April 29, 2013, 4:54 p.m.
The Role of Technology in the Outcome of the Kenyan General Election Ayesha Chugh and Katherine Krueger This article was originally posted on the ACE Electoral Knowledge Network 'Elections Today' feature. Introduction Kenyan citizens voted in a historic general election on March 4, 2013. It was a significant contest in light of the mass violence following the 2007 presidential election, claiming over 1,000 lives and displacing over 600,000 people. This year, Election Day was largely peaceful and Kenya demonstrated its ability to move beyond 2007’s ethnic turmoil. According to official results released by Kenya’s Independent Elections and Boundaries Commission (IEBC), Uhuru Kenyatta, leader of the Jubilee Coalition, defeated his opponent, Raila Odinga of the Coalition for the Restoration of Democracy (CORD). Odinga subsequently challenged these results by petitioning the Supreme Court, which ultimately upheld the IEBC’s ruling. Despite the peaceful outcome of this election, the international community and news media continue to focus on its failures, and specifically, technological ones. Some tend to assume that malfunctions of Kenyan voter technology caused Kenya’s polling crisis. On the contrary, we must understand the relationship of technology to Kenya’s vote counting process as a managerial failure before a technological one. Background: Kenya’s Election Management Overhaul Kenya’s 2007 humanitarian crisis sparked a multi-year overhaul of Kenya’s political and electoral system to stabilize the country.[i] By 2013, the reform process was underway by the IEBC, and the body had largely implemented a plan to improve polling stations using three pieces of technology: Biometric Voter Registration (BVR), Electronic Voter Identification (EVID, or pollbooks), and an electronic Results Transmission System (RTS). Yet these technological enhancements were a large challenge for the IEBC to implement. Election Day Challenges As voters took to the polls on March 4, several technological problems arose. Laptops and cell phones used for the RTS ran out of battery power; additionally, some polling stations (particularly in the rural areas) had no outlets.[ii] Many poll workers were also poorly trained on how to use the new election software, lacking access to (or, in some cases, forgetting) basic PIN numbers and passwords needed to operate. Additionally, the electronic voter identification system (EVID) had not been set up, distributed, and supplied with enough battery power to last a whole day. This affected more remotely located polling stations. Furthermore, secure servers intended for results transmission were unable to handle the volume of data being uploaded, leading to the breakdown of the RTS.iii An error with the RTS source code also led many to believe there were 300,000 invalid ballots, when in reality the source code error had just multiplied the actual number of invalid ballots by 8 (an ‘8x error’).iv The Failure of Project Management We must view the problems faced by the IEBC above as managerial, rather than technological. This is exemplified by the delays in key processes that adversely impacted the management of election technology. The short timeframe between the development of the RTS and the election limited the amount of testing for the system before the election. Delays in several key processes, which include—the competitive procurement process for a BVR system, a three month delay in the commencement of voter registration, and the procurement process for EVID — directly impacted the ability of the IEBC to test all three of these technologies. Although the 2011 Elections Act mandated a 90-day period between the end of voter registration and the election, Kenya’s National Assembly compressed it to 60 days, giving the IEBC a tight timeline in which to train their staff and prepare for any contingencies. The Role of Technology in the Outcome of the Election Planning more thoroughly for technological failures may have prevented the abandonment of the RTS. However, the relationship of technology to the overall outcome of the election, and the ensuing political conflict over votes, is less apparent. Although the Supreme Court in March ruled the election was indeed free, fair, and credible, some continue to associate technological failures with the outcome of this election. For example, some Odinga supporters use the failure of technology to bolster claims that election results were manipulated, citing the fact that the election commission was unwilling to publish election results by polling stations, and repeatedly changed the official number of registered voters.v As controversial as this question has become, these technological failures ultimately did not compromise the sanctity of the election. Regardless of whether technical difficulties had occurred or not, all electronically transmitted results were provisional, and while this provided the IEBC with an accountability mechanism, the official results were solely based on paper. Each presiding officer of a polling station would have had to fill out a paper form with results details, and official results would have had to be determined by manually tallying all votes. Therefore, the official results declaring Uhuru Kenyatta as the winner were tabulated by hand.vi As the Carter Center notes in their final observation report of the election, “[d]espite serious shortcomings of the IEBC's management of technology and release of information, [we find] that the paper-based procedure for counting and tallying presented enough guarantees to preserve the expression of the will of Kenyan voters.” [iii] Conclusion Kenya’s many challenges on Election Day, therefore, must be understood as problems of project management, rather than a consequence of voter technology. Kenya’s Supreme Court finally ordered a recount of 22 polling stations in late March, after Raila Odinga alleged voter fraud and petitioned for the nullification of the official results. Considering multiple non-technological problems faced on Election Day, Odinga may have still demanded a recount regardless of the technological setbacks. For instance, Kenya experienced an unprecedentedly high voter turnout of 86 percent, contributing to long lines and voting stations voting four hours after polls had officially closed. Additionally, Kenya’s RTS had to deal with six different elections as stipulated by the 2010 Constitution.vii Notwithstanding, Odinga may have also alleged fraud without any of these challenges arising, noting the incentives for losers to declare fraud in a high stakes winner-takes-all system like Kenya’s. The tightness of an electoral race, guaranteed by Kenya’s majoritarian electoral system, which stipulates that the winner must receive 50 percent of the national vote +1 (+ 25 percent of the vote in at least half its counties) may encourage candidates to reject election results on grounds of manipulation. Finally, compared to the multitude of problems faced in 2007, the March 4th 2013 election (and its use of technology) can be seen as a success. For example, The Kriegler Report’s evaluation of the management of the 2007 polls revealed mass incidents of ballot stuffing and 1.2 million dead people on the voter roll. Thus the technologies in question during this election (EVID, BVR, and RTS) were introduced to address these larger problems, and must be understood in this context.