Proportional representation and the German elections: round three

Jarndyce hits back elegantly and concisely at those of us who have pointed out  that the German election results are a sharp lesson in the defects of proportional representation (PR) when it comes to electing a government with clear and effective policies.  He claims that “The German result is irrelevant to the PR debate here in the UK, for at least three reasons:” —

1. Look at the numbers: two blocs of 35% are [?and] three of 10%, roughly. Would any reasonable electoral system dish out a majority government based on that spread? Should it, and still be able to call itself a democracy?

This correctly asserts the desirability of proportionality in a body which has to be broadly representative of the opinions of those of the electorate who vote, i.e. one which is there principally to voice opinions on behalf of the population at large.  A good example would be a British elected second chamber.  But the House of Commons and the Bundestag are not bodies like that. Proportionality and representativeness are not necessary, or even always desirable, in a body whose main purpose is to choose and then hold to account a government that can govern stably and effectively, carrying out the policies which it has offered to the electorate, almost always based on the party which has won more votes than any other.  

Jarndyce refers with apparent approval to this from Paul Davies in Make My Vote Count:

Unlike Canadian elections, the German results actually mirror public opinion. The German electorate is genuinely split on the country’s future direction. But at least voters can begin dealing with their options in a political arena with a level playing field. It’s called democracy.

Again, this reflects the essentially confused view that an electoral college with the duty of choosing a government needs to ‘mirror public opinion’, whereas if it does so at all precisely, it will produce, as often as not, either fudge or paralysis or both.  In normal circumstances it broadly mirrors the votes cast by installing a government from the party that won more votes than any other.  Moreover it isn’t the voters at election time who can ‘begin dealing with their options’ under PR, but the politicians after the election.   The distinguished economist Anatole Kaletsky writes, in a postscript to the same piece, that –

The German election was a triumph of democracy in the same way as the referendums in France and the Netherlands. Just like those referendums, it has created a political stalemate, neutered diplomacy and paralysed the economic reform process. But political paralysis was exactly what German voters wanted — and quite rightly so.  Germans were right to vote for political paralysis for the same reason that the French and the Dutch were right to immobilise Europe…

That’s a good example of another common fallacy in the PR armoury of arguments.  Germans manifestly didn’t vote for political paralysis, any more than they will have voted for a Grand Coalition of the SDP with the CDU if that’s what emerges eventually from the smoke-filled rooms.  They voted for political parties, none of which advocated political paralysis. The overwhelming majority of German voters voted in the way they thought likeliest to produce a government formed or led by each voter’s party of choice.  If the overall result had been one that enabled his or her party of choice to put into effect precisely the policies that party had put forward in its election programme, those voting for that party would obviously have been delighted.  It’s possible that some voters are happy with the prospect of protracted coalition-making, fudge and paralysis:  but if so, they had no way of reflecting that hope in the way they voted: nor could they have done so under any system yet devised, apart from a referendum.   An economist as skilled as Mr Kaletsky ought to have been able to work that out.  Perhaps he was being ironical.

Jarndyce continues:

Anyway, I thought the complaint was that Germany’s FDP held blackmail kingmaker powers and pivoted the system around them. Now we’re complaining when they don’t.

Who are these ‘we’ who are complaining that the FDP no longer has the numbers to act as kingmaker?  I’m certainly not.  I have pointed out that in UK circumstances, the LibDems, winning fewer votes than either of the two bigger parties, would almost always be able to act as kingmakers, not only deciding for themselves, almost on a whim, whether Labour or the Tories should occupy No. 10, but virtually always claiming a place in the resultant government as junior coalition partner, as the FDP did for decades in Germany.

2. Leaving aside the odd quirk, Germany’s electoral system delivers near-perfect proportionality. 10% of the votes entitles you to 10% of Bundestag seats. The proportional part of the vote is compensatory rather than parallel, to use the jargon. But nobody serious is suggesting a perfectly proportional system for the UK. Both AV+ and the system I suggested here in June would deliver majorities on large pluralities, as would the Single Transferable Vote. The Jamaica, Grand or traffic-light games are a diverting but irrelevant sideshow for UK psephologists.

I, like many others, say that the German election results represent a dire warning of the defects and perils of PR, which the fanatics constantly demanding PR for elections to the House of Commons ought to heed.  Jarndyce, your answer to this is now to claim that ‘nobody serious’ is arguing for ‘perfect’ PR for the UK.  That’s news to me!  Anyway, I’m glad that you concede the point which some of us serious people have been making, i.e. that PR (perfect or im-) would be deeply damaging for Britain as a system for electing the House of Commons.  I’m quite happy to talk about non-proportional systems for the House of Commons such as AV, AV+ and other variants, each of which has its pros and cons, just as I would welcome PR for the House of Lords, or Senate, if (as I hope) we are ever allowed to elect it.

3. The “ooh, a coalition” bogeyman is just that — a bogeyman. The correct response is: “so what?” Germany’s main, stable conservative bloc is effectively a coalition. Whatever comes out of the bargaining, it won’t be an entirely new programme or set of policies, completely unmandated by voters, as PR opponents always suggest. Germany will get a marriage of two or three parties’ existing mandates, a consensus of support on a narrower agenda, perhaps with a specific time limit. It might wobble, it might not. We elect MPs as proxies. The German ones are just going to have to earn their money the hard way for a change.

Alas, this just won’t wash.  The policies emerging from the compromises squeezed out of the  coalition partners in the course of the horse-trading required to form it – dropping a key proposal here in exchange for agreement on a completely new one there – are unlikely to resemble anything that any single party was offering during the election.  It is absolutely right to point out that with PR – which, remember, no serious person advocates for the UK anyway – not only is the senior governing party in a coalition chosen after the election by politicians, not at the election by the electorate:  but also the coalition’s policies are likewise hammered out after the election by politicians, not selected from the several on offer at the election by the electorate.

Here is Tim Garton Ash in today’s Guardian on the prospects for Germany under whatever coalition may emerge after what may be several weeks of bargaining and horse-trading:

Any of the now possible coalitions will be alliances of chalk and cheese, if not of fire and water. They will involve extraordinarily painful compromises on policy. They will be plagued by personality clashes and parties jockeying for position in an election everyone will expect to come sooner rather than later. The results in economic and social policy – and probably in foreign policy – will be more of that soft fudge in which German attempts at reform have been suffocating for more than a decade. This will be bad for Germany, bad for Europe and bad for the world economy.

The most likely fudge-factory would be a so-called grand coalition between Social and Christian Democrats. Schröder has said he won’t serve under Merkel, nor will Merkel under Schröder, so that (unless they change their tune) a double decapitation would be needed before the grand coalition could even begin. With the parties having diametrically opposed policies in areas such as health-service reform, fudge mountains would be called for.

Bad for Germany, bad for Europe and bad for the world economy.  That seems right.  And PR would be bad for Britain.  

22 September 2005.

German elections

5 Responses

  1. Jarndyce says:

    Thanks for the response. Obviously, we’ll never agree here. Just a couple of brief replies:

    1. I accept that representativeness isn’t the only criterion for electing a legislative chamber. But there has to be a floor. For me, FPTP doesn’t pass that threshold, neither does any system that can elect a government on around a third of the vote. That’s 2:1 against whatever policies are to be imposed.

    2. Coalition does involve compromise, but when we elect MPs as proxies we give them that task across the board. (Major had to do it with the Unionists, without permission from voters. It didn’t spell the end of British democracy. In fact, politicians do things all the time that aren’t in the manifesto, and haven’t been mandated.) Anyway, rather a few weeks of compromise and bargaining than 5 years of policies opposed 2:1. And if coalition naturally smoothes out extremes, then all the better for it. Though even empirically, I’m not sure that’s a correct prediction. Denmark has largely adopted “Thatcherite”-style economic policies in the last decade despite an electoral system that is hyper-proportional.

    3. I don’t accept that anyone serious is punting perfect proportionality for the UK. STV, AV+, regional open lists: all of these deliver legislative majorities to large pluralities, and still pass the representativeness test.

    4. It’s possible that some voters are happy with the prospect of protracted coalition-making, fudge and paralysis: but if so, they had no way of reflecting that hope in the way they voted

    That’s not the case in Germany: voters have two votes and research has shown that they use them to express coalition preferences. (I can get the reference for that assertion if you wish.)

    Also worth taking into account: turnout last Sunday was 77%. It’s a long, long time since we got close to that here. And as Martin has shown here, the problem in Germany isn’t the system, but the divided polity. FPTP at this election would have delivered almost exactly the same result. (Martin has figures.)

    Still, we’ll not be reaching agreement, I guess…

  2. Brian says:


    Thanks for that. I agree that we’re not going to agree. Just one mini-reply about your argument that the second vote can be used to indicate a wish for coalition: I dealt with that in my much earlier piece about the Jenkins Commission report, pointing out that the second ‘top-up’ vote can’t be cast in a way guaranteed to achieve any specific purpose since it’s impossible to know in advance which party’s candidates will qualify for second votes: and that research shows that some people split their two votes between different parties, not because they want those two to form a coalition, but because they don’t expect their first choice party to need topping up, meaning that to cast the second vote for one’s first choice party is to waste it. In those circumstances, the second vote becomes in effect a second preference vote, as in AV. And the problem with that is that second preferences ought not to be given the same weight as first preferences, for obvious reasons. However, I accept your point that some voters may well use their two votes to reflect a hope that the two parties for which they vote will form a coalition. I’m not sure how much further that takes us, though, since all German voters know that under their PR system there will have to be a coalition government, however they cast their votes. To express a preference for one possible coalition over another is not the same thing as voting in favour of coalition government as a good in itself.

    But there’s enough substance (and enough misguided argument) on both sides of this argument to keep us busy for years!


  3. Brian says:

    This is a comment which I have been unable, for some reason, to post on Martin Stabe’s post about Germany and PR because of formatting problems:


    You say: >> I can’t understand why some people like, for example, Brian Barder or Tom Watson MP, think an electoral system that forces ‘a clear and decisive result’ is preferable to one that leads to a weak executive forced to compromise. The latter, it seems to me, is a more accurate reflection of the will of the electorate when they return results like those in Germany – and Britain.

    But the fact that under PR (and much more rarely under FPTP) an almost evenly divided electorate votes in a way that produces ‘a weak executive forced to compromise’ doesn’t mean that that outcome was desired by any single voter. When political opinions diverge widely, it’s meaningless to talk about ‘the will of the electorate’. The German and UK election results are perfectly consistent with the vast majority of voters ‘ i.e. the 60 to 70 percent of them who voted for one or other of the two biggest parties ‘all having wanted their chosen party to win a safe majority in ‘a clear and decisive outcome’ so that the country might be governed stably and with a clear sense of direction.

    You say: >> Both Tony Blair and Angela Merkel lead parties that have the support of a minority: Labour’s 35.3 per cent is nearly the same as 35.2 per cent the CDU/CSU took in Germany. Yet the almost identical minority result guarantees Tony Blair a clear Parliamentary majority, while Angela Merkel won’t be able to form a government unless she is willing to water down her policies until they are acceptable to enough parties to reflect a majority of the electorate.

    This seems to brush aside the fact that since 1935 no British political party has ever won half or more of the total votes cast. Every government since 1935 has been in the position that more voters voted against it (or at any rate voted for a different party) than for it. Under the German system of PR every single British government would have been forced to form a coalition (formally or informally) in order to be able to govern. This is a recipe for fudged policies for which no-one voted at the election, for messy compromise, for instability (it’s always open to the smaller party in the coalition to throw its weight behind another larger party and thus precipitate a change of government without even an election to endorse or justify it, as has happened in Germany) and for unpredictability. No business or finance company can safely plan ahead in such circumstances. The likeliest outcome is virtual paralysis.

    You write: >> In an entirely FPTP German election the 2 October by-election would become a circus like no other: the balance of power in Berlin would depend on the outcome. If, in this bizzaro-world, the CDU won Dresden, Merkel would form a government with a one-seat majority. If, however, the SPD or Linkspartei took the seat, there would again be a hung parliament with all the same arguments as exist in real-world Germany today. A minority government would be the likely outcome. New elections would only be a matter of time. Proportional representation did not cause the current situation. The voters did. Democracy is a bitch, huh?

    All you are saying here is that because the two main parties have received such a similar level of support, with the third party not all that far behind, the result would have been a hung parliament and the need for coalition-mongering after the election whether the election had been held under PR or FPTP. But that proves nothing, except the obvious point that FPTP can in exceptional circumstances produce a hung parliament. We knew that already. We also knew, didn’t we?, that on all the copious evidence, even when the main parties are very close in their level of support, FPTP very rarely indeed produces a hung parliament or the risk of fudge and paralysis that goes with it, whereas PR is guaranteed to produce that outcome every single time. And, as noted earlier, it’s quite wrong to say that the voters caused the current situation in Germany: the system was absolutely guaranteed to produce it, whereas the main alternative system available would probably have produced it too on this occasion, but very rarely on others, even if the voters behaved in much the same way.

    There are useful discussions of these issues at (for example) , including the comments, following on from the post of Jarndyce’s which you cite. For an especially interesting exchange, see

  4. Neil Harding says:

    Brian, you say that FPTP only produces coalition governments in ‘exceptional’ circumstances. With the rise of the Lib Dems to 62 seats, the chances of coalition govts have risen 10 fold. If this continues (and change is slow and steady under FPTP), we could get hung parliaments over and over for a long time.

    You also astoundingly say that PR gives weak fudged govts. This is the same PR that has given Germany post war economic growth, a health service, transport infrastructure and social welfare to put ours to shame. In fact all the long term decisions on economic management and welfare get taken far more efficiently than us with our stop-go electoral system stifling debate and progress. I could go on…

  1. 23 September, 2005

    Germany, PR, parliaments, and presidents

Leave a Reply

Your email address will not be published. Required fields are marked *