I've been listening to lots of interesting comments on the Independence Referendum debate recently and thought it was probably worth getting all my notes somewhere in a coherent form. I've shied away from developments on the substance of independence or devolution towards issues relating the process and form of the referendum itself (and the post is large enough just focusing on those questions). The following is based on an event at the Constitution Unit with Prof. Robert Hazell and Alan Trench and two panels before the Scottish Affairs Committee, one quizzing academics (for the purposes of this post, Panel A) and the other quizzing both sides of the AV campaign (Panel B).
Professor Vernon Bogdanor, Research Professor at King’s College London
Professor John Curtice, Professor of Politics University of Strathclyde
Peter Kellner, President, YouGov
Professor Iain McLean, Official Fellow in Politics, Nuffield College, University of Oxford
Matthew Elliott - Campaign Director of NOTOAV
Katie Ghose - Chair of Yes To Fairer Votes, Chief Executive of Electoral Reform Society (ERS)
William Norton - Responsible Person for NOTOAV and referendum agent for No campaign for NE Referendum 2004
Willie Sullivan - Head of Field Operations for yes to fairer votes, Scottish director of Electoral Reform Society (ERS).
I've also included some brief comments from the Electoral Commission evidence, but for the most part the debate has changed enough to make much of that session irrelevant.
Role of the Electoral Commission
The general consensus from Panels A & B that the currently proposed question is unfair, but Curtice points out that Electoral Commission corrections haven't always been completely unobjectionable so wouldn't go as far as to make the commission's comments on questions binding. Along those lines Norton argued it'd be unnecessary as proceeding with a question the Electoral Commission had labelled dodgy would be a gift to the disadvantaged campaign anyway so sufficient political incentives exist without mandating use of the electoral commission formulation. Kellner made the case that the question doesn't actually make that much difference, something interesting for a pollster to say as question effects can have a large effect on how people answer. However I can see there being a case for not applying polling logic to referenda as most voters will be exposed to the campaigns long before the ballot and so question-wording effects could be far less strong than in polling. He argued what was far more important was that the question be perceived to be fair, and the Electoral Commission role in this regard is creating demonstrable justice rather than creating an unproblematic question.
There was a general feeling in panel A however that in the event of referendum where there might be other contentious structural issues other than the question (such as the form of questions and counting system to choose between three or more options) that the commission's advisory role should be expanded to include those issues. In Panel B there was opposition to the idea of the electoral commission producing more material, with both sides pointing out they objected to the neutral material during the AV campaign. Sullivan argued that it probably wasn't cost-effective and the money might be better spent on more broadcasts - which Elliot argued there should be more of, but shorter to fit in better with how people understand adverts.
Does a 'yes/no' format matter?
The former Scottish First Minister Lord McConnell has suggested that votes should select from duelling statements (i.e. something like 'Scotland should be an independent country'', 'Scotland should remain part of the UK'') rather than answering a yes/no question to deprive the SNP of the advantage of running a 'yes' campaign, but from the AV teams there wasn't an impression that this was hugely important. Elliot argued that actually yes/no isn't hugely important (or at least that slanting is variable on economic conditions and national mood rather than an inherent bias either way) and the debate can still be defined in those terms even if the question isn't, with the example of how the 1997 Scottish referendum was covered by the media as a yes/no campaign despite it a question of duelling statements. Sullivan thought that the big difference was 'change/status quo' which exists however the question is worded, and that the advantage will rest with the status quo. The take away message here being that there probably isn't a lot of point in trying to avoid a 'yes/no' question.
Should there be multiple options?
Kellner said it would be 'perverse' if some kind of devo-max wasn't an option and as Bogdanor nicely put it "it'd be very odd that an instrument designed to discover what Scottish people think should exclude the option that they most favour". Trench argued that most Scots want around 65% change which is closer to independence than not, so it could well in the unionist's interests to put forward a devolution option and that the current Unionist pursuit of a 'jam tomorrow' option where a No vote on independence would be rewarded with further devolution could well run into a credibility problem as the recommendations of Calman Commission have yet to pass into law - giving a bad record for the UK government's ability to proactively manage further devolution.
There's a related point that even if there's only one question on the ballot, there's still a decent case to be made for having multiple referendums on that single question. As it was unlikely that negotiations or consensus on indepedence or even agreement on certain flavours of devolution could be reached before the vote (although Kellner pointed out, a lot of it will become apparent within the campaign) Panel A were quite keen on the need for two referendums, one enabling negotiations and another post-negotiations binding referendum for agreement to terms and Trench similarly supports this. In Panel B, Sullivan indicated that the ERS wasn't wholly supportive of that (or at least, much less than their evidence would suggest).
Panel A raised concerns about the content of Devo-Max, mainly that it's as yet ill-defined and needs to be more so before the referendum (the arrival of devo-plus on the scene may or may not change this situation). Bogdanor put forward that if it's an option it not only needs to be defined but needs to be shown what is on offer is genuinely an option from Westminster. Noton pointed out that the North-East devolution, welsh powers and AV referendums had outcomes that were well-known and thought out before the election. He argues that it's fundamentally a unionist proposal and it's up to the unionist side to define it (which raises the question of who's actually for the status quo, but we'll get to that in a bit). McLean again argued that it isn't realistic for all the negotiations to happen ahead of time, hence the idea of a second confirming referendum.
Ghose and Sullivan agreed that a three option referendum could probably have three campaigns, which brings up the issue that the Political Parties, Elections and Referendums Act 2000 (PPERA) has an 'all or nothing' rule for having a designated lead campaign, something that could be interesting in a three-way referendum as it's not at all clear who the groups that would argue for each option would be - you'd expect a common unionist position to emerge and it's not clear if that would be devo-max or the status quo. Depending on the exact details of a devolution option, there might of course be a three-way split but in the event that this doesn't happen it's not clear if any lead campaigns are allowed (Referendum legislation could of course amend PPERA to allow a minimum of two).
How should a multi-question referendum be conducted?
In terms of precedent, the New Zealand electoral system referendum was mentioned by Kelner, Sullivan and Hazell as an example of how a multi-option referedum might be conducted. This is is one sense good precedent being a multi-option referendum from a Westminster style parliamentary democracy, but as it's the example that comes up again and again it does illustrate that this is relatively new ground and the door is open for pretty much anything to happen here and create precedent. Kellner also pointed out the 1997 Scottish devolution referendum was effectively a three option vote, so it's not completely without UK precedent, although in this case all three camps would stronger options. Similarly Norton brought up the North East 2004 referendum as an example but also pointed out as the second question was a subsidiary question, all campaigning focused on the first question.
The SNP's double question (the problems with which I went into here) seems to have mostly gone away at this point and Panel A were divided on the best way to go foward. Kellner has previously written something helpful about the five possible ways of doing thisand he and McLean both favoured a ballot that allowed people to rank '1' and '2' which is then counted by the Condorcet method. In the unlikely event of a Condorcet loop, McLean would want something in the referendum legislation to deal with it and suggests defaulting to the status quo. Bogdanor suggested a three option ballot with a later run-off election and Curtice a gateway question approach, with the first question giving a choice for independence and if that fails, question two giving the choice between status quo and devo max. The ERS favours a different gateway question, with the first being a question about change in principle, the second then being a choice between the two change options.
As Alan Reid called Sullivan on, the problem with the ERS question is it only allows the status quo to a voter's first or last preference and 'Devo Max, status quo, independence' is a perfectly likely ordering (and such a voter could end up empowering their least favoured option by voting in accordance with their most) - a similar problem holds for Curtice (it's the flaw of the gateway question in principle) but in this case it's independence that can only be placed first or last, something that's possibly more in key with actual preferences. McLean and Kellner were opposed to runoffs and AV on the grounds that an option like Devo Max could be the Condorcet winner and the consensus choice but if it came third in the first round it'd be removed. Panel B very briefly ventured onto this territory giving us the memorable moment of a former member of the Yes to Fairer Votes campaign saying 'it wouldn't work' but alas they didn't go further into it. Trench questioned if you could get a mandate from preferential voting, but this would presumably be less of an issue with a run-off referendum.
Bogdanor argues that the referendum should happen sooner rather than later as the issues aren't that complicated and the continued uncertainty could be causing economic damage (a point he'd repeat from the audience at the Con Unit). McLean argued in response that the biggest economic issue is currency for Scotland's large financial industry - and that this is an issue that could be addressed (and there is incentive for both sides to do so) well before the referendum, making timing less of an issue.
Hazell disagreed with Bogdanor that economic grounds were enough to justify Westminster interference in picking a date, and argued that there are two legitimate concerns for the UK government regarding the time - avoiding a day that clashes with another election, and also one that would come close to the next general election as this would create constitutional awkwardness in the event of a vote for independence if a general election occurred between the yes vote and independence - Scottish MPs would be returned and potentially help form a government before disappearing again.
This suggests that a date should be picked with regard to being able to wrap up potential negotiations before May 2015 - however the assumption here is that the fixed parliament will last that long - over at the LSE Politics and Policy blog Patrick Dunleavy is not at all convinced that the coalition will go the distance, expecting a dissolution and Conservative and Labour agreement on an early election sometime in 2014 - in which case it's probably too late to avoid the problem above. However this exact question and an ongoing referendum through 2014 might actually either reinforce the coalition or at least create disincentives for Labour to agree to dissolve parliament (Keeping Scotland in the union is strongly in the party's interest and may deplete campaigning funds to an extent that they don't want to run a general election the same year) and hence make the coalition more likely to run it's full course - so it's possibly everything cancels out here.
On a related time issue is the length of control period and the campaign in general, which there was a general consensus in panel B that a longer period was better and that the campaign needed at least twelve months. Norton argued that 10 weeks control period simply wasn't enough time to deal with the practicalities and time taken along the critical path to the free post mailshots - in 2004 his campaign almost missed the deadline and the NotoAV campaign this time around started printing before they were officially the designated campaign. The main problem (that's probably unresolvable) with referenda as opposed to Scottish parliament elections (and now possibly Westminster) is there's no fixed time for the election to work towards and there’s an incentive to do what the SNP are doing and have a long pre-legislation period (still a year to go before that according to their timetable) where campaign actions can’t be monitored because there’s officially nothing to campaign about going on.
Bogdanor thinks the voting age should be 16, but that referenda shouldn't be special and the existing franchise should apply. Curtice argues that changing the francise would be opening a can of worms in terms of ideas about nationality and that it'd be technically difficult to get people on the register anyway. Andrew Scallan (Director of Electoral Administration at The Electoral Commission) spent a long time not wanting to give a precise answer as to how long the process of expanding the registers would take (more than week, less than a year) but did suggest the best scenario would be legislation in the summer followed by a household canvas in the Autumn, but at other times of year the process might be more difficult/result in a less complete register.
Interestingly as it's a local government francise EU citizens would be able to vote but as typically they're bared from national elections (which elections for a Scottish parliament would become in the event of an independence win) means that they could effectively vote to remove their own representation. Curtice points out however that they're unlikely to vote at all so it's a bit of a moot point.
Reduction of Westminster MPs
I've moved this section into a separate post.
Status of rest of UK
The role of the rest of the UK in this decision is variable on outcome, Mclean argues that Scotland could declare independence unilateral but could not declare devo max unilateral, as devolution would be a gift of powers from Westminster. This distinction is less than it seems though as the kind of independence that the SNP describes includes some quite complex inter-dependence that would similarly require UK consent. Trench said that a referendum on devo-max could include the rest of the UK (I think I'd be correct in describing his view as falling far shorter than 'should'), Hazell was content that it could be negotiated by the government. If the precedent in general has been for referendums to devolve powers, the precedent is also that only the affected region votes.
Will the English bomb Scottish airports?
No one brought up this pressing issue.