How much of a DUP/SF swing *would* it take for Martin McGuinness to be First Minister?

In her speech to the DUP party conference today, First Minister Arlene Foster called attention to the narrowness of the margin between her party and Sinn Féin:

A swing of only two votes in every hundred from the DUP to Féin would see Martin McGuinness become the next First Minister.

In the last Assembly election in 2011, the DUP won 38 seats with 30.0% of first preferences, and Sinn Féin won 29 seats with 26.9%. At a first glance, that 3.1% margin between the two parties’ vote shares is even closer than the First Minister claimed; a uniform swing of a mere 1.6% would be enough to make SF the largest party by votes, and as we all know the largest party by seats gets to choose the First Minister.

But there’s an important difference between seats and votes.

Looking at the 2011 results for each constituency, and applying a (highly improbable) uniform shift of votes from the DUP to Sinn Fein while keeping the votes for other parties at the 2011 levels, it seems that the real figure required to give SF more seats than the DUP is more like 5% than 2%; the DUP could actually trail SF by more than 6% in first preferences overall, and still win more seats. This is partly because the DUP’s stronger constituencies have smaller electorates, and partly because in the last election the DUP tended to get elected with votes to spare while a number of successful SF candidates had tighter squeaks to get in.

To be specific.

1% shift: DUP 37 (-1), SF 30 (+1)

The first consequential change as a result of a DUP->SF shift actually involves neither party directly. A mere 460 votes separated David McNarry, then of the UUP, from Joe Boyle of the SDLP in Strangford. A 0.8% swing from DUP to SF would have given the UUP fewer transfers, and the SDLP more, causing a different result.

0.9% extra votes for SF gives them a second seat at the expense of the SDLP in Upper Bann; a slightly bigger shift against the DUP loses them one of their three seats in East Londonderry to the UUP.

2% shift (the “Foster line”): DUP 35 (-3), SF 30 (+1)

I see two more DUP seats at risk in this range – their third seat in North Belfast, which would have remained with the UUP in 2011 if the DUP had 1.3% fewer votes, and the third seat in South Antrim, which would have remained with the SDLP if the DUP had 1.8% fewer votes and there were also 1.8% more Nationalist transfers to go round.

3% shift: DUP 34 (-4), SF 30 (+1)

If the DUP vote shifts to SF uniformly by about 3%, I think that they lose the second seat in Upper Bann to SF and the SDLP keep theirs (rather than the SDLP losing to SF as imagined above). This is the first case of a direct transfer between the two parties.

4% shift: DUP 34(-4), SF 30 (+1)

I don’t see any more seats falling in this range, though the DUP would be trailing SF by at least 5% in total vote share by now.

5% shift: DUP 30 (-8), SF 34 (+5)

This is the real tipping point, with four direct transfers between the two parties. On a 5% shift, I think the DUP would likely lose seats directly to SF in South Down, Foyle, Lagan Valley and Mid Ulster. Even if only two of those four were to shift, it is enough to give SF the the right to choose the First Minister – they would be tied on 32 seats each, but the number of votes would then be taken into account. Sinn Fein would have 31.9% of the vote to the DUP’s 25.0%; the total Nationalist vote would be around 46% and the total Unionist vote around 43%.

So, basically, the DUP can afford to lose a few votes and still be the largest party. Understandably enough, this is not their preferred scenario.