Voter Choice
138. Under a reformed system it is crucial that the voters' right to express their view of individual candidates should be at least maintained and preferably enhanced. FPTP does retain the right, theoretical in most cases but occasionally practical as the last election showed, to get rid of a deeply distasteful candidate even in a nominally safe seat. It would be a count against a new system if any candidate, by gaining party machine endorsement for being at the head of a list, were to achieve a position of effective immunity from the preference of the electorate. This is the essence of the case for open as opposed to closed lists for Top-up members.
139. The practical importance of the issue can be exaggerated under a Top-up system as devolved as that which we propose. If there is in most cases no more than one Top-up seat for which to compete, and in no case more than two, parties are unlikely at the maximum to put forward a top-up list of more than three. That it should slightly exceed the number of seats available is desirable in order to provide for list vacancies between general elections, which will be dealt with in paragraph 143. Nevertheless it remains essential that the elector should have two rights; first to bolt the party ticket completely with his or her second vote, in other words to vote for a candidate of one party for the constituency and then to cast his or her vote in a different direction for the Top-up representative or representatives. Without this right the new system would not fulfil the objective of freeing the voter from the prison of having to suffer an unwanted candidate for the constituency in order to get a desired government. Second, however, it is equally desirable that the voter should be able to discriminate between the candidates put forward for the list by the party for which he or she wishes to cast the second vote. Only if this is so does the Commission feel that it will have sufficiently discharged its third requirement of providing for an extension of voter choice.
140. Equally, however, it should be recognised that all electors will not always wish to discriminate between candidates of the same party, perhaps for the very good reason that they know nothing about either or any of them. Where this is so a meaningless choice should not be enforced and the ticking of a party box should be an option. The desirability of freedom of choice is not invalidated by the fact that not everybody wishes to exercise it.
141. Such freedom does involve a more complicated ballot paper than would be necessary under a more party-dominated system. If the right of the voter to split votes between constituency and list choices were eliminated only a single vote need be cast, with its consequence following through to determine the Top-up as well as the constituency result, and with the ballot paper looking exactly the same as in a FPTP election. But to abrogate a right to differentiate which 37% of New Zealanders and 12 1/2 % of Germans have chosen to exercise would be an unacceptable and unnecessary piece of caucus authoritarianism. In the same way the absence of choice between individual list candidates within a party would also somewhat simplify the ballot paper. Freedom always comes at a price, but this in our view is one which is worth paying. A typical sample ballot paper, assuming five parties, the average number of candidates for a constituency at the last election, is at Annex B.
Northern Ireland
142. In recommending an alternative voting system for Westminster it was clearly essential for the Commission to be confident that it could be applied with equal validity throughout the United Kingdom. Our attention focused particularly on Northern Ireland where the political landscape differs in a number of ways from the rest of the United Kingdom. We found no sufficient case for recommending a different system for Northern Ireland. We were much persuaded to this view by the evidence submitted to us by the Northern Ireland Forum for Political Dialogue, and in particular the main conclusion of the Northern Ireland Committee on Electoral Reform that, while they hoped for a recommendation in favour of STV (as already used in Northern Ireland for local elections, elections to the European Parliament and the Assembly) it would in any event be preferable that the same system should be applied for parliamentary elections throughout the United Kingdom. Such uniform application we recommend, subject only to a special provision that there should be a minimum of four Top-up members (divided between two Top-up areas) for Northern Ireland. We believe this is necessary to accommodate the more complex party system which there operates.
By-Elections
143. By-elections for constituency seats under the recommended system present no problems. They will be fought exactly as they are today except for the substitution of AV for FPTP. Where vacancies occur for Top-up MPs the position is more complicated, but it is hardly a major issue, for it would be surprising if more than two or three such vacancies were to occur throughout the country in the course of a parliament. Election to a representative position should wherever possible be preferred to selection but it is difficult to see how one could be made to work in these rare circumstances. If a straight city or county-wide contest were to take place it would almost by definition result in the victory of the predominant party in the area, thus negating the essential purpose of the Top-up seats. If a highly complicated formula were evolved by which an election could take place but the results were adjusted so as to prevent this negation the outcome would suffer all and more of the disadvantages of turning best (or second or third best) losers into winners. In these isolated cases we therefore recommend that the next candidate on the Top-up list should move up. If for any reason there is no available person in this position the seat should remain vacant until the next general election.
Thresholds
144. Is a threshold necessary in order to deter the securing of Top-up seats by very small splinter parties such as has tended to discredit the Israeli and the previous Italian systems? They of course are or were full list systems and thus very different from the limited Top-up which we are advocating. Nevertheless it has generally been accounted a favourable aspect of the German system that it has required a party to secure either 5% of the total vote or success in their direct constituency elections before it can qualify for any Top-up seats. This has been an effective safeguard against the fissiparousness of the system prevailing under the ill-fated Weimar Republic, although 5% necessarily has an arbitrary quality about it. Nonetheless if we believed that a threshold was necessary to prevent the evils of excessive splintering, we would certainly propose one. But all the indications are that it is unnecessary. At the last election under our highly devolved system with no more than one or two list seats for an area the lowest percentage of the total vote which would have placed a party in likely contention for a Top-up seat would have been the 10.9% scored by the Liberal Democrats in Nottinghamshire. As no-one could reasonably advocate a threshold nearly as high as that it would be a classic example of imposing an otiose constraint. We do not therefore recommend a formal threshold; our system itself imposes an informal one more severe than the Germans' 5%.
Tactical Voting
145. Before we come to estimating the likely effects of the new system it is necessary to discuss one count against it which has been raised in advance. This is that it offers scope for tactical voting on a scale which would damage both the greater proportionality of the new system and its ability to counteract the "electoral deserts" for major parties scenario. Thus, to take an extreme case, Labour voters in Glasgow, knowing that because of their party's constituency dominance, it would have no chance of winning a compensating Top-up seat, might on a massive scale switch their Top-up vote to the Liberal Democrats, thereby depriving the Conservatives, whose real strength across the city is stronger than that of the Liberal Democrats, of the Top-up seat to which they should be entitled. It is easily possible to see the theory of the argument. The Commission, however, having examined it carefully, believe that its practical effects can be grossly exaggerated. Its comments on the issue follow in the next five paragraphs.
146. All electoral systems are open to a degree of tactical voting. This is certainly true of FPTP, where tactical voting was fairly widely practised in the special circumstances of the 1997 election, as was expounded in paragraph 128. There is nothing morally wrong about either informal tactical voting or the formalisation of alternative choices under AV. In many situations of life a decision has to be made in favour of a second or third best choice and there is no inherent reason why what has often to be applied to jobs, houses, even husbands and wives should be regarded as illegitimate when it comes to voting. The point at issue is the narrower one of whether with an Additional Member/Top-up system tactical voting can block the objective of the corrective mechanism giving greater proportionality.
147. However the evidence is that effective tactical voting is very much a minority occupation. Not much more than one in ten voters attempts it, and a much smaller proportion achieve the result they intend. To suggest against this background that under a new system and in the fog of battle which accompanies an election, parties are going to be able to manoeuvre their votes, not in their own favour but in favour of another party, with all the precision of guards' battalions on a parade-ground, seems to us distinctly far-fetched.
148. For this to happen three unlikely conditions would have to be met. First, each party with votes to spare would need to find and convey to its supporters a complete and fairly precise confidence in the outcome before it had taken place. In retrospect the result of the 1997 election looks one of the most certain in living memory. Yet there was much nervousness and uncertainty of mood in the Labour party during the campaign, just as there was in the Conservative party (to which eloquent testimony is paid in the memoirs of the three members of the high command at the time) in the run-up to the almost equally inevitable 1987 victory. This is what is meant by the 'the fog of battle'. Second, the properties and likely result of a new and somewhat more complicated system would have to be understood and foreseen with a clinical precision which has rarely been associated with the old familiar system. And third, the orders based on this precise appreciation would have to command the obedience not just of militant cadres but of a somewhat inchoate mass of voters.
149. Any reformed system is likely to be attacked by its opponents on the ground of complications which will confuse the voters. This in our view is not a valid point, as is shown by the much higher turnout in, for example, Germany in September, 1998 as compared with the very low turnout under the simplifications of the United States' system. But also to claim that it will be subject to the most sophisticated manipulation not only by party organisers but the voters themselves is surely a classic example of trying to have it both ways.
150. Furthermore the indications are that the maximum probable amount of tactical vote-splitting would have been unlikely to make the results less proportional in 1997, and in the more normal conditions of 1992 would have made it distinctly more proportional. There may, compatibly with this, be some mild adverse effect upon a party which gets itself into as unelectable a position as the Conservatives did in 1997 (and the Labour party in 1983), but it is difficult with most electoral systems, certainly including FPTP, wholly to protect a party against the consequences of such positions.
The Number of Top-up Members
151. We now come to the next crucial question of what should be the size of the Top-up, and the best, but not necessarily certain, estimates which we can make (on the basis of psephological advice) of what likely results this combination of AV and limited Top-up would have produced for 1992 and 1997. We also give an indication of what might have been the 1983 and 1987 outcomes although it must be appreciated that the further back you go (rather in the way that the further forward a weather forecast goes) so the reliability of the estimates becomes less.
152. In considering the level of Top-up we are required to balance carefully the potentially competing criteria set out in our terms of reference. On the one hand the importance of maintaining the link between MPs and their constituencies and the need to ensure stable government - to the arguable extent that this requires single party majority government most of the time - pushes towards keeping the level of Top-up as low as possible. On the other hand the requirement to deliver broad proportionality would push us towards a larger Top-up sufficient to correct, or at least substantially to ameliorate, potential disproportional outcomes on the constituency side.
153. We were further constrained in our deliberations by the unavoidable reality that the political landscape in which we are operating is not static, and that, regrettably, we cannot realistically expect our recommendations to be in operation at a general election in much less than eight years. Changes to the landscape are already clearly visible on the horizon. Legislation to remove the statutory minimum for parliamentary constituencies in Scotland is currently before Parliament. This will inevitably result in a reduction in the number of the Scottish MPs at some point in the future - probably following the next Scottish Boundary Commission Review which is due to start in 2000. The level of over-representation in Wales is less dramatic but, in the wake of the creation of the Welsh Assembly, Parliament may want to address it as well. Furthermore it is impossible to judge the extent to which changes to the criteria which the Boundary Commission use for setting parliamentary boundaries, and which were the subject of criticism in an Home Affairs Select Committee Report in February 1987, can reduce the bias which, as we have described earlier, can affect different parties at different times, and which is currently working so forcefully against the Conservatives. A second stage of House of Lords reform is also clearly a strong possibility within the time-scale we are considering.
154. This has not been an easy circle to square. We feel we can best do so by identifying a narrow range within which that level should be set in the light of developments outlined above. Our investigations (see Annex A) suggest that a Top-up of between 15% and 20% of MPs would do sufficient justice to the three competing criteria discussed above to be acceptable. It will be for Parliament to decide after the referendum (if favourable to change) on the basis of the evidence before it at the time at what point in that range the specific limit should be set. It will be crucial that the evidence provided to Parliament for this purpose is soundly-based, fair and demonstrably non-partisan. In our view this evidence would be best provided by an independent body such as an Electoral Commission. We discuss this in the context of the recent Neill Committee recommendations in paragraphs 166-168.
155. For the sake of simplicity we think it best to give our estimates of the likely 1992 result under this recommended system at the middle point of the bracket, that is 17.5%. We obviously would not claim full precision for the exact numbers of seats which would have been won by each party, even though they have been arrived at with professional and impartial advice. We think it highly unlikely, however, that any margin of error for any party would exceed a handful of seats.
Table 2-1992 Election*
|
|
CON |
LAB |
LIB DEM |
SNP/PC |
Various
NI parties |
|
| FPTP |
336 |
271 |
20 |
7 |
17 |
| AV Top-up |
316 |
240 |
74 |
11 |
18 |
|
156. Instead of the weak and eroding Conservative majority which characterised the next five years, Mr Major would therefore have found himself from the start in a hung parliament, and a truly hung one, for a Labour/Liberal Democrats partnership would have been short of a majority, indeed just short of the Conservative total, and the Liberal Democrats had already moved to a sufficiently anti- Conservative position, not surprisingly perhaps after three Conservative parliaments, that a Major/Ashdown coalition, which could have commanded a majority, would have been impossible. The probable outcome would therefore have been an early second election, for which there have of course been several precedents under FPTP. It could easily be argued, however, that this might have been preferable from the point of view of decisive government than the five years of uncertain power which followed. It could also be argued that such an uncertain sound of the trumpet would have been a true reflection of the national mood in 1992 - a feeling that it was time for a change accompanied by a hesitation about entrusting power to the only partially reformed Labour party of the time, and that there is no need to apologise for an electoral system which would have accurately have reflected this uncertainty.
157. On the same basis our estimates for the 1997 result are:
Table 3
|
|
CON |
LAB |
LIB DEM |
SNP/PC |
Various
NI parties |
OTHER |
|
| FPTP |
165 |
419 |
46 |
10 |
18 |
1 |
| AV Top-up |
168 |
368 |
89 |
15 |
18 |
1 |
|
| *It should be noted that the projection prepared for the Commission superimposed 1992 voting patterns on to the scheme put forward by the Commission and therefore assume a House of Commons of 659 members. |
158. As will be seen this would not have prevented the Labour Party retaining a substantial overall majority of 77 - and one of 200 over the Conservatives - although it would of course have reduced the 'swollen' swing in seats. It would have substantially although not wholly eliminated the injustice to the Liberal Democrats (their strictly proportional entitlement was 111) and it would very marginally have improved the Conservative representation even at a time when their fortunes were nearly beyond the help of any electoral system.
159. A further insight into the proportionality of our recommended system can be provided by the test of a statistical measuring rod known as a DV score, which measures the degree of deviation between a party's share of the vote and its share of seats. Again this rod does not have absolute validity but it is a useful indicator. Using this rod our researches show that when compared with FPTP our Top-up system reduced DV by one half (from 18 to 9) in 1992 conditions, and by just over one third (from 21 to 13. 2) in 1997 conditions. While these outcomes fall to a greater or lesser extent short of full proportionality (which, however, is generally considered to be achieved as fully as is normally practicable if the figure falls in the range of 4 to 8 ) this reflects our wish to minimise geographical disturbance and the prospect of constant coalition. The 1992 score also compares remarkably favourably with the outcome in the last Irish election, when their DV was actually higher at 9.8. The comparison is remarkable because STV (there operated) is generally considered by the most austere electoral reformers to be the epitome of desirability. It should however be noted that in the last but one Irish election the DV score was down to 6.8 and that in 1997 the British estimate is a good deal higher at 13.2. But 1997 in Britain was a 'bucking bronco' of an election which was very difficult for any system fully to control.
160. Looking further back to 1983 and 1987 our own estimates are that our recommended system would on both occasions have produced overall Conservative majorities, of 30 in 1983 and 20 in 1987. Even allowing for a wider margin of error it is improbable that the governing party would have been overturned. These majorities, despite the Conservatives' vote shares in 1983 and 1987 being not very different from that achieved by the Labour party at the last election, would be considerably smaller than that of Labour in 1997. This must be in part due to the persistence of bias in any system largely founded on single member constituencies. The need to address this bias is integral to the successful implementation of our system (see paragraph 164).
161. Our recommendation would therefore have produced single party majority Government in three out of the last four elections, with the only exception being a parliament which, even under the old system, exhibited many of the features of uncertain command. It is therefore difficult to argue that what we propose is a recipe either for a predominance of coalitions or for producing a weakness of government authority, except when it springs out of a hesitancy of national mood which may rightly show itself through any electoral system.
|
