We hear many supporters of PR (whichever flavour of the month) complaining that FTTP is "unfair", and "wastes votes", and claiming proportional representation is the "fairest" system. I am challenging this claim and arguing that PR is far worse than other systems.
1) Starting with a fact. All parties are actually coalitions. Heseltine did not agree with all of Thatcher's policies nor did Benn agree with all of Wilson's policies.
2) Another fact. When a country implements the "purest" forms of PR, using a single nationwide constituency (The Netherlands, Israel), MPs will be elected with under 1% of the vote in the country, and less that 5% of the votes from any part of the country.
3) A working hypothesis. In a two-party system, ignoring any effects of gerrymandering (e.g. population decline means that Labour seats tend to have smaller electorates), the voting system has little effect on the election result. Even the US presidential system, which has no pretensions to fairness, usually delivers the most popular candidate.
4) An empirical observation. Where coalitions are declared before the election (as in Italy and Australia), then for electoral purposes at the time of the election, they are behaving as a single party – it may be a makeshift party (Italy) or a permanent coalition (Australia) but the result of pre-election coalitions can be seen as a two-party election. In fact, since they are campaigning on a common manifesto, the existence of the two parties, rather than a single party, is expressly aimed at exploiting faults in the electoral system and to get the coalition additional votes by misleading the voters on the policies that the coalition will implement. Example: Australia's National (country) & Liberal (town) coalition is essentially the same as the UK's Tories choosing different 'flavours' of candidates for shire and city constituencies - why are they not a single party? There must be an advantage to compensate for the costs of running two party organizations, and this must be in the form of additional votes.
Applying the working hypothesis (3) to the empirical observation (4), I conclude that pre-election coalitions result in a system where "proportional representation" adds no true value, but peversely results in a system where some voters are more equal than others.
5) A working hypothesis. PR results in coalitions between parties. This is not an absolute truth. There have been a few times when the Swedish Social Democrats have managed an absolute majority, but even in Sweden a coalition is the most probable outcome (if only between the Social Democrats and the Communists)
Reasoning from the above, the distinguishing feature of any form of proportional representation is the Post-election coalition. This coalition can comprise as few as two parties (as is the case in Germany, where the system is designed to be flawed in favour of large parties) or it may need as many as five parties (as is the case in the current Belgian government, where separate parties for linguistic groups result in a coalition of three Flemish-speaking parties and two French-speaking parties) but the defining characteristic in practice, and the fundamental flaw of PR, is the post-election coalition.
Moving on to a second issue, what do people vote for (or against)?
The simplistic answer is "parties". But what does this mean? At an election, a party offers the voter a set of polices, and a general "philosophy" that will guide their decisions on issues not addressed in the election. No-one votes for Labour or the Tories without an opinion that they generally believe in the policies & philosophy. (Or, more cynically, that their policies & philosophy are less bad than the alternatives).
However, few people agree with every part of a party's policies. Example: I may agree with the Tories & LibDems that ID cards must be scrapped, but I may disagree with Tories about local income tax. Regardless of the Parties' stand on local income tax, I am far more concerned with avoiding a 1984-like command state. So, following a simplistic model, I use a weighted count on each Party's policies and make a choice. In a PR world, this would result in a weighting LibDem 5, Tory 4, Labour 0, and voting for them 1, 2, 3. The alternative, in a FTTP world, is to say which parties are likely to be competitive, and choosing the best alternative between them. In my constituency, that is likely to result in a vote for the Tories.
Unfortunately, after the election, comes the coalition, and then your vote is stolen! The coalition talks are hidden away from the public. The public cannot affect these talks. Being excessively cynical, one could say that after an election, only 2 votes (Germany) or 5 votes (Belgium) need to be counted. Even taking an optimistic view, we are talking about counting (at most) about 400 votes, if every MP in the coalition parties was allowed to influence the outcome. (A most unlikely assumption!) Is this fair? Has my vote been wasted?
Continuing my example above, Clegg could do a deal with Brown, accept ID cards in exchange for some other policy he wants, and sell me, and any other supporters of freedom down the river. Even worse, we wouldn't even know whether the LibDems were ever telling the truth when they claimed they were opposed to ID cards. Clegg could have done an under the table deal with Brown well before the election for a post-election coalition, and never intended to oppose ID cards, but just put a "paper" opposition in their manifesto, in order to skim some anti-ID cards votes away from the Tories. How would the public know?
Not a single voter will have voted for the set of policies implemented by a post-election coalition. It is even possible (though improbable) that not a single voter would have voted for the coalition policies, if these had been known before the election. There is zero evidence to support any assertion that if coalition polices had been in the manifesto of each member their voters would have supported them. In particular, noting that when politicians vote on a "non-party" basis, they tend to support policies that the public opposes (e.g. their own salaries & expenses; stopping capital punishment), I would suspect that they would choose policies that the supporters of each party opposes. In Belgium, blaming unpopular policies on the coalition partner is the same as the UK blaming unpopular policies on the EU.
The effects of coalition policy get even more bizarre as you approach "perfect" PR. In Israel, a one-man party, whose sole representative was an extremist rabbi, Meir Kahane, held the balance of power between two possible coalitions. His major policy was the forcible deportation of every Arab in Israel and the occupied territories to Jordan and Egypt. Israel pulled back from the brink, but they were right on the edge of the chasm. This would not have been the policy of the majority of Israelis. It wouldn't even have been a case of the tail wagging the dog. It would have been the tip of the tail wagging the dog!
Contrast this to FTTP. In FTTP, each party's policies were stated, are not fudged by coalition deals, and voters can hold individual MPs to account. The policies, as a set, have to be crafted to be consistent, and to attract the greatest number of voters. Voters for the governing party actually voted for the policies the government promised to implement, which is untrue for any coalition government. Every MP who abstained or voted against the EU referendum has been caught holding their electorate in contempt, and may be called to account.
My conclusion from the above analysis is that PR provides a false illusion of "improved" representation, which does not exist when you analyse the consequences of the proposal. This illusion is, in fact, used by a political elite to avoid having to meet the voters' demands. FTTP, though not perfect, in practise gives the electorate more influence over policies.
Of course, both FTTP and PR suffer from a common underlying flaw. They are both "representative" systems, where the voter is expected to hand over, if not a blank cheque, then a check with plenty of gaps for the MP to abuse. I believe that "representative" democracy is a defective, time-limited, system, forced on people by a consequence of the number of voters, and the time/distance from the voter to the seat of government. This limitation no longer holds, as Switzerland and California demonstrate through the use of referenda. We could move to direct democracy, though the political classes may resist. However, in the past the political classes resisted a wider franchise, and were forced to surrender.
The Shipwright - 'This ship is sound. It is a miracle of modern engineering. It is, to all intents and purposes, unsinkable'.