Monday, August 18, 2008




Jesus College, University of Oxford

ABSTRACT The importance of the Kenya crisis for the African continent is not that Kenya may become
‘another Rwanda’, but that it reveals how fragile Africa’s new multi-party systems may be when weak
institutions, historical grievances, the normalization of violence, and a lack of elite consensus on the
‘rules of the game’, collide. This paper provides an overview of the election campaign and the results,
highlighting the major areas of continuity and change with Kenya’s recent past, and identifying the key
dynamics within the presidential and parliamentary contests. In doing so, it focuses on the ongoing
process of coalition-building that underpinned both the rise and fall of NaRC and the emergence of
Odinga as a major political player. I suggest that an appreciation of the difficulties of coalition building
central to understanding the campaign and the results, and will be key to any long-term resolution of
the Kenya crisis.

The contrast could not have been greater. On 29 December 2002, Mwai Kibaki was sworn
in as Kenya’s third president in front of a million jubilant citizens in Nairobi’s Uhuru
Park. The Kenya African National Union (KANU) government had been swept from
power at the polls by Kibaki’s National Rainbow Coalition (NaRC); Kibaki had crushed
retiring President Daniel arap Moi’s anointed successor, Uhuru Kenyatta, in the
presidential election. On the podium, Kibaki was joined by his leading allies, notably
Raila Odinga who had masterminded the election campaign while Kibaki convalesced
after a road accident. A multi-ethnic alliance, which promised far-reaching constitutional
reform and robust action against corruption, seemed to have finally defeated entrenched
authoritarianism in Kenya.

On 30 December 2007, Kibaki’s second term began in a very different way. This time
the President was inaugurated in a hastily arranged ceremony at his official residence,
State House. Only an hour before, the presidential election result had been announced in
a small private room with only the state broadcaster (Kenya Broadcasting Corporation) in
attendance. Previous attempts to publicly declare the result had been abandoned amidst
chaotic protests against alleged electoral malpractice. Guests at Kibaki’s inauguration
were hastily assembled, many arriving late to be seen live on television scuttling across the
lawns of State House to take their seats while the President was already swearing the oath
of office. No one remembered to invite the band, and so no national anthem was played.

At that same moment, from Kisumu on the shores of Lake Victoria to the coastal city
of Mombasa, the first wave of violence began. But it was not just the spectre of violence

Correspondence Address: Nicholas Cheeseman, African Studies Centre, University of Oxford, 92 Woodstock
Road, Oxford OX2 7ND, UK. Email:

ISSN 1753-1055 Print/1753-1063 Online/08/02016619 # 2008 Taylor & Francis
DOI: 10.1080/17531050802058286

The Kenyan Elections of 2007: An Introduction 167

that differentiated 2007 from 2002. The cast of characters had undergone a significant
realignment. Supported by both former President Moi and his own 2002 opponent,
Uhuru Kenyatta, Kibaki had this time defeated his former key NaRC allies. His main
opponent in 2007 was Odinga, his most significant ally in 2002. Furthermore, Odinga’s
Orange Democratic Movement (ODM) represented a broad coalition dominated by a
combination of former allies of Kibaki, such as Charity Ngilu and Najib Balala, as well as
former Moi loyalists, such as Musalia Mudavadi.

Writing in late February 2008, any attempt to comprehensively introduce a special
edition examining the election of two months previous and the violent aftermath will
necessarily be incomplete. Indeed, this entire issue represents only a partial account of the
dynamics at play in Kenya, as its imbalances in local studies and particular thematic
interests leave many key issues unaddressed. For instance, there is much need for research
and analysis of the role of the Kenyan police during the violence and of the impact of
international actors during the mediation process. It is also too early to speculate about
whether a political solution will end the violence or if the current commitment to a
coalition government will hold. This volume does not act as judge and jury of electoral
malpractice in the December elections, nor is its intent prescriptive. Rather, our aim is
simply to provide a range of perspectives on the constellation of factors that together
underpinned the Kenya crisis . to offer a first attempt at understanding this most
complex, important, and devastating, of events. Much of the hyperbolic media coverage
has suggested that the past two months represent a unique period in Kenyan history,1 but
in the pages that follow, close observers of Kenyan politics argue that a more sober
analysis of the 2007 campaign reveals many continuities with the recent past.

The importance of the Kenya crisis for the African continent is not that Kenya may
become ‘another Rwanda’, but that it reveals how fragile Africa’s new multi-party systems
may be when weak institutions, historical grievances, the normalization of violence, and a
lack of elite consensus on the ‘rules of the game’, come together to form what Mwangi and
Holmquist have called a ‘perfect storm’.2 Significantly, many of Africa’s other fragile
democracies share some of Kenya’s most significant political characteristics. Salient
ethno-regional identities reinforced by historical grievances over land ownership,
economic inequality, and political exclusion, are central to an understanding of the
Kenya crisis, and are present in one form or another in countries as diverse as Namibia
and Cote d’Ivoire.3 At the same time, the spread of the violence in Kenya owed much to
the informalization of the state, the diffusion of violence beyond central control, and the
rise of militias connected to the political elite; all developments that will be familiar to
students of Nigeria and the ‘Collapsing States’ identified by Reno.4 Furthermore, the lure
of the imperial presidency, and the difficulty of achieving elite agreement on key
constitutional questions . which Chege has termed the ‘elusiveness of cohesion’ . was a
major contributor to the elite deadlock which followed the election, and is a problem that
has also bedevilled a number of other countries including Zambia and Uganda. Finally,
the difficulty of forming an effective coalition government from a weak, fluid, and heavily
divided party system, was an issue that bookended the Kenya crisis and has also
hampered effective governance in Benin and Malawi.5

This introduction does not attempt the impossible task of individually introducing the
15 diverse papers that comprise this volume, many of which do an impressive job of
connecting the events of the last two months to the historical development of Kenya’s
political economy. Rather, I aim to set out some of the main themes of the Kenyan

168 N. Cheeseman

election of 2007, beginning with a brief overview of its more distinctive features: the
increasing salience of religion and age, the impact of opinion polls, the rise of grass-roots
party organization, and the role of technology and rumour in shaping public
understandings of the election. I also provide an overview of the results based on the
official figures,6 in order to contextualize the subsequent contributions for readers
unfamiliar with Kenya. However, I am most concerned to demonstrate the structural
foundations of the 2007 elections, and to connect this to the ongoing process of coalition-
building that underpinned both the rise and fall of NaRC and the emergence of Odinga as
a major political player. I suggest that an appreciation of the difficulties of coalition
building is central to any understanding of the election results and will be key to any
long-term resolution of the Kenya crisis.

Continuity and Change in Kenyan Politics

Even while stressing the long-term roots of the Kenya crisis, the distinctive features of the
2007 polls should not be forgotten. Ethnicity may have stolen the headlines, but this
election saw the rise to prominence of a number of social cleavages previously thought to
be marginal to Kenyan voting habits. Following Odinga’s signing of a Memorandum of
Understanding (MOU) with representatives of the Muslim community, religion became a
mobilizing issue. Both parties attempted to spin the debate to manipulate the religious
cleavage to their own advantage.7 Moreover, religious leaders participated in the elections
to a previously unseen extent.8 Age also became an important predictor of voting
behaviour, suggesting that in 2007 the Kenyan ‘youth’ finally lost patience with the older
generation.9 Six million voters joined the voter register and the numbers of youths under
the age of 35 increased by 57 per cent.10 All the main parties and presidential candidates
made concerted efforts to appeal to these new voters. Odinga, presenting himself as the
‘change’ candidate, proved more successful. Opinion polls consistently found that the
electorate was split, voters over 50 predominantly supporting Kibaki and those aged 18.
35 overwhelmingly supporting Odinga.11 If ethnic tension was the main story of the 2007
election, then generational competition was a crucial sub-plot. The defeat of many of the
country’s most prominent political dynasties by younger challengers suggested the
potential for a real transformation of the identities, if not the strategies, of Kenya’s
political elite. Nowhere was this trend better illustrated than in the failure of the Moi
family, Simeon Nyachae, and Musikari Kombo, to mobilize support for Kibaki in their
home areas.

Of course, the 2007 election campaign was also different because at the presidential
level it was such a close race. The competitiveness of the two main candidates, combined
with the rise of an independent opinion polling industry over the last five years, gave
prominence to opinion polls which dominated the front pages in the final months of the
campaign. These polls, conducted by companies including the Steadman Group, Infotrak
Harris, Consumer Insight, and Strategic Research, were painstakingly pored over by voter
and aspiring politicians alike. The opinion polls themselves became the subject of great
controversy and disagreement. No Kenyan politician was prepared to let an unfavourable
poll go unchallenged, in large part motivated by the fear that rival candidates would gain
a momentum boost, and a potential bandwagon effect, from favourable poll headlines.12
Significantly, while polls conducted by Steadman and Consumer Insight matched many
observers’ assessment that the election was too close to call, polls by Strategic Research

The Kenyan Elections of 2007: An Introduction 169

and Infotrak Harris using smaller and more tightly clustered samples consistently gave
Odinga a sizeable majority.13 These misleading polls contributed to the disappointment
and outrage at the declaration of a Kibaki victory in pro-Odinga areas. A similar
phenomenon occurred in Zambia in 2006, when polls mainly conducted in urban areas
predicted a landslide victory for opposition leader Michael Sata, prompting incredulity
among his supporters when the incumbent, Levy Mwanawasa, was declared the victor.14
Opinion polls play an important role in representing the ‘voice’ of ordinary people, but
where polls are not scientifically conducted or are poorly reported they may contribute to
political instability.15

Despite bullish public pronouncements, the campaign teams on both sides appreciated
how tight the race was: increasing turnout by 1 per cent might have been enough to secure
victory. Party activists on both sides feared that apathy and overconfidence in their home
areas would depress turnout, so employed strategies designed to ensure ‘100 per cent
turnout’. This involved forming local ‘cells’, each given responsibility for a small number of
polling stations. Party agents took a systematic approach to identifying and targeting
marginal, undecided, and infirm voters to make sure that they made it to the polling
stations.16 Voter mobilization was also sometimes coercive. In Central Province and
Nyanza voters were informed that they would be denied access to shops and transport if
they could not show their ‘inky finger’ as proof that they had voted.17

This was also Kenya’s first election in the Information Age. The spread of access to
technology and the youthful electorate ensured that the battle for votes, and then lives,
was fought through the internet, emails and (especially) SMS text messages. Rumour
played a central role in local ‘understanding’ of the unfolding political events, and the
spread of telecommunications capacity enhanced the power of rumour.18 Significantly,
the communicative capacity of the ordinary citizen exceeded by far even the most
intrusive efforts by the state to control flows of information, as was demonstrated by the
refusal of Kenya’s main mobile phone companies to cease delivering SMS text messages.
When rumours began circulating in early January that the Chief of Police and the head of
the Kenyan army had resigned in protest at Kibaki’s handling of the election, the rapid
spread of the ‘news’ via text message forced the government to hold a press conference to
deny the claims.19

These new means of communication could be used to pass truth, rumour, and
deliberate accusations. Thus, while the main media outlets adopted a responsible tone
both before and during the violence, the use of new technologies to circulate
unsubstantiated allegations and falsified documents ensured that this was an election
characterized by misinformation. Prior to Election Day, many forged documents were
‘leaked’ onto the internet to discredit presidential candidates, the most famous example
being a fake Memorandum of Understanding between Raila Odinga and Muslim leaders .
intended to suggest to Christian voters that Odinga would turn Kenya into a Sharia state.
The real MOU, an extremely anodyne document, simply committed Odinga to
devolution and greater public spending on the poor coastal and northern regions.20
After the polls, these same media were filled with both messages of hate and pleas for
reconciliation. Mobile phones were used to mobilize violence, and to coordinate relief.
More forged documents were generated, to persuade domestic and international
observers that the blame for the post-election violence should be placed on political
rivals.21 Never before have Kenyan voters been privy to so much election-related
information, and never before has it been harder to sort fact from fiction.

170 N. Cheeseman

Going by the media coverage of the Kenya crisis, one would think that both the postelection
violence and the mediation efforts of international actors were dramatic new
developments in Kenyan politics which represented a sharp break from a stable and
peaceful past.22 However, while the conflict has been tragic, it is far from unprecedented.
By the end of February 2008, the violence had claimed over 1,000 lives and left over
500,000 displaced. In the ethnic clashes of 1992 and 1997, over 3,000 lives were lost, and
over 300,000 displaced. The greater intensity of the recent violence is undeniable, yet the
‘ethnic clashes’ of the 1990s were the forerunners of the present crisis. The large-scale
instrumental use of political violence, structured along ethnic lines and fuelled by notions
of majimbo, was introduced to Kenya by Daniel arap Moi and then fostered by the
willingness of political leaders of all persuasions to sponsor and tolerate the existence of
militias.23 The strong continuities from 1992 become more obvious when one realizes
that, contrary to the general depiction of the sequencing of the 2007 election, violence
was one of the themes that ran right through the whole campaign.

Most reports on the election violence begin their analysis on 30 December, failing to
connect the pre-and post-poll periods.24 Yet there were a number of dramatic examples
of political violence in the last few months of 2007. In October, party nominations
resulted in violent skirmishes between supporters of the same party, with supposed allies
coming to blows.25 Although the police and the Electoral Commission of Kenya (ECK)
had to be brought in to calm the situation, the unrest was not considered to be
exceptional by either political leaders or the media. In November, police searching for
members of the Mungiki gang are alleged to have executed an estimated 500 suspected
Mungiki members without trial during raids on slum areas around Nairobi.26 The killings
sparked outrage among Kenyan human rights groups, but the national media and the
international community proved remarkably apathetic.27 In early December, violence
flared up in a number of locations, most notably in Mt Elgon, claiming 70 or more lives
and displacing thousands more.28 Despite evidence of the involvement of local politicians
in inciting the violence, an ECK hearing issued fines but allowed the guilty to continue
campaigning.29 In total, almost 600 people were killed in the three months before polling
day.30 That this death toll received little comment speaks to the remarkable normalization
of violence within Kenyan political life.

The coverage of the post-election violence has tended to homogenize an extremely
heterogeneous process; there is not one battle, but many overlapping conflicts with
differing motivations and dynamics31: the violence between Kalenjin and Kikuyu
communities in Eldoret is very different from the violence perpetrated by the state in
Kisumu, and the violence between Luo residents and Kikuyu members of the Mungiki
gang in Kibera bore no relation to the struggles that engulfed Mt Elgon. It is important to
underline that, though political violence was triggered by the election, the spread of the
conflict reflects long-term popular frustrations. The Kenya crisis needs to be placed in the
context of local understandings of citizenship, belonging, and exclusion.32 The anger of
ODM supporters at the perceived theft of the 2007 election cannot be separated from
their perception that they have been excluded from the political process for many years;
the declaration of Kibaki as president was such a powerful trigger for the violence
precisely because it tapped into a rich mine of strong historical grievances.

In order to fully understand the dynamics of the conflict, and the state’s inability to
respond, we must connect these local debates and frustrations to the way in which the
Kenyan state structures both political competition and the means of coercion. The

The Kenyan Elections of 2007: An Introduction 171

structural foundations of the Kenya crisis lie in the state’s loss of the monopoly over the
legitimate use of force, the inadvertent diffusion of violence outside of central control, a
lack of credible institutions autonomous from the executive, and the focus on an ethnic
winner-takes-all view of politics in which control of the presidency becomes the only
prize that matters.33 These processes are not unique to Kenya, and comparisons to other
cases in which formal institutions have been deliberately weakened and violence has
become normalized are instructive; for example, Reno’s argument that in Nigeria violence
‘comes from the collapse of the regime’s patronage-based monopoly over commerce,
coupled with elite co-optation of armed groups’, has many synergies with Mueller’s
understanding of the conflict in Kenya.34 As elsewhere, a long-term ‘solution’ to the
Kenya crisis will need to address the rise of militias and their connections to the political
elite as well as issues of coalition politics, constitutional reform, and truth, justice, and

Does the mediation of the Kenya crisis represent a new contribution by international
actors to Kenyan democracy? It is certainly true that pressure from the US and the EU has
contributed to Kibaki’s willingness to compromise, and few would deny the important
role played by Kofi Annan in bringing the two sides together. But the role of foreign
governments is complex. George Bush was the first foreign leader to congratulate Kibaki,
and it initially appeared that the United States intended to endorse the election results.35
It was only after the EU and domestic observers highlighted the blatant examples of
electoral manipulation that the US adopted a more critical tone. The initial acquiescence
of the US reflects Kenya’s geo-strategic import and America’s desire to retain a key ally in
the war on terror. This suggests continuities to the ‘realist’ foreign policy that saw many
governments support Moi’s authoritarian regime in the 1980s.

International commitment to promoting free and fair elections has at times been
compromised by concerns for regional stability. The best example of this is the farce
surrounding the suppression of the exit poll conducted by the International Republic
Institute (IRI). The official IRI explanation is that the poll, which interviewed 5,500
respondents in 179 constituencies, has been pulled because the ‘IRI does not have
confidence in the integrity of the data and therefore believes the poll is invalid’.36
However, opposition supporters and human rights leaders allege that the poll was not
released because it found evidence of an overwhelming victory for Raila Odinga.37 The
poll was co-ordinated by James Long and Clark Gibson, researchers at the University of
California, San Diego. According to Gibson, the poll ‘conforms to social science
standards’ and there is no methodological reason for the IRI not to release the findings.38
The initial reluctance of the US government to criticize the polls, and the decision of IRI
to bury the exit poll findings, mirror the complex mix of motivations that have
underpinned the interventions of international actors over the past 15 years, as
documented in the work of Stephen Brown.39 While the role of international actors
may have been more prominent in early 2008, it is far too early to suggest a radical break
with the murky compromises of the past.

So, while much was new about the recent elections, we should resist the temptation to
see the crisis as a new phase of Kenyan political history. We will understand the process
much better if we examine the traces of the past that are so visible today. The elections of
2007 marked the fiftieth anniversary of the beginning of experiments with democracy in
Kenya. At times during the two-year long preamble, which stretched back to the defeat of
the government in the constitutional referendum of 2005, the historical hinterland

172 N. Cheeseman

seemed set to define this election. The language of ‘tribalism’, embedded with historical
grievances, dominated debates at the grassroots. The final rush to the start-line of the
official campaign period witnessed the now mandatory rapid turnover in party names as
fragile coalitions were broken and made. Other features of past elections remained in
place. The overwhelming ‘home’ support for presidential candidates in their provinces of
origin, rumours of vote buying, and a chaotic nominations process prior to the polls
suggested this was a very Kenyan election. The names of many of the most prominent
individuals in the campaigns were familiar to any student of post-colonial Kenya .
Kibaki, Moi, Kenyattta, Nyagah, Odinga, Awori and Pattni . only the generation had
changed in some cases. Even Kenneth Matiba made another sad bid for the presidency.
One could not escape the sense that a stale cadre of political leaders of the country was
out of kilter with a youthful electorate. However, it was not just the identities of the
politicians that reminded one of the past, but also the way in which the business of
politics was conducted.

Coalitions and Campaigning

A key theme which runs through this edition . and which has underpinned high politics
since the re-introduction of multi-partyism . relates to the challenges of coalition
building. Kenya’s political geography ensures a tendency towards coalition politics. The
country is extremely ethnically diverse, with no one ethnic group comprising more than
22 per cent of the population. Parties do not simply map onto ethnic groups and the 2007
election cannot simply be understood as an ethnic census.40 However, many regional
leaders, empowered through ties of ethnicity, lineage, and neo-patrimonial networks
maintained by their vast personal wealth, have proved able to mobilize vast communities,
even in the absence of formal party structures.41 For example, according to the official
results for the 2007 election, Kibaki secured over 95 per cent of the votes cast in 26
constituencies in his Central Province base, while Odinga repeated the feat in 21
constituencies in Nyanza. The capacity of regional leaders to construct extremely
personalized political machines means that aspiring presidential candidates must recruit
the support of a number of ethno-regional leaders if they are to emerge victorious at the
ballot box. Consequently, coalition politics is a dominant feature of Kenyan political life.

However, putting together an effective campaign team is far easier said than done . the
refusal of many political leaders to accept subordinate positions, combined with the
failure of victorious candidates to construct power sharing structures, means that Kenyan
political alliances have a tendency to fall apart. In large part, it was the fragmentation
of the opposition which enabled Moi and KANU to retain power until 2002. What
rendered Kibaki vulnerable in 2007 was that despite enjoying the benefits of incumbency
and a strong economic performance, he failed to construct a viable alliance to replace

Aside from widespread popular frustration with the incumbent KANU regime, NaRC’s
victory in 2002 was built on three key foundations.42 Firstly, the leaders that comprised
NaRC were senior political figures of some standing. Having contested for the Presidency
in 1997, Charity Ngilu, Raila Odinga, and Mwai Kibaki, could all claim to be genuine
political heavyweights. Secondly, NaRC represented a broad and inclusive alliance. On the
one hand, the party brought together a broad range of ethno-regional communities
including the Kikuyu, the Luo, and the Kamba. On the other, NaRC represented a

The Kenyan Elections of 2007: An Introduction 173

mixture of the old and the new, illustrated by Odinga’s vigorous campaigning on behalf of
the ‘elder statesman’ Kibaki. Finally, NaRC leaders were able to arrive at a pre-election
agreement to support Kibaki for the Presidency, and to unite behind a single list of
parliamentary candidates, on the basis of the now infamous MOU which set out how
power would be shared if NaRC proved victorious. In short, NaRC’s victory in 2002 was
based on the effective integration of a diverse alliance into a multi-ethnic movement for

However, immediately following the election the alliance fragmented over the
implementation of the MOU, most notably around the question of constitutional review.
The failed review process had three significant consequences. 43 Firstly, the election of
2007 was contested under the highly centralized constitution inherited from the one-
party era. Secondly, Kibaki’s manipulation of the process undermined what little trust
had developed between the Kenyan political elite. Finally, many of the key members of the
NaRC coalition, including Odinga, Ngilu, and Kalonzo Musyoka, moved into opposition
to campaign against the government in the constitutional referendum. The fragmentation
of NaRC necessitated another round of coalition building. Lacking sufficient firepower in
his remaining faction in NaRC to mount a credible campaign on his own, the President
was forced to fall back on old friends, forming an electoral pact with the Kenyan African
National Union (KANU), FORD-Kenya, and a number of smaller parties including Safina
and New FORD-Kenya.

Kibaki’s new alliance was significantly weaker than NaRC in at least three ways. Firstly,
the loss of Odinga, Musyoka, and Ngilu, left Kibaki with a much narrower ethno-regional
support base which was heavily reliant on Central Province. Not all Kenyan communities
vote en bloc, and not all political outcomes are determined by ethnicity, especially at the
parliamentary level.44 However, at the presidential level Kenyan communities have
historically proved unwilling to vote for a party that does not contain any of their aspiring
leaders within its ranks. Secondly, forming an alliance with ex-president Daniel arap Moi
and his controversial ally Nicholas Biwott highlighted Kibaki’s failure to break with the
past and to tackle corruption, and gave the impression that the President was out of
touch with the mood of the country. It was in part this failure which enabled Odinga both
to present himself as the ‘change’ candidate and to mobilize younger voters who were
frustrated at the refusal of the old guard to ‘move over’.

Finally, and most significantly, Kibaki failed to integrate the various components of his
alliance into an effective election vehicle. At the presidential level KANU, FORD-K,
NaRC-Kenya, and Safina, all agreed to support Kibaki’s candidacy under the Party of
National Unity (PNU) banner. Although senior PNU figures initially declared that the
party would only field one common candidate in each election, coalition members
including KANU rejected this plan, recognizing that having their ‘own’ MPs in parliament
would give them greater bargaining power in the event of a close election. After a
particularly heated meeting in early October housing minister Soita Shitanda stated that
having decided not to work with the ODM because ‘they had insisted we dissolve our
parties first’, his New FORD-Kenya party would not agree to a common list of
parliamentary candidates under any circumstances because it would mean ‘the death of
other political parties’.45 Significantly, Kibaki had already become too dependent on his
new partners to insist, and as a result PNU affiliated parties ran parliamentary candidates
against each other.

174 N. Cheeseman

Central to PNU’s difficulties was the failure to manage the process of nominations. In
many cases, the nomination procedure proved intensely divisive, with three or four
credible candidates demanding selection in the most attractive seats. Amid widespread
claims of favouritism and electoral malpractice, the nominations procedure exacerbated
the tensions within PNU. Losing candidates sought out smaller parties to represent,
resulting in a proliferation of the number of competing pro-Kibaki candidates dividing
the ‘PNU’ vote at the parliamentary level.46 In total, 11 parties affiliated to the PNU won
seats in the 2007 election, including KANU, Safina, NaRC-Kenya, FORD-K, Sisi Kwa Sisi,
the DP, Kenda, FORD-A, FORD-P, New FORD-K, and the Mazingira Green Party. In
some cases, KANU, NaRC-Kenya, PNU, the DP, and FORD-K all contested for the same
parliamentary seat. The title ‘party’ is misleading for some of the smaller entities . many
were little more than the ‘paper parties’ described so evocatively by Ngugi wa Thiong’o.47
Nonetheless, these highly personalized political machines proved able to mobilize large
constituencies, playing a significant role in dividing the ‘pro-Kibaki’ vote.

Following a series of public clashes in early December between parties and senior
leaders within the PNU alliance, the Sunday Nation concluded that ‘the rivalry is hurting
the President’s campaign’.48 In extreme cases the infighting escalated into an ‘internal
war’, with candidates embarking on smear campaigns against their supposed allies.49
Labour Minister Dr Newton Kulundu was reportedly ‘furious that a Cabinet colleague in
Western Province is sponsoring violence and spoiling PNU’s chances of winning the polls’
because ‘he wants to become the political king in the region’.50 Similarly, rivalry between
Cabinet Ministers Mukhisa Kituyi, and FORD-Kenya leader Musikari Kombo served to
seriously disrupt PNU activities, and meant that more energy was ‘spent on personal
survival than fighting off the ODM wave.’51 On one occasion an ‘explosive encounter’
between Kituyi and Kombo erupted during one of Kibaki’s campaign tours, forcing the
President to divert time from campaigning to act as peacekeeper for his own troops.52

In one of the more bizarre twists of the election campaign, the main parties associated
with the PNU claimed that their membership of the PNU alliance entitled their
parliamentary candidates to receive PNU funds. It is indicative of Kibaki’s reliance on his
new partners that on 4 December he agreed to provide KSh. 250,000 for each
parliamentary candidate in the PNU alliance, half the KSh. 500,000 given to each official
PNU candidate.53 Given that the total number of parliamentary candidates standing for
affiliated parties was more than double the number on the PNU ticket, this committed the
party to spending more on ‘rival’ parliamentary candidates than on it’s own. The failure
to effectively integrate the PNU coalition resulted in the waste of campaign resources,
created opportunities for ODM to prosper at the parliamentary level, and undermined the
attempt of PNU campaigners to present a united front in support of Kibaki.54

In contrast to the chaos of the PNU campaign, Raila Odinga proved adept at knitting
his patchwork quilt of diverse allies into an effective organization. The group of dissident
leaders that left NaRC in 2005 over the question of constitutional reform united to
campaign against the draft constitution under the banner of the ODM. Strengthened by
the defection of William Ruto, following splits within KANU, and emboldened by the
defeat of the Government in the 2005 constitutional referendum, Odinga and Musyoka
agreed to transform ODM into a political party to contest the 2007 elections. The Orange
grouping itself then split over a leadership contest which neither Odinga nor Musyoka
were willing to lose. Two factions . Odinga’s ODM and Kalonzo Musyoka’s ODM-Kenya
(ODM-K) . registered to contest the polls. The split mirrored the division of the

The Kenyan Elections of 2007: An Introduction 175

opposition Forum for the Restoration of Democracy (FORD) into FORD-Asili and
FORD-Kenya in 1992, enabling Moi to remain in power. As in 1992, it is clear that had
Odinga and Musyoka remained united, defeat of the incumbent would have been assured.

Despite the loss of Musyoka, ODM emerged from the process of candidate selection in
a stronger and more unified position than PNU, especially at the parliamentary level.
While the nominations process did present difficulties,55 ODM never fragmented in the
manner of PNU; the only ODM-affiliated parties to win seats in 2007 were NaRC and the
United Democratic Movement (UDM). Odinga designed a well co-ordinated three-tiered
party structure which empowered distinct campaign teams to operate at the presidential,
provincial, and constituency level. Most significantly, Odinga’s campaign was centred on
ODM’s so-called Pentagon of regional leaders selected for their ability to deliver support
across the country . Mudavadi (Western), Ruto (Rift Valley), Nyaga (Eastern), Ngilu
(Eastern) and Balala (Coast). Rather than the glorious isolation of Kibaki’s campaign,
Odinga campaigned as first among equals.

As well as being more united, ODM’s campaign was more energetic and more creative.
Although ODM was outspent by PNU and received far less coverage in the state media,
the ODM campaign reached parts of Kenya that other parties could not.56 In part, this
was due to ODM’s carefully targeted message. Rather than release one manifesto for the
whole country, ODM devised a series of regional manifestos specifically tailored to
address local debates in each province. Odinga reached out to specific communities,
outlining policies designed to meet local grievances, most controversially signing a
memorandum of understanding with the National Muslim Leaders Forum.

Odinga’s considerable organizational capacities and passion for electioneering underpinned
the effectiveness of the ODM campaign. However, it was the perception of a
common interest that united the disparate factions of the ODM alliance. By creating a
coalition of groups which believe they have been historically disenfranchised and
promising to free them from central domination . which in many cases was interpreted as
Kikuyu domination . Odinga effectively gave his alliance partners common cause. ODM’s
most significant election pledge, which dominated the campaign, was to introduce a
majimbo (regionalist) constitution that would transfer power away from Nairobi to the
provinces. Majimbo is a heavily loaded term in the Kenyan context, and has been used by
politicians of different persuasions to refer to both greater local autonomy and to the use
of ‘ethnic cleansing’ to bring about a transition to mono-ethnic locations.57 By playing on
the widespread demand for devolution of power, and by campaigning on a slogan that
many local populations interpreted as promising greater control over land, Odinga was
able to mobilize a support base with little history of unity. However, the call to majimbo
motivated Odinga’s enemies as well as his allies; PNU supporters’ fear of what an Odinga
presidency would mean for Kenya was demonstrated by the frequent comparisons
between the ODM leader and Idi Amin.58

Opinion polls conducted over the last few years reveal the extent to which Odinga’s
coalition building exercise transformed him from political hopeful to presidential front
runner. When Odinga ran for the Presidency in 1997, he received just 10.8 per cent of the
vote. In 2005 opinion polls placed his national support at around 8 per cent, and many
commentators suggested that he was unelectable. But in the course of 2005 and 2006
Odinga’s ability to attract, organize, and mobilize communities across Kenya transformed
him into the favourite for the presidency. As Figure 1 shows, the final Steadman poll,
conducted on 19 December 2007 and based on 6,100 interviews in every district of the

176 N. Cheeseman

Figure 1. Presidential popularity by province. Source: Daily Nation, ‘With Eight Days to Go, Race
Too Close to Call’, 19 December 2007.

country based on a nationally representative sample, placed Odinga in the lead in six of
Kenya’s eight provinces.59 Towards the end of the campaign all of the main polling
companies suggested that Odinga’s greater spread of support had engineered a small but
robust national lead: however, the polling companies with larger sample sizes and more
dispersed interview locations concluded that the presidential contest was simply too close
to call. The final Steadman poll placed Odinga on 45 per cent of the vote, with Kibaki just
behind on 43 per cent. While outside of the margin of error for that poll, Steadman were
well aware that variations in turnout could overturn Odinga’s narrow lead.60 Given that
Kibaki’s strong areas of Central Province have historically turned out at higher levels than
Odinga’s strong areas of Nyanza and the Coast, a victory for either candidate was
perfectly plausible.

The Results

The count itself became a piece of major political theatre, and as the different acts of the
play unfolded, accusations of electoral manipulation mounted and the likelihood of postelection
unrest escalated.61 On the basis of the first batch of declared results, which
overwhelmingly came from ODM strongholds, Odinga supporters initially thought that
their man was on his way to State House. However, as results from constituencies
favourable to Kibaki began to be announced by the ECK, it seemed that Kibaki was set for
a narrow re-election. Despite the protestations of the ODM and the EU that irregularities
in the results had not been sufficiently investigated, the chair of the ECK, Kivuitu,
eventually declared Kibaki the winner with 4,584,721 to Odinga’s 4,352,993. In the days
following the declaration the Kenyan National Human Rights Commission (KNHCR)
and the domestic observer group (KEDOF) held press conferences to voice concerns
about the presidential result.62 Most damning was the report of the EU Election
Observers Mission, the largest international observer group, which documented a
number of startling verified irregularities.63 In Molo and Kieni constituencies the figures
given to EU observers were over 20,000 lower than the figures later released by the ECK.
In a further eight Central Province constituencies the EU observers were not even allowed
in to monitor the vote count. The head of the EU observation team personally reported
having seen result forms for the Lari and Kandara constituencies in which the final

The Kenyan Elections of 2007: An Introduction 177

numbers had been changed to favour Kibaki. In both cases the ECK had failed to provide
an adequate explanation for the vote inflation.

Neither side had fought an entirely clean campaign. Suspicions that Odinga’s vote may
have been artificially increased by his supporters in Nyanza Province seem to have been
confirmed by reports on NTV on Thursday 10 January that the ODM’s own ‘audit of the
polls’ file admitted that the vote for Odinga was inflated in at least 10 constituencies.
Despite the fact that both sides engaged in vote inflation, the decision of four ECK
Commissioners to express their own doubts about the electoral process, followed by
Kivuitu’s explosive admission that he was placed under great pressure to release the
results and that even he did not actually know who had won, convinced ODM supporters
that Kibaki allies had systematically manipulated the polls at the very highest level of the
electoral process.64 The critical statements by domestic observer teams also made front
page news, and the combined effect was to instantly delegitimize the new government.

A review of the official ECK figures is consistent with the interpretation that there was
rigging on both sides. It seems, however, that electoral manipulation was most extensive
in locations loyal to the PNU. Vote inflation is usually indicated by exceptionally high
levels of turnout and a discrepancy between the parliamentary and presidential ballot.
Identifying all constituencies in which turnout stood at over 85 per cent, an arbitrary cutoff
but one that reflects those locations where turnout is unusually high given the
national average, generates a list of 25 constituencies within which there is no dominant
pattern. Nine of these constituencies were in Central Province, and Kibaki was the leading
presidential candidate in 10. However, 11 of the constituencies were in Nyanza Province,
and Odinga was the leading candidate in 15. In Mbita and Karachuonyo constituencies in
Nyanza, turnout was recorded at an unprecedented 95 per cent.

The pattern of discrepancies between presidential and parliamentary elections is more
suggestive.65 Given that Kenyans vote simultaneously in the parliamentary and
presidential elections, and almost all voters cast both ballots, there should be a roughly
similar number of total presidential and parliamentary votes.66 Large discrepancies
between the two ballots are therefore indicative of the inflation of the presidential vote.
Overall, there were 206,897 more ballots cast in the presidential than the parliamentary
one, representing a total discrepancy of 2 per cent.67 In 19 constituencies there were over
10 per cent more presidential ballots cast than parliamentary ones (again this is an
arbitrary threshold but it highlights those constituencies in which there is strong evidence
that votes were added to presidential totals). These 19 constituencies account for 165,333
of the total additional presidential vote. Of these, a majority (12) were won by Kibaki,
including three constituencies in Central Province. Only six were won by Odinga, none of
which were in Nyanza.68 The evidence therefore suggests that the majority of vote
inflation took place in PNU areas.

But we will never know for sure: calculating ‘real’ vote totals for Odinga and Kibaki
based on official ECK figures and information from election observers is not feasible. Nor
is it possible to reconstruct the election by conducting a re-count or a re-tallying because
of a widespread failure to follow basic procedures. In some constituencies, ballot boxes
were not kept in secure locations. In others, they were destroyed in the violence that
followed the polls. Many of the forms recording inconsistencies, which were sent to the
ECK by returning officers, have since gone missing. Furthermore, in some cases there are
a number of different versions of each form circulating, making it difficult to tell the real
from the fake.

178 N. Cheeseman

Sadly, it now looks like the best evidence available bar the actual ballots themselves, the
IRI exit poll, will never be officially released. It is widely believed that the exit poll found
that Odinga won the election by 7 per cent.69 Adjusting these figures to weight for actual
voter turnout using official ECK figures, the poll found that Odinga would have won by

2.7 per cent. In other words, even taking the ECK turnout figures . which provide a net
benefit to Kibaki . at face value, the exit poll predicted an Odinga victory. It is important
to note that there are many problems with conducting an exit poll in the Kenyan context.
Many Kenyans feared reprisals for voting the ‘wrong way’, and Kibaki supporters believed
that they were more likely to be the target of violence.70 As a result, it is plausible that
more Kibaki voters than Odinga voters either refused to answer or misrepresented their
voting behaviour.71 Nevertheless, the low refusal rate of respondents and the wide margin
between Kibaki and Odinga recorded in the poll are suggestive of an ODM victory.
The presidential and the parliamentary contests were very different elections. Although
there are 22 election petitions outstanding, election observers are agreed that the
parliamentary polls were significantly more free and fair.72 As expected, the ODM
dominated the parliamentary contest, although the party was unable to secure an
outright majority in the legislature. With the nominated seats appointed by party leaders
on the basis of party size, ODM won more than twice the number of seats that the PNU
took (Table 1). ODM’s strength in the parliament remains even if we take into account
the groupings of parties that have historically been allies or have entered into postelection
pacts. Counting KANU, FORD-K, Safina, NaRC-Kenya, along with a number of
smaller ‘one seat’ parties, brings the total number of ‘PNU affiliated’ MPs up to 78. Even
after ODM-Kenya had been persuaded to join the government, adding another 18 seats,
PNU could only rely on the support of 96 MPs. Including the seats won by NaRC and
UDM, ODM commands 109 seats in the legislature, giving the ‘opposition’ a clear
advantage . although the government has since proved able to recruit additional support
within the legislature when necessary.

Some commentators have taken ODM’s dominance at the parliamentary level to be
evidence of vote rigging at the presidential level, on the basis that there appears to be a
massive disparity between the support for ODM in the two elections. However, this is a
misleading interpretation of the results: many Kenyans who voted for Kibaki in the
presidential race did not vote for a PNU candidate in the parliamentary race. A careful
analysis of the results suggests that ODM’s parliamentary dominance is perfectly
compatible with an incredibly tight presidential contest.

The self-destructive nature of the PNU campaign was most evident in Kenya’s fiercely
contested multi-ethnic constituencies in urban locations. The parliamentary contest in
Embakasi is a perfect example. The ‘pro-Kibaki’ vote was divided between the President’s

Table 1. Distribution of parliamentary seats, largest five parties

Party Seats Nominated Total Total %

ODM 99 6 10547
PNU 43 3 46 21
ODM-K 16 2 18 8
KANU 14 1 15 7
Safina 5 . 52
Vacant seats 3 . 31
Others 31 . 31 14

The Kenyan Elections of 2007: An Introduction 179

current electoral vehicle, PNU, and his old one, the Democratic Party (DP). The three
main PNU-affiliated candidates in the constituency . David Mwenje, Ferdinand Waititu,
and John Ndirangu Kariuki . were all significant political players. Former Nairobi Mayor
Kariuki secured the nomination on the PNU ticket, and received 21,153 votes. Former
Nairobi Deputy Mayor Waititu fared better, receiving 28,833 votes having decided to
stand on the DP ticket after losing out in the PNU nominations. Finally, the controversial
incumbent MP and sometime sponsor of Mungiki, David Mwenje, polled 3,197 having
moved from the PNU to the National Reform Congress Party and finally to Chama Cha
Mwananchi.73 The total vote of the three main PNU affiliated candidates was 53,183.
ODM candidate Melitus Were, who was later assassinated during the post-election
violence, won the seat with just 35,345 votes.

The example of Embakasi clearly demonstrates a number of key features of the Kenyan
election. Firstly, Kibaki and his associates failed to imbue the PNU coalition with any
great meaning and even traditional Kibaki supporters felt little impulse to buy into the
emperor’s new clothes. Constructed in a rush and lacking any real sense of purpose, many
PNU candidates lost out to ‘pro-Kibaki’ rivals from smaller parties with far fewer
resources; the ability of Safina candidates to defeat official PNU candidates in Central
Province typifies this trend. Secondly, Embakasi shows how the deep divisions in the
PNU’s campaign laid the foundation for ODM’s electoral success. Of course, we must
always be careful when dealing with counterfactuals, and we cannot simply assume that a
single ‘PNU’ candidate would have received all of the votes that accrued to Mwenje,
Waititu, and Kariuki. Nonetheless, the magnitude of the vote accruing to PNU affiliated
candidates suggests that the ODM would have been unlikely to defeat a more unified
PNU parliamentary campaign.

Finally, the Embakasi example illustrates the point that the poor showing of PNU at the
parliamentary level tells us little about the presidential race. A simple reading of the
Embakasi results might suggest that it is peculiar that Kibaki could secure 72,376 votes to
Odinga’s 50,001 in a constituency where ODM won the parliamentary election. However,
if, as one would expect, the supporters of Mwenje, Waititu, and Kariuki, voted
overwhelmingly for Kibaki in the presidential election, then Kibaki should have won a
comfortable majority of the presidential votes cast in the constituency. In short, a closer
inspection of the parliamentary vote suggests that, just like the presidential contest, at the
parliamentary level support for the government and the opposition was finely balanced.
While there were also cases in which a more unified ODM campaign would have taken
seats from PNU, on balance a more centralized party system would have generated a net
gain for the PNU.

Looking at all 210 constituencies, and assuming that votes would have been transferred
from the affiliated party’s candidate to an official ‘unified’ candidate, the results suggest
that divisions within the PNU may have cost the party around 60 seats.74 It is important
to note that 35 of those seats were actually won by PNU affiliated parties in the election,
and would simply have been transferred to direct PNU control. However, the remaining
25 constituencies are cases where the party was likely to have been successful had its
affiliates not divided the vote. Remarkably, even against a wholly united ODM (including
ODM, ODM-K, and all affiliated parties), a united PNU may have been able to win an
additional 55 seats, 19 of which were actually won by the ODM in the election. While the
split in ODM had significant implications at the presidential level, it was far less
significant than the divisions within the PNU in terms of the parliamentary vote.

180 N. Cheeseman

As crude as these projections are, they provide a necessary corrective to the view that
there is a massive disjuncture between the ‘pro-Kibaki’ vote in the parliamentary and
presidential elections. When read correctly, the evidence of the parliamentary campaign
suggests that this was an incredibly tight election at every level, and the process of
coalition building played a key role in shaping both the campaign, and the outcomes.

The Aftermath

The roots of the Kenya crisis are too deep to be resolved by a political settlement between
Odinga and Kibaki. Successful coalition politics will not be sufficient to safeguard Kenya’s
fragile democracy, but it is a necessary step that must precede more ambitious reforms.
This is why the mediation talks brokered by Annan focused on identifying a form of
coalition government acceptable to both sides as the most viable path to end the violence.
While the PNU attempted to maintain its hold over power, the ODM demanded that its
strong showing in parliament be rewarded with the creation of a post of Prime Minister
for Raila Odinga, and a healthy proportion of cabinet seats.

The deal signed in early March masks considerable disagreement over issues such as
constitutional reform, the powers of the prime minister, devolution of power, and the
best time to hold fresh elections. Tackling these divisive questions will be all the more
difficult because of Odinga’s unhappy experience of coalition government last time
around, the antagonism and low levels of trust between the main participants, and the
limitations of party structures and formal political institutions. Coalition building is far
harder in a context of low levels of interpersonal trust where formal institutions such as
the legislature and the judiciary cannot be relied upon to enforce the ‘rules of the game’.
Relations between rival political leaders appear to be at an all time low, and public
pressure for negotiation may be compromised by the fact that trust between different
ethnic communities in Kenya has evaporated.75 Moreover, the failure of the ECK has
reminded the Kenyan opposition of the need to prevent government interference with key
democratic institutions. Significantly, some of the key demands made by ODM leaders
during the negotiations were designed to limit the power of the executive and to protect
the opposition in the event of the collapse of the coalition.

The problems of coalition building are here to stay. The last ten years of Kenyan
political history have been characterized by cycles of elite fragmentation and concentration.
Ultimately, the instability of political alliances has translated into a weak party
system. Kenya needs a stable government, but constructing durable political alliances
takes real skill and resources. Kibaki failed to manage this process in the build up to the
election campaign, and unless the Kenyan elite can remember what it is they have in
common, there is no guarantee that his latest attempt will fare any better. But, this should
not be taken as evidence that democracy is an inherently destabilizing system of political
governance in Africa.76 Kenya’s crisis was not inevitable and might have been avoided.
Recent events in Kenya are an important reminder that if elections are to have democratic
outcomes, they must be embedded in a web of supportive institutions. Democratic
elections do not sit well with an authoritarian constitution, and coalition building and
elite compromise are undermined by the dominant nature of the presidency and the lack
of institutions that operate independently of the executive. The potential for conflict in
Kenya would have been significantly diffused if the government had not deliberately
manipulated the process of constitutional reform to prevent the decentralization of power

The Kenyan Elections of 2007: An Introduction 181

away from the executive. The crisis might also have been more quickly resolved if
opposition leaders had been able to rely on key democratic institutions to function in the
forthcoming parliament. With the benefit of hindsight, the lack of international criticism
of Kibaki’s failure to dismantle the authoritarian structure inherited from the one-party
state is striking.

There are many other countries in which elections are routinely being held while
comprehensive constitutional reforms are routinely delayed, most obviously Zambia and
Uganda. As in Kenya, the absence of reform in these cases threatens to undermine public
confidence in the electoral process, and the willingness of opposition leaders to place their
trust in the legislature and the judiciary; consequently, it reduces the likelihood of
peaceful elections. The Kenya crisis is a timely reminder to both international actors and
African governments that the continent’s new multi-party states may be extremely fragile,
no matter how calm they appear on the surface, and that in the tricky process of
democratization, sequencing matters.


The author would like to thank David Anderson, Adrienne Lebas, James Long, John Lonsdale, and Justin Willis
for their helpful comments on this paper. The arguments presented here were developed in conversations with
Daniel Branch and many friends in Nairobi who made fieldwork during the election possible and bearable.



The use of the term ‘genocide’ in early reports was the best example of this. For one example see Cape Times,


See their paper in this volume.


Flint, ‘State-Building’; UNDP, ‘Horizontal Inequalities’.


Reno, ‘The Politics’.


Anderson, ‘Kenya on the Brink’.


Weis, Available at


Willis, this volume.


Kavulla, this volume.


Mwangola, ‘Leaders of Tomorrow’.


International Crisis Group, ‘Kenya in Crisis’, 4.


Daily Nation, ‘Raila is the Youth’s Favourite’.


Wolf, forthcoming.


Sunday Nation, ‘Raila’.


Larmer and Fraser, ‘Of Cabbages’.


Wolf, forthcoming.


At a meeting of the Kibaki Tena Fundraising Campaign at the PanAfric Hotel on 8 December 2008, the
organizers focused on an ‘adopt a polling scheme’ strategy.


This point was reiterated to a number of election observers at polling stations in different provinces. At the
Olympic polling station voters were particularly distressed when their names did not appear on the electoral
register because without an ‘inky finger’ they would not be let on the bus to get home.


Osborn, this volume.


The Standard, ‘Ali’.


The Standard, ‘Raila’s Secret MoU’.


These documents and many more were released via ‘wikileaks’, see


See, for example, Voice of America, ‘Kenya Unrest’.


Anderson and Lochery, this volume.


This is true of the otherwise excellent report by the International Crisis Group, ‘Kenya in Crisis’.

182 N. Cheeseman


European Union EOM, ‘Preliminary Statement’, 7.


KNHRC, ‘Preliminary Report’.


The Standard, ‘It’s Bloodbath’.


KNHRC, ‘Still Behaving Badly’; Sunday Nation, ‘Kuresoi’.


Daily Nation, ‘ECK Summons’.


These are set out in the KNHCR, ‘Preliminary Report’. They have been contested by the government.


For an example, see Daily Telegraph, ‘15 Killed’.


Lonsdale, this volume.


Mueller, this volume.


Reno, ‘The Politics’, 841.


San Francisco Bay View, ‘Sham Election’. The US government has since claimed that it called to offer its
congratulations to the Kenyan people, and not to Kibaki specifically.


IRI, ‘IRI Statement’.


McLatchy, ‘Kenyan President’.


Nairobi Star, ‘IRI’.


See Brown, ‘From Demiurge to Midwife’.


Bratton and Kimenyi, this volume.


For more on this point see Cheeseman, ‘Kenya Since 2002’; Branch and Cheeseman, ‘The Politics of Control’.


See Anderson, ‘Kenya’s Elections’.


Cheeseman, ‘Kenya Since 2002’.


Gona, this volume; MacArthur, this volume.


Daily Nation, ‘PNU’.


European Union EOM, ‘Preliminary Statement’.


Sunday Nation, ‘Elections’.


Sunday Nation, ‘No Ebb’.


Sunday Standard, ‘Western Vote’.


Sunday Nation, ‘No Ebb’.


Sunday Standard, ‘Western Vote’.




Daily Nation, ‘Election Funds’.


Sunday Standard, ‘Stand-off’.


Daily Nation, ‘ODM’.


KNHCR, ‘Still Behaving Badly’.


Ghai, this volume.


Flyers depicting Odinga as Idi Amin were handed out at the final PNU rally at Uhuru Park by PNU


Significantly, the sample size was large enough to allow us to look at the provincial breakdown of presidential
support with some confidence.


Daily Nation, ‘With Eight Days’.


Throup, this volume.


KNHRC, ‘General Elections 2007’; KPTJ Countdown to Deception; KNHCR, ‘Behaving Badly’; KPTJ, ‘Kenyan
Election’; UNDP, ‘Election 2007’ and ‘Interim Report’.


European Union EOM, ‘Preliminary Statement’.


Daily Nation, ‘Kivuitu Calls for Probe’.


Throup and Hornsby, Multi-Party Politics.


Observing 16 polling stations during election day in a variety of Nairobi polling stations, I, nor any of the
observers I came into contact with, saw anyone fail to vote in all three elections (civic/parliamentary/


Calculated as (total presidential votes . total parliamentary votes)/parliamentary votes.


Weis, Available at


Nairobi Star, ‘IRI’; Slate, ‘What’s Really Going On’.


Dercon et al., ‘Violence’.


It is worth noting that there is currently no concrete evidence to support this hypothesis.


African Press International, ‘22 Kenyan MPs’.


Daily Nation, ‘Losers in Primaries’.

The Kenyan Elections of 2007: An Introduction 183


The party groupings used to calculate this are: ODMODM, NaRC, UDM; PNUPNU, Safina, KANU,
NaRC-KENYA, Sisi Kwa Sisi, DP, KENDA, FORD-A, FORD-K, FORD-P, NEW FORD-K, Green Party. A fully
united ODM combines the ODM group and ODM-K.


Bratton and Kimenyi, this volume.


Githongo, this volume.


Election Reports

European Union Election Observation Mission (EOM). ‘‘Preliminary Statement: Doubts about the Credibility of
the Presidential Results Hamper Kenya’s Democratic Progress.’’ 1 January 2008. Available from http://; INTERNET.

International Crisis Group. ‘‘Kenya in Crisis.’’ 21 February 2008. Available from
home/index.cfm?id 5316; INTERNET.

Kenya Elections Domestic Observation Forum. ‘‘Preliminary Press Statement and Verdict of 2007 Kenya’s
General Elections.’’ 31 December 2008. Available from
31-12-07.pdf; INTERNET.

Kenya National Human Rights Commission (KNHRC). ‘‘Behaving Badly: Deception, Chauvinism and
Waste during the Referendum Campaigns.’’ September 2006. Available from
documents/referendumreport.pdf; INTERNET.

* . ‘‘Preliminary Report on Alleged Executions and Disappearance of Persons Between June and October.’’ 4
November 2007. Available from; INTERNET.
* . ‘‘Still Behaving Badly: Second Periodic Report of the Election Monitoring Project.’’ December 2007.
Available from; INTERNET.
Kenyans for Peace with Truth and Justice (KPTJ). ‘‘Count Down to Deception: 30 Hours that Destroyed Kenya.’’
18 January 2007. Available from; INTERNET.

* . ‘‘Kenyan Election Observers’ Log: December 2930, 2007.’’ January 2007. Available from http://; INTERNET.
Steadman Group. ‘‘Political Campaigns Media Coverage and Advertising Expenditure.’’ 7 December 2007.
Available from; INTERNET.
United Nations Development Programme (UNDP). CAPF Election Monitoring Initiative. ‘‘Election 2007:
Interim Report on National Voter Bribery Survey.’’ Available from www.; INTERNET.

* . ‘‘Horizontal Inequalities and Violent Conflict. Coˆ te d’Ivoire Country Paper.’’ Human Development Report
Office Occasional Paper 2005.32. Available from
hdr2005_langer_arnim_32.pdf; INTERNET.
* . ‘‘Interim Report on Presidential Campaign Expenditure.’’ 17 December 2007. Available from www.; INTERNET.

African Press International. ‘‘22 Kenyan Members of Parliament to Face Court Battles.’’ 27 January 2007.
Cape Times. ‘‘Africa Faces Genocide in Kenya.’’ 1 February 2008.
The Daily Nation. ‘‘ECK Summons Eight.’’ 21 December 2007.

* . ‘‘Election Funds.’’ 4 December 2008.
* . ‘‘Kivuitu Calls for Probe into Tallying of Presidential Votes.’’ 3 January 2008.
* . ‘‘Losers in Primaries Work Overdrive.’’ 19 November 2008.
* . ‘‘ODM Rules out 3-piece Voting Model.’’ 7 December 2007.
* . ‘‘PNU Strikes Deal on Nomination Method.’’ 11 October 2007.
* . ‘‘Raila is the Youth’s Favourite Candidate.’’ 8 October 2007.
* . ‘‘With Eight Days to Go, Race Too Close to Call.’’ 19 December 2007.
Daily Telegraph. ‘‘15 Killed in Kenya Violence.’’ 4 March 2008.
International Republican Institute (IRI). ‘‘IRI Statement on Kenya Exit Poll.’’ 7 February 2008. Available from; INTERNET.
McClatchy. ‘‘Kenyan President Lost Election.’’ 14 January 2008.Nairobi Star. ‘‘IRI.’’ 14 February 2008.
San Francisco Bay View. ‘‘Sham Election in Kenya.’’ 9 January 2008.

184 N. Cheeseman

Slate. ‘‘What’s Really Going On in Kenya?’’ 2 January 2008.
The Standard. ‘‘Ali Denies Resignation Claims.’’ 1 January 2008.

* . ‘‘It’s Bloodbath as Police Strike Back at Mungiki.’’ 6 June 2007.
* . ‘‘Raila’s Secret Moll with Muslims.’’ 28 November 2008.
Sunday Nation. ‘‘Elections: Why I Maintain Faith in my Country.’’ 9 December 2007.
* . ‘‘Kuresoi Cries for Peace.’’ 9 December 2007.
* . ‘‘No Ebb in PNU Parties’ Rivalry.’’ 9 December 2007.
* . ‘‘Raila in 10-point Lead.’’ 15 December 2007.
Sunday Standard. ‘‘Stand-off among PNU Affiliates Far from Over.’’ 9 December 2007.
* . ‘‘Western Vote Proves Elusive as PNU Tussle.’’ 9 December 2007.
Voice of America. ‘‘Kenya Unrest.’’ 4 February 2008.

Anderson, D. ‘‘Kenya on the Brink.’’ Prospect Magazine, January, 142 (2008). Available from http:// 9983; INTERNET.

* . ‘‘Kenya’s Elections 2002 . The Dawning of a New Era?’’ African Affairs 102, no. 407 (2003): 33142.
Branch, D., and N. Cheeseman. ‘‘The Politics of Control in Kenya: Understanding the Bureaucratic-Executive
State.’’ Review of African Political Economy 33, no. 107 (2006): 1131.
Brown, S. ‘‘From Demiurge to Midwife: Changing Donor Roles in Kenya’s Democratisation Process.’’ In Kenya:
The Struggle for Democracy, edited by Godwin Murunga and Shadrack Nasong’o. London: Zed Books, 2007.
Cheeseman, N. ‘‘Kenya Since 2002: The More Things Change the More they Stay the Same.’’ In Africa Since 2002,
edited by Lindsay Whitfield and Raufu Mustapha. London: James Currey, 2008.
Dercon, S., M. Bratton, M. Kimeryi, R. Gutierrez-Romero and T. Bold. ‘‘Ethnicity, Violence and the 2007
Elections in Kenya.’’ Unpublished Conference Paper, Oxford University, 17 January 2008.
Flint, L. ‘‘State-Building in Central Southern Africa: Citizenship and Subjectivity in Barotseland and Caprivi.’’
International Journal of African Historical Studies 36, no. 2 (2003): 393428.
Larmer, M., and A. Fraser. ‘‘Of Cabbages and King Cobra: Populist Politics and Zambia’s 2006 Election.’’ African
Affairs 425, no. 106 (2007): 61137.
Mwangola, M. ‘‘Leaders of Tomorrow.’’ In Kenya: The Struggle for Democracy, edited by Godwin Murunga and
Shadrack Nasong’o. London: Zed Books, 2007.
Reno, W. ‘‘The Politics of Insurgency in Collapsing States.’’ Development and Change 33, no. 5 (2002): 83758.
Throup, D., and C. Hornsby. Multi-Party Politics in Kenya: The Kenyatta and Moi States and the Triumph of the

System in the 1992 Election. Athens: Ohio University Press, 1998.
Wolf, Tom. ‘‘The Public, Politicians and Polling in the 2007 Kenya Election.’’ Journal of Eastern African Studies,