Thursday, 2 December 2010

Open Primaries

Open primaries are a really great idea. The problem with party politics is that most people end up voting for the semi-conscious hulk of flesh wearing the right coloured rosette rather than an actual competent human being. The idea of an open primary is to have an election within the party so that the electorate can choose the best candidate for the party.

The big problem with PR, one of the main arguments against it in fact, is that exactly this lack of a constituency MP and any personal responsibility that makes the people elected personally unaccountable. This is a real danger, the European elections are a perfect example, the parties choose a list of candidates and voters choose their preferred parties. Nowhere does any member of the general population get any say in who shall be their MEP. This is great for giving sinecures to loyal party members, or for getting weird and embarrassing people far away from Westminster.

This is, from the point of view of accountable and democratic government, obviously not satisfactory. It's the victory of the party over the people.

A solution to this is an open primary. That is, before the main election there are open, public elections to decide who will be the candidate from each party. The most well known example of this is in the US presidential election. For months before the final presidential elections there are votes for who will be the candidate from the two main parties. This way the people dont simply have to choose between two possible candidates, the final two are chosen by a wide electorate to then be put to the whole population.

The UK

Ok, enough of theory. Practice is needed. We have 3 real groups of legislators to select (I dont care here (as always) about sub-national and local elections) ie, the Lords, the commons, and the MEPs.

The Lords

We are promised in the holy gospel according to Clegg that the government will present:

proposals for a wholly or mainly elected upper chamber on the basis of Proportional representation.

It is likely that this will advocate single long terms of office.

Ok. So we're picking people on the basis of PR, serving single long terms. This is not, and is not capable of being, a recipe for personal accountability. Single long terms means that after a Lord is elected they will have almost no incentive to do anything (short of recall powers). With this kind of a disadvantage for good governance right at the start we need something to redress the balance.

To introduce the system currently used to elect MEPs for the Lords would be an unambiguous disgrace. There is no accountability at any stage for MEPs. If we are to have either a national or (more likely) a regional PR system then some kind of primary election is an absolute must. There must be some way to say who the representative from each party will be. Now we could argue till we're blue in the face about details, but the gist has to be something like this:

An election campaign over the whole constituency (probably the regions as with European elections, but some other convenient sub-division of people works just as well) where each candidate from each party can explain why they should be chosen above everyone else in their party.

An election, within that area, with a wide electorate. I dont mind if that means you have a fully open primary (ie everyone regardless of party preference can vote), a restricted vote (ie you can vote for candidates from only one party, or a properly closed primary (ie you must be a fully paid up member of a party to vote). But the election must be held with all the normal precautions (and sadly until there's been dramatic reform that means you cant have a postal vote - I'm sorry, you may as well not bother our postal voting system is so hopelessly corrupt) and must output an ordered list of people.

The winners of these elections (however they are decided) determine who gets the seats if that party wins n seats. This way the number from each party is decided by PR and is properly fair and representative, and the candidates have at least some personal credibility.

The MEPs

Exactly as above.

The Commons

There is a huge problem with running this kind of thing in the commons. Which is that UK constituencies are tiny. They cant support a large party membership in most constituencies. The first and so far only primary election in the UK for an MP cost £38,000 for one party for one constituency. Very very few areas have the capacity for local parties to run and pay for such things, and it seems hard to imagine that much support for the cost to be supported by government. As a very rough estimate suppose that each of the (say 600) constituencies to have such an election for say 5 parties, gets as a very rough figure £100 million. Obviously if this is institutionalised it can be done far more efficiently, but say we cut this down to the £80 million of a general election this still doubles costs which might be difficult.

There is also the practical problem that it's hard to find enough candidates to make it worth while in many areas. I imagine if you put forward all 5 members of the Glasgow conservative party (I jest I jest) you wouldn't have enough of a contest to make it worth while.

I would love to see open primaries wherever they can happen easily. The problem is that many people in the Tory and Labour parties see this as the be all and end all of political reform. Simple they say, have open primaries and you can keep FPTP.

Umm, not really, they are way more expensive than they need to be, and wont be very effective considering the small number of quality applicants in many areas. (Dont get me wrong there are many local parties with many important and viable candidates, and I would love to see open primaries for them, but this cant be universal). I would like to suggest that in many area the intra-party competition will only ever be of limited importance and the inter-party competition is key.

And this is exactly the area (hello high horse) where AV can help. AV can allow you to kill two birds with one stone. The key areas of real political turmoil and controversy where our current system are largely seen to fail are in two parts with, it seems to me, one cause.

All women shortlists

Ok, not just them, but they are a good example, what I mean here is central party choosing one candidate who is not wanted locally. This means that either the local party voters must vote for a candidate they dont want or (as with the Blaenau Gwent 2005 election) put up an independent alternative candidate. This has obvious dangers of vote splitting. Obviously Blaenau Gwent was a safe labour seat, but if the same thing happened in an even slightly marginal seat the two labour candidates would half the number of votes they got. This vote splitting could have been disastrous for Labour meaning they would have been almost forced to stick with the unwanted candidate Maggie Jones. This is undesirable.

Tactical voting

A lot is talked about this thing called tactical voting, I want to explain what I mean when I say tactical voting, why it's bad.

Suppose for simplicity we consider a seat with 3 major parties. Everyone has a favourite and a second favourite party and it is known by opinion polls and history that one of two parties will win. Voters of the third party can then either vote genuinely for their own party or vote for their second choice. The question for them is which way will (on the balance of probabilities) give them the best expected outcome. I.e., does a vote for their third party candidate make it sufficiently more likely that the third party will win to make it a better decision than a vote for the second preference making it more likely that they will win.

This is a choice that often comes down on the side of voting for the second party. This means that many people vote for their second choice, which genuinely distorts the true result. There was a poll shortly before the last election (anomalous time I know) asking "If you believed the Lib Dems had a genuine chance of victory how would you vote?" The polls suggested the Lib Dems winning a landslide majority, in fact, they lost seats. And this isn't just a Lib Dem point, I dont consider myself loyal to the Lib Dems, there are countless seats where the third party in elections is often just as popular as the two top parties but is nowhere near them in votes cast. Examples of this can be found for all parties.

Splitting costs and AV

Both of these questions boil down to one thing. Vote splitting. In first past the post if you run against an identical candidate you get half the vote, so it is very hard to get a diverse collection of politically similar candidates to stand as they fear eroding the support of people similar to them. This is a problem as it reduces voter choice and leads to poor candidates and poor representation of parties similar to existing ones. The bigger labour is the harder it is for the Greens or other "left" parties to gain power, the bigger the Tories are the harder it is for UKIP or other "right" parties.

The answer, as I see it, to both of these failures can be provided by AV.

In the first case the independent Peter Law could quite happily run along side Maggie Jones, complimenting her, explaining why she was a great choice and why people should put her second, but ultimately why he was the better candidate.

In the second case everyone can honestly vote for people in the order they want them. That way you get both advantages of the two way decision above, you increase the chance your candidate will win and increase the chance that your second choice will win.

This way the big two problems are solved in one fell swoop.


Is it not contradictory of me to say that Open primaries are awesome and the way to go for MEPs and Lords elections but that AV is better for MPs? No, and I'd like to suggest why.

The difference between full national or regional PR systems and electing MPs is the size of the constituency and thus the number of people involved.

The problems with FPTP in the commons are significant, but this is spread over a huge number of constituencies, so that the problems in each constituency are small, the second best candidate winning not the first by some measure, or a party getting a thousand fewer votes than they should etc. This kind of problem is not really well dealt with by an open primary, they are hard to run.

The problems with PR are significant in each constituency. And the AV equivalent, STV, would be far too hard to be practical with potentially dozens and dozens of candidates. So here open primaries are vital, this may be just about the only contact that most people will ever get with their representative.


  1. PR takes many different forms. The sort advocated by the Electoral Reform Society is the Single Transferable Vote. This vote is used to elect several MP's for each constituency, which was still the case in the UK until 1951 for some constituencies. Candidates are voted for by name so if the constituency returns 4 MP's, party 'A' can put up 4 candidates, just as they would for a single seat primary and the voters can rank them in order. Each party contesting the constituency will, or should, field the same number of candidates so voters are presented with a real choice.

  2. A quite correct explanation. STV over areas about 4 or 5 times the size of a current constituency clearly fall about half way between the regional PR of European elections and the current commons system. So it's perhaps slack of me to ignore that possibility.

    Clearly you cant quite assume that this will be too small to make open primaries pointless, or that the STV election paper would be too large to manageable. (I'm not sure it's impossible for people to handle say 5 names from 4 parties). In this case I think that both can have their place. If you're in an area where one party will win say 4 or 5 of a 5 member constituency, or even 3 or 4 of a 4 member one then I feel it's probably for the best that those 3 or 4 people be selected by primaries beforehand, but if each party wins only 1 or 2 members say this is less important.

    In short, yes STV is a valid electoral system, and one I shouldn't neglect. It would work, and would only sometimes need primaries.

    I dont think this changes my analysis of single member constituencies or of regional or national sized constituencies.

  3. No it doesn't change your analysis of single member constituencies at all.

    My concern with open primaries is how they would be legislated for and who would oversee them?

    To be truly open, anyone should be able to stand for any party and anyone should be able to vote in any party's primary.

    A lot of problems here.

  4. I agree that you need to draft relevant legislations/codes of practice very carefully. I've not gone into detail because I've not really got the expertise to say much and haven't thought enough about it. But I dont think there's anything that a few smart lawyers couldn't work out given a month or two.

  5. As noted above there are several types of PR which still involve constituencies which you haven't discussed.

    More importantly though, you seem to think that parties are somehow irrelevant or obstructive in politics. Do you think that someone should be able to put themselves forward for election without having the support of a grass roots movement which will campaign for them? If you would make candidate selection entirely open then how would you propose to get more people engaged and active in politics? How would you encourage more membership of parties if you're taking away the one thing that members can still really decide in their parties, ie select candidates? Would you like to see a political system where no-one really bothers campaigning, people just vote in open selections from their armchairs, and entirely independent candidates fund their campaigns by courting big businesses rather than from a movement of activists?

    Personally, I'd like to see a kind of politics where people will get involved in a movement, campaign for it, support it and feel strongly about it, and where candidates are produced by those movements. Ideally, with the most grass roots support from politically engaged people in the local community should win. Otherwise we will end up with a system where ordinary people don't get involved beyond voting for someone, disinterestedly, without knowing or caring very much about the whole thing, and where campaigning is about billboards, TV ads and automated phone calls rather than community activism.

    Political parties are important.

    This was a bit of an unstructured rant but hope I got the points across!

  6. After consideration I don't like open primaries at all - they are an expensive 'sledge hammer to miss the nut'.

    With AV you can have 'portfolio candidates' with bespoke agendas (probably based on a party agenda) instead of having to put up with slightly adjusted 'off the peg' party manifestos.

    MPs should be in-hoc to their electorate, not an unaccountable party leader.

    (mis-edited previous post..)


Feedback always welcome.