Peter Kellner of the respected polling company YouGov has a persuasive analysis here of how Scottish votes could deprive the Labour party of an overall U.K. majority in next year’s general election. It’s based on the undoubted fact that because the next election will include seven parties that have the prospect of winning at least one parliamentary seat, the old concept of a “uniform swing” across the entire country won’t apply. There will be different swings between different parties in different constituencies. (Northern Ireland is excluded from these calculations.) In particular Scotland will have a very different political outcome from England and Wales because of the rise of the Scottish National party’s threatening Labour’s tally of 41 seats out of the country’s 49.
First, however, here is YouGov’s projection of the election result — a narrow Labour victory of one — if the old uniform swing were still applicable across the whole of Britain:
Kellner then breaks the U.K. (or rather GB for Great Britain) into two separate electoral battlegrounds, namely (1) England and Wales and (2) Scotland. That is reasonable because the SNP scarcely registers in the rest of Britain where it doesn’t even field candidates, but it is currently scoring about 43 percent of the Scottish vote in polls. On that basis he calculates how many seats a “swing” in each battleground would produce and then, in order to get the final GB total for all parties, he adds up the two sets of results. Here is that final tally:
As you can see, that makes Labour the single largest party but one that is still 20 seats short of an absolute majority. The Lib Dems fall from 56 to eleven seats, the SNP rises to 47 seats, and Plaid Cymru has a modest three. UKIP gets nothing.
But though two uniform swings are better than one, so to speak, the strong likelihood is that in an election with so many “real” parties, there will be a much greater diversity of swings than that. As I argued here last Friday, it is almost the case that each constituency will be an election in itself with next-door constituencies producing very different results and “swings.” Kellner therefore adds a third forecast based in which he “corrects” his previous results on the basis of other such factors as the tendency for first-time MPs to enjoy a 2 percent gain in votes in their second election. When these factors are taken into account, the final projected result is extraordinarily close:
If that were the result, Labour and the SNP would be the only two parties which could form a majority coalition — but that coalition would have a majority of two (or three when the Speaker is removed from the opposition total.) A single by-election defeat would leave the government dependent upon the Speaker’s casting vote. In such circumstances the Liberal Democrats might give the government a good working majority by joining it, but that would make them vulnerable to attacks from both UKIP and the Tories in highly uncertain poltical conditions. They would probably shy away from that and repair themselves in Opposition. So the most likely constitutional result would be a minority Labour government operating on a “confidence and supply” basis, i.e., keeping the Queen’s government going but unable to pass controversial measures, however necessary. In the past (1923, 1929, 1950) such unstable arrangements have led to the dissolution of Parliament and an early election in a relatively short time.
As Kellner makes very clear, however, there are several ifs, whethers, and maybes in this projection. Slightly different assumptions at several points produce very different results. My own shortlist of queries and corrections, some echoing Kellner, others peculiar to me, is as follows:
‐ The bedrock calculation underpinning these projections is based on several current polls showing the Labour party enjoying a small lead over the Tories. That could and probably will change. Most observers think that the Tories will improve on their showing because the economy is improving. Real events change the voters’ minds. But we can’t be sure that will be the result of an improving economy or even that the economy will continue improving.
‐ The Scottish polls showing a massive SNP lead that would produce a catastrophic loss of Labour seats are a snapshot of opinion in the aftermath of a referendum campaign that has given a massive boost to Scottish nationalists. Several signs are that this is not a short-term phenomenon. But the passion, energy, and activism fuelling the SNP’s resurgence probably will decline in the next six months. Also, as the main Unionist party in Scotland, Labour might gain from tactical voting to keep out the Nats. In short, Labour will lose Scottish seats, but probably not as many as polls currently suggest.
‐ This will be the most exciting election since the Second World War. No one will be able to forecast the result with any confidence; voters of almost every opinion will find someone simpatico to vote for. As a result the turnout will rise, perhaps considerably, bringing many voters who now feel disenfranchised back into the voting booth. All parties will gain some supporters from this, but the great likelihood is that UKIP, the Greens, and the SNP will benefit much more substantially than Labour, the Lib-Dems, or the Conservatives. That is likely to compensate — and perhaps more than compensate–for the usual tendency of small parties to lose ground as the election approaches.
How significant might this be? Well, between 1945 and 1997, the turnout in British general elections never fell below 70 per cent, varying between 71.4 and 83.9 percent. In the last three elections, however, turnout has shrunk to between 59.1 and 65.4 percent. An increase of 10 percent in the 2010 turnout, while far from impossible, would bring almost three million additional voters into the polling stations as many other millions were changing parties.It is hard to believe that such a large expansion plus “churning” of the electorate would not make massive waves.
YouGov has painted a plausible picture of “stable instability”– major changes in voting behavior leading to minor changes in party representation at Westminster. It could well prove to be correct. Somehow, though, it doesn’t feel right.