On second places

Thinking about the elections the other week, and in particular the amount of noise that was made about the Tories taking ‘second place’ from Labour in Scotland. From a Manchester perspective, this chimed with the comments we’ve heard from Liberal Democrat sources about the City Council being a ‘one-party state’, on the basis that all 96 seats were occupied by Labour. (‘Were’ being the operative word; we now have one (1) Liberal Democrat councillor, former MP John Leech, who can thus consider himself the leader of the opposition (and probably does).) The implicit suggestion was that the local Lib Dems were snapping at Labour’s heels – or rather, that they would be, if only the electoral system allowed it – in much the same way that the Scottish Tories are supposedly on the SNP’s tail.

The problem with this kind of argument is that not all second places are equal. (Essentially, talking about ‘places’ in an election – instead of votes or shares of the vote – is converting an interval/ratio variable to an ordinal ranking; you inevitably lose information in the process.) Here are a couple of charts for you to compare and contrast.

Screen Shot 2016-05-21 at 15.38.02

Screen Shot 2016-05-21 at 15.38.16

What’s going on there? Clearly we’ve got two different distributions, both normal-ish and with a right skew; one has nothing under the 10% mark and is truncated on the right (at 100%), the other has nothing over 80% and is truncated on the left (at zero). Whatever we’re measuring, there’s a lot less of it in the second chart.

What we’re measuring is opposition. Specifically, the first chart is based on the votes received by second-placed candidates, in English constituencies in the 2015 General Election, as a proportion of the winning candidate’s vote. So, for example, there were 35 seats (6.6% of the total of 533) at which the election was close enough for the second-placed candidate to receive 90% or more of the winning candidate’s vote – as against 17 (3.2%) where the vote was so one-sided that the runner-up got less than 20% of the winner’s vote.

As for the second chart, those are second places in the 32 council seats that were contested in Manchester this May. As you can see, there were no seats in which the runner-up came as close as 80% of the winner’s vote, let alone 90%; only two of the 32 exceeded 60%. Both of these, it’s worth stressing, are the product of a big campaigning push by the local Lib Dems at this election in particular; one seat they won, with Labour on 69% of the Lib Dem vote, while in the other the Lib Dems came second with 76% of the Labour vote. In none of the other 30 seats did the runner-up’s vote exceed 52% of the (Labour) winner’s. In 9 of the 32 seats – 28% of the total – the runner-up vote was 20% or less of the winner’s; the 20-30% range accounts for another nine.

Three points. Firstly, this is not a political earthquake waiting to happen, for anyone; those are some distant second places. Secondly, in a situation like this it doesn’t much matter who occupies second place. Another quick and dirty chart:

Screen Shot 2016-05-21 at 17.43.10

That’s number of seats vs winner’s share of the vote; it starts at 50%-60% because the lowest winning vote share is in that range. (To be precise, the Lib Dems got 52.6% in the one seat they won; in their other main target seat they pushed the Labour vote right down to 50.3%.) As you can see, the seats where Labour got as little as three-fifths of the vote are in the minority; the mean winning vote share is over 65%, and the median is just under. In a situation where, on average, Labour are getting two votes for every one cast for all the other parties put together, caring very much about who’s in second place demonstrates either wild optimism or innumeracy.

And a whole range of people are in second place. In this table of runners-up – screenshotted from Excel, because I couldn’t be bothered to sit here for ten minutes typing in <tr> and <td> tags – I’m going back to the ‘share of the winner’s vote’ metric. In other words, ‘51%’ in this table represents getting over 20% of the vote when Labour get 40%, or over 30% to Labour’s 60%; it’s really the bare minimum to have any kind of shot at ever actually winning the seat.

So, who are the runners-up, and how are they doing?

Screen Shot 2016-05-21 at 18.05.47

(Sums to 30, not 32; one seat was won by the LDs, as mentioned above, and in one other the runner-up was an Independent.)

That’s an awful lot of not a lot going on, particularly considering that two of the three Lib Dem runner-up scores in the rightmost column were 50.1% and 51.6% (of the Labour vote). Yes, there are Kipper runners-up – quite a few of them: ten to the Greens’ eight, and a couple of them not too far below 50% of the Labour vote – but really, there’s nothing here to worry about. What we’re looking at here isn’t the rise of UKIP – their single best vote share was 27.4%, in a seat where Labour took 59.4% – but the total collapse of the local Tories and the (more recent and more dramatic) near-total collapse of the Lib Dems. In some parts of the city, in fact, the collapse of the old opposition parties is all that’s happened. Look at that top left square: four seats where the Tories were in second place to Labour, with 10% or less of the vote. (It wasn’t for want of alternatives, either; five candidates stood in three of those four seats, six in the other.) In other parts, Greens have started work on replacing the left-liberal Lib Dem opposition voice, or Kippers on replacing the Tories. But they’re in for an awfully long haul, with no guarantee of any success at all – particularly now that the Lib Dems are starting to pick themselves up again.

And that’s the third point I wanted to make: there’s a big difference between getting 40-50% of the winning party’s vote and getting votes in the 70-80% range which put you properly in contention. And a large part of what makes the difference is party organisation: having party members willing to put posters in their windows, chip in to support party funds, let you know what local people are worried about and (not least) go out on the knocker, just to make sure everybody knows that there’s an election on and that your party’s standing. Another screenshot, this one from the Manchester Evening News local elections liveblog:

Screen Shot 2016-05-06 at 14.31.49

You need people, in short – and Manchester Labour’s got plenty of those, particularly since last September.

Two final thoughts, one about electoral systems and one about Scotland. Given that, in 31 wards out of 32, Labour took more than 50% of the vote – with a winning margin (the difference between Labour’s and the runner-up’s share of the vote) ranging between 12% and 74%(!) – you might think that proportional representation wouldn’t have a lot to offer. And you’d be half right, but only half. A strictly proportional allocation of the votes cast – say, a party list system electing to a single 32-seat constituency – would give 21 Labour seats (instead of 31), 4 Lib Dems (1), 2 Tories (0), 3 Greens (0) and 2 UKIP (0). Split the seats elected more or less in two and use an additional member system – as seen in the Scottish Parliament – and you get 21 Labour, 4 Lib Dems, 1 Tory, 4 Greens and 2 Kippers (but please don’t ask me to show my working). Multiply by three for the full council, and we have Labour occupying 63 seats out of 96 (instead of 95). I admit, it’s a bigger impact than I’d anticipated before I did the number-crunching. Whether it would make Manchester any less of a ‘one-party state’ is another question. Labour would effectively be faced with three separate opposition groups, numbering 12, 12, and 9 – all of which they could outvote, jointly or severally, till the cows came home.

The lesson for Scotland, meanwhile, is to look at the big picture and not get distracted by minutiae of electoral arithmetic. Whether the Tories or Labour are in second place in Scotland is about as significant as whether the Greens or UKIP have more second places in the Manchester council results – which is to say, not significant at all. Manchester is Labour, and the party has an army of people devoted to keeping it that way; any challenger has more than one mountain to climb. Since 2015, exactly the same things can be said of Scotland and the SNP. We can argue about who threw Labour Scotland away, and whether it can ever be restored in its old form, but the political reality is that it’s gone – and that it won’t be recreated easily or soon, by anyone. This, of course, has implications for how Labour goes into the next General Election campaign – but that’s a subject for another post.

Advertisements

One Comment

  1. metatone
    Posted 23 May 2016 at 20:54 | Permalink | Reply

    Thanks for this. Much more cogent than my Twitter rantings about how the Tory 2nd place in Scotland was basically just a big distraction…

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

%d bloggers like this: