Ballot boxes at Mikindani Primary School in Mombasa County, August 11, 2022. [Kelvin Karani, Standard]

“Kivumbi 2022” is done. It is time to begin initial reflections on the 2022 General Election. This is similar to the interim reports issued by foreign and local Election Observation Missions after they traversed Kenya and sampled our electoral terrain.

The typical interim mission report uses a template of observations on electoral preparations, voter registration and education, party nominations and candidate registration, electoral campaigns and election day events but concludes just when vote counting is ongoing. 

The roles of media and civil society, plus the electoral commission, Judiciary, security and other State agencies are captured in this report, with recommendations to these institutions supporting calls on Kenyans to exercise calm and patience during the vote count.

This year’s election has been evaluated as “satisfactory to good” in the reports in the public domain. Electoral preparations were better than 2017, though voter registration and education didn’t quite pan out as planned, party candidate nominations were more civil than they have been in the past, violence was less pervasive during the campaigns and events on election day were OK, save for KIEMS kit hitches that affected voting in places.

This early stance by missions reflects lessons from past elections. In 2007, positive assessments were reversed by the post-election violence. In 2017, the Supreme Court annulled the presidential poll. It is an evolving approach that appreciates that everything up to voting is fine, but the results phase is always tense.

‘Takes it all’

Though this is beyond the scope of these missions, it is fair to state that 90 per cent of this tension revolves around the results of the presidential vote. This week in discussions with colleagues, we asked if our “winner-takes-all” presidency is a divisive, rather than unifying, figure in Kenya, despite devolution. Look at the past week and ask - is our show-stopping presidential election the real threat to peace and progress? 

Let’s pursue a slightly different perspective today. I have argued before that the complete transformation of our electoral process goes beyond modernising IEBC. It needs a voter-centric lens that steps out of the irritating soccer analogy in which IEBC sets rules and referees a game between teams with players who decide which rules to follow and which refereeing decisions to respect. 

In this current politician/candidate-centric approach to elections, it is generally assumed that we all want to be fans (voters) to be herded to the sports arena (polling booth) like livestock to a cattle dip. This is the result of a missing end-citizen (similar to end-consumer) lens that focuses politics over processes. It is also the result of culture that prefers seeks “quick temporary fixes” to “sustained and continuous improvement”.

This brings us to my outline “Kenyan Voter’s Observation Report”. It is not a citizen testimonial, but a macro-perspective using data to tease out questions along the five steps of the voter experience: registration as a Kenyan; identification as a voter; voting in the booth; tracking results and accepting electoral choices. 


The supply-side, as mentioned last week, is about institutional steps to process Kenyans into IDs, IDs into voters, voters into voting, voting into votes, votes into counts, counts into tallies, tallies into results, and results into announcements and announcements into certificates.  Let’s stick with the demand-side. 

Registration as a Kenyan is where it starts. According to the 2019 census, 54 per cent of Kenyans were aged 18 and over, which is roughly 26 out of 48 million people at the time.  Applying the same proportion to 2022 with a population around 52 million, suggests an adult population universe of 28 million.

So, the voter’s pre-requisite is an identity. Do we have 28 million active ID cards in Kenya? A question for the National Registration Bureau today; a serious question on our national identification strategy for the future after the floundering efforts of Huduma Namba. From an electoral perspective, we do not spend enough time thinking strategically about identification and registration. Indeed, this calls for better “joined-up” government on day-to-day work and operations, and not simply episodic events such as elections.

From Kenyan ID (including passport), we move to becoming a Kenyan voter. Remember, the estimate is 28 million Kenyans with IDs in 2022. The IEBC target was six million new voters from 2017’s 19.6 million voter roll. Factoring in a million deaths over five years to be expunged from the register, this implied a new voter total of 24.6 million. The actual voter roll came in at 22.1 million for 2022; roughly 2.5 million (not 3.5 million) short. But the real projection suggests it was six million short if we are 28 million adults!

Then there’s the youth question.  Apparently, IEBC’s 6 million target was all youth; only 2.5 young people signed up. It isn’t clear that we are getting deep enough into this question when our 18-35s who are 60 per cent of the adult population take up only 40 per cent of the voter roll (meaning 40:60 for over 35 adults). 

Before we get into voting next, here’s a final number to consider: in 2017, 9.4 million young people were registered to vote; in 2022, the number was down to 8.8 million.

Two steps

Whatever the reason, it’s long before the vote.  Broadly, where and why do we lose the overall 28 million voter potential?  It is easy to talk about voter apathy. Is there space for (voluntary) incentives, (affirmative) guidelines and (compulsory) mandates? All questions so far around these first two steps must apply with equal emphasis to women, the marginalized and vulnerable and any other groups facing disadvantage or discrimination.

Let’s get to voting in 2022. A lazy narrative has already emerged that the relatively low voter turnout is explained by apathy among the youth. To begin, this not necessarily factual until a profile of the actual voters is known; as could be established based on the biometric registrations during voting. But was the turnout, estimated at 66 per cent in this article, low and disappointing, as many claimed last week? Especially when we learn that between half and a million fewer people voted this year than in 2017?

First round

Many Kenyans will remember the 2002 NARC moment as a “wave”, right? Well, the turnout then was 57 per cent.  In fact, this year’s turnout is closest to 1992 (66 per cent), 1997 (68 per cent) and 2007 (69 per cent). And it is nowhere close to the incredulous 86 per cent turnout of 2013, or the astonishing 78 per cent the first round of 2017? Why don’t the change moments reflect in the voter numbers? Do the high turnouts tell us more about the electoral stakes than the state of nation? Do these turnouts really matter at all?

Let us complicate this further. The voter experience so far is register-identify, then register-voter then vote.  Does this mean that the collective voter experience (like an aggregated customer experience) depends on registration and turnout?  Take the multi-party era. 

In 1992, 72 per cent of eligible voters registered and turnout was 66 per cent. Does this make the effective vote the multiple – in this case 48 per cent?  Think of this number as saying that just under half of eligible votes actually voted in 1992. 

Contrast with euphoric 2002 with 66 per cent of eligible voters and 57 per cent turnout – effective vote 39 per cent.  Or the highest multiple to date – in 2017, 78 per cent of eligible voters, 78 per cent of whom voted – which effectively means 61 per cent of eligible voters actually voted.  1992 and 2002 were the only years in the multiparty era where less than half of eligible voters actually voted. At a 66 per cent turnout, 2022 suggests between 52 and 56 per cent of eligible votes voted; only 2013 and 2017 have been higher.

As we consider the results and acceptance phases that we are currently in the middle of, let’s complete this calculation set for you to ponder. Again, 2002 is remembered for its 61 per cent winning vote. Adjusted for registered voters, this falls to 35 per cent, then 23 per cent of eligible voters – the lowest level since Kanu, but yet the most popular (and most convincing) result in our electoral history. 

Naturally, 2013 and 2017 had the highest percentages, but the 2022 number might settle closer to 2002-2007 levels. What does this apparent randomness of numbers tell us about the voter experience? Did it become a high volume, high maintenance – low quality, low return exercise over time? And does the experience of poor returns last time discourage new investments in candidates next time?

Maybe we need to dig deeper into the voter experience before rushing into mobile voting, KIEMS kits, APIs or OCR-ready documents, live streaming and other technology quick-fixes as many suggest. Isn’t getting this citizen experience right the “continuous improvement” pathway to a better electoral process?