The republic of uganda in the supreme court of uganda at kampala

In my view in the phrase “affected the result” the word “result” means not only the result in the sense that a certain candidate won and another candidate lost. The result may be said to be affected if after making adjustments for the effect of proved irregularities the contest seems much closer than it appeared to be when first determined. But when the winning majority is so large that even a substantial reduction still leaves the successful candidate a wide margin, then it cannot be said that the result of the election would be affected by any particular non-compliance of the rules.”

Mr. Walubiri contended that the decision of the Court of Appeal of Tanzania in ATTORNEY-GENERALV KABOUROU (supra) reflects a different view. That contention appeared to be based on the following statement of the Court in its judgment at p.772 d-e of the report:

“….taking into account the principle which underlies the Constitution and the 1985 Act that elections shall be free and fair, we are of the considered opinion that an election which is generally unfree and unfair is not an election at all as invisaqed by the Constitution and the 1985 Act, and consequently anything which renders the elections unfree and/or unfair is in law valid ground for nullification of such purported election. We are further of the considered opinion that any law which seeks to protect unfree and/or unfair elections from nullification would be unconstitutional” (emphasis is added)

That statement however should not be taken out of context.

It is evidently to me that both the original court and the appellate court in coming to their decisions in ATTORNEY-GENERAL VS KABOUROU took the effect of the non-compliance in issue into account. The subject matter of the petition was the result of a by-election. The Petitioner established two major aspects of non-compliance with the principle of “fair election” namely-
(a) that in supporting the candidate of the ruling CCM party, the central government had, with corrupt motive to influence voters, undertaken to repair a road in the constituency in consideration of the constituents voting for him; and

(b) that a government radio gave to the CCM party’s campaign, more air-time, than it gave to other competing parties combined, and had shown in its own broadcasts bias in favour of the CCM party.

In assessing whether the non-compliance affected the result in favour of the ruling party’s candidate, the courts took into account the number of voters that were exposed to those acts which constituted non-compliance. Thus in respect of the undertaking to repair the road, the courts had to be, and were satisfied that, the undertaking was promised by cabinet ministers at very well attended rallies of people in the constituency. And as regards the broadcasts, the Court of Appeal took judicial notice of the common knowledge that many people in rural and urban areas in Tanzania possessed radio sets, and regularly listened to the government radio. From that the Court inferred that a large number of people in the Constituency listened to the broadcasts regarding the by-election. But the Court went further to pose the question: “But did these broadcasts affect the results of the by election in favour of the CCM candidate?” which it answered thus:

“Having examined the contents of various broadcasts and bearing in mind the time tested maxim that information is power, we are bound to conclude to the effect that these broadcasts in favour of CCM must have influenced the by- election results in favour of the CCM candidate.”

I take that to be the ratio decidendi of that case, rather than the opinion that the court expressed about “a generally unfree and unfair election,” which I reproduced earlier. In my considered opinion therefore, that decision is not authority for the proposition that non-compliance with a principle laid down in the Constitution is sufficient ground for annulment of an election without proof that it affected the result of the election.

In IBRAHIM vs. SHAGARI & OTHERS (1985) LRC (Const.) 1, the Supreme Court of Nigeria considered a law, expressed in the negative form, stating when an election is not to be invalidated. It provided that “an election shall not be invalidated by reason of non-compliance with Part II of the Act if it appears to the Court…………that the election was conducted substantially in accordance with provisions of the said Part II and that non-compliance did not affect the result of the election.” Reflecting on that provision, Nnamani J.S.C, said in his judgment at p.21:

“……the court is the sole judge and if it is satisfied that the election has been conducted substantially in accordance with Part II of the Act it will not invalidate it. The wording of section 123 is such that it presumes that there will be some minor breaches of regulations but the election will only be avoided if the non-compliance so resulting and established in court by credible evidence is substantial Further the court will take into account the effect if any which such noncompliance with provisions of Part II of the Electoral Act 1982 has had on the result of the election (from that consideration) the duty to satisfy the court that a particular non-compliance with the provisions of Part II of the Electoral Act which he has averred in his petition, lies on the Petitioner.”

Several relatively recent decisions of the English Courts referred to us were concerned with legislation similarly expressed in negative form, to the effect that an election shall not be declared invalid by reason of breach of duty in connection with the election, or breach of the election rules, “if it appears to the Court that the election was so conducted as to be substantially in accordance with the law as to elections and that the (breach) did not affect the result.”

All those decisions revolved around the effect of the breach on the result of the election. Two of the decisions demonstrate this point very well. MORGAN vs. SIMPSON (1974) QB 344 and GUNN vs. SHARPE (1974) 1 QB 808 were decided by the Divisional Court about the same time. In MORGANS case the petitioners sought annulment of an election on the ground that 44 ballot papers had been rejected as invalid because they were not stamped with the official mark. It was proved that if the 44 ballot papers had not been invalidated for want of the official mark, the petitioner would have won with a majority of 7 votes, instead of the respondent winning with 11 votes as declared. The failure to stamp the ballot papers had therefore affected the result. However, the Divisional Court dismissed the petition, holding that since the election was conducted substantially in accordance with the law as to elections, the fact that a small number of errors had affected the result was not a sufficient reason for declaring the election invalid. The petitioners appealed. While that appeal was pending the Divisional Court on 10th April 1974 gave judgment in GUNN vs. SHARE (supra) which was based on a similar complaint, in that failure to stamp 102 ballot papers with the official mark, had similarly and demonstrably affected the result. The court chose not to follow its decision in MORGAN’ case (supra) but to distinguish it on facts. The difference was that in the case of the 44 ballot papers, the failure to stamp them had been spread over 18 polling stations, whereas in GUNN’S case, out of the 102 unstamped ballot papers from 10 polling stations, 98 ballot papers were from only one polling station, so that over half of the voters who sought to vote at that particular polling station had been disfranchised. On that account, the Divisional Court held the failure to stamp the ballot papers with the official mark, to amount to conduct of an election which was not substantially in accordance with the law as to election. And because it affected the result, the election was held to be invalid. A few months later, the Court of Appeal delivered judgment in the MORGAN vs. SIMPSON appeal, allowing the appeal. The head note of the report in (1975) 1QB 151 reads in part:

“……..the election must be declared invalid for although it had been conducted substantially in accordance with the law as to elections the breach of the rules in omitting to stamp the 44 papers had affected the result....”

In his judgment Lord Denning M.R., commented on the decision of the Divisional Court in GUNN’s (supra). He said at p.164:

“They (Judges in GUNN’s Case) put it on the ground that the election was not so conducted as to be substantially in accordance with the law as to elections. But I think it should have been put on the ground that the mistake did affect ii result of the election.”

In the later case of RUFFLE vs. ROGERS (1982)1 GB 1220, the same problem of omitting to stamp ballot papers with the official mark had led to a declaration of one candidate as elected by majority of two, whereas the result would have been a tie between the two candidates if those ballot papers had been duly stamped. Because the result was thus affected by the failure to stamp the ballot papers, the election was held to be void.

Among the English precedents referred to this Court, I did not come across any which decided to annul an election on ground of a breach of the law or rules which did not affect the result. However in MORGAN vs. SIMPSON. (supra) the Court of Appeal made observations that were supportive of that proposition. Although counsel for the Petitioner in this petition, did not seek to rely on them, the observations helped me in focusing on the interpretation to put on S.58 (6) (a) of the Act. Lord Denning M.R., after reviewing the history of the law on elections, suggested, inter alia, that if an election was conducted so badly that it was not substantially in accordance with the law as to elections, the election is vitiated, irrespective of whether the result was affected or not. Stephenson L.J. interpreting the provision that no election shall be invalidated “if it appears that the election was so conducted as to be substantially in accordance with the law as to elections and that the (breach) did not affect its result” said at p.16l:

“……an election will stand if there have been breaches of the law but they are not substantial or they have not affected the result.

We are not required to read the conjunction “and” disjunctively if we are to give effect to the intention of Parliament………. This construction seems to be in accordance with the common law and common sense and with decisions that an election which is conducted in violation of the principles of an election by ballot is no real election and should be declared void even though it may not or could not have affected the result ………There is no case cited to us which conflicts with the conclusion that such a substantial misconduct of an election must avoid it.”

The common sense view referred to me to, seems to be this: Where an election is substantially conducted in compliance with the law it is not to be invalidated because the noncompliance is not substantial. It should therefore, follow that where the noncompliance was substantial the election ought to be invalidated. That reasoning appears to fit in well with the statutory provision expressed in the negative form, because it has the effect of leaving non-Compliance open- ended as a ground of validating an election, providing only exceptional circumstances when an election will not invalidated despite noncompliance The same reasoning, however, does not appear to suit the statutory provision such as S.58(6) (a), which limits the circumstances when non-compliance will operate as a ground for annulment of the election. The limitation is the effect on the result. To my understanding, for the ground under S.58 (6) (a) to succeed, the Court has to be satisfied on two things, not on only one of the two. The Court has to be satisfied-

that the election was not conducted in accordance with the principles laid down in the provisions of the Act that the noncompliance affected the result of the election in a substantial manner.

There is only one ground set out in paragraph (a) of s.58 (6) of the Act. In my view to hold that an election could be annulled on the ground that non-compliance was substantial would be adding a fresh ground for annulment. It is in that regard that ATTORNEY-GENERAL vs. KABOUROU (supra) is distinguishable. There, the statutory provision was expressed in the negative formular which I have just described. The trial court held, and it was upheld by the Court of Appeal, that the list of grounds for annulment of an election set out in the statute was not exhaustive. Clearly the same cannot be said of S.58 (6) of the Act, where the list is expressly made exhaustive by the phrase that (the election) “shall only be annulled on any of the following grounds.”

Those grounds defined in paragraphs (a) (b) and (c) do not include “substantial non-compliance” with provisions or principles in the Act. Parliament, either deliberately or inadvertently did not include it. I incline to the view that it was deliberately omitted. I am strengthened in that view by the difference in wording in paragraph (a) from that in paragraph (b) and (C). Whereas each of the latter two grounds (which relate to qualification and conduct by a candidate) is sufficient irrespective of the effect on the result, the former is conditional on the effect it has on the result of the election. This Court cannot include the ground of substantial non-compliance either by removing or otherwise altering the limitation in paragraph (a), or by adding it to the list as a fourth ground.

I am of course mindful of what Stephenson L.J. called an election “which is no real election”, in MAROGAN’S case; and also of what was described in ATTORNEY-GENERAL vs. KABOUROU by the Court of Appeal of Tanzania as “a purported election that is not an election at all as invisaged by the Constitution.” That, however, was not the scenario presented in the instant petition. What was presented in this petition was an election invisaged in the Constitution, which, however was faulted by failure to abide by the law and principles which govern the proper conduct of elections. Moreover in the Ugandan judicial context, the appropriate court to determine and declare that a purported Presidential election was not an election invisaged by the Constitution, and is therefore, unconstitutional, would, in my considered opinion, be the Constitutional Court, moved under Art.137 of the Constitution. Similarly, whilst I respectfully share the opinion expressed in ATTORNEY-GENERAL VS KABOUROU, that “any law which seeks to protect unfree and/or unfair elections from nullification would be unconstitutional”, in Uganda it is the same Constitutional Court that has the competence to declare that by so seeking to protect such election, the law is inconsistent with, or in contravention of, the Constitution and therefore unconstitutional.

I now turn to the reasons for the conclusion that, in the instant petition, the non-compliance found to have occurred, was not proved to have affected the result of the election in a substantial manner. I have already indicated the reasons why I rejected the proposition that it was not necessary to prove the effect of non-compliance on the election result. I had, therefore, to consider the alternative, namely whether in absence of direct proof, the effect could have been inferred from the proved non-compliance. In my view, for the Petitioner to succeed that way, the Court would have to find that the only irresistible inference to be drawn from the evidence on the several aspects that constituted non-compliance is that the non-compliance affected the result of the election in a substantial manner. I was not convinced that the Court could so find in the instant case. I will comment on each aspect separately and finally on the overall non-compliance, in relation to the election result.

Failure to comply with ss.28 and 32(5) of the Act:

The Petitioner did not adduce evidence on the effect of non-compliance with S.28 of the Act. However, by way of arguments it was contended that because the Petitioner was not made aware of all the polling stations in time, he was deprived of the opportunity to appoint agents at the new polling stations. Even that argument was not seriously canvassed. No evidence was brought to show, in actual terms, at how many polling stations the Petitioner was not represented because he was not notified, or was belatedly notified of them. In my view that was necessary before inviting the Court to infer that the omission affected the result. But an attempt was made through Mukasa David Bulonge, Head of Election Management in the Petitioner’s Task Force. He deponed in a supplementary affidavit that he selected as sample from tally sheets from Mbarara, Bushenyi, Kamuli and Pader districts, results of twenty polling stations which were not gazetted. He concluded from those results that the l Respondent “received a far higher percentage of the votes cast in the newly created polling stations (ranging between 72% and 100/%) than he did nationwide (69.3%).” Even if that conclusion had been taken as correct, and there were several reasons to suggest that it was not accurately representative of 1,176 new polling stations, it is not a fact from which any reasonable court would infer, conclusively that if the polling stations had been gazetted and the Petitioner was represented thereat by polling agents, the voting, and consequently the result would have been different. Nor can a court of law infer from that conclusion that all or any votes cast in those polling stations were irregularly cast, as was submitted by Mr. Mbabazi in respect of what he called “ghost” polling stations. I venture to say that the evidence would have carried more weight if Mukasa David Bulonge had focused on new polling stations in respect of which there was evidence of irregular voting.

Similarly I did not find evidence that proved directly, or from which it could be inferred, that failure on the 2nd Respondent’s part to avail copies of voters’ roll for use by the Petitioner’s agents as required under s.32(5) affected, or could have affected, the result of the election.

Effect of intimidation:

Intimidation in the electoral process may, as I have indicated manifest in diverse forms, from acts of violence and harassment, to invasion of secrecy of voting. It can affect the result in two ways. In one way, it may prevent fair competition between or among the contesting candidates. Secondly it may cause voters not to vote according to their free will, either by compelling them not to vote or to vote for a candidate they do not freely choose. The extent, to which fair competition is prevented by intimidation, can be proved by direct evidence. Similarly proof of the extent to which voters are prevented from voting because of intimidation, can be by direct evidence. However the Court cannot demand, and does not expect, proof of the effect of intimidation as would require a voter to disclose how, or for whom, he or she voted, as that would be a violation of the principle of secret ballot. Learned counsel for the Petitioner contended that in those circumstances, it is not possible to quantify the effect of intimidation in terms of figures and numbers of voters or votes. I agreed with that contention in as much as there is bound to be invisible effect of intimidation which is not seen or easily perceived. That, however, would not be reason for a court to readily conclude that any amount of intimidation affected the result. Nor, conversely, should it be ground for the court to lightly dismiss evidence of intimidation because its effect on the final result is not established in figures and numbers. Ultimately, what the court must determine judicially is whether in view of the proved intimidation, the election result is a consequence of intimidation, or whether despite the intimidation the result is a choice made freely by the majority of the voters. Needless to say therefore, that proof of the level and extent of the intimidation is very material for that determination. On the one hand the intimidation may be so grave and extended to such large proportion of the electorate that it becomes compelling or irresistible to infer that it affected the result. On the other hand, the intimidation may have been such as would not compel an average voter to act against his or her will, or may have been confined to a relatively small proportion of the electorate. In such eventuality the court would not infer that the intimidation affected the result, except where the contest was so close that the court is led to the conclusion that this balance in the context was swung or tilted by the intimidation.

This was very clearly pointed out by Bramwell B. in the North Durhan Case 2 O’M & H. 152 at p.156 where after reference to intimidation perpetrated by a candidate or his agent he went on to say:-

“…….there is another intimidation that has been called a common law intimidation, and it applies to a case where the intimidation is of such a character, so general and extensive in its operation that it cannot be said that the polling was a fair representation of the opinion of the Constituency. If the intimidation was local or partial, for instance, if in this case it had been limited to one district as Hetton is, I have no doubt that in that case it would have been wrong to have set aside this election, because one could have seen to demonstration that the result could not possibly have been brought about by intimidation, and that the result would not have been different if it had not existed. I do not mean the result of the polling in that particular district, but the general result of the majority for the Respondents. But where it is of such a general character that the result might have been affected in my judgment it is no part of the duty of Judge to enter into a kind of scrutiny to see whether possibly, or probably even, or as a matter of conclusion upon the evidence if that intimidation had not existed, the result would have been different.”

And later he observed:

“….I think if it were otherwise, and if one were told that partial intimidation would avoid an election, although it were certain that it had affected the result of the election, the consequence would be that a few mischievous persons might upset every election-”

I respectively agree but with one necessary qualification in view of the standard proof required under S.58 (6) of the Act. The Court must be satisfied as matter of conclusion upon the evidence that the intimidation did affect the result in a substantial manner.

In this petition, as I have already indicated, intimidation was proved to my satisfaction to have occurred in a few areas. There was however scanty direct evidence on the effect the intimidation had on the voting. Only a few deponents averred that they did not vote, either because they were intimidated or because they were physically prevented as they were unlawfully detained. The evidence of intimidation in Rukungiri, however, was of such gravity and so generalised that I concluded it must have had effect on the voting there. It was not directed at the Petitioner and his agents alone, but it also reached out to the supporters and other voters who turned up to meet the Petitioner or to attend his rally, and the rallies addressed by his campaign agents, and were violently dispersed by soldiers. The violence and harassment led to personal injuries and deprivation of liberty of the victims and must have caused apprehension and fear of the same by those who witnessed the occurrences as well as the close neighbours who heard of it. As late as just over a week prior to polling day, a person was shot dead in a violence that erupted after a rally addressed by the Petitioner in Rukungiri Town. On polling day the Petitioner’s polling agents were openly harassed out of some polling stations they were supposed to oversee.

In Kamwenge District intimidation was more individualised. Apart from the incident when the Petitioner went to hold a rally in Kamwenge Town, for the rest of the campaign period harassment was against the Petitioner’s agents and known supporters only. Some suffered physical assault, unlawful detention and damage to property while others were subjected to threats.

The evidence adduced showed that on polling day there were two forms of intimidation generally. One was the harassment of some of the Petitioner’s polling agents. Significantly apart from that this was no evidence of physical intimidation directed at voters on polling day. The other form of intimidation on polling day was the interference with the right of the voter to vote in secret. The latter form of intimidation was manifested not only in the two districts but also in few other places, particularly in the special polling stations for soldiers.