The Computer Society of Kenya

Since 1986


Tuesday December 20, 2016

Parliament has been summoned to meet today and discuss urgent matters, one of which is about changing the law to allow the Independent Electoral and Boundaries Commission (IEBC) to fall back to manual systems in the event electronic systems fail.

The battle lines have already been drawn, with each side of the political divide screaming valid points for or against the motion. 

On one hand, we have the self-declared ‘digital’ government wishing to have manual systems on standby, ready to take over in the event electronic systems fail.

Curiously, on the other hand we have a presumed ‘analogue’ opposition fronting for a strictly electronic system, arguing quite correctly that the manual system is notoriously prone to manipulation by incumbents in Africa.

The political temperatures are rising rapidly and moving dangerously towards explosive.  At this rate, we are likely to witness the return of street battles, unless both sides reason together and resolve the stalemate.

We have extensively analysed the electronic electoral systems on this column since the last election, with the most recent post speaking to this question while providing background knowledge.

So who is right and who is wrong?  Unfortunately, both sides are right and at the same time, very wrong.


First let us deconstruct the Jubilee argument that all systems must have a backup in case of failure. Whereas this is true, the Jubilee side prefers to be conveniently silent on the opportunities for fraud that arise from such manual interventions, and there are many.

Most critically, ghost voters can be resurrected to cast their votes, particularly if both the Electronic Voter Identification system fails in conjunction with a failed Results Transmission System.

Indeed, nothing stops the opposition Cord side from also abusing a failed system, but the incumbent always has an upper hand when it comes to leveraging institutional failures.

Jubilee should therefore take the conversation beyond agitating for a manual backup system, and include conversations around how much electronic failure is tolerable, beyond which the results are no longer tenable or credible.

One approach would be to recognise that the electronic electoral system is not one monolithic system, but consists of three key elements, namely the biometric voter registration (BVR), the electronic voter identification (EVID) and the electronic results transmission system (RTS).

Further, we must agree that the possibility of electoral fraud increases exponentially according to the number of electronic elements that fail. 

If, for example, the electronic voter identification component fails and the voter is subsequently identified manually, the chances of fraud are quite limited, as long as the other two components, the BVR and the RTS work effectively.  

Conversely, if the RTS fails but the BVR and EVID work perfectly, there should be less cause for alarm.  Essentially, the three subsystems have a symbiotic relationship that can be used to validate or cross-check each other.


Which brings us now to the opposition position. They are right to the extent that manual systems are prone to exploitation, particularly by the incumbent government and should therefore be avoided.  

However, they are wrong to expect electronic systems to execute flawlessly during the voting exercise.  Reality has always dictated otherwise, given that any man-made system is prone to failure.

Sometime the failure is maliciously engineered, while other times it is a reflection of the genuine weakness inherent within man-made systems.

Because it will be difficult to differentiate between the two types of failures during the voting period, it is better to agree on what should be done in the event of failure.

So Cord, just as prescribed for Jubilee, should be discussing what level of electronic failure is acceptable, beyond which the results can no longer be acceptable given the potential exposure to manipulation that would arise from the manual alternatives.

My own take is that biometric voter registration must work since the registration time frame is quite spread out, allowing for any necessary repairs of faulty equipment.  It therefore should not be negotiable.

On the other hand, electronic voter identification (EVID) and the results transmission system (RTS) are quite time-sensitive. If they failed, manual intervention may be the only option available.

However, both of them should not be allowed to fail at the same time, since they play a critical role of validating each other.

In cases where both have failed, a repeat exercise should be considered particularly, in areas that are traditional strongholds for either party. This could be the compromise position for both parties.

Share this page