A reply to James Kelly (Scot goes Pop)

James Kelly continues his #bothvotesSNP contention in a blog entitled “It doesn’t matter whether you’re an optimist or a pessimist about the SNP’s chances : “tactical voting on the list” is a bad idea either way“. James is entitled to his point of view, but having excoriated (rightly) the Sunday Herald for misrepresenting the views of Professor Curtice, I think at times in this blog he arguably does the same thing, though in different ways.

First of all, he contends that Curtice has “made clear” that “there is a chance the election result may differ somewhat from current opinion polls.” And so Curtice would. Indeed, he is quite transparent how he has come up with the numbers in his paper (see page 9, http://electoral-reform.org.uk/sites/default/files/files/publication/The-2016-Scottish-Election-Briefing.pdf) where he sets out the variety of assumptions that have been made to develop his forecast. As with any forecast, if the assumptions are wrong then the likelihood is that the outcome will be wrong as well. Indeed, opinion polls are usually only accurate to within a margin of +/- 3% (e.g. a forecast of 50% is usually 95% likely to be somewhere between 47 and 53%). If the SNP were within 3% of their Unionist rivals – or even something like this – then the problem of error would be acute. But the fact is that poll after poll tends to put the forecast vote for the SNP at often more than Labour + Conservative and Lib Dem votes put together. There is such a quantum difference between the parties that such considerations are substantially niceties.

But more seriously, while Curtice would be, I am sure, among the first to acknowledge margin of error as an inevitable feature of opinion polls and political forecasting, he has hardly said that the numbers on which he based his paper were thought up one night after work in the pub with a few of the boys. In other words, while Curtice did acknowledge that the result may differ from current opinion polls, his use of the word “somewhat” would suggest a degree of caution in that caveat which I do not think Kelly follows.

Secondly he argues “Crucially, the direction in which the polls lead us astray doesn’t really matter – if the SNP are being overestimated by the current polls, they’ll need list votes and list seats simply to retain their majority, but if they’re being underestimated on the list vote, they stand to win a decent number of list seats even if they take a clean sweep of constituency seats.  Both of those possibilities are very real, but if you were a gambling man/woman, you’d probably be betting more on the latter, simply on the basis of past history.”

Well actually though, it does matter. If the SNP are overestimated, then the problem of winning no list seats that Curtice describes becomes less acute to the degree they have been overestimated (on the constituency vote). The specific problem that Curtice (the “Curtice problem”) refers to is based on an analysis of Lothian region where the SNP win every constituency, but no regional list seats because they begin with their regional list vote being divided by 9 (constituency seats) +1 – i.e. 10.

In fact, the reduction in constituency seats won would have to be substantial to significantly change the regional list situation the SNP face. For instance, Curtice works through the consequences of them winning every constituency in Lothian (which is what his forecast suggests), which means their regional list vote is divided by ten, reducing it from 41% to 4.1%. But suppose they win “only” seven of the nine, then their vote would be divided “only” by eight (7+1), which would reduce their regional list vote to 5.12%, which would win them no Regional List seats either. In fact, with a regional list vote of 41% they would have to win no more than four seats to win any Regional List seats (4+1, so meaning their regional list vote would be divided by 5, which would be 8.2%, so all other things equal on Curtice’s figures, they would win the 4th of the 7 seats, at the expense of the Labour Party. But more than four seats would mean they would win no Regional List seats with a vote of 41%. Of course cet.par. would not apply in reality as if the SNP only won 4 seats, other parties would between them win 5. However, set against that, interestingly in 2011 the SNP won 8 of the 9 constituency seats, and with 39% of the vote won absolutely no regional list seats at all in Lothian (though Margo did secure a regional list seat).

Of course then the argument might be that those SNP voters who have been seduced by the siren voices of the smaller parties would have been better to vote SNP on the regional list. Let’s assume then that the entire Green vote goes to the SNP (which is kind of unlikely, but let’s take the limit case). This would give the SNP a Regional List vote of 52.8% (41.2 + 11.6), but with 9 constituencies in the bag, they still would not win a single regional list seat. In fact, the independence movement might well be one seat down, for the Greens would not have a member elected at stage 3. Instead that would go to the Labour Party. The SNP might win a seat on the last, seventh round. The SNP (with Green ) vote would be reduced to 5.28%, but the Conservatives would be running them close with 5.2%, so it would depend on very few votes. In any event we are well in to margin of error, and the point to note is that maximising the SNP regional vote is not guaranteed to achieve very much at all. From the wider point of view of achieving independence, there may be better uses for it.

Underestimation of the SNP vote is easier to deal with – quite simply it makes the “Curtice problem” all the more acute. In his paper, there are at least two regions where the SNP do not win every seat and they win two regional list seats. The extra seats could be at the expense of these two regional list seats (not entirely a bad swap though!). Moreover, without a very substantial increase in their regional list vote, if the SNP won every constituency seat, they are unlikely to win “a decent number” of regional list seats as Kelly blithely tells us. Indeed if Curtice is right, they are unlikely to win any.

Moreover, what is meant by “past history”? Is it 2007 (when 26 of the 47 the seats they won were regional list), or, to much a lesser extent, 2011 (when only 16 of the 69 seats were won on the regional list)? Or is it May last year when the SNP won 56 of 59 Westminster seats, which proportionately to Holyrood’s 73 constituency seats, would see them win 69.3% – or more reasonably 69 or 70 (coincidentally the latter being what Curtice forecasts).

Thirdly, we are told there is “an implausibly wide gap between the SNP’s support” between constituency votes and regional list votes, as well as the forecasting error of the SNP Regional List vote in 2007 and 2011, leading Kelly to argue “So in trying to interpret what the polls are really telling us about the current state of play on the list, there are two basic options – either we can assume that the polls are broadly right about the constituency vote but are underestimating the SNP on the list, or that the polls may be somewhat wrong about both ballots”.

I think I have been clear about the latter contention, but with regard to the former, this is quite a different election from 2007 and 2011. The former saw the SNP squeak in to minority government by a single seat, and in 2011 left us all – and Labour in particular – quite gobsmacked by winning the “impossible” overall majority. The present election is coming a year after the SNP came within less than 5,000 votes of taking every one of the 59 Westminster seats in Scotland, and after a year of polls that vary really only in respect of how far they suggest the SNP are ahead, or, if you want to look at it another way, how far behind the others are.

In turn this has led an argument (exemplified by Curtice), which I know Kelly rejects, that SNP voters should think about how they use their regional list vote. If the SNP were to achieve the level of support in constituencies that has been forecast, then their chances of achieving much on the regional list are limited. Maybe, just maybe, a minority element of the SNP support has listened to that, and decided to vote for someone else with their regional list vote? I wonder why this was not considered on the list of explanations? Seems plausible to me.

Kelly then goes on to argue, “If we can’t entirely trust what the polls are saying on either ballot, then all of the incredibly precise “predictions” (ie. projections) of seat numbers that are being used to make the case for tactical voting are completely meaningless”. Well that is more than a bit of an exaggeration, is it not? It’s not that they would just have to be wrong (and in a very precise sense they probably will be – they are an indicator and little more), they would have to be totally and utterly wrong to make much difference to the SNP’s difficulties on the regional list. Curtice suggests that they would win every seat in Lothian, but as I pointed out already their constituency seats would have to be reduced not by one or two, but by about half before they would win a regional list seat on the basis of the regional list share of votes that Curtice works from. I would suggest that it is not so much a matter of tactical voting being meaningless, but that the polls would have to be meaningless before the loss of SNP votes would have substantial effect.

I am in any sense rather puzzled why someone like James Kelly, who has built a very successful and well-respected blog on the basis of interpreting opinion polls and trends in these, is casting such doubt on a forecast by a respected psephologist. Or is it just the question that he poses – “That this situation could arise in a number of regions, given the SNP’s current standing in the polls, has led to speculation that nationalist supporters might be wise on the second ballot to vote tactically for a different party, such as the Greens or the left-wing RISE grouping, both of which also support independence. That way their vote might contribute to the election of another independence supporting MSP rather than apparently being wasted.” Kelly rightly excoriated the Sunday Herald for reading much more into this than Curtice would have intended, but is Kelly himself not trying to undermine the point by challenging the data at its foundation? Just why would he do this?

It is fairly well-known that, in addition to having a system with some proportionately, one of the aims of the voting system we have for Holyrood, was to make it difficult – perhaps it was seen as impossible – for a single party to achieve a majority on its own – even Labour. However, the SNP have thoroughly disproved that. In short, the problem of how SNP voters should use their second vote is quite a deliberate outcome for the electoral system in Scotland. The aim always was that when a party did well in constituencies that they should be disadvantaged – or looked at the other way, the parties which had done less well, should be advantaged – in the allocation of regional list seats. This is not an accident, but something designed into the system.

The use of Curtice’s work by the Sunday Herald was wrong – and Kelly is to be congratulated for his earlier blog pointing that out – but the issue for SNP voters that Curtice points to is not a matter of chance, but a function of how the system works (deliberately? For another day). It is wrong that it has been ambushed by the Sunday Herald, and Cat Boyd’s piece in the National (http://www.thenational.scot/comment/cat-boyd-split-your-two-votes-to-revive-the-spirit-of-2014.16491) is no more than I would expect from a politician playing politics (however elegantly she might put it). But Nicola Sturgeon is doing basically the same thing when she says the election “is not ….. a game of chance with the electoral system – it is about choosing a government and a First Minister to lead the country forward for the next five years and into a new decade.” ((http://www.heraldscotland.com/news/14438134.Sturgeon_warns_SNP_voters_not_to_play__quot_game_of_chance_quot__with_electoral_system/?action=success#comment_16048992).

Yet if we follow the #bothvotesSNP recommendation then, given the difficulties the SNP will have with winning regional list seats, then the advantage goes to the Unionist parties who polls suggest will have been almost wiped out at constituency level. That is how the system works. But, in the case of the Lothian region, if a small number of SNP voters (less than 1% in fact) transferred their regional list vote to the Greens this could see Andy Wightman rather than Sarah Boyack elected. I think that would be a good thing. Would you not agree?

Dogs bark, and political parties seek to maximise their vote, but do we not have to ask whether the aim of an electorate is only to support a political party or support a cause, in this case independence? Are we best served by maximising the vote for the SNP on both votes, even if it means allowing Unionist politicians (e.g. Sarah Boyack) rather than an independence supporting candidate (e.g. Andy Wightman) to be elected? Sure there are risks. As Benjamin Franklin wrote, “the only certainties in life are death and taxes” (and maybe not even the latter these days if you have enough money). But maybe, to maximise our impact, the independence supporting electorate will have “to live a little dangerously” as Sir Charles Gray put it, in the immediate aftermath of the 1992 General Election.

One thing is certain, if the polls are accurate – within margins of error – then if a portion of the SNP vote does not transfer to other parties for the regional list, then we will have missed a chance to minimise the Unionist involvement at Holyrood. Ideally there would be a single beneficiary, but that chance is long gone (something to talk about for next time?), but even if spread over three parties, it could be the difference between 8 Green MSPs – Curtice’s projection – and 16 (one more in 8 of 9 regions) which would mean more than 20% fewer Unionist MSPs. Is a Holyrood with 86 (70 SNP + 16 Greens) independence supporting MSP not better than 78? Is 66% of that loaf not better than 60% of it? None of this will happen without appropriate advice, and #bothvotesSNP, since it will allow more Unionists to be elected, is not appropriate advice in my opinion for the reasons set out here. Scotland goes Pop is both helpful and though provoking, but in this instance, I think James Kelly is wrong.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s