Where’s Westminster?

2019 is off to a quiet start in terms of Scottish polling and elections. There are no upcoming or anticipated by-elections for the first time since the local elections in 2017. No one commissioned any public Holyrood polling to kick-start the new year. Given the upcoming Scottish budget and ongoing train-wreck that is Brexit, this may be the calm before the storm. But for now and for nerds like me, it’s a bit boring. To fill the gap for my first 2019 post, I thought I’d go behind the scenes of this project again to explain why, as you might have noticed, there’s a Westminster seat projection shaped hole in my output.

So far the only time I’ve given explicit projections for Westminster seats based on polling came as part of my 2018 in Review posts. By contrast, I always give Holyrood seat projections, complete with shiny maps. Especially when not embedded in a thread of tweets with lots of different polling figures, people often end up specifically asking – and I, awkwardly, have to reply “I haven’t actually calculated.” Given my intention of being a one-stop-shop for all Scottish elections data, that might seem like an odd omission. There are a few overlapping reasons for this, that basically come down to;

  • Westminster is widely covered by others as it is
  • Pure First Past the Post is difficult to reliably project
  • First Past the Post projections dramatise small shifts in vote share yet are intensely boring due to non-impact of new and/or smaller parties
  • I don’t have an in-house calculator for Westminster

Even 20 years into devolution, Westminster is still often seen as the “main” parliament, and Holyrood as a distant second. This is despite an already weighty set of initial powers over everyday issues like education, healthcare, transport and justice being augmented recently with further major powers over the likes of taxation and welfare. Although there does feel like there’s an increasing disconnect between the Scottish and whole-UK spheres of politics, that hasn’t come about from Scots losing interest in Westminster. Instead, we seem to have managed the quite remarkable task of keeping Scottish MPs firmly entrenched in our Scottish bubble whilst still sending them off down to London most of the time.

As a result there are still plenty of column-inches devoted to Scots MPs in Westminster, and lots of excited rune-reading of every poll that comes out for it. Any number of places feature Westminster seat calculators, and there are at least a few that allow for Scotland-specific projections. And there are lots of excellent UK-wide politics Twitter accounts that at least touch on Scotland’s impact on Westminster. I’ve always been clear that rather than adding to that, I’m quite happy to keep my focus primarily on the Scottish Parliament and Councils.

Additionally, the purely First Past the Post nature of Westminster elections makes projections quite unreliable. We all know that voter behaviour cannot be perfectly modelled (except by Our Lord and Saviour, Professor Sir John Curtice) by simply taking a national poll and applying it to 59 discrete local areas with 59 local campaigns and 59 collections of issues. Individual seats can and do weather national waves, and there’s always a surprise or two elsewhere.

The other big issue with First Past the Post is that it’s just objectively a bad system for representing a diverse electorate. It’s a simple fact that FPTP encourages tactical voting and discourages voting for smaller parties or the emergence of new ones, because it’s not a proportional system. For Labour to have polled a quarter of the vote in 2015 yet win just a single seat was nonsense, for example. As spectators we get so caught up in the drama of a half-dozen seats changing hands that we forget it’s primarily a sign of a broken electoral system, and not a reliable indicator of support.

Throw in the tendency towards maps toned to show who won each constituency – the simplest and most obvious way to represent the data -and this feeds an approach to politics which effectively denies the existence never mind the voice of anyone who didn’t vote their constituency’s colour. And ironically enough, because it’s so dramatic to watch two horse SNP-Lab races in Glasgow or Con-SNP races in Aberdeenshire, it ends up being really boring because you don’t care who the other actors are. At Holyrood, it’s exciting to see where those parties that clearly aren’t in the running for FPTP seats are nonetheless winning list seats because they do have supporters there, and those supporters do get a voice in parliament.

“Ah,” I hear you cry, “but you do do First Past the Post projections for Scottish Parliament constituencies with those toned maps!” I do, but remember, Holyrood has that proportional layer from the lists. Although imperfectly proportional, it’s enough of a balancing factor that I feel slightly more confident about the overall number of seats than I do with pure FPTP. Beyond that, the fact I have my own calculator means can I produce additional maps showing the projected runners up and margins in each constituency. That angles those projections slightly away from the monotone, “this area is an X party area and no one else is there” visual of typical Westminster maps towards actual nuance.

It’s true that if I had a DIY Westminster calculator I could at least do the runners up and margins. But the lack of in-house calculator is effectively the result of everything else – there are already loads out there, projections are particularly dubious, and FPTP is a rubbish voting system that distorts our politics even if you present the data in the best possible way. I don’t consider it worth the effort to create an own-brand version of something plenty of other folk have already done, to feed into that distortion.

So for the time being, I’m still going to leave folks guessing on seat counts anytime I report a Westminster poll – but if you’re really keen, then remember, Electoral Calculus is always there!