THE SUPREME COURT

Between/

1 I agree with my colleagues McKechnie, Clarke and Charleton JJ. on the principal ground of appeal, namely, that in this case a number other than 1 or “one” in a numerical sequence on a ballot paper is not a “mark clearly indicating a first preference” within the meaning of article 80(2)(b) of the Local Election Regulations 1995 (S.I. No. 297 of 1995) (“the Regulations”). I wish to add some brief observations to explain why I consider the petitioner here is entitled to relief notwithstanding the significant weight that I consider should be attached to the certainty of outcome in elections, and which was identified, admittedly in the context of constitutional referenda, in the judgments of the members of this Court in Jordan v. Minister for Children and Youth Affairs [2015] I.E.S.C. 33. I also wish to explain why I respectfully differ from Charleton J. and consider that it has been established that the issue identified here is “likely to have affected the result of the election” within the meaning of s.5 of the Local Elections (Petitions and Disqualifications) Act 1974 such as to entitle the appellant to relief under s.8(1)(a) of the same Act, namely, that the votes cast in the election shall be counted afresh.

2 This is an appeal by case stated pursuant to s.7(3) of the 1974 Act, which provides that where a petition challenging the outcome of a local election has been tried in the relevant Circuit Court, a party to the petition may appeal on a question of law to the Supreme Court. In simple terms, the question is whether an individual ballot paper conforms to the requirement of the Regulations. Article 80(2) of the Regulations, which is contained in Part XIII of those Regulations, headed “Rules for the Counting of the Votes”, is in fact expressed in negative terms. It states:

“Any ballot paper –

(a) which does not bear the official mark; or

(b) on which the figure 1 standing alone, or the word ‘one’ or any other mark which, in the opinion of the returning officer, clearly indicates a first preference, is not placed at all or is not so placed as to indicate a first preference for some candidate; or

(c) on which the figure 1 standing alone indicating a first preference, or the word ‘one’ or any other mark which, in the opinion of the returning officer, clearly indicates a first preference, is set opposite the name of more than one candidate; or

(d) on which anything is written or marked which, in the opinion of the returning officer, is calculated to identify the elector;

shall be invalid and not counted, but the ballot paper shall not be invalid by reason only of carrying the words ‘one’, ‘two’, ‘three’ (and so on) or any mark which, in the opinion of the returning officer, clearly indicates a preference or preferences.”

3 Since the time of the great Reform Act, it has been the case that disputes about elections are brought to an electoral court. While, in principle, it might have been possible to provide that a returning officer could allocate votes according to his or her best estimate of the likely intentions of the voter, the legislation does not do so. Instead, it requires the exclusion of any ballot paper which does not carry the figure “1”, the word “one”, or a mark clearly indicating a first preference. The fact that the Regulations are framed negatively in this way is significant. A ballot is invalid unless it satisfies one of only three criteria: in this case, the only possible one is that it contains a mark constituting a clear indication of first preference.

4 For the reasons set out by my colleagues, with which I agree, the sequence 3, 4, 5, or any similar sequence, does not satisfy the statutory requirement of clearly indicating a first preference. In this regard, it is important, firstly, that the voter has failed to comply with the instruction to place a 1 (whether in numeral or word) on the ballot paper. Secondly, it is significant that what is alleged to be a mark, in this case indicating a first preference, is itself a numeral other than 1. The number 3 only makes sense in the context of other numerals. To that extent, the number 3 implies the existence of a number 1 and also that they are different, and that they are mutually exclusive. In the same way, “third” implies the existence of “first” and that they are, by definition, different things. Normally, to say that something is third means that is not, and cannot be, first.

5 This, however, is only a starting point. It might be thought of as the default position. There are a number of circumstances where the context might lead to a different conclusion. In simple terms, if a person explains that every time they say three they mean one, then the default position might be overridden. Similarly, if it is known that a person has an irrational fear of the numbers 1and 2, or has a computer where the striking the key “1” always produces the number “3” then once again the default position might be overcome. The question of interpretation of any communication depends firstly on the words used, but also on the context in which they were used, and finally on the degree of certainty required.

6 Here, however, it is a requirement of the process that the decision maker (and therefore interpreter) be unaware of the identity of the person completing the ballot paper, and furthermore know nothing about them. The method of communication is deliberately reduced to one of the most simple and normally unambiguous forms known: in a first past the post election, an X in the appropriate box will suffice: in a proportional representation election, the use of numbers indicating preference. Furthermore, the returning officer (or the court) must consider that the mark clearly indicates a first preference.

7 It was accepted that a numeric sequence like 3, 4, 5, or 4, 5, 6, might occur on a ballot paper in a single election held on a given day in circumstances where the number 3 or 4, as the case may be, was the lowest number (and therefore the highest preference) on the ballot paper completed. In such a circumstance, a returning officer has only the numbers and the ballot paper for context. It was accepted by all parties that it was at least possible that such an event could occur, and that there were possible explanations for it which were not consistent with the candidate, against whom the number 3 had been marked, being the voter’s first preference. It is not necessary to assign any particular likelihood to the different possible explanations: it is enough that it is possible that the number 3 might suggest a first preference, but might also suggest, as indeed the number normally implies, a third preference. The ‘Memorandum for Guidance of Local Authority Returning Officers at the Local Elections, Friday the 23rd of May, 2014’, issued by the Department of Environment, Community and Local Government, made the practical suggestion that in such circumstances, the returning officer should discuss the matter with agents. This might resolve the matter one way or another on a practical basis, but it is, at a minimum, indicative that the matter is open to doubt. I also doubt that it is in fact a matter for individual agreement: if it is the case that such a sequence might be taken as indicating a first preference, but might not, then it is difficult to say that it indicates such a preference clearly. Counsel other than counsel for the Minster seemed to take the position that in such circumstances, while the number 3 in a sequence 3, 4, 5, etc. might suggest a first preference, and even indicate such a preference, it could not be said to “clearly indicate” a first preference, and such a vote would be invalid. If this was the position they inclined to, I think they were correct to do so.

8 Accordingly, in my view, the case narrows to the point where it must be argued that the fact of another election being held on the same day, which would give a plausible reason for the existence of ballots in the local election where the highest preference on the ballot paper was a number such as 3, is sufficient to satisfy the statutory test and clearly indicate a first preference. In my view, this is not sufficient to satisfy the statutory requirement. Once it is accepted that it is possible to have the sequence 3, 4, 5, on a ballot paper without the number 3 indicating the voter’s first preference, then it does not seem to me to be sufficient that it can be said that it is more likely than not that the candidate with the 3 marked beside them was, in fact, the voter’s first preference. Nor is it sufficient that it can be said that the vast majority marked in such a way are ballots in which the candidate with 3 marked beside them is in fact the voters’ first preference. Each ballot paper must be taken on its own. If some of them could have been marked in this way in circumstances where the voter intended by the number 3 to indicate that that candidate was his or her third preference, then to determine in all cases that the number clearly indicated a first preference would be simply wrong. As McKechnie J. points out, the terms of the Memorandum for Guidance, which advises that such votes should be admitted, does so on the basis that the sequence “suggests” that voters followed on from preferences recorded on ballot papers with another election held on the same day. There is, however, a significant difference between a suggestion and a clear indication.

9 It is, I think, important that this is what might be described as a mechanical case: the question for the Court is the correct way to determine the outcome of the poll held on Friday the 23rd of May, 2014. There is no question, therefore, of the difficulties discussed in Jordan which may arise if the remedy sought is the holding of a fresh poll. Subject to the qualification that the proportionate representation system has an inbuilt, and perhaps unavoidable, element of randomness at the point of distribution of surplus, the issue here is merely the proper counting of the votes cast, and can be readily ascertained by carrying out that exercise now. Where real doubt exists as to the outcome which can be dispelled definitively by the recounting of the votes (which are kept after the election to facilitate that very exercise) it seems to me that the values of certainty and confidence in the outcome tip in favour of dispelling doubt.

10 In the judgment he delivers, Charleton J. considers that the evidence does not go so far as establishing, however, that the mistake in relation to these votes is one which is likely to have affected the result of the election. As I understand it, he considers that the evidence goes so far as to show that there may be more invalid votes included than the margin in this case, but that the evidence does not establish that those votes affected the result in the sense that it is, in theory, at least possible that all of the invalid votes cast should be treated as first preferences for the petitioner or some other candidate, which, if removed, would only have the result of increasing the margin by which the petitioner lost. He considers that the petitioner ought to have sought inspection of the ballot papers under article 93 of the Regulations as part of the preparation for the petition, and would then have been in a position to go further on the evidence. In future petitions, should an issue such as this arise, it may indeed be prudent to seek to inspect, but I am not convinced that failure to do so is a bar to the petition in this case. First, in a simple case, the identification and removal of invalid votes may make it clear what the correct outcome was. For such a case, the impact on the election will not be likely, it will be certain, one way or the other. I am not convinced that a petitioner has to demonstrate certainty in order to succeed. In other circumstances, the consequences of the removal of the invalid votes may indeed be more complex. Proportional representation is a notoriously sensitive system and the outcome of an election may depend not just on the numbers, the votes, and the preferences, but also on the sequence in which they are distributed. Even if on their face the removal of invalid votes might appear broadly neutral, their removal may have an impact on the distribution of votes, the elimination of candidates, and therefore the election of others. In such a case, it is not possible to be certain as to the outcome without rerunning the entire count. I am not satisfied that it is necessary to do so in order to seek relief on the petition, or indeed that inspection under article 93 would necessarily permit such an exercise. Inspection, therefore, might have made it more (or less) likely that the inclusion of these votes had an impact on the result, but that does not mean that the present, limited ,evidence does not indicate that effect on the outcome was likely. This was an extremely close election, and in the circumstances of this case, I am satisfied that the evidence, limited as it was, is sufficient for the petitioner to succeed.

11 I am influenced in this regard by analogy with the test applied in Jordan. Once this Court has found that there was a mistake with the sequence which did not commence with the number 1, and given the fact that the election was so close, then the point has been reached where it can be said that a reasonable person could be in doubt about, and no longer trust, the outcome of the election. That doubt can be readily resolved by having the ballots already cast counted afresh. If the Circuit Court judge had considered that the view of the law that the Court has taken was even possible, then I have little doubt that he would have ordered the counting of the votes in the same way as occurred in in the Matter of Election to European Parliament for Constituency of Leinster held on June 15, 1989, Petition of Micheal Bell (Unreported, High Court, Hamilton P., 24th November, 1989), if only to eliminate the possibility that the arguably invalid votes had affected the outcome. That would have been a sensible approach to the case and I see no reason why it should not be taken now. The position has been reached where all participants in the election, and indeed the relevant electorate, are entitled to know, rather than speculate about, the ballots cast.

12 I agree with my colleagues that the consequence of an order that the votes be counted afresh means that the process should be restarted under the supervision of the Court from the point of the opening of the ballot boxes. Because there is an inherent element of randomness in the functioning of the proportional representation system, this may mean that other variables are introduced which may have an impact on the outcome. There may be nothing inherently wrong in this: it is a feature of the system for elections which is used in Ireland. However, I agree with the observations of Clarke J. that it may be possible, if considered desirable, to seek to provide a remedy which would ensure that, on an order that the votes be counted afresh, the only variable involved would be the removal of the factor which the Court had found to constitute a mistake. That, however, is a matter of policy and practicality, and may be the subject of dispute in future litigation. On the view I take of the statute, it does not arise here.