Tuesday, August 11, 2009



The General Elections in Kenya,
December 2007
Clark C. Gibson, James D. Long
Department of Political Science,
University of California,
San Diego, 9500 Gilman Drive,
#0521, La Jolla, CA 92093-0521, USA

On 27 December 2007, Kenyan voters participated in the fourth elections since the re-introduction of multi-party Politics in 1991, casting ballots for the president and members of parliament. Little violence occurred on election day, and observers considered the voting process well organized. However, delay by the Electoral Commission of Kenya (ECK) in announcing the results incited serious unrest in the country. Challenger Raila Odinga had taken an early lead on the first day of counting, but incumbent President Mwai Kibaki erased that margin and went on to win by 2%.

The hasty inauguration of Kibaki on 30 December, and claims of vote rigging, fostered violence throughout the country. The fighting resulted in nearly 1200 deaths, 500,000 displaced persons, and the widespread destruction of land and property.

Mediation efforts by former United Nations Secretary General, Kofi Annan, led to a power-sharing coalition government in February, 2008 leaving Kibaki with the presidency while creating the post of prime minister for Odinga.

Independent analyses and the work of the Independent Review Commission to study the election revealed deep flaws in the vote tally process, undermining the credibility of the ECK’s official results. Since the ballot counts are untrustworthy, we present the results of our nation-wide exit poll. These data provide the only reliable information on the results of Kenya’s 2007 elections.

1. Background

Like most African countries, Kenya re-introduced multiparty competitive elections in the early 1990s. Daniel arap Moi of the Kenya African National Union (KANU) assumed the presidency in 1978 under one-party rule; he also prevailed over a divided opposition in the 1992 and 1997 multi-party elections. Moi abided by constitutional limits,appointing a political neophyte, Uhuru Kenyatta, as the KANU candidate in the 2002 election.

The opposition National Rainbow Coalition (NARC), headed by Kibaki, defeated Kenyatta in a landslide. Kibaki’s electoral coalition included Odinga from Nyanza province and Kalonzo Musyoka from Eastern. Standing as a ‘change’candidate in the 2002 election, Kibaki promised important political and economic reforms after Moi’s lackluster
performance throughout the 1980s and 1990s. He promised to deliver a new constitution that would decentralize Kenya’s bloated central government by, among other things, creating the post of prime minister (slated for Odinga)and strengthening parliament. Echoing the reforms promised elsewhere in Africa, Kibaki also advocated universal
primary education for Kenyan children.

Kibaki inherited a negative growth rate, and alarmingly high rates of unemployment and poverty. Importantly, he turned Kenya’s economy around; the agricultural and services sectors grew impressively, as did the flow of international aid (severely reduced during the Moi era). By 2007, the economy rocketed to a growth rate of nearly 7%.Kibaki also delivered on his promise of universal primary education.

The NARC coalition, however, disintegrated over the issue of constitutional revision. The draft of the new constitution, produced at the Bomas Convention, hewed closely to the promises made by Kibaki ahead of the 2002 election. But Kibaki erased the curtailment of presidential powers in revisions of the draft, which was
presented to voters in a 2005 referendum. Kibaki’s reversal led to Odinga, Musyoka, and others defecting from NARC and forming the Orange Democratic Movement (ODM). ODM campaigned vigorously against Kibaki’s version of the constitution, which was rejected 58–42%.

2. Electoral system
Kenya’s electoral rules set three requirements to be elected president. A candidate must win the most votes in a nation-wide count, and secure at least 25% of the vote in five of any eight provinces. The presidential victor must also win the parliamentary seat in their own constituency.

Kenya’s National Assembly is chosen by a first-past-the post system in 210 single-member constituencies. In addition to these elected seats, 12 MPs are allocated to parties in proportion to their elected seat total, along with two ex-officio members

3. Party contenders

The political landscape shifted ahead of the 2007 Election with the demise of NARC and the emergence of ODM. Kibaki reconstituted his coalition as the Party of National Unity (PNU), which included members of Moi and Kenyatta’s KANU. In the meantime, ODM constructed a coalition of leaders – labeled the ‘Pentagon’– from the spread of Kenyan provinces: Odinga from Nyanza province, Musalia Mudavadi (Western), William
Ruto (Rift Valley), Joseph Nyagah (Eastern), Najib Balala (Coast), and eventually Charity Ngilu (Eastern).

Falling out with ODM, Kalonzo Musyoka formed the splinter ODM-Kenya (ODM-K), but garnered little support outside the Ukambani region in Eastern province, Musyoka’s home.

Given the fracturing of NARC and the embryonic nature of PNU, ODM, and ODM-K, the National Assembly election featured a shifting set of parties and candidates. PNU fielded candidates under its own party label and relied on support from a coalition of smaller parties (including KANU). ODM and ODM-K largely contested alone. A disorderly nominations process locked out many candidates from the main parties,
leading to many defections to smaller parties. Altogether 2548 candidates stood for the National Assembly’s 210 elected seats.

4. Electoral campaign

The presidential campaign centered heavily on Kibaki’s performance. His 2002 victory had generated such tremendous ‘‘euphoria’’ that Kenyans registered some of the highest support for democracy in the Afrobarometer survey in 2003.

By 2007, however, many citizens were far less enthusiastic about Kibaki’s achievements in office as well as democracy in general. Both Odinga and Kibaki took positions that played on their policy strengths, highlighting differences over salience and solutions. In appealing to Kenya’s wealthier and growing urban middle class, Kibaki touted his performance in achieving a robust growth rate. Odinga fashioned a more populist message, charging Kibaki with helping the rich and ignoring the country’s poor. Thus Kenya’s economic successes and challenges featured significantly throughout the campaign period.

Constitutional revision also remained important to many voters.

The Kibaki version of the constitution confirmed many voters’ fears that the government remained too strong and centralized. Odinga seized upon the issue, promising to prioritize constitutional revision when he took office. Kibaki sought to downplay the importance of the constitution and the loss of the 2005 referendum.

Central government power also undergirded the issue of majimbo (Swahili fo‘regions’), which implied thedevolution of political responsibilities to lower levels of government. For Odinga, majimbo was a platform from which he could assert various campaign themes: reducing the power of the executive; giving greater political and
fiscal authority to local units; and an alternative vision of economic development that placed more decision-making responsibilities in local communities. In response, Kibaki charged Odinga’s coalition with tribalism, suggesting that majimbo was a strategy to divide Kenya along ethno regional lines.

Education remained a benchmark for Kibaki’s performance in office, having provided universal primary education during his first term. While challenges persisted on teacher training, supplies, and infrastructure, Kibaki received local and international praise for his efforts.

Playing to these strengths, Kibaki promised to provide universal secondary education. Odinga, however, downplayed the education issue or appeared inconsistent, sometimes promoting universal coverage but at other times advocating co-financing between parents and the government. However, ODM did make corruption a key issue.
As in most African elections, challengers frequently point to corruption as a significant failing of incumbents.

Although Kibaki stood on a reformist platform in 2002, promising to pursue Kenya’s corrupt ruling class and creating new institutions to investigate graft, he was generally considered to have failed. Many of Kenya’s most notoriously corrupt parliamentarians remained integral to his ruling coalition. Moreover, Kibaki did not act decisively in the two most highly publicized cases, the Anglo-Leasing and Goldenberg scandals. Unsurprisingly, Odinga focused heavily on Kibaki’s record on corruption and promised to run a ‘‘clean government’’ if elected.

Ethnicity also remained relevant to the election. Both Kibaki and Odinga’s coalitions actively sought votes beyond their Kikuyu and Luo bases to appeal to Kenya’s many swing groups. High profile leaders from the same ethnic groups joined both parties, making it difficult to predict votes based on ethnicity alone. For example, Kibaki, Odinga, and Musyoka all chose members of the Luhya community as vice-presidential candidates; William Ruto, a Kalenjin, was part of Odinga’s ‘Pentagon’ but former President Moi (also a Kalenjin) and his family supported Kibaki. Typical in Kenyan campaigns, both sides charged the other with using tribal appeals while boasting of their own multi-ethnic credentials.

Age and youth vote also mattered, with each campaign actively courting the vote of young Kenyans. Kibaki tried to win them with his ‘‘Vijana na Kibaki’’ (Young People for Kibaki) movement; but Odinga was more successful, tying youth issues to those of the jobless and urban poor. Odinga also defined the contest as a ‘change’
election, through which Kenya’s young voters could throw out an aging and corrupt political class.

5. Election results

International and domestic observers are near unanimous that vote counting in the 2007 elections was flawed. Because it is impossible to re-count the actual ballots, we use data from our exit poll to analyze the outcome. A comparison of our exit poll with official results highlights important areas of discrepancy. The ECK reported that Kibaki, Odinga, and Musyoka all won parliamentary seats. It also reported that Kibaki tallied 225,174 more votes than Odinga, producing a 46.4–44.1% win while Musyoka polled a distant third with 8.9% of the vote. The ECK reported that Kibaki met the 25% threshold in every province except Nyanza, Odinga’s homeland; and
that Odinga surpassed 25% in all provinces except Central and Eastern, Kibaki’s homeland.

The official ECK results VS the exit poll.

The first difference to note is the presidential vote: Odinga won the exit poll 46.1% to Kibaki’s 40.2%, a victory lying outside the poll’s margin of error (G1.32%). The second important difference is the set of provincial tallies: the official results claimed that Kibaki won Northeastern province with 50.5%, whereas the poll shows an Odinga win with 76%. Moreover, in seven out of eight provinces, Kibaki registered more votes in the official results than in the exit poll. The differences, in percentage points, are greatest in North Eastern (33.3 points), Nairobi (14.6), Coast (8.5), Western (8), and Eastern (7.8). In Rift Valley, Kibaki registered 7.7 points fewer votes in the official results than in the exit poll.

Conversely, in seven of eight provinces Odinga registered fewer votes in the official tally than in the exit poll. These differences, in percentage points, mirror Kibaki’s additions: North Eastern (28.8 points), Nairobi (10.6), and Coast
(7.8), and Western (6.9). In Rift Valley, Odinga registered 10 points more official ballots than the exit poll votes.

Taken together, Kibaki benefited from producing additional votes in seven provinces while losing votes in one; Odinga benefited from additional votes in one province and lost in seven. If we aggregate these net differences across provinces, we find that Kibaki benefited from 355,843 extra votes in the official tally when compared to the exit poll, while Odinga lost 57,951 votes, for a total of 413,794.

If we assume that the exit poll is the more valid tally and compare the 413,794 difference with Kibaki’s margin of victory (225,174 votes), it is clear that the ECK results are off by enough of a margin to have declared the wrong winner.
Because these differences in vote totals are biased in Kibaki’s favour, a comparison of the exit poll to ECK results casts serious doubt on a Kibaki victory.

The exit poll also casts doubt on whether Kibaki fulfilled the 25% of the vote in the five provinces rule. He unambiguously passed the bar in Nairobi, Eastern, Central, and Rift Valley, while clearly failing in Nyanza. But in the exit
poll, Kibaki falls below 25% in Coast, Northeastern, and Western. Because these counts are within the provincial margins of error, however, the exit poll cannot claim that he did or did not meet the provincial requirement. Our data,
however, find that Odinga clearly met the requirement in six provinces – Nairobi, Coast, Northeastern, Rift Valley, Western, and Nyanza – while failing in Central and Eastern.

6. Voting behaviour

Our exit poll also sought to uncover the determinants of Kenyans’ vote in the presidential election, especially how government performance, campaign issues, and ethnicity impacted their choices. If performance and issues did not
matter to vote choice, there should be no difference in these evaluations across each candidate’s supporters.

Voters who chose Odinga or Musyoka had generally negative evaluations of the nation’s economy and the economic situation of their family. While Kibaki’s voters had positive evaluations of the nation’s economy, a majority still had negative evaluations of their family’s economic situation, but to a lesser degree than opposition supporters. Odinga’s and Musyoka’s voters gave mostly negative ratings whilst Kibaki’s voters thought the government had done an excellent or good job.

Had Kibaki fulfilled his promises since the last election in 2002?

Voters for the challengers were clearly dissatisfied with Kibaki’s job in office (conversely, his supporters thought he had mostly fulfilled his promises). These results on government performance present a clear pattern: Kibaki drew support from those with positive ratings while Odinga and Musyoka drew support from those who were
less sanguine. These outcomes suggest that Kibaki’s performance, and Odinga’s campaign drawing attention to it had an important effect on their vote.

Voter’s ‘left–right’ position.

We asked Respondents whether they thought that economic growth or employment should be a bigger priority for the government. We anticipated that Odinga’s voters were more likely to choose employment and Kibaki’s supporters were more
likely to choose growth. The results reflect closely the issues the candidates highlighted in their campaigns.

In assessing voters’ desire for change, we asked whether it was more important for candidates to have experience or new ideas. We expected voters valuing experience to favour the incumbent while those seeking new ideas would favour challengers. Odinga and Musyoka’s supporters were, indeed, more likely to desire change while Kibaki’s voters generally preferred experience. The overall response rate leads us to label this contest as a ‘change’ election.

The vote breakdown of candidate support by respondents’ most important issue.

Kibaki clearly won among those voters who listed the economy and education as the most important issue. Odinga won on the issues of employment, majimbo (federalism), corruption, and constitutional revision. These results mirror the salient
issues stressed by the presidential candidates.

The vote by ethnicity.

As expected, the main candidates performed well within their own ethnic groups: Kibaki garnered 94% of the Kikuyu vote (and 88% of the related Meru), Odinga 98% of the Luos, and Musyoka 86% of the Kambas. Despite candidates’ attempts to lure Luhya voters, they overwhelmingly supported Odinga (75%). There was also a fierce campaign for the Kalenjin vote, but, similarly, Kalenjins voted overwhelmingly for Odinga (88%). Groups such as the Kisii, Mijikenda, and Maasai split their votes more evenly, suggesting the importance of capturing swing votes.

Incumbent parliamentarians.

First, we probed how well their MPs had spent money on services in their communities, a proxy for performance over the disbursal of Constituency Development Funds (CDFs). Most voters had negative evaluations (75%). We also asked respondents how well MPs had kept their promises since the last election. Again, the vast majority (79%) responded negatively, which was reflected in the defeat of 65% of incumbents, including 20 former members of Kibaki’s cabinet. For the most part, Kenyans thought poorly of their legislators and tossed them out of office.

7. Aftermath

The ECK’s inability to maintain confidence in the vote produced widespread protest and violence. During December 2007–February 2008, Kenya experienced shocking levels of post-election hostility: battles between government officers and ODM supporters; between members of both main political coalitions; and between various ethnic communities, particularly over longstanding land disputes. Crimes of opportunity also
spread, adding to the intensity of disorder.

The power-sharing agreement, brokered by Kofi Annan in February 2008, helped to quell the violence. In brief, the agreement was for Kibaki to remain president and Odinga to become prime minister. The PNU–ODM accord recommended a new government of national unity that should pursue the needed reforms: investigation of the ECK; drafting a new constitution; the formulation of a new land policy; and the creation of a truth and reconciliation commission.

8. Outlook

The power-sharing agreement brought political adversaries together in a ‘grand coalition’ and reduced violence,but important challenges remain. The government faces the immediate difficulties of resettling the displaced, and meeting the goals outlined in the agreement.

The Independent Review Commission (‘‘Kriegler Commission’’) investigated the ECK’s conduct of the elections and although it did not find evidence of systematic cheating, their September 2008 report found enough ECK incompetence to recommend its dissolution and reconstitution.

The Waki Commission investigated the post-election violence and has evidence implicating six cabinet members and five sitting MPs in planning and executing the violence. While the Commission has recommended a special tribunal to try
these suspects, leading members of both parties have so far rejected the report.

With Kibaki, Odinga, and Musyoka all in government, the absence of an opposition party undermines reform efforts. The Waki Commission charges that powerful members in government encouraged the post-election violence, but the parties are so far unrepentant and refuse to go forward to try these suspects. Some actions have been taken towards reforming the institutions targeted in the power-sharing accord, but the government has thus far moved slowly. Both ODM and PNU appear to be more concerned with preparing for the 2012 elections. In particular, battles are already underway for party leadership, especially in PNU where Kibaki cannot stand for a third term.