Yesterday, Northern Ireland’s voters went to the polls to elect a new assembly. Counting is now underway in early elections that were sparked by a row between Sinn Féin and the Democratic Unionist Party (DUP) over the Renewable Heat Incentive scheme.  The scandal, however, revealed much wider differences between the partners in the North’s compulsory governing coalition. And so for the fifth time in 34 months, Northern Ireland’s voters went off to the voting booths.

There has been some novelty to this election. Two of the parties have new leaders (Michelle O’Neill for Sinn Féin and Naomi Long for Alliance). And women lead three of the parties (O’Neill, Long, and Arlene Foster of the DUP), a change from the male dominated politics Northern Ireland has experienced up to now. However, the assembly election campaign has borne all the traditional hallmarks of previous campaigns.  Green versus orange: chasms between both communities aggravated by backbiting over who was responsible for the collapse of the power-sharing executive and sniping between leading personalities of both Sinn Féin and the DUP. Further, the parties of nationalism and unionism have taken pot-shots at each other in the furious scramble for votes. With the assembly reducing from 108 seats to 90, this fight has been especially intense, both between parties and between candidates of the same party.

One notable departure from the conventional trajectory of Northern Ireland elections was the comments by the Ulster Unionist Party (UUP) leader Mike Nesbitt on 13 February. In an interview on the BBC, Mr. Nesbitt announced he would be transferring his preference vote to the nationalist SDLP once he had voted for all UUP candidates.[1] While Nesbitt stopped short of calling on all UUP supporters to follow his lead, senior party figures criticised his call. Mr. Nesbitt also earned a strong rebuke from first minister and DUP leader Arlene Foster who said: “It is dangerous to advocate supporting candidates who are Pro-United Ireland above preferences for fellow unionists. The greater the number of nationalist MLAs elected the stronger the push will be for a Border Poll.”[2]

Mr. Nesbitt’s decision to make public his own preference vote has brought into the spotlight the issue of transfers and raises the vexed question of how important are they? Conventional wisdom, among many media commentators (e.g.: Coghlan 2002, Collins 2015) is that they play an important role in deciding the result of an election. Evidence from the Irish National Election Study (INES) shows that voters do cast lower preferences when given the chance (Marsh et al. 2008). Yet few studies have explored the impact of transfers on elections. Among those that have, these analyses focused on one country, were descriptive, and are now dated. Thus, a more comprehensive analysis of the decisive impact of transfers on elections in single transferable vote (STV) systems is required.

A research project by Quinlan and Schwarz exploring the mechanical impact of transfers in STV systems seeks to fill this void. We have collected data from multiple elections over five decades in four countries: Ireland, Northern Ireland, Scotland, and Malta (spanning a near universe of the countries that use STV for voting). In this contribution, we focus on preliminary results from our analysis of transfers in Northern Ireland. Our data focuses on election results from the five assembly elections held between 1998 and 2016. Our objective is two-fold: 1) to decipher if transfers have a decisive impact on the result and 2) if so, do transfers benefit one party over another?

We suggest that transfers have a decisive impact on the result when transfer distribution changes the result. In other words, if preferences hadn’t been given out, would the result have remained the same? Two examples, detailed in Table 1 below helps us to explain this. On the left-hand side, we detail the results of the 2011 Northern Ireland assembly election in North Antrim, where six seats were up for grabs. On the right, we show the results of the 1989 Irish general election in Cork South Central, where five seats were on offer. In North Antrim, the candidates who occupied the winning positions on the first count (i.e.: positions 1-6) went on to be elected after transfers had been shared. Thus we assume that transfers in this situation had no decisive impact on the result. However, on the right-hand side of Table 1, Cork South Central tells a different story. We see that two candidates who occupied winning positions after count 1 failed to be elected. Instead, candidates Dennehy (FF) and Wyse (PD), who occupied sixth and seventh position on the first count, overtook O’Keefe (FF) and Corr (FG) to win election. In this circumstance, we can say that transfers had a decisive impact as the transfer distribution changed the result from count 1.[3]

So to the key question: have transfers had a decisive impact in Northern Ireland? Figure 1 shows that the decisive impact of transfers in Northern Ireland has varied between elections. On average transfers have had a decisive impact in 10 per cent of cases but this ranged from as low as 5 per cent in 2007 to 14 per cent in 1998 and 2003. In last year’s assembly election, transfers had a decisive impact on the destination of 11 seats (11/108=10.1%). This preliminary data suggests two things. First, transfers can and do have a decisive impact on elections. Second, this effect however, must be kept in perspective – transfers having a pivotal impact are the exception and not the rule and we should be careful not to overemphasize their influence.

Figure 1 Proportion of candidates decisively elected by transfer votes in Northern Ireland assembly elections 1998-2016. Source of data: Quinlan & Schwarz (2017).

The second question is: are certain parties in Northern Ireland benefiting when transfers are decisive? There are two ways of looking at this. Figure 2 does this by taking a descriptive look at whether parties gained or lost seats because of the distribution of transfers in the five assembly elections. A quick review shows that the Ulster Unionists have consistently gained seats because of transfer distribution, gaining four seats in 2016. In most elections, Sinn Féin have been penalized by transfers, losing seats in four of the five elections, including six seats in 2003. Over time, the SDLP has gained more than it has lost, although in the two most recent Stormont elections it has lost seats because of transfers.

Figure 2 Net seats gained and lost by parties after distribution of transfers in Northern Ireland assembly elections 1998-2016. Source of data: Quinlan and Schwarz (2017). Please note: Instances where transfers resulted in a change of personnel from the same party but not a change in a party’s net seats are not included.

However, to answer the question convincingly, we turn to multivariate analysis. This allows us to control for the impact of other reasons that might influence both transfers being decisive and the likelihood of being elected, namely:  a candidate’s sex, whether they were an incumbent, and the number of candidates standing in the constituency. Figure 3 confirms the descriptive analysis about the Ulster Unionists. It shows that compared with the DUP (our reference category in the multivariate models), UUP candidates were more likely to benefit from transfers. This is to the tune of 12 percentage points – in other words when transfers matter, the Ulster Unionists are more likely to benefit than the DUP. We see a similar situation for the SDLP. Compared with the DUP, SDLP candidates are more likely to benefit by about 11.5 percentage points.

Figure 3 Average predicted effect of transfers being decisive for the Ulster Unionist Party in Northern Ireland assembly elections 1998-2016. Source of data: Quinlan & Schwarz (2017). Please note: Predicted effect based on logit model considering other variables which are held at their mean values (n=535).

Our analysis is preliminary and we must recognize the number of cases we have to explore is limited compared to the data we have available on other countries. Yet we can draw some tentative conclusions. Transfers do have a decisive impact in Northern assembly elections. But their effect should not be overstated, as the circumstances in which they matter are the exception rather than the rule. Our data implies when transfers are decisive, they benefit the moderate parties, the UUP and the SDLP. This is consistent after multivariate analysis but we must admit that our estimates yielded large confidence intervals so the effects are likely to be modest.  Nonetheless, it does suggest that First Minister Foster’s concern over Mike Nesbitt’s comments was not without foundation given that transfers tend not to benefit her party. However, her anxiety might be overblown because transfers tend not to come into play often, and therefore, while important, might not be as imperative as commentators sometimes imagine. We will know on Saturday evening, when all the votes have been counted in Vote 2017 if the patterns we have outlined above are repeated.

Post by Dr. Stephen Quinlan and Hannah Schwarz of the GESIS-Leibniz Institute for the Social Sciences, Germany.

 BIOGRAPHIES AND ACKNOWLEDGEMENTS

Dr. Stephen Quinlan is Senior Researcher at the GESIS – Leibniz Institute for Social Sciences in Mannheim. He is the Project Manager of the Comparative Study of Electoral Systems (CSES, http://www.cses.org) project, which explores electoral behaviour in over forty states worldwide.

Hannah Schwarz is a Data Processing Specialist at the GESIS – Leibniz Institute for Social Sciences in Köln, and works primarily on the Comparative Study of Electoral Systems  project.

The authors are sincerely grateful to Annika Stein for research assistance on our project.

REFERENCES& FOOTNOTES

Coghlan, D., 2002. Fianna Fail is knocking on the door of an overall Dáil majority, according to latest poll. The Irish Times, 3 May.

Collins, S., 2015. The next election will be all about transfers. The Irish Times, 31 Oct.

Marsh, M., Sinnott, R., Garry, J., and Kennedy, F., 2008. The Irish Voter. Manchester, UK: Manchester University Press.

[1]     See: http://www.bbc.com/news/uk-northern-ireland-38960977

[2]     See:  http://www.belfasttelegraph.co.uk/news/northern-ireland-assembly-election/dup-leader-claims-it-is-dangerous-to-transfer-votes-to-nationalists-35460405.html

[3]     We do not assume that the outcome would have been the same under a system where only the first preferences taken into account. Rather, our focus is on what happens to transfers.