England and Wales High Court (Queen's Bench Division) Decisions

IN THE HIGH COURT OF JUSTICE
QUEEN S BENCH DIVISION
DIVISIONAL COURT
IN THE MATTER OF THE REPRESENTATION OF THE PEOPLE ACT 1983
AND IN THE MATTER OF A LOCAL GOVERNMENT ELECTION FOR THE
KINSON SOUTH WARD OF BOURNEMOUTH BOROUGH COUNCIL HELD ON 7 MAY 2015

		Petition No. M/352/15 Royal Courts of Justice Strand, London, WC2A 2LL
		30/10/2015

MR JUSTICE JAY:

Introduction

1. On 7^th May 2015 combined parliamentary and local government elections were held in Bournemouth. There are 18 local government Wards of the Bournemouth Borough Council, each retaining 3 councillors. Two of these Wards are Kinson North and Kinson South. On 7^th May 2015 there were 7,664 registered electors eligible to vote in the Kinson South Ward.
2. The result of the election for Kinson South Ward, following a recount, was declared on 8^th May 2015. Laurence Fear (UKIP) secured 1,224 votes, Roger Marley (Conservative Party) 1,188 votes, Norman Decent (Conservative Party) 1,147 votes and Beryl Baxter (Labour Party) 1,083 votes. Accordingly, Laurence Fear, Roger Marley and Norman Decent were elected. It may be seen on the elementary arithmetic that the fourth candidate, Beryl Baxter received 64 votes fewer than Norman Decent, 105 votes fewer than Roger Marley and 141 votes fewer than Laurence Fear. There were a number of other candidates, but the gulf between them and Ms Baxter's tally is too great to have any material impact in the circumstances which have arisen.
3. Beryl Baxter has petitioned this Court seeking a declaration that the local government election was invalid, on the basis of alleged acts and omissions perpetrated by the Returning Officer, the Fourth Respondent to these proceedings. Helpfully, the parties are agreed that the issues raised by this Petition can be determined by this Court by way of a Special Case pursuant to section 146 of the Representation of the People Act 1983. This has eliminated the need for the finding of facts in an intricate evidential setting.
4. The two issues raised by this Petition are important and novel. Before addressing them, it is convenient to distil the essential ingredients of what occurred on 7^th and 8^th May 2015.

Essential Factual Background

5. The problems arose owing to a simple, unfortunate, printing error. "Books" or batches of ballot papers were printed and assembled in advance of election day. Whilst the cover of the Kinson South ballot books showed the correct information for the Ward, the ballot papers inside were for Kinson North; and vice-versa.
6. At around 7.15 on polling day, the Presiding Officer at one of the polling stations in Kinson South telephoned the Elections office, contained within the Council's offices, informing it of the problem. It very soon became clear that this problem affected each of the nine polling stations in both Wards, and steps were promptly taken to start moving ballot papers from the incorrect polling stations to the correct ones. As this was happening, all polling stations in Kinson North and Kinson South were contacted and stopped issuing ballot papers for the local government elections for these Wards until the correct ballot papers had been provided.
7. By 10.15, all polling stations in Kinson North and Kinson South had been issued with the correct ballot papers. In one of the nine Wards, the correct ballot papers could be issued immediately, and so there was no period of postponement or interregnum. However, the maximum hiatus was one of 2 hours 32 minutes.
8. Unfortunately, some electors provided with the wrong ballot papers nonetheless placed them in a Kinson South ballot box before the pause in voting.
9. Since the ballot boxes originally provided now contained ballots for the wrong election, each polling station was furnished with a new ballot box. In the meantime, lists had been prepared of those affected by the ballot paper problem. These were compiled by checking the corresponding number lists in the station for the elector numbers of those impacted, which information was then cross-referred to the register of electors to obtain the names and addresses of the affected electors.
10. At around 13.00 on 7^th May 2015, a decision was made to send a "door knocking team" to the addresses of the affected electors, with instructions to invite the latter to return to vote. The exercise began at around 13.30, and continued until around 16.30 when a new team, armed with the same instructions, was despatched to continue the exercise, which was completed at around 21.20.
11. After the result of the election was declared on 8^th May, an unofficial investigation was undertaken to ascertain the numbers of electors affected by the ballot paper error. The outcome of this investigation, taken in conjunction with the formal inspection and re-count which took place in the High Court before Master Eastman on 17^th September 2015, may be summarised in the following way.
12. The number of electors in the Kinson South polling stations issued with incorrect ballot papers was 76. Of these, 56 returned to vote, and completed ballot papers, having been invited to do so pursuant to the exercise mentioned under paragraph 10 above. The number of electors who attended Kinson South polling stations after the problem had been identified, were issued only with parliamentary ballot papers at that time, and later returned and voted in the local election was 65. The number of electors in this category who did not return to vote was 95. Thus, the total number of electors affected by the ballot paper error to the extent that they were effectively disenfranchised was 115 (i.e. 20 plus 95). The 56 returning electors were affected to the extent that their second votes may have been in breach of the rules, and should therefore be disregarded.
13. An analysis has also been undertaken of how the 56 returning electors voted. The result may be encapsulated in this manner. Laurence Fear's 1,224 votes equated to 31.85% of the ballot papers included in the recount on 8^th May. He received 37.5% of the votes on the ballot papers submitted by the 56 returning voters. The corresponding figures for Roger Marley are 30.91% and 33.92%, for Norman Decent are 29.94% and 26.78%, and for Beryl Baxter are 29.15% and 26.78%. Thus, even if the returning electors' votes are not taken into account, they make no material difference to the arithmetical calculations set out at paragraph 2 above.

The First Issue

14. Section 48 of the Representation of the People Act 1983 provides, insofar as is material:

"48 Validity of local elections, and legal costs

(1) No local government election shall be declared invalid by reason of any act or omission of the returning officer or any other person in breach of his official duty in connection with the election or otherwise or rules under section 36 or section 42 above if it appears to the tribunal having cognizance of the question that -

(a) the election was so conducted as to be substantially in accordance with the law as to elections; and

(b) the act or omission did not affect the result.

(2) A local government election, unless questioned by an election petition within the period fixed by law for these proceedings, shall be deemed to have been to all intents a good and valid election."

15. It is common ground that the Returning Officer's relevant acts and omissions were perpetrated in good faith. Their precise definition and ambit is in dispute, hence the second issue arising on this Petition. It is also common ground that, notwithstanding these acts and omissions, however characterised, this local government election was conducted as to be substantially in accordance with the governing law as to elections. Accordingly, the first issue engages sub-paragraph (b) of section 48(1) and not sub- paragraph (a).
16. The Petitioner accepts that it must appear to this Court that the Returning Officer's acts and omissions did not affect the result of the election as regards Laurence Fear. Even if all 115 of the affected electors had voted for the Petitioner, and none had voted for him, she would still have fallen short of the votes he had already accumulated. It follows that the election of Mr Fear cannot be declared invalid, and for this straightforward reason he did not attend the hearing to defend a position which could not be assailed.
17. However, the basic arithmetic does not lead to the same ineluctable conclusion as regards Norman Decent and Roger Marley. This is because the gap Ms Baxter needs to make up is less than 115. That said, assuming for the purposes of argument that a statistical or psephological approach is appropriate (and this is the very issue that arises for decision), I accept the submission of Mr Francis Hoar for the Second and Third Respondents that it is unlikely that the 115 affected electors would have voted in such a way as to lift the Petitioner to third place. Not merely (on this footing) may reasonably robust inferences be drawn from the voting behaviour of the 56 returning voters - who were widely distributed across the Ward, and voted in a broadly similar manner to the much larger cohort of original voters - sub-paragraphs (4) and (5) of paragraph 3 of Mr Hoar's well-reasoned Skeleton Argument makes these telling points:

"For Cllr Decent's election to have been affected by the error Ms Baxter would have to have obtained 65 more votes than him among the 115 returning voters. It is important to consider that some of the affected voters could have cast a vote for Cllr Decent, either instead of or (given each voter had three votes) as well as Ms Baxter. Even on the strained assumption that none of the affected voters would have voted for Cllr Decent at all, Ms Baxter would still have required an advantage of 57.39% of the votes (as opposed to the 1.59% fewer votes she received overall and the identical number of votes she received among the returning voters); this does not change if the votes of the 56 returning voters are removed, as Cllr Decent and Ms Baxter received 15 votes each from those voters.

For Cllr Marley's election to have been affected, he would have had to have polled lower than both Ms Baxter and Cllr Decent and both the following would have to have occurred:

(a) Ms Baxter would have had to have obtained 106 votes out of 115 more than Cllr Marley (an advantage of 92.17%, as opposed to the 2.76% fewer votes she received overall and the 7.14% fewer votes received from returning voters) (or 102 votes more, an advantage of 88.70%, if the 56 votes are disregarded, Cllr Marley having received four more votes than Ms Baxter amongst those voters); and

(b) Cllr Decent would have to have obtained 41 votes out of 115 more than Cllr Marley (an advantage of 33.65%, as opposed to the 1.7% fewer votes Cllr Decent received than Cllr Marley overall and the 7.14% fewer votes received from returning voters) (or 37 votes more, an advantage of 32%, if the 56 votes are disregarded, Cllr Marley having received four more votes than Cllr Decent amongst those voters)."

18. It follows that on a strong preponderance of probabilities the result would have been the same had the acts and omissions in question not occurred. This conclusion may obviously be expressed with an even higher degree of confidence as regards Mr Marley.
19. However, these conclusions may only be drawn if the stated premise is correct, namely that a statistical or psephological approach is appropriate. It is hotly disputed. In these circumstances, the issue for this Court may be formulated in these terms: for the purposes of section 48 of the 1983 Act, is it permissible for the Court to undertake an inferential analysis of how disenfranchised electors would or might have voted?
20. There is no decided authority which directly answers this question. Accordingly, recourse must be had to the language and purpose of section 48 of the 1983 Act, and to first principles.
21. The leading authority on the correct, general approach to this provision remains the decision of the Court of Appeal in Morgan v Simpson [1975] 1 QB 151. The provision under scrutiny there was section 37(1) of the Representation of the People Act 1949, but it was in identical terms. According to Lord Denning M.R., the section should be read in such a way as to confer a positive obligation on the Court to declare an election invalid if it appears that the election was not concluded so as to be substantially in accordance with the law, or that the act or omissions did affect the result. For Stephenson and Lawton LJJ, the section should retain its negative form, but the conjunction "and" should not be read disjunctively. However, the practical import of these textual considerations dwindles to next to nothing in view of the twin considerations that the parties are agreed that there is no formal burden of proof, and as Stephenson LJ's observed:

"For the negative form of the section provides that both substantial compliance with the law and no effect upon the result are required in conjunction to save breaches of duty or of the rules from avoiding an election, as is pointed out in the judgments of Lord Denning M.R and Lawton LJ" (at 167E/F)

22. In both Edgell v Glover [2003] EWHC 2566 (QB) and Considine v Didrichsen [2004] EWHC 2711 (QB) this Court proceeded on the basis that the reasoning of Lord Denning M.R. and Stephenson LJ could be synthesised into the following two statements of principle:

"(1) Section 48 can be translated and understood as creating a positive duty with the consequence that an election must be declared invalid by reason of any act or omission of the returning officer if it appears that the election was not so conducted as to be substantially in accordance with the law as to elections or that the act or omission did not affect the result.

(2) The negative form of the section means that both substantial compliance with the law and no effect upon the result are required to save breaches of duty or the rules from voiding the election." (per Newman J in Edgell at paragraph 23)

23. In all the relevant jurisprudence in this jurisdiction, whereas courts have been prepared to analyse how, if at all, the relevant act or omission has affected the number of votes which would or might have been cast, there is no example of courts being disposed to analyse how (i.e. for whom) these hypothetical votes would or might have been distributed. On my understanding, on no previous occasion has the court been asked to undertake such an exercise.
24. Mr Hoar drew our attention to the decision of the High Court of Justice in Northern Ireland, in Re The Parliamentary Election for Fermanagh and South Tyrone [unreported, 19^th October 2001]. In that case, a polling station was kept open for up to 15 minutes beyond the appointed time. Carswell LCJ performed some rough-and- ready calculations of the number of voting papers which could have been issued during this period, and concluded that it could not have been materially more than thirty - still well short of the successful candidate's majority of 53 votes. Thus:

"We do not need to determine the matter by resort to the burden of proof, for we are affirmatively satisfied on the balance of probabilities that materially fewer than 53 voting papers were issued after 10pm on polling day. We therefore hold that the breach of the regulations did not affect the result of the election."

25. Mr Hoar's core submission is that the same analytical, probabilistic approach may be applied to the present situation, with the same outcome. He is fortified in that submission by his compelling analysis of the figures (see paragraph 17 above), as well as by the decision of the Court of Appeal of the East Caribbean States in Ouinn- Leandro v Jonas [2010] 78 WIR 216, where a close scrutiny of voter distribution and voting patterns was undertaken in order to draw inferential conclusions as to how votes would have been cast (see, in particular, paragraphs 181, 191-192 and 201 of the judgment of Rawlins CJ). However, this authority is not persuasive, still less binding, in this jurisdiction, and it appears that no contrary argument was advanced.
26. Mr Gavin Millar QC for the Petitioner, and Mr Timothy Straker QC for the Returning Officer, advanced broadly similar submissions to the effect that for reasons of statutory language, and on grounds of principle, it was simply impermissible notionally to penetrate the lock and seal of the ballot box by speculating how excluded voters would or might have cast their ballots at any particular time.
27. In my judgment, the Petitioner's case on this issue is correct. Does it appear to this Court that these acts and omissions affected the result? My initial reaction to the statutory language was to focus on the word "appears" in section 48, the use of which tends to suggest that the approach should not be overly analytical and should not travel deeper than the external or superficial appearance of the matter. However, I am persuaded by Mr Millar's argument that the whole of the relevant phrase should be considered, including the verb "affect In the context of the respective positions of the Petitioner, the Second Respondent and the Third Respondent, it appears that the result has been affected because 115 electors have been effectively disenfranchised by the Fourth Respondent's acts and omissions. It is, of course, true that both these Respondents could not come fourth in the poll, but the role of the court is not to attempt to reach any conclusion as to the hypothetical result, absent the relevant acts and omissions having taken place. This would be to enter terra incognita, namely the very territory which in terms of principle and policy must be regarded as constitutionally sacred. By enacting section 48 of the 1983 Act, Parliament cannot be treated as somehow empowering the judicial arm of Government to peer into the voting booth, whether by drawing informed, probabilistic inferences or otherwise.
28. This conclusion supports the delicate constitutional balance which clearly exists in this domain, and achieves practical and legal certainty. If Mr Hoar's exercise were permissible, it would have to be undertaken in all situations where similar problems were created, even if the numbers were more finely balanced. It is no answer to this objection to say that the facts of this instant case are particularly strong. Furthermore, there would in principle be no constraint on the type of evidence the court might receive: e.g. psephological (on a micro or macro level), geographical and behavioural. In my view, it seems obvious that Parliament could not have intended to mandate such a potentially far-reaching, penetrating and invidious level of inquiry by the judiciary.
29. The position is clearly different as regards Mr Fear, because the basic arithmetic yields the same result in his case, however the figures may be finessed. This is not a question of drawing inferences but of simple subtraction; the gap to be bridged is more than 115. Nor can the present case be equated with Re The Parliamentary Election for Fermanagh and South Tyrone, being probably the most helpful authority from Mr Hoar's perspective, because the common sense conclusions of a deductive nature that the Court was prepared to make were entirely straightforward, and did not go beyond a simple head count.
30. There are dicta in Edgell which at least on one reading might suggest that the court has a deeper role. At paragraph 29 of his judgment, Newman J said that the words "it appears" means something equivalent to "the court must make up its mind on the evidence", without it seems applying rigid evidential considerations or formally applying a standard of proof. This phraseology is capable of lending some support for Mr Hoar's case, but ultimately it is only limited. The facts of Edgell were very different, and it is not an example of an instance of the court being prepared to draw inferences as to how missing votes would or might have been cast.
31. There are many situations where courts are prepared to draw inferences from evidence of a statistical nature, and I have in mind in particular the field of personal injuries' litigation. However, in those domains the policy considerations are rather different, not least because no quasi-constitutional issues arise. In Sienkiewicz v Grief (UK) Ltd [2011] 2 AC 229 the Supreme Court stated that epidemiological evidence should be deployed, if at all, with considerable care, because it is inherently generic. In the present context, the practical and policy arguments militating against the use of this type of evidence appear overwhelming. Voting behaviour can be extremely unpredictable across a Ward, and may depend on factors about which it is difficult to be precise. Ultimately, the policy of section 48(1) is that an investigation of likely voting behaviour is tantamount to an exercise in pure speculation, and must be avoided.
32. This Petition must succeed on the basis that it appears to the Court that the acts and omissions for which the Fourth Respondent is responsible, effectively disenfranchising 115 electors, affected the result of the election of the Second and Third Respondents.
33. Strictly speaking there is no need to consider the second issue which arises. However, we heard detailed argument on it, and given the importance of the point and that this case may go further, it is appropriate that it be addressed.

The Second Issue

34. In relation to this second issue, the fault lines between the parties are differently distributed. Mr Hoar sided with Mr Straker, and it was only Mr Millar who argued that the ballots of the 56 returning electors should not have been included.
35. The legal backdrop to this issue is the relevant provisions of the Local Elections (Principal Areas) (England and Wales) Rules 2006 [2006 S.I. No. 3304] ("LEPAR"), in particular Rules 15, 16, 33 and 35 of Schedule 3. LEPAR contains a specific and detailed set of mandatory rules governing the method of poll, the form of the ballot paper, the questions to be put to voters, and the voting procedure.
36. I agree with Mr Millar that it is necessary to be specific as to the breaches of LEPAR which were perpetrated by the Returning Officer in this case.
37. In my judgment, these breaches comprise:

(i) issuing an erroneous or invalid ballot paper to 76 electors;

(ii) failing to issue a ballot paper to the 160 electors who applied to vote during the interregnum;

(iii) issuing the 56 returning electors with a second ballot paper.

38. Items (i) and (ii) are common ground, but item (iii) is in dispute and needs to be addressed.
39. Mr Millar submitted that there is no provision under LEPAR, which is a comprehensive and self-contained code of rules, empowering a Returning Officer to cause or countenance a preliminary investigation to be undertaken, and a "door knocking team" to be despatched. In my view, this submission is clearly correct, but it does not strike at the real vice of what happened here. Although the 56 returning electors would not have returned to the polling stations in the absence of these affirmative, purportedly curative steps, the real point under LEPAR is whether there exists any power to issue a second ballot paper in these circumstances. If there was, the taking of these pragmatic steps would make no difference to the legal analysis.
40. Mr Millar's submissions have two fundamental ramifications: first, that there is no "self-help" remedy under the regulatory scheme; and, secondly, that these acts and omissions must be dealt with, if at all, through the judicial process. Mr Millar accepted that the corollary of his argument was that it must be deemed to be the policy of LEPAR to fix the voter with responsibility for examining the ballot paper for its correctness.
41. Mr Straker's brief and attractive riposte to Mr Millar's technical argument is that these 56 returning electors had never cast a legally valid vote "first time round" because the ballot paper was non-compliant with LEPAR, from which it followed that their second vote was in fact their first legally cognisable ballot. It was as if, first time round, these electors had been given a ballot paper for Bourneville.
42. The parties are in complete agreement that on the first occasion the relevant electors were not given a valid ballot paper. For Mr Millar, this is one of the "acts or omissions" founding his section 48 complaint; for Mr Straker, this would have been an act or omission falling in that category were it not for what happened subsequently. Thus, Mr Straker's submission has the uncomfortable appearance of seeking to define the legal character of what occurred by the happenstance of subsequent executive action by the Returning Officer.
43. Furthermore, Mr Straker's submission is based on the twin premises that the first ballot cast may be deemed to have no legal effect, and the Returning Officer is effectively empowered to say so. Mr Millar hinted that the analogy with public law may not be complete, but in my view Viscount Radcliffe's famous dictum in Smith v East Elloe RPC [1956] AC 736 (at 769) provides a valuable analysis in the present situation:

"An order, even if not made in good faith, is still an act capable of legal consequences. It bears no brand of invalidity upon its forehead. Unless the necessary proceedings are taken at law to establish the cause of invalidity and to get it quashed or otherwise upset, it will remain as effective for its ostensible purpose as the most impeccable of orders."

44. There are numerous judicial utterances of the highest authority to broadly similar effect. The old void/voidable distinction is, of course, no longer regarded by public lawyers as a helpful analytical or juridical tool. The more parsimonious and straightforward analysis is that the ballots of the 76 electors, all cast on invalid ballot papers, fell under the provisions of LEPAR nonetheless to be taken into account until an issue was raised as to their validity (see, further, the express wording of s.48(2)). These ballots were invalid, and could (and should) be adjudicated as such by the Returning Officer at the count; or, if necessary, upon an Election Petition. However, LEPAR contains no mechanism for correcting errors once the ballot paper has been folded and placed into the ballot box. This is treated by the rules as a legally irrevocable act - at least until the poll closes and the ballots are scrutinised.
45. I agree with Mr Millar that there is no provision in LEPAR for errors of any sort to be corrected by the delivery of a second ballot paper to the voter. Indeed, the tenor of Rules 33(l)(b) and 35(l)(c) is to the contrary effect, because these provisions are predicated on there being one vote, not two. The answer to this submission is not to be found in interpolating the adjective "valid", or the adverb "validly", in the appropriate place in the rules, since this is entirely question-begging.
46. The upshot is that LEPAR, properly construed and applied, leaves no room for discretionary decision-making by returning officers, let alone for "self-help" remedies which entail ascertaining who has voted incorrectly, and then making contact with them. However well-intentioned, what happened here had the tendency to undermine rather than to safeguard the integrity and secrecy of the whole voting process: the philosophy of LEPAR is to regulate and constrain contact between presiding officers (or their agents) and voters, and to postpone dealing with errors and irregularities until after the closing of the polls.
47. I accept that the consequence of Mr Millar's submission being correct is that voters only had one bite at the metaphorical cherry in circumstances where they could have no responsibility for the initial printing error. However, although all 76 voters appear not to have spotted the mistake, each was responsible for ensuring that his or her individual ballot paper could be and was completed correctly. The delays which would have flowed from their recognising the error, although constituting an act or omission under the Rules, would have been highly unlikely to have affected the result.
48. In short, I am driven to conclude that Mr Millar's submissions under this rubric are well-founded, and I would uphold this Petition on the second issue as well as the first.

MR JUSTICE WILKIE

49. I agree.