There has been some electoral analysis since results came through on Friday which has focused on transfers. However, transfers are like away goals – they only matter if it is close in the first place. In fact, the story of this election is as much if not more about first preference votes, which are the real determiner of where most seats go.
Clearly the SDLP did benefit from transfers; notably a lot of transfers originating from ex-DUP Ulster Unionist candidate Jenny Palmer did not go to the DUP, enabling the SDLP to take an unexpected Lagan Valley seat. Fermanagh & South Tyrone is arguably more freaky, with an SDLP elimination helping the Ulster Unionist over the line (but, without that elimination occurring so soon, it would in fact have been an Ulster Unionist helping the SDLP over the line). It was always likely the SDLP would sneak the last seat in East Londonderry and Upper Bann on the back of Ulster Unionist transfers and the simple fact it is very hard to balance two candidates ahead of one even if you have close to two-and-a-half times as many first preference votes. So it is not that transfers do not matter, but it is worth noting that these are almost always transfers from eliminated candidates.
In the end, however, you get nowhere without first preference votes, and it is here that the Ulster Unionists were unlucky or foolish, depending on how you look at it. The Ulster Unionist first preference vote share actually rose by 0.3 points to 12.9% and back over 100,000 votes, yet yielded a calamitous result. Here, we need to look at the geography.
I have noted before that Northern Ireland can be split electorally into four distinct unit – trends in each unit are usually the same way, but can vary significantly between them. These units are:
- Belfast City (the four Belfast constituencies);
- Belfast suburbs (the five constituencies around Belfast);
- Border/Rural (South Down, Upper Bann, Newry/Armagh, Fermanagh/South Tyrone, West Tyrone); and
- North Coast/Lough (Foyle, East Londonderry, Mid Ulster, North Antrim).
The Ulster Unionist performance in these is most marked. In its best ever election for first preference vote, in 2003, the party scored 18.0%, 33.6%, 21.1% and 16.5% in these areas respectively, adding up to 22.7% overall; but by 2011 these had changed to 8.8%, 17.4%, 16.9% and 7.7% for 13.2%. Clearly this was a dip everywhere, but it was much less pronounced in Border/Rural. However, fast forward to 2017 and we find 7.3%, 22.1%, 12.6% and 8.1% for 12.9% – the same vote share more or less, but yielded in quite different locations. The marked rise to 22.1% in the Belfast suburbs, the best for 14 years, saw only one gain (East Antrim) and actually two losses (Strangford and Lagan Valley) in the five-seaters. It also constituted stacking up votes in places like North Down where they were not required. However, in the Border/Rural areas disaster struck – the marked comparative decline saw a near wipe-out, with only one seat out of five retained. We can see here, therefore, how the different geographic trends led to markedly different outcomes – had the party been able to “lend” some votes from the Belfast suburbs to Border/Rural, that would have cost it little in the former but saved it a lot in the latter. To a degree, this is luck of the draw, but it again points to what a strategic error Leader Mike Nesbitt’s transfer remarks (broadly welcomed in the Unionist majority suburbs but not at all in the Nationalist majority border area) were.
Another note here is the beginning of the evening up of the Alliance performance. These figures for the party in 2011 were 13.1%, 14.9%, 2.8% and 2.9% for 7.7% overall; in a similarly impressive election they are now 14.9%, 15.0%, 4.5% and 3.6% for 9.1%, with the added 1.4 points thus coming primarily away from Greater Belfast. On this occasion, this yielded no more seats (although the eight holds were very comfortable), but even a slight improvement in future surely would.
We live in interesting times…