Nairobi — The Supreme Court's judgment regarding Kenya's 2013 election leaves many questions unanswered and provides little motivation for electoral improvements in the future.
The Supreme Court of Kenya has released its detailed judgment on the landmark case challenging the 2013 presidential election. Previously, on 30 March, the Court ruled that the country's elections had been free, fair and in compliance with the constitution; that Uhuru Kenyatta and William Ruto had been legitimately elected; and that rejected ballots should not have been included in calculating the final tallies. On 16 April, the Court finally released a detailed judgment explaining its reasoning. But unfortunately for those looking for answers, the 113-page document made for disappointing reading.
To its credit, the Court recommended an investigation into the Independent Electoral and Boundaries Commission's (IEBC) less-than-transparent procurement of electronic equipment. But on the other hand, the Court failed to address the majority of evidence questioning the integrity of the voter register and the forms used to tally the results, both of which were at the heart of the petitioners' case.
Numbers that don't add up
The voter register is a crucial part of any electoral process. It allows for an analysis of voter turnout, voting patterns and trends over time. It also provides an important check against potential fraud. Without a verifiable register, it is difficult to ensure an electoral process is legitimate.
Indeed, the Court's judgment underscored the importance of the voter register, stating that the rights to vote and be registered as a voter "give life" to every other procedure in the electoral process. It is surprising, then, that the remainder of the document either excused many significant problems with the register or ignored them all together.
The most fundamental question is the following: Why weren't all registered voters, regardless of the availability of their biometric data, included in the legally published register? The Elections Act clearly states that after the close of the public inspection of the register, the IEBC is to combine all changes and publish a notice that the final compilation is complete. How is it then that the IEBC, after having made official the final register, pulled out additional lists of registered voters that were not included in the published version?
This question remains unanswered in the Court's judgment. The judgment says that the petitioners' argument that the presidential election could only have been based on the Biometric Voter Registration element of the Principal Register of Voters is not tenable. But the petitioners did not make that argument. In reality, the petitioners questioned the changes between the provisional and final registers, arguing that any changes that were made to the register should have all been included in the Principal Register at the time of its legal publication. They asked for an explanation as to why this was not the case.
But instead of answering these questions, the Court justifies the need for multiple registers. It explains that multiple registers are permitted by law and outlines how once the register is finalised, it is divided up and distributed to various electoral units, including polling stations, wards, constituencies, counties and diaspora voting centres. But this line of reasoning does not negate the necessity of one overarching register. After all, it is only from that one document that the other registers are derived. This was the petitioners' main argument: that the various registers did not add up to the IEBC's national number of registered voters, as announced on 9 March.
Instead of questioning the fact the numbers don't add up, the Court simply cites the IEBC's explanation which does not clarify anything.
The IEBC went to great lengths to explain why there seemed to be a fluid number of registered voters. The Principal Register, it explained, was not a comprehensive list of eligible voters. It did not include the "Special" register of 36,236 individuals whose biometric information could not be captured, nor did it contain 12 voters who had not been transferred to the main register from a previously used "test" account. Notably, however, this still does not result in a final number of registered voters announced on 9 March: 14,352,545 (from the Principal Register) + 36,236 (from the Special Register) + 12 (from the test account) ≠ 14,352,536.
Double counting and fluid numbers
Responding to questions surrounding additions to the register, which appear to have been made after the close of the voter registry, the Court includes another faulty IEBC explanation.
The Court cites an affidavit from the IEBC, in which the Director of the Voter Registration Programme explained that voters who had inadvertently been assigned to the wrong polling stations were assigned to the correct stations and that this "naturally resulted in a variation between the number of registered voters in the provisional register and the Principal Register". If it was only transfers that were being made, however, the overall number of registered voters should not change. In reality, though, the Principal Register contained 12,509 more voters than the provisional list.
The Commission went on to contend that 30,000 names had to be deleted because they represented double-entries. Of those 30,000, 13,237 were then added to the register. Why were they added if they had been wrongfully registered more than once? The IEBC goes on to say that on 18 February, it was discovered that "several names" had been left out the Principal Register. How many of these names were there? And why is it that these names were discovered so late in the process? Shouldn't the IEBC have been aware of these names before they published the final register?
Even leaving that aside, however, the numbers still do not add up. Subtracting the multiple entries (and re-adding the 13,237) along with the Special Register and the "test" entries still does not result in a number that matches the number announced by the IEBC on March 9: 14,337,399 - 30,000 + 13,237 + 36,236 + 12 ≠ 14,352,536.
Where's the Green Book?
The Court also does not address the petitioners' regional analysis of changes to the voter register. How does the Court resolve the fact that 68,846 voters were added to Central and Rift Valley provinces while 14,122 voters were taken away from Coast and Nyanza provinces?
Furthermore, there is also the issue of the Green Book, which, according to the IEBC, is the only comprehensive list of all registered voters in Kenya, compiled by IEBC agents who manually recorded every single registered voter at the registration centres. If this is true, then shouldn't the Green Book have been the legally published register? Moreover, the IEBC has yet to share the number of registered voters as recorded in the Green Book. Shouldn't this have been the ultimate check on the validity of the register? The number in the Green Book should match the number announced by the IEBC on March 9. Yet the Court does not ask for this number. In fact, in its judgment, the Court goes so far as to say that the Green Book, despite the fact that it was used as a primary document to originate the primary register of voters, is not required to be provided for by law.
At the very least, one should be able to go to the polling station level and add up the number of voters in the principal register and special register to check whether this equals the number in the Green Book. Since the Green Book numbers are not publically available, this is impossible.
But do such changes to the register even matter? The Court seems unconvinced, but consider this: An analysis that eliminates the changes made to the register shows that presidential victor Uhuru Kenyatta received about 21,000 more votes than he otherwise would have and defeated candidate Raila Odinga received about 11,000 fewer votes than otherwise. Given that Kenyatta exceeded the 50% threshold by only around 8,500 votes, it is easy to see what impact the changes to the register could have had.
The Court judgement also fails to substantively comment on its own investigations, which showed several important inconsistencies. Without explaining its reasoning, the Court simply states that 5 out of the 22 polling stations it reviewed showed discrepancies in the number of votes cast, as recorded on Form 34 and Form 36. In fact, there were 6 polling stations for which the numbers differed, including the one in which the comparison could not be done because Form 36 did not contain an entry for the number of votes cast.
Perhaps more importantly, the judgment also did not include even a mention of the fact that the Court found there to have been over 100% voter turnout in 16 of the 22 polling stations when using Form 34 and the Principal Register, and over 100% voter turnout in 18 of the 22 polling stations when using Form 36 and the Principal Register. Even when using the Green Book as the basis for comparison, there were two polling stations that showed more than 100% voter turnout. According to the law, results from polling stations that show more than 100% voter turnout must be cancelled. This did not happen.
The Court's report and the judgment further failed to mention the plethora of other problems with the forms themselves. While the report pointed to ten polling stations for which a Form 34 was missing, observers' reports showed at least 83 missing forms. Aside from the instances of over 100% voter turnout mentioned above, observers also noted nine other Form 34s showing this problem.
The Court's judgment also failed to mention all the forms on which the numbers did not add up; cases in which the forms were not signed by party agents; blank Form 34s; changes made to figures on the forms without an authorising signature; illegible figures on the forms; and forms on which results for candidates are missing and multiple copies, both identical and non-identical, of the same Form 34s.
A disappointing judgment
Overall the, the Supreme Court's detailed judgment is disappointing, mostly for its simple failure to confront the evidence questioning the credibility of the electoral process. When considered in the context of Kenyan political history, this is not just disappointing but negligent.
The report issued by the Independent Review of the Electoral Commission, better known as the Kriegler report, cited so many mathematical errors and inconsistencies on the primary tallying forms that there was no way to know who actually won the 2007 election. It is not hard to see how the process described above mirrors that scenario.
Sadly, the Court's judgment implies that it is acceptable to run a deeply flawed election. The precedent has now been set, and there is little incentive for the IEBC to improve its conduct in the future. Can Kenyans expect this to be the template for all future elections? Only time will tell, but if so, it is a sorry fate for Kenyan democracy.
Seema Shah earned her PhD in Political Science from UCLA. She currently works as a Researcher for AfriCOG in Kenya, with a focus on elections and ethnic violence in the developing world.