even further because they are denied official group status and the benefits that entails.

The Green Party suffers particularly badly from this phenomenon in London. It polled 7.9 per cent of the London vote in 2006 but won only 12 seats – 0.6 per cent of representation. There were ten boroughs where they exceeded 10 per cent of the vote, but only two in which they won more than a single councillor, namely Lewisham (six seats) and Camden (two seats). In Hackney 20.6 per cent only secured a single Green seat, despite the party being the second most popular in the council election. The Conservatives in Barking & Dagenham won a single seat despite polling a higher share of the vote than the BNP, who secured 12 seats. Local election results give a heavily distorted picture of local political opinion, and can falsely polarise local political debate, as in Barking (and Newham, where Labour are the only major political party to be represented).

Wipe-out results in local elections have serious implications for the life of political parties. Sustaining a healthy presence on the council is an important part of the effort to mount a good campaign at general election time, and local political issues keep parties connected with what the electorate are most concerned about. Losing all its councillors cuts a party off from the mainstream of civic and local life. The instability in local government representation – at first the destruction of the Conservatives in the mid-1990s and now Labour losses – is damaging to the parties as institutions and accelerates their decline. The consequences are felt on the national political scene as well as in local government.

Exaggerated mandates and excessive swings

The current electoral system (MNTV but also FPTP) frequently produces outcomes in which one party dominates to the extent that it is difficult for opposition groups to make themselves heard and organise effectively. Comfortable – or even narrow – leads in terms of votes can translate into massive majorities in terms of seats.

Councils dominated by a single party can sometimes produce effective government, but they lack the institutional checks and balances to ensure that power is exercised with accountability. Since the Local Government Act 2000 there has been a change in the role of councillors. The formation of council executives has meant a shift in decision-making power to the leadership of the largest party in councils that have a majority. The role of backbench and opposition-party councillors revolves to a greater degree than before around scrutiny and overview of executive decisions. Scrutiny panels can review and challenge executive decisions. Councillors are also involved in reviewing policy, non-political work such as sitting on panels such as licensing and planning, and representational functions such as casework. When scrutiny is such an important function of the councillor, it is important that the electoral system should facilitate proper scrutiny and this requires a significant body of opposition representatives (should, as they do nearly everywhere, enough people vote for it).

A different – but sometimes related – problem to exaggerated mandates is that in councils where the social and political landscape is fairly homogenous, small swings in votes can lead to massive swings in seats. The classic case of this phenomenon is the London borough of Richmond. A large number of wards are marginal and the council has swung back and forth between a two-to-one Lib Dem majority and a two-to-one Conservative majority on relatively small changes in votes in the last two elections.

Tamworth also has violent swings in seats between Labour and Conservative in local elections for similar reasons. Excessive swings can harm long-term stability because experienced councillors are often swept out indiscriminately.

A constituency link

Councillors serve an important function in representation. They represent relatively small electoral areas, and councils are responsible for a large proportion of services that directly affect people and communities and generate casework. Much casework is in fact more appropriately handled at councillor rather than MP level. The link between councillor and ward is a closer and more direct one than that between MP and constituency – after all, the councillor still normally lives and works full-time in the community defined by the local authority, and is generally more accessible.

Local councillors often have good links with their wards – many live in their wards, know the area and its problems well and are quite wellknown amongst their constituents. These links appear to be no less strong in multi-member wards and this strength of identification is evidence against the idea that a monopoly of representation by a single individual is necessary for a strong constituency link to be maintained. However, these links are not perfect. In safe seats members may not feel the same incentive to maintain their links with all communities in their wards, and constituents whose views differ widely from their councillors’ can feel unrepresented. Voters in single member wards lack a choice of representative to approach, and those in multi-member wards usually only have councillors from one party.

Nor do existing ward boundaries necessarily reflect the communities that people identify with, given that they have to be drawn for a set number of members and within acceptable limits as regards equality of numbers. In rural and small town areas in particular, ward boundaries can frequently cut across communities.

What are the implications for government stability?

Stable governing arrangements for local authorities, while not as important as in national government that deals with the big issues of economic and foreign policy, are still important. Many local government decisions in terms of planning and service provision have long term effects and some stability in local leadership is a desirable criterion.

However, in some areas there is an excess of stability, in that councils are ruled for extremely long periods by the same political group. Labour has run County Durham since 1919 with minimal opposition, and the London Borough of Barking & Dagenham since its creation in 1964. Conservative rule in the city of Westminster dates back more than 100 years. In some of these areas long continuous rule is justified because the voters have consistently given the party more than 50 per cent of the vote, but often such strong support gives a ruling group a near-monopoly on council representation. A very long time in power can result in a closed style of government, disconnection with local needs and even – as the Poulson scandals in the 1970s showed – corruption.

Less defensible are those cases where a party has continuous control without majority support. Labour has run Rotherham continuously since 1973 and currently enjoys a massive majority (55 out of 63 councillors) but in terms of votes scored only 38.2 per cent in 2004 and 42.9 per cent in 2006. Labour has enjoyed majority control of Nuneaton & Bedworth council since the first elections in 1973, but the borough has given the Conservatives the lead in terms of popular vote in three of the four elections since 2000 (in 2000, 2004 and 2006) compared to a Labour lead in just one (2002).

In councils with annual elections under FPTP there can be frequent alternation in power, often based on very small changes in voting patterns. In Cheltenham, for instance, the Liberal Democrats lost the council to no overall control in 1999, the Conservatives won control in 2000, lost it to the Liberal Democrats in 2002, who then lost to no overall control in 2004. In Welwyn Hatfield the council flipped back and forth between Labour and Conservative with three changes in control between 1999 and 2002. If stable arrangements in local government are to be a priority, annual FPTP elections seem an impediment. The options are to move to four-year terms, or at least to have more frequent elections for part of the council by a system that does not generate such wild swings in election results as FPTP.

The introduction of mayoral elections from 2000 onward in some local authorities has had implications for the question of government stability in local electoral arrangements (the system used to elect mayors is the subject of another section of this report). Executive mayors appoint the ‘government’ of the local authority, and this does not depend on a majority in the council chamber. Questions of ‘government stability’ become irrelevant in councils which exist alongside executive mayors.

In the absence of government formation, the functions of the council in a mayoral system are based primarily on representation and scrutiny. As well as geographic representation of each ward, there is a broader sense in which the council represents local society as a whole and can therefore be expected to reflect significant shades of local opinion as well as gender and ethnic balances. In several local authorities with elected mayors, the mayor’s party has a grossly excessive majority, as in Newham where Labour has the mayoralty and 54 out of 60 seats (there are two other groups of three each) and in Watford where the Liberal Democrats have the mayoralty and 29 out of 36 seats (the next largest group has three councillors). With these situations it is extremely difficult for opposition parties to exercise their constitutional role of scrutinising the executive and providing alternative local political leadership.

Are women and ethnic minorities fairly represented?

Women, ethnic minorities, young people and those not from a professional background remain significantly under-represented on local authorities, and in some cases are conspicuously absent. At that rate of growth, it would take 113 years for councils to reach parity between men and women. The Electoral Reform Society’s own research4 suggests that the 2006 local elections resulted in practically no increase in the proportion of women on councils.

In Scotland, the number of women on councils has been stuck at 22 per cent and even declined slightly since 1997 In Wales, 21.8 per cent of councillors are women and Northern Ireland, where local government is elected by STV, has 22 per cent, up from 19 per cent in 2001.5 With some 460 local councils across the UK, and differences in authority type and electoral arrangements, it should come as no surprise that there are huge variations in terms of diversity. London and metropolitan boroughs tend to have more, while English county councils tend to have fewer women at just 24.3 per cent. Two councils, Redcar & Cleveland and Islington, have more women than men in their council chamber. Three councils have just one woman elected: Inverclyde in Scotland, Merthyr Tydfil in Wales and Strabane in Northern Ireland. Tamworth is England’s poorest local authority for gender balance, with only three women among its 30 councillors. The picture also varies a little with party, though no party is close to balance.

4. Electoral Reform Society, 2006, The Great Local Vote Swindle.
5. COSLA, 2003, Scotland’s Councillors 2003 http://www.cosla.g ov.uk/attachments/ publications/cllrsurv ey2003.pdf; Welsh Local Government Association, 2004, Census of Local Authority Councillors; http://www.wlga.go v.uk/uploads/public ations/932.pdf; NILGA, 2005, National Census of Local Government Councillors in Northern Ireland 2005 Key Findings www.nilga.org/cens us05 and Northern Ireland Assembly Research & Library Services Research Paper 28/2 (April 2002) http://www.niasse mbly.gov.uk/io/rese arch/2802.pdf.

Ethnic diversity

In terms of ethnicity, 96.5 per cent of councillors are white, while 3.5 per cent come from an ethnic-minority background. National census figures estimate that 7.9 per cent of the UK population is from a BME (Black and Minority Ethnic) background. In London, the figure rises to 14 per cent for ethnic-minority councillors, well under half the proportion of the population.²⁵Greater London Authority Race Equality Scheme 2005–2008 http://www.london.g ov.uk/mayor/equaliti es/docs/race_equali ty_scheme_full.pdf].

In Scotland and Wales, the ethnic balance of councils is also a poor reflection of the ethnic make-up of the population as a whole: 1.1 per cent of Scotland’s councillors are from minority ethnic backgrounds, compared to 2 per cent of the Scottish population. In Wales, 0.8 percent of councillors had an ethnic-minority background, in a population of around 2 per cent.

Age

Councils also poorly reflect the age profile of the wider population. In every UK nation, the average councillor is in their 50s and, with the exception of Northern Ireland, younger people are very poorly represented.

The fact that the balance of the sexes is somewhat better on councils than in the House of Commons is of little encouragement, when we consider the lack of competition for local election nominations. The question of the under-representation of women on councils has not received anything like as much attention from the media, women’s lobby or political parties. The Labour Party has used positive action to ensure that one-third of candidates selected are women in some areas. This has resulted in an increase in the party’s proportion of women, particularly in London (where all-out elections make it easier to ensure that one woman is included on the ballot paper among the party’s three candidates). Other parties have been reluctant to follow suit.

A key feature in local government elections which discourages diversity among candidates and representatives are single-member elections. (Although many UK councils operate on the basis of multi-member wards, election by thirds means parties are only selecting one candidate at a time for each contest and voters can support only one candidate; it works essentially in the same way as a single-member constituency election).

When parties choose just one candidate per seat, they tend to play it safe, picking a stereotypical candidate. It can be hard for women, minority ethnic or young candidates to break through. In wards where parties select a team of three candidates, it does make it easier to spot under-representation, for example if a party puts forward three men and no women. However, voters are reliant on parties to select a more diverse team, and this is not always the case. The system offers limited voter choice or flexibility. Women’s representation is thus only a little better in London’s three-member MNTV elections than in the rest of the country where election takes place by thirds under a straightforward FPTP system.

Other features of local government also dissuade a wider range of people from considering standing. The timing of meetings, workload, level of support and perceived low status of local government can particularly discourage women, those with caring responsibilities and those who work full-time (indeed a high proportion of councillors are self-employed or retired). It may be that the nature and organisation of council work, as well as terms and conditions for councillors will have to be rethought if council chambers are to be fully accessible to candidates from across society.

A more proportional system would favour women and other under-represented groups by giving parties an electoral incentive to select a balanced team of candidates and give voters the chance to prioritise candidates they identify with.

Alternatives

More mayors?

One possible means of reforming local electoral structures is to create more directly elected mayors. It is questionable, given the rather weak effects of this reform on turnout, whether this will accomplish the aim of increasing participation. It is also not a course of action which has met approval in the local authorities that have recently had referendums on whether to adopt the mayoral system. In total, of 35 propositions to establish an elected mayoralty, only 12 have been passed. The success rate since the first elections (outside London) in May 2002 has been particularly low, with a 55 per cent ‘Yes’ vote in Torbay being the only win, for five losses. The Local Government White Paper of October 2006 proposed making the process easier if a council wishes to adopt the model.

If the mayoral model is to be more widely used, however, the question of how councils should be elected will not go away.

Annual or all-out?

The principal argument for annual elections is the increased accountability this offers to local electors. All-out elections give a ruling administration four years in which its priorities may diverge rapidly from those of the electorate. Government thinking in the late-1990s was in favour of annual elections because of this increased accountability. The February 1998 Green Paper Modernising Local Government – Local Democracy and Community Leadership, floated the idea of annual elections for London borough councils, which would necessitate a pattern of three-member wards. The subsequent July 1998 White Paper, Modernising Local Government – In Touch with the People, proposed election by thirds for all unitary authorities including London boroughs. The Local Government Commission for England worked during its review of London ward boundaries between 1998 and 2002 to produce a pattern of wards that would permit annual elections. The government dropped the plan to introduce annual elections, but its consequences are still there in the map of London’s wards.

Government thinking now seems to be more inclined towards all-out elections for local authorities, having absorbed evidence that these elections seem to attract higher turnout and it would be easier to rationalise the complicated calendar of local elections. Most proportional systems would require a move to either all-out, or at most by halves, elections unless the size of local authorities is to increase markedly. However, the rationale for frequent appeals to the electorate may be weakened if all-out elections gave a more representative result than is possible under FPTP.

Single-member FPTP

Rationalisation of FPTP through the removal of MNTV wards is an attractive option for opponents of electoral reform. It was recommended by the minority on the Sunderland Commission and various other inquiries. The Wheatley Commission on local government in Scotland in the 1970s strongly recommended a singlemember pattern and strongly condemned all-out elections in multi-member wards.

Single-member FPTP would mark a move away from the blatant irrationality of MNTV and solve some of the problems associated with it, such as excessive swings in representation and a party enjoying a monopoly of representation on perhaps a small plurality vote. There are many examples of how MNTV can give a monopoly of local representation to a party with a very small share of support. An example was the 2004 election in the Calder ward of Calderdale metropolitan borough council, where 28.5 per cent of the vote was enough for the Liberal Democrats to win all three seats in the ward.

However, single-member FPTP might introduce other problems. Permitting MNTV wards in counties and districts electing all-out increases the flexibility with which the boundaries can be drawn. There will be cases in which singlemember district boundaries will be artificial and

not reflect community identity, while allowing a multi-member ward could permit more sensible lines to be drawn. This was the rationale behind permitting multi-member wards in county council elections for the first time (in recent years) in 2005.

MNTV also does more to encourage parties to put forward balanced and diverse slates of candidates than single-member FPTP, although the effects in practice seem to be limited. It also gives voters a choice of which representative to approach and does not seem to result in a dilution of the quality of representation. An ERS survey²⁶Electoral Reform Society, 2003 What Councillors Think (What councillors think, 2003) found that, if anything, councillors representing multi-member wards felt that they had better contact with electors than those in single-member wards. Having several members may improve the quality of representation by giving representatives an opportunity to specialise in particular aspects of the role and become expert in, say, local housing issues.

Single-member FPTP might lead to an increase in the number of wards going uncontested, for two reasons. One is that the smaller the constituency size, the likelier it is to be socially and politically homogeneous. Cutting a local authority into, say, 60 small areas rather than 20 rather larger ones, is likely to create more wards in which there is no effective political competition. The other reason is simply that parties have only modest campaigning resources and potential candidates. With single-member wards in a 60-member council, a party would have to find 60 candidates to give every voter a chance to support it, which weaker parties are unlikely to muster. With three-member wards, the target is a more manageable 20 in order to achieve coverage of every ward.

It therefore seems apparent that moving to single-member FPTP is not the simple answer that some of its proponents claim.

Party list PR

It is technically possible for party lists to be used in local government elections, and this is not uncommon in other European countries, such as France, the Netherlands and Italy.

However, wards are a long-established feature of British local government, and representation of constituents at ward level is a very important feature of the councillor’s role. The functions of local government in Britain are such that dealing with housing and social services casework by an individual who knows the immediate locality is vital. The abolition of wards and election of councillors from lists covering the entire local authority would be a radical – and surely unwelcome – step, and increase the distance between local government and the governed.

A closed party list would also reduce accountability of councillors to voters; it would be possible for a member giving poor service to constituents to be confident of re-election because of his or her standing in the party hierarchy.

In addition, party list systems make it difficult for independent candidates to compete on equal terms with the parties.

For all these reasons, party list systems appear most unsuitable for use in local government elections in Britain.

Mixed Member Systems

MMP (also referred to in Britain as AMS, Additional Member System) for local elections would mark less of a break with local representation than party list PR. In areas with existing multi-member wards it would not even involve having larger wards – in fact, it might be possible to elect local councillors from smaller areas than at present.

In the metropolitan boroughs, instead of having three members per ward, and elections by thirds in three years of a four-year cycle, each ward could have two members and elections every two years. The other councillors would come from the city-wide top-up. For instance, in Manchester, with 32 wards, there would be 16 city-wide councillors elected alongside the 32 ward representatives every two years. Alternatively, the council could elect all-out every four years and use single-member wards half the size of current wards.

A case can be made for city-wide representatives alongside ward representatives in order to obtain a broad strategic view of the needs of the local area as distinct from an aggregation of parochial concerns. This would not be a completely new departure; elected Mayors are supposed to do this. Before the Local Government Act 1972 councils did have members bearing the title of alderman, who were indirectly elected (by the councillors) who served the locality as a whole. Using a modernised title indicating this senior status for list members might be useful in dispelling the idea that members elected from the list are second-class representatives.

Election results under city-wide AMS in Manchester would seem to deliver outcomes that are reasonably proportional without fragmenting political representation unduly. Assuming that the list votes would be the same as the total votes in the ward elections for each party (not a safe assumption, as more voters would have the opportunity to choose minor parties), the results would be as shown on the following page.

The Greens and Conservatives would win representation for their votes. Labour’s stronger showing in 2006 took the party to only just short of winning a majority of seats even with a proportional top-up, so an overall majority in Manchester for Labour would be an achievable goal. The same is true in other areas where one party has a clear lead in the popular vote.

However, while MMP would work in areas where party politics is well established, there would be problems in transferring it to areas where councillors’ local credentials are often more important than the party label. There is often no sense, of the sort required by a top-up list, of a districtwide vote for one party or another and this would be an entirely artificial introduction. It could end up forcing the politicisation of councils where people do not really want it.

MMP would also introduce the vexed issues of relations between listand constituencymembers into local authorities. Patterns of conflict and ‘cherry-picking’ of casework in marginal wards could develop if list members were permitted to do casework. If protocols could be devised to ban list members from doing casework this could cut the members off from contact with the issues citizens care about, and leave it unclear what such members were supposed to do. It is likely that list members would tend to represent opposition parties, because a party with a majority or nearmajority of seats would be likely to win that majority from the constituency-based elections, as in Scotland and Wales. A ward-oriented administration and a strategic-oriented opposition seems to be the worst of both worlds.

In 2000 the Kerley Working Group in Scotland reported that:

…we do consider that the two classes of member that AMS would produce, with the possible disadvantages [noted by the Working Group], are distinctly less attractive than the single type of member that STV provides.

Similarly, the Sunderland Commission concluded in 2002 that it could not support MMP for use in local government in Wales:

The objections to list systems… apply with equal force to the Additional Member component here. Also, we could not see how a local authority could benefit from a division of members, some of whom would have constituency responsibilities and some not.

While there is a case for MMP in local government, there seem to be severe drawbacks that make the system unattractive for this use unless there is no acceptable alternative.

Single Transferable Vote

Unlike other proportional systems, STV would not require candidates to line up in parties. Preferences are cast for candidates as individuals. Voters may well, in many areas, choose to follow party loyalties but this is not enforced by the electoral system. STV would allow independents and minor parties to come forward and give them a fair chance of winning election. In many areas the parties will stand fewer candidates than there are vacancies available (because they do not stand a chance of winning every seat). This would allow the boundaries between partisan and non-partisan politics to be blurred: independents who share many political views with a party, but do not wish to be under party discipline, could stand in alliance with that party.

STV would also preserve the positive features of multi-member representation that are visible even with MNTV elections. Parties would be encouraged to offer representative slates of candidates. Competition and choice would give incentives to individual councillors to offer good service to local electors. STV would enable the pattern of specialisation and ward links that currently exists in multi-member wards to continue (although in modified form, as multi-member wards represented entirely by one party will change from being the norm to being a very rare phenomenon).

STV would also get around the boundary definition problem of a uniform single-member FPTP system. Boundaries can be drawn to reflect genuine communities of different sizes, because the number of councillors per ward is flexible.

Reviews of local government electoral systems in Scotland and Wales have both concluded that STV is the preferred system. The Sunderland Commission in Wales supported STV because:

1. STV is a proportional representation system. As such, it has far greater capacity than do majoritarian electoral systems to secure representation of the diversity of views within communities… STV has the potential to address the ‘single-party state’ problem, and facilitate the election of members who can provide the strong check and balance which we believe the introduction of new political management structures in local authority requires. 2. STV has the potential to address other aspects of diversity. 3. STV offers voters a far greater opportunity than does FPTP to express preferences between candidates. 4. STV is a wholly constituency-based system, and as such it meets local people’s concerns to have identifiable local elected representatives to speak on behalf of themselves and their areas.
2. We see no reason to think that the introduction of STV will result in lower turnouts and fewer contested elections – indeed, we think the opposite may be true.

Political circumstances being different in Scotland than Wales, the case for STV proposed cumulatively by McIntosh and Kerley was accepted and legislated for in 2004.

Conclusion

We conclude that both FPTP and MNTV are unsatisfactory means of electing local authorities.

We also note that several alternative systems appear to be unsuitable, the principal stumbling block being that they would be inimical to the position of independent councillors because they would not enhance voter choice of candidate.

The Single Transferable Vote, however, does not have these drawbacks and has positive merits of its own. This has been recognised in the use of STV in Northern Ireland local elections, the Sunderland, McIntosh and Kerley reports in Wales and Scotland and the subsequent introduction of STV in local elections in Scotland from 2007. The Electoral Reform Society urges the introduction of STV for local government elections in England and Wales as well as Scotland and Northern Ireland. We are confident that it would be a major contribution to renewing and modernising local government.

We do not necessarily expect immediate acceptance of our position. But we would welcome the opening up of an informed debate and wide public consultation about the local government electoral system in England. Should a general reform be delayed, we see no reason at all why there should not be ‘pilots’ of different methods of elections in authorities that so choose.

Where now?

In this section we outline the current state of affairs with the electoral systems in operation in Britain, recent official reviews and how policy might develop from this point, before drawing attention to a model of developing future policy in a genuinely democratic and deliberative way.

First Past the Post – broken beyond repair

We believe the case for changing the system of election for the House of Commons is truly compelling.

Earlier in the report we have spelled out quite how badly it fails to satisfy the government’s original criteria given to Jenkins. The original Jenkins criteria were defined with reference to elections for the House of Commons, and they remain a good statement of the desirable features of an electoral system for Westminster. To recapitulate, the criteria are:

• the requirement for broad proportionality;
• the need for stable government;
• an extension of voter choice; and
• the maintenance of a link between MPs and geographical constituencies.

As we have seen, the current FPTP system decisively fails on the first and third counts, is increasingly unlikely to pass on the second, and lack of majority winners at constituency level is undermining the fourth count. It also sets up barriers for the representation of women and minorities. An increasing concern is the differential turnout and concentration on marginal seats that gives different social groups unequal power – the old more than the young, the middle class more than the working poor, the white more than the black, the south more than the north, the motorist more than the public transport user. The political process therefore magnifies, rather than compensates for, existing inequalities in society.

More broadly, FPTP has fostered a boring, sterile, adversarial mode of politics. It has a direct impact on elections, making then uninteresting by being predictable and putting many electors in a position in which their votes have little chance of influencing the outcome. It also shapes our political culture by forcing parties to compete for the swing voters in marginal constituencies, marginalising the role of local candidates, feeding them back the slogans that work best in focus groups and giving politics an artificially adversarial nature.

For most voters, the ya-boo of present party politics is a turn-off: polling has shown that one thing voters really want is for politicians to “stop their bickering” and start working together in tackling problems. But at the same time people do want politicians and a political culture that facilitates an honest exchange of views. Proportionality could encourage parties to ‘be themselves’ and by doing so could encourage more differentiation between the parties, widening the choices presented to electors, while at the same time discouraging a negative, adversarial approach to putting their case before the electors. Parties which might need to find coalition partners after an election are less likely during an election campaign to unnecessarily slag off opponents whose support they might need in forming a government. With coalition government more of a possibility, parties will still want to present their distinctive policies and to argue why they are better than others, but in doing so they are likely to be more measured in their approach (as was seen to some extent in the Labour and Liberal Democrat campaigns in the run up to the 1997 general election, and in the 2003 Scottish Parliament election campaign).

There is, therefore, an urgent need to complete the unfinished business of the Jenkins Commission.

The government should redeem its 1997 promise to consult the electorate on a replacement for the FPTP electoral system to be used in the House of Commons.

How to decide on a new system

Major constitutional reforms are, it is generally conceded, best settled in a referendum. The votes on devolution to Scotland, Wales and Northern Ireland in 1997 and 1998 established or confirmed a precedent that consulting the electorate is appropriate in cases of major change. Even relatively minor changes involving elected mayors, and regional government for London and the North East have been the subject of referendums. The electoral system for the House of Commons is perhaps the issue on which taking the issue to the electorate is most appropriate, because otherwise MPs remain the only group who decide their own electoral system.

The referendum, however, can be a blunt instrument. It is arguably not appropriate for every case. Issues need to reach a certain threshold of significance before it becomes worthwhile, and they are not used (and need not be used in British constitutional practice) for every single local government reorganisation, electoral change or European treaty. A winning party with a firm manifesto commitment to proceed on something does not require a referendum, as with the abolition of the hereditary peers in 1999 (and again in 2007…) even if it is a significant constitutional change. Some possible electoral system changes are arguably too minor to merit a referendum, for instance a move to another majoritarian system, the Alternative Vote.

Referendums are a particularly imperfect way of deciding between more than two possible courses of action. With electoral systems, there is a multiplicity of theoretical options for change and there needs to be some way of narrowing down the choice. One potential course of action is a multi-choice referendum, but these have severe drawbacks. Narrowing down the choice to three or four systems to put on the ballot still requires a decision procedure which could be contentious. Particularly in this context, a multioption referendum could produce a situation where there is no agreed ‘winner’, i.e. if several alternative systems are offered and First Past the Post is the choice of the largest single group of voters but well short of a majority. Even if the multi choice referendum rules are AV, FPTP supporters could claim that by the logic of their favourite system, they are the real winners! In a situation such as FPTP 35 per cent, STV 40 per cent, MMP 25 per cent there are concurrent majorities for proportionality and single member constituencies, but the system that combines these is actually the least popular choice.

Multi-choice referendums are therefore likely to confuse the issue and possibly create a smokescreen for a retreat from electoral reform. Otherwise, there are two broad ways of proceeding. One, as took place in Italy in the early 1990s, was to approve the principle of change first and sort out the details in the next stage. The other, which took place slightly later in New Zealand, was to decide which change option was to go forward through a ‘preferendum’ and later pit the winner of that ballot against the status quo in the final referendum. Although reform happened in both countries, each is not without its risks. As one has seen so often with Lords reform, agreement on the principle of change can often fall apart when it comes to specifics. It is also possible that opponents of change can sufficiently affect the reform proposal that it is weakened and unable to gain enough popular support to pass. In Australia, for instance, the monarchy survived despite a theoretical republican majority because the form of republican presidency offered (indirect election) was not attractive to voters.

Italian constitutional arrangements made it procedurally relatively easy to strike down the status quo because there existed a right of popular veto – i.e. electors could insist via a petition that repealing an existing law be put to a vote. In Britain there is as yet no such procedure that can be launched from outside government, although there is a precedent in the form of the Scottish devolution referendum which passed the principle but left the detail to parliament. However, in the case of electoral reform where there are several ways of proceeding, such a procedure would not entrench it enough and risk a future government rowing back on essential properties of the new system.

If the choice is to be between the status quo and a solid alternative proposal, the question arises as to how that alternative is reached. Deciding which questions are asked is an even stronger form of political power than deciding what the answers may be, and for this reason some sort of external authority – beyond government’s traditional decision making processes – is needed to establish a proposal for a permanent reform that can be put before the electorate. There are three broad ways that could be used to formulate a proposal – agreement between political leaders, an expert commission or a Citizens’ Assembly.

Speaker’s conference

In the past, important electoral reforms have been discussed and brought forward through several mechanisms, including straightforward legislation. But major changes have often involved a wider process of some sort to build in consensus and legitimacy at an early stage. One sort of institution that has been used in the past is the Speaker’s Conference. Such a conference first took place in 1910 to attempt to resolve the bitterly fought issues around the Budget and the powers of the House of Lords, although it produced no consensus. In 1917 there was another such Conference specifically

about the electoral system, but in the end the chances of electoral reform were lost during Commons proceedings in 1917-18. The 1944 Speaker’s Conference on boundary reviews was successful and established the system we still have for drawing constituency boundaries. Subsequent Speakers’ Conferences have been less ambitious and focused on smaller electoral reforms such as changes to electoral law and the voting age.

The ostensible advantage of a Speaker’s Conference is that it binds in the party leaderships to its conclusions, and also allows a frank exchange of views to take place in private before conclusions are reached. On the other hand, history shows that even when the party leaderships are bound in, the recommendations can still fail to be implemented. Before the Electoral Commission was established, they may have been a good way to reach agreement on relatively peripheral issues but the likelihood of producing an inter-party consensus on electoral reform is remote. The process also has a rather elite-based, topdown flavour of sensible chaps coming together to agree how things should be run, which is no longer a model that looks likely to entrench long term changes.

Jenkins Mark 2

Another option for deciding the way forward would be to establish another official commission of inquiry. Much of the Jenkins report stands the test of time. Its analysis – to say nothing of its elegant prose – is still a vital contribution. So are the criteria set by the government in giving Jenkins his remit. However, there are a number of areas where additional evidence has emerged since 1998 and review and updating in the light of this evidence is necessary. The Independent Commission on PR’s report in 2004, and this report, have assembled some of that evidence. Some questions of principle emerge:

1. First among these is the experience of how mixed member systems have operated in a British political context in Scotland, Wales and London. The rancorous experience in Wales in particular poses questions as to the viability of creating two routes into the House of Commons. Is British political culture, with its tendency to privilege a constituency mandate over any other form of election (particularly in the case of Westminster), capable of adapting to a mixed system? 2. Another consideration is ballot complexity, which unfortunately seems to be a barrier to participation of some voters in politics. MMP in Scotland and Wales seems to need explication at each use, as the rise in spoilage in 2003 showed. The London elections of 2004, in which voters completed one List PR ballot for the European Parliament, one List PR ballot for the Assembly list, one FPTP ballot for Assembly constituency, and the double-column semi-preferential SV ballot for Mayor, was probably too much. The Jenkins system of AV+ would involve a preferential constituency ballot and an open list PR ballot for the county units. 3. In 1998 Jenkins considered the Alternative Vote in the light of the historically unusual 1997 election, in which AV would have swollen the Labour majority and harmed the ability of the Conservatives to form a large enough parliamentary opposition. With some distance and perspective on 1997, the arguments on AV are different, particularly as so few MPs have the support of the majority of the voters in their constituencies.

It is open to the government to form another Royal Commission, a Jenkins Mark 2, to consider the electoral reform debate in the light of new evidence and the new political context. This option has merit in providing another expert view, but on the other hand official reviews and inquiries, however meritorious, can often end up on the shelf. One only needs look at the Jenkins report, and the Richard report on Wales, to see that. Circumstances change, priorities emerge, and all the report has achieved is a contribution to the academic literature – and a stay of execution for the electoral system and a breathing space for the government of the day. An inquiry as such is not worth much without a concrete pledge for the government to hold a referendum, or legislate, based on its recommendations. And even then, concrete pledges are sometimes not worth much – as with the 1997 commitment to hold a referendum on Jenkins.

Deliberative democracy: the Citizens’ Assembly

British Columbia The Citizens’ Assembly concept first took root in the Canadian province of British Columbia. Before the 2001 election, the provincial Liberal Party made a commitment to:

• Appoint a Citizens’ Assembly on Electoral Reform to assess all possible models for electing Members of the Legislative Assembly (MLAs), including preferential ballots, proportional representation, and the current electoral system;
• Give the Citizens’ Assembly a mandate to hold public hearings throughout British Columbia;
• Hold a province-wide referendum on any change of electoral system that the Citizens’ Assembly might recommend.

To their credit, even after gaining considerably under the present system (the Liberals won 77 out of 79 seats on only 57 per cent of the vote), they fulfilled their pledge. Gordon Gibson was appointed to advise on the mandate and make up of the Citizens’ Assembly, which he did, in the Report on the Constitution of the Citizens’ Assembly on Electoral Reform, published in December 2002. In April 2003, the Assembly was set up, with Jack Blaney, former president of Simon Fraser University, unanimously elected as chair.

The Citizens’ Assembly marked a new approach to representative democracy. As Blaney noted, “never before in modern history has a democratic government given to unelected, ‘ordinary’ citizens the power to review an important public policy, then seek from all citizens approval of any proposed changes to that policy. The British Columbia Citizens’ Assembly on Electoral Reform has had this power and responsibility and, throughout its life, complete independence from government.” The 160 members of the Assembly comprised of a man and a woman from each of the 79 provincial ridings chosen randomly from the populace, plus two Aboriginal members. The chairman was the 161st member. As the Assembly’s final report asserts:

The members of the Citizens’ Assembly presented a wide variety of backgrounds and experience: they reflected the diversity of this province. They also had varying degrees of knowledge and understanding of electoral systems, so a three month Learning Phase was provided to prepare members for the tasks and challenges represented by the mandate.

As a result, six sessions were held between January and March 2004, where Assembly members heard lectures and participated in discussion groups, and where they discussed the principles and practices of electoral systems.

Following this, fifty public hearings were held during May and June, where between four and sixteen Assembly members attended. In the course of the hearings, 3,000 British Columbians attended presentations and made comments and suggestions. According to the final report, “the dominant themes of the presentations included the need for change, more proportionality, local representation and increased voter choice”. Meanwhile nearly 1500 individuals made written submissions to the Assembly. At the conclusion of the public hearings, between September and November, the Assembly met to review what they had read and heard, with discussions framed “within a well-defined set of democratic values: fair (proportional) electoral results, effective local representation, and greater voter choice”.

In October, the Assembly compared two prospective models for a new electoral system – the Single Transferable Vote and the Mixed Member Proportional system. According to the final report, “members explored not only how each system worked and the consequences of adopting one or the other, they also considered the effect each system would have on how our political parties work, on the legislature, and on the pattern of government in the province”. On 23 October, the Assembly voted on the questions “which of the two alternatives would best serve British Columbia?”, and by a margin of 123 to 31 voted in favour of STV.

The next stage was to consider STV vis-à-vis the present First Past the Post system. “Members had decided to recommend a change only if they were convinced that the proposed alternative was demonstrably superior to the current system”. Thus on 24 October, the Assembly took two votes. On the question “do we recommend retaining the current FirstPast-the-Post electoral system in British Columbia?” 11 members voted Yes, and 142 voted No. On the question “Do we recommend the STV (BC-STV) system to the people of British Columbia in a referendum on May 17, 2005?”, 146 said Yes, and only 7 said No.

Finally, on 10 December, the Assembly published its final report, Making Every Vote Count: The case for electoral reform in British Columbia, recommending that STV be adopted in British Columbia. The specific system for use in British Columbia, BC-STV for short, would see between two and seven MLAs returned by each electoral district, with the proviso that “given that achieving proportional electoral outcomes is a primary reason for recommending BC-STV, using larger rather than smaller numbers of members per district should always be preferred when drawing district boundaries. While some very sparsely populated areas may require districts with as few as two members, the principle of proportionality dictates that, in the most densely populated urban areas, districts should be created at the upper end of the range.”

The Assembly’s proposal was put to a referendum in May 2005. In the run-up to the referendum many Assembly members were in effect ‘citizen advocates’ for the system they had designed. This made the referendum campaign itself more meaningful and connected to the people than many such campaigns – the proposal was not advocated by ‘experts’ flying in from far away, but by local teachers, farmers and home-makers in each district. By using a participatory mechanism in the first place, the Assembly process continued the dynamic of involving the electorate in the decision.

The BC-STV proposal was approved with a 57 per cent vote in favour in the province, and the support of the majority of voters in all but two of the existing legislative districts. However, the government had stipulated before the process began that it would need a 60 per cent vote in favour to pass, and the reform process has unfortunately not yet concluded.

Beyond British Columbia The experience of British Columbia has attracted interest and approval from many observers and political scientists, and has been followed by the establishment of similar bodies in other Canadian provinces, including the largest, Ontario. The first European Citizens’ Assembly sat in 2005-06 in the Netherlands to debate that country’s electoral system.

A Citizens’ Assembly is very different from a referendum, although not of course an alternative or replacement. It is deliberative as well participatory, and enables an argument to be developed beyond the Yes/ No simplicities of a referendum. Control of the agenda of politics is just as crucial as the process of making a decision, and the Assembly mechanism turns the agenda behind the referendum over to the people.

The idea of a Citizens’ Assembly may seem risky, or overly idealistic, but the Canadian experience suggests that it works. It is capable of catching the imagination, and democratising politics in a way that an official review – particularly a desk-bound, internal official review – cannot. It is capable of exploring and making accessible complex issues, in a way a referendum alone cannot. It is, in itself, a radical step towards renewing democracy. Why not do it?

Alternatives to FPTP

The Jenkins Commission itself conducted a thorough, thoughtful review of alternative electoral systems and much of the logic of the case it outlined in 1998 remains valid. The government’s own review needs to respond to the arguments laid out by Jenkins and his colleagues. Because that case has still not been answered, and because we believe that there should now be a full public review, in this report we will not engage in an exhaustive dissection of alternative systems but outline the arguments the Electoral Reform Society finds persuasive.

Party List PR

List PR has never exercised much attraction for electoral reformers as a model for the House of Commons. At one end of the scale is a nationwide PR system as in Israel or the Netherlands. At the other is Spain’s, based on relatively small local electoral areas.

A system with large lists would be highly proportional, and very likely fair to women and minorities, but would be likely to lead to a fragmented parliament (making stable government significantly more difficult) and destroy the relationship between MPs and geographical constituencies.

Small lists would be less proportional, and not do much to extend voter choice, but would preserve some sort of constituency link. Most variants of List PR would also effectively deprive voters of much choice over which individuals, as opposed to parties, end up representing them. If closed lists are used, this means an overt lack of voter choice. It would not enable voters to choose candidates on grounds other than party allegiance. It does not seem suited to the functions and nature of the House of Commons and we do not recommend it.

Alternative Vote (AV)

While List PR would be a complete break with past practice, the Alternative Vote (AV) is a very cautious step towards a reformed electoral system. It would retain the current pattern of single member constituencies and therefore the single-channel constituency link that many MPs value so highly. It would give this link greater legitimacy by providing each MP with some sort of majority mandate. As a majoritarian system, it is probably more conducive to stable government than FPTP, because it seems most likely to produce either comfortable majorities or situations where a proper coalition is required, rather than minority government. It extends voter choice a little, by enabling voters to choose their real first choice party without fear of wasting their vote or letting in their least favourite party – although it does not give people much choice of candidate.

AV is also not reliably more proportional than FPTP. Although it may be in most elections, it would not have been in 1997 – which concerned the Jenkins Commission. As a single member district system it does not remove the causes of unfair outcomes for women and minorities. It would perpetuate, and maybe worsen, the pattern of regional deserts and strongholds for each party that is already apparent.

Mixed Member Proportional and other Additional Member Systems

There are three variants of AMS that are relevant to discussing the House of Commons – Mixed Member Proportional (MMP, as in Scotland), its close cousin AV+ (AV with a small proportional top-up, as recommended by Jenkins), and Mixed Member Majoritarian (MMM, as in Japan).

This sort of system preserves single-member constituencies, but it does so at the expense of also having list members who have little direct accountability to the electorate. This is not an attractive proposition if voter disengagement is considered to be in part a consequence of the perceived distance between electors and their representatives. Constituency contests would still suffer from the prevalence of safe seats and lack of real competition, from high numbers of wasted votes and the need for tactical voting to make vote count.

Certainly some of these problems could be overcome by the AV+ system recommended by the Jenkins Commission, but there are serious objections to that system on account of its complexity, requiring voters to vote preferentially for constituency candidates as well as for a party list. It would be especially complex if, as recommended by Jenkins, people are given the option of voting for particular list candidates on their party’s list.

As well as these objections, the experience of the Scottish Parliament and, in particular, the Welsh Assembly, offer evidence on how mixed systems might operate in the British context. Relations between constituency and list members in Wales have been particularly conflict-ridden, and both in Wales and Scotland there is a lack of clarity about how the list members go about representing the electorate. In a system with such a strong tradition of constituency representation there is an obvious danger of these problems being reproduced in the House of Commons. In a situation where MPs are accustomed to notifying each other when they make public appearances in a colleague’s territory, free-floating regional members would be an uneasy fit. In both Scotland and Wales, which are new institutions, although inheriting British traditions – the implicit higher status of constituency members is acknowledged. How much more so would it be if regional or county members were grafted onto a constituency-minded place like the Commons?

MMP also involves slimming down the number of constituencies, unless there is a large (and presumably unpopular) increase in the number of MPs. Even the relatively cautious AV+ would require a full review of constituency boundaries and a round of musical chairs which would undoubtedly worry incumbent MPs. While STV and List PR would involve larger units, these could – in the first instance in any case – be based on existing seats. This would remove an anxiety-inducing phase of the reform process.

These objections need not be terminal, as MMP gives the electorate more choice and more power than FPTP and it has been introduced in another Westminster system, namely New Zealand. Certainly, the assumptions MPs make about the way they do their job, and any element of wounded constituency pride, should take second place behind the will of the electorate if MMP or AV+ emerges as the change option and people support it.

Single Transferable Vote

The Single Transferable Vote (STV) is relatively simple for the voter to use, requiring him or her only to number individual candidates in order of preference. It retains a strong link to geographical constituencies and gives a strong incentive for all MPs to be attentive to the people who elected them. It is broadly proportional but does not involve a splintering of representation between small parties.

Our organisation was formed as the Proportional Representation Society, but with increased use of list systems in Europe the name was changed to the Electoral Reform Society, believing that there is more to electoral reform than the quest for proportionality. For us, it is important that voters are able to choose the people who will represent them, and not just the parties, and it is the desire for broad proportionality as well as increased voter choice that leads us to STV. Arguments in favour of STV include:

1. STV, by using ‘preference voting’, allows the voter to express views on the candidates in a more sophisticated way. Voters mark candidates in order of preference. Then, if a voter’s first choice of candidate does not need the vote (either because the candidate has no chance of election or already has sufficient votes), the vote can be transferred to the next preference, and transferred again if necessary until the vote reaches a candidate for whom it can be used.
2. Preference voting also removes the need for most forms of ‘tactical voting’, something that has become common in British elections. Instead of electors voting for the candidate they want, they may vote for the candidate with the best chance of defeating the person they least want. Preference voting allows voters to vote naturally without fear of their votes being wasted.
3. STV has the advantage that it makes all elections competitive. Even in areas where a party has strong support, voters are likely to be able to choose between the candidates of that party, and seldom will a party have sufficient support to win all of the seats in a constituency. As a result, there are no ‘safe’ seats (thereby increasing accountability to the electorate) and very few candidates without any hope of being elected.
4. With STV most votes will count towards the result. As a consequence, most voters will find that they have voted for a successful candidate, although not necessarily with their first preference vote.
5. As a result, STV can strengthen the link between elected members and their constituents. People are more likely to feel an attachment to a politician whom they have helped elect than to one they voted against. In Ireland, politicians have twice tried to change from STV because of the level of accountability it produces, but twice in referendums the electorate have chosen to retain STV.
6. STV, like list systems, can require parties to select more than one candidate in any electoral area. This is more likely to lead to a diverse group of candidates as parties, in seeking with widen their appeal, have an incentive to field teams of candidates that are more balanced in terms of gender, age, ethnicity, etc. STV has the added advantage that it will produce broadly proportional outcomes by whatever criteria significant groups of voters consider important, and not just by party. If, for example, a significant number of voters choose women candidates, irrespective of party affiliation, then a corresponding proportion of women candidates are likely to be elected.
7. Furthermore, it is our belief, supported by evidence from Ireland where STV is used, that STV has the potential to encourage a more positive form of campaigning. With FPTP candidates can win, and often do, through negative campaigning: it is sufficient for them to attack opponents as they do not need to win a majority of votes but only more than their rivals. Under STV candidates are more likely to be successful if they can appeal beyond their own party support so that they benefit from transfers from others.

While we expect political debates to be robust, unnecessary attacks on opponents are unlikely to attract transferred votes should these opponents be eliminated during the count.

In short, the Electoral Reform Society believes that the Single Transferable Vote clearly satisfies the Jenkins criteria of broad proportionality, voter choice and constituency links. By its relatively high effective threshold for representation it also prevents unduly fragmented parliaments that make government formation difficult. By using preferential voting it encourages a more consensual approach to politics and helps make coalition formation more transparent to the electorate and less difficult for politicians. It is also compatible with diversity and equality in representation. As a proportional system giving voter choice, it is in the family of electoral systems that produces more participation than FPTP, although the relationship between these factors is by no means simple.

We believe that the case for STV is strong and can withstand scrutiny (as it has repeatedly in recent years in British Columbia and the assorted Kerley, Sunderland, Richard and – to a degree – Arbuthnott commissions in Britain). By contrast, the case for FPTP tends to fall apart under examination. It has not been recommended by any serious study of electoral systems anywhere for a long time. We look forward with relish to putting the case for STV to a democratically convened, deliberative commission or Citizens’ Assembly. Can the remaining defenders of FPTP make the same claim?

Other elected bodies

While General Elections are the most important – and most problematic – form of election in the UK, there is room for examining elections for other levels of government. In many of them there has been welcome progress since 1997, although there is room for further reform, although in English local government the need for change is urgent. For many of these levels of government there are already independent reports available which point the way to future reform – Arbuthnott in Scotland, Richard in Wales and Sunderland for Welsh local government.

Scottish Parliament The Arbuthnott Commission reported on Scottish electoral systems in January 2006, and concentrated for the most part on the electoral system for the Scottish Parliament. In the medium term it recommended reforms to the existing AMS/ MMP system but reserved the possibility of reconsidering the system and changing to STV at some future stage. The Commission’s recommendations, other than some fairly minor changes to ballot design, have been on the shelf since and there will be no significant change before 2011 at the earliest. The Commission’s brief was perhaps unsatisfactory, and there is room for revisiting electoral arrangements for the Parliament in the light of the 2007 election and its results.

Welsh Assembly The Richard Commission reported on the government of Wales in March 2004, recommending several steps to put devolution for Wales on a firmer footing. Some of these were legislated in 2006 as the Government of Wales Act, but this legislation ignored the Richard Commission’s considered views on the electoral system and chose instead to tinker with ‘dual candidacy’. In order to cope with enhanced powers, Richard recommended that the size of the Assembly be enlarged from 60 to 80 AMs. To accommodate this expansion, a change to the Single Transferable Vote appeared to be the best option.

There seems relatively little that can be said in addition to the Richard Commission’s recommendations, and the logic that Richard followed – somewhat reluctantly – to a larger Assembly and STV will probably influence any future review or inquiry.

Scottish local government

The deliberation process in Scotland has, after two reviews and legislation, resulted in the Local Governance Act 2004 that changed the electoral system with effect from May 2007. The performance of STV in local elections in this first set of polls will no doubt be keenly studied and observed and there will be lessons to learn, but some of the longer-term changes brought about by the new system will only become apparent with future sets of elections. There would seem to be no need to think about further changes and refinements to the system for at least a couple of electoral cycles.

Welsh local government

The report of the Sunderland Commission has remained on the table since 2002 and the broad issues concerning elections for Welsh local authorities have not changed since then – uncontested seats and a lack of competition in many areas are still a problem. There is therefore already a considered proposal for change, and an option to retain the status quo. The Welsh Assembly – and the Westminster government – may have to revisit these issues in future.

English local government

The government’s 2006 White Paper ‘Strong and Prosperous Communities’ was disappointingly lacking in depth when it came to local authority electoral arrangements. The main recommendation was an – optional – move to whole council elections on single member wards. While this is a more rational form of FPTP than using multi-member wards, it does diminish the continuous accountability of election by thirds and may introduce even worse problems of uncompetitive local elections. The enhanced role of parish councils and the extension of the parish model to the big cities make the anomalies in elections for these bodies all the more pressing. Some of the other proposals such as directly elected executives were sketchy in the extreme. The White Paper showed no signs of having thought through the electoral issues or having worked in a ‘joined up’ way about accountability.

Devolution of significant powers to local level should be accompanied by enhanced democracy and accountability of local institutions. Without addressing the issues and seeking meaningful public consent, the government’s blueprint for local authorities may last no longer than the other grand designs offered in 1972 and 1991. While there may be a case for local options and different models in different areas, the current pattern has developed piecemeal and not according to any rational plan or in response to local opinion.

There should be an open review of local government electoral arrangements, with the participation of stakeholders and the public in general.

Mayors

We have presented our case that the Supplementary Vote has failed to live up to the hopes originally vested in it. It has proved a confusing system for voters to use (as shown in the large proportion of spoiled and incomplete ballots), it has not produced majority winners in most contests, and it restricts voter choice in what is an increasingly diverse and plural local political environment.

We await the government’s own verdict on the Supplementary Vote system, but believe there is a clear case that it has outlived its usefulness and should be replaced by the Alternative Vote.

Greater London Assembly

The GLA is essentially part of the same system of London government as the Mayor, offering checks and balances to the running of the region. Elections to it take place concurrently with those for the Mayor. Having recommended against retaining the two-column SV ballot for Mayor and replacing it with a preferential system, there is logic in having a uniform ballot procedure for the voter at London elections. The 2000 and 2004 elections produced strong evidence that the multiplicity of ballot papers and a mixture of disguised preferential, singlecandidate and party list voting was confusing.

Introducing the Single Transferable Vote for the GLA would mean that voters would be faced with only two ballot papers, and would complete them in the same way (i.e. by numbering preferences).

European Parliament

The European Parliament as an institution has other issues of accountability and public perception than just the electoral system, but the electoral system is certainly part of the way in which it relates to the public. The Arbuthnott Commission in Scotland has already recommended that Scotland should join Northern Ireland in using the Single Transferable Vote for European elections. The position in Wales and England is also ripe for review. While some of Arbuthnott’s arguments are made specifically with Scotland in mind, much of the logic is also applicable to Wales and the English regions. There is also the example of European elections in Northern Ireland which seem to attract greater public interest.

The system of election for the European Parliament has not been subject to a recent review, other than the Department for Constitutional Affairs internal review.

The best way forward for the European Parliament electoral system would be a fully open review, with terms of reference that would prevent it from being hijacked into being a re-hashing of tired arguments about the overall relationship between Britain and the EU. Evidence should be sought from MEPs, the parties, academic experts, civil society institutions and the public about how to introduce a greater element of accountability and choice to the European Parliament electoral system while retaining proportionality, and to find a more comfortable role for the institution in British political life.

A new start for politics

The government’s review has been ‘desk research’ since it was established in February 2005 and it is certainly necessary to conduct desk research. The Electoral Reform Society’s review, encapsulated here, is also desk research. However, no review can be considered complete until it is opened out for public consultation. We have set out our views and invite them to be tested against expert and public opinion, and hope that our ideas can contribute to a government review which has not reached its final stage.

The complex of issues around disillusionment and disengagement from politics has deservedly attracted attention recently, and as noted by many commentators including the POWER report the electoral system is at the heart of the problem. The right sort of electoral reform, we believe, is a good in itself and will help to bring politicians and people closer. It is a necessary, but not sufficient, condition before politics in Britain can be renewed and make a fresh start.

However, the process by which reform comes about should also be of a high standard and represent the best rather than the worst of politics as it is now. There is, rightly, a suspicion of reforms that come from the top downwards, and a quick fix or Westminster deal is unlikely to cure the dissatisfaction with the political system.

There is a need to involve the wider public in the debate, and the process of making a decision.

Because Westminster is the central institution of the British state, such a model of decision making needs to be all the more inclusive and democratic.

We look forward to the next stage of the review of electoral systems which will broaden the debate and take it to the electorate.