In his bi-monthly column, our Co-editor Nic Cheeseman argues that, despite the glitches that occurred in Kenya’s recent elections, electoral technology can and should be made to work in the country as part of a broad set of checks and balances in the election.
It has been hard to find anyone in Kenya with a good word to say about the new technology introduced to act as check on electoral manipulation in the wake of the March 4 elections. Cord leaders included the failure of the new measures as one of the main planks of their petition against the official results. International and domestic election observers noted their disappointment that none of the new systems functioned as intended. Journalists questioned whether the system had been tampered with: was this “conspiracy” or “cock-up”? Donors have wondered how such fragile equipment could cost so much money. Even Twitter got in on the act; following yet another delay in the counting process, one sharp-witted commentator wondered whether it was actually the IEBC’s clock that had been hacked.
These criticisms are understandable. The elections could have run much more smoothly than they did — which would have gone a long way to persuading losing candidates to accept the results. But the notion that Kenya is somehow not ready for electoral technology is wide off the mark because it rests on a misunderstanding about what went wrong in the election and underestimates what Kenya is capable of. By suggesting that election technology still has much to offer Kenya I do not mean to gloss over how badly the new mechanisms for voter verification and the transmission of results actually worked. The biometric verification of voters by checking their fingerprints using an electronic scanner failed as soon as it had begun. In a small number of polling stations the equipment never appeared. In many more polling streams the kits did not work — either because they lost power, or because the scanner could not be made to function. According to ELOG, the domestic election group, the kits failed in a majority of polling stations at some point during the day. As a result IEBC officials did as they had been instructed: they reverted to the same manual system used in 2007. This does not mean that rigging necessarily occurred — manual systems can work perfectly well and deliver free and fair results — but it does mean that one of the important checks that was supposed to stop multiple voting and ballot box stuffing was wholly ineffective.
The transmission of provisional results via mobile phone fared little better. The idea was that the results from each of the 33,000 polling streams would be texted in and automatically broadcast live at Bomas, on the Internet, and on TV. By comparing these results with the official aggregate figures for each constituency it would have been possible to check for errors, or manipulation, in the counting and tabulation process. But once the first flurry of results had come in the system slowly ground to a halt. Not only did results for the majority of polling streams fail to appear, but the results that were transmitted raised serious problems. Instead of the gap between the two candidates fluctuating over time as results from different leaders’ strongholds came in — as the results did in 2007 — the text results seemed to give Kenyatta a constant lead over Odinga. Even more mysterious was the case of the rejected votes. On the basis of the mobile phone transmissions it appeared that these would play a major role in the election. Not only did the rejected vote tally outperform Musalia Mudavadi, but once the chair of the IEBC had declared that rejected votes would be included in the calculation of the 50%+1 threshold required for a first round victory it seemed that the incredibly high number of rejected votes might deny Uhuru Kenyatta a first round victory.
In the end, of course, when the manual count started to come in it turned out that the actual number of rejected votes was much smaller. The IEBC explained this discrepancy away by saying that there had been a “glitch” in the technology that had multiplied all rejected votes by eight — but this raised more questions than it answered, and further undermined the credibility of the text transmission system.
So what is the case for giving electoral technology another chance? Well for one thing, not all of the technology failed. It is important to remember that the use of biometric registration to create electoral register was actually a success. Civil society groups and Cord have understandably complained about the use of multiple registers and the possibility that some people who did not register actually voted, but no one can deny that the voter register in 2013 represented a vast improvement on previous years. Most significantly, over a million “ghost voters” were excluded, which meant that it was not possible to dramatically inflate the votes for a candidate by adding ballot papers in the name of the dead who could not make it to the polls. More work needs to be done to generate a unified and transparent electoral register, but this was an important step in the right direction.
The second important thing to keep in mind is why the technology failed. The systems put in place in 2013 did not fail because they were implemented in Africa, or because Kenya is not sufficiently technologically savvy to make them work. They failed because they were not introduced with sufficient time to test them and to build in contingencies. Neither the kits to scan voters’ fingerprints nor the handsets used to text in results were distributed in time to allow for rigorous testing.
This was not all the IEBC’s fault. In the fall-out from the election, many people have forgotten that following the failure of the first procurement process, the IEBC announced it was abandoning its plans to introduce new technology. One reason that the commission took this decision was that they realised that the electoral timeline risked becoming impossibly tight. On this point they were right, and it was the intense time pressure, combined with the failure of a number of key players to pick up the pace, which meant that the election went ahead with technology that had not been rigorously trialed.
But this does not mean that new technology cannot be made to work. The failure of these systems was not due to anything unique to Kenya — with so little preparation they would have failed anywhere in the world. And there are good reasons for thinking that Kenya is actually fairly well placed to make new electoral technology work. Mobile phone penetration is incredibly high, while Safaricom remains a world leader in terms of innovation. Kenyans are more likely to use their phones for services such as mobile money transfers than their American or German counterparts. If individuals in Nairobi can use the mobile network to send money home to Coast, Eastern, Nyanza, North Eastern and so on, there is no reason that a returning officer cannot text in results of an election. Of course, the lack of electricity in some areas presents a logistical challenge, as does the need to distribute kits to remote areas. But the IEBC largely managed this during voter registration.
This is not to say that it is easy to introduce new technology. Making voter registration and results transmission processes stronger and more reliable will be costly and will require both political will and commitment of the IEBC. Even then, it is risky to rely too heavily on BVR kits and mobile phones; technology is no panacea and individual systems can always fail. For this reason, it is a good idea to invest in a series of checks and balances including domestic monitors and party agents. Used in this way, new methods of identifying voters and counting ballots can increase the confidence of political leaders and the public in the electoral process — as they did before the systems started to fail last time round. So, although it is important not to fetishize the digital world, the best way to improve the quality of elections in Kenya is not to abandon technology, but to make it work.
This column was written for the Daily Nation and appeared 26th April 2013. The Daily Nation is the largest newspaper in East Africa with a daily circulation of around 205,000.
