Electoral Reform Society
I will begin by introducing the representatives from the Electoral Reform Society: Ken Ritchie, the chief executive, James Gilmour and Peter Facey. They will give a presentation, after which there will be an opportunity for questions.
Ken Ritchie (Electoral Reform Society):
Thank you for that welcome and the invitation. We have circulated copies of some of the transparencies I am going to use. We have been asked to talk about electoral systems and we will focus on the three that have been suggested by the McIntosh commission as worth attention: AMS, the additional member system, AV-plus, the alternative vote plus and STV, single transferable vote. Research for the commission pointed out that a number of systems could be devised for local government in Scotland. If you would like us to comment on other systems, we are happy to do so.
The McIntosh report pointed out that choice of an electoral system depends on how you want it to perform, not in terms of how people will vote and outcomes, but according to what are seen as the important criteria. I propose that we briefly run through the criteria that were identified as important by McIntosh. We can then move on to a consideration of the systems. If members feel that there are certain things that they would like us to focus on, they should let us know and we will adapt accordingly.
The criteria are proportionality, the councillor-ward link, a fair provision for independents to ensure that any system does not discriminate against independents, the allowance for geographic diversity, and the close fit between council wards—what we have called "natural communities". I shall return to those criteria in more detail.
Most needs to be said about the first one, proportionality. We are talking about proportional representation. Electoral reform is not just about proportional representation, and some people might argue the converse. Clearly, we want fairness for all parties. In many areas, parties get a significant number of votes but little representation. Most voters vote according to party affiliation and if we do not have fairness for parties we do not have fairness for voters.
Not all voters vote simply by party. McIntosh makes the point that we should also seek proportional representation to bring in other community interests. People might select the candidates whom they want to support by gender, age, ethnicity, or particular policies that lie outside the normal party debate. All sorts of factors might enter into their choice. Some electoral systems aim to provide fairness and proportional representation simply by party, but others try to go further, to allow, to some extent, representation by the criteria that the voters think are important.
An important aspect of proportionality is that, in a proportional system, many more votes are going to count. In far too many wards, voters can vote one way or another futilely, as those wards are always won by the same party. That is a huge disincentive for the voter, as voting does not make much sense. There is no great incentive for parties to campaign in those areas, and party organisation tends to become moribund. We do not claim that proportional representation will solve all those problems, but we want all voters to have an interest in voting. Their votes will contribute to a larger total that will determine how seats are allocated, and there will be a better chance of breathing new life into local democracy where there are appallingly low turnouts.
The retention of the councillor-ward link is clearly important. We feel that the elected representative should be accountable to the voters so that if the voters are not satisfied they can remove and change their representative at a future election. The link must be seen as two-sided. On the one hand the councillor must be linked to an identified ward; on the other, voters are entitled to feel that somebody has the job of representing their interests.
Slide 4 on the retention of the councillor-ward link says:
"All wards should have (a) councillor(s) responsible for raising ward issues".
I put councillor or councillors, because one cannot have a proportional system without having more than one representative for an area. With single-member wards, the person who has an absolute majority—or a majority over the other parties—is the representative; voters for other parties cannot be represented.
Proportionality can be achieved in different ways. As members know, the Scottish Parliament has a system whereby a number of members are identified with constituencies, while others are elected through top-up lists. That model, with two categories of representation, is one option. The other option is to have multi-member wards in which all councillors are elected on the same basis. I will return to that point when I talk about systems, as there is a fairly important difference between the two.
It has been argued that there ought to be fair provision for independents in local government. We accept that independents are often disadvantaged because they do not have a party machine to campaign on their behalf. However, certain types of system, which focus on proportionality by party, tend to view independents as one-person parties or may put pressure on independents to combine into a type of pseudo-party, which many independents do not want to do. As a result, independents can be discriminated against.
The McIntosh commission pointed to the great diversity that exists in our country: there are areas in the Highlands and Islands where wards are geographically large and areas in cities with high concentrations of population. McIntosh was keen to consider how we can achieve a system that copes with both extremes. Linked to that was the fit between council wards and natural communities—we all know of ward boundaries that run through the middle of a street or divide areas that clearly have the same problems and characteristics. Some systems allow boundaries to be drawn more flexibly than others—with multi-member wards, for example, there can be flexibility on size—and so get round such problems.
Those are the five criteria proposed by McIntosh for choosing a voting system. I do not suggest for one moment that all five are as important as each other. There may be other factors that the committee or the working group on renewing local democracy will want to take into consideration.
I move on now to discuss the three systems that were recommended. Members will be familiar with the additional member system, as it was used for elections to the Scottish Parliament. The additional member system has been used since the war in Germany and was recently introduced in New Zealand. However, we have no knowledge of its being used at local government level. In Germany, the system is used for elections in the Länder, but those tend to be much larger units. The system tends to be used in situations where people are being elected as legislators, rather than as ward representatives. A move towards the additional member system would, therefore, be a bit of a first at local government level.
The role of top-up councillors—the people who are elected from the lists as distinct from those who are elected from wards—under the additional member system would have to be considered. I know that the two categories of elected representative in the Parliament are a point of discussion and that it may be some time before the roles are properly developed. What would happen in local government under the additional member system is perhaps a separate issue. I know from discussions on modernising local government that there are new models involving elected provosts and cabinet structures within councils, so it could be argued that there will be two categories of member anyway—those who deal more with the ward duties and those who deal with more strategic policies.
The additional member system does not provide those two categories. The largest party would tend to get most of the ward seats and the smaller parties would tend to win their seats through the top-up lists. If one compares the make-up of the Scottish Parliament with what might happen in a council, it can be seen that most of the Labour members might end up with ward case work, while other parties could be exclusively concerned with political strategy—that would clearly make a nonsense of local government.
The ratio of ward to top-up councillors is what determines the level of proportionality. It also depends partly on the absolute number of seats as it would be possible, as slide 9 shows, to have a top-up list for an entire council or to split the top-up seats by area. The Scottish Parliament does not have a single top-up list for the whole of Scotland, but has parcels of roughly 12 seats of which seven are constituency seats and five are top-up seats. The more top-up members there are, the more proportional the system can be. Of course, the downside is that the more top-up seats there are, the more councillors there are who do not have a strong link with the electorate. They will be people who owe their seats more to their positions in the party, perhaps, than to direct election. There are implications for boundaries, to which I will return in a moment.
If the top-up ratio is small, most councillors will have good links, but proportionality will not be that great. As I have said on slide 10, there is the increased risk of tactical voting. When one party is expected to win so many of the ward seats that it does not mean much to vote for that party in the top-up lists, there is a danger that voters will vote for another party in the top-up lists.
Large top-up ratios will have good proportionality. As the slide shows, councillors without a ward link are much more numerous and ward sizes are larger. If there is no increase in the number of councillors, the number of ward councillors has to be reduced to make space for top-up councillors; the number of wards has to be reduced and the size of wards increased.
Unless there is a full-scale redrawing of boundaries, which would be a massive exercise, the options for change are more limited. Possible options include combining every four wards into three down to combining pairs of wards. In each case there are implications for the ratio and number of ward councillors. Combining pairs of wards need not mean as many councillors from the top-up list as I have suggested; you could have 30 and 20 and reduce the overall number.
To summarise, the additional member system gives proportionality but by party because calculating top-up seats is by party. For that reason AMS discriminates against independents. They may have a fair chance in the wards but not in top-up seats. A downside is that it creates two categories of councillor, with the top-up group of councillors having only weak links to the electorate.
The second system that the McIntosh commission suggested was worth looking at is alternative vote plus. The terminology is new and comes from the Jenkins commission, which recommended AV-plus for Westminster elections. It has not yet been used but that is not a downside, as we should manufacture the system that meets our particular needs. AV-plus is a form of additional member system but improves on it by the use of the alternative vote in electing ward representatives, and when it comes to the top-up lists it gives people a choice not just of parties but of the candidates within the party.
Some of you will be familiar with the alternative vote as many parties use it internally in the form of an eliminating ballot where it means going through a process until one candidate has not only a majority over the others but more than 50 per cent of the vote. On slide 14 there is an example of how it would work with four candidates: Forsyth may start as the front runner with 73 votes, but out of 200 that is well under half, although in a first-past-the-post election Forsyth would win. Thompson, with only 42 votes, is the first to be eliminated.
No he is not—Campbell is.
I beg your pardon, Campbell with 18 votes is. We then redistribute the second preferences on the ballot papers for Campbell. Still no one passes the 100 mark so Thompson is eliminated and in the end Munro wins. Many political parties use it and I hope it is fairly easy to understand. The advantage of this system is that whoever wins a ward seat has the positive support of more than half of the voters—not necessarily as their first choice but at least their support. That means that at least half the voters will feel satisfaction in the result and have a councillor in whom they can have some confidence.
In addition, there is no need for the tactical voting that mars so many of our elections. People do not have to think, "Should I vote for my first choice of party, because it may not do very well? I am more intent on beating another party. Let me use my vote for the party that I think can defeat the one that I most want to be defeated." With the alternative vote plus system, people can vote for the candidates who they really want to win.
A voter may want the Natural Law party to win, but if it is eliminated, the vote has not disappeared—it simply moves on to the voter's second choice, and so on until there are only two candidates left. The system asks voters, "If the choice is between those two candidates, who would you choose?" The system increases the choice open to voters and means that voters do not need to go through a guessing game. It means that parties can campaign honestly for positive support rather than the usual round-the-streets campaigning of, "I know you support them but they don't stand a chance, so why not support us?" The alternative vote plus system can lead to a much more positive approach to an election campaign.
In giving voters a choice of who they vote for in the top-up lists, we are moving from what are called closed lists, which were used for the European elections where the choice of candidate was closed to the voter, to lists where that choice is open to the voter. A ballot paper might look something like the one shown on slide 15, so that a voter wanting to support party A could choose which candidate to support. The votes for the parties count in exactly the same way as before—the seats are allocated. However, even if party A's favoured candidate was Anderson, if McLeod received more votes than Anderson and there was only one seat available, the seat would go to McLeod. The voters determine the order of the list, which means that there is that extra accountability to the electorate from those who are elected.
To sum up the alternative vote plus system, there is more choice for the voters, tactical voting is eliminated and there are better prospects for independents. In many cases, people may vote for a party and if they are to be given a second choice, they may not wish to move to another party, preferring to recognise a strong independent candidate who might be worth supporting. There are indications that, under this sort of system, independents do not lose out. The open lists increase accountability and, even if there were additional members or top-up members, at least they have more democratic legitimacy than might be the case otherwise.
I now move to the third system—the single transferable vote, which is quite different from the other two systems. It has been used in Northern Ireland for both local government and European elections and has been used for all elections in the Republic of Ireland since the 1920s. In Scotland, it was used until 1928 for separate elections for education authorities. There are many professional organisations, such as trade unions and the National Trust that also make use of it.
Slide 18 refers to the main features of the system. In the additional member system, there are two ballot papers—one for the constituency and one for the party. In the single transferable vote system, there is only one ballot paper, which would look the same as the alternative vote paper used for ward elections using alternative vote plus. All members are elected on the same basis. There would be a group—typically four to six members, although there may be fewer in rural areas—each member of which has equal legitimacy. It is broadly proportional by party, but it allows the voter as much choice as possible, so that if voters decided that they wanted to vote on, say, fox hunting, or Europe as their major concern, they would be able to do that. The result would be broadly proportional by these other criteria.
All the people are elected individually, which preserves the element of accountability. Candidates may be elected on the basis of their party affiliations, but the electoral system treats them as individuals. That does not discriminate against independents.
A final point to consider is that significant boundary changes would not be needed. We are talking about multi-member wards in which there can be flexibility in the number of seats in each ward—a flexibility that can be extended even in a local authority ward. By a process of amalgamating wards, we could form new multi-member wards without needing to enter into a drastic exercise of changing existing boundaries.
I propose to give members a simplified account of STV, to give an impression of how it produces particular results. Please tell me if this is old hat or unnecessary. For simplicity, I have assumed that there are eight candidates contesting four seats. Two parties have each put up two candidates, and there are a couple of independents.
There are 500 votes in total. The first thing to do is to calculate a quota that is the number of votes that a candidate must receive to be elected. With 500 votes, the quota is, in this case, 100 votes for a seat. If four candidates receive at least 100 votes each—we will assume that five candidates would not receive exactly the same number of votes—a fifth cannot receive 100, which is the target. According to the voting figures shown on the slide, Fraser, who has received 150 votes, would be elected straight away.
In a first-past-the-post election, if a candidate wins a huge majority, there are many wasted votes; moreover, that candidate would be elected even if the majority was only one vote. With a single transferable vote, votes are not wasted in the same way. If somebody has a surplus of votes, that surplus can be transferred. In this case, Fraser has 150 votes although he needed only 100. One third of those—50 votes—must be transferred, which gives us this result: Fraser is elected, but let us assume that even with the transferred votes nobody else has reached the 100 quota.
Let us also assume that most of Fraser's 150 votes would be transferred to candidates of party B. As would happen under the alternative vote system, the person who has the smallest number of votes is eliminated. In the case illustrated on the slide, that would now be Evans in party A. If he is eliminated, his 35 votes would generally be transferred to the other candidates within his or her party—Allan and Campbell.
The quota is 100 votes, so no more candidates have been elected. We therefore continue to eliminate candidates, in this case Robertson, although not as many of his votes will go back to his party, as there has already been a fairly big transfer. Nevertheless, with 105 votes, McLeod has more than the 100 mark and we continue with the process. If there is a surplus, it is transferred, as it may help somebody at the bottom of the list. If a candidate cannot be saved by a transfer, the person at the bottom is eliminated.
I am a bit lost, so this may be an extremely stupid question. Which 50 votes are transferred?
We do not transfer 50 individual votes. We split up the votes and transfer a third of them. So, if a person votes for Fraser on the first preference vote and for McLeod on the second preference, two thirds of each single vote would be given to Fraser and one third to McLeod.
I want to understand this properly. The votes are allocated in proportion to the second preference votes. That is the key. If everyone voted for one person on the second preference, a third of the 150 votes—50 votes—would go to that person. The allocation is in proportion. That is why the second slide adds up to 150 votes—it shows the proportion of Fraser's votes given to each candidate.
My apologies. I see that a column has been left off the slide, showing the process for splitting up Fraser's 150 votes, of which we want to transfer only 50. The danger if we simply had a lucky dip into the voting papers would be that we could pull out a sample of votes in favour of one party or another. What we do instead is to look at all the votes that have generated the surplus and transfer the corrected proportion of the votes to the other candidates—a fraction of the votes is carried forward to each.
The example that I have used shows how the process might carry on until there is a close contest at the bottom of the list between two independent candidates, Smith and Williamson, and between two people from party A, Allan and Campbell. An independent, Williamson, loses out. Many people might prefer to transfer their votes to the other independent candidate, so that Smith is elected, but some help to elect Allan at the top.
What are the messages? If we take the three candidates from party B together, party B has more than 200 votes in total, which is enough to get two candidates elected. However, all the votes were stacked up on one candidate—Fraser, with 150 votes. The party does not lose out, however, because if people's second votes are transferred to other candidates of the same party—most people vote with some party loyalty—the party can use that surplus to ensure that two of its people are elected. In party A, which had the three names at the top, there was a fairly close contest between Allan and Campbell, which, in the end, was resolved by votes from outside the party.
Although people will mainly vote by party, STV also favours candidates who have a broad base of support and can attract second and subsequent preferences. For that reason, the system is useful for independents, as they will often pick up second and subsequent preferences from other parties. In the Irish Republic, about 10 per cent of councillors are independents, which more or less matches the number of first preference votes that they received. It is also possible for small parties to do well. The Green party has succeeded in getting members elected to the Irish Parliament even though it had a small number of first preference votes, because the party has been given subsequent preference votes and have won through in that way. The system gives a broadly proportional outcome allowing representation of different opinions, including smaller parties.
As I said, the system provides some flexibility in implementation. We recommend that four to six seats per ward would normally be ideal to give a good level of proportionality. In a rural area, a six-member ward could be geographically rather large. There is a trade-off between the geographic size of the constituency and the desired level of proportionality. In a rural area, a ward of either three or four seats might be preferable, and in extreme examples there might be a two or even a single-member seat. I am thinking of an island that constitutes one ward or of areas in the Highlands where multi-member wards can be so big that councillors would find it difficult to get round the entire ward.
I have summarised our comparison of the systems. The additional member system clearly gives proportionality by party, as does AV-plus. STV is the only system that gives broader proportionality, if voters use other significant criteria to decide how to vote. Under STV, everybody is elected on the same basis and has the same sort of link with their ward. Under AMS and AV-plus, some are ward councillors with a strong link, although perhaps not as strong as the current link because wards will be larger. The downside is that the additional members do not have as strong a link. AMS is disadvantageous to independents; the transferable voting in AV-plus or STV gives them a better chance.
As I said, under STV there can be flexibility in how constituencies are put together. Unless the option of pairing wards—with half the number of ward councillors—is chosen, ward boundaries would to some extent have to be redrawn in order to implement either AMS or AV-plus.
As the committee will can see from these criteria, the Electoral Reform Society comes down in favour of STV. In saying that, we welcome the lead that Scotland has taken in going for a proportional system and we believe that any of the systems that we have looked at have a lot more to offer than the existing first-past-the-post system. We will be happy to discuss other systems, or to help in any further way that you feel appropriate.
Thank you, Ken. That was interesting. This is a complex issue—the more one hears about it, the more complex it seems to become. Does anyone have any questions?
In the elections for the Scottish Parliament and for the European Parliament—the two elections in Scotland that have had an element of proportionality—the turnouts were lower than at the general election and at the previous European election. I am dubious about the link between proportionality and encouraging the electorate to come out and vote. I am also concerned about the obsession with lists. The alternative vote system can be applied in many ways. Why do we have to have AV-plus?
Peter Facey (Electoral Reform Society):
Turnout for local government elections in England and Wales was about 29 per cent. A month later, it fell to about 23 or 24 per cent. There is an indication that turnout has something to do with how people rank institutions—the importance with which the institutions are regarded. Westminster is still regarded as the most important institution, with the Scottish Parliament next and the Welsh Assembly after that and so on.
Turnout can also be affected by the work that political activists do on the ground. Some systems encourage activists to work in marginal wards. In local government, for example, wards that are marginal tend to have higher turnouts than safe wards, simply because of the amount of political activity that has gone into that ward—the amount of canvassing and the number of bits of paper that have gone through people's doors. If more wards become marginal, political activity is spread and that may have some effect on turnout. The problem with closed-list systems is that they do not increase activity among political activists. The experience of the European election was that political activists did not know for whom they were campaigning, so there was no increase in campaigning and therefore no increase in turnout.
The alternative vote system on its own can be less proportional than first past the post and can produce greater disproportionality. For example, the Labour party would probably have had a much larger majority in the House of Commons if the general election had been held under AVS. Whatever the complaints at the moment, they are not that the Government has too small a majority; many people believe that it has too large a majority. The question is whether you believe that that is a good thing.
I would like a little more information about the systems that are operated in other countries. First, you mentioned various countries that use STV at the local level. What is the turnout in those countries? Secondly, there has not been much mention of Europe. What systems operate in European countries?
I cannot give you an exact answer on the continental question—the necessary academic work does not exist. In most places, variations of list systems are used—usually across the whole of a council area. One thing to be borne in mind is that council areas in Europe seem to be much smaller than they are in Britain, Ireland or the United States. The difference tends to be whether open or closed lists are used.
There is a complete range of turnouts under STV. Some countries have compulsory voting, so turnouts there are not particularly relevant. Australia has compulsory voting and has turnouts of more than 90 per cent. Malta has a turnout of 96 per cent, which is the highest in the world for a country in which voting is not compulsory. One could argue that Malta is a special case because the country has a population of roughly 300,000 people. It has the highest turnout of any free democratic state. Only a country such as North Korea beats Malta on turnout.
Ireland has turnouts of about 60 per cent for local government elections. The turnouts in Ireland are about 60 per cent for all elections—parish elections, European elections and Dail elections. Turnouts in Northern Ireland are also quite static—around the 55 to 65 per cent mark.
Until recently, turnouts in Britain for local government elections were increasing. From a very low base there was a small, gradual increase. We used to say that the average turnout in Britain was about 40 per cent but, in the most recent local elections, there seems to have been a downward trend. It is too early to say whether that is a long-term phenomenon. Some academics argue that, because of the change in central Government, there are fewer protest votes. It is certainly true that turnout in local government elections in the UK is below continental norms.
If we had a series of different systems for the different elections, do you think that that would affect the turnout? We already have four different systems: European, Scottish Parliament, Westminster and local government.
Secondly, must AMS have two separate votes? Some people misunderstand the system and consider the second vote to be a second-choice vote. There is confusion surrounding that second vote.
My third question is practical and does not relate to the electoral mechanics. As you will be aware, tensions exist between constituency MSPs and list MSPs about work load—turf wars. Would that be more exaggerated at local government level if there were to be two different classes of councillor?
I will try to answer a couple of those questions, although my colleagues might also want to come in.
We sometimes make the mistake of thinking that our electors are not as bright as they are. The experience of the Scottish and Welsh elections showed not only that people understood how to vote, but that many of them understood how to make the best use of their vote within the voting system. A few people may not have fully understood the significance of the second vote, but people did not need to know the full details of the d'Hondt system to know what to do with the party vote.
In Northern Ireland, people have been using first past the post for Westminster elections and STV for local government and European elections—there have even been list systems for the elections to some bodies. People have coped with the range of electoral systems without any bother. I am not arguing that using a range of different systems is desirable, but that should not constrain us too much in choosing which system is the best for the job that has to be done.
It would be possible to construct an additional member system in which each elector had only one vote—a vote for the Labour party in the constituency contest would be assumed to be a list vote for the Labour party. That system has been discussed for many years, but it significantly reduces the choices that are open to the voter.
The difficulty with such an additional member system is that it maintains a first-past-the-post system in the constituency. A voter might want to support, say, a Green party candidate, but he or she would know that, in the constituency election, the contest will be between, say, Labour and the SNP and that a vote for anyone other than Labour or the SNP would be ineffective.
With an additional member system, it might be possible for a voter to split his or her vote. The evidence shows that many people did that in the elections to the Scottish Parliament. People seemed to be intelligent enough to make the best use of the system. If we adopted an additional member system that used only one vote, we would restrict the choice.
There are tales of political turf wars, both in Scotland and in Wales. MSPs' decisions and legislative scrutiny affect the whole of Scotland, but elected representatives in local government deal with local problems and my guess is that the problem of turf wars would become more acute if a number of councillors had no clear ward responsibility.
One of the things that concerned me about the Scottish Parliament elections was the length of the second ballot paper: the Lothians one was about 18 in long and featured 16 or 17 choices, some of which made a mockery of the democratic process. How would that tendency be controlled? I am not talking only about the joke candidates: in Lothians, 10,000 votes went to the Socialist Labour party, which has no strong base in the country. I believe that many who voted for that party were trying to vote for the Scottish Labour party. A few years ago, somebody stood as a Literal Democrat and took a large number of votes from the Liberal Democrats.
The problem with the Literal Democrat situation has been resolved now that we have registration of parties. People will no longer be able to use that name.
Bristow Muldoon makes a more general point about the number of candidates. Even in first-past-the-post by-elections, there is a huge number of candidates. That problem cannot be tackled by the choice of electoral system. Many people want to stand as candidates to draw attention to themselves, to their cause or to their business, rather than to seek election. Citizens have a right to stand, but there may have to be, for example, a much tougher test in terms of the number of people who are required to propose and second a candidate.
Any proportional system will result in the election of more candidates at a time and will require a larger ballot paper. The ballot papers that were used in many areas for the European elections were long but the process was simple. As far as I am aware, there is no evidence to suggest that people were put off voting by the length of the ballot paper. It is unlikely that they would have been, as they had not seen the ballot paper until they went to cast their vote.
We do not want voter choice extended so far that there is no longer a meaningful choice, but the problem is not as great as some make out. I know that James has views on this matter.
James Gilmour (Electoral Reform Society):
There is a difference with systems that include a party list element. Our experience with AMS—which crosses over into AV-plus and the closed lists for the European Parliament—is that the party list element implicitly encourages parties to nominate as many candidates as there are potential vacancies. For example, in the Scottish Parliament elections, each party could nominate up to 12 candidates for the regional list. However, the many parties that did so had no prospect of winning all 12 seats.
We can contrast that experience with the experience of countries that have used STV. As Ken has indicated, under that system parties give their supporters a wide choice. However, parties hoping to win three seats in a six-member ward would put up only four or five candidates. Very rarely would parties put up six candidates. If one party did so, all the other parties that were contesting the ward would not do the same. There is a practical realism in STV. The system offers a realistic choice to party supporters, and business managers do not waste effort putting in extra nominations to make up a full list. That is a difference between the two types of system.
I have always thought that a way round having all these dafties on the list is by stipulating that a candidate cannot ride on the back of the list if his or her party does not pay a deposit and has a candidate standing for the first-past-the-post seat. However, that was not going to be my suggestion. Parties are eternal optimists. The SNP put 12 candidates on the Glasgow list. We also had 10 candidates running for first-past-the-post seats, so I suppose we were trying to hedge our bets.
McIntosh did us a favour by recommending only three electoral systems to choose from. I have a list of the electoral systems used in each European Union country. Each country uses some form of proportional representation in local government, all of which seem to be completely different. I will not bore the committee by reading out the details of those systems, but if members want to see the list, I can hand copies around. How closely has the Electoral Reform Society examined other European systems, and is the organisation convinced that the three recommended systems are the best systems to consider in detail for Scotland?
The fundamental difference with local government and politics in Britain and Ireland is that we have a culture of ward-based politics. In most other electoral systems, candidates are elected to represent a whole geographic area. Sometimes that is broken down, but the idea of the ward, as we mean it, is quite Anglo-Saxon—or Anglo-Saxon/Celtic, at least.
You saved yourself there.
Yes, you just about saved yourself.
I made a quick revision.
You were right the first time.
I think that McIntosh went for the three systems because of the element of a ward-based link that they include.
The committee could certainly examine the possibility of list systems operating in smaller areas, but that would produce its own problems. The most realistic options are in front of the committee, and they are the ones that are most acceptable to politicians and to the electorate in Scotland.
I would not like to try to sell you a list system for the whole of Glasgow or, even worse, for the whole of the Highlands and Islands. That would not get very far.
So, in other words, you think that those are the three best options.
I think that they are the most realistic options for Scottish politics.
One of the issues that you have highlighted is that there is a problem in the way that people participate in the democratic process of the first-past-the-post system. You mentioned that there were wards in which it is pointless to vote because an overwhelming number of people will vote for one candidate. In certain geographical areas they vote for a candidate because of historical links.
That is not an inflexible situation. If people participate actively in the democratic process, that can change. It is not an inevitable part of the first-past-the-post system. People are not always blind to the possibility of changing their party loyalties. There have been by-elections in which the shift has been quite dramatic. We must be careful that we do not associate problems with a system that are not the result of that system but of something historical. There is a base vote for a number of political parties and that creates problems in attempting to achieve proportionality.
It is interesting to me that McIntosh identifies a series of criteria, which, as they are broken down, result in STV appearing to be the most effective system. Are there any other criteria that you would include, and which do you think are the most important? Is there a case to be made for including accountability?
I made a point about a base vote that can be shifted. That is the easiest way to remove someone who is not doing his or her job. Local people can identify such a representative and put pressure on them; it is simple to vote them out of office. Do the other systems that you have identified allow for that? What are the difficulties with that, and do you think that it is important?
We must distinguish between how voters vote and how they would like to vote under a particular system. In the first-past-the-post-system we know that there are wards where people joke that if a certain party put a donkey up for election, it would be elected. It is not a criticism of the party to say that that party dominates the ward. That party might be in the fortunate position of being in an area in which it enjoys very strong support. Other parties do not, therefore, have a chance of removing a candidate from that party. That reduces accountability.
Under any proportional system, people who are not supporters of that dominant party still have a reason to vote. In the additional member system, some parties might have no hope of winning the constituency contest, but votes for them will contribute to the top-up list calculation, which will determine how well the other parties might do on a wider basis. There are people who might feel that the contest in their ward is effectively between two parties and, therefore, they might as well go out and vote for one of those two parties. Their decision will be essentially tactical, but their underlying support might be for something quite different.
When it comes to the ability to remove a representative, there is more criticism of the closed list system, as the electorate did not vote for the individual but for the party. In other systems, particularly the single transferable vote system where candidates stand as individuals, if they do not perform as the electorate would wish them to, they will not receive the same number of votes next time. While one cannot make automatic comparisons with the different political culture in Ireland, it is interesting to note that, since the war, there have been two referendums in the Republic of Ireland as politicians tried to move away from the STV system, which they felt had created too strong a link of accountability. Those referendums resulted in the retention of the system.
Johann Lamont asked whether there were other criteria. After proportionality, I would go for voter choice, which is implicit in some of the McIntosh recommendations. If the right of the voter to choose their representatives is stressed, that also respects the right of the voter to choose when they want to change that representative. I see voter choice as being just on the other side of accountability.
I do not think that you can have it both ways. You cannot say that, on the one hand, the problem with the first-past-the-post system is that people are stupid and that they are blinded by party loyalty and that, on the other hand, they are bright enough to deal with the sophistication of whatever political system we might present.
You identified that one of the problems with the first-past-the-post system is that people's loyalty and the donkey vote distort the political process. Do you agree that, when choosing from electoral systems, the choice is not one of a perfect system against inadequate systems? In a sense, one is choosing the least bad system. On one level, voters are able to make a pure choice for their first preference for their party and to have that marked in some way. However, one's vote may not necessarily have the same weight in terms of the hard choice about who will make the decisions in local government.
The SNP has experienced that in this Parliament. There is a significant body of SNP members in the Parliament that is not part of the ruling coalition. When the system is more proportional, it does not necessarily mean that the pure vote is reflected in influence, in terms of what happens after the election.
We have not tried today to address the arguments for a change away from the first-past-the-post system to a more proportional system. I have been more concerned to deal with the arguments about which proportional system might be appropriate. I agree that no system will achieve absolutely everything. I hope that it does not become a choice of the least bad system, as it should be a choice of the most good—if members will excuse the grammar.
I do not think that there is any question of saying that voters are stupid under the first-past-the-post system: they are not stupid, as they know how to use their vote tactically. They see that, by voting for certain candidates, they can make a protest gesture and many will decide to make that gesture. However, if they want to influence the result, they know that there are many wards in Scotland that are very strong Labour wards and there are others that are equally strongly held by other parties and voters cast their votes accordingly. We want a system that liberates the voters so that they can vote on the basis of what they believe—we want a system that will take account of that.
I will let Kenny in as he wishes to say something on this point. I will then call Jamie to speak.
Sometimes the first-past-the-post system offers no choice. In Alex Salmond's constituency, there were five wards in which there was only an SNP candidate. That would not happen under the STV system. It cannot be that 100 per cent of the voters in those wards would endorse those five SNP councillors. That takes away choice and the only question that remains is who the SNP will select as a candidate in those wards. There are other cases involving other political parties where folk are effectively unopposed because there is no structural opposition to one particular party.
Johann made the point that if a candidate is abysmal, they can be put out of office. However, that depends on whether the number of people who have contact with that councillor—or who are even aware of their existence—is sufficient to overcome traditional loyalties. Under an STV system, if a party puts up three candidates and one of them has a poorer reputation than the others, those with traditional loyalties will post their votes for the two and perhaps use their third choice for someone from a completely different party. That way they can get rid of people who are not performing well at the same time as being able to exercise choice. People can be loyal to a party without having to vote for someone who they think is useless.
The issues about seats being uncontested and weak party structures are about more than the electoral system and people's understanding of it. I agree with many of the points that have been made. I do not see the first-past-the-post system as being perfect. There are other problems, but we are not necessarily identifying the solutions by focusing on electoral systems alone.
I want to pursue Kenny's point about the fact that it is hard to get rid of a clown under the present system. If a seat is contested at local council or Westminster elections, it is technically possible to get rid of a bad candidate. I want to explore how much research the Electoral Reform Society has done following on from the Scottish elections. Under AMS, the list is ranked according to the system used by individual parties. With all due respect to list members here, it is possible that a clown could be No 1 on the list, simply because the party was not particularly in touch. I am sure that the Liberal Democrats would insist that Donald is a good example. [Laughter.]
That is the kind of loyalty that we enjoy. The Liberal Democrats got it right.
One of the criticisms is that someone who could not face the electorate on a first-past-the-post vote could nevertheless get in as first or second on the list. Has the Electoral Reform Society examined the results of the Scottish election to find out whether, when voters went into the booth and looked at the list of names, they considered who were No 1 and No 2 for Labour, Liberal Democrat and SNP? I think that that is important when we weigh up the advantages of the additional member system.
I am shuddering about the idea that we might have to decide which of the people at the top of the lists would have been unelectable under another system.
I am giving an extreme example. I want to know what you have done to investigate whether the public took into account the list rankings and the names on the lists.
As far as I am aware, no study has been undertaken as yet to ascertain that. It would be very difficult. We would have to grab voters immediately after they leave a polling station, put them into a focus group and talk through those issues. As it is the first election, we do not know what the Scottish voters think about the system.
There was some work done in the European election to discover whether voters would prefer a closed or open list. About 50 per cent of the voters gave a strong indication that they would have liked to have more say about who they could vote for and where their votes went. We know that voters are concerned about the compilation of lists, but I cannot give you an exact answer as to how that worked out under AMS in the Scottish elections.
That would be an extremely difficult study to undertake, because at the time the voters look at the list, they do not know whether the people at the top of list are likely to have already won first-past-the-post constituency seats. The relevance of the ranking on the lists is not immediately apparent.
With respect, it was for the Labour party. So ranking is important.
How the lists are ranked is important to the parties, but you were asking about analysis of voter reaction.
It is important to the electorate. For example, in the Highlands, the electorate knew that Labour's Peter Peacock was very likely to get in. Are the names taken into consideration in your analysis? They made a big difference to the ranking, especially with the Labour party. With the SNP, I accept your point; with the Liberals, we were a mixture.
There's a surprise. [Laughter.]
If we take into account the experience in other countries, academics in Britain will take a number of years to catch up with what happened—academics usually look backwards rather than forwards. The names at the top of lists make a difference. Most countries that operate list systems have what they call vote getters—big names who are put at the top of the list because they drag in extra votes. Sometimes though, the people below them on the list may not be the most popular individuals. However, it will literally take years for academics to find out the effect that that had in Scotland.
I would like to go back to the idea of accountability and responsibility. Using the example that you gave of the single transferable vote, I am not convinced that being the fourth person elected, having received only a fraction of the votes in the first count, makes that person any more accountable or responsible than someone who is elected by the biggest number on the first count of the ballot. What makes someone who comes fourth, and has to wait until the fractions are added up, more accountable or responsible?
It is possible that the person who is elected by those final fractions is even more accountable. I see accountability in terms of the electorate being able to remove a councillor who they do not feel has been performing as they would like. The electoral system has to give the electorate the power to do that.
It does.
If you are elected after several rounds of counting and vote transfers, you are in what might be called a vulnerable position. You would want to improve on that before the next elections.
Someone who has been repeatedly elected under the first-past-the-post system, albeit by a minority of the people, must be doing something right. If people do not come together to remove that person, but continually re-elect him or her under the first-past-the-post system, is that person not accountable and responsible in the same way as he or she would be if elected by a small percentage in the fourth count?
We are not talking about absolute blacks and whites. If, for example, in a three-way contest, someone was elected with 40 per cent of the vote, and 60 per cent of the voters absolutely detested that candidate—
Mrs Thatcher?
I shudder at the thought of mentioning names or parties, but that 60 per cent of voters may be prepared to settle for anything other than the person supported by 40 per cent. However, because there are two other candidates, the 60 per cent may well be split between them and each may get 30 per cent. Under the first-past-the-post system, it is not only a question of the person with the most votes winning the election, it could be a question of the least popular person winning. That does not seem a sensible idea.
How does STV change that?
Even under the alternative vote system, if the vote for the three candidates was split 40:31:29, the candidate with 29 per cent would be eliminated, and—if the mood of the electorate was against the candidate with 40 per cent, as I have suggested—that 29 would be added to the 31, giving a result of 40:60, and the person who originally got 40 per cent would be defeated.
I am still not convinced that that would lead to increased accountability.
But we would not have had the poll tax.
I have a question about accountability from a slightly different angle. On the model on the slide, is it correct that, as there are eight councillors for four wards, there would be four councillors—perhaps from four different parties—who would be accountable to the four-ward area but not to any single ward?
The accountability would be to the wider ward of four wards put together.
In the Highlands and Islands, that wider ward might be the size of Belgium. Would it not be better to take STV to its logical conclusion, and allow votes to transfer within a single ward? There would be the benefit of the first-past-the-post system, in that a single person would be elected, but the result would be based on the most popular person among all the voters.
If it is used in single wards, the single transferable vote boils down to the alternative vote—as I have described, that is the system that one might have if there were three candidates. As Peter has said, the alternative vote is not proportional, and can be even less proportional than the first-past-the-post system. Labour would have done better in 1997 with the alternative vote because there were many constituencies in which Labour almost pipped the Conservatives. If Labour had had transfers of votes from Liberal Democrats, it is likely that Labour would have won more seats.
I do not suggest that one should opt for four-member wards in parts of the Highlands. Wards could be reduced to three members, or even to two in places. At lunch, we discussed Tiree, which might need to be a one-member ward. However, to have proportionality, there needs to be more than one member. There has to be a sensible compromise between proportionality and the geographic size of wards that councillors have to cover.
This is a supplementary point to the question that Jamie asked. You said that one of the problems with the elections in the spring was that voters who were voting in the first-past-the-post election did not know whether that person would be worth voting for on the list because they did not know the outcome. In certain areas of the country, electors are sophisticated enough to know in their marrows—because of national polls—that the person for whom they voted in the first-past-the-post election would not make it. They therefore made a separate judgment on people on the lists.
As several people have said, the electorate is far more sophisticated than we think. There are places where the result is a racing certainty. Alex Salmond's area is an example, but there are other places where it is equally a racing certainty that the Labour party cracks it. Voters for other parties probably know that and vote out of loyalty, but make a choice on the lists that is totally unrelated to the first-past-the-post election.
I agree. ICM and The Herald published figures just before the Scottish Parliament elections that suggested that, on average, something like 23 per cent of voters were going to vote tactically; they were going to vote differently in the regional lists and the single-member constituencies. The difficulty is that, although one might know what will happen in a single-member constituency, to vote tactically with complete confidence, one needs to know what will happen in all the constituencies in that region.
Or perhaps just most of them.
At the Scottish Parliament election, there were two regions of Scotland in which one party took all the constituency seats. According to the best of the opinion polls that were published before the election, that was likely to happen in five, six or seven of the regions, but it did not. One needs precise information at a local level if one is to exploit that peculiarity of the two-vote proportional representation systems such as the additional member system and alternative vote plus. The issue does not arise with the single transferable vote system, in which there is only one vote and only one kind of member, and people vote in a positive way. If people want to vote against particular candidates, they must ensure that they put all the other candidates in front of them.
I was involved in some of the early discussions about electoral systems for the Scottish Parliament. One of the arguments for the two-vote system was that, because of tactical voting, if people had only one vote and the proportional vote was taken off that, the result would not be a true reflection of people's opinions. I used to tell people that having two votes would allow people to vote with their heads in the first ballot and with their hearts in the second. The second vote should have given a pure representation of people's instincts. In fact, it did not happen like that, because people used their votes in a perfectly legitimate, but different, way.
You said that in 1997, if the alternative vote system had been used, there would not have been a proportional result. However, people voted tactically in 1997 in greater numbers than at any time before then. For historical reasons, the progressive forces—if that is how one wants to describe them—were split, and people all over the country made hard decisions about how to vote, based on what they saw in the opinion polls. If the alternative vote system had been used, would the results that you have extrapolated have been different if people had not voted tactically? Is the AV system flawed, or do the flaws become evident because we are dealing with figures produced by a distorted first-past-the-post system in which people vote tactically?
It is important to remember that the introduction of a new voting system might make it necessary to rethink the starting point from which people vote. The outcome of an election does not necessarily represent people's views. It is true that a lot of people voted tactically in the previous general election. The result may therefore not properly reflect the balance of support for the different parties.
However, many people did not vote tactically. There are many constituencies in which a lot of people voted Liberal Democrat and felt loyalty to that party, despite the fact that, according to the voting history of that constituency, a Liberal Democrat vote was likely to get nowhere. There are places in which Labour people continued to vote Labour although, historically, Labour was in third or fourth position, making a Labour vote purely a vote of party loyalty. Many people stuck to loyalty to their party, rather than voting in a way that could be effective. Of course, all parties were out telling their own people to boost the party vote.
With the alternative vote system, that dilemma does not arise. People who want to vote Labour, Conservative, SNP or whatever, can go out and vote for their preferred party in the knowledge that, even if, historically, their party does not have any chance whatever in that area, the vote is not wasted. Once their party is knocked out, they still have a chance of saying whom they would prefer among the candidates who remain.
That is why I was asking why you would roll that out.
On the grounds of proportionality.
Are you saying that it is not proportional on figures that would not give you proportionality because people are voting tactically?
Australia has used that system for about 60 years and, in terms of proportionality, it probably does worse than the United Kingdom in some elections. When Lord Jenkins looked at different electoral systems, he said that AV on its own has the effect of quickening the waters of political change. It tends to speed things up, so that there can be larger swings against parties. It has the advantage of giving a majority in a constituency, but it can produce larger swings in favour of and against parties, and can have a more exaggerated impact than even the first-past-the-post system can have.
In terms of the overall result, AV tends not to reflect how the nation, or the council area in this case, voted as a whole. It is a more accurate reflection of individuals' votes, but it does not represent the community as a whole. Then again, it is not supposed to: it is a majoritarian system.
Bristow, could you make your point quickly.
With regard to the point that was made about people's votes counting and people voting for what they believe in, is not it also the case that any top-up system has the problem that people may vote tactically? For example, if someone in Glasgow believed that it was likely that the Labour party would not get any top-up seats, they would say, "I will vote for my second-choice party in order to get the result that I desire." I am aware of people who strongly support Labour but who took that view in the elections.
That undoubtedly is a problem with the system. It only happens in areas in which one party is very strong. The more top-up seats that are available, the more likely it is that people will still be inclined towards supporting their party at the top-up level, because their party can still win seats. However, if the top-up is much smaller, a person may be sure that their party will not gain from the top-up and therefore may vote for what they regard as their second-choice party.
Thank you, Ken, James and Peter. I hope that you do not mind that we spent longer on this issue than we thought we would, but members wanted to get their teeth into this issue.
McIntosh's recommendations are for PR for local government. It is my view that the committee should recommend a voting system. We may recommend the first-past-the-post system: if one member has anything to do with it, we certainly will. [Laughter.] However, we are examining all systems, and I am glad that you were asked questions about other voting systems. I have questions about what McIntosh meant by the councillor-ward link, and I will pursue them with him when we meet him again. Thank you for coming along and for your presentation.
We shall have a five-minute comfort break and return at quarter to 4. The clerks told me that they would put members in alphabetical order. You moved around, so I am going to separate Bristow and Kenny. We may put that to the vote when we come back: the first vote that we have ever had will be to separate those two.
Meeting suspended.
On resuming—