Special report: Why Kenyans need clean data to be sure real people vote in August

A depressed vote is as much about turnout as it is about initial registration. [Denish Ochieng, Standard]

When we delved into the tyranny of numbers a fortnight ago, excitement centred around unaudited Independent Boundaries and Electoral Commission (IEBC) data on 22.1 million voters registered for the 2022 election.

For the Presidency at 80 per cent voter turnout, 150,000 votes short of nine million gets you the job, while 10 million is the Holy Grail; the unequivocally decisive statement not of victory, but annihilation.

As discussed before, Kenya’s 2022 focus is - rightly or wrongly - on the Presidency.

For the avoidance of doubt, 2013 was decided on 86 per cent turnout, while 2017 dropped to 78 per cent for the first round; and half of that for the repeat election.

For greater clarity, a 75 per cent turnout in 2022 requires just over President Uhuru Kenyatta’s 2017 first round 8.2 million. It was subsequently annulled by the Supreme Court. A 2013-level 85 per cent raises the stakes to 9.4 million, and an impossible 91 per cent turnout takes it to the 10 million Holy Grail. 

Will 2022 repeat 2013, or 2017? 

There is more to consider. Contrary to popular belief around the NARC wave, voter turnout in the 2002 transition from Kanu was only 57 per cent.

Going by the official numbers, this was actually lower than the previous two multi-party elections in 1992 and 1997 which hit 66 and 68 per cent respectively.

Or the higher 69 per cent that represented turnout at the infamous 2007 election that took us to war.  This is what makes Jubilee’s 2013 and 2017 turnout numbers of 86 and 78 per cent so incredible, and incredulous. 

Before you all jump in, two quick points must be made.  First, smart candidates are probably busy drilling down and rolling up these macro-numbers across Kenya’s diverse geographies and demographies; hence the coalitions, alliances and transient “come we stay” political arrangements currently in vogue.

This is the unfortunate and discomfiting reason why our election has been termed an “ethnic census”. 

Second, the political scientists will tell you that turnout, especially within regions, tends to link directly to whether or not there is a “home” candidate on the ballot. 

An elderly voter being assisted to cast her ballot at Bidii Primary School, Nairobi, 2013. [Boniface Okendo, Standard]

If, as it seems, Mount Kenya will not have a viable home candidate, then one expects turnout to be lower than usual; which because of the size of this voting bloc in relation to the total, suggests overall turnout at less than 80 per cent.  

But leave those worries for the influencers, pollsters and number-crunchers sitting in presidential campaign secretariats; safe in the knowledge that, as is the case world over, there is always one side that prefers a higher turnout, or more darkly, a suppressed  - more politely, depressed - vote works better for some than for others.  Let us instead stick with macro-numbers and pursue this further line of thought. 

Turnout vs Signup

A depressed vote is as much about turnout as it is about initial registration.

While we tend to focus on voter turnout (the number who vote against the number who register to vote), we pay less attention to voter signup (the number who register to vote against the number who are eligible to register to vote). IEBC tends to set its registration targets based on this eligibility number.  

So, for example, for 2013, its target was 18 million out of 21.8 million eligible (an 83 per cent target signup rate) voters.  It registered 14.4 million voters – a signup rate of 66 per cent; also 80 per cent of their target – of whom 12.4 million voted. 

Effectively, the 2013 election result was determined by 57 per cent of what the vote might have been if voting in Kenya was actually compulsory for everyone of voting age. 

If we look at 2017, a similar picture emerges. From a target of 22 million out of 25.2 million eligible (target signup rate of 87 per cent) voters, IEBC registered 19.6 million (signup rate of 78 per cent; 89 per cent of target), of whom 15.2 million voted in the first round presidential election (60 per cent of eligible voters). 

In 2022, IEBC’s ambitious 26 million voter registration target was set against an eligible voter population of 30.3 million today (target 86 per cent signup), of whom the current unaudited voter data says 22.1 million, or 73 out of every 100 eligible voters (85 per cent of overall target) in the register.  

Yes, there will be murmurs about IEBC’s poor new voter registration.

August 2022 will simply give us voter turnout and the effective vote for poll legitimacy.

Let us put these numbers into perspective.  At 57 and 60 per cent respectively, the effective vote (signup times turnout) in 2013 and 2017 was relatively high in a country without compulsory voting. Only in 1997 and 2007 has the effective vote exceeded 50 per cent in our history (2002 came in at 37 per cent).

Yet, how credible are our voter (and people) numbers, especially in more recent times? As we consider this question in the full knowledge that the IEBC voter register is yet to be audited and cleaned, here are four wider angles to this question.

Voter records

The first harks back to the audit of the 2017 register, and issues raised. 

Disparities between ID cards in circulation and ID-eligible citizens are still an area of concern.

Difficulties with data on deaths in a country in which only four out of every 10 deaths is officially registered while between three and four of every 10 births are not.

Ballot boxes at NSSF Grounds polling station, Nairobi, 2013. [Boniface Okendo, Standard]

There is also the issue of voter records with duplicate or invalid identification numbers or documents.

IEBC might usefully advise us on how the 2017 register was cleaned given that the pre-audit number of voters was pretty much the same as the post-audit number of voters in the actual polls register. 

This first audit perspective will be complicated by a second related one; the advent of new responsibilities that fall under the Office of the Registrar for Political Parties (ORPP). 

An immediate one from the electoral perspective is the requirement for political parties to submit their complete party membership lists to the office by March 26 – less than a month from today. 

These lists will be verified and certified by ORPP within seven days, after which they must be submitted to the IEBC by April 9. 

Although this is a matter now in court, the key point to remember is that only party members can engage in direct nominations of party candidates; a process that should be completed by the end of April.

Numbers games

Now here is the twister.  While there is nothing that says one must be a party member in order to vote in a general election, or one must be a voter in order to join a party, the numbers games begin again. 

Recent reports from the ORPP suggest we have 24.8 million party members today (when we have 22.1 million voters); up from 14.6 million in June 2020 (when we had 19.6 million voters) and 7.6 million in 2013 (when we had 12.2 million voters).  What explains this dramatic rise in party membership since 2020? 

And, given past questions about persons being registered in parties without their knowledge, how will data verification of party membership be synchronised with data cleaning of the voter register into a single version of the truth? 

Do April nominations look like the main election in this new schema? 

Speaking of a single version of the truth, is this not what Huduma Namba was all about; the single unique identity that the Central Bank has this week positioned as the future of our digital payments ecosystem in its 2022-25 National Payments Strategy? 

 Huduma Namba

It is impossible to revisit and repeat the whole idea of Huduma Namba as the core of a National Integrated Identity Management System (NIIMS) that would have squeezed our multiple identities – foundational and functional; biographical and biometric – into a single digital identity number.  It’s too late for 2022, but there must be lessons here for the future. 

This brings us to an uncomfortable, final point on our tyranny of numbers.  We may be 24 million-plus party members, or 22 million-plus registered voters in an adult population just above 30 million people. 

Yet 2022 and the future isn’t an “adults-only” affair, it’s about all of Kenya, so how many are we? 

We counted ourselves as 47 million in 2019, when earlier estimates showed we had hit that population number even before the 2017 election.