Following the historic, unprecedented decision by the Supreme Court to nullify Kenya’s presidential election, which international election observers had largely commended, the credibility of observers, and the practice of observation, was seriously damaged. In this article, Aly Verjee identifies six lessons for international election observers ahead of the re-run of polls planned for 26 October.
It might be shorter to list those who were not critical of international election observers in Kenya. Following the historic decision of Kenya’s Supreme Court to nullify the August 8 presidential elections, international election observation missions (EOMs) have been pummelled.
Leading human rights activist Maina Kiai said: ‘There is something incredibly wrong with the way international election observation is happening…the international observers came in for a short time, saw something glossy, something shiny on the day of election and then they make a decision. That is patently shallow.’
Muthoni Wanyeki, formerly of Amnesty International’s Nairobi office, opined: ‘there’s a problem with the election observer industry. They focus too much on the pre-electoral process and the process of voting.’
John Kerry, who co-led the Carter Center’s EOM, came in for particular opprobrium. One opposition official snorted moments after the Supreme Court ruling that Kerry had come to Kenya to see baby elephants at Sheldrick’s. (This is false: Kerry did not visit the elephant sanctuary.) But the image is beguiling: privileged foreigners using the pretext of an election to be tourists in a country with no shortage of attractions. (The same sentiment was offered by pundits in Uganda last year.)
Criticism of observation is not new, nor limited to Kenya. Earlier this year, Stephen Chan argued ‘electoral observation, while not worthless, has limited worth…the observers are not so much the intrepid pioneers in the dark as well-accustomed members of a bus tour party.’
Writing about Uganda’s 2016 elections, Nic Cheeseman, Gabrielle Lynch and Justin Willis argued observers had pulled their punches, and went on to conclude: ‘international observers have frequently refused to give elections the evaluations they deserve for fear of offending incumbent governments and triggering political instability — and, also, it would seem, because they apply lower standards on the continent.’
Shortly after the ruling of the Supreme Court in Kenya, Cheeseman said: ‘we need to rethink the way we are doing election monitoring…whether we can make the international monitoring system more dynamic, more responsive, more cutting-edge and more digital.’
This article is not an unconditional, uncritical defence of international EOMs in Kenya in 2017. Instead, I want to reflect on what can be learned from Kenya. Can any hard truths emerge for international observers as they prepare to observe again in Kenya, and indeed elsewhere? By dismissing the value of observation, are critics missing something essential? I would argue yes to both questions.
Not all observers are created equal
Any critique of observation should first acknowledge the great diversity amongst EOMs. It therefore follows that there may be different solutions for improving the practice of election observation, depending on the organisation in question.
For an African election, there are the regional intergovernmental organisations, such as the East African Community (EAC), the Intergovernmental Authority on Development (IGAD) and the International Conference on the Great Lakes Region (ICGLR), which observe elections in their member states and, with few exceptions, tend to shy away from criticising the elections organised by their neighbours. These missions tend to be led by foreign ministers or senior ambassadors.
There is the African Union (AU), which is bound by treaty to observe elections in its member states, but is hampered by a lack of resources. The AU deploys relatively few observers, and has only recently begun to send observers for longer periods. Typically, AU missions are headed by a former, respected head of state, often from a country outside the immediate region, and thus thought to be less constrained in what he (for it is almost always a man) might be willing to say.
The European Union (EU) observes many, but not all, African elections, and usually has the most money to organise its missions, allowing it the biggest footprint and the best logistics, and sometimes, but not always, the best coverage of a given country. However, it tends to have the weakest leadership profile, with missions headed by relatively unknown European parliamentarians.
The Commonwealth observes elections in its member states, with a small but typically high-level delegation of prominent figures, often headed by a former head of state or senior minister.
The Carter Center, the National Democratic Institute (NDI) and the International Republican Institute (IRI) are non-governmental, non-profit organisations that observe elections globally, and tend to appoint prominent retired political leaders to head their missions. Democracy International (DI) is a for-profit company that organises EOMs, often in countries where NDI or IRI are not active or have determined that they cannot deploy an EOM.
Outside of the Organisation for Security and Cooperation in Europe (OSCE) region, The Carter Center is the only international observer group that conducts true long-term observation, sometimes years, but more often months, in advance of election day. The European Union also claims it observes elections for the long-term, but defines a long-term presence in more constrained terms: ‘ideally, EU observers will be present from the opening of the campaign to the announcement of final results and the adjudication of any election-related complaints.’
Generally, the Carter Center arrives earlier and stays longer than anybody else, but usually has a smaller team on the ground than the EU.
The take away from this lengthy description of the different observer groups should be that lumping all international EOMs together, as much as they suggest they do things the same way, is overly simplistic. Missions can rarely be authoritative about all aspects of an election, nor should they be. Missions deploy at different times and vary considerably in their composition and expertise. EOM methodologies vary in their rigour and depth. Individual observers, most of whom are volunteers, vary in their professionalism and commitment.
The pre-election period still matters
The criticism of observers focusing on the pre-election period ignores the fact that in most countries, elections that lack credibility (as in Uganda) are set on rotten foundations, which deteriorate long before election day.
Whether it is insufficient respect for fundamental freedoms, a legally problematic electoral framework, problems with voter registration, restrictions on the campaign, or improper use of state resources by the incumbent, all these problems occur before a voter goes anywhere near a polling station.
It has taken much effort and many years to shift the practice of international observation away from fetishizing the ballot. In countries where electoral practice is less institutionalized, there can still be a tendency for EOMs to comment on the stoic patience of voters and the long lines they endured to vote, as if the patience of voters was an indicator of the election’s credibility. If an election is a process, then the process before election day is as important as the process after election day.
Arguably, in Kenya, the pre-election period should have been even more comprehensively assessed by observers. The politics of the party primaries, which most observers were not present to see, were and are critical to understanding local politics. Controversies over voter registration, ballot procurement, and earlier electoral preparations cannot be assessed if observers are not present to assess them.
When the politicians arrive: what they say, and how EOMs are managed
For most of the life of an EOM, the mission is run by career electoral specialists. These are a mix of consultants, country office staff and headquarters personnel from Atlanta (The Carter Center), Brussels (the EU), London (The Commonwealth) or Washington (NDI, IRI, DI). There is great overlap in these professional circles and it is not uncommon for someone who worked for the EU in one election to be working for NDI in the next.
Most of these professionals are well versed in the practice of elections. And yet, at the critical moment where the judgments made are the ones that are heard by the voters and by the media, African EOMs are entrusted to the short-term arrivals: the high-level politicians invited to lead the mission.
These prominent figures, former presidents and prime ministers and political grandees, add credibility to an EOM: they have themselves been democratically elected, peacefully retired from office, and in some cases (e.g. Kerry, Goodluck Jonathan) been defeated in especially humbling circumstances, in prominent races.
But it does not follow that such politicians are always well grounded in the practice of electoral management: their background in the democratic process is running in elections, not the running of elections.
Nor, more critically from the perspective of EOM methodology, are political figures necessarily the best placed to offer technical judgments, even if they subscribe to them. Nuances matter, and politicians will speak as politicians do. It is a double-edged sword to have leadership of this kind, and it depends on what the objectives are: to deliver a technical assessment, or to give a political verdict. Nuances matter. No matter the caveats and reservations in an EOM’s preliminary statement, it is the soundbite delivered by the mission leadership that is remembered and reported.
Flexibility in the timing of the first statement
EOMs prize themselves on methodological consistency. One EOM convention is that a preliminary statement of findings is made 48 hours after the polls close. As Holly Ruthrauff notes, ‘it is already difficult for international observers to wait 48 hours to announce their findings - as is common practice – as national and international media, as well as the public, are waiting impatiently for their assessments.’
And while there is certainly eagerness to hear assessments, perhaps, in some elections, there is cause to keep people waiting a bit longer, especially in countries where the tabulation and transmission of results is expected or projected to be particularly vulnerable. For while a later statement or final report may nuance findings considerably, by then it is too late to change the first impression the EOM has created.
If another day or two is needed to deliver a more comprehensive assessment, then methodological consistency should give way to what best serves the process in the country in question. In the grand scheme of things, getting the judgment right is better than getting the judgment quickly.
Back to basics: assessing vote counts, tabulation and transmission
The majority opinion of the Supreme Court of Kenya states: ‘no evidence has been placed before us to suggest that…vote counting [was] not conducted in accordance with the law.’
The Court’s decision turns on the paragraph immediately preceding:
“We find and hold…that the [the chair of the electoral commission], declared the final results for the election of the president, before the [electoral commission] had received all the results…from all the…polling stations contrary to the Constitution and the applicable electoral law….We further find that the [electoral commission] in disregard of the provisions of Section 39 (1C), of the Elections Act, either failed, or neglected to electronically transmit, in the prescribed form, the tabulated results of an election of the president, from many polling stations to the National Tallying Centre.”
“And that therefore:
“whether or not [Uhuru Kenyatta] received a large number of votes becomes irrelevant…
“For the above reasons, we find that the 2017 presidential election was not conducted in accordance with the principles laid down in the Constitution and the written law on elections in that it was, inter alia, neither transparent nor verifiable. On that ground alone…we have no choice but to nullify it.”
These failures largely occurred out of sight of the observers, who lack both the technical expertise and a systematic statistical methodology to interrogate such processes. The question, therefore, is what observers, can, or should, attempt in future, while not downplaying their other tasks. Indeed, the Court commented:
“…we must also state that whereas the role of observers and their interim reports were heavily relied upon by the respondents as evidence that the electoral process was free and fair, the evidence before us points to the fact that hardly any of the observers interrogated the process beyond counting and tallying at the polling stations. The interim reports cannot therefore be used to authenticate the transmission and eventual declaration of results.”
In the 2016 Zambian elections, the EU EOM, in which I participated, conducted a systematic assessment of the deviations in turnout between presidential and parliamentary elections. The findings raised questions about the credibility of the results in some constituencies, but these findings were only published in the mission’s final report, months after the election.
International election observation is necessarily sample based. Sometimes, the sample is pre-determined and, sometimes, it is left to individual observer teams. But compared to the observation of polling, in which a pair of observers may see a dozen polling stations over the course of the day, a sample based approach means the observation of counting, where only one count is observed, will always be statistically insignificant, given that observers will always be few and polling stations always many.
The observation of tabulation at constituency centres could be statistically significant, depending on the proportion of centres observed. It may therefore make sense for EOMs to focus more on tabulation and the subsequent transmission, rather than on counting.
EOMs could use statistical methods to assess election results, turnout and invalid or rejected ballots more systematically, as social scientists have demonstrated (see, for example, Mebane’s second digit test on Mexico; Carriquiry’s survey of statistical analyses in Venezuela, and my critique of the elections in South Kordofan in 2011.)
In his 2017 critique, Chan makes several useful suggestions, although parallel vote tabulations are beyond the capacity of international observers (the use of this method by domestic observers is discussed below.) But international observers cannot expect to forensically analyse all election data; statistical investigation must be informed by political context, be selective, timely and accurate. Replacing observers with statisticians would miss the point; but that said, the techniques employed by an EOM must adapt and evolve.
Protecting the civic space of observation; in defence of domestic observers
Lost in the furore around international EOMs is consideration of the implications criticism of EOM methodology has for domestic observers. These groups, usually drawn from national civil society, borrow heavily from the techniques of international EOMs. What they lack in mobility they make up for in numbers, and they cover many more polling places than international observers can hope to reach. Domestic observers are the only entities that can mount parallel vote tabulations, a proven technique that can help statistically demonstrate the credibility (or lack thereof) of results.
In Kenya, the Election Observation Group (ELOG) conducted a PVT, which showed, on the basis of independently collected polling station results forms, that the August 2017 results were credible, irrespective of the problems that were later faced in the transmission of results by the IEBC. The drive-by criticism of ELOG was off-base and unwarranted.
Domestic observers irritate election commissions and ruling regimes, even when they validate the results, because as citizens, they are invested in the findings of their election missions and have the advocacy networks to sustain pushing for their recommendations and future improvements. Discrediting the practice of observation harms the credibility of domestic observers, too, and, in societies already tending towards repression, may ultimately contribute to reducing the space in which civil society can operate.
Election observation still has a place
The shortcomings of EOMs in Kenya raise tough questions about methodology and purpose. But undifferentiated criticism of international EOMs will not contribute to strengthening electoral practice. It has long been recognised that observation can always improve. Yet observation is and always will be only one piece of an elections puzzle. EOMs need to accept criticism where it is due; critics should not write off EOMs as irrelevant. Election observation, done well, still has a place.
* THE VIEWS OF THE ABOVE ARTICLE ARE THOSE OF THE AUTHOR AND DO NOT NECESSARILY REFLECT THE VIEWS OF THE PAMBAZUKA NEWS EDITORIAL TEAM
* BROUGHT TO YOU BY PAMBAZUKA NEWS
* Please do not take Pambazuka for granted! Become a Friend of Pambazuka and make a donation NOW to help keep Pambazuka FREE and INDEPENDENT!