What if the Alternative Vote system had been used?

A psephological 'what if' – 'what if' the Alternative Vote system had been used on May 6? Would the result have been significantly different?

AV is basically the well-known PR-STV system, but operated in single-member constituencies. If no candidate is the first preference of a majority of voters, the candidate with the fewest number of first preference rankings is eliminated and that candidate's ballots are redistributed at full value to the remaining candidates according to the next ranking on each ballot. This process is repeated until one candidate obtains a majority of votes among candidates not eliminated.

Obviously this particular 'what if' requires a certain number of assumptions to be made, but luckily Northern Ireland has a rich seam of previous elections – using PR-STV – to mine. So, for simplicity, the approximate transfer behaviour of the 2007 Assembly elections will be used. The TUV, of course, did not exist in 2007, but the transfer behaviour of like-minded unionists will suffice as a proxy.

West of the Bann Sinn Féin and SDLP transfers will go to each other – but only 60% of them – the rest do not transfer. East of the Bann (or, more accurately, in Greater Belfast), SDLP transfers will be split 50/50 between Sinn Féin and Alliance. Unionist transfers will remain within the family, but TUV votes will not all transfer – perhaps only 80% of them. Alliance transfers will be split 50/50 between the SDLPO and the UUP.

In addition, of course, account must be taken for the tactical voting in 2010, where voters had basically one chance to make it count, and may thus have already voted for their 'second, but most realistic, preference', rather than their real first preference (if he/she was perceived to have no chance of being elected).

With these brave assumptions, it is possible to assess the AV winners of each of the 18 seats as follows:

• West Belfast, Mid Ulster and North Down would be won outright, as their winner received over 50% of the 'First Preferences'.
• East Belfast: it is likely here that the Alliance first preference vote would not have been so high. The nationalist candidates would be quickly eliminated and their votes transferred to Long, but would make only a marginal difference. Vance's TUV votes would have gone in the opposite direction, adding slightly to both Ringland's and Robinson's total. The question of who would next be eliminated decides this contest. Probably all three – UCUNF, Alliance and DUP – are within a short distance of each other. If, at this point, Ringland is ahead of Long, then she is eliminated and many of her votes transfer to Ringland, giving him the winning margin. But if Long is ahead of Ringland, then it is the UCUNF votes that are transferred – and they may tend to stay in the unionist family, giving the victory – a narrow one – to Robinson.
• North Belfast: this would be close. All candidates except Dodds and Kelly would be eliminated – giving Dodds the UCUNF transfers, and Kelly some of the Alliance and SDLP transfers. But probably not enough, so Dodds would win.
• South Belfast: the Green and Alliance transfers would have seen McDonnell over the line without much difficulty.
• East Antrim: when the nationalists are eliminated a sufficient number of their votes would not transfer, and with a small number of transfers from the TUV Sammy Wilson would easily win.
• East Derry: Alliance and the TUV would be eliminated first, giving a small boost to the DUP and the SDLP. The SDLP would probably be next to go, giving Sinn Féin a boost, and in much smaller numbers, the UUP. Last to be eliminated, though, is UCUNF, who carry Campbell over the winning line.
• Fermanagh and South Tyrone; the SDLP transfers would see Michelle Gildernew safely home.
• Foyle: there was some unionist tactical voting here, but this merely pre-empted the transfers that would have happened under AV. Durkan would get in safely with unionist transfers.
• Lagan Valley: Donaldson is so close that the non-transferable votes after the elimination of Sinn Féin, the SDLP and the TUV would carry him mathematically over 50%.
• Newry and Armagh: this offers the interesting possibility of serious unionist tactical voting. Conor Murphy got 42% of the 'first preferences', but it would take 54% of all the other votes to transfer to the SDLP for Bradley would take the seat. The probability of such coordination amongst unionists is fairly small, though, so Murphy would probably retain the seat.
• North Antrim: nobody was within an ass's roar of Paisley, so he would take the seat with ease.
• South Antrim: the Alliance votes and some of the SDLP votes would probably transfer to Empey, and there is little chance that Sinn Féin transfers would go to McCrea. McCrea would get some TUV transfers, but there weren't enough TUV votes to change the outcome. Empey would have won.
• South Down: as in Foyle, Ritchie's 2010 vote was swollen by unionist tactical votes. But these would happen under AV, just one step later.
• Strangford: although Shannon may not have gotten many transfers, the non-transferrable votes from Sinn Féin and SDLP eliminations would see him safely over 50% of those remaining.
• Upper Bann: Alliance and SDLP elimination would take Sinn Féin beyond the UCUNF (and maybe DUP) candidates, but the elimination of UCUNF would have brought the DUP's Simpson out ahead.
• West Tyrone: Sinn Féin would be over 50% with the help of only a few transfers. It would have taken over 92% of all of the other transfers to have gone to the SDLP for them to have beaten Doherty, and that is nigh-on impossible.

So AV would, in fact, have changed relatively little. Peter Robinson probably would have retained his seat, and Empey would have gained one. That, of course, is just for the 2010 election – in the future Northern Ireland's relatively sophisticated electorate may learn to play the system better, and use their transfers more decisively. Perhaps in 2015 the electorate will get a real chance to try.