Scottish Parliament elections: the disastrous PR arithmetic
The results of the elections to the Scottish parliament, held on 3 May 2007 under a spectacularly complicated form of Proportional Representation (PR), have created a more chaotic situation than even the most vigorous critics of PR — such as me — could have predicted. Contrary to the opinion polls before the election, which promised the Scottish National Party with its programme of independence for Scotland a decisive lead over Labour, in the end the SNP was only one seat ahead with 47 to Labour's 46, making it the biggest single party, but only just.
The assumption had been that no one party would win an overall majority in the new parliament, which proved correct, but that either the SNP or perhaps Labour would be able to command an overall majority by recruiting the LibDems into a coalition — which did not prove correct, since the Lib Dems couldn't achieve an overall majority (requiring 65 seats or more) by adding their 16 to either the SNP's 47 or the Labour Party's 46. As a result, an overall majority for a new Scottish government now requires a coalition or at least an alliance of at least three of the parties. The Lib Dems have not made the horse-trading (which customarily follows a PR election to a government-forming assembly) any easier by declaring that they won't go into a coalition with either the SNP (unless its lively leader, Alex Salmond, drops his key policy of holding a referendum on Scottish independence during the life of the current parliament), or with Labour, its senior coalition partner in government until the election. This makes the formation of a workable coalition almost impossible, unless we can imagine a German-style 'rainbow coalition' of Labour (46) plus the Conservatives (17) plus either the two Greens or the [three] independents [actually only one independent, so still not enough for an overall majority — see Comment below] or some combination of these, sufficient to bring the total up to 65 or more. Such a coalition would find it hard to agree on almost any policies whatever, apart from keeping the SNP out and opposing its proposal for a referendum on independence. But since the SNP doesn't in any case have the votes to get the legislation for a referendum through the parliament, the agonising compromises required by all the partners in a rainbow coalition wouldn't seem to be justified. Anyway, neither the Greens nor the independents seem likely to share the tent with the Tories, or even indeed with Labour. Moreover, a coalition of natural enemies designed to deprive the biggest single party of the chance to form a government would seem to much of Scottish public opinion to be a dirty trick, whatever the arithmetic.
Unless the Lib Dems relent and join the SNP and (probably) the Greens in a coalition, it now appears likely that Mr Salmond will try to form a minority SNP government which will not command an overall majority in the parliament, and so will be forced to negotiate with other parties over every major policy initiative or piece of legislation in order to cobble together an ad hoc majority, issue by issue. The result, if it works at all, will be a patchwork quilt of unpredictable policies, collectively incoherent and unrelated to any single party's manifesto commitments for which Scottish voters voted. Since any proposal by the minority government will be subject to the veto of almost any combination of two other parties (or more), nothing remotely radical or controversial will get through. Policy outcomes will reflect the lowest common denominator of consensus, a recipe for almost no change in any direction. And it will depend on superhuman restraint on the part of the other parties in not putting down a vote of No Confidence in the SNP government which, if carried, would force the Salmond administration to resign. (This would not necessarily lead to fresh elections, especially if it happened soon, since the Labour leader could still try to put together a coalition in place of the SNP's minority government.) But if within 28 days of polling day Salmond fails to form a government able to win an initial vote of confidence (the Lib Dems and perhaps Labour probably abstaining in order not to wreck the enterprise right at the outset), fresh elections will have to be held — quite possibly leading to the same stalemate.
The implications of all this for both the Liberal Democrats (with the rude awakening from their dreams, not of winning the election, but of finding themselves king-makers, a perfectly understandable if disreputable ambition for the party in third or fourth place in a PR election) and for Proportional Representation generally, are spelled out by the ever reliable Simon Jenkins in a magnificent philippic in the Guardian of 9 May with the stimulating title: Nice but hopeless, the Lib Dems should call it a day. This has predictably prompted loud cries of unconvincing protest from the PR fanatics, including notably the inexhaustible Polly Toynbee and the so-called Electoral Reform Society, both seeking to persuade us that the kind of aimless, inconsistent ad-hockery which PR has wished on the unfortunate Scots in lieu of coherent government is somehow preferable to government by a single party which can be held accountable to the electorate for the fulfilment of its election promises — the commonest outcome of elections by First Past the Post.
With the future of the United Kingdom and its union between Scotland and England in the balance, these are not matters to be lightly dismissed. A party bent on Scottish secession from the UK has emerged as the biggest party in a democratic election. The consolation is that the parties opposed to secession comfortably outvoted the SNP. But with the current PR system, and the possible permutations and deals leading to the formation of a government so hopelessly uncertain, anything could happen.
As if this potentially disastrous situation were not bad enough, there's the added consideration that the PR system used for the Scottish parliament is almost unintelligible. One of the conditions of a functioning democracy is that ordinary people should be able to understand the system used to choose a legislature and a government: if they can't, it's impossible for them to make rational choices about how best to vote. Contrary to much media theorising and a large number of spoiled ballot papers, the ballot papers themselves are simple enough. But you need to be a Senior Wrangler to understand the way they are counted to produce the result:
The formula used for deciding which parties win regional top-up seats is known as the d’Hondt system. First, party list votes are totalled from each of the constituencies making up the region.These totals are then divided by the number of constituency seats each party has won, plus one. The party with the highest total after this calculation elects one additional member. That party’s divisor is then increased by one (because of its victory in the first round) and new figures calculated. Again, the party with the highest total wins a seat. The process is then repeated until all seven additional members are elected.
Under the d’Hondt system, it is intended that the parties which perform well in terms of constituency votes, but fail to translate that success into elected constituency members, will be rewarded via the additional member system. Conversely, parties which do well in terms of securing constituencies, will win fewer top-up seats. The effects of the AMS system are best illustrated by example. In Lothian [in the previous elections] Labour obtained 6 seats with 35.4% of the constituency vote. The SNP on the other hand gained no constituency seats in Lothian despite obtaining 21.1% of the constituency votes. The “top-up” seats aim to rectify this disproportionality; however, with over half of the seats allocated by the FPTP system,the regional vote cannot fully compensate for disproportionality on the constituency vote. Thus of the 16 seats available in Lothian, the Labour Party obtained 6 seats or 38% of the available seats. The SNP were not allocated any seats from the Lothian constituency vote, but gained 2 seats from the Lothian regional list equating to 12.5% of the total available seats in Lothian, correcting some but not all of the over-representation of Labour in the first vote. [http://tinyurl.com/2pvnlm (PDF file)]
I hope that's clear….
Brian
The d'Hondt system is used in Spain and other European countries. The single transferable vote is used in Ireland. France has a deferred alternate vote system. The Netherlands uses a proportional system with the whole country as one constituency. Belgium has a dual system of parties for the French- and Flemish-speaking parts of the country. Germany has a dual system whereby half of the seats in federal and regional assemblies are elected by FPTP in constitutencies and half are elected under a proportional system (with a complicated topping up procedure). All of these countries have stable governments, usually involving coalitions.
But anything other than straight FPTP and one-party majority government is too subtle and complicated for the English mind. Ho hum.
Brian writes: Peter, I take it that you mean the Scottish mind rather than the English? I'm familiar with the d'Hondt system from experience of it in Australia (which is a real test-bed for different voting systems), and it seems to me to represent the triumph of loony theory over experience and common sense, however many of our fellow-Europeans may use it. (I wonder how many of them really understand its ramifications well enough to calculate how best to cast their top-up votes!) There are plenty of examples of unstable government resulting from PR: perhaps the most extravagant is the occasion in 1982 when the Free Democratic Party (FDP), the German equivalent of our Lib Dems (the smallest of the three main parties), switched its support from the ruling Social Democrats (SDP) to the opposition Christian Democrats (CDU/CSU) and thus triggered a change of government from left to right without a single German voter having any say in the matter. PR has also forced the Germans and other countries to resort to left/right Grand Coalitions which largely deprive the electorate of the benefits of an active opposition to hold government to account. PR is fine for deliberative assemblies that have no government-creating role, such as a properly reformed second chamber at Westminster. Incidentally I have no objections to the Single Transferable Vote system for any kind of elected body, but anyway STV isn't PR. However, all these are stale old arguments that have been extensively aired in comments on earlier Ephems posts: the point of main interest now is the ramifications of the Scottish election results.
Brian, as I think I've said here before, your basic complaint against PR, that it involves essentially unaccountable coalition-building and horse-trading, just doesn't hold water, if that is intended to show that FPTP is superior. FPTP also involves coalition-building and horse-trading, just at a different, and markedly less accountable, stage from the coalition-building and horse-trading of PR, that of the creation of political parties. PR mitigates this by reducing the costs, in terms of the likelihood of having influence on the political process, of defecting from large parties, and so reducing the incentive to form winning coalitions before the vote. The Scottish election results strike me as markedly less disturbing the British ones for the past thirty years or so, which have been marked by generally strong central governments, none of which ever won the votes of the mahority of the British electorate.
Brian writes: Rob, thanks for that, but please forgive me if I don't deploy here the counter-arguments to what you say all over again. I'm interested here in the specifics of the Scottish election results, how they have been affected by the particular form of PR under which they were conducted, and their implications for Scotland's future as part of the UK. Even the most dedicated enthusiast for PR must surely recognise that in this very particular case, the election results have produced a set of problems that will make it very difficult indeed for Scotland to enjoy stable and coherent government while its parliament is composed as it is at present. We have had a generally illuminating but inevitably inconclusive debate elsewhere on this blog on the pros and cons of PR in general: it's how Scotland is going to cope now, after the elections of 3 May, that seems to me of main interest now. Simon Jenkins's article cited in this post is required reading, I submit, on that subject.
Brian, that's fair enough, as long as you understand that the desire for a stable and coherent government is not, and may be in direct tension with, the desire for an accountable or representative one. Indeed, insofar as stable and coherent government is threatened by coalition-building, I'd submit that it's unlikely that FPTP has a decisive advantage here, because it also involves coalition-building. Jenkins, after all, is cherry-picking his examples: Israeli politics are notoriously corrupt, I think, and I'm sure we could find FPTP systems that don't work as well as one might like. I don't know enough about German or Dutch politics to be sure of this, but I'd guess that neither of them does particularly poorly on sensible criteria of stability or coherence of governments despite PR. Either way, none of this really addresses what you say is your main point, the outcome of the Scottish elections. I suppose I am just much more sanguine about this than you; if the Scots wanted to divide up their votes in this way, then it is only fair that they get a legislature made up of parties divided up in this way. They'll sort themselves out in the end, I'm sure, either by forming some kind of coalition, or by having new elections, which if the electorate is as disturbed as you by the lack of strong central authority, presumably has a good chance of producing such an authority.
Brian writes: Rob, I don't want to prolong this unnecessarily, but I would just take issue with you on the question of the electorate "wanting to divide up their votes" in a particular way in order to produce a particular result, or "having a good chance of producing … a[n strong central] authority [if disturbed by the lack of one following the recent Scottish elections]." But even with the top-up system now used for elections to the Scottish parliament, it's impossible for any individual voter to vote in such a way as to produce any particular outcome in terms of the distribution of seats among the parties and the consequences of that for stable or unstable government. All the evidence is that most voters — i.e. the majority which votes for one of the two parties theoretically capable of forming a government — would like their chosen party to win with an overall majority that would enable it to govern without the need to compromise on policy with a coalition partner. In British general elections, you can in principle vote for the Lib Dem candidate in the hope that the Lib Dems will be able to blackmail one of the two bigger parties in a hung parliament into agreeing to introduce PR for future elections as the price of being put into No. 10 by the Lib Dems as king-makers (with rarely more than 20 per cent of the votes cast!). A Lib Dem vote might also in some circumstances make a hung parliament likelier than a vote for Labour or the Tories (or, in Scotland, for Labour or the SNP), so those like yourself who value a 'representative' (coalition) government more highly than a stable and accountable one might rationally vote Lib Dem. But the fact that 70 to 80 per cent of the electorate votes for parties other than the Lib Dems suggests, doesn't it?, that this is very much a minority view. (I say 'stable and accountable' because a one-party government with an overall majority in the relevant legislature can be held to account by the electorate for its success or failure in carrying out its manifesto promises and policies: whereas a coalition government with policies determined by horse-trading after the election is a hybrid that not a single voter voted for and one that can't be held to account over its election manifesto because by definition it didn't have one.)
I agree with most of what you say, and I'm surprised at the way the Libdems seem to be behaving.
But are you sure the Scots would be any better off with FPTP?
Brian writes: No, I can't be sure of that. All one can say is that they would stand a much better chance of getting a clear-cut result and an executive able to govern under FPTP than under the present system (which, as the quotation in my post here shows, is anyway far from guaranteeing exact proportionality). As it is, no-one will have a clue from one week to the next which of the SNP government's proposals will attract enough support from enough of the other parties to get through and which will not. This will make it very difficult for businesses and indeed for ordinary citizens to plan since no-one will be able to predict what tax or planning or education policies will be in place more than a month or two ahead. The safest guess is that not much will change as this stalemate looks like a recipe for any proposals for change to be rejected. What a mess! But the PR enthusiasts will have the dubious satisfaction of knowing that the Scottish parliament reflects the balance of opinion in the electorate somewhat more accurately than it would have done under FPTP, although what concrete advantages will flow from that remains to me a mystery.
Incidentally a knowledgeable Scottish friend said in an e-mail to me the other day:
Dear Brian
I've just caught up with your take on the Scottish election result-on the day Mr Salmond's minority administration was sworn in.
There is a small error in your analysis of possible parliamentary permutations. You refer above to the difficulty of coalition-formation "unless we can imagine a German-style 'rainbow coalition' of Labour (46) plus the Conservatives (17) plus either the two Greens or the three independents (my italics)," I suspect that you conflate the two greens with the sole remaining independent [Margo MacDonald] to arrive at the total of three. To be totally clear, if pedantic, the final result was:-
SNP: 47 seats
Labour: 46 seats
Conservatives: 17 seats
LibDems: 16 seats
Greens: 2 seats
Ind.: 1 seat
129 seats
Brian writes: Thomas, many thanks for the correction — which is not 'pedantic' at all, but which indicates that the difficulty of assembling an overall majority for anything is going to be even more difficult than I had thought. My recollection is that I originally took the figures from a table on the Scottish parliament website, now apparently replaced by another (see page 48 of this [PDF file]) which of course confirms your own figures. Or perhaps I simply copied them down wrong.