The 2019 Westminster election in Northern Ireland and the impact of the electoral system

The headline stories from the results of last week’s Westminster election in Northern Ireland have been well aired. A setback for the DUP, treading water by SF, a revival by the SDLP and Alliance, the continued marginalisation of the UUP, and the near irrelevance of the other smaller parties.

Closer inspection of the figures delivers a more nuanced picture. The DUP vote fell in every constituency (apart from a tiny and irrelevant increase in West Belfast), but it should be borne in mind that it fell from a historic high: the 36 per cent of the votes the party won at the 2017 Westminster election is its best ever performance at a domestic parliamentary election (leaving aside EP elections where candidate factors loom especially large). Last week it won 31 per cent of the votes, a figure it has exceeded only twice in the 19 contests of the past forty years (in 2017 and 2005). In terms of seats, 8 out of 18 preserves the party’s position as the strongest in NI, and this figure too is among the DUP’s best, exceeded only in 2005 and 2017. Still, there is no denying the blow to the party’s morale dealt by the defeat of its deputy leader Nigel Dodds in North Belfast, and the disappearance of its previously influential position at Westminster with the Conservatives’ achievement of an overall majority may increase its readiness to restore the devolved Assembly.

Sinn Féin retained its 7 seats, the most the party has ever won at a Westminster election, and John Finucane’s victory in North Belfast, following Mark Ward’s surprise victory in the Dublin MW by-election at the end of November, is clearly a morale booster. As against this, it lost Foyle to the SDLP not by a narrow margin, as had been widely expected, but by a landslide. And looking at the broader picture, SF lost more votes than the DUP did (it fell by over 6 per cent, even allowing that it stood its candidates down in three constituencies, and its vote declined in every constituency bar North Belfast), and its 23 per cent represents its poorest performance since the 2001 Westminster election.

The SDLP is understandably upbeat about capturing 2 seats, having been wiped out at the 2017 Westminster contest. The party has never come close to achieving the 7 seats won by SF last week, its Westminster peak being 4 seats in 1992. In terms of votes it was up a bit on recent performances, though it still received fewer than 15 per cent of the votes, as has been the case at every contest since 2010. Still, its 2 Westminster seats will re-establish the party’s relevance, though it is unfortunate for the province that its nationalist pro-Remain voices were not heard in the Commons chamber in the previous parliament, where they could have made a difference.

Alliance (APNI) had its best election ever. It won a Westminster seat for only the second time ever, as well as being not too far off recapturing the East Belfast seat that it held between 2010 and 2015, and it more than doubled its overall vote. The 17 per cent of the votes that it won is by some way its highest level ever, taking it to the position of third strongest party (in terms of votes) for the first time ever. With the two main unionist parties taking only 42 per cent of the votes and the two main nationalist parties just 38 per cent, the ‘centre ground’ has never been stronger.

The Westminster single-member plurality, or ‘first-past-the-post’ (FPTP), electoral system is notoriously poor at converting voter preferences accurately into seats in multi-party systems, and Northern Ireland exemplifies this very well. Only 3 of NI’s 18 Westminster MPs won a majority of votes in their constituency, and the winner in South Down received fewer than a third of the votes there. Even a modest change of electoral system to the alternative vote would almost certainly have given the SDLP an additional seat at the expense of SF in South Down, while the UUP (thanks to Alliance transfers) rather than the DUP might have prevailed in South Antrim and there is an outside possibility that UUP transfers would have given the North Down seat to the DUP at the expense of Alliance.

More fundamentally, proportional representation in NI would have effected major changes in seat distribution over the years. How would the parties have fared if Northern Ireland had been treated as a single constituency, with seats allocated by the Sainte-Laguë formula? (The Sainte-Laguë method is the ‘fairest’ method of allocating seats, in the sense that it is even-handed between small and large parties; the D’Hondt formula, in contrast, tends to favour the larger parties. It is not known whether in 1998 D’Hondt was chosen as the basis for awarding positions in the NI Executive because of a lack of awareness of its large-party bias or because, on the contrary, this property was only too well known to the key decision-makers.)

We assume for the purpose of this exercise that all votes in the hypothetical large constituency would have been cast just as they actually were. Of course, that would not have been the case, given that under first-past-the post supporters of smaller parties have an incentive not to vote sincerely but to vote strategically in order to prevent their most-disliked candidate from winning. In addition, patterns of selective withdrawal distort the picture to a degree, though to some extent this evens itself out, as in 2019 the SDLP’s withdrawal in favour of SF in North Belfast was balanced by SF’s reciprocation in South Belfast, and the UUP’s withdrawal in favour of the DUP in North Belfast was balanced by reciprocation in Fermanagh–South Tyrone. In any case, the actual voting figures are really all we have to go on.

Comparing the seats actually won with the ‘fair’ outcomes under province-wide PR (see Table below) shows how much FPTP has distorted the outcomes. In the early years the UUP were the main beneficiaries, taking 65 per cent of the seats with 34 per cent of the votes in 1983, for example. More recently the UUP has become one of the smaller parties and hence has been under-represented, and it is now the DUP that achieves a sizeable seat bonus, in 2017 taking 56 per cent of the seats with 36 per cent of the votes. SF has been reasonably accurately represented over the period as a whole. The main victims have been the moderate and centrist parties, Alliance and the SDLP. The SDLP has taken only 26 seats over these 11 elections compared to the 36 that its votes would have earned it on a proportional basis, while Alliance ‘should’ have received at least one seat at every election instead of winning just 2 over the entire period. Over the 5 elections of the period 2005–2019 inclusive, the three most centrist parties (UUP, SDLP, Alliance) have won just 16 seats compared with the 35 that their votes would have earned them on a proportional basis. (If D’Hondt rather than Sainte-Laguë were employed, the aggregate figures over this 40-year 11-election period would be DUP 47, UUP 52, SF 39, SDLP 37, Alliance 13, and Others 1 (the Conservatives in 1992).)

NISeatsPR.pngIn short, FPTP, or single-member plurality as it is better termed, has consistently worked against the representation of centrist and moderate opinion in Northern Ireland and has operated to the benefit of the DUP in particular. However detrimental this may be to the interests of Northern Ireland, there seems very little likelihood of NI receiving a special ‘opt-out’ of the sort that it has enjoyed for European Parliament elections; this could take the form not necessarily of a province-wide 18-seat constituency but, for example, three 6-seat constituencies. Clearly, allowing the people of Northern Ireland to choose their representatives by PR would raise the question of why the people of the rest of the UK should be denied this opportunity, and neither main party in Britain has shown any kind of enthusiasm for PR. The unsuitability of FPTP for any multi-party polity is well attested, and its consistent weakening of the centre ground in Northern Ireland makes it a particularly inappropriate method of determining the representation of voters there.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your 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 )

Connecting to %s