Hansard

Thank you. The committee has asked me to do three things: first, to provide an overview of the electoral systems that are mentioned in the Executive's white paper; secondly, to address the question of which system best meets the criteria that are laid out in the white paper in order to judge which electoral system ought to be used in future Scottish local government elections; and thirdly, to address the implications of the Executive's decision not to reduce the number of councillors—such a reduction having been recommended by the Kerley working party. While doing those three things, I was asked also to address any recent research that I might have conducted. I will do that in the course of my presentation.Given the pressure of time, and having made some assumptions about the committee's familiarity with the various electoral systems, I will go through the first part of my presentation very quickly. The second part is undoubtedly more important.I remind members quickly that the single transferable vote is used in multimember wards. Voters are requested to place candidates in rank order, after which a counting process takes place in which votes are redistributed as necessary in accordance with the second and later preferences that are expressed by voters, until a statutory quota is reached. There is experience of such systems in both the north and south of Ireland, in Malta and in the Australian Senate.The second system that the Executive mentions is the additional member system. I assume that members are wholly familiar with that because it is, of course, the system that is used to elect the Scottish Parliament.There are two variants of those systems, which I mention in my submission. The first is the alternative vote. Conceptually, it is no more than the single transferable vote in single-member constituencies. The quota in the alternative vote is 50 per cent plus 1, which is exactly the quota that you would get if you were to work through the mathematics of STV. The alternative vote is used in Australia, and what I would consider to be a bastardised version of it is now used to elect directly elected mayors south of the border.AV-plus is simply a variant of the additional member system. Instead of what happens in Holyrood, with directly elected members being elected by the first-past-the-post system, they are elected using the alternative vote. To my knowledge, AV-plus is not used anywhere in the world, but it was recommended by Lord Jenkins as part of his commission's work into a future electoral system for the House of Commons.Those are the systems that we are considering—along with the current system that is used in Scottish local government, which is the first-past-the-post system.The Executive mentions three criteria in its white paper, although the second really comprises three criteria in itself. The first criterion is the importance of maintaining the councillor-ward link. Secondly, the white paper says:"Any new system should ensure that voters' preferences are clearly reflected … should be capable of being used … in both urban and rural council areas. A new system should not unduly act … against the interests of independent candidates".Thirdly, it says: "Any new system must manifestly be seen to have clear support".To what degree—if at all—do those criteria differ from those that have been the subject of debate ever since the McIntosh report was published in 1999? The answer is, "Not a great deal." The councillor-ward link, the need to deal with both urban and rural areas and the considerations for independent candidates were all covered in that report. The one consideration of the McIntosh commission that has been dropped is the requirement for a close fit between wards and communities which, as I explain in my written paper, is probably sensible. The one criterion that was not among the commission's criteria and that has been added is that there should be evidence of clear support.In my view, there is a substantial ambiguity in the white paper, which is the idea that"Any new system should ensure that voters' preferences are clearly reflected in the result of an election". Under one interpretation, that means proportionality but, under another interpretation, it means ensuring that there is a clear overall majority. To be frank, anyone who wishes to determine which of the systems meet the Executive's criteria faces a serious problem, because the criteria in the white paper are not sufficiently unambiguous for us to be able to find that out.In order to be helpful to the committee, I propose to indicate the degree to which AMS and STV meet the criterion of proportionality, and under what circumstances they do so. I will also address the question of the extent to which FPP meets the criterion of majoritarianism, which is usually set out for that system. That seems to be the best that one can do, given the nature of the white paper.The first criterion that the Executive says is vital is the idea of the councillor-ward link. Alas, in the debate that has been going on over the past three years, the councillor-ward link has been frequently invoked but rarely defined. It seems often to be a question of a definitional fiat; by the councillor-ward link, we mean simply that there should be one councillor to one ward. If that is what we mean, we can end all debate, because no system of proportional representation can operate where all councillors are elected in single-member wards.The alternative is to identify what are thought to be the attributes of single-member wards. That argument is mentioned in the white paper. It relates, in essence, to the degree to which single-member wards provide incentives for local councillors to act as local advocates on behalf of individual constituents and communities, vis-à-vis the local council, and to the degree to which those councillors are clearly accountable to all the voters in their wards, as a result of which voters should feel a clearer sense of representation.That is what I will mean by the councillor-ward link. It is a set of understandings about the advantages of councillors taking on roles as local advocates. It is argued by the advocates of single-member representation that that role would best be advanced through single-member wards. Equally, the advocates of multimember wards have a theoretical argument about why their favoured system promotes local service. Crucially, they rely on the argument that multimember wards mean competition between councillors, possibly even between councillors of the same party, and that that competitive process provides an incentive for individual councillors to engage in a service function that is at least as strong as that which can be found in single-member wards. I have been doing research on the subject, not so much in Scotland, but in a comparative context, and I would like to share my findings with the committee.Another piece of evidence comes from the Scottish Executive itself: the Scottish household survey. The survey for 1999-2000 asked two questions that are relevant to the debate. The first was to ask people whether they had been in contact with a local councillor in the past year about a problem or query. The answer was that 8 per cent of people had been in contact with a councillor.Throughout the debat e, one must remember that, although the probably constant flow of people coming to elected representatives with queries or problems seems terribly important to those representatives, it is relatively unimportant from the perspective of the electorate, because most people do not approach their councillors most of the time.If, as is claimed, under the existing system there is a clear councillor-ward link, one would expect most people to know who their local councillor is. However, according to the Scottish household survey, only 43 per cent of people claim to know who their councillor is. It should be noted that no attempt was made to verify the accuracy of that claim.I have been involved in comparative international work that has allowed us to ask the same questions about countries that use widely varying electoral systems. Some have single-member districts, some have multimember districts and some have mixed systems. Some of the questions that were asked in the survey are relevant to the issue that we are debating.One question relates to the extent to which people report contact with their elected representatives. As members can see, people who live in countries that have single-member districts report higher average levels of contact than those who live in countries that have mixed or multimember systems. However, one key message is that most people are not in contact with their elected representatives most of the time.If the claims that are made about single-member districts were correct, one would expect people to be able to name correctly one of the candidates in their district. If single-member districts promote a personal vote, it is to be presumed that people will know for whom they are voting. However, that claim is not substantiated by international evidence. In fact, people who live in countries that have single-member districts were less likely than those who live in countries that have multimember districts to be able accurately to remember candidates' names.As I suggested earlier, if the link that is generated by single-member systems is valuable and valued by electors, they should feel that their elected representatives are more likely to understand what they think. On that issue, the result is a no-score draw among all the various systems.Although the Executive believes that the councillor-ward link is vital and that claim has been central to the debate about electoral systems, there seems—according to that criterion—to be no significant empirical evidence that any of the systems under consideration has a clear advantage over the others. The evidence is nothing like strong enough to suggest that a judgment should be made on the basis of the councillor-ward link alone.I will try to unravel the problem that is created by the ambiguity of the Executive's definition of its second criterion. Members will not be surprised when I say that, whether we choose AMS or STV, the outcome of elections is likely to be significantly more proportional than is the case under the existing system. One of the key characteristics of Edinburgh, for example, is that currently the SNP gets one fifth of the vote there, but has only one seat on the council. The Labour party receives less than one third of the vote, but has more than half the seats on the council.If the Executive is concerned that electoral outcomes should be more proportional, the real issue is not a choice between STV and AMS; rather it is how those systems are implemented. In the case of the single transferable vote system, the size of wards is crucial to determining how proportional the system will be. In the case of AMS, the key factor is the proportion of seats that are additional. I illustrate that point by showing what would happen to Labour's position in Edinburgh under different systems. Under a four-member STV system, or an additional member system in which 43 per cent of the seats were allocated as additional seats, the Labour party would receive one third of the seats from one third of the vote. Under a three-member STV system or an additional member system in which only 25 per cent of the seats were top-up seats, the Labour party would end up with nearly 40 per cent of seats. I have provided the committee with further examples in my written evidence. If our concern is proportionality, we should focus on the size of wards under STV and the size of the top-up under AMS, rather than on the choice between AMS and STV. Each system can be made almost as proportional as the other, as long as the system is implemented in a way that is designed to achieve that.However, there is one clear difference between the two systems. Under the single transferable vote system, results are influenced not only by the first preferences of voters, but by their second preferences. For that reason, the system's impact will depend on the distribution of second preferences among voters. The slide on the screen now shows the second preferences of voters in the 1999 elections. For example, 19 per cent of Conservative first-preference voters indicated that their second preference was for the Labour party.There are three crucial facts to derive from the table. First, nobody loves the Conservatives; therefore, they will be disadvantaged at present by any transferable vote system. Secondly, the Liberal Democrats tend to be relatively liked by everybody else. Thirdly and not least, the Labour party is preferred to the SNP among all the other parties' voters. Therefore, in a narrow contest between Labour and the SNP, the Labour party benefits from transfers.The alternative vote system that the Executive's report mentions is also influenced by second preferences. It is therefore worth considering what would happen under that system, given those preferences. Given what I have said, what members see on the screen will not surprise them. The slide entitled "AV in Edinburgh" is a simulation of how the alternative vote system would operate in Edinburgh. The SNP's one ward in Edinburgh was won only narrowly. Under the alternative vote system, the SNP would be left with no seats in Edinburgh, and the Labour party would be slightly better off than it is under the current system, as would the Liberal Democrats. One can say simply and readily that, if the aim is to achieve proportionality, the alternative vote system should not be considered further. It is in no sense a proportional system and is capable, as in the case of Edinburgh, of producing results that are even less proportional than the current system. How does the first-past-the-post system do at delivering majorities and how good is it at ensuring that a majority goes to the right party? Does it treat large and small parties, in its own terms, fairly? The Executive asks all those questions and they are all crucial to the claims that are made for first past the post. I argue that, whatever one thinks about STV and AMS, it is not evident that FPP meets the criteria that its supporters commonly uphold for it.For example, advocates of FPP tend to dislike coalition Governments or hung councils, but alas—from their perspective—we already have them in Scotland. The 1999 elections generated 10 hung councils and only half the councils in Scotland had a majority single party running them. It is already the case that FPP relatively frequently generates non-majority councils in Scotland. It is also the case that, if the recommendations of the Kerley report were implemented, eight or nine councils would still have a majority party using the figures from the 1999 elections.In truth, the debate between FPP and STV—and probably also AMS, depending on how it was implemented—is not that one is clearly majoritarian and the other is clearly proportional. It is an argument about degrees of proportionality and majoritarianism. The difference between them is much narrower than the debate commonly assumes.The second problem that FPP faces is that there is no guarantee that the party that gets the most seats is the party that has the most votes. That is clearly a failure of the system according to the criteria that its advocates set up. The slide entitled "Wrong Winners" gives three examples of where that has happened in Scotland.The third failure of FPP is that it is not the case that it simply discriminates against small parties and is kind to big parties. If that were true, for example, the Liberal Democrats and the Conservatives should be treated more or less equally in Scotland. They are not; the Liberal Democrats basically get their proportionate share and the Conservatives are discriminated against. If it were the case that FPP advantaged big parties and not small parties, it should disadvantage the SNP less than it does the Liberal Democrats. As members can see from the slide entitled "FPP Scotland-wide in 1999", it fails that test as well. Therefore, I suggest that if, by saying that any system should reflect clearly the views of the electorate, the Executive means that councils in Scotland should have majorities that are clear reflections of at least the rank orders of the parties in terms of shares of the vote, FPP is clearly not the right instrument through which to achieve that objective.Two other criteria are embedded in the Executive's second criterion for a new system. I will dispose of those briefly. One is that the system should be capable of being adapted to urban and rural areas. My view is that it would be possible to do that and that that would be made easier by the Executive's decision not to reduce the number of councils. It could be done simply by changing the sizes of wards and the proportions of additional seats. Secondly, STV is clearly rather more friendly to independents than AMS is. That is probably the one clear way in which one can distinguish between the two systems.The third criterion is "clear support". It is worth noting that in saying that a"new system must … have clear support",the Executive does not say support by whom—it could simply mean that all the political parties are in favour of it. I have assumed that by stating that criterion, the Executive means that a new system will require significant evidence of public support, rather than simply support among politicians.There is relatively little research on the public's attitudes to electoral reform for Scottish local government—it is not the world's most exciting subject. However, I am aware of two pieces of survey research evidence. The first comes from a question that was put in a System Three poll by the Electoral Reform Society. As the slide shows, there was heavy support in favour of the principle of proportionality.The second piece of evidence comes from research that I was involved in at the time of the 1999 Holyrood election. We used the fact that we had a group of voters in Scotland that were used to both a proportional system and a first-past-the-post system to ask more questions about proportional representation than respondents had ever been subjected to previously. For the most part, the questions were about Holyrood and Westminster rather than about local government, but one question asked whether the Holyrood method of voting should be used in future local government elections and there was fairly substantial support for that proposition.It would be difficult to argue that there is clear evidence that the electorate does not want a system that is more proportional. However, I acknowledge—my submission goes into more detail on this—that the answer to such questions depends on how they are asked. If we ask people whether the UK should introduce PR for the House of Commons, everybody says yes. If we ask them about the relative merits of proportionality versus strong government, we get a different answer. Having said that, there is a difference between the two sets of answers for Holyrood and Westminster. It may be that concerns about strong government are particular to Westminster and do not necessarily apply to Scottish local government.What can give you some indication of the aspects of the debate about electoral reform that resonate with voters? Some of the debates do not resonate as much as people think, whereas others do. For example, the arguments that proportional representation leads to unstable government or that people prefer single-party government do not resonate. Only 16 per cent and 14 per cent of Scots voters respectively endorsed those propositions.In contrast, the idea of having one MSP for an area rather than several is something that does resonate with the public. The evidence is that although there is probably broad support for the idea of proportionality, there are aspects of alternative systems that are less popular and it is not the case that all the evidence supports the argument of one side against the other. However, I have come to the view that it would be difficult to argue that we could not change the system at all because there was strong public opposition to such a change.The second aspect of public support that might be considered to be important is the argument that voters might withdraw from participating in elections because the complexity of the system meant that they did not understand what they were doing any more. My next two slides show that most people did not find complexity a problem. Very few people had trouble filling in the ballot paper in 1999, and although rather more people said that it was difficult to understand the relationship between votes and seats, there is no evidence to show that that led people not to vote.Moreover, on one of the rare occasions that academics agree with one another, almost all the international academic evidence on electoral systems and turnout suggests that turnout in countries that have PR tends to be higher than in countries that do not. The academics differ about the size of the effect and some of the evidence shows a difference that is not particularly large; however, they all agree about the direction of the effect. It would be very difficult to argue that we should not introduce either STV or AMS to Scottish local government on the ground that the complexity will put off voters. That argument does not seem to have much basis.The third thing that I was asked to do was to consider the implications of the Executive's decision to keep the same number of councillors, regardless of the choice of alternative system. That would make it easier to introduce either of the alternative systems under consideration, compared to Kerley's proposals. If we do not attempt to reduce the number of councillors in an area, under STV one or other or a combination of two things could be done. We could have more councillors per ward, which would make the system more proportional. Alternatively, we could have a smaller average electorate per ward. Four-member wards could have a smaller electorate, because we would have more councillors overall.If we are concerned about having big wards in rural areas, the Executive's decision has made it easier to implement STV than it would have been under Kerley's original recommendations. However, the Executive's decision also provides more flexibility for AMS. What could be called the extra councillors that the Executive has brought back into the system could be used to provide more additional seats for any given ward size under AMS—that would create more additional top-up councillors, which would make the system more proportional. Alternatively, the extra councillors could be used to create smaller wards with a smaller average electorate size in the single-member districts, which might make the system easier to implement in rural areas.Apart from what might be considered its political wisdom of not asking turkeys to vote for Christmas, the decision makes it easier to deal with the potential conflict between the wish not to have wards that are too big, particularly in rural areas, and the wish for greater proportionality, which may or may not be embedded in the Executive's white paper.